Cross Reference: /src/sys/fs/puffs

History log of /src/sys/fs/puffs
Revision	Date	Author	Comments
1.1	22-Oct-2006	pooka	branches: 1.1.2; 1.1.6; 1.1.8; kernel portion of puffs - the Pass-to-Userspace Framework File System. It contains the VFS attachment and userspace message-passing interface. This work was initially started and completed for Google SoC 2005 and tweaked to work a bit better in the past few weeks. While being far from complete, it is functional enough to be able and stable to host a fairly general-purpose in-memory file system in userspace. Even so, puffs should be considered experimental and no binary compatibility for interfaces or crash-freedom or zero security implications should be relied upon just yet. The GSoC project was mentored by William Studenmund and the final review for the code was done by Christos.
1.1.8.2	30-Dec-2006	yamt	sync with head.
1.1.8.1	22-Oct-2006	yamt	file Makefile was added on branch yamt-lazymbuf on 2006-12-30 20:50:00 +0000
1.1.6.2	10-Dec-2006	yamt	sync with head.
1.1.6.1	22-Oct-2006	yamt	file Makefile was added on branch yamt-splraiseipl on 2006-12-10 07:18:38 +0000
1.1.2.2	18-Nov-2006	ad	Sync with head.
1.1.2.1	22-Oct-2006	ad	file Makefile was added on branch newlock2 on 2006-11-18 21:39:20 +0000
1.1	18-May-2009	pooka	branches: 1.1.2; add some todo-items, based on a file which was lingering in my local tree for apparently almost two years now
1.1.2.2	20-Jun-2009	yamt	sync with head
1.1.2.1	18-May-2009	yamt	file TODO was added on branch yamt-nfs-mp on 2009-06-20 07:20:30 +0000
1.10	05-Feb-2019	pgoyette	It turns out we do want the puffs compat code in any kernel which has built-in compat_50 regardless of whether the kernel also has puffs. Should finally fix PR kern/53943
1.9	04-Feb-2019	wiz	try '&' instead of '&&'
1.8	04-Feb-2019	pgoyette	Don't include puffs_compat in a kernel unless the filesystem is selected along with COMPAT_50. Also, don't include puffs_compat in the main puffs filesystem module; it is part of the compat_50 module. Should address PR kern/53943
1.7	27-Jan-2019	pgoyette	Merge the [pgoyette-compat] branch
1.6	11-Oct-2014	uebayasi	branches: 1.6.18; 1.6.20; Define filesystem attributes with vfs dependency.
1.5	06-Jul-2010	pooka	branches: 1.5.18; remember to add the new file to the build
1.4	10-Nov-2007	pooka	branches: 1.4.18; 1.4.38; 1.4.40; Part 2/n of extensive changes to request transport to/from userspace: Rip the transport code completely out of puffs and generalize it into an independent module which will be used for multiple purposes in the future. This module is called the Pass-to-Userspace Transporter (known as "putter" among friends). This is very much work-in-progress and one dependency with puffs remains: the request framing format. The device name is still /dev/puffs, but that will change soon. Users of puffs need the following in their kernel configs now: pseudo-device putter
1.3	27-Sep-2007	pooka	branches: 1.3.2; 1.3.4; Split routines handling nodes from puffs_subr to puffs_node. No functional change.
1.2	05-Dec-2006	pooka	branches: 1.2.2; 1.2.4; 1.2.10; 1.2.22; 1.2.24; 1.2.26; shuffle functions around a bit: move the transport (/dev/puffs) to a different file from the messaging (request contents). no functional change
1.1	22-Oct-2006	pooka	branches: 1.1.2; kernel portion of puffs - the Pass-to-Userspace Framework File System. It contains the VFS attachment and userspace message-passing interface. This work was initially started and completed for Google SoC 2005 and tweaked to work a bit better in the past few weeks. While being far from complete, it is functional enough to be able and stable to host a fairly general-purpose in-memory file system in userspace. Even so, puffs should be considered experimental and no binary compatibility for interfaces or crash-freedom or zero security implications should be relied upon just yet. The GSoC project was mentored by William Studenmund and the final review for the code was done by Christos.
1.1.2.3	12-Jan-2007	ad	Sync with head.
1.1.2.2	18-Nov-2006	ad	Sync with head.
1.1.2.1	22-Oct-2006	ad	file files.puffs was added on branch newlock2 on 2006-11-18 21:39:20 +0000
1.2.26.1	06-Oct-2007	yamt	sync with head.
1.2.24.2	09-Jan-2008	matt	sync with HEAD
1.2.24.1	06-Nov-2007	matt	sync with HEAD
1.2.22.2	11-Nov-2007	joerg	Sync with HEAD.
1.2.22.1	02-Oct-2007	joerg	Sync with HEAD.
1.2.10.1	09-Oct-2007	ad	Sync with head.
1.2.4.4	15-Nov-2007	yamt	sync with head.
1.2.4.3	27-Oct-2007	yamt	sync with head.
1.2.4.2	30-Dec-2006	yamt	sync with head.
1.2.4.1	05-Dec-2006	yamt	file files.puffs was added on branch yamt-lazymbuf on 2006-12-30 20:50:00 +0000
1.2.2.2	10-Dec-2006	yamt	sync with head.
1.2.2.1	05-Dec-2006	yamt	file files.puffs was added on branch yamt-splraiseipl on 2006-12-10 07:18:38 +0000
1.3.4.1	19-Nov-2007	mjf	Sync with HEAD.
1.3.2.1	13-Nov-2007	bouyer	Sync with HEAD
1.4.40.1	05-Mar-2011	rmind	sync with head
1.4.38.1	17-Aug-2010	uebayasi	Sync with HEAD.
1.4.18.1	11-Aug-2010	yamt	sync with head.
1.5.18.1	03-Dec-2017	jdolecek	update from HEAD
1.6.20.1	10-Jun-2019	christos	Sync with HEAD
1.6.18.2	22-Sep-2018	pgoyette	Include the compat code whether or not the calling device or filesystem exists.
1.6.18.1	24-Mar-2018	pgoyette	Add fs/puffs compat_50 to the modules
1.8	12-Dec-2019	pgoyette	Rather than keeping a separate mutex, condvar, and pserialize for each module hook, we can share a common set of synchronization structures. This cuts the amount of cacheline_aligned data for these structures by 50%. Note that we still have a per-hook localcount, since we need to count individual references. As discussed with riastradh@ Welcome to 9.99.22 !
1.7	01-Mar-2019	pgoyette	Rename the MODULE__HOOK() macros to MODULE_HOOK_() as briefly discussed on irc. NFCI intended. Ride the earlier kernel bump - it;s getting crowded.
1.6	29-Jan-2019	pgoyette	Normalize all the compat hooks' names to the form <subsystem>_<function>_<version>_hook NFCI XXX Note that although this introduces a change in the kernel-to- XXX module interface, we are NOT bumping the kernel version number. XXX We will bump the version number once the interface stabilizes.
1.5	27-Jan-2019	pgoyette	Merge the [pgoyette-compat] branch
1.4	22-Apr-2015	pooka	branches: 1.4.16; 1.4.18; sprinkle COMPAT_50
1.3	10-Nov-2014	maxv	branches: 1.3.2; Do not uselessly include <sys/malloc.h>.
1.2	11-Jul-2010	pooka	branches: 1.2.2; 1.2.4; 1.2.10; 1.2.24; 1.2.40; Do fhtovp compat translation only for fhtovp ops, not all vfs ops. Allocate tailing extra buffer for compat op too.
1.1	06-Jul-2010	pooka	Add compat to enable running puffs in a 64bit time_t kernel against a server which runs in 32bit time_t namespace.
1.2.40.1	17-Jan-2015	martin	Pull up following revision(s) (requested by maxv in ticket #427): sys/compat/svr4/svr4_schedctl.c: revision 1.8 sys/netinet/tcp_timer.c: revision 1.88 sys/miscfs/genfs/layer_vfsops.c: revision 1.45 sys/compat/svr4/svr4_ioctl.c: revision 1.37 sys/ufs/chfs/chfs_vfsops.c: revision 1.14 sys/miscfs/fdesc/fdesc_vfsops.c: revision 1.91 sys/compat/linux/arch/i386/linux_ptrace.c: revision 1.30 sys/compat/common/kern_time_50.c: revision 1.28 sys/netinet6/ip6_forward.c: revision 1.74 sys/miscfs/umapfs/umap_vnops.c: revision 1.57 sys/compat/svr4/svr4_fcntl.c: revision 1.74 distrib/sets/lists/comp/mi: revision 1.1931 sys/netinet6/udp6_output.c: revision 1.46 sys/fs/puffs/puffs_compat.c: revision 1.3 sys/fs/udf/udf_rename.c: revision 1.11 sys/compat/svr4/svr4_filio.c: revision 1.24 sys/fs/udf/udf_rename.c: revision 1.12 sys/netinet/tcp_usrreq.c: revision 1.202 sys/miscfs/umapfs/umap_subr.c: revision 1.29 sys/compat/linux/common/linux_fadvise64.c: revision 1.3 sys/netinet/if_atm.c: revision 1.34 sys/miscfs/procfs/procfs_subr.c: revision 1.106 sys/miscfs/genfs/layer_subr.c: revision 1.37 sys/netinet/tcp_sack.c: revision 1.30 sys/compat/freebsd/freebsd_misc.c: revision 1.33 sys/compat/freebsd/freebsd_file.c: revision 1.33 sys/ufs/chfs/chfs_vnode.c: revision 1.12 sys/compat/svr4/svr4_ttold.c: revision 1.34 sys/compat/linux/common/linux_file.c: revision 1.114 sys/compat/linux/arch/mips/linux_machdep.c: revision 1.43 sys/compat/linux/common/linux_signal.c: revision 1.76 sys/compat/common/compat_util.c: revision 1.46 sys/compat/linux/arch/arm/linux_ptrace.c: revision 1.18 sys/compat/svr4/svr4_sockio.c: revision 1.36 sys/compat/linux/arch/arm/linux_machdep.c: revision 1.32 sys/compat/svr4/svr4_signal.c: revision 1.66 sys/kern/kern_exec.c: revision 1.410 sys/fs/puffs/puffs_vfsops.c: revision 1.115 sys/compat/svr4/svr4_exec_elf64.c: revision 1.15 sys/compat/linux/arch/i386/linux_machdep.c: revision 1.159 sys/compat/linux/arch/alpha/linux_machdep.c: revision 1.50 sys/compat/linux32/common/linux32_misc.c: revision 1.24 sys/netinet/in_pcb.c: revision 1.153 sys/sys/malloc.h: revision 1.116 sys/compat/common/if_43.c: revision 1.9 share/man/man9/Makefile: revision 1.380 sys/netinet/tcp_vtw.c: revision 1.12 sys/miscfs/umapfs/umap_vfsops.c: revision 1.95 sys/ufs/ext2fs/ext2fs_vfsops.c: revision 1.186 sys/compat/common/uipc_syscalls_43.c: revision 1.46 sys/ufs/ext2fs/ext2fs_vnops.c: revision 1.115 sys/fs/puffs/puffs_msgif.c: revision 1.97 sys/compat/svr4/svr4_ipc.c: revision 1.27 sys/compat/linux/common/linux_exec.c: revision 1.117 sys/ufs/ext2fs/ext2fs_readwrite.c: revision 1.66 sys/netinet/tcp_output.c: revision 1.179 sys/compat/svr4/svr4_termios.c: revision 1.28 sys/fs/udf/udf_strat_bootstrap.c: revision 1.4 sys/fs/puffs/puffs_subr.c: revision 1.67 sys/fs/puffs/puffs_node.c: revision 1.36 sys/miscfs/overlay/overlay_vnops.c: revision 1.21 sys/fs/cd9660/cd9660_node.c: revision 1.34 sys/netinet/raw_ip.c: revision 1.146 sys/sys/mallocvar.h: revision 1.13 sys/miscfs/overlay/overlay_vfsops.c: revision 1.63 share/man/man9/malloc.9: revision 1.50 sys/netinet6/dest6.c: revision 1.18 sys/compat/linux/common/linux_uselib.c: revision 1.33 sys/compat/linux/common/linux_socket.c: revision 1.120 share/man/man9/malloc.9: revision 1.51 sys/netinet/tcp_subr.c: revision 1.257 sys/compat/linux/common/linux_socketcall.c: revision 1.45 sys/compat/linux/common/linux_fadvise64_64.c: revision 1.3 sys/compat/freebsd/freebsd_ipc.c: revision 1.17 sys/compat/linux/common/linux_misc_notalpha.c: revision 1.109 sys/compat/linux/arch/alpha/linux_pipe.c: revision 1.17 sys/netinet6/in6_pcb.c: revision 1.132 sys/netinet6/in6_ifattach.c: revision 1.94 sys/compat/svr4/svr4_exec_elf32.c: revision 1.15 sys/miscfs/nullfs/null_vfsops.c: revision 1.90 sys/fs/cd9660/cd9660_util.c: revision 1.12 sys/compat/linux/arch/powerpc/linux_machdep.c: revision 1.48 sys/compat/freebsd/freebsd_exec_elf32.c: revision 1.20 sys/miscfs/procfs/procfs_vfsops.c: revision 1.94 sys/compat/linux/arch/powerpc/linux_ptrace.c: revision 1.28 sys/compat/linux/common/linux_sched.c: revision 1.67 sys/compat/linux/common/linux_exec_aout.c: revision 1.67 sys/compat/linux/common/linux_pipe.c: revision 1.67 sys/compat/linux/common/linux_llseek.c: revision 1.34 sys/compat/linux/arch/mips/linux_ptrace.c: revision 1.10 Do not uselessly include <sys/malloc.h>. Cleanup: - remove struct kmembuckets (dead) - correctly deadify MALLOC_XX - remove MALLOC_DEFINE_LIMIT and MALLOC_JUSTDEFINE_LIMIT (dead) - remove malloc_roundup(), malloc_type_setlimit(), MALLOC_DEFINE_LIMIT() and MALLOC_JUSTDEFINE_LIMIT() from man 9 malloc New sentence, new line. Bump date for previous. Obsolete malloc_roundup(9), malloc_type_setlimit(9) and MALLOC_DEFINE_LIMIT(9) man pages.
1.2.24.1	03-Dec-2017	jdolecek	update from HEAD
1.2.10.2	05-Mar-2011	rmind	sync with head
1.2.10.1	11-Jul-2010	rmind	file puffs_compat.c was added on branch rmind-uvmplock on 2011-03-05 20:55:07 +0000
1.2.4.2	17-Aug-2010	uebayasi	Sync with HEAD.
1.2.4.1	11-Jul-2010	uebayasi	file puffs_compat.c was added on branch uebayasi-xip on 2010-08-17 06:47:19 +0000
1.2.2.2	11-Aug-2010	yamt	sync with head.
1.2.2.1	11-Jul-2010	yamt	file puffs_compat.c was added on branch yamt-nfs-mp on 2010-08-11 22:54:34 +0000
1.3.2.1	06-Jun-2015	skrll	Sync with HEAD
1.4.18.2	08-Apr-2020	martin	Merge changes from current as of 20200406
1.4.18.1	10-Jun-2019	christos	Sync with HEAD
1.4.16.6	23-Jan-2019	pgoyette	Convert the macros for setting and unsetting a hook to generate in-line code rather than using an intermediary hook##set routine. Hooks are set and unset only in one place, so the intermediary routine provides no benefit. IMHO using the macro at the point- of-call is more readable than using it elsewhere in the code and then calling the generated intermediary routine (for which you won't even find its declaration or definition unless you remember to search for the HOOK_SET macro instead). NFC intended, will verify with a bulk build and an atf test run.
1.4.16.5	14-Jan-2019	pgoyette	Create a variant of the HOOK macros that handles hook routines of type void, and use them where appropriate.
1.4.16.4	13-Jan-2019	pgoyette	Remove the HOOK2 versions of the MODULE_HOOK macros. There were only a few uses, and using them led to some lack of clarity in the code. Instead, we now use two separate hooks, with names that make it clear(er) what we're doing. This also positions us to start unraveling some of the rtsock_50 mess, which will need (at least) five hooks.
1.4.16.3	18-Sep-2018	pgoyette	The COMPAT_HOOK macros were renamed to MODULE_HOOK, adjust all callers
1.4.16.2	17-Sep-2018	pgoyette	Adapt (most of) the indirect function pointers to the new MP-safe mechanism. Still remaining are the compat_netbsd32 stuff, and some usb subroutines.
1.4.16.1	24-Mar-2018	pgoyette	Add fs/puffs compat_50 to the modules
1.108	01-Feb-2025	andvar	s/furher/further/ in comment.
1.107	09-Feb-2024	andvar	branches: 1.107.2; fix spelling mistakes, mainly in comments and log messages.
1.106	15-May-2020	maxv	hardclock_ticks -> getticks()
1.105	23-Feb-2020	ad	UVM locking changes, proposed on tech-kern: - Change the lock on uvm_object, vm_amap and vm_anon to be a RW lock. - Break v_interlock and vmobjlock apart. v_interlock remains a mutex. - Do partial PV list locking in the x86 pmap. Others to follow later.
1.104	01-Mar-2019	pgoyette	branches: 1.104.6; Rename the MODULE__HOOK() macros to MODULE_HOOK_() as briefly discussed on irc. NFCI intended. Ride the earlier kernel bump - it;s getting crowded.
1.103	29-Jan-2019	pgoyette	Normalize all the compat hooks' names to the form <subsystem>_<function>_<version>_hook NFCI XXX Note that although this introduces a change in the kernel-to- XXX module interface, we are NOT bumping the kernel version number. XXX We will bump the version number once the interface stabilizes.
1.102	27-Jan-2019	pgoyette	Merge the [pgoyette-compat] branch
1.101	17-Apr-2017	hannken	branches: 1.101.10; 1.101.12; Add vfs_ref(mp) and vfs_rele(mp) to add or remove a reference to struct mount. Rename vfs_destroy(mp) to vfs_rele(mp) and replace incrementing mp->mnt_refcnt with vfs_ref(mp).
1.100	26-Dec-2016	skrll	branches: 1.100.2; Hold the interlock when calling cv_broadcast as per condvar(9)
1.99	07-Jul-2016	msaitoh	branches: 1.99.2; KNF. Remove extra spaces. No functional change.
1.98	06-May-2015	hannken	Remove miscfs/syncfs and - move the syncer into kern/vfs_subr.c. - change the syncer to process the mountlist and VFS_SYNC as appropriate. - use an API for mount points similiar to the API for vnodes: - vfs_syncer_add_to_worklist(struct mount mp) to add - vfs_syncer_remove_from_worklist(struct mount mp) to remove a mount. No objections on tech-kern@
1.97	10-Nov-2014	maxv	branches: 1.97.2; Do not uselessly include <sys/malloc.h>.
1.96	05-Sep-2014	matt	Don't use C++ class and this keywords as variables.
1.95	28-Aug-2014	hannken	Change puffs from hashlist to vcache. - field "pa_nhashbuckets" of struct "puffs_kargs" becomes a no-op. and should be removed on the next protocol version bump.
1.94	17-Oct-2013	christos	branches: 1.94.4; - remove unused variables - add _NOERROR flavor macros for the case where errors are ignored.
1.93	05-Nov-2012	dholland	branches: 1.93.2; Excise struct componentname from the namecache. This uglifies the interface, because several operations need to be passed the namei flags and cache_lookup also needs for the time being to be passed cnp->cn_nameiop. Nonetheless, it's a net benefit. The glop should be able to go away eventually but requires structural cleanup elsewhere first. This change requires a kernel bump.
1.92	27-Jul-2012	manu	branches: 1.92.2; Rename slow sopreq queue into node sopreq queue, to refet the fact that is only intended for postponed node reclaims. When purging the node sopreq queue, do not call puffs_msg_sendresp(), as it makes no sense.
1.91	22-Jul-2012	manu	Fix hang unmount bug introduced by last commit. We introduced a slow queue for delayed reclaims, while the existing queue for unmount, flush and exist has been renamed fast queue. Both queues had timestamp for when an operation should be done, but it was useless for the fast queue, which is always used to run an operation ASAP. And the timestamp test had an error that turned ASAP into "at next tick", but nobody what there to wake the thread at next tick, hence the hang. The fix is to remove the useless and buggy timestamp test for fast queue.
1.90	21-Jul-2012	manu	- Improve PUFFS_KFLAG_CACHE_FS_TTL by reclaiming older inactive nodes. The normal kernel behavior is to retain inactive nodes in the freelist until it runs out of vnodes. This has some merit for local filesystems, where the cost of an allocation is about the same as the cost of a lookup. But that situation is not true for distributed filesystems. On the other hand, keeping inactive nodes for a long time hold memory in the file server process, and when the kernel runs out of vnodes, it produce reclaim avalanches that increase lattency for other operations. We do not reclaim inactive vnodes immediatly either, as they may be looked up again shortly. Instead we introduce a grace time and we reclaim nodes that have been inactive beyond the grace time. - Fix lookup/reclaim race condition. The above improvement undercovered a race condition between lookup and reclaim. If we reclaimed a vnode associated with a userland cookie while a lookup returning that same cookiewas inprogress, then the kernel ends up with a vnode associated with a cookie that has been reclaimed in userland. Next operation on the cookie will crash (or at least confuse) the filesystem. We fix this by introducing a lookup count in kernel and userland. On reclaim, the kernel sends the count, which enable userland to detect situation where it initiated a lookup that is not completed in kernel. In such a situation, the reclaim must be ignored, as the node is about to be looked up again.
1.89	19-Oct-2011	manu	branches: 1.89.2; 1.89.8; Remove #ifdef DIAGNOSTIC guards around KASSERT, as the macro contains them
1.88	18-Oct-2011	manu	Make sure pagedaemon does not sleep for memory in puffs_vnop_sleep. Add KASSERT on any sleeping memory allocation to check it cannot happen again.
1.87	03-Jul-2011	mrg	avoid some uninitialised variable warnings from GCC. at least the puffs one seems valid, but i'm not 100% sure.
1.86	12-Jun-2011	rmind	Welcome to 5.99.53! Merge rmind-uvmplock branch: - Reorganize locking in UVM and provide extra serialisation for pmap(9). New lock order: [vmpage-owner-lock] -> pmap-lock. - Simplify locking in some pmap(9) modules by removing P->V locking. - Use lock object on vmobjlock (and thus vnode_t::v_interlock) to share the locks amongst UVM objects where necessary (tmpfs, layerfs, unionfs). - Rewrite and optimise x86 TLB shootdown code, make it simpler and cleaner. Add TLBSTATS option for x86 to collect statistics about TLB shootdowns. - Unify /dev/mem et al in MI code and provide required locking (removes kernel-lock on some ports). Also, avoid cache-aliasing issues. Thanks to Andrew Doran and Joerg Sonnenberger, as their initial patches formed the core changes of this branch.
1.85	11-Feb-2011	yamt	branches: 1.85.2; puffs_msg_wait: check PARKFLAG_HASERROR before PARKFLAG_CALL. PR/44240.
1.84	15-Nov-2010	pooka	branches: 1.84.2; 1.84.4; Apply patch from PR kern/44093 by yamt: Interrupt server wait only on certain signals (same set at nfs -i) instead of all signals. According to the PR this helps with "git clone" run on a puffs file system.
1.83	12-Nov-2010	pooka	Allow clients to reuse a "park". Patch from <yamt>, fixes PR kern/44086 by him.
1.82	06-Jul-2010	pooka	Remove groolingly spooky variable which has been haunting us for several years without doing anything useful.
1.81	06-Jul-2010	pooka	Add compat to enable running puffs in a 64bit time_t kernel against a server which runs in 32bit time_t namespace.
1.80	14-Jan-2010	pooka	branches: 1.80.2; 1.80.4; In case the operations thread has exited, do not queue any more operations. This prevents kernel memory leaks (one of which happened every time the file system was unmounted via PUFFSOP_UNMOUNT ... and incidentally would've been trivially caught with the old malloc(9) interface. I wonder if the message is to use a ton of pools instead of regression-attractive kmem interface).
1.79	07-Jan-2010	pooka	Rename PUFFS_SOPREQ_EXIT to PUFFS_SOPREQSYS_EXIT to better signal it comes from within the kernel instead of as a direct result of a user request. no functional change
1.78	07-Jan-2010	pooka	Fix variable name in my commit tree too.
1.77	07-Jan-2010	pooka	Add a PUFFS_UNMOUNT server->kernel request, which causes the kernel to initiate self destruct, i.e. unmount(MNT_FORCE). This, however, is a semi-controlled self-destruct, since all caches are flushed before the (possibly) violent unmount takes place.
1.76	07-Dec-2009	pooka	Process flush requests from the file server in a separate thread context. This fixes a long-standing but seldomly seen deadlock, where the kernel was holding pages busy (due to e.g. readahead request) while waiting for the server to respond, and the server made a callback into the kernel asking to invalidate those pages. ... or, well, theoretically fixes, since I didn't have any reliable way of repeating the deadlock and I think I saw it only twice.
1.75	07-Dec-2009	pooka	Need to send protocol layer response instead of transport layer return value. While there, just collapse all non-supported types into one entry.
1.74	05-Nov-2009	pooka	Kill suspend support. It was never implemented correctly: * it depended on the biglock (in a very cruel way) * it was attached to userspace transactions rather than logical fs operations (If someone wants to revisit it some day, most of the stuff can be reused from cvs history)
1.73	18-Mar-2009	cegger	Ansify function definitions w/o arguments. Generated with sed.
1.72	25-Sep-2008	ad	branches: 1.72.2; 1.72.4; 1.72.8; 1.72.12; PR kern/39307 (mfs will sometimes panic at umount time) Change dounmount() so that it never drops the caller provided reference. Garbage collecting 'struct mount' is up to the caller.
1.71	06-May-2008	ad	branches: 1.71.2; 1.71.6; PR kern/38141 lookup/vfs_busy acquire rwlock recursively Simplify the mount locking. Remove all the crud to deal with recursion on the mount lock, and crud to deal with unmount as another weirdo lock. Hopefully this will once and for all fix the deadlocks with this. With this commit there are two locks on each mount: - krwlock_t mnt_unmounting. This is used to prevent unmount across critical sections like getnewvnode(). It's only ever read locked with rw_tryenter(), and is only ever write locked in dounmount(). A write hold can't be taken on this lock if the current LWP could hold a vnode lock. - kmutex_t mnt_updating. This is taken by threads updating the mount, for example when going r/o -> r/w, and is only present to serialize updates. In order to take this lock, a read hold must first be taken on mnt_unmounting, and the two need to be held across the operation. One effect of this change: previously if an unmount failed, we would make a half hearted attempt to back out of it gracefully, but that was unlikely to work in a lot of cases. Now while an unmount that will be aborted is in progress, new file operations within the mount will fail instead of being delayed. That is unlikely to be a problem though, because if the admin requests unmount of a file system then s(he) has made a decision to deny access to the resource.
1.70	30-Apr-2008	ad	PR kern/38135 vfs_busy/vfs_trybusy confusion The previous fix worked, but it opened a window where mounts could have disappeared from mountlist while the caller was traversing it using vfs_trybusy(). Fix that.
1.69	29-Apr-2008	ad	kern/38135 vfs_busy/vfs_trybusy confusion The symptom was that sometimes file systems would occasionally not appear in output from 'df' or 'mount' if the system was busy. Resolution: - Make mount locks work somewhat like vm_map locks. - vfs_trybusy() now only fails if the mount is gone, or if someone is unmounting the file system. Simple contention on mnt_lock doesn't cause it to fail. - vfs_busy() will wait even if the file system is being unmounted.
1.68	31-Jan-2008	tnn	branches: 1.68.6; 1.68.8; 1.68.10; - Needs sys/atomic.h for atomic_inc_uint() - Quench compiler warning about signed/unsigned mismatch when building LKM
1.67	30-Jan-2008	ad	Expunge references to lockmgr.
1.66	30-Jan-2008	ad	Make it compile. I'll leave it to pooka to figure out what is the correct thing here because I don't understand what this code is doing.
1.65	30-Jan-2008	ad	PR kern/37706 (forced unmount of file systems is unsafe): - Do reference counting for 'struct mount'. Each vnode associated with a mount takes a reference, and in turn the mount takes a reference to the vfsops. - Now that mounts are reference counted, replace the overcomplicated mount locking inherited from 4.4BSD with a recursable rwlock.
1.64	28-Jan-2008	pooka	For code clarity typedef void *puffs_cookie_t. No functional change.
1.63	02-Jan-2008	pooka	silence gcc about break type-punning
1.62	02-Jan-2008	ad	Merge vmlocking2 to head.
1.61	05-Dec-2007	pooka	branches: 1.61.4; Send a response message for flush operations from the kernel instead of abusing the return value of write(2).
1.60	26-Nov-2007	pooka	branches: 1.60.2; In case the userspace wait is interrupted, don't use ERESTART as the return value, rather use EINTR. reported by Reinoud
1.59	20-Nov-2007	pooka	Retire M_PUFFS, use kmem(9) instead.
1.58	17-Nov-2007	pooka	fix some debug prints
1.57	16-Nov-2007	pooka	Restructure the messaging interface a bit more: make all interfacing with the file server happen through puffs_msg_enqueue() and puffs_msg_wait() instead of having a billion different routines. Build the existing system upon these two. Most importantly though, decouple insertation into the op queue from the actual wait. This is useful for a number of reasons coming soon to a cvs repo near you.
1.56	12-Nov-2007	pooka	Bounds-check responses from userspace.
1.55	12-Nov-2007	pooka	* split the putter header into a kernel version and a userland version + install latter to /usr/include/dev/putter * remove last dependencies to puffs from putter, it's completely independent now
1.54	12-Nov-2007	pooka	Move putter code from directly under dev/ to dev/putter/ no functional change
1.53	10-Nov-2007	pooka	Part 2/n of extensive changes to request transport to/from userspace: Rip the transport code completely out of puffs and generalize it into an independent module which will be used for multiple purposes in the future. This module is called the Pass-to-Userspace Transporter (known as "putter" among friends). This is very much work-in-progress and one dependency with puffs remains: the request framing format. The device name is still /dev/puffs, but that will change soon. Users of puffs need the following in their kernel configs now: pseudo-device putter
1.52	07-Nov-2007	ad	Merge from vmlocking: - pool_cache changes. - Debugger/procfs locking fixes. - Other minor changes.
1.51	04-Nov-2007	pooka	branches: 1.51.2; Make some comments match current reality. No functional change.
1.50	25-Oct-2007	pooka	Reference mountpoint when fetching operations and release waiters in unmount.
1.49	21-Oct-2007	pooka	Always provide caller information from the kernel based on curlwp. (but don't deprecate the old puffs_cid interface just yet)
1.48	19-Oct-2007	pooka	When doing a read operation, don't copy the whole kernel buffer to userspace, since it doesn't contain any information yet. I should still rework this more so this is just a quickie to get the read/write style interface more up to speed with the ioctl version.
1.47	11-Oct-2007	pooka	branches: 1.47.2; Handle suspend and flush requests from the file server.
1.46	11-Oct-2007	pooka	Part 1/n of some pretty extensive changes to how the kernel module interacts with the userspace file server: * since the kernel-user communication is not purely request-response anymore (hasn't been since 2006), try to rename some "request" to "message". more similar mangling will take place in the future. * completely rework how messages are allocated. previously most of them were borrowed from the stack (originally all of them), but now always allocate dynamically. this makes the structure of the code much cleaner. also makes it possible to fix a locking order violation. it enables plenty of future enhancements. * start generalizing the transport interface to be independent of puffs * move transport interface to read/write instead of ioctl. the old one had legacy design problems, and besides, ioctl's suck. implement a very generic version for now; this will be worked on later hopefully some day reaching "highly optimized". * implement libpuffs support behind existing library request interfaces. this will change eventually (I hate those interfaces)
1.45	09-Oct-2007	pooka	g/c vntouser_req(), it's not used anymore
1.44	04-Oct-2007	pooka	g/c the "sizeop" code previous used for ioctl/fcntl. It was already commented out and has bitrotted beyond all recognition, so it needs complete rethinking.
1.43	02-Oct-2007	pooka	If kernel resource allocation fails after the file server has committed something, issue an abort. The abort is done through the regular op channel, e.g. failed mkdir leads to regular rmdir, inactive and reclaim. No internal interface is planned currently for the one file system out of a million which would implement it to benefit from the one case in a billion where kernel resource allocation actually does fail and out of that one case in a trillion where internal vs. external would make a difference.
1.42	01-Oct-2007	pooka	* better error checking: validate error values received from userland to be vaild errno values * include string describing error in PUFFS_ERR * get rid of union in puffs_req, it's nothing but trouble * pass pmp to async i/o callbacks
1.41	27-Sep-2007	pooka	Add error notifications, which are used to deliver errors from the kernel to the file server for silly things the file server did, e.g. attempting to create a file with size VSIZENOTSET. The file server can handle these as it chooses, but the default action is for it to throw its hands in the air and sing "goodbye, cruel world, it's over, walk on by".
1.40	19-Jul-2007	pooka	branches: 1.40.4; 1.40.6; 1.40.8; 1.40.10; add debug printf
1.39	09-Jul-2007	ad	branches: 1.39.2; Merge some of the less invasive changes from the vmlocking branch: - kthread, callout, devsw API changes - select()/poll() improvements - miscellaneous MT safety improvements
1.38	06-Jun-2007	pooka	Move puffs to a two clause license where it already isn't so. And as agc pointed out, even files with the third clause were already effectively two clause because of a slight bug in the language...
1.37	18-May-2007	pooka	Introduce noref setbacks, which the file server can use to signal the kernel it has 0 references to the node in question. In other words, this can be used to avoid inactive(), or, if the file server does not implement inactive, prompt reclaim for removed nodes.
1.36	08-May-2007	pooka	If the op was interrupted, decrease ops waiting for fetch from the file server only if the op was still waiting for fetch (as opposed to waiting for the response). Also, properly flag the possible following inactive as an op for which we do not want to wait for the response from the file server.
1.35	07-May-2007	pooka	Introduce puffs "setbacks", which can be used to set certain flags for nodes upon return from the userspace. Currently it can be used to indicate that the file server should be notified of "inactive" in case the file server has opted to not receive inactive every time the reference count for a vnode drops to zero. (inactive is a common event, almost never requires any action and must be executed sychronously, so it is wasteful). While doing this, cleanup the release-relock nonsense from the vntouser*() arguments. It was never enabled and the whole LOCKEDVP() concept was very broken to begin with.
1.34	01-May-2007	pooka	Fix a problem introduced when I converted puffs to use newlock2: when unmounting the file system in case of a certain timing (and possibly some other conditions), a thread would wait on a condition variable, while another thread broadcast the cv and immediately proceeded to destroy it. The result was a system frozen completely solid shorly after the process waiting for the cv woke up. So introduce reference counting to synchronize destruction of the resources in unmount. I was able to repeat the problem only on my laptop in some special cases, so I do not know how common it was. Ironically, killing the file server process violently instead of unmount() didn't have this problem because it never entered the unmount path from two directions.
1.33	24-Apr-2007	pooka	remember to flag park as done when we're done with it
1.32	22-Apr-2007	pooka	Now that puffs_park is allocated from the heap and actually freed by the userdead routine, don't do a TAILQ_FOREACH but rather an honest for loop.
1.31	21-Apr-2007	pooka	Take care not to access park->park_preq if the waiter is gone, as that memory is no longer available.
1.30	20-Apr-2007	pooka	don't mutex_enter() manually, we've already park_reference()d a few lines earlier for entering the same mutex
1.29	11-Apr-2007	pooka	make overspammy debug printf less overspammy
1.28	04-Apr-2007	pooka	Fix one more bug from today's commit: don't remove the op for which getops runs out of file server buffer space from the request queue. Otherwise that operation silently vanishes and things go, well, quite wrong.
1.27	04-Apr-2007	pooka	fix two loop mutex botches in previous
1.26	04-Apr-2007	pooka	Make it possible to interrupt waiters for fs operation completion again. This is useful until locking is further developed and basically any deadlocks can be solved by killing appropriate processes. Thanks especially to Tommi Kyntola and Antti Louko for sitting down with me and discussing resource ownership and locking strategies in implementing this.
1.25	04-Apr-2007	pooka	s/ppark/park/ to make all the variable names consistent - park is always a pointer now. no functional change
1.24	30-Mar-2007	pooka	* abstract ASYNCBIOREAD and let callers freely issue a callback called from putop. even though there's only one user currently, makes code more readable * move "delta" to a standard parameter in vntouser and get rid of the specialcase vntouser_delta
1.23	29-Mar-2007	pooka	in userdead assign waiter return value only if there is a waiter for a particular request
1.22	29-Mar-2007	pooka	Convert spinlocks & sleep/wakeup to newlock2 locking stuff. Fix a bunch of bugs. * park structures are now always allocated from a pool instead of a mixed stack/malloc allocation * get rid of the whole adjbuf concept, always just alloc the maximal amount of memory to satisfy a request * little regression: don't allow interrupting wait from file system to userspace; this had problems already before, but now the problems really started to shine through. I'll try to make this work again some day. * fix bmap to return a sensible value in runp
1.21	20-Mar-2007	pooka	* rework the page cache interaction a bit: cache metadata in the kernel and flush it out all at once instead of continuous updating * add support for delivering notifications to the file server about when a page was written to (but disabled by default for now). the file server can use this to request flushing or invalidating the kernel page cache
1.20	14-Mar-2007	pooka	branches: 1.20.2; Support B_READ\|B_ASYNC in strategy by calling biodone() directly when the file server puts the result.
1.19	27-Feb-2007	pooka	branches: 1.19.2; 1.19.4; Make wait for the user file server PCATCHable. This makes it possible to recover the system by just killing processes in case a file server manages to recurse into itself either by fault of file server implementation or by pilot error. The downside is that the code is extremely hard to follow and practically screams out for newlock2 (in addition to screaming "bug here"). The whole PCATCH nonsense and induced megacomplexity can hopefully be avoided in the future by tweaking other parts of the implementation.
1.18	03-Feb-2007	pooka	branches: 1.18.2; fstrans owner automatically gets a normal lock, don't need to lazy lock pointed out by hannken
1.17	29-Jan-2007	hannken	Change fstrans enum types to upper case. No functional change. From Antti Kantee <pooka@netbsd.org>
1.16	26-Jan-2007	pooka	Initial attempt at suspend/snapshot support for userspace file servers. This is still pretty much on the level "if it breaks ...". It should work for single-threaded servers which handle one operation from start to finish in one go. Also, it does not yet totally correctly synchronize metadata and data in some cases. So needless to say, it needs improvement, but it is possible that will have to wait for some lock revampage.
1.15	19-Jan-2007	pooka	debug print requests going into the queue
1.14	15-Jan-2007	pooka	Store puffs_node's on lists hashed with the cookie value instead of just one flat list.
1.13	29-Dec-2006	pooka	branches: 1.13.2; Don't allow calls to be queued while MOUNTING. We don't make any kernel->server calls at that time and it allows a window where operations use an incorrect root node cookie. XXX: there's still a (very much smaller and biglock safe) race, but that's going to be solved by some more thorough restructuring
1.12	10-Dec-2006	pooka	Don't return EWOULDBLOCK in case we have delivered some requests even if we are operating on a nonblocking descriptor.
1.11	10-Dec-2006	pooka	PCATCH in tsleep while waiting for operations in getop. Otherwise we could end up in an unkillable deadlock if GETOP was called when an operation that had locked the root vnode was already in userspace.
1.10	05-Dec-2006	pooka	branches: 1.10.2; shuffle functions around a bit: move the transport (/dev/puffs) to a different file from the messaging (request contents). no functional change
1.9	05-Dec-2006	pooka	Allow multiple requests to be transferred in each GET/PUTOP. For a single request, the performance is still the same.
1.8	21-Nov-2006	pooka	if we are going to bail due to the mountpoint being gone from under us while waiting for syncer lock, release the newly acquired syncer lock prior to bailing
1.7	21-Nov-2006	pooka	cosmetics
1.6	14-Nov-2006	pooka	branches: 1.6.2; Fix a race condition with unmount where the mountpoint might disappear from under us while waiting for syncer_lock and before we got to vfs_busy. This happens easily e.g. when the userspace server loses its will to live in VOP_RECLAIM, which is called from vflush() in VFS_UNMOUNT. We get two competing unmounters. When the first one finishes, it releases syncer_lock. Now the second one tries to vfs_busy(), but is greeted with garbage in *mp. XXX: Technically this is a more general issue and should be fixed elsewhere, but it's hard to trigger it with normal file systems unless they are unmounted "simultaneously" twice and are dirty enough for flushing to take a while. So make a note about it in the little black book next to the poems and postpone the crusade for now.
1.5	09-Nov-2006	pooka	few renames to better differentiate between mount & start.. plus some other renaming
1.4	07-Nov-2006	pooka	attach to genfs & support page cache. most noticeable effect is mmap and therefore execution of binaries starting to work, some speed improvements with large file I/O also. caching semantics and error case handling most likely need revisiting.
1.3	06-Nov-2006	pooka	puffs_park always contains a specific puffs_req, so make it a member instead of a pointer
1.2	25-Oct-2006	pooka	If the control descriptor is closed, mark userspace dead and wakeup all waiters before trying to get the syncer lock necessary for dounmount(). This prevents a deadlock if the userspace server dies while the syncer is running.
1.1	22-Oct-2006	pooka	kernel portion of puffs - the Pass-to-Userspace Framework File System. It contains the VFS attachment and userspace message-passing interface. This work was initially started and completed for Google SoC 2005 and tweaked to work a bit better in the past few weeks. While being far from complete, it is functional enough to be able and stable to host a fairly general-purpose in-memory file system in userspace. Even so, puffs should be considered experimental and no binary compatibility for interfaces or crash-freedom or zero security implications should be relied upon just yet. The GSoC project was mentored by William Studenmund and the final review for the code was done by Christos.
1.6.2.5	09-Feb-2007	ad	Sync with HEAD.
1.6.2.4	01-Feb-2007	ad	Sync with head.
1.6.2.3	12-Jan-2007	ad	Sync with head.
1.6.2.2	18-Nov-2006	ad	Sync with head.
1.6.2.1	14-Nov-2006	ad	file puffs_msgif.c was added on branch newlock2 on 2006-11-18 21:39:20 +0000
1.10.2.3	18-Dec-2006	yamt	sync with head.
1.10.2.2	10-Dec-2006	yamt	sync with head.
1.10.2.1	05-Dec-2006	yamt	file puffs_msgif.c was added on branch yamt-splraiseipl on 2006-12-10 07:18:38 +0000
1.13.2.9	04-Feb-2008	yamt	sync with head.
1.13.2.8	21-Jan-2008	yamt	sync with head
1.13.2.7	07-Dec-2007	yamt	sync with head
1.13.2.6	15-Nov-2007	yamt	sync with head.
1.13.2.5	27-Oct-2007	yamt	sync with head.
1.13.2.4	03-Sep-2007	yamt	sync with head.
1.13.2.3	26-Feb-2007	yamt	sync with head.
1.13.2.2	30-Dec-2006	yamt	sync with head.
1.13.2.1	29-Dec-2006	yamt	file puffs_msgif.c was added on branch yamt-lazymbuf on 2006-12-30 20:50:00 +0000
1.18.2.5	17-May-2007	yamt	sync with head.
1.18.2.4	07-May-2007	yamt	sync with head.
1.18.2.3	15-Apr-2007	yamt	sync with head.
1.18.2.2	24-Mar-2007	yamt	sync with head.
1.18.2.1	12-Mar-2007	rmind	Sync with HEAD.
1.19.4.1	11-Jul-2007	mjf	Sync with head.
1.19.2.8	12-Oct-2007	ad	Sync with head.
1.19.2.7	09-Oct-2007	ad	Sync with head.
1.19.2.6	01-Sep-2007	ad	Update for pool_cache API changes.
1.19.2.5	20-Aug-2007	ad	Sync with HEAD.
1.19.2.4	09-Jun-2007	ad	Sync with head.
1.19.2.3	08-Jun-2007	ad	Sync with head.
1.19.2.2	10-Apr-2007	ad	Sync with head.
1.19.2.1	05-Apr-2007	ad	Compile fixes.
1.20.2.1	29-Mar-2007	reinoud	Pullup to -current
1.39.2.1	15-Aug-2007	skrll	Sync with HEAD.
1.40.10.2	19-Jul-2007	pooka	add debug printf
1.40.10.1	19-Jul-2007	pooka	file puffs_msgif.c was added on branch matt-mips64 on 2007-07-19 22:05:23 +0000
1.40.8.2	14-Oct-2007	yamt	sync with head.
1.40.8.1	06-Oct-2007	yamt	sync with head.
1.40.6.4	23-Mar-2008	matt	sync with HEAD
1.40.6.3	09-Jan-2008	matt	sync with HEAD
1.40.6.2	08-Nov-2007	matt	sync with -HEAD
1.40.6.1	06-Nov-2007	matt	sync with HEAD
1.40.4.10	09-Dec-2007	jmcneill	Sync with HEAD.
1.40.4.9	27-Nov-2007	joerg	Sync with HEAD. amd64 Xen support needs testing.
1.40.4.8	21-Nov-2007	joerg	Sync with HEAD.
1.40.4.7	14-Nov-2007	joerg	Sync with HEAD.
1.40.4.6	11-Nov-2007	joerg	Sync with HEAD.
1.40.4.5	04-Nov-2007	jmcneill	Sync with HEAD.
1.40.4.4	28-Oct-2007	joerg	Sync with HEAD.
1.40.4.3	26-Oct-2007	joerg	Sync with HEAD. Follow the merge of pmap.c on i386 and amd64 and move pmap_init_tmp_pgtbl into arch/x86/x86/pmap.c. Modify the ACPI wakeup code to restore CR4 before jumping back into kernel space as the large page option might cover that.
1.40.4.2	07-Oct-2007	joerg	Sync with HEAD.
1.40.4.1	02-Oct-2007	joerg	Sync with HEAD.
1.47.2.4	21-Nov-2007	bouyer	Sync with HEAD
1.47.2.3	18-Nov-2007	bouyer	Sync with HEAD
1.47.2.2	13-Nov-2007	bouyer	Sync with HEAD
1.47.2.1	25-Oct-2007	bouyer	Sync with HEAD.
1.51.2.3	18-Feb-2008	mjf	Sync with HEAD.
1.51.2.2	08-Dec-2007	mjf	Sync with HEAD.
1.51.2.1	19-Nov-2007	mjf	Sync with HEAD.
1.60.2.1	08-Dec-2007	ad	Sync with head.
1.61.4.2	08-Jan-2008	bouyer	Sync with HEAD
1.61.4.1	02-Jan-2008	bouyer	Sync with HEAD
1.68.10.4	11-Aug-2010	yamt	sync with head.
1.68.10.3	11-Mar-2010	yamt	sync with head
1.68.10.2	04-May-2009	yamt	sync with head.
1.68.10.1	16-May-2008	yamt	sync with head.
1.68.8.1	18-May-2008	yamt	sync with head.
1.68.6.2	28-Sep-2008	mjf	Sync with HEAD.
1.68.6.1	02-Jun-2008	mjf	Sync with HEAD.
1.71.6.1	19-Oct-2008	haad	Sync with HEAD.
1.71.2.1	10-Oct-2008	skrll	Sync with HEAD.
1.72.12.1	21-Apr-2010	matt	sync to netbsd-5
1.72.8.1	13-May-2009	jym	Sync with HEAD. Commit is split, to avoid a "too many arguments" protocol error.
1.72.4.5	02-Nov-2011	riz	Pull up following revision(s) (requested by manu in ticket #1679): sys/fs/puffs/puffs_vnops.c: revision 1.157 sys/fs/puffs/puffs_vnops.c: revision 1.158 sys/fs/puffs/puffs_vnops.c: revision 1.159 sys/fs/puffs/puffs_vfsops.c: revision 1.97 sys/fs/puffs/puffs_vfsops.c: revision 1.99 sys/fs/puffs/puffs_vnops.c: revision 1.160 sys/fs/puffs/puffs_vfsops.c: revision 1.100 sys/miscfs/syncfs/sync_subr.c: revision 1.47 sys/fs/puffs/puffs_node.c: revision 1.21 sys/fs/puffs/puffs_node.c: revision 1.22 sys/fs/puffs/puffs_msgif.c: revision 1.88 sys/fs/puffs/puffs_msgif.c: revision 1.89 sys/fs/puffs/puffs_vnops.c: revision 1.156 Make sure ioflush does not sleep in PUFFS code path, waiting for a mutex, a memory allocation, or a response from the filesystem. This avoids deadlocks in the following situations: 1) when memory is low: ioflush waits the fileystem, the fielsystem waits for memory 2) when the filesystem does not respond (e.g.: network outage ona distributed filesystem) Fix the build that was broken by struct lwp *updateproc reference in RUMP-visible code. Instead of checking that updateproc (aka ioflush, aka syncer) will not sleep in PUFFS code, I check for any kernel thread: after all none of them are designed to hang awaiting for a remote filesystem operation to complete. Roll back the change that forced kernel threads to not sleep in PUFFS. The change does not make consensus, since only pagedaemon should need it. Other threads will tolerate sleeping, and problems here are only symptoms that something is going wrong in memory management. The cause, not the symptoms, need to be fixed. Make sure pagedaemon does not sleep for memory in puffs_vnop_sleep. Add KASSERT on any sleeping memory allocation to check it cannot happen again. Remove #ifdef DIAGNOSTIC guards around KASSERT, as the macro contains them
1.72.4.4	15-Jul-2011	riz	Pull up following revision(s) (requested by manu in ticket #1604): sys/fs/puffs/puffs_msgif.c: revision 1.84 Apply patch from PR kern/44093 by yamt: Interrupt server wait only on certain signals (same set at nfs -i) instead of all signals. According to the PR this helps with "git clone" run on a puffs file system.
1.72.4.3	20-May-2011	bouyer	Revert ticket 1604, it does't build.
1.72.4.2	19-May-2011	bouyer	Pull up following revision(s) (requested by manu in ticket #1604): sys/fs/puffs/puffs_msgif.c: revision 1.84 via patch Apply patch from PR kern/44093 by yamt: Interrupt server wait only on certain signals (same set at nfs -i) instead of all signals. According to the PR this helps with "git clone" run on a puffs file system.
1.72.4.1	09-Jan-2010	snj	Pull up following revision(s) (requested by pooka in ticket #1212): sys/fs/puffs/puffs_msgif.c: revision 1.76 via patch sys/fs/puffs/puffs_sys.h: revision 1.73 via patch sys/fs/puffs/puffs_vfsops.c: revision 1.84 via patch Process flush requests from the file server in a separate thread context. This fixes a long-standing but seldomly seen deadlock, where the kernel was holding pages busy (due to e.g. readahead request) while waiting for the server to respond, and the server made a callback into the kernel asking to invalidate those pages. ... or, well, theoretically fixes, since I didn't have any reliable way of repeating the deadlock and I think I saw it only twice.
1.72.2.1	28-Apr-2009	skrll	Sync with HEAD.
1.80.4.2	05-Mar-2011	rmind	sync with head
1.80.4.1	16-Mar-2010	rmind	Change struct uvm_object::vmobjlock to be dynamically allocated with mutex_obj_alloc(). It allows us to share the locks among UVM objects.
1.80.2.1	17-Aug-2010	uebayasi	Sync with HEAD.
1.84.4.1	17-Feb-2011	bouyer	Sync with HEAD
1.84.2.1	06-Jun-2011	jruoho	Sync with HEAD.
1.85.2.1	23-Jun-2011	cherry	Catchup with rmind-uvmplock merge.
1.89.8.1	12-Aug-2012	martin	Pull up following revision(s) (requested by manu in ticket #438): lib/libperfuse/perfuse_priv.h: revision 1.31 sys/fs/puffs/puffs_msgif.h: revision 1.80 sys/fs/puffs/puffs_vnops.c: revision 1.171 lib/libpuffs/puffs_ops.3: revision 1.31 sys/fs/puffs/puffs_vnops.c: revision 1.172 sys/fs/puffs/puffs_vnops.c: revision 1.173 sys/fs/puffs/puffs_vnops.c: revision 1.174 usr.sbin/perfused/perfused.c: revision 1.24 sys/fs/puffs/puffs_sys.h: revision 1.80 sys/fs/puffs/puffs_sys.h: revision 1.81 sys/fs/puffs/puffs_sys.h: revision 1.82 lib/libperfuse/subr.c: revision 1.19 lib/libperfuse/perfuse.c: revision 1.30 sys/fs/puffs/puffs_msgif.c: revision 1.90 sys/fs/puffs/puffs_msgif.c: revision 1.91 sys/fs/puffs/puffs_msgif.c: revision 1.92 lib/libperfuse/ops.c: revision 1.59 lib/libpuffs/puffs.3: revision 1.53 lib/libperfuse/debug.c: revision 1.12 lib/libpuffs/puffs.3: revision 1.54 sys/fs/puffs/puffs_vnops.c: revision 1.167 sys/fs/puffs/puffs_msgif.h: revision 1.79 usr.sbin/perfused/msg.c: revision 1.21 sys/fs/puffs/puffs_vfsops.c: revision 1.102 sys/fs/puffs/puffs_vfsops.c: revision 1.103 sys/fs/puffs/puffs_vfsops.c: revision 1.105 lib/libpuffs/puffs.h: revision 1.123 lib/libperfuse/perfuse_if.h: revision 1.20 lib/libperfuse/perfuse.c: revision 1.29 lib/libpuffs/dispatcher.c: revision 1.42 lib/libpuffs/dispatcher.c: revision 1.43 - Fix same vnodes associated with multiple cookies The scheme used to retreive known nodes on lookup was flawed, as it only used parent and name. This produced a different cookie for the same file if it was renamed, when looking up ../ or when dealing with multiple files associated with the same name through link(2). We therefore abandon the use of node name and introduce hashed lists of inodes. This causes a huge rewrite of reclaim code, which do not attempt to keep parents allocated until all their children are reclaimed - Fix race conditions in reclaim There are a few situations where we issue multiple FUSE operations for a PUFFS operation. On reclaim, we therefore have to wait for all FUSE operation to complete, not just the current exchanges. We do this by introducing node reference count with node_ref() and node_rele(). - Detect data loss caused by FAF VOP_PUTPAGES causes FAF writes where the kernel does not check the operation result. At least issue a warning on error. - Enjoy FAF shortcut on setattr No need to wait for the result if the kernel does not want it. There is however an exception for setattr that touch the size, we need to wait for completion because we have other operations queued for after the resize. - Fix fchmod() on write-open file fchmod() on a node open with write privilege will send setattr with both mode and size set. This confuses some FUSE filesystem. Therefore we send two FUSE operations, one for mode, and one for size. - Remove node TTL handling for netbsd-5 for simplicity sake. The code still builds on netbsd-5 but does not have the node TTL feature anymore. It works fine with kernel support on netbsd-6. - Improve PUFFS_KFLAG_CACHE_FS_TTL by reclaiming older inactive nodes. The normal kernel behavior is to retain inactive nodes in the freelist until it runs out of vnodes. This has some merit for local filesystems, where the cost of an allocation is about the same as the cost of a lookup. But that situation is not true for distributed filesystems. On the other hand, keeping inactive nodes for a long time hold memory in the file server process, and when the kernel runs out of vnodes, it produce reclaim avalanches that increase lattency for other operations. We do not reclaim inactive vnodes immediatly either, as they may be looked up again shortly. Instead we introduce a grace time and we reclaim nodes that have been inactive beyond the grace time. - Fix lookup/reclaim race condition. The above improvement undercovered a race condition between lookup and reclaim. If we reclaimed a vnode associated with a userland cookie while a lookup returning that same cookiewas inprogress, then the kernel ends up with a vnode associated with a cookie that has been reclaimed in userland. Next operation on the cookie will crash (or at least confuse) the filesystem. We fix this by introducing a lookup count in kernel and userland. On reclaim, the kernel sends the count, which enable userland to detect situation where it initiated a lookup that is not completed in kernel. In such a situation, the reclaim must be ignored, as the node is about to be looked up again. Fix hang unmount bug introduced by last commit. We introduced a slow queue for delayed reclaims, while the existing queue for unmount, flush and exist has been renamed fast queue. Both queues had timestamp for when an operation should be done, but it was useless for the fast queue, which is always used to run an operation ASAP. And the timestamp test had an error that turned ASAP into "at next tick", but nobody what there to wake the thread at next tick, hence the hang. The fix is to remove the useless and buggy timestamp test for fast queue. Rename slow sopreq queue into node sopreq queue, to refet the fact that is only intended for postponed node reclaims. When purging the node sopreq queue, do not call puffs_msg_sendresp(), as it makes no sense. Fix race condition between (create\|mknod\|mkdir\|symlino) and reclaim, just like we did it between lookup and reclaim. Missing bit in previous commit (prevent race between create\|mknod\|mkdir\|symlink and reclaim) Bump date for previous. New sentence, new line; remove trailing whitespace; fix typos; punctuation nits. Add PUFFS_KFLAG_CACHE_DOTDOT so that vnodes hold a reference on their parent, keeping them active, and allowing to lookup .. without sending a request to the filesystem. Enable the featuure for perfused, as this is how FUSE works. Missing bit in previous commit (PUFFS_KFLAG_CACHE_DOTDOT option to avoid looking up ..)
1.89.2.3	22-May-2014	yamt	sync with head. for a reference, the tree before this commit was tagged as yamt-pagecache-tag8. this commit was splitted into small chunks to avoid a limitation of cvs. ("Protocol error: too many arguments")
1.89.2.2	16-Jan-2013	yamt	sync with (a bit old) head
1.89.2.1	30-Oct-2012	yamt	sync with head
1.92.2.3	03-Dec-2017	jdolecek	update from HEAD
1.92.2.2	20-Aug-2014	tls	Rebase to HEAD as of a few days ago.
1.92.2.1	20-Nov-2012	tls	Resync to 2012-11-19 00:00:00 UTC
1.93.2.1	18-May-2014	rmind	sync with head
1.94.4.2	17-Jan-2015	martin	Pull up following revision(s) (requested by maxv in ticket #427): sys/compat/svr4/svr4_schedctl.c: revision 1.8 sys/netinet/tcp_timer.c: revision 1.88 sys/miscfs/genfs/layer_vfsops.c: revision 1.45 sys/compat/svr4/svr4_ioctl.c: revision 1.37 sys/ufs/chfs/chfs_vfsops.c: revision 1.14 sys/miscfs/fdesc/fdesc_vfsops.c: revision 1.91 sys/compat/linux/arch/i386/linux_ptrace.c: revision 1.30 sys/compat/common/kern_time_50.c: revision 1.28 sys/netinet6/ip6_forward.c: revision 1.74 sys/miscfs/umapfs/umap_vnops.c: revision 1.57 sys/compat/svr4/svr4_fcntl.c: revision 1.74 distrib/sets/lists/comp/mi: revision 1.1931 sys/netinet6/udp6_output.c: revision 1.46 sys/fs/puffs/puffs_compat.c: revision 1.3 sys/fs/udf/udf_rename.c: revision 1.11 sys/compat/svr4/svr4_filio.c: revision 1.24 sys/fs/udf/udf_rename.c: revision 1.12 sys/netinet/tcp_usrreq.c: revision 1.202 sys/miscfs/umapfs/umap_subr.c: revision 1.29 sys/compat/linux/common/linux_fadvise64.c: revision 1.3 sys/netinet/if_atm.c: revision 1.34 sys/miscfs/procfs/procfs_subr.c: revision 1.106 sys/miscfs/genfs/layer_subr.c: revision 1.37 sys/netinet/tcp_sack.c: revision 1.30 sys/compat/freebsd/freebsd_misc.c: revision 1.33 sys/compat/freebsd/freebsd_file.c: revision 1.33 sys/ufs/chfs/chfs_vnode.c: revision 1.12 sys/compat/svr4/svr4_ttold.c: revision 1.34 sys/compat/linux/common/linux_file.c: revision 1.114 sys/compat/linux/arch/mips/linux_machdep.c: revision 1.43 sys/compat/linux/common/linux_signal.c: revision 1.76 sys/compat/common/compat_util.c: revision 1.46 sys/compat/linux/arch/arm/linux_ptrace.c: revision 1.18 sys/compat/svr4/svr4_sockio.c: revision 1.36 sys/compat/linux/arch/arm/linux_machdep.c: revision 1.32 sys/compat/svr4/svr4_signal.c: revision 1.66 sys/kern/kern_exec.c: revision 1.410 sys/fs/puffs/puffs_vfsops.c: revision 1.115 sys/compat/svr4/svr4_exec_elf64.c: revision 1.15 sys/compat/linux/arch/i386/linux_machdep.c: revision 1.159 sys/compat/linux/arch/alpha/linux_machdep.c: revision 1.50 sys/compat/linux32/common/linux32_misc.c: revision 1.24 sys/netinet/in_pcb.c: revision 1.153 sys/sys/malloc.h: revision 1.116 sys/compat/common/if_43.c: revision 1.9 share/man/man9/Makefile: revision 1.380 sys/netinet/tcp_vtw.c: revision 1.12 sys/miscfs/umapfs/umap_vfsops.c: revision 1.95 sys/ufs/ext2fs/ext2fs_vfsops.c: revision 1.186 sys/compat/common/uipc_syscalls_43.c: revision 1.46 sys/ufs/ext2fs/ext2fs_vnops.c: revision 1.115 sys/fs/puffs/puffs_msgif.c: revision 1.97 sys/compat/svr4/svr4_ipc.c: revision 1.27 sys/compat/linux/common/linux_exec.c: revision 1.117 sys/ufs/ext2fs/ext2fs_readwrite.c: revision 1.66 sys/netinet/tcp_output.c: revision 1.179 sys/compat/svr4/svr4_termios.c: revision 1.28 sys/fs/udf/udf_strat_bootstrap.c: revision 1.4 sys/fs/puffs/puffs_subr.c: revision 1.67 sys/fs/puffs/puffs_node.c: revision 1.36 sys/miscfs/overlay/overlay_vnops.c: revision 1.21 sys/fs/cd9660/cd9660_node.c: revision 1.34 sys/netinet/raw_ip.c: revision 1.146 sys/sys/mallocvar.h: revision 1.13 sys/miscfs/overlay/overlay_vfsops.c: revision 1.63 share/man/man9/malloc.9: revision 1.50 sys/netinet6/dest6.c: revision 1.18 sys/compat/linux/common/linux_uselib.c: revision 1.33 sys/compat/linux/common/linux_socket.c: revision 1.120 share/man/man9/malloc.9: revision 1.51 sys/netinet/tcp_subr.c: revision 1.257 sys/compat/linux/common/linux_socketcall.c: revision 1.45 sys/compat/linux/common/linux_fadvise64_64.c: revision 1.3 sys/compat/freebsd/freebsd_ipc.c: revision 1.17 sys/compat/linux/common/linux_misc_notalpha.c: revision 1.109 sys/compat/linux/arch/alpha/linux_pipe.c: revision 1.17 sys/netinet6/in6_pcb.c: revision 1.132 sys/netinet6/in6_ifattach.c: revision 1.94 sys/compat/svr4/svr4_exec_elf32.c: revision 1.15 sys/miscfs/nullfs/null_vfsops.c: revision 1.90 sys/fs/cd9660/cd9660_util.c: revision 1.12 sys/compat/linux/arch/powerpc/linux_machdep.c: revision 1.48 sys/compat/freebsd/freebsd_exec_elf32.c: revision 1.20 sys/miscfs/procfs/procfs_vfsops.c: revision 1.94 sys/compat/linux/arch/powerpc/linux_ptrace.c: revision 1.28 sys/compat/linux/common/linux_sched.c: revision 1.67 sys/compat/linux/common/linux_exec_aout.c: revision 1.67 sys/compat/linux/common/linux_pipe.c: revision 1.67 sys/compat/linux/common/linux_llseek.c: revision 1.34 sys/compat/linux/arch/mips/linux_ptrace.c: revision 1.10 Do not uselessly include <sys/malloc.h>. Cleanup: - remove struct kmembuckets (dead) - correctly deadify MALLOC_XX - remove MALLOC_DEFINE_LIMIT and MALLOC_JUSTDEFINE_LIMIT (dead) - remove malloc_roundup(), malloc_type_setlimit(), MALLOC_DEFINE_LIMIT() and MALLOC_JUSTDEFINE_LIMIT() from man 9 malloc New sentence, new line. Bump date for previous. Obsolete malloc_roundup(9), malloc_type_setlimit(9) and MALLOC_DEFINE_LIMIT(9) man pages.
1.94.4.1	29-Aug-2014	martin	Pull up following revision(s) (requested by hannken in ticket #67): sys/fs/puffs/puffs_sys.h: revision 1.86 sys/fs/puffs/puffs_vfsops.c: revision 1.114 sys/fs/puffs/puffs_msgif.c: revision 1.95 sys/fs/puffs/puffs_node.c: revision 1.32 sys/fs/puffs/puffs_vnops.c: revision 1.184 Change puffs from hashlist to vcache. - field "pa_nhashbuckets" of struct "puffs_kargs" becomes a no-op. and should be removed on the next protocol version bump.
1.97.2.4	28-Aug-2017	skrll	Sync with HEAD
1.97.2.3	05-Feb-2017	skrll	Sync with HEAD
1.97.2.2	09-Jul-2016	skrll	Sync with HEAD
1.97.2.1	06-Jun-2015	skrll	Sync with HEAD
1.99.2.2	26-Apr-2017	pgoyette	Sync with HEAD
1.99.2.1	07-Jan-2017	pgoyette	Sync with HEAD. (Note that most of these changes are simply $NetBSD$ tag issues.)
1.100.2.1	21-Apr-2017	bouyer	Sync with HEAD
1.101.12.2	08-Apr-2020	martin	Merge changes from current as of 20200406
1.101.12.1	10-Jun-2019	christos	Sync with HEAD
1.101.10.11	22-Jan-2019	pgoyette	Convert the MODULE_{,VOID_}HOOK_CALL macros to do everything in-line rather than defining an intermediate hook##call function. Almost all of the hooks are called only once, and although we lose the ability of doing things like if (MODULE_HOOK_CALL(...) == 0) ... we simplify things quite a bit. With this change, we no longer need to have both declaration and definition macros, and the definition no longer needs to have both prototype argument list and a "real" argument list. FWIW, the above if now needs to written as int ret; MODULE_HOOK_CALL(..., ret); if (ret == 0) ... with appropriate use of braces {}.
1.101.10.10	21-Jan-2019	pgoyette	No need to declare the hook_call() function for void hooks. So remove and simplify.
1.101.10.9	18-Jan-2019	pgoyette	Don't restrict hooks to having only int or void types. Pass the hook's type to the various macros, as needed. Allows us to reduce diffs to original in at least one or two places (we no longer have to provide an additional parameter to the hook routine for returning a non-int return value).
1.101.10.8	14-Jan-2019	pgoyette	Create a variant of the HOOK macros that handles hook routines of type void, and use them where appropriate.
1.101.10.7	13-Jan-2019	pgoyette	Remove the HOOK2 versions of the MODULE_HOOK macros. There were only a few uses, and using them led to some lack of clarity in the code. Instead, we now use two separate hooks, with names that make it clear(er) what we're doing. This also positions us to start unraveling some of the rtsock_50 mess, which will need (at least) five hooks.
1.101.10.6	29-Sep-2018	pgoyette	In MODULE_HOOK_CALL_DECL we don't need to provide the actual argument list for calling the hook function, nor do we need to provide the default value (for when the hook has not been set).
1.101.10.5	18-Sep-2018	pgoyette	The COMPAT_HOOK macros were renamed to MODULE_HOOK, adjust all callers
1.101.10.4	18-Sep-2018	pgoyette	Split the COMPAT_CALL_HOOK to separate the declaration from the implementation. Some hooks are called from multiple source files, and the old method resulted in duplicate implementations. Implement MP-safe hooks for the usb_subr_30 code. Pass the helper functions as arguments to the compat code so it does not have to determine if the kernel contains usb code.
1.101.10.3	17-Sep-2018	pgoyette	Adapt (most of) the indirect function pointers to the new MP-safe mechanism. Still remaining are the compat_netbsd32 stuff, and some usb subroutines.
1.101.10.2	24-Mar-2018	pgoyette	Use function pointers to call the compatability functions.
1.101.10.1	24-Mar-2018	pgoyette	Add fs/puffs compat_50 to the modules
1.104.6.1	29-Feb-2020	ad	Sync with head.
1.107.2.1	02-Aug-2025	perseant	Sync with HEAD
1.87	03-Dec-2021	pho	Avoid using register_t in <fs/puffs/puffs_msgif.h> The purpose of this header file is to interface between the kernel-space and user-space, and is #include'd by a user-space header <puffs.h>. It should therefore not use any of kernel-only types, as it's not reasonable to require user-land filesystems to #define _KERNTYPES.
1.86	08-Mar-2021	christos	give names to the enums so we can cast by name for lint
1.85	23-Sep-2019	christos	branches: 1.85.8; Restore binary compatibility by using the statvfs90 structure internally.
1.84	15-Feb-2015	manu	branches: 1.84.18; Add PUFFS_KFLAG_NOFLUSH_META to prevent sending metadata flush to FUSE FUSE filesystems do not expect to get metadata updates for [amc]time and size, they updates the value on their own after operations. The PUFFS PUFFS_KFLAG_NOFLUSH_META option prevents regular metadata cache flushes to the filesystem , and libperfuse uses it to match Linux FUSE behavior. While there, fix a bug in SETATTR: do not update kernel metadata cache from SETATTR reply when the request is asynchronous, as we do not have the reply yet.
1.83	31-Oct-2014	manu	branches: 1.83.2; Add PUFFS_HAVE_FALLOCATE in puffs_msgif.h so that filesystem can decide at build time wether fallocate is usable
1.82	31-Oct-2014	manu	Add PUFFS support for fallocate and fdiscard operations
1.81	16-Aug-2014	manu	Add a oflags input field to open requests so that the filesystem can pass back information about the file. Implement PUFFS_OPEN_IO_DIRECT, which will force direct IO (bypassing page cache) for the file.
1.80	10-Aug-2012	manu	branches: 1.80.2; 1.80.14; Add PUFFS_KFLAG_CACHE_DOTDOT so that vnodes hold a reference on their parent, keeping them active, and allowing to lookup .. without sending a request to the filesystem. Enable the featuure for perfused, as this is how FUSE works.
1.79	21-Jul-2012	manu	- Improve PUFFS_KFLAG_CACHE_FS_TTL by reclaiming older inactive nodes. The normal kernel behavior is to retain inactive nodes in the freelist until it runs out of vnodes. This has some merit for local filesystems, where the cost of an allocation is about the same as the cost of a lookup. But that situation is not true for distributed filesystems. On the other hand, keeping inactive nodes for a long time hold memory in the file server process, and when the kernel runs out of vnodes, it produce reclaim avalanches that increase lattency for other operations. We do not reclaim inactive vnodes immediatly either, as they may be looked up again shortly. Instead we introduce a grace time and we reclaim nodes that have been inactive beyond the grace time. - Fix lookup/reclaim race condition. The above improvement undercovered a race condition between lookup and reclaim. If we reclaimed a vnode associated with a userland cookie while a lookup returning that same cookiewas inprogress, then the kernel ends up with a vnode associated with a cookie that has been reclaimed in userland. Next operation on the cookie will crash (or at least confuse) the filesystem. We fix this by introducing a lookup count in kernel and userland. On reclaim, the kernel sends the count, which enable userland to detect situation where it initiated a lookup that is not completed in kernel. In such a situation, the reclaim must be ignored, as the node is about to be looked up again.
1.78	08-Apr-2012	manu	Add name and atttribute cache with filesytem provided TTL. lookup, create, mknod, mkdir, symlink, getattr and setattr messages have been extended so that attributes and their TTL can be provided by the filesytem. lookup, create, mknod, mkdir, and symlink messages are also extended so that the filesystem can provide name TTL.
1.77	27-Sep-2011	christos	branches: 1.77.2; 1.77.6; 1.77.8; don't get affected by the NAME_MAX bump. Use the same constant as the rest of the extrattr code.
1.76	04-Jul-2011	manu	Add a flag to VOP_LISTEXTATTR(9) so that the vnode interface can tell the filesystem in which format extended attribute shall be listed. There are currently two formats: - NUL-terminated strings, used for listxattr(2), this is the default. - one byte length-pprefixed, non NUL-terminated strings, used for extattr_list_file(2), which is obtanined by setting the EXTATTR_LIST_PREFIXLEN flag to VOP_LISTEXTATTR(9) This approach avoid the need for converting the list back and forth, except in libperfuse, since FUSE uses NUL-terminated strings, and the kernel may have requested EXTATTR_LIST_PREFIXLEN.
1.75	06-Jul-2010	pooka	Add compat to enable running puffs in a 64bit time_t kernel against a server which runs in 32bit time_t namespace.
1.74	07-Jun-2010	pooka	Make retval argument for pathconf a register_t to match VOP_PATHCONF. This makes the size the same on 64bit archs. Don't bother bumping any version, since you'd have explicitly had to jump through some hoops to use pathconf before.
1.73	21-May-2010	pooka	add option string for no attribute cache (foreseeing the odd event I might actually implement one some day)
1.72	21-May-2010	pooka	Since libpuffs needs a major bump for extattr support anyway, make some changes to the user-kernel protocol. Namely, try to be a little more resilient some future changes.
1.71	21-May-2010	pooka	Support extended attributes.
1.70	20-May-2010	pooka	Fix typo.
1.69	07-Jan-2010	pooka	branches: 1.69.2; 1.69.4; Add a PUFFS_UNMOUNT server->kernel request, which causes the kernel to initiate self destruct, i.e. unmount(MNT_FORCE). This, however, is a semi-controlled self-destruct, since all caches are flushed before the (possibly) violent unmount takes place.
1.68	17-Oct-2009	pooka	Bump protocol version once more to allow for previous to be pulled to netbsd-5 (protocols are not compatible due to time_t/dev_t change).
1.67	17-Oct-2009	pooka	Transmit VOP_ABORTOP() to the server.
1.66	12-Jan-2009	pooka	Bump interface version number for the time_t/dev_t changes.
1.65	28-Jan-2008	pooka	branches: 1.65.6; 1.65.10; 1.65.18; 1.65.20; 1.65.26; For code clarity typedef void *puffs_cookie_t. No functional change.
1.64	08-Dec-2007	pooka	Now that "l" is gone both as an argument to operations and from componentname, remove all vestiges of puffs_cid.
1.63	05-Dec-2007	pooka	Send a response message for flush operations from the kernel instead of abusing the return value of write(2).
1.62	04-Dec-2007	pooka	Add a bit to differentiate if a message is a request or a response.
1.61	27-Nov-2007	pooka	branches: 1.61.2; Remove "puffs_cid" from the puffs interface following l-removal from the kernel vfs interfaces. puffs_cc_getcaller(pcc) can be used now should the same information be desired.
1.60	12-Nov-2007	pooka	* split the putter header into a kernel version and a userland version + install latter to /usr/include/dev/putter * remove last dependencies to puffs from putter, it's completely independent now
1.59	21-Oct-2007	pooka	branches: 1.59.2; Always provide caller information from the kernel based on curlwp. (but don't deprecate the old puffs_cid interface just yet)
1.58	19-Oct-2007	pooka	When doing a read operation, don't copy the whole kernel buffer to userspace, since it doesn't contain any information yet. I should still rework this more so this is just a quickie to get the read/write style interface more up to speed with the ioctl version.
1.57	11-Oct-2007	pooka	branches: 1.57.2; g/c garbage
1.56	11-Oct-2007	pooka	Part 1/n of some pretty extensive changes to how the kernel module interacts with the userspace file server: * since the kernel-user communication is not purely request-response anymore (hasn't been since 2006), try to rename some "request" to "message". more similar mangling will take place in the future. * completely rework how messages are allocated. previously most of them were borrowed from the stack (originally all of them), but now always allocate dynamically. this makes the structure of the code much cleaner. also makes it possible to fix a locking order violation. it enables plenty of future enhancements. * start generalizing the transport interface to be independent of puffs * move transport interface to read/write instead of ioctl. the old one had legacy design problems, and besides, ioctl's suck. implement a very generic version for now; this will be worked on later hopefully some day reaching "highly optimized". * implement libpuffs support behind existing library request interfaces. this will change eventually (I hate those interfaces)
1.55	04-Oct-2007	pooka	g/c the "sizeop" code previous used for ioctl/fcntl. It was already commented out and has bitrotted beyond all recognition, so it needs complete rethinking.
1.54	02-Oct-2007	pooka	If kernel resource allocation fails after the file server has committed something, issue an abort. The abort is done through the regular op channel, e.g. failed mkdir leads to regular rmdir, inactive and reclaim. No internal interface is planned currently for the one file system out of a million which would implement it to benefit from the one case in a billion where kernel resource allocation actually does fail and out of that one case in a trillion where internal vs. external would make a difference.
1.53	01-Oct-2007	pooka	* better error checking: validate error values received from userland to be vaild errno values * include string describing error in PUFFS_ERR * get rid of union in puffs_req, it's nothing but trouble * pass pmp to async i/o callbacks
1.52	27-Sep-2007	pooka	nuke trailing , from enum. spotted by xtraeme
1.51	27-Sep-2007	pooka	Add error notifications, which are used to deliver errors from the kernel to the file server for silly things the file server did, e.g. attempting to create a file with size VSIZENOTSET. The file server can handle these as it chooses, but the default action is for it to throw its hands in the air and sing "goodbye, cruel world, it's over, walk on by".
1.50	23-Aug-2007	pooka	branches: 1.50.2; 1.50.4; Add a third type of fh option, passthrough, where the kernel does not attempt to handle struct fid at all and passes it as such to userspace.
1.49	22-Aug-2007	pooka	Mimic namei structure changes for puffs. bump both kernel & lib version.
1.48	15-Aug-2007	pooka	Nuke PUFFSLOOKUP_FOO and move to NAMEI_FOO
1.47	30-Jul-2007	pooka	branches: 1.47.4; 1.47.6; Move PUFFS_TYPEPREFIX to puffs_msgif.h since it's used in a macro there.
1.46	27-Jul-2007	pooka	include <uvm/uvm_prot.h>
1.45	27-Jul-2007	pooka	Change unused fflags parameter in VOP_MMAP to prot and pass in desired vm protection.
1.44	19-Jul-2007	pooka	define PUFFSREQSIZEOP ioctl, which can be used to fetch the maximum request size
1.43	18-Jul-2007	pooka	kill MFSNAMELEN limit
1.42	17-Jul-2007	pooka	branches: 1.42.2; Set a file server supplied file system type in the type field and set the mntfromname to be the place mounted from instead of the type.
1.41	16-Jul-2007	pooka	1\|2 is more correct when it's 3 instead of 2. This makes calls to the file server inactive less over-eagerly executed and masks some problems with the new mounting style. Effectively, it makes some file systems such as psshfs mountable again (only without -o allops).
1.40	02-Jul-2007	pooka	support turning REQUIREDIR off and extra consume in lookup
1.39	02-Jul-2007	pooka	Get rid of the "int *refs" parameter to inactive: the same can be accomplished now with puffs_setbacks.
1.38	01-Jul-2007	pooka	Give the file server to ability to request the entire pathname buffer under lookup by using PUFFS_KFLAG_LOOKUP_FULLPNBUF instead just the current component.
1.37	01-Jul-2007	pooka	Instead of supplying a plain pid, supply an abstract struct puffs_cid *, which can currently be used to query the pid and lwpid.
1.36	01-Jul-2007	pooka	make puffs_cred an opaque type
1.35	24-Jun-2007	pooka	Actually, keep PUFFS_KFLAG_NOCACHE and -o cache around as shorthand to neither page- nor namecache.
1.34	24-Jun-2007	pooka	Split the NOCACHE option in twain: NOCACHE_NAME & NOCACHE_PAGE.
1.33	06-Jun-2007	pooka	Move puffs to a two clause license where it already isn't so. And as agc pointed out, even files with the third clause were already effectively two clause because of a slight bug in the language...
1.32	18-May-2007	pooka	Introduce noref setbacks, which the file server can use to signal the kernel it has 0 references to the node in question. In other words, this can be used to avoid inactive(), or, if the file server does not implement inactive, prompt reclaim for removed nodes.
1.31	18-May-2007	pooka	Support VOP_POLL. This requires some acrobatics on the puffs_node, as we give a reference to userspace for the puffs_node for the duration of the poll call. So reference count puffs_node separately from the parent vnode. vref()/vrele() is not possible due to a possible surprise visit from VOP_INACTIVE.
1.30	17-May-2007	pooka	Make it possible for the file server to specify the root vnode type and other information instead of always using VDIR. To make this possible without races, require all root node information already in puffs_mount() and nuke puffs_start2() and the associated start operation completely. requested/inspired by Tobias Nygren
1.29	07-May-2007	pooka	Introduce puffs "setbacks", which can be used to set certain flags for nodes upon return from the userspace. Currently it can be used to indicate that the file server should be notified of "inactive" in case the file server has opted to not receive inactive every time the reference count for a vnode drops to zero. (inactive is a common event, almost never requires any action and must be executed sychronously, so it is wasteful). While doing this, cleanup the release-relock nonsense from the vntouser*() arguments. It was never enabled and the whole LOCKEDVP() concept was very broken to begin with.
1.28	22-Apr-2007	pooka	define PUFFS_KFLAG_WTCACHE, which makes the page cache write-through
1.27	16-Apr-2007	pooka	Give the file server the ability to specify the file handle length instead of defining a static length file handle on the framework-level.
1.26	13-Apr-2007	pooka	Allow file servers to request the number of hash cookie buckets for pnode -> vnode reverse lookup.
1.25	13-Apr-2007	pooka	* add fhlen to kernel argument structure * rename it to puffs_kargs instead of puffs_args
1.24	11-Apr-2007	pooka	* support VFS_FHTOVP and VFS_VPTOFH * support cookies in for VOP_READDIR nfs exporting puffs file systems works now
1.23	06-Apr-2007	pooka	actually, we don't need a separate op for flushing the whole page cache of a node, just use the range op with endoff = 0
1.22	29-Mar-2007	pooka	Convert spinlocks & sleep/wakeup to newlock2 locking stuff. Fix a bunch of bugs. * park structures are now always allocated from a pool instead of a mixed stack/malloc allocation * get rid of the whole adjbuf concept, always just alloc the maximal amount of memory to satisfy a request * little regression: don't allow interrupting wait from file system to userspace; this had problems already before, but now the problems really started to shine through. I'll try to make this work again some day. * fix bmap to return a sensible value in runp
1.21	20-Mar-2007	pooka	export puffs version of namei ISLASTCN macro to userspace
1.20	20-Mar-2007	pooka	* rework the page cache interaction a bit: cache metadata in the kernel and flush it out all at once instead of continuous updating * add support for delivering notifications to the file server about when a page was written to (but disabled by default for now). the file server can use this to request flushing or invalidating the kernel page cache
1.19	26-Jan-2007	pooka	branches: 1.19.2; 1.19.6; 1.19.8; 1.19.10; Initial attempt at suspend/snapshot support for userspace file servers. This is still pretty much on the level "if it breaks ...". It should work for single-threaded servers which handle one operation from start to finish in one go. Also, it does not yet totally correctly synchronize metadata and data in some cases. So needless to say, it needs improvement, but it is possible that will have to wait for some lock revampage.
1.18	16-Jan-2007	pooka	g/c revoke msg structure
1.17	09-Jan-2007	pooka	comment out flushmulti for now, it's not done and kdump will complain as mjf noted
1.16	09-Jan-2007	pooka	Introduce flush operations, which the fs server can use to control kernel caching. Currently supported are only flushing the name cache for a directory or flushing the name cache for the entire fs. Also, get rid of PNODE_INACTIVE status, since it was racy and essentially didn't work. All this on top of being useless in the first place ....
1.15	07-Jan-2007	pooka	vfs sync, flushes regular file data only (user server can take care of flushing any metadata it might have hidden away)
1.14	02-Jan-2007	pooka	* check userspace version and prevent incompatible mount * some general maintenance
1.13	29-Dec-2006	pooka	branches: 1.13.2; rename the kernel-provided componentname to puffs_kcn; libpuffs now provides puffs_cn built on top of it
1.12	07-Dec-2006	pooka	branches: 1.12.2; let implementation ultimately decide if mmap is supported - pass VOP_MMAP to fs server
1.11	05-Dec-2006	pooka	Allow multiple requests to be transferred in each GET/PUTOP. For a single request, the performance is still the same.
1.10	01-Dec-2006	pooka	prefix kernel flags with PUFFS_KFLAG to have a separate namespace from the library flags
1.9	01-Dec-2006	pooka	don't call the fs server for all operations, only those it has told us that it implements
1.8	18-Nov-2006	pooka	branches: 1.8.2; Require statvfs info from startreq so that we have that info available. Also, don't pass fsid to userspace and just fill it in the kernel.
1.7	17-Nov-2006	pooka	Introduce uncached operation, makes sense when the file system backend can be modified from elsewhere than the file system interface
1.6	09-Nov-2006	pooka	few renames to better differentiate between mount & start.. plus some other renaming
1.5	07-Nov-2006	pooka	attach to genfs & support page cache. most noticeable effect is mmap and therefore execution of binaries starting to work, some speed improvements with large file I/O also. caching semantics and error case handling most likely need revisiting.
1.4	26-Oct-2006	pooka	support specfs
1.3	25-Oct-2006	pooka	pass VOP_INACTIVE() to userspace
1.2	23-Oct-2006	pooka	bump the reqstruct minsize to something more believable (but I should really fix this, still)
1.1	22-Oct-2006	pooka	kernel portion of puffs - the Pass-to-Userspace Framework File System. It contains the VFS attachment and userspace message-passing interface. This work was initially started and completed for Google SoC 2005 and tweaked to work a bit better in the past few weeks. While being far from complete, it is functional enough to be able and stable to host a fairly general-purpose in-memory file system in userspace. Even so, puffs should be considered experimental and no binary compatibility for interfaces or crash-freedom or zero security implications should be relied upon just yet. The GSoC project was mentored by William Studenmund and the final review for the code was done by Christos.
1.8.2.4	01-Feb-2007	ad	Sync with head.
1.8.2.3	12-Jan-2007	ad	Sync with head.
1.8.2.2	18-Nov-2006	ad	Sync with head.
1.8.2.1	18-Nov-2006	ad	file puffs_msgif.h was added on branch newlock2 on 2006-11-18 21:39:20 +0000
1.12.2.2	10-Dec-2006	yamt	sync with head.
1.12.2.1	07-Dec-2006	yamt	file puffs_msgif.h was added on branch yamt-splraiseipl on 2006-12-10 07:18:38 +0000
1.13.2.9	04-Feb-2008	yamt	sync with head.
1.13.2.8	21-Jan-2008	yamt	sync with head
1.13.2.7	07-Dec-2007	yamt	sync with head
1.13.2.6	15-Nov-2007	yamt	sync with head.
1.13.2.5	27-Oct-2007	yamt	sync with head.
1.13.2.4	03-Sep-2007	yamt	sync with head.
1.13.2.3	26-Feb-2007	yamt	sync with head.
1.13.2.2	30-Dec-2006	yamt	sync with head.
1.13.2.1	29-Dec-2006	yamt	file puffs_msgif.h was added on branch yamt-lazymbuf on 2006-12-30 20:50:00 +0000
1.19.10.1	29-Mar-2007	reinoud	Pullup to -current
1.19.8.1	11-Jul-2007	mjf	Sync with head.
1.19.6.7	12-Oct-2007	ad	Sync with head.
1.19.6.6	09-Oct-2007	ad	Sync with head.
1.19.6.5	20-Aug-2007	ad	Sync with HEAD.
1.19.6.4	15-Jul-2007	ad	Sync with head.
1.19.6.3	09-Jun-2007	ad	Sync with head.
1.19.6.2	08-Jun-2007	ad	Sync with head.
1.19.6.1	10-Apr-2007	ad	Sync with head.
1.19.2.4	17-May-2007	yamt	sync with head.
1.19.2.3	07-May-2007	yamt	sync with head.
1.19.2.2	15-Apr-2007	yamt	sync with head.
1.19.2.1	24-Mar-2007	yamt	sync with head.
1.42.2.2	03-Sep-2007	skrll	Sync with HEAD.
1.42.2.1	15-Aug-2007	skrll	Sync with HEAD.
1.47.6.2	30-Jul-2007	pooka	Move PUFFS_TYPEPREFIX to puffs_msgif.h since it's used in a macro there.
1.47.6.1	30-Jul-2007	pooka	file puffs_msgif.h was added on branch matt-mips64 on 2007-07-30 09:04:59 +0000
1.47.4.8	09-Dec-2007	jmcneill	Sync with HEAD.
1.47.4.7	27-Nov-2007	joerg	Sync with HEAD. amd64 Xen support needs testing.
1.47.4.6	14-Nov-2007	joerg	Sync with HEAD.
1.47.4.5	26-Oct-2007	joerg	Sync with HEAD. Follow the merge of pmap.c on i386 and amd64 and move pmap_init_tmp_pgtbl into arch/x86/x86/pmap.c. Modify the ACPI wakeup code to restore CR4 before jumping back into kernel space as the large page option might cover that.
1.47.4.4	07-Oct-2007	joerg	Sync with HEAD.
1.47.4.3	02-Oct-2007	joerg	Sync with HEAD.
1.47.4.2	03-Sep-2007	jmcneill	Sync with HEAD.
1.47.4.1	16-Aug-2007	jmcneill	Sync with HEAD.
1.50.4.2	14-Oct-2007	yamt	sync with head.
1.50.4.1	06-Oct-2007	yamt	sync with head.
1.50.2.3	23-Mar-2008	matt	sync with HEAD
1.50.2.2	09-Jan-2008	matt	sync with HEAD
1.50.2.1	06-Nov-2007	matt	sync with HEAD
1.57.2.2	13-Nov-2007	bouyer	Sync with HEAD
1.57.2.1	25-Oct-2007	bouyer	Sync with HEAD.
1.59.2.4	18-Feb-2008	mjf	Sync with HEAD.
1.59.2.3	27-Dec-2007	mjf	Sync with HEAD.
1.59.2.2	08-Dec-2007	mjf	Sync with HEAD.
1.59.2.1	19-Nov-2007	mjf	Sync with HEAD.
1.61.2.2	26-Dec-2007	ad	Sync with head.
1.61.2.1	08-Dec-2007	ad	Sync with head.
1.65.26.1	21-Apr-2010	matt	sync to netbsd-5
1.65.20.3	17-Jul-2011	riz	Pull up following revision(s) (requested by manu in ticket #1645): lib/libc/sys/Makefile.inc 1.207 via patch lib/libc/sys/extattr_get_file.2 patch lib/libpuffs/dispatcher.c 1.34,1.36 via patch lib/libpuffs/puffs.c 1.107 via patch lib/libpuffs/puffs.h 1.115,1.118 via patch sys/fs/puffs/puffs_msgif.h 1.71,1.76 via patch sys/fs/puffs/puffs_vfsops.c 1.88 via patch sys/fs/puffs/puffs_vnops.c 1.145,1.154 via patch sys/kern/vfs_xattr.c 1.24-1.27 via patch sys/kern/vnode_if.c 1.87 via patch sys/sys/Makefile 1.133 via patch sys/sys/extattr.h 1.6 via patch sys/sys/vnode_if.h 1.81 via patch sys/ufs/ffs/ffs_vnops.c patch sys/ufs/ufs/ufs_extattr.c 1.31,1.34 via patch * support extended attributes * bump major due to structure growth * add some spare space * remove ABI sillyness Support extended attributes. Fix multiple non compliances in our Linux-like extattr API, and make it public so that it can be used. Improve a bit listxattr(2). It attemps to list both system and user extended attributes, and it faled if calling user did not have privilege for reading system EA. Now we just lise user EA and skip system EA in reading them is not allowed. Fix bug introduced in previous commuit: Do not vrele() a vnode we did not obtained. Improve UFS1 extended attributes usability - autocreate attribute backing file for new attributes - autoload attributes when issuing extattrctl start - when autoloading attributes, do not display garbage warning when looking up entries that got ENOENT Add a flag to VOP_LISTEXTATTR(9) so that the vnode interface can tell the filesystem in which format extended attribute shall be listed. There are currently two formats: - NUL-terminated strings, used for listxattr(2), this is the default. - one byte length-pprefixed, non NUL-terminated strings, used for extattr_list_file(2), which is obtanined by setting the EXTATTR_LIST_PREFIXLEN flag to VOP_LISTEXTATTR(9) This approach avoid the need for converting the list back and forth, except in libperfuse, since FUSE uses NUL-terminated strings, and the kernel may have requested EXTATTR_LIST_PREFIXLEN.
1.65.20.2	14-Dec-2009	sborrill	Revert previous version bump which should not have been in the supplied patch. This maintains compatibility between 5.0 and 5.1 (at the cost of needing userland libraries recompiled if one's been tracking netbsd-5).
1.65.20.1	18-Oct-2009	sborrill	Pull up the following revisions(s) (requested by pooka in ticket #1100): lib/libpuffs/dispatcher.c: revision 1.33 lib/libpuffs/puffs.c: revision 1.99 lib/libpuffs/puffs.h: revision 1.111 sys/fs/puffs/puffs_msgif.h: revision 1.67 via patch sys/fs/puffs/puffs_vnops.c: revision 1.136 Support VOP_ABORTOP() in puffs.
1.65.18.1	19-Jan-2009	skrll	Sync with HEAD.
1.65.10.3	11-Aug-2010	yamt	sync with head.
1.65.10.2	11-Mar-2010	yamt	sync with head
1.65.10.1	04-May-2009	yamt	sync with head.
1.65.6.1	17-Jan-2009	mjf	Sync with HEAD.
1.69.4.3	05-Mar-2011	rmind	sync with head
1.69.4.2	03-Jul-2010	rmind	sync with head
1.69.4.1	30-May-2010	rmind	sync with head
1.69.2.1	17-Aug-2010	uebayasi	Sync with HEAD.
1.77.8.4	27-Feb-2015	martin	Pull up following revision(s) (requested by manu in ticket #1260): lib/libpuffs/puffs.3: revision 1,55,1.60 sys/fs/puffs/puffs_msgif.h: revision 1.84 lib/libperfuse/ops.c: revision 1.83 sys/fs/puffs/puffs_sys.h: revision 1.89 sys/fs/puffs/puffs_vfsops.c: revision 1.116 lib/libperfuse/perfuse.c: revision 1.36 sys/fs/puffs/puffs_vnops.c: revision 1.200-1.202 Use more markup. New sentence, new line. Bump date for previous. Add PUFFS_KFLAG_NOFLUSH_META to prevent sending metadata flush to FUSE FUSE filesystems do not expect to get metadata updates for [amc]time and size, they updates the value on their own after operations. The PUFFS PUFFS_KFLAG_NOFLUSH_META option prevents regular metadata cache flushes to the filesystem , and libperfuse uses it to match Linux FUSE behavior. While there, fix a bug in SETATTR: do not update kernel metadata cache from SETATTR reply when the request is asynchronous, as we do not have the reply yet. Update file size after write without metadata flush If we do not use metadata flush, we must make sure the size is updated in the filesystem after a write, otherwise the next GETATTR will get us a stale value and the file will be truncated.
1.77.8.3	03-Nov-2014	msaitoh	Pull up following revision(s) (requested by manu in ticket #1140): lib/libperfuse/ops.c 1.63-1.69 lib/libperfuse/perfuse.c 1.32-1.33 lib/libperfuse/perfuse_priv.h 1.32-1.34 lib/libperfuse/subr.c 1.20 lib/libpuffs/creds.c 1.16 lib/libpuffs/dispatcher.c 1.47 lib/libpuffs/puffs.h 1.125 lib/libpuffs/puffs_ops.3 1.37-1.38 lib/libpuffs/requests.c 1.24 sys/fs/puffs/puffs_msgif.h 1.81 sys/fs/puffs/puffs_sys.h 1.85 sys/fs/puffs/puffs_vnops.c 1.183 usr.sbin/perfused/msg.c 1.22 Bring libpuffs, libperfuse and perfused on par with -current: - implement FUSE direct I/O - remove useless code and warnings - fix missing GETATTR bugs - fix exended attribute get and list operations
1.77.8.2	12-Aug-2012	martin	Pull up following revision(s) (requested by manu in ticket #438): lib/libperfuse/perfuse_priv.h: revision 1.31 sys/fs/puffs/puffs_msgif.h: revision 1.80 sys/fs/puffs/puffs_vnops.c: revision 1.171 lib/libpuffs/puffs_ops.3: revision 1.31 sys/fs/puffs/puffs_vnops.c: revision 1.172 sys/fs/puffs/puffs_vnops.c: revision 1.173 sys/fs/puffs/puffs_vnops.c: revision 1.174 usr.sbin/perfused/perfused.c: revision 1.24 sys/fs/puffs/puffs_sys.h: revision 1.80 sys/fs/puffs/puffs_sys.h: revision 1.81 sys/fs/puffs/puffs_sys.h: revision 1.82 lib/libperfuse/subr.c: revision 1.19 lib/libperfuse/perfuse.c: revision 1.30 sys/fs/puffs/puffs_msgif.c: revision 1.90 sys/fs/puffs/puffs_msgif.c: revision 1.91 sys/fs/puffs/puffs_msgif.c: revision 1.92 lib/libperfuse/ops.c: revision 1.59 lib/libpuffs/puffs.3: revision 1.53 lib/libperfuse/debug.c: revision 1.12 lib/libpuffs/puffs.3: revision 1.54 sys/fs/puffs/puffs_vnops.c: revision 1.167 sys/fs/puffs/puffs_msgif.h: revision 1.79 usr.sbin/perfused/msg.c: revision 1.21 sys/fs/puffs/puffs_vfsops.c: revision 1.102 sys/fs/puffs/puffs_vfsops.c: revision 1.103 sys/fs/puffs/puffs_vfsops.c: revision 1.105 lib/libpuffs/puffs.h: revision 1.123 lib/libperfuse/perfuse_if.h: revision 1.20 lib/libperfuse/perfuse.c: revision 1.29 lib/libpuffs/dispatcher.c: revision 1.42 lib/libpuffs/dispatcher.c: revision 1.43 - Fix same vnodes associated with multiple cookies The scheme used to retreive known nodes on lookup was flawed, as it only used parent and name. This produced a different cookie for the same file if it was renamed, when looking up ../ or when dealing with multiple files associated with the same name through link(2). We therefore abandon the use of node name and introduce hashed lists of inodes. This causes a huge rewrite of reclaim code, which do not attempt to keep parents allocated until all their children are reclaimed - Fix race conditions in reclaim There are a few situations where we issue multiple FUSE operations for a PUFFS operation. On reclaim, we therefore have to wait for all FUSE operation to complete, not just the current exchanges. We do this by introducing node reference count with node_ref() and node_rele(). - Detect data loss caused by FAF VOP_PUTPAGES causes FAF writes where the kernel does not check the operation result. At least issue a warning on error. - Enjoy FAF shortcut on setattr No need to wait for the result if the kernel does not want it. There is however an exception for setattr that touch the size, we need to wait for completion because we have other operations queued for after the resize. - Fix fchmod() on write-open file fchmod() on a node open with write privilege will send setattr with both mode and size set. This confuses some FUSE filesystem. Therefore we send two FUSE operations, one for mode, and one for size. - Remove node TTL handling for netbsd-5 for simplicity sake. The code still builds on netbsd-5 but does not have the node TTL feature anymore. It works fine with kernel support on netbsd-6. - Improve PUFFS_KFLAG_CACHE_FS_TTL by reclaiming older inactive nodes. The normal kernel behavior is to retain inactive nodes in the freelist until it runs out of vnodes. This has some merit for local filesystems, where the cost of an allocation is about the same as the cost of a lookup. But that situation is not true for distributed filesystems. On the other hand, keeping inactive nodes for a long time hold memory in the file server process, and when the kernel runs out of vnodes, it produce reclaim avalanches that increase lattency for other operations. We do not reclaim inactive vnodes immediatly either, as they may be looked up again shortly. Instead we introduce a grace time and we reclaim nodes that have been inactive beyond the grace time. - Fix lookup/reclaim race condition. The above improvement undercovered a race condition between lookup and reclaim. If we reclaimed a vnode associated with a userland cookie while a lookup returning that same cookiewas inprogress, then the kernel ends up with a vnode associated with a cookie that has been reclaimed in userland. Next operation on the cookie will crash (or at least confuse) the filesystem. We fix this by introducing a lookup count in kernel and userland. On reclaim, the kernel sends the count, which enable userland to detect situation where it initiated a lookup that is not completed in kernel. In such a situation, the reclaim must be ignored, as the node is about to be looked up again. Fix hang unmount bug introduced by last commit. We introduced a slow queue for delayed reclaims, while the existing queue for unmount, flush and exist has been renamed fast queue. Both queues had timestamp for when an operation should be done, but it was useless for the fast queue, which is always used to run an operation ASAP. And the timestamp test had an error that turned ASAP into "at next tick", but nobody what there to wake the thread at next tick, hence the hang. The fix is to remove the useless and buggy timestamp test for fast queue. Rename slow sopreq queue into node sopreq queue, to refet the fact that is only intended for postponed node reclaims. When purging the node sopreq queue, do not call puffs_msg_sendresp(), as it makes no sense. Fix race condition between (create\|mknod\|mkdir\|symlino) and reclaim, just like we did it between lookup and reclaim. Missing bit in previous commit (prevent race between create\|mknod\|mkdir\|symlink and reclaim) Bump date for previous. New sentence, new line; remove trailing whitespace; fix typos; punctuation nits. Add PUFFS_KFLAG_CACHE_DOTDOT so that vnodes hold a reference on their parent, keeping them active, and allowing to lookup .. without sending a request to the filesystem. Enable the featuure for perfused, as this is how FUSE works. Missing bit in previous commit (PUFFS_KFLAG_CACHE_DOTDOT option to avoid looking up ..)
1.77.8.1	23-Apr-2012	riz	Pull up following revision(s) (requested by manu in ticket #195): lib/libskey/skeysubr.c: revision 1.27 lib/libkvm/kvm_getloadavg.c: revision 1.11 lib/libwrap/update.c: revision 1.9 lib/liby/yyerror.c: revision 1.9 lib/libpuffs/puffs_ops.3: revision 1.30 lib/libwrap/misc.c: revision 1.10 lib/libwrap/hosts_access.c: revision 1.20 lib/libpuffs/pnode.c: revision 1.11 lib/libperfuse/subr.c: revision 1.17 lib/libpuffs/pnode.c: revision 1.12 lib/libperfuse/subr.c: revision 1.18 lib/libwrap/options.c: revision 1.15 lib/libwrap/fix_options.c: revision 1.11 lib/libperfuse/ops.c: revision 1.52 lib/libperfuse/ops.c: revision 1.53 lib/libperfuse/ops.c: revision 1.54 lib/libwrap/hosts_ctl.c: revision 1.5 lib/libintl/gettext.c: revision 1.27 lib/libwrap/shell_cmd.c: revision 1.6 lib/libpuffs/dispatcher.c: revision 1.39 lib/libperfuse/perfuse_priv.h: revision 1.27 lib/libwrap/socket.c: revision 1.19 lib/libpuffs/puffs.3: revision 1.50 lib/libperfuse/perfuse_priv.h: revision 1.28 lib/libpuffs/puffs_priv.h: revision 1.45 lib/libpuffs/puffs.3: revision 1.51 lib/libperfuse/perfuse_priv.h: revision 1.29 lib/libwrap/percent_x.c: revision 1.5 lib/libpuffs/puffs.3: revision 1.52 lib/libperfuse/debug.c: revision 1.11 sys/fs/puffs/puffs_vnops.c: revision 1.165 lib/libwrap/tcpd.h: revision 1.13 sys/fs/puffs/puffs_vnops.c: revision 1.166 lib/libwrap/eval.c: revision 1.7 sys/fs/puffs/puffs_msgif.h: revision 1.78 sys/fs/puffs/puffs_vfsops.c: revision 1.101 lib/libwrap/rfc931.c: revision 1.9 lib/libwrap/clean_exit.c: revision 1.5 lib/libpuffs/puffs.h: revision 1.120 lib/libc/stdlib/jemalloc.c: revision 1.27 lib/librmt/rmtlib.c: revision 1.26 lib/libpuffs/puffs.h: revision 1.121 sys/fs/puffs/puffs_sys.h: revision 1.79 lib/librumpclient/rumpclient.c: revision 1.48 lib/libwrap/refuse.c: revision 1.5 lib/libperfuse/perfuse.c: revision 1.26 lib/libperfuse/perfuse.c: revision 1.27 tests/fs/puffs/t_fuzz.c: revision 1.5 lib/libperfuse/perfuse.c: revision 1.28 lib/libpuffs/dispatcher.c: revision 1.40 sys/fs/puffs/puffs_node.c: revision 1.24 lib/libwrap/diag.c: revision 1.9 lib/libintl/textdomain.c: revision 1.13 Use C89 function definition Add name and atttribute cache with filesytem provided TTL. lookup, create, mknod, mkdir, symlink, getattr and setattr messages have been extended so that attributes and their TTL can be provided by the filesytem. lookup, create, mknod, mkdir, and symlink messages are also extended so that the filesystem can provide name TTL. Add PUFFS_KFLAG_CACHE_FS_TTL flag to puffs_init(3) to use name and attribute cache with filesystem provided TTL. lookup, create, mknod, mkdir, symlink, getattr and setattr messages have been extended so that attributes and their TTL can be provided by the filesytem. lookup, create, mknod, mkdir, and symlink messages are also extended so that the filesystem can provide name TTL. The filesystem updates attributes and TTL using puffs_pn_getvap(3), puffs_pn_getvattl(3), and puffs_pn_getcnttl(3) Use new PUFFS_KFLAG_CACHE_FS_TTL option to puffs_init(3) so that FUSE TTL on name and attributes are used. This save many PUFFS operations and improves performances. PUFFS_KFLAG_CACHE_FS_TTL is #ifdef'ed in many places for now so that libperfuse can still be used on netbsd-5. Split file system. Comma fixes. Remove dangling "and". Bump date for previous. - Makesure update_va does not change vnode size when it should not. For instance when doing a fault-issued VOP_GETPAGES within VOP_WRITE, changing size leads to panic: genfs_getpages: past eof. -Handle ticks wrap around for vnode name andattribute timeout - When using PUFFS_KFLAG_CACHE_FS_TTL, do not use puffs_node to carry attribute and TTL fora newly created node. Instead extend puffs_newinfo and add puffs_newinfo_setva() and puffs_newinfo_setttl() - Remove node_mk_common_final in libperfuse. It used to set uid/gid for a newly created vnode but has been made redundant along time ago since uid and gid are properly set in FUSE header. - In libperfuse, check for corner case where opc = 0 on INACTIVE and RECLAIM (how is it possible? Check for it to avoid a crash anyway) - In libperfuse, make sure we unlimit RLIMIT_AS and RLIMIT_DATA so that we do notrun out of memory because the kernel is lazy at reclaiming vnodes. - In libperfuse, cleanup style of perfuse_destroy_pn() Do not set PUFFS_KFLAG_CACHE_FS_TTL for PUFFS tests
1.77.6.1	29-Apr-2012	mrg	sync to latest -current.
1.77.2.2	30-Oct-2012	yamt	sync with head
1.77.2.1	17-Apr-2012	yamt	sync with head
1.80.14.3	27-Feb-2015	martin	Pull up following revision(s) (requested by manu in ticket #555): lib/libpuffs/puffs.3: revision 1.60 sys/fs/puffs/puffs_msgif.h: revision 1.84 lib/libperfuse/ops.c: revision 1.83 sys/fs/puffs/puffs_sys.h: revision 1.89 sys/fs/puffs/puffs_vfsops.c: revision 1.116 lib/libperfuse/perfuse.c: revision 1.36 sys/fs/puffs/puffs_vnops.c: revision 1.200-1.202 Add PUFFS_KFLAG_NOFLUSH_META to prevent sending metadata flush to FUSE FUSE filesystems do not expect to get metadata updates for [amc]time and size, they updates the value on their own after operations. The PUFFS PUFFS_KFLAG_NOFLUSH_META option prevents regular metadata cache flushes to the filesystem , and libperfuse uses it to match Linux FUSE behavior. While there, fix a bug in SETATTR: do not update kernel metadata cache from SETATTR reply when the request is asynchronous, as we do not have the reply yet. Update file size after write without metadata flush If we do not use metadata flush, we must make sure the size is updated in the filesystem after a write, otherwise the next GETATTR will get us a stale value and the file will be truncated.
1.80.14.2	05-Nov-2014	snj	Pull up following revision(s) (requested by manu in ticket #181): lib/libperfuse/fuse.h: revision 1.6 lib/libperfuse/ops.c: revision 1.78 lib/libperfuse/perfuse.c: revision 1.35 lib/libperfuse/perfuse_priv.h: revision 1.36 lib/libpuffs/dispatcher.c: revision 1.48 lib/libpuffs/opdump.c: revision 1.37 lib/libpuffs/puffs.c: revision 1.118 lib/libpuffs/puffs.h: revision 1.126 lib/libpuffs/puffs_ops.3: revisions 1.40-1.41 sys/fs/puffs/puffs_msgif.h: revision 1.82-1.83 sys/fs/puffs/puffs_msgif.h: revision 1.82 sys/fs/puffs/puffs_vnops.c: revision 1.196 Add PUFFS support for fallocate and fdiscard operations -- libpuffs support for fallocate and fdiscard operations -- Add PUFFS_HAVE_FALLOCATE in puffs_msgif.h so that filesystem can decide at build time wether fallocate is usable -- FUSE fallocate support There seems to be no fdiscard FUSE operation at the moment, hence that one is left unused.
1.80.14.1	26-Aug-2014	riz	Pull up following revision(s) (requested by manu in ticket #52): sys/fs/puffs/puffs_msgif.h: revision 1.81 sys/fs/puffs/puffs_sys.h: revision 1.85 sys/fs/puffs/puffs_vnops.c: revision 1.183 Add a oflags input field to open requests so that the filesystem can pass back information about the file. Implement PUFFS_OPEN_IO_DIRECT, which will force direct IO (bypassing page cache) for the file.
1.80.2.1	03-Dec-2017	jdolecek	update from HEAD
1.83.2.1	06-Apr-2015	skrll	Sync with HEAD
1.84.18.1	13-Apr-2020	martin	Mostly merge changes from HEAD upto 20200411
1.85.8.1	03-Apr-2021	thorpej	Sync with HEAD.
1.38	08-Feb-2018	dholland	Typos.
1.37	20-Aug-2016	hannken	Remove now obsolete operation vcache_remove(). Welcome to 7.99.36
1.36	10-Nov-2014	maxv	branches: 1.36.2; Do not uselessly include <sys/malloc.h>.
1.35	04-Nov-2014	manu	Fix PUFFS node use-after-reclaim When puffs_cookie2vnode() misses an entry, vcache_get() creates a new node (puffs_vfsop_loadvnode being called to initialize the PUFFS part), then it discovers it is VNON, and tries to vrele() it. vrele() calls VOP_INACTIVE(), which led us in puffs_vnop_inactive() where we sent a request to the filesystem for a node that already had been reclaimed. The fix is to check for VNON nodes in puffs_vnop_inactive() and to return without doing anyting. This is suboptimal, but a better workaround would probably need to modify vcache API, with an impact on other filesystems. Let us keep it simple.
1.34	30-Sep-2014	hannken	Fix the puffs_sop_thread -> puffs_cookie2vnode path: - pass the cookie by reference - add missing mutex_exit() - update assertion for VNON typed vnodes
1.33	05-Sep-2014	manu	When changing a directory content, update the ctime/mtime in kernel cache, otherwise the updated ctime/mtime appears after the cached entry expire.
1.32	28-Aug-2014	hannken	Change puffs from hashlist to vcache. - field "pa_nhashbuckets" of struct "puffs_kargs" becomes a no-op. and should be removed on the next protocol version bump.
1.31	23-Jan-2014	hannken	branches: 1.31.4; Change vnode operations create, mknod, mkdir and symlink to return the resulting vnode *vpp unlocked. Discussed on tech-kern@ Welcome to 6.99.30
1.30	17-Oct-2013	christos	- remove unused variables - add _NOERROR flavor macros for the case where errors are ignored.
1.29	06-Mar-2013	yamt	branches: 1.29.6; comments use sizeof(var) instead of sizeof(type) where possibly confusing
1.28	05-Nov-2012	dholland	Excise struct componentname from the namecache. This uglifies the interface, because several operations need to be passed the namei flags and cache_lookup also needs for the time being to be passed cnp->cn_nameiop. Nonetheless, it's a net benefit. The glop should be able to go away eventually but requires structural cleanup elsewhere first. This change requires a kernel bump.
1.27	23-Jul-2012	manu	branches: 1.27.2; Backout NCHNAMLEN check for cache_enter. That change collided with rmind's move of this exact check into cache_enter
1.26	23-Jul-2012	manu	Di not call cache_enter with path components bigger than NCHNAMLEN, as it panics the kernel.
1.25	22-Jul-2012	rmind	Move some the test for MAKEENTRY into the cache_enter(9). Make some variables in vfs_cache.c static, __read_mostly, etc. No objection on tech-kern@.
1.24	08-Apr-2012	manu	Add name and atttribute cache with filesytem provided TTL. lookup, create, mknod, mkdir, symlink, getattr and setattr messages have been extended so that attributes and their TTL can be provided by the filesytem. lookup, create, mknod, mkdir, and symlink messages are also extended so that the filesystem can provide name TTL.
1.23	19-Jan-2012	manu	branches: 1.23.2; Fix a race condition where the filesystem lookups a vnode that is being recycled, producing ENOENT while the file does exist. Approved by yamt
1.22	19-Oct-2011	manu	branches: 1.22.2; 1.22.6; Remove #ifdef DIAGNOSTIC guards around KASSERT, as the macro contains them
1.21	18-Oct-2011	manu	Make sure pagedaemon does not sleep for memory in puffs_vnop_sleep. Add KASSERT on any sleeping memory allocation to check it cannot happen again.
1.20	29-Aug-2011	manu	Add a mutex for operations that touch size (setattr, getattr, write, fsync). This is required to avoid data corruption bugs, where a getattr slices itself within a setattr operation, and sets the size to the stall value it got from the filesystem. That value is smaller than the one set by setattr, and the call to uvm_vnp_setsize() trigged a spurious truncate. The result is a chunk of zeroed data in the file. Such a situation can easily happen when the ioflush thread issue a VOP_FSYNC/puffs_vnop_sync/flushvncache/dosetattrn while andother process do a sys_stat/VOP_GETATTR/puffs_vnop_getattr. This mutex on size operation can be removed the day we decide VOP_GETATTR has to operated on a locked vnode, since the other operations that touch size already require that.
1.19	30-Jun-2011	wiz	dependant -> dependent
1.18	12-Jun-2011	rmind	Welcome to 5.99.53! Merge rmind-uvmplock branch: - Reorganize locking in UVM and provide extra serialisation for pmap(9). New lock order: [vmpage-owner-lock] -> pmap-lock. - Simplify locking in some pmap(9) modules by removing P->V locking. - Use lock object on vmobjlock (and thus vnode_t::v_interlock) to share the locks amongst UVM objects where necessary (tmpfs, layerfs, unionfs). - Rewrite and optimise x86 TLB shootdown code, make it simpler and cleaner. Add TLBSTATS option for x86 to collect statistics about TLB shootdowns. - Unify /dev/mem et al in MI code and provide required locking (removes kernel-lock on some ports). Also, avoid cache-aliasing issues. Thanks to Andrew Doran and Joerg Sonnenberger, as their initial patches formed the core changes of this branch.
1.17	25-Jul-2010	hannken	branches: 1.17.6; It makes no sense to call vget() with LK_RETRY.
1.16	21-Jul-2010	hannken	Make holding v_interlock mandatory for callers of vget(). Announced some time ago on tech-kern.
1.15	05-Nov-2009	pooka	branches: 1.15.2; 1.15.4; Kill suspend support. It was never implemented correctly: * it depended on the biglock (in a very cruel way) * it was attached to userspace transactions rather than logical fs operations (If someone wants to revisit it some day, most of the stuff can be reused from cvs history)
1.14	30-Sep-2009	pooka	* fix a race i introduced almost two years ago in rev 1.116: operations creating a node cannot be considered outgoing operations, since after return from userspace they modify file system state by creating a new node. if we do not protect the file system by holding the directory lock, a lookup operation might race us into the kernel and create the node earlier. * remove pnode from hashlish before sending the reclaim faf off to userspace. also, hold pmp_lock while frobbing the list.
1.13	06-May-2008	ad	branches: 1.13.10; 1.13.18; PR kern/37950 Unmounting psshfs immediately panics the machine puffs_getvnode() was inserting vnodes into mnt_vnodelist without taking a reference to the mount for each. When vnodes are scrubbed, refs to the vnodes mount structure are dropped => boom.
1.12	01-Mar-2008	rmind	branches: 1.12.2; 1.12.4; Welcome to 4.99.55: - Add a lot of missing selinit() and seldestroy() calls. - Merge selwakeup() and selnotify() calls into a single selnotify(). - Add an additional 'events' argument to selnotify() call. It will indicate which event (POLL_IN, POLL_OUT, etc) happen. If unknown, zero may be used. Note: please pass appropriate value of 'events' where possible. Proposed on: <tech-kern>
1.11	28-Jan-2008	pooka	branches: 1.11.2; 1.11.6; For code clarity typedef void *puffs_cookie_t. No functional change.
1.10	24-Jan-2008	ad	specfs changes for PR kern/37717 (raidclose() is no longer called on shutdown). There are still problems with device access and a PR will be filed. - Kill checkalias(). Allow multiple vnodes to reference a single device. - Don't play dangerous tricks with block vnodes to ensure that only one vnode can describe a block device. Instead, prohibit concurrent opens of block devices. As a bonus remove the unreliable code that prevents multiple file system mounts on the same device. It's no longer needed. - Track opens by vnode and by device. Issue cdev_close() when the last open goes away, instead of abusing vnode::v_usecount to tell if the device is open.
1.9	02-Jan-2008	ad	Merge vmlocking2 to head.
1.8	17-Nov-2007	pooka	branches: 1.8.2; 1.8.6; Make puffs_updatenode() take a puffs_node instead of a vnode. This way we don't need to worry if a vnode has been reclaimed from under us.
1.7	16-Nov-2007	pooka	Restructure the messaging interface a bit more: make all interfacing with the file server happen through puffs_msg_enqueue() and puffs_msg_wait() instead of having a billion different routines. Build the existing system upon these two. Most importantly though, decouple insertation into the op queue from the actual wait. This is useful for a number of reasons coming soon to a cvs repo near you.
1.6	11-Oct-2007	pooka	branches: 1.6.2; 1.6.4; 1.6.6; 1.6.8; Part 1/n of some pretty extensive changes to how the kernel module interacts with the userspace file server: * since the kernel-user communication is not purely request-response anymore (hasn't been since 2006), try to rename some "request" to "message". more similar mangling will take place in the future. * completely rework how messages are allocated. previously most of them were borrowed from the stack (originally all of them), but now always allocate dynamically. this makes the structure of the code much cleaner. also makes it possible to fix a locking order violation. it enables plenty of future enhancements. * start generalizing the transport interface to be independent of puffs * move transport interface to read/write instead of ioctl. the old one had legacy design problems, and besides, ioctl's suck. implement a very generic version for now; this will be worked on later hopefully some day reaching "highly optimized". * implement libpuffs support behind existing library request interfaces. this will change eventually (I hate those interfaces)
1.5	10-Oct-2007	ad	Merge from vmlocking: - Split vnode::v_flag into three fields, depending on field locking. - simple_lock -> kmutex in a few places. - Fix some simple locking problems.
1.4	02-Oct-2007	pooka	branches: 1.4.2; 1.4.4; 1.4.6; If kernel resource allocation fails after the file server has committed something, issue an abort. The abort is done through the regular op channel, e.g. failed mkdir leads to regular rmdir, inactive and reclaim. No internal interface is planned currently for the one file system out of a million which would implement it to benefit from the one case in a billion where kernel resource allocation actually does fail and out of that one case in a trillion where internal vs. external would make a difference.
1.3	01-Oct-2007	pooka	* better error checking: validate error values received from userland to be vaild errno values * include string describing error in PUFFS_ERR * get rid of union in puffs_req, it's nothing but trouble * pass pmp to async i/o callbacks
1.2	27-Sep-2007	pooka	comments & other minor maintenance
1.1	27-Sep-2007	pooka	Split routines handling nodes from puffs_subr to puffs_node. No functional change.
1.4.6.6	28-Oct-2007	ad	Fix up mnt_vnodelist handling.
1.4.6.5	23-Oct-2007	ad	Sync with head.
1.4.6.4	12-Oct-2007	ad	Sync with head.
1.4.6.3	09-Oct-2007	ad	Sync with head.
1.4.6.2	09-Oct-2007	ad	Sync with head.
1.4.6.1	02-Oct-2007	ad	file puffs_node.c was added on branch vmlocking on 2007-10-09 13:44:18 +0000
1.4.4.3	14-Oct-2007	yamt	sync with head.
1.4.4.2	06-Oct-2007	yamt	sync with head.
1.4.4.1	02-Oct-2007	yamt	file puffs_node.c was added on branch yamt-x86pmap on 2007-10-06 15:29:48 +0000
1.4.2.4	21-Nov-2007	joerg	Sync with HEAD.
1.4.2.3	26-Oct-2007	joerg	Sync with HEAD. Follow the merge of pmap.c on i386 and amd64 and move pmap_init_tmp_pgtbl into arch/x86/x86/pmap.c. Modify the ACPI wakeup code to restore CR4 before jumping back into kernel space as the large page option might cover that.
1.4.2.2	02-Oct-2007	joerg	Sync with HEAD.
1.4.2.1	02-Oct-2007	joerg	file puffs_node.c was added on branch jmcneill-pm on 2007-10-02 18:28:52 +0000
1.6.8.4	23-Mar-2008	matt	sync with HEAD
1.6.8.3	09-Jan-2008	matt	sync with HEAD
1.6.8.2	06-Nov-2007	matt	sync with HEAD
1.6.8.1	11-Oct-2007	matt	file puffs_node.c was added on branch matt-armv6 on 2007-11-06 23:31:15 +0000
1.6.6.2	18-Feb-2008	mjf	Sync with HEAD.
1.6.6.1	19-Nov-2007	mjf	Sync with HEAD.
1.6.4.6	17-Mar-2008	yamt	sync with head.
1.6.4.5	04-Feb-2008	yamt	sync with head.
1.6.4.4	21-Jan-2008	yamt	sync with head
1.6.4.3	07-Dec-2007	yamt	sync with head
1.6.4.2	27-Oct-2007	yamt	sync with head.
1.6.4.1	11-Oct-2007	yamt	file puffs_node.c was added on branch yamt-lazymbuf on 2007-10-27 11:35:10 +0000
1.6.2.1	18-Nov-2007	bouyer	Sync with HEAD
1.8.6.1	02-Jan-2008	bouyer	Sync with HEAD
1.8.2.1	04-Dec-2007	ad	Pull the vmlocking changes into a new branch.
1.11.6.2	02-Jun-2008	mjf	Sync with HEAD.
1.11.6.1	03-Apr-2008	mjf	Sync with HEAD.
1.11.2.1	24-Mar-2008	keiichi	sync with head.
1.12.4.3	11-Aug-2010	yamt	sync with head.
1.12.4.2	11-Mar-2010	yamt	sync with head
1.12.4.1	16-May-2008	yamt	sync with head.
1.12.2.1	18-May-2008	yamt	sync with head.
1.13.18.1	21-Apr-2010	matt	sync to netbsd-5
1.13.10.4	25-Jan-2012	riz	Pull up following revision(s) (requested by manu in ticket #1714): sys/fs/puffs/puffs_node.c: revision 1.23 Fix a race condition where the filesystem lookups a vnode that is being recycled, producing ENOENT while the file does exist. Approved by yamt
1.13.10.3	02-Nov-2011	riz	Pull up following revision(s) (requested by manu in ticket #1679): sys/fs/puffs/puffs_vnops.c: revision 1.157 sys/fs/puffs/puffs_vnops.c: revision 1.158 sys/fs/puffs/puffs_vnops.c: revision 1.159 sys/fs/puffs/puffs_vfsops.c: revision 1.97 sys/fs/puffs/puffs_vfsops.c: revision 1.99 sys/fs/puffs/puffs_vnops.c: revision 1.160 sys/fs/puffs/puffs_vfsops.c: revision 1.100 sys/miscfs/syncfs/sync_subr.c: revision 1.47 sys/fs/puffs/puffs_node.c: revision 1.21 sys/fs/puffs/puffs_node.c: revision 1.22 sys/fs/puffs/puffs_msgif.c: revision 1.88 sys/fs/puffs/puffs_msgif.c: revision 1.89 sys/fs/puffs/puffs_vnops.c: revision 1.156 Make sure ioflush does not sleep in PUFFS code path, waiting for a mutex, a memory allocation, or a response from the filesystem. This avoids deadlocks in the following situations: 1) when memory is low: ioflush waits the fileystem, the fielsystem waits for memory 2) when the filesystem does not respond (e.g.: network outage ona distributed filesystem) Fix the build that was broken by struct lwp *updateproc reference in RUMP-visible code. Instead of checking that updateproc (aka ioflush, aka syncer) will not sleep in PUFFS code, I check for any kernel thread: after all none of them are designed to hang awaiting for a remote filesystem operation to complete. Roll back the change that forced kernel threads to not sleep in PUFFS. The change does not make consensus, since only pagedaemon should need it. Other threads will tolerate sleeping, and problems here are only symptoms that something is going wrong in memory management. The cause, not the symptoms, need to be fixed. Make sure pagedaemon does not sleep for memory in puffs_vnop_sleep. Add KASSERT on any sleeping memory allocation to check it cannot happen again. Remove #ifdef DIAGNOSTIC guards around KASSERT, as the macro contains them
1.13.10.2	17-Sep-2011	bouyer	Pull up following revision(s) (requested by manu in ticket #1666): sys/fs/puffs/puffs_sys.h: revision 1.78 via patch sys/fs/puffs/puffs_node.c: revision 1.20 via patch sys/fs/puffs/puffs_vnops.c: revision 1.155 via patch Add a mutex for operations that touch size (setattr, getattr, write, fsync). This is required to avoid data corruption bugs, where a getattr slices itself within a setattr operation, and sets the size to the stall value it got from the filesystem. That value is smaller than the one set by setattr, and the call to uvm_vnp_setsize() trigged a spurious truncate. The result is a chunk of zeroed data in the file. Such a situation can easily happen when the ioflush thread issue a VOP_FSYNC/puffs_vnop_sync/flushvncache/dosetattrn while andother process do a sys_stat/VOP_GETATTR/puffs_vnop_getattr. This mutex on size operation can be removed the day we decide VOP_GETATTR has to operated on a locked vnode, since the other operations that touch size already require that.
1.13.10.1	03-Oct-2009	snj	Pull up following revision(s) (requested by pooka in ticket #1042): sys/fs/puffs/puffs_node.c: revision 1.14 sys/fs/puffs/puffs_vnops.c: revision 1.134 * fix a race i introduced almost two years ago in rev 1.116: operations creating a node cannot be considered outgoing operations, since after return from userspace they modify file system state by creating a new node. if we do not protect the file system by holding the directory lock, a lookup operation might race us into the kernel and create the node earlier. * remove pnode from hashlish before sending the reclaim faf off to userspace. also, hold pmp_lock while frobbing the list.
1.15.4.3	19-May-2011	rmind	Implement sharing of vnode_t::v_interlock amongst vnodes: - Lock is shared amongst UVM objects using uvm_obj_setlock() or getnewvnode(). - Adjust vnode cache to handle unsharing, add VI_LOCKSHARE flag for that. - Use sharing in tmpfs and layerfs for underlying object. - Simplify locking in ubc_fault(). - Sprinkle some asserts. Discussed with ad@.
1.15.4.2	05-Mar-2011	rmind	sync with head
1.15.4.1	16-Mar-2010	rmind	Change struct uvm_object::vmobjlock to be dynamically allocated with mutex_obj_alloc(). It allows us to share the locks among UVM objects.
1.15.2.1	17-Aug-2010	uebayasi	Sync with HEAD.
1.17.6.1	23-Jun-2011	cherry	Catchup with rmind-uvmplock merge.
1.22.6.2	29-Apr-2012	mrg	sync to latest -current.
1.22.6.1	18-Feb-2012	mrg	merge to -current.
1.22.2.4	22-May-2014	yamt	sync with head. for a reference, the tree before this commit was tagged as yamt-pagecache-tag8. this commit was splitted into small chunks to avoid a limitation of cvs. ("Protocol error: too many arguments")
1.22.2.3	16-Jan-2013	yamt	sync with (a bit old) head
1.22.2.2	30-Oct-2012	yamt	sync with head
1.22.2.1	17-Apr-2012	yamt	sync with head
1.23.2.3	03-Nov-2014	msaitoh	Pull up following revision(s) (requested by manu in ticket #1149): sys/fs/puffs/puffs_node.c: revision 1.33 sys/fs/puffs/puffs_vnops.c: revision 1.185 When changing a directory content, update the ctime/mtime in kernel cache, otherwise the updated ctime/mtime appears after the cached entry expire.
1.23.2.2	12-Aug-2012	martin	Pull up following revision(s) (requested by manu in ticket #484): sys/fs/nilfs/nilfs_vnops.c: revision 1.18 sys/ufs/ufs/ufs_lookup.c: revision 1.117 sys/nfs/nfs_vnops.c: revision 1.295 sys/ufs/chfs/chfs_vnops.c: revision 1.8 sys/ufs/ext2fs/ext2fs_lookup.c: revision 1.70 sys/fs/unionfs/unionfs_vnops.c: revision 1.6 sys/kern/vfs_cache.c: revision 1.89 sys/fs/efs/efs_vnops.c: revision 1.26 sys/fs/hfs/hfs_vnops.c: revision 1.26 sys/fs/adosfs/adlookup.c: revision 1.16 sys/fs/puffs/puffs_vnops.c: revision 1.168 sys/fs/tmpfs/tmpfs_vnops.c: revision 1.98 sys/fs/ntfs/ntfs_vnops.c: revision 1.52 sys/fs/cd9660/cd9660_lookup.c: revision 1.20 sys/fs/msdosfs/msdosfs_lookup.c: revision 1.24 sys/fs/smbfs/smbfs_vnops.c: revision 1.80 sys/fs/udf/udf_vnops.c: revision 1.72 sys/fs/filecorefs/filecore_lookup.c: revision 1.14 sys/fs/puffs/puffs_node.c: revision 1.25 Move some the test for MAKEENTRY into the cache_enter(9). Make some variables in vfs_cache.c static, __read_mostly, etc. No objection on tech-kern@.
1.23.2.1	23-Apr-2012	riz	Pull up following revision(s) (requested by manu in ticket #195): lib/libskey/skeysubr.c: revision 1.27 lib/libkvm/kvm_getloadavg.c: revision 1.11 lib/libwrap/update.c: revision 1.9 lib/liby/yyerror.c: revision 1.9 lib/libpuffs/puffs_ops.3: revision 1.30 lib/libwrap/misc.c: revision 1.10 lib/libwrap/hosts_access.c: revision 1.20 lib/libpuffs/pnode.c: revision 1.11 lib/libperfuse/subr.c: revision 1.17 lib/libpuffs/pnode.c: revision 1.12 lib/libperfuse/subr.c: revision 1.18 lib/libwrap/options.c: revision 1.15 lib/libwrap/fix_options.c: revision 1.11 lib/libperfuse/ops.c: revision 1.52 lib/libperfuse/ops.c: revision 1.53 lib/libperfuse/ops.c: revision 1.54 lib/libwrap/hosts_ctl.c: revision 1.5 lib/libintl/gettext.c: revision 1.27 lib/libwrap/shell_cmd.c: revision 1.6 lib/libpuffs/dispatcher.c: revision 1.39 lib/libperfuse/perfuse_priv.h: revision 1.27 lib/libwrap/socket.c: revision 1.19 lib/libpuffs/puffs.3: revision 1.50 lib/libperfuse/perfuse_priv.h: revision 1.28 lib/libpuffs/puffs_priv.h: revision 1.45 lib/libpuffs/puffs.3: revision 1.51 lib/libperfuse/perfuse_priv.h: revision 1.29 lib/libwrap/percent_x.c: revision 1.5 lib/libpuffs/puffs.3: revision 1.52 lib/libperfuse/debug.c: revision 1.11 sys/fs/puffs/puffs_vnops.c: revision 1.165 lib/libwrap/tcpd.h: revision 1.13 sys/fs/puffs/puffs_vnops.c: revision 1.166 lib/libwrap/eval.c: revision 1.7 sys/fs/puffs/puffs_msgif.h: revision 1.78 sys/fs/puffs/puffs_vfsops.c: revision 1.101 lib/libwrap/rfc931.c: revision 1.9 lib/libwrap/clean_exit.c: revision 1.5 lib/libpuffs/puffs.h: revision 1.120 lib/libc/stdlib/jemalloc.c: revision 1.27 lib/librmt/rmtlib.c: revision 1.26 lib/libpuffs/puffs.h: revision 1.121 sys/fs/puffs/puffs_sys.h: revision 1.79 lib/librumpclient/rumpclient.c: revision 1.48 lib/libwrap/refuse.c: revision 1.5 lib/libperfuse/perfuse.c: revision 1.26 lib/libperfuse/perfuse.c: revision 1.27 tests/fs/puffs/t_fuzz.c: revision 1.5 lib/libperfuse/perfuse.c: revision 1.28 lib/libpuffs/dispatcher.c: revision 1.40 sys/fs/puffs/puffs_node.c: revision 1.24 lib/libwrap/diag.c: revision 1.9 lib/libintl/textdomain.c: revision 1.13 Use C89 function definition Add name and atttribute cache with filesytem provided TTL. lookup, create, mknod, mkdir, symlink, getattr and setattr messages have been extended so that attributes and their TTL can be provided by the filesytem. lookup, create, mknod, mkdir, and symlink messages are also extended so that the filesystem can provide name TTL. Add PUFFS_KFLAG_CACHE_FS_TTL flag to puffs_init(3) to use name and attribute cache with filesystem provided TTL. lookup, create, mknod, mkdir, symlink, getattr and setattr messages have been extended so that attributes and their TTL can be provided by the filesytem. lookup, create, mknod, mkdir, and symlink messages are also extended so that the filesystem can provide name TTL. The filesystem updates attributes and TTL using puffs_pn_getvap(3), puffs_pn_getvattl(3), and puffs_pn_getcnttl(3) Use new PUFFS_KFLAG_CACHE_FS_TTL option to puffs_init(3) so that FUSE TTL on name and attributes are used. This save many PUFFS operations and improves performances. PUFFS_KFLAG_CACHE_FS_TTL is #ifdef'ed in many places for now so that libperfuse can still be used on netbsd-5. Split file system. Comma fixes. Remove dangling "and". Bump date for previous. - Makesure update_va does not change vnode size when it should not. For instance when doing a fault-issued VOP_GETPAGES within VOP_WRITE, changing size leads to panic: genfs_getpages: past eof. -Handle ticks wrap around for vnode name andattribute timeout - When using PUFFS_KFLAG_CACHE_FS_TTL, do not use puffs_node to carry attribute and TTL fora newly created node. Instead extend puffs_newinfo and add puffs_newinfo_setva() and puffs_newinfo_setttl() - Remove node_mk_common_final in libperfuse. It used to set uid/gid for a newly created vnode but has been made redundant along time ago since uid and gid are properly set in FUSE header. - In libperfuse, check for corner case where opc = 0 on INACTIVE and RECLAIM (how is it possible? Check for it to avoid a crash anyway) - In libperfuse, make sure we unlimit RLIMIT_AS and RLIMIT_DATA so that we do notrun out of memory because the kernel is lazy at reclaiming vnodes. - In libperfuse, cleanup style of perfuse_destroy_pn() Do not set PUFFS_KFLAG_CACHE_FS_TTL for PUFFS tests
1.27.2.4	03-Dec-2017	jdolecek	update from HEAD
1.27.2.3	20-Aug-2014	tls	Rebase to HEAD as of a few days ago.
1.27.2.2	23-Jun-2013	tls	resync from head
1.27.2.1	20-Nov-2012	tls	Resync to 2012-11-19 00:00:00 UTC
1.29.6.1	18-May-2014	rmind	sync with head
1.31.4.5	17-Jan-2015	martin	Pull up following revision(s) (requested by maxv in ticket #427): sys/compat/svr4/svr4_schedctl.c: revision 1.8 sys/netinet/tcp_timer.c: revision 1.88 sys/miscfs/genfs/layer_vfsops.c: revision 1.45 sys/compat/svr4/svr4_ioctl.c: revision 1.37 sys/ufs/chfs/chfs_vfsops.c: revision 1.14 sys/miscfs/fdesc/fdesc_vfsops.c: revision 1.91 sys/compat/linux/arch/i386/linux_ptrace.c: revision 1.30 sys/compat/common/kern_time_50.c: revision 1.28 sys/netinet6/ip6_forward.c: revision 1.74 sys/miscfs/umapfs/umap_vnops.c: revision 1.57 sys/compat/svr4/svr4_fcntl.c: revision 1.74 distrib/sets/lists/comp/mi: revision 1.1931 sys/netinet6/udp6_output.c: revision 1.46 sys/fs/puffs/puffs_compat.c: revision 1.3 sys/fs/udf/udf_rename.c: revision 1.11 sys/compat/svr4/svr4_filio.c: revision 1.24 sys/fs/udf/udf_rename.c: revision 1.12 sys/netinet/tcp_usrreq.c: revision 1.202 sys/miscfs/umapfs/umap_subr.c: revision 1.29 sys/compat/linux/common/linux_fadvise64.c: revision 1.3 sys/netinet/if_atm.c: revision 1.34 sys/miscfs/procfs/procfs_subr.c: revision 1.106 sys/miscfs/genfs/layer_subr.c: revision 1.37 sys/netinet/tcp_sack.c: revision 1.30 sys/compat/freebsd/freebsd_misc.c: revision 1.33 sys/compat/freebsd/freebsd_file.c: revision 1.33 sys/ufs/chfs/chfs_vnode.c: revision 1.12 sys/compat/svr4/svr4_ttold.c: revision 1.34 sys/compat/linux/common/linux_file.c: revision 1.114 sys/compat/linux/arch/mips/linux_machdep.c: revision 1.43 sys/compat/linux/common/linux_signal.c: revision 1.76 sys/compat/common/compat_util.c: revision 1.46 sys/compat/linux/arch/arm/linux_ptrace.c: revision 1.18 sys/compat/svr4/svr4_sockio.c: revision 1.36 sys/compat/linux/arch/arm/linux_machdep.c: revision 1.32 sys/compat/svr4/svr4_signal.c: revision 1.66 sys/kern/kern_exec.c: revision 1.410 sys/fs/puffs/puffs_vfsops.c: revision 1.115 sys/compat/svr4/svr4_exec_elf64.c: revision 1.15 sys/compat/linux/arch/i386/linux_machdep.c: revision 1.159 sys/compat/linux/arch/alpha/linux_machdep.c: revision 1.50 sys/compat/linux32/common/linux32_misc.c: revision 1.24 sys/netinet/in_pcb.c: revision 1.153 sys/sys/malloc.h: revision 1.116 sys/compat/common/if_43.c: revision 1.9 share/man/man9/Makefile: revision 1.380 sys/netinet/tcp_vtw.c: revision 1.12 sys/miscfs/umapfs/umap_vfsops.c: revision 1.95 sys/ufs/ext2fs/ext2fs_vfsops.c: revision 1.186 sys/compat/common/uipc_syscalls_43.c: revision 1.46 sys/ufs/ext2fs/ext2fs_vnops.c: revision 1.115 sys/fs/puffs/puffs_msgif.c: revision 1.97 sys/compat/svr4/svr4_ipc.c: revision 1.27 sys/compat/linux/common/linux_exec.c: revision 1.117 sys/ufs/ext2fs/ext2fs_readwrite.c: revision 1.66 sys/netinet/tcp_output.c: revision 1.179 sys/compat/svr4/svr4_termios.c: revision 1.28 sys/fs/udf/udf_strat_bootstrap.c: revision 1.4 sys/fs/puffs/puffs_subr.c: revision 1.67 sys/fs/puffs/puffs_node.c: revision 1.36 sys/miscfs/overlay/overlay_vnops.c: revision 1.21 sys/fs/cd9660/cd9660_node.c: revision 1.34 sys/netinet/raw_ip.c: revision 1.146 sys/sys/mallocvar.h: revision 1.13 sys/miscfs/overlay/overlay_vfsops.c: revision 1.63 share/man/man9/malloc.9: revision 1.50 sys/netinet6/dest6.c: revision 1.18 sys/compat/linux/common/linux_uselib.c: revision 1.33 sys/compat/linux/common/linux_socket.c: revision 1.120 share/man/man9/malloc.9: revision 1.51 sys/netinet/tcp_subr.c: revision 1.257 sys/compat/linux/common/linux_socketcall.c: revision 1.45 sys/compat/linux/common/linux_fadvise64_64.c: revision 1.3 sys/compat/freebsd/freebsd_ipc.c: revision 1.17 sys/compat/linux/common/linux_misc_notalpha.c: revision 1.109 sys/compat/linux/arch/alpha/linux_pipe.c: revision 1.17 sys/netinet6/in6_pcb.c: revision 1.132 sys/netinet6/in6_ifattach.c: revision 1.94 sys/compat/svr4/svr4_exec_elf32.c: revision 1.15 sys/miscfs/nullfs/null_vfsops.c: revision 1.90 sys/fs/cd9660/cd9660_util.c: revision 1.12 sys/compat/linux/arch/powerpc/linux_machdep.c: revision 1.48 sys/compat/freebsd/freebsd_exec_elf32.c: revision 1.20 sys/miscfs/procfs/procfs_vfsops.c: revision 1.94 sys/compat/linux/arch/powerpc/linux_ptrace.c: revision 1.28 sys/compat/linux/common/linux_sched.c: revision 1.67 sys/compat/linux/common/linux_exec_aout.c: revision 1.67 sys/compat/linux/common/linux_pipe.c: revision 1.67 sys/compat/linux/common/linux_llseek.c: revision 1.34 sys/compat/linux/arch/mips/linux_ptrace.c: revision 1.10 Do not uselessly include <sys/malloc.h>. Cleanup: - remove struct kmembuckets (dead) - correctly deadify MALLOC_XX - remove MALLOC_DEFINE_LIMIT and MALLOC_JUSTDEFINE_LIMIT (dead) - remove malloc_roundup(), malloc_type_setlimit(), MALLOC_DEFINE_LIMIT() and MALLOC_JUSTDEFINE_LIMIT() from man 9 malloc New sentence, new line. Bump date for previous. Obsolete malloc_roundup(9), malloc_type_setlimit(9) and MALLOC_DEFINE_LIMIT(9) man pages.
1.31.4.4	09-Nov-2014	msaitoh	Pull up following revision(s) (requested by manu in ticket #194): sys/fs/puffs/puffs_vnops.c: revision 1.197 sys/fs/puffs/puffs_node.c: revision 1.35 Fix PUFFS node use-after-reclaim When puffs_cookie2vnode() misses an entry, vcache_get() creates a new node (puffs_vfsop_loadvnode being called to initialize the PUFFS part), then it discovers it is VNON, and tries to vrele() it. vrele() calls VOP_INACTIVE(), which led us in puffs_vnop_inactive() where we sent a request to the filesystem for a node that already had been reclaimed. The fix is to check for VNON nodes in puffs_vnop_inactive() and to return without doing anyting. This is suboptimal, but a better workaround would probably need to modify vcache API, with an impact on other filesystems. Let us keep it simple.
1.31.4.3	30-Sep-2014	martin	Pull up following revision(s) (requested by hannken in ticket #67): sys/fs/puffs/puffs_node.c: revision 1.34 sys/fs/puffs/puffs_vnops.c: revision 1.187 Fix the puffs_sop_thread -> puffs_cookie2vnode path: - pass the cookie by reference - add missing mutex_exit() - update assertion for VNON typed vnodes
1.31.4.2	10-Sep-2014	martin	Pull up following revision(s) (requested by manu in ticket #79): sys/fs/puffs/puffs_node.c: revision 1.33 sys/fs/puffs/puffs_vnops.c: revision 1.185 When changing a directory content, update the ctime/mtime in kernel cache, otherwise the updated ctime/mtime appears after the cached entry expire.
1.31.4.1	29-Aug-2014	martin	Pull up following revision(s) (requested by hannken in ticket #67): sys/fs/puffs/puffs_sys.h: revision 1.86 sys/fs/puffs/puffs_vfsops.c: revision 1.114 sys/fs/puffs/puffs_msgif.c: revision 1.95 sys/fs/puffs/puffs_node.c: revision 1.32 sys/fs/puffs/puffs_vnops.c: revision 1.184 Change puffs from hashlist to vcache. - field "pa_nhashbuckets" of struct "puffs_kargs" becomes a no-op. and should be removed on the next protocol version bump.
1.36.2.1	05-Oct-2016	skrll	Sync with HEAD
1.67	10-Nov-2014	maxv	Do not uselessly include <sys/malloc.h>.
1.66	16-Nov-2008	pooka	branches: 1.66.26; 1.66.42; more <sys/buf.h> police
1.65	01-Mar-2008	rmind	branches: 1.65.4; 1.65.10; 1.65.12; Welcome to 4.99.55: - Add a lot of missing selinit() and seldestroy() calls. - Merge selwakeup() and selnotify() calls into a single selnotify(). - Add an additional 'events' argument to selnotify() call. It will indicate which event (POLL_IN, POLL_OUT, etc) happen. If unknown, zero may be used. Note: please pass appropriate value of 'events' where possible. Proposed on: <tech-kern>
1.64	28-Jan-2008	pooka	branches: 1.64.2; 1.64.6; For code clarity typedef void *puffs_cookie_t. No functional change.
1.63	02-Jan-2008	pooka	More type-punning workarounds. Curiously the kernel compilation flags cause gcc to not complain.
1.62	08-Dec-2007	pooka	branches: 1.62.4; Now that "l" is gone both as an argument to operations and from componentname, remove all vestiges of puffs_cid.
1.61	08-Dec-2007	pooka	Remove cn_lwp from struct componentname. curlwp should be used from on. The NDINIT() macro no longer takes the lwp parameter and associates the credentials of the calling thread with the namei structure.
1.60	17-Nov-2007	pooka	branches: 1.60.2; Make puffs_updatenode() take a puffs_node instead of a vnode. This way we don't need to worry if a vnode has been reclaimed from under us.
1.59	17-Nov-2007	pooka	Implement a biodone callback for async writes similar to reads and use that when possible.
1.58	16-Nov-2007	pooka	Restructure the messaging interface a bit more: make all interfacing with the file server happen through puffs_msg_enqueue() and puffs_msg_wait() instead of having a billion different routines. Build the existing system upon these two. Most importantly though, decouple insertation into the op queue from the actual wait. This is useful for a number of reasons coming soon to a cvs repo near you.
1.57	11-Oct-2007	pooka	branches: 1.57.2; 1.57.4; Part 1/n of some pretty extensive changes to how the kernel module interacts with the userspace file server: * since the kernel-user communication is not purely request-response anymore (hasn't been since 2006), try to rename some "request" to "message". more similar mangling will take place in the future. * completely rework how messages are allocated. previously most of them were borrowed from the stack (originally all of them), but now always allocate dynamically. this makes the structure of the code much cleaner. also makes it possible to fix a locking order violation. it enables plenty of future enhancements. * start generalizing the transport interface to be independent of puffs * move transport interface to read/write instead of ioctl. the old one had legacy design problems, and besides, ioctl's suck. implement a very generic version for now; this will be worked on later hopefully some day reaching "highly optimized". * implement libpuffs support behind existing library request interfaces. this will change eventually (I hate those interfaces)
1.56	10-Oct-2007	ad	Merge from vmlocking: - Split vnode::v_flag into three fields, depending on field locking. - simple_lock -> kmutex in a few places. - Fix some simple locking problems.
1.55	01-Oct-2007	pooka	* better error checking: validate error values received from userland to be vaild errno values * include string describing error in PUFFS_ERR * get rid of union in puffs_req, it's nothing but trouble * pass pmp to async i/o callbacks
1.54	29-Sep-2007	pooka	kill trailing whitespace
1.53	27-Sep-2007	pooka	Split routines handling nodes from puffs_subr to puffs_node. No functional change.
1.52	27-Sep-2007	pooka	Revert previous, it makes no sense whatsoever.
1.51	27-Sep-2007	pooka	Undo state created in cookie2vnode if an error is returned.
1.50	27-Sep-2007	pooka	Differentiate between cookie2vnode returning an error and return to caller, address unknown: no such cookie, no such node. Make the callers use this info to either create a new vnode or bail.
1.49	27-Sep-2007	pooka	Add error notifications, which are used to deliver errors from the kernel to the file server for silly things the file server did, e.g. attempting to create a file with size VSIZENOTSET. The file server can handle these as it chooses, but the default action is for it to throw its hands in the air and sing "goodbye, cruel world, it's over, walk on by".
1.48	27-Sep-2007	pooka	Don't forget to insert the root node on the hash list. ... I should remember to test also if unmounting a file system works before I commit stuff.
1.47	27-Sep-2007	pooka	Fix a race in how new cookies are checked. Previously the checking was done separate of inserting the cookie into the lookup structure and without any form of interlock. This could lead to the same cookie pointing to two different nodes. Remedy the race by creating a separate "checked and ready to be inserted" cookie list which serves as an interlock without having to hold a fs-global creation lock.
1.46	24-Sep-2007	pooka	add a few comments and g/c dead code
1.45	04-Sep-2007	pooka	branches: 1.45.2; * don't allow the file server to specify a node size to be VSIZENOTSET * KASSERT that VNOVAL == VSIZENOTSET
1.44	01-Aug-2007	pooka	branches: 1.44.2; 1.44.4; 1.44.6; add comment to flag a slight problem
1.43	29-Jul-2007	ad	It's not a good idea for device drivers to modify b_flags, as they don't need to understand the locking around that field. Instead of setting B_ERROR, set b_error instead. b_error is 'owned' by whoever completes the I/O request.
1.42	22-Jul-2007	pooka	Keep track of the maximum size we have supplied the file server (or it has supplied us). If we fault pages which are at offset >= server size, but less than the in-kernel vnode size, inform the file server of the latest developments in file size before issueing the fault. The avoids confusion with files which are not written start to finish. fixes kern/36429 by yamt
1.41	19-Jul-2007	pooka	Initialize pnode to 0 after fetching it from the pool. At least one effect is poll() working much better, as selinfo doesn't contain random bits.
1.40	09-Jul-2007	ad	branches: 1.40.2; Merge some of the less invasive changes from the vmlocking branch: - kthread, callout, devsw API changes - select()/poll() improvements - miscellaneous MT safety improvements
1.39	02-Jul-2007	pooka	check for invalid vtype
1.38	01-Jul-2007	pooka	Give the file server to ability to request the entire pathname buffer under lookup by using PUFFS_KFLAG_LOOKUP_FULLPNBUF instead just the current component.
1.37	01-Jul-2007	pooka	Instead of supplying a plain pid, supply an abstract struct puffs_cid *, which can currently be used to query the pid and lwpid.
1.36	01-Jul-2007	pooka	make puffs_cred an opaque type
1.35	24-Jun-2007	pooka	Split the NOCACHE option in twain: NOCACHE_NAME & NOCACHE_PAGE.
1.34	21-Jun-2007	pooka	Refactor the pnode2vnode translation slightly so that VFS_ROOT can use it directly.
1.33	21-Jun-2007	pooka	Reorganize how the root vnode is fetched so that it doesn't always go through VFS_ROOT() and allow to fetch it without locking it. This allows us to call the cache flush operations also for the root vnode and most notably fixes e.g. a "No such file or directory" for a psshfs root directory ls -l when a file was locally deleted and remotely re-created. Also fix some sloppy programming in root node fetch (mostly cosmetic).
1.32	06-Jun-2007	pooka	Move puffs to a two clause license where it already isn't so. And as agc pointed out, even files with the third clause were already effectively two clause because of a slight bug in the language...
1.31	18-May-2007	pooka	Support VOP_POLL. This requires some acrobatics on the puffs_node, as we give a reference to userspace for the puffs_node for the duration of the poll call. So reference count puffs_node separately from the parent vnode. vref()/vrele() is not possible due to a possible surprise visit from VOP_INACTIVE.
1.30	17-May-2007	pooka	Make it possible for the file server to specify the root vnode type and other information instead of always using VDIR. To make this possible without races, require all root node information already in puffs_mount() and nuke puffs_start2() and the associated start operation completely. requested/inspired by Tobias Nygren
1.29	08-May-2007	pooka	Adventures in file systems, part (u_quad_t)-1: we can't use the file system value for the size of device special files, as that comes from specfs instead of the "host" file system. Therefore, take care that getattr doesn't override the value of vp->v_size.
1.28	01-May-2007	pooka	Fix a problem introduced when I converted puffs to use newlock2: when unmounting the file system in case of a certain timing (and possibly some other conditions), a thread would wait on a condition variable, while another thread broadcast the cv and immediately proceeded to destroy it. The result was a system frozen completely solid shorly after the process waiting for the cv woke up. So introduce reference counting to synchronize destruction of the resources in unmount. I was able to repeat the problem only on my laptop in some special cases, so I do not know how common it was. Ironically, killing the file server process violently instead of unmount() didn't have this problem because it never entered the unmount path from two directions.
1.27	30-Mar-2007	pooka	* abstract ASYNCBIOREAD and let callers freely issue a callback called from putop. even though there's only one user currently, makes code more readable * move "delta" to a standard parameter in vntouser and get rid of the specialcase vntouser_delta
1.26	29-Mar-2007	pooka	Convert spinlocks & sleep/wakeup to newlock2 locking stuff. Fix a bunch of bugs. * park structures are now always allocated from a pool instead of a mixed stack/malloc allocation * get rid of the whole adjbuf concept, always just alloc the maximal amount of memory to satisfy a request * little regression: don't allow interrupting wait from file system to userspace; this had problems already before, but now the problems really started to shine through. I'll try to make this work again some day. * fix bmap to return a sensible value in runp
1.25	20-Mar-2007	pooka	* rework the page cache interaction a bit: cache metadata in the kernel and flush it out all at once instead of continuous updating * add support for delivering notifications to the file server about when a page was written to (but disabled by default for now). the file server can use this to request flushing or invalidating the kernel page cache
1.24	14-Mar-2007	pooka	branches: 1.24.2; Support B_READ\|B_ASYNC in strategy by calling biodone() directly when the file server puts the result.
1.23	12-Mar-2007	ad	branches: 1.23.2; Pass an ipl argument to pool_init/POOL_INIT to be used when initializing the pool's lock.
1.22	27-Feb-2007	pooka	branches: 1.22.2; Make wait for the user file server PCATCHable. This makes it possible to recover the system by just killing processes in case a file server manages to recurse into itself either by fault of file server implementation or by pilot error. The downside is that the code is extremely hard to follow and practically screams out for newlock2 (in addition to screaming "bug here"). The whole PCATCH nonsense and induced megacomplexity can hopefully be avoided in the future by tweaking other parts of the implementation.
1.21	20-Feb-2007	ad	Call genfs_node_destroy() where appropriate.
1.20	16-Feb-2007	pooka	branches: 1.20.2; Check against root node cookie when fetching a new vnode and invoke VFS_ROOT() if the cookies match. Without this fix, if the root vnode was reclaimed, doing lookups for dotdot from the root vnode was possible. In practice this occured only through getcwd.
1.19	15-Feb-2007	pooka	Hide the debug prints behind PUFFSDEBUG instead of DEBUG. Make the latter define the former.
1.18	26-Jan-2007	pooka	Initial attempt at suspend/snapshot support for userspace file servers. This is still pretty much on the level "if it breaks ...". It should work for single-threaded servers which handle one operation from start to finish in one go. Also, it does not yet totally correctly synchronize metadata and data in some cases. So needless to say, it needs improvement, but it is possible that will have to wait for some lock revampage.
1.17	25-Jan-2007	pooka	don't hold spinlocks (except vnode interlock) when doing vget()
1.16	15-Jan-2007	pooka	Store puffs_node's on lists hashed with the cookie value instead of just one flat list.
1.15	15-Jan-2007	pooka	* do not accept the directory cookie as the result of a lookup (otherwise we'd be locking against ourselves) * do not accept duplicate cookies when creating new nodes
1.14	09-Jan-2007	pooka	Introduce flush operations, which the fs server can use to control kernel caching. Currently supported are only flushing the name cache for a directory or flushing the name cache for the entire fs. Also, get rid of PNODE_INACTIVE status, since it was racy and essentially didn't work. All this on top of being useless in the first place ....
1.13	30-Dec-2006	pooka	branches: 1.13.2; * use PUFFS_KFLAG_NOCACHE to also signal that we don't want the namecache * enter files into the namecache immediately when new nodes are created (if it's a caching mount, of course)
1.12	29-Dec-2006	pooka	rename the kernel-provided componentname to puffs_kcn; libpuffs now provides puffs_cn built on top of it
1.11	05-Dec-2006	pooka	branches: 1.11.2; shuffle functions around a bit: move the transport (/dev/puffs) to a different file from the messaging (request contents). no functional change
1.10	05-Dec-2006	pooka	Allow multiple requests to be transferred in each GET/PUTOP. For a single request, the performance is still the same.
1.9	18-Nov-2006	pooka	branches: 1.9.2; As a first generation best-effort hack, use NOCACHE to mean "file size can change without the kernel knowing" and therefore query the file size before invoking read or write operations.
1.8	16-Nov-2006	christos	__unused removal on arguments; approved by core.
1.7	07-Nov-2006	pooka	attach to genfs & support page cache. most noticeable effect is mmap and therefore execution of binaries starting to work, some speed improvements with large file I/O also. caching semantics and error case handling most likely need revisiting.
1.6	27-Oct-2006	pooka	fix checkalias true branch: don't unlock or lock twice
1.5	27-Oct-2006	pooka	support fifos
1.4	26-Oct-2006	pooka	support specfs
1.3	26-Oct-2006	pooka	Fix operations creating new nodes to honor the vnode locking protocol if the userspace server returns an error. Fixes lockups if any of the following operations failed: create, mknod, mkdir, symlink
1.2	23-Oct-2006	pooka	Apply a little eggwash to a deadlock condition where calling getnewvnode() while holding on to any vnode lock deadlocks the system if the file system is being forcibly unmounted. Normal file systems don't trigger this problem because of two reaons: 1) they don't hold on to vnode locks while idling who-knows-where, so the race doesn't trigger 2) they aren't usually unmounted with FORCE; puffs is, in case "someone" manages to make a crashy userspace server Nevertheless, a real solution is slowly being braised.
1.1	22-Oct-2006	pooka	kernel portion of puffs - the Pass-to-Userspace Framework File System. It contains the VFS attachment and userspace message-passing interface. This work was initially started and completed for Google SoC 2005 and tweaked to work a bit better in the past few weeks. While being far from complete, it is functional enough to be able and stable to host a fairly general-purpose in-memory file system in userspace. Even so, puffs should be considered experimental and no binary compatibility for interfaces or crash-freedom or zero security implications should be relied upon just yet. The GSoC project was mentored by William Studenmund and the final review for the code was done by Christos.
1.9.2.4	01-Feb-2007	ad	Sync with head.
1.9.2.3	12-Jan-2007	ad	Sync with head.
1.9.2.2	18-Nov-2006	ad	Sync with head.
1.9.2.1	18-Nov-2006	ad	file puffs_subr.c was added on branch newlock2 on 2006-11-18 21:39:20 +0000
1.11.2.2	10-Dec-2006	yamt	sync with head.
1.11.2.1	05-Dec-2006	yamt	file puffs_subr.c was added on branch yamt-splraiseipl on 2006-12-10 07:18:38 +0000
1.13.2.9	17-Mar-2008	yamt	sync with head.
1.13.2.8	04-Feb-2008	yamt	sync with head.
1.13.2.7	21-Jan-2008	yamt	sync with head
1.13.2.6	07-Dec-2007	yamt	sync with head
1.13.2.5	27-Oct-2007	yamt	sync with head.
1.13.2.4	03-Sep-2007	yamt	sync with head.
1.13.2.3	26-Feb-2007	yamt	sync with head.
1.13.2.2	30-Dec-2006	yamt	sync with head.
1.13.2.1	30-Dec-2006	yamt	file puffs_subr.c was added on branch yamt-lazymbuf on 2006-12-30 20:50:01 +0000
1.20.2.6	17-May-2007	yamt	sync with head.
1.20.2.5	07-May-2007	yamt	sync with head.
1.20.2.4	15-Apr-2007	yamt	sync with head.
1.20.2.3	24-Mar-2007	yamt	sync with head.
1.20.2.2	12-Mar-2007	rmind	Sync with HEAD.
1.20.2.1	27-Feb-2007	yamt	- sync with head. - move sched_changepri back to kern_synch.c as it doesn't know PPQ anymore.
1.22.2.14	12-Oct-2007	ad	Sync with head.
1.22.2.13	09-Oct-2007	ad	Sync with head.
1.22.2.12	09-Oct-2007	ad	Sync with head.
1.22.2.11	16-Sep-2007	ad	Checkpoint work in progress on the vnode lifecycle and reference counting stuff. This makes it work properly without kernel_lock and fixes a few quite old bugs. See vfs_subr.c 1.283.2.17 for details.
1.22.2.10	20-Aug-2007	ad	Sync with HEAD.
1.22.2.9	19-Aug-2007	ad	- Back out the biodone() changes. - Eliminate B_ERROR (from HEAD).
1.22.2.8	15-Jul-2007	ad	Sync with head.
1.22.2.7	17-Jun-2007	ad	- Increase the number of thread priorities from 128 to 256. How the space is set up is to be revisited. - Implement soft interrupts as kernel threads. A generic implementation is provided, with hooks for fast-path MD code that can run the interrupt threads over the top of other threads executing in the kernel. - Split vnode::v_flag into three fields, depending on how the flag is locked (by the interlock, by the vnode lock, by the file system). - Miscellaneous locking fixes and improvements.
1.22.2.6	09-Jun-2007	ad	Sync with head.
1.22.2.5	08-Jun-2007	ad	Sync with head.
1.22.2.4	10-Apr-2007	ad	Sync with head.
1.22.2.3	05-Apr-2007	ad	Compile fixes.
1.22.2.2	13-Mar-2007	ad	Pull in the initial set of changes for the vmlocking branch.
1.22.2.1	13-Mar-2007	ad	Sync with head.
1.23.2.1	11-Jul-2007	mjf	Sync with head.
1.24.2.1	29-Mar-2007	reinoud	Pullup to -current
1.40.2.2	10-Sep-2007	skrll	Sync with HEAD.
1.40.2.1	15-Aug-2007	skrll	Sync with HEAD.
1.44.6.2	01-Aug-2007	pooka	add comment to flag a slight problem
1.44.6.1	01-Aug-2007	pooka	file puffs_subr.c was added on branch matt-mips64 on 2007-08-01 14:20:46 +0000
1.44.4.3	23-Mar-2008	matt	sync with HEAD
1.44.4.2	09-Jan-2008	matt	sync with HEAD
1.44.4.1	06-Nov-2007	matt	sync with HEAD
1.44.2.4	09-Dec-2007	jmcneill	Sync with HEAD.
1.44.2.3	21-Nov-2007	joerg	Sync with HEAD.
1.44.2.2	26-Oct-2007	joerg	Sync with HEAD. Follow the merge of pmap.c on i386 and amd64 and move pmap_init_tmp_pgtbl into arch/x86/x86/pmap.c. Modify the ACPI wakeup code to restore CR4 before jumping back into kernel space as the large page option might cover that.
1.44.2.1	02-Oct-2007	joerg	Sync with HEAD.
1.45.2.2	14-Oct-2007	yamt	sync with head.
1.45.2.1	06-Oct-2007	yamt	sync with head.
1.57.4.3	18-Feb-2008	mjf	Sync with HEAD.
1.57.4.2	27-Dec-2007	mjf	Sync with HEAD.
1.57.4.1	19-Nov-2007	mjf	Sync with HEAD.
1.57.2.1	18-Nov-2007	bouyer	Sync with HEAD
1.60.2.1	26-Dec-2007	ad	Sync with head.
1.62.4.1	08-Jan-2008	bouyer	Sync with HEAD
1.64.6.2	17-Jan-2009	mjf	Sync with HEAD.
1.64.6.1	03-Apr-2008	mjf	Sync with HEAD.
1.64.2.1	24-Mar-2008	keiichi	sync with head.
1.65.12.1	19-Jan-2009	skrll	Sync with HEAD.
1.65.10.1	13-Dec-2008	haad	Update haad-dm branch to haad-dm-base2.
1.65.4.1	04-May-2009	yamt	sync with head.
1.66.42.1	17-Jan-2015	martin	Pull up following revision(s) (requested by maxv in ticket #427): sys/compat/svr4/svr4_schedctl.c: revision 1.8 sys/netinet/tcp_timer.c: revision 1.88 sys/miscfs/genfs/layer_vfsops.c: revision 1.45 sys/compat/svr4/svr4_ioctl.c: revision 1.37 sys/ufs/chfs/chfs_vfsops.c: revision 1.14 sys/miscfs/fdesc/fdesc_vfsops.c: revision 1.91 sys/compat/linux/arch/i386/linux_ptrace.c: revision 1.30 sys/compat/common/kern_time_50.c: revision 1.28 sys/netinet6/ip6_forward.c: revision 1.74 sys/miscfs/umapfs/umap_vnops.c: revision 1.57 sys/compat/svr4/svr4_fcntl.c: revision 1.74 distrib/sets/lists/comp/mi: revision 1.1931 sys/netinet6/udp6_output.c: revision 1.46 sys/fs/puffs/puffs_compat.c: revision 1.3 sys/fs/udf/udf_rename.c: revision 1.11 sys/compat/svr4/svr4_filio.c: revision 1.24 sys/fs/udf/udf_rename.c: revision 1.12 sys/netinet/tcp_usrreq.c: revision 1.202 sys/miscfs/umapfs/umap_subr.c: revision 1.29 sys/compat/linux/common/linux_fadvise64.c: revision 1.3 sys/netinet/if_atm.c: revision 1.34 sys/miscfs/procfs/procfs_subr.c: revision 1.106 sys/miscfs/genfs/layer_subr.c: revision 1.37 sys/netinet/tcp_sack.c: revision 1.30 sys/compat/freebsd/freebsd_misc.c: revision 1.33 sys/compat/freebsd/freebsd_file.c: revision 1.33 sys/ufs/chfs/chfs_vnode.c: revision 1.12 sys/compat/svr4/svr4_ttold.c: revision 1.34 sys/compat/linux/common/linux_file.c: revision 1.114 sys/compat/linux/arch/mips/linux_machdep.c: revision 1.43 sys/compat/linux/common/linux_signal.c: revision 1.76 sys/compat/common/compat_util.c: revision 1.46 sys/compat/linux/arch/arm/linux_ptrace.c: revision 1.18 sys/compat/svr4/svr4_sockio.c: revision 1.36 sys/compat/linux/arch/arm/linux_machdep.c: revision 1.32 sys/compat/svr4/svr4_signal.c: revision 1.66 sys/kern/kern_exec.c: revision 1.410 sys/fs/puffs/puffs_vfsops.c: revision 1.115 sys/compat/svr4/svr4_exec_elf64.c: revision 1.15 sys/compat/linux/arch/i386/linux_machdep.c: revision 1.159 sys/compat/linux/arch/alpha/linux_machdep.c: revision 1.50 sys/compat/linux32/common/linux32_misc.c: revision 1.24 sys/netinet/in_pcb.c: revision 1.153 sys/sys/malloc.h: revision 1.116 sys/compat/common/if_43.c: revision 1.9 share/man/man9/Makefile: revision 1.380 sys/netinet/tcp_vtw.c: revision 1.12 sys/miscfs/umapfs/umap_vfsops.c: revision 1.95 sys/ufs/ext2fs/ext2fs_vfsops.c: revision 1.186 sys/compat/common/uipc_syscalls_43.c: revision 1.46 sys/ufs/ext2fs/ext2fs_vnops.c: revision 1.115 sys/fs/puffs/puffs_msgif.c: revision 1.97 sys/compat/svr4/svr4_ipc.c: revision 1.27 sys/compat/linux/common/linux_exec.c: revision 1.117 sys/ufs/ext2fs/ext2fs_readwrite.c: revision 1.66 sys/netinet/tcp_output.c: revision 1.179 sys/compat/svr4/svr4_termios.c: revision 1.28 sys/fs/udf/udf_strat_bootstrap.c: revision 1.4 sys/fs/puffs/puffs_subr.c: revision 1.67 sys/fs/puffs/puffs_node.c: revision 1.36 sys/miscfs/overlay/overlay_vnops.c: revision 1.21 sys/fs/cd9660/cd9660_node.c: revision 1.34 sys/netinet/raw_ip.c: revision 1.146 sys/sys/mallocvar.h: revision 1.13 sys/miscfs/overlay/overlay_vfsops.c: revision 1.63 share/man/man9/malloc.9: revision 1.50 sys/netinet6/dest6.c: revision 1.18 sys/compat/linux/common/linux_uselib.c: revision 1.33 sys/compat/linux/common/linux_socket.c: revision 1.120 share/man/man9/malloc.9: revision 1.51 sys/netinet/tcp_subr.c: revision 1.257 sys/compat/linux/common/linux_socketcall.c: revision 1.45 sys/compat/linux/common/linux_fadvise64_64.c: revision 1.3 sys/compat/freebsd/freebsd_ipc.c: revision 1.17 sys/compat/linux/common/linux_misc_notalpha.c: revision 1.109 sys/compat/linux/arch/alpha/linux_pipe.c: revision 1.17 sys/netinet6/in6_pcb.c: revision 1.132 sys/netinet6/in6_ifattach.c: revision 1.94 sys/compat/svr4/svr4_exec_elf32.c: revision 1.15 sys/miscfs/nullfs/null_vfsops.c: revision 1.90 sys/fs/cd9660/cd9660_util.c: revision 1.12 sys/compat/linux/arch/powerpc/linux_machdep.c: revision 1.48 sys/compat/freebsd/freebsd_exec_elf32.c: revision 1.20 sys/miscfs/procfs/procfs_vfsops.c: revision 1.94 sys/compat/linux/arch/powerpc/linux_ptrace.c: revision 1.28 sys/compat/linux/common/linux_sched.c: revision 1.67 sys/compat/linux/common/linux_exec_aout.c: revision 1.67 sys/compat/linux/common/linux_pipe.c: revision 1.67 sys/compat/linux/common/linux_llseek.c: revision 1.34 sys/compat/linux/arch/mips/linux_ptrace.c: revision 1.10 Do not uselessly include <sys/malloc.h>. Cleanup: - remove struct kmembuckets (dead) - correctly deadify MALLOC_XX - remove MALLOC_DEFINE_LIMIT and MALLOC_JUSTDEFINE_LIMIT (dead) - remove malloc_roundup(), malloc_type_setlimit(), MALLOC_DEFINE_LIMIT() and MALLOC_JUSTDEFINE_LIMIT() from man 9 malloc New sentence, new line. Bump date for previous. Obsolete malloc_roundup(9), malloc_type_setlimit(9) and MALLOC_DEFINE_LIMIT(9) man pages.
1.66.26.1	03-Dec-2017	jdolecek	update from HEAD
1.91	27-Jan-2019	pgoyette	Merge the [pgoyette-compat] branch
1.90	07-Jul-2016	msaitoh	branches: 1.90.16; 1.90.18; KNF. Remove extra spaces. No functional change.
1.89	15-Feb-2015	manu	Add PUFFS_KFLAG_NOFLUSH_META to prevent sending metadata flush to FUSE FUSE filesystems do not expect to get metadata updates for [amc]time and size, they updates the value on their own after operations. The PUFFS PUFFS_KFLAG_NOFLUSH_META option prevents regular metadata cache flushes to the filesystem , and libperfuse uses it to match Linux FUSE behavior. While there, fix a bug in SETATTR: do not update kernel metadata cache from SETATTR reply when the request is asynchronous, as we do not have the reply yet.
1.88	05-Oct-2014	apb	branches: 1.88.2; Add close brace, accidentally omitted from previous change.
1.87	05-Oct-2014	apb	Safer definitions of DPRINTF and DPRINTF_VERBOSE. In the PUFFSDEBUG case, wrap do { ... } while (/CONSTCOND/0) around the definitions. In the non-PUFFSDEBUG case, define them as ((void)0) instead of as empty.
1.86	28-Aug-2014	hannken	Change puffs from hashlist to vcache. - field "pa_nhashbuckets" of struct "puffs_kargs" becomes a no-op. and should be removed on the next protocol version bump.
1.85	16-Aug-2014	manu	Add a oflags input field to open requests so that the filesystem can pass back information about the file. Implement PUFFS_OPEN_IO_DIRECT, which will force direct IO (bypassing page cache) for the file.
1.84	17-Oct-2013	christos	branches: 1.84.4; - remove unused variables - add _NOERROR flavor macros for the case where errors are ignored.
1.83	06-Mar-2013	yamt	branches: 1.83.6; comment
1.82	11-Aug-2012	manu	branches: 1.82.2; Missing bit in previous commit (PUFFS_KFLAG_CACHE_DOTDOT option to avoid looking up ..)
1.81	27-Jul-2012	manu	Rename slow sopreq queue into node sopreq queue, to refet the fact that is only intended for postponed node reclaims. When purging the node sopreq queue, do not call puffs_msg_sendresp(), as it makes no sense.
1.80	21-Jul-2012	manu	- Improve PUFFS_KFLAG_CACHE_FS_TTL by reclaiming older inactive nodes. The normal kernel behavior is to retain inactive nodes in the freelist until it runs out of vnodes. This has some merit for local filesystems, where the cost of an allocation is about the same as the cost of a lookup. But that situation is not true for distributed filesystems. On the other hand, keeping inactive nodes for a long time hold memory in the file server process, and when the kernel runs out of vnodes, it produce reclaim avalanches that increase lattency for other operations. We do not reclaim inactive vnodes immediatly either, as they may be looked up again shortly. Instead we introduce a grace time and we reclaim nodes that have been inactive beyond the grace time. - Fix lookup/reclaim race condition. The above improvement undercovered a race condition between lookup and reclaim. If we reclaimed a vnode associated with a userland cookie while a lookup returning that same cookiewas inprogress, then the kernel ends up with a vnode associated with a cookie that has been reclaimed in userland. Next operation on the cookie will crash (or at least confuse) the filesystem. We fix this by introducing a lookup count in kernel and userland. On reclaim, the kernel sends the count, which enable userland to detect situation where it initiated a lookup that is not completed in kernel. In such a situation, the reclaim must be ignored, as the node is about to be looked up again.
1.79	08-Apr-2012	manu	Add name and atttribute cache with filesytem provided TTL. lookup, create, mknod, mkdir, symlink, getattr and setattr messages have been extended so that attributes and their TTL can be provided by the filesytem. lookup, create, mknod, mkdir, and symlink messages are also extended so that the filesystem can provide name TTL.
1.78	29-Aug-2011	manu	branches: 1.78.2; 1.78.6; 1.78.8; Add a mutex for operations that touch size (setattr, getattr, write, fsync). This is required to avoid data corruption bugs, where a getattr slices itself within a setattr operation, and sets the size to the stall value it got from the filesystem. That value is smaller than the one set by setattr, and the call to uvm_vnp_setsize() trigged a spurious truncate. The result is a chunk of zeroed data in the file. Such a situation can easily happen when the ioflush thread issue a VOP_FSYNC/puffs_vnop_sync/flushvncache/dosetattrn while andother process do a sys_stat/VOP_GETATTR/puffs_vnop_getattr. This mutex on size operation can be removed the day we decide VOP_GETATTR has to operated on a locked vnode, since the other operations that touch size already require that.
1.77	11-Jan-2011	kefren	add advlock to puffs. ok pooka@ should fix kern/43321
1.76	06-Jul-2010	pooka	Add compat to enable running puffs in a 64bit time_t kernel against a server which runs in 32bit time_t namespace.
1.75	07-Jan-2010	pooka	branches: 1.75.2; 1.75.4; Rename PUFFS_SOPREQ_EXIT to PUFFS_SOPREQSYS_EXIT to better signal it comes from within the kernel instead of as a direct result of a user request. no functional change
1.74	07-Jan-2010	pooka	Add a PUFFS_UNMOUNT server->kernel request, which causes the kernel to initiate self destruct, i.e. unmount(MNT_FORCE). This, however, is a semi-controlled self-destruct, since all caches are flushed before the (possibly) violent unmount takes place.
1.73	07-Dec-2009	pooka	Process flush requests from the file server in a separate thread context. This fixes a long-standing but seldomly seen deadlock, where the kernel was holding pages busy (due to e.g. readahead request) while waiting for the server to respond, and the server made a callback into the kernel asking to invalidate those pages. ... or, well, theoretically fixes, since I didn't have any reliable way of repeating the deadlock and I think I saw it only twice.
1.72	05-Nov-2009	pooka	Kill suspend support. It was never implemented correctly: * it depended on the biglock (in a very cruel way) * it was attached to userspace transactions rather than logical fs operations (If someone wants to revisit it some day, most of the stuff can be reused from cvs history)
1.71	05-Nov-2009	pooka	Reinstante PNODE_DYING. vmlocking had a brief hiatus when it was not a valid optimization, but that's long gone and once VOP_INACTIVE is called and the file server says that the vnode is going to be recycled, it really is going to be recycled extra references gained or not.
1.70	28-Jan-2008	pooka	branches: 1.70.10; 1.70.20; 1.70.28; For code clarity typedef void *puffs_cookie_t. No functional change.
1.69	02-Jan-2008	pooka	More type-punning workarounds. Curiously the kernel compilation flags cause gcc to not complain.
1.68	02-Jan-2008	ad	Merge vmlocking2 to head.
1.67	08-Dec-2007	pooka	branches: 1.67.4; Now that "l" is gone both as an argument to operations and from componentname, remove all vestiges of puffs_cid.
1.66	05-Dec-2007	pooka	Send a response message for flush operations from the kernel instead of abusing the return value of write(2).
1.65	20-Nov-2007	pooka	branches: 1.65.2; Retire M_PUFFS, use kmem(9) instead.
1.64	17-Nov-2007	pooka	Make puffs_updatenode() take a puffs_node instead of a vnode. This way we don't need to worry if a vnode has been reclaimed from under us.
1.63	17-Nov-2007	pooka	Implement a biodone callback for async writes similar to reads and use that when possible.
1.62	16-Nov-2007	pooka	Restructure the messaging interface a bit more: make all interfacing with the file server happen through puffs_msg_enqueue() and puffs_msg_wait() instead of having a billion different routines. Build the existing system upon these two. Most importantly though, decouple insertation into the op queue from the actual wait. This is useful for a number of reasons coming soon to a cvs repo near you.
1.61	12-Nov-2007	pooka	Bounds-check responses from userspace.
1.60	10-Nov-2007	pooka	Part 2/n of extensive changes to request transport to/from userspace: Rip the transport code completely out of puffs and generalize it into an independent module which will be used for multiple purposes in the future. This module is called the Pass-to-Userspace Transporter (known as "putter" among friends). This is very much work-in-progress and one dependency with puffs remains: the request framing format. The device name is still /dev/puffs, but that will change soon. Users of puffs need the following in their kernel configs now: pseudo-device putter
1.59	11-Oct-2007	pooka	branches: 1.59.2; 1.59.4; Part 1/n of some pretty extensive changes to how the kernel module interacts with the userspace file server: * since the kernel-user communication is not purely request-response anymore (hasn't been since 2006), try to rename some "request" to "message". more similar mangling will take place in the future. * completely rework how messages are allocated. previously most of them were borrowed from the stack (originally all of them), but now always allocate dynamically. this makes the structure of the code much cleaner. also makes it possible to fix a locking order violation. it enables plenty of future enhancements. * start generalizing the transport interface to be independent of puffs * move transport interface to read/write instead of ioctl. the old one had legacy design problems, and besides, ioctl's suck. implement a very generic version for now; this will be worked on later hopefully some day reaching "highly optimized". * implement libpuffs support behind existing library request interfaces. this will change eventually (I hate those interfaces)
1.58	09-Oct-2007	pooka	g/c more unused stuff
1.57	09-Oct-2007	pooka	g/c vntouser_req(), it's not used anymore
1.56	04-Oct-2007	pooka	g/c the "sizeop" code previous used for ioctl/fcntl. It was already commented out and has bitrotted beyond all recognition, so it needs complete rethinking.
1.55	02-Oct-2007	pooka	If kernel resource allocation fails after the file server has committed something, issue an abort. The abort is done through the regular op channel, e.g. failed mkdir leads to regular rmdir, inactive and reclaim. No internal interface is planned currently for the one file system out of a million which would implement it to benefit from the one case in a billion where kernel resource allocation actually does fail and out of that one case in a trillion where internal vs. external would make a difference.
1.54	01-Oct-2007	pooka	* better error checking: validate error values received from userland to be vaild errno values * include string describing error in PUFFS_ERR * get rid of union in puffs_req, it's nothing but trouble * pass pmp to async i/o callbacks
1.53	27-Sep-2007	pooka	Split routines handling nodes from puffs_subr to puffs_node. No functional change.
1.52	27-Sep-2007	pooka	Differentiate between cookie2vnode returning an error and return to caller, address unknown: no such cookie, no such node. Make the callers use this info to either create a new vnode or bail.
1.51	27-Sep-2007	pooka	Add error notifications, which are used to deliver errors from the kernel to the file server for silly things the file server did, e.g. attempting to create a file with size VSIZENOTSET. The file server can handle these as it chooses, but the default action is for it to throw its hands in the air and sing "goodbye, cruel world, it's over, walk on by".
1.50	27-Sep-2007	pooka	Fix a race in how new cookies are checked. Previously the checking was done separate of inserting the cookie into the lookup structure and without any form of interlock. This could lead to the same cookie pointing to two different nodes. Remedy the race by creating a separate "checked and ready to be inserted" cookie list which serves as an interlock without having to hold a fs-global creation lock.
1.49	24-Sep-2007	pooka	add a few comments and g/c dead code
1.48	30-Jul-2007	pooka	branches: 1.48.4; 1.48.6; 1.48.8; 1.48.10; Move PUFFS_TYPEPREFIX to puffs_msgif.h since it's used in a macro there.
1.47	22-Jul-2007	pooka	Keep track of the maximum size we have supplied the file server (or it has supplied us). If we fault pages which are at offset >= server size, but less than the in-kernel vnode size, inform the file server of the latest developments in file size before issueing the fault. The avoids confusion with files which are not written start to finish. fixes kern/36429 by yamt
1.46	17-Jul-2007	pooka	branches: 1.46.2; Set a file server supplied file system type in the type field and set the mntfromname to be the place mounted from instead of the type.
1.45	01-Jul-2007	pooka	Give the file server to ability to request the entire pathname buffer under lookup by using PUFFS_KFLAG_LOOKUP_FULLPNBUF instead just the current component.
1.44	01-Jul-2007	pooka	Instead of supplying a plain pid, supply an abstract struct puffs_cid *, which can currently be used to query the pid and lwpid.
1.43	01-Jul-2007	pooka	make puffs_cred an opaque type
1.42	24-Jun-2007	pooka	Split the NOCACHE option in twain: NOCACHE_NAME & NOCACHE_PAGE.
1.41	21-Jun-2007	pooka	Refactor the pnode2vnode translation slightly so that VFS_ROOT can use it directly.
1.40	21-Jun-2007	pooka	Reorganize how the root vnode is fetched so that it doesn't always go through VFS_ROOT() and allow to fetch it without locking it. This allows us to call the cache flush operations also for the root vnode and most notably fixes e.g. a "No such file or directory" for a psshfs root directory ls -l when a file was locally deleted and remotely re-created. Also fix some sloppy programming in root node fetch (mostly cosmetic).
1.39	06-Jun-2007	pooka	Move puffs to a two clause license where it already isn't so. And as agc pointed out, even files with the third clause were already effectively two clause because of a slight bug in the language...
1.38	19-May-2007	pooka	forgot to commit this with puffs_vnops.c 1.72: Actually, we do need separate "no references in file server" and "noref + inactive" flags if we wish to correctly support unix open file semantics and optimize away pre-reclaim cache flushes. So, add PNODE_DYING which stands for norefs + inactive.
1.37	18-May-2007	pooka	Introduce noref setbacks, which the file server can use to signal the kernel it has 0 references to the node in question. In other words, this can be used to avoid inactive(), or, if the file server does not implement inactive, prompt reclaim for removed nodes.
1.36	18-May-2007	pooka	Support VOP_POLL. This requires some acrobatics on the puffs_node, as we give a reference to userspace for the puffs_node for the duration of the poll call. So reference count puffs_node separately from the parent vnode. vref()/vrele() is not possible due to a possible surprise visit from VOP_INACTIVE.
1.35	17-May-2007	pooka	Make it possible for the file server to specify the root vnode type and other information instead of always using VDIR. To make this possible without races, require all root node information already in puffs_mount() and nuke puffs_start2() and the associated start operation completely. requested/inspired by Tobias Nygren
1.34	07-May-2007	pooka	Introduce puffs "setbacks", which can be used to set certain flags for nodes upon return from the userspace. Currently it can be used to indicate that the file server should be notified of "inactive" in case the file server has opted to not receive inactive every time the reference count for a vnode drops to zero. (inactive is a common event, almost never requires any action and must be executed sychronously, so it is wasteful). While doing this, cleanup the release-relock nonsense from the vntouser*() arguments. It was never enabled and the whole LOCKEDVP() concept was very broken to begin with.
1.33	01-May-2007	pooka	Fix a problem introduced when I converted puffs to use newlock2: when unmounting the file system in case of a certain timing (and possibly some other conditions), a thread would wait on a condition variable, while another thread broadcast the cv and immediately proceeded to destroy it. The result was a system frozen completely solid shorly after the process waiting for the cv woke up. So introduce reference counting to synchronize destruction of the resources in unmount. I was able to repeat the problem only on my laptop in some special cases, so I do not know how common it was. Ironically, killing the file server process violently instead of unmount() didn't have this problem because it never entered the unmount path from two directions.
1.32	16-Apr-2007	pooka	Give the file server the ability to specify the file handle length instead of defining a static length file handle on the framework-level.
1.31	13-Apr-2007	pooka	* add fhlen to kernel argument structure * rename it to puffs_kargs instead of puffs_args
1.30	04-Apr-2007	pooka	Make it possible to interrupt waiters for fs operation completion again. This is useful until locking is further developed and basically any deadlocks can be solved by killing appropriate processes. Thanks especially to Tommi Kyntola and Antti Louko for sitting down with me and discussing resource ownership and locking strategies in implementing this.
1.29	30-Mar-2007	pooka	* abstract ASYNCBIOREAD and let callers freely issue a callback called from putop. even though there's only one user currently, makes code more readable * move "delta" to a standard parameter in vntouser and get rid of the specialcase vntouser_delta
1.28	29-Mar-2007	pooka	Convert spinlocks & sleep/wakeup to newlock2 locking stuff. Fix a bunch of bugs. * park structures are now always allocated from a pool instead of a mixed stack/malloc allocation * get rid of the whole adjbuf concept, always just alloc the maximal amount of memory to satisfy a request * little regression: don't allow interrupting wait from file system to userspace; this had problems already before, but now the problems really started to shine through. I'll try to make this work again some day. * fix bmap to return a sensible value in runp
1.27	20-Mar-2007	pooka	* rework the page cache interaction a bit: cache metadata in the kernel and flush it out all at once instead of continuous updating * add support for delivering notifications to the file server about when a page was written to (but disabled by default for now). the file server can use this to request flushing or invalidating the kernel page cache
1.26	14-Mar-2007	pooka	branches: 1.26.2; Support B_READ\|B_ASYNC in strategy by calling biodone() directly when the file server puts the result.
1.25	27-Feb-2007	pooka	branches: 1.25.2; 1.25.4; Make wait for the user file server PCATCHable. This makes it possible to recover the system by just killing processes in case a file server manages to recurse into itself either by fault of file server implementation or by pilot error. The downside is that the code is extremely hard to follow and practically screams out for newlock2 (in addition to screaming "bug here"). The whole PCATCH nonsense and induced megacomplexity can hopefully be avoided in the future by tweaking other parts of the implementation.
1.24	15-Feb-2007	pooka	branches: 1.24.2; Hide the debug prints behind PUFFSDEBUG instead of DEBUG. Make the latter define the former.
1.23	29-Jan-2007	hubertf	Remove more duplicate headers. Patch by Slava Semushin <slava.semushin@gmail.com> Again, this was tested by comparing obj files from a pristine and a patched source tree against an i386/ALL kernel, and also for src/sbin/fsck_ffs, src/sbin/fsdb and src/usr.sbin/makefs. Only changes in assert() line numbers were detected in 'objdump -d' output.
1.22	26-Jan-2007	pooka	Initial attempt at suspend/snapshot support for userspace file servers. This is still pretty much on the level "if it breaks ...". It should work for single-threaded servers which handle one operation from start to finish in one go. Also, it does not yet totally correctly synchronize metadata and data in some cases. So needless to say, it needs improvement, but it is possible that will have to wait for some lock revampage.
1.21	21-Jan-2007	pooka	optimize a bit: don't flush pages for vnodes which have no references in the kernel or links in the backend
1.20	15-Jan-2007	pooka	Store puffs_node's on lists hashed with the cookie value instead of just one flat list.
1.19	15-Jan-2007	pooka	* do not accept the directory cookie as the result of a lookup (otherwise we'd be locking against ourselves) * do not accept duplicate cookies when creating new nodes
1.18	09-Jan-2007	pooka	Introduce flush operations, which the fs server can use to control kernel caching. Currently supported are only flushing the name cache for a directory or flushing the name cache for the entire fs. Also, get rid of PNODE_INACTIVE status, since it was racy and essentially didn't work. All this on top of being useless in the first place ....
1.17	02-Jan-2007	pooka	* check userspace version and prevent incompatible mount * some general maintenance
1.16	30-Dec-2006	pooka	branches: 1.16.2; * use PUFFS_KFLAG_NOCACHE to also signal that we don't want the namecache * enter files into the namecache immediately when new nodes are created (if it's a caching mount, of course)
1.15	29-Dec-2006	pooka	rename the kernel-provided componentname to puffs_kcn; libpuffs now provides puffs_cn built on top of it
1.14	10-Dec-2006	pooka	Fix a race condition that would cause the mountpoint to be cleaned from under someone waiting for the fs server response in puffs_unmount() if the descriptor was closed during the response wait (such as bug leading to a crash in fs implementation unmount()).
1.13	05-Dec-2006	pooka	branches: 1.13.2; shuffle functions around a bit: move the transport (/dev/puffs) to a different file from the messaging (request contents). no functional change
1.12	05-Dec-2006	pooka	Allow multiple requests to be transferred in each GET/PUTOP. For a single request, the performance is still the same.
1.11	01-Dec-2006	pooka	prefix kernel flags with PUFFS_KFLAG to have a separate namespace from the library flags
1.10	01-Dec-2006	pooka	don't call the fs server for all operations, only those it has told us that it implements
1.9	18-Nov-2006	pooka	branches: 1.9.2; As a first generation best-effort hack, use NOCACHE to mean "file size can change without the kernel knowing" and therefore query the file size before invoking read or write operations.
1.8	17-Nov-2006	pooka	Introduce uncached operation, makes sense when the file system backend can be modified from elsewhere than the file system interface
1.7	09-Nov-2006	pooka	few renames to better differentiate between mount & start.. plus some other renaming
1.6	07-Nov-2006	pooka	attach to genfs & support page cache. most noticeable effect is mmap and therefore execution of binaries starting to work, some speed improvements with large file I/O also. caching semantics and error case handling most likely need revisiting.
1.5	06-Nov-2006	pooka	puffs_park always contains a specific puffs_req, so make it a member instead of a pointer
1.4	06-Nov-2006	pooka	make it possible to build & load puffs as an LKM by Lubomir Kundrak, PR kern/35000
1.3	27-Oct-2006	pooka	support fifos
1.2	26-Oct-2006	pooka	support specfs
1.1	22-Oct-2006	pooka	kernel portion of puffs - the Pass-to-Userspace Framework File System. It contains the VFS attachment and userspace message-passing interface. This work was initially started and completed for Google SoC 2005 and tweaked to work a bit better in the past few weeks. While being far from complete, it is functional enough to be able and stable to host a fairly general-purpose in-memory file system in userspace. Even so, puffs should be considered experimental and no binary compatibility for interfaces or crash-freedom or zero security implications should be relied upon just yet. The GSoC project was mentored by William Studenmund and the final review for the code was done by Christos.
1.9.2.4	01-Feb-2007	ad	Sync with head.
1.9.2.3	12-Jan-2007	ad	Sync with head.
1.9.2.2	18-Nov-2006	ad	Sync with head.
1.9.2.1	18-Nov-2006	ad	file puffs_sys.h was added on branch newlock2 on 2006-11-18 21:39:20 +0000
1.13.2.3	18-Dec-2006	yamt	sync with head.
1.13.2.2	10-Dec-2006	yamt	sync with head.
1.13.2.1	05-Dec-2006	yamt	file puffs_sys.h was added on branch yamt-splraiseipl on 2006-12-10 07:18:38 +0000
1.16.2.9	04-Feb-2008	yamt	sync with head.
1.16.2.8	21-Jan-2008	yamt	sync with head
1.16.2.7	07-Dec-2007	yamt	sync with head
1.16.2.6	15-Nov-2007	yamt	sync with head.
1.16.2.5	27-Oct-2007	yamt	sync with head.
1.16.2.4	03-Sep-2007	yamt	sync with head.
1.16.2.3	26-Feb-2007	yamt	sync with head.
1.16.2.2	30-Dec-2006	yamt	sync with head.
1.16.2.1	30-Dec-2006	yamt	file puffs_sys.h was added on branch yamt-lazymbuf on 2006-12-30 20:50:01 +0000
1.24.2.5	17-May-2007	yamt	sync with head.
1.24.2.4	07-May-2007	yamt	sync with head.
1.24.2.3	15-Apr-2007	yamt	sync with head.
1.24.2.2	24-Mar-2007	yamt	sync with head.
1.24.2.1	12-Mar-2007	rmind	Sync with HEAD.
1.25.4.1	11-Jul-2007	mjf	Sync with head.
1.25.2.8	12-Oct-2007	ad	Sync with head.
1.25.2.7	09-Oct-2007	ad	Sync with head.
1.25.2.6	16-Sep-2007	ad	Checkpoint work in progress on the vnode lifecycle and reference counting stuff. This makes it work properly without kernel_lock and fixes a few quite old bugs. See vfs_subr.c 1.283.2.17 for details.
1.25.2.5	20-Aug-2007	ad	Sync with HEAD.
1.25.2.4	15-Jul-2007	ad	Sync with head.
1.25.2.3	09-Jun-2007	ad	Sync with head.
1.25.2.2	08-Jun-2007	ad	Sync with head.
1.25.2.1	10-Apr-2007	ad	Sync with head.
1.26.2.1	29-Mar-2007	reinoud	Pullup to -current
1.46.2.1	15-Aug-2007	skrll	Sync with HEAD.
1.48.10.2	30-Jul-2007	pooka	Move PUFFS_TYPEPREFIX to puffs_msgif.h since it's used in a macro there.
1.48.10.1	30-Jul-2007	pooka	file puffs_sys.h was added on branch matt-mips64 on 2007-07-30 09:04:59 +0000
1.48.8.2	14-Oct-2007	yamt	sync with head.
1.48.8.1	06-Oct-2007	yamt	sync with head.
1.48.6.3	23-Mar-2008	matt	sync with HEAD
1.48.6.2	09-Jan-2008	matt	sync with HEAD
1.48.6.1	06-Nov-2007	matt	sync with HEAD
1.48.4.7	09-Dec-2007	jmcneill	Sync with HEAD.
1.48.4.6	21-Nov-2007	joerg	Sync with HEAD.
1.48.4.5	14-Nov-2007	joerg	Sync with HEAD.
1.48.4.4	11-Nov-2007	joerg	Sync with HEAD.
1.48.4.3	26-Oct-2007	joerg	Sync with HEAD. Follow the merge of pmap.c on i386 and amd64 and move pmap_init_tmp_pgtbl into arch/x86/x86/pmap.c. Modify the ACPI wakeup code to restore CR4 before jumping back into kernel space as the large page option might cover that.
1.48.4.2	07-Oct-2007	joerg	Sync with HEAD.
1.48.4.1	02-Oct-2007	joerg	Sync with HEAD.
1.59.4.4	18-Feb-2008	mjf	Sync with HEAD.
1.59.4.3	27-Dec-2007	mjf	Sync with HEAD.
1.59.4.2	08-Dec-2007	mjf	Sync with HEAD.
1.59.4.1	19-Nov-2007	mjf	Sync with HEAD.
1.59.2.3	21-Nov-2007	bouyer	Sync with HEAD
1.59.2.2	18-Nov-2007	bouyer	Sync with HEAD
1.59.2.1	13-Nov-2007	bouyer	Sync with HEAD
1.65.2.3	26-Dec-2007	ad	Sync with head.
1.65.2.2	08-Dec-2007	ad	Sync with head.
1.65.2.1	04-Dec-2007	ad	Pull the vmlocking changes into a new branch.
1.67.4.2	08-Jan-2008	bouyer	Sync with HEAD
1.67.4.1	02-Jan-2008	bouyer	Sync with HEAD
1.70.28.1	21-Apr-2010	matt	sync to netbsd-5
1.70.20.3	17-Sep-2011	bouyer	Pull up following revision(s) (requested by manu in ticket #1666): sys/fs/puffs/puffs_sys.h: revision 1.78 via patch sys/fs/puffs/puffs_node.c: revision 1.20 via patch sys/fs/puffs/puffs_vnops.c: revision 1.155 via patch Add a mutex for operations that touch size (setattr, getattr, write, fsync). This is required to avoid data corruption bugs, where a getattr slices itself within a setattr operation, and sets the size to the stall value it got from the filesystem. That value is smaller than the one set by setattr, and the call to uvm_vnp_setsize() trigged a spurious truncate. The result is a chunk of zeroed data in the file. Such a situation can easily happen when the ioflush thread issue a VOP_FSYNC/puffs_vnop_sync/flushvncache/dosetattrn while andother process do a sys_stat/VOP_GETATTR/puffs_vnop_getattr. This mutex on size operation can be removed the day we decide VOP_GETATTR has to operated on a locked vnode, since the other operations that touch size already require that.
1.70.20.2	18-Jun-2011	bouyer	Pull up following revision(s) (requested by manu in ticket #1623): lib/libpuffs/puffs.c: revision 1.116 sys/fs/puffs/puffs_vnops.c: revision 1.151 Call advlock method if supplied
1.70.20.1	09-Jan-2010	snj	Pull up following revision(s) (requested by pooka in ticket #1212): sys/fs/puffs/puffs_msgif.c: revision 1.76 via patch sys/fs/puffs/puffs_sys.h: revision 1.73 via patch sys/fs/puffs/puffs_vfsops.c: revision 1.84 via patch Process flush requests from the file server in a separate thread context. This fixes a long-standing but seldomly seen deadlock, where the kernel was holding pages busy (due to e.g. readahead request) while waiting for the server to respond, and the server made a callback into the kernel asking to invalidate those pages. ... or, well, theoretically fixes, since I didn't have any reliable way of repeating the deadlock and I think I saw it only twice.
1.70.10.2	11-Aug-2010	yamt	sync with head.
1.70.10.1	11-Mar-2010	yamt	sync with head
1.75.4.1	05-Mar-2011	rmind	sync with head
1.75.2.1	17-Aug-2010	uebayasi	Sync with HEAD.
1.78.8.4	27-Feb-2015	martin	Pull up following revision(s) (requested by manu in ticket #1260): lib/libpuffs/puffs.3: revision 1,55,1.60 sys/fs/puffs/puffs_msgif.h: revision 1.84 lib/libperfuse/ops.c: revision 1.83 sys/fs/puffs/puffs_sys.h: revision 1.89 sys/fs/puffs/puffs_vfsops.c: revision 1.116 lib/libperfuse/perfuse.c: revision 1.36 sys/fs/puffs/puffs_vnops.c: revision 1.200-1.202 Use more markup. New sentence, new line. Bump date for previous. Add PUFFS_KFLAG_NOFLUSH_META to prevent sending metadata flush to FUSE FUSE filesystems do not expect to get metadata updates for [amc]time and size, they updates the value on their own after operations. The PUFFS PUFFS_KFLAG_NOFLUSH_META option prevents regular metadata cache flushes to the filesystem , and libperfuse uses it to match Linux FUSE behavior. While there, fix a bug in SETATTR: do not update kernel metadata cache from SETATTR reply when the request is asynchronous, as we do not have the reply yet. Update file size after write without metadata flush If we do not use metadata flush, we must make sure the size is updated in the filesystem after a write, otherwise the next GETATTR will get us a stale value and the file will be truncated.
1.78.8.3	03-Nov-2014	msaitoh	Pull up following revision(s) (requested by manu in ticket #1140): lib/libperfuse/ops.c 1.63-1.69 lib/libperfuse/perfuse.c 1.32-1.33 lib/libperfuse/perfuse_priv.h 1.32-1.34 lib/libperfuse/subr.c 1.20 lib/libpuffs/creds.c 1.16 lib/libpuffs/dispatcher.c 1.47 lib/libpuffs/puffs.h 1.125 lib/libpuffs/puffs_ops.3 1.37-1.38 lib/libpuffs/requests.c 1.24 sys/fs/puffs/puffs_msgif.h 1.81 sys/fs/puffs/puffs_sys.h 1.85 sys/fs/puffs/puffs_vnops.c 1.183 usr.sbin/perfused/msg.c 1.22 Bring libpuffs, libperfuse and perfused on par with -current: - implement FUSE direct I/O - remove useless code and warnings - fix missing GETATTR bugs - fix exended attribute get and list operations
1.78.8.2	12-Aug-2012	martin	Pull up following revision(s) (requested by manu in ticket #438): lib/libperfuse/perfuse_priv.h: revision 1.31 sys/fs/puffs/puffs_msgif.h: revision 1.80 sys/fs/puffs/puffs_vnops.c: revision 1.171 lib/libpuffs/puffs_ops.3: revision 1.31 sys/fs/puffs/puffs_vnops.c: revision 1.172 sys/fs/puffs/puffs_vnops.c: revision 1.173 sys/fs/puffs/puffs_vnops.c: revision 1.174 usr.sbin/perfused/perfused.c: revision 1.24 sys/fs/puffs/puffs_sys.h: revision 1.80 sys/fs/puffs/puffs_sys.h: revision 1.81 sys/fs/puffs/puffs_sys.h: revision 1.82 lib/libperfuse/subr.c: revision 1.19 lib/libperfuse/perfuse.c: revision 1.30 sys/fs/puffs/puffs_msgif.c: revision 1.90 sys/fs/puffs/puffs_msgif.c: revision 1.91 sys/fs/puffs/puffs_msgif.c: revision 1.92 lib/libperfuse/ops.c: revision 1.59 lib/libpuffs/puffs.3: revision 1.53 lib/libperfuse/debug.c: revision 1.12 lib/libpuffs/puffs.3: revision 1.54 sys/fs/puffs/puffs_vnops.c: revision 1.167 sys/fs/puffs/puffs_msgif.h: revision 1.79 usr.sbin/perfused/msg.c: revision 1.21 sys/fs/puffs/puffs_vfsops.c: revision 1.102 sys/fs/puffs/puffs_vfsops.c: revision 1.103 sys/fs/puffs/puffs_vfsops.c: revision 1.105 lib/libpuffs/puffs.h: revision 1.123 lib/libperfuse/perfuse_if.h: revision 1.20 lib/libperfuse/perfuse.c: revision 1.29 lib/libpuffs/dispatcher.c: revision 1.42 lib/libpuffs/dispatcher.c: revision 1.43 - Fix same vnodes associated with multiple cookies The scheme used to retreive known nodes on lookup was flawed, as it only used parent and name. This produced a different cookie for the same file if it was renamed, when looking up ../ or when dealing with multiple files associated with the same name through link(2). We therefore abandon the use of node name and introduce hashed lists of inodes. This causes a huge rewrite of reclaim code, which do not attempt to keep parents allocated until all their children are reclaimed - Fix race conditions in reclaim There are a few situations where we issue multiple FUSE operations for a PUFFS operation. On reclaim, we therefore have to wait for all FUSE operation to complete, not just the current exchanges. We do this by introducing node reference count with node_ref() and node_rele(). - Detect data loss caused by FAF VOP_PUTPAGES causes FAF writes where the kernel does not check the operation result. At least issue a warning on error. - Enjoy FAF shortcut on setattr No need to wait for the result if the kernel does not want it. There is however an exception for setattr that touch the size, we need to wait for completion because we have other operations queued for after the resize. - Fix fchmod() on write-open file fchmod() on a node open with write privilege will send setattr with both mode and size set. This confuses some FUSE filesystem. Therefore we send two FUSE operations, one for mode, and one for size. - Remove node TTL handling for netbsd-5 for simplicity sake. The code still builds on netbsd-5 but does not have the node TTL feature anymore. It works fine with kernel support on netbsd-6. - Improve PUFFS_KFLAG_CACHE_FS_TTL by reclaiming older inactive nodes. The normal kernel behavior is to retain inactive nodes in the freelist until it runs out of vnodes. This has some merit for local filesystems, where the cost of an allocation is about the same as the cost of a lookup. But that situation is not true for distributed filesystems. On the other hand, keeping inactive nodes for a long time hold memory in the file server process, and when the kernel runs out of vnodes, it produce reclaim avalanches that increase lattency for other operations. We do not reclaim inactive vnodes immediatly either, as they may be looked up again shortly. Instead we introduce a grace time and we reclaim nodes that have been inactive beyond the grace time. - Fix lookup/reclaim race condition. The above improvement undercovered a race condition between lookup and reclaim. If we reclaimed a vnode associated with a userland cookie while a lookup returning that same cookiewas inprogress, then the kernel ends up with a vnode associated with a cookie that has been reclaimed in userland. Next operation on the cookie will crash (or at least confuse) the filesystem. We fix this by introducing a lookup count in kernel and userland. On reclaim, the kernel sends the count, which enable userland to detect situation where it initiated a lookup that is not completed in kernel. In such a situation, the reclaim must be ignored, as the node is about to be looked up again. Fix hang unmount bug introduced by last commit. We introduced a slow queue for delayed reclaims, while the existing queue for unmount, flush and exist has been renamed fast queue. Both queues had timestamp for when an operation should be done, but it was useless for the fast queue, which is always used to run an operation ASAP. And the timestamp test had an error that turned ASAP into "at next tick", but nobody what there to wake the thread at next tick, hence the hang. The fix is to remove the useless and buggy timestamp test for fast queue. Rename slow sopreq queue into node sopreq queue, to refet the fact that is only intended for postponed node reclaims. When purging the node sopreq queue, do not call puffs_msg_sendresp(), as it makes no sense. Fix race condition between (create\|mknod\|mkdir\|symlino) and reclaim, just like we did it between lookup and reclaim. Missing bit in previous commit (prevent race between create\|mknod\|mkdir\|symlink and reclaim) Bump date for previous. New sentence, new line; remove trailing whitespace; fix typos; punctuation nits. Add PUFFS_KFLAG_CACHE_DOTDOT so that vnodes hold a reference on their parent, keeping them active, and allowing to lookup .. without sending a request to the filesystem. Enable the featuure for perfused, as this is how FUSE works. Missing bit in previous commit (PUFFS_KFLAG_CACHE_DOTDOT option to avoid looking up ..)
1.78.8.1	23-Apr-2012	riz	Pull up following revision(s) (requested by manu in ticket #195): lib/libskey/skeysubr.c: revision 1.27 lib/libkvm/kvm_getloadavg.c: revision 1.11 lib/libwrap/update.c: revision 1.9 lib/liby/yyerror.c: revision 1.9 lib/libpuffs/puffs_ops.3: revision 1.30 lib/libwrap/misc.c: revision 1.10 lib/libwrap/hosts_access.c: revision 1.20 lib/libpuffs/pnode.c: revision 1.11 lib/libperfuse/subr.c: revision 1.17 lib/libpuffs/pnode.c: revision 1.12 lib/libperfuse/subr.c: revision 1.18 lib/libwrap/options.c: revision 1.15 lib/libwrap/fix_options.c: revision 1.11 lib/libperfuse/ops.c: revision 1.52 lib/libperfuse/ops.c: revision 1.53 lib/libperfuse/ops.c: revision 1.54 lib/libwrap/hosts_ctl.c: revision 1.5 lib/libintl/gettext.c: revision 1.27 lib/libwrap/shell_cmd.c: revision 1.6 lib/libpuffs/dispatcher.c: revision 1.39 lib/libperfuse/perfuse_priv.h: revision 1.27 lib/libwrap/socket.c: revision 1.19 lib/libpuffs/puffs.3: revision 1.50 lib/libperfuse/perfuse_priv.h: revision 1.28 lib/libpuffs/puffs_priv.h: revision 1.45 lib/libpuffs/puffs.3: revision 1.51 lib/libperfuse/perfuse_priv.h: revision 1.29 lib/libwrap/percent_x.c: revision 1.5 lib/libpuffs/puffs.3: revision 1.52 lib/libperfuse/debug.c: revision 1.11 sys/fs/puffs/puffs_vnops.c: revision 1.165 lib/libwrap/tcpd.h: revision 1.13 sys/fs/puffs/puffs_vnops.c: revision 1.166 lib/libwrap/eval.c: revision 1.7 sys/fs/puffs/puffs_msgif.h: revision 1.78 sys/fs/puffs/puffs_vfsops.c: revision 1.101 lib/libwrap/rfc931.c: revision 1.9 lib/libwrap/clean_exit.c: revision 1.5 lib/libpuffs/puffs.h: revision 1.120 lib/libc/stdlib/jemalloc.c: revision 1.27 lib/librmt/rmtlib.c: revision 1.26 lib/libpuffs/puffs.h: revision 1.121 sys/fs/puffs/puffs_sys.h: revision 1.79 lib/librumpclient/rumpclient.c: revision 1.48 lib/libwrap/refuse.c: revision 1.5 lib/libperfuse/perfuse.c: revision 1.26 lib/libperfuse/perfuse.c: revision 1.27 tests/fs/puffs/t_fuzz.c: revision 1.5 lib/libperfuse/perfuse.c: revision 1.28 lib/libpuffs/dispatcher.c: revision 1.40 sys/fs/puffs/puffs_node.c: revision 1.24 lib/libwrap/diag.c: revision 1.9 lib/libintl/textdomain.c: revision 1.13 Use C89 function definition Add name and atttribute cache with filesytem provided TTL. lookup, create, mknod, mkdir, symlink, getattr and setattr messages have been extended so that attributes and their TTL can be provided by the filesytem. lookup, create, mknod, mkdir, and symlink messages are also extended so that the filesystem can provide name TTL. Add PUFFS_KFLAG_CACHE_FS_TTL flag to puffs_init(3) to use name and attribute cache with filesystem provided TTL. lookup, create, mknod, mkdir, symlink, getattr and setattr messages have been extended so that attributes and their TTL can be provided by the filesytem. lookup, create, mknod, mkdir, and symlink messages are also extended so that the filesystem can provide name TTL. The filesystem updates attributes and TTL using puffs_pn_getvap(3), puffs_pn_getvattl(3), and puffs_pn_getcnttl(3) Use new PUFFS_KFLAG_CACHE_FS_TTL option to puffs_init(3) so that FUSE TTL on name and attributes are used. This save many PUFFS operations and improves performances. PUFFS_KFLAG_CACHE_FS_TTL is #ifdef'ed in many places for now so that libperfuse can still be used on netbsd-5. Split file system. Comma fixes. Remove dangling "and". Bump date for previous. - Makesure update_va does not change vnode size when it should not. For instance when doing a fault-issued VOP_GETPAGES within VOP_WRITE, changing size leads to panic: genfs_getpages: past eof. -Handle ticks wrap around for vnode name andattribute timeout - When using PUFFS_KFLAG_CACHE_FS_TTL, do not use puffs_node to carry attribute and TTL fora newly created node. Instead extend puffs_newinfo and add puffs_newinfo_setva() and puffs_newinfo_setttl() - Remove node_mk_common_final in libperfuse. It used to set uid/gid for a newly created vnode but has been made redundant along time ago since uid and gid are properly set in FUSE header. - In libperfuse, check for corner case where opc = 0 on INACTIVE and RECLAIM (how is it possible? Check for it to avoid a crash anyway) - In libperfuse, make sure we unlimit RLIMIT_AS and RLIMIT_DATA so that we do notrun out of memory because the kernel is lazy at reclaiming vnodes. - In libperfuse, cleanup style of perfuse_destroy_pn() Do not set PUFFS_KFLAG_CACHE_FS_TTL for PUFFS tests
1.78.6.1	29-Apr-2012	mrg	sync to latest -current.
1.78.2.3	22-May-2014	yamt	sync with head. for a reference, the tree before this commit was tagged as yamt-pagecache-tag8. this commit was splitted into small chunks to avoid a limitation of cvs. ("Protocol error: too many arguments")
1.78.2.2	30-Oct-2012	yamt	sync with head
1.78.2.1	17-Apr-2012	yamt	sync with head
1.82.2.3	03-Dec-2017	jdolecek	update from HEAD
1.82.2.2	20-Aug-2014	tls	Rebase to HEAD as of a few days ago.
1.82.2.1	23-Jun-2013	tls	resync from head
1.83.6.1	18-May-2014	rmind	sync with head
1.84.4.3	27-Feb-2015	martin	Pull up following revision(s) (requested by manu in ticket #555): lib/libpuffs/puffs.3: revision 1.60 sys/fs/puffs/puffs_msgif.h: revision 1.84 lib/libperfuse/ops.c: revision 1.83 sys/fs/puffs/puffs_sys.h: revision 1.89 sys/fs/puffs/puffs_vfsops.c: revision 1.116 lib/libperfuse/perfuse.c: revision 1.36 sys/fs/puffs/puffs_vnops.c: revision 1.200-1.202 Add PUFFS_KFLAG_NOFLUSH_META to prevent sending metadata flush to FUSE FUSE filesystems do not expect to get metadata updates for [amc]time and size, they updates the value on their own after operations. The PUFFS PUFFS_KFLAG_NOFLUSH_META option prevents regular metadata cache flushes to the filesystem , and libperfuse uses it to match Linux FUSE behavior. While there, fix a bug in SETATTR: do not update kernel metadata cache from SETATTR reply when the request is asynchronous, as we do not have the reply yet. Update file size after write without metadata flush If we do not use metadata flush, we must make sure the size is updated in the filesystem after a write, otherwise the next GETATTR will get us a stale value and the file will be truncated.
1.84.4.2	29-Aug-2014	martin	Pull up following revision(s) (requested by hannken in ticket #67): sys/fs/puffs/puffs_sys.h: revision 1.86 sys/fs/puffs/puffs_vfsops.c: revision 1.114 sys/fs/puffs/puffs_msgif.c: revision 1.95 sys/fs/puffs/puffs_node.c: revision 1.32 sys/fs/puffs/puffs_vnops.c: revision 1.184 Change puffs from hashlist to vcache. - field "pa_nhashbuckets" of struct "puffs_kargs" becomes a no-op. and should be removed on the next protocol version bump.
1.84.4.1	26-Aug-2014	riz	Pull up following revision(s) (requested by manu in ticket #52): sys/fs/puffs/puffs_msgif.h: revision 1.81 sys/fs/puffs/puffs_sys.h: revision 1.85 sys/fs/puffs/puffs_vnops.c: revision 1.183 Add a oflags input field to open requests so that the filesystem can pass back information about the file. Implement PUFFS_OPEN_IO_DIRECT, which will force direct IO (bypassing page cache) for the file.
1.88.2.2	09-Jul-2016	skrll	Sync with HEAD
1.88.2.1	06-Apr-2015	skrll	Sync with HEAD
1.90.18.1	10-Jun-2019	christos	Sync with HEAD
1.90.16.3	14-Jan-2019	pgoyette	Create a variant of the HOOK macros that handles hook routines of type void, and use them where appropriate.
1.90.16.2	17-Sep-2018	pgoyette	Adapt (most of) the indirect function pointers to the new MP-safe mechanism. Still remaining are the compat_netbsd32 stuff, and some usb subroutines.
1.90.16.1	24-Mar-2018	pgoyette	Add fs/puffs compat_50 to the modules
1.28	10-Nov-2007	pooka	Part 2/n of extensive changes to request transport to/from userspace: Rip the transport code completely out of puffs and generalize it into an independent module which will be used for multiple purposes in the future. This module is called the Pass-to-Userspace Transporter (known as "putter" among friends). This is very much work-in-progress and one dependency with puffs remains: the request framing format. The device name is still /dev/puffs, but that will change soon. Users of puffs need the following in their kernel configs now: pseudo-device putter
1.27	11-Oct-2007	pooka	branches: 1.27.2; 1.27.4; Handle suspend and flush requests from the file server.
1.26	11-Oct-2007	pooka	Part 1/n of some pretty extensive changes to how the kernel module interacts with the userspace file server: * since the kernel-user communication is not purely request-response anymore (hasn't been since 2006), try to rename some "request" to "message". more similar mangling will take place in the future. * completely rework how messages are allocated. previously most of them were borrowed from the stack (originally all of them), but now always allocate dynamically. this makes the structure of the code much cleaner. also makes it possible to fix a locking order violation. it enables plenty of future enhancements. * start generalizing the transport interface to be independent of puffs * move transport interface to read/write instead of ioctl. the old one had legacy design problems, and besides, ioctl's suck. implement a very generic version for now; this will be worked on later hopefully some day reaching "highly optimized". * implement libpuffs support behind existing library request interfaces. this will change eventually (I hate those interfaces)
1.25	04-Oct-2007	pooka	g/c the "sizeop" code previous used for ioctl/fcntl. It was already commented out and has bitrotted beyond all recognition, so it needs complete rethinking.
1.24	27-Sep-2007	pooka	Differentiate between cookie2vnode returning an error and return to caller, address unknown: no such cookie, no such node. Make the callers use this info to either create a new vnode or bail.
1.23	27-Sep-2007	pooka	Fix a race in how new cookies are checked. Previously the checking was done separate of inserting the cookie into the lookup structure and without any form of interlock. This could lead to the same cookie pointing to two different nodes. Remedy the race by creating a separate "checked and ready to be inserted" cookie list which serves as an interlock without having to hold a fs-global creation lock.
1.22	19-Jul-2007	pooka	branches: 1.22.4; 1.22.6; 1.22.8; 1.22.10; define PUFFSREQSIZEOP ioctl, which can be used to fetch the maximum request size
1.21	09-Jul-2007	ad	branches: 1.21.2; Merge some of the less invasive changes from the vmlocking branch: - kthread, callout, devsw API changes - select()/poll() improvements - miscellaneous MT safety improvements
1.20	21-Jun-2007	pooka	Refactor the pnode2vnode translation slightly so that VFS_ROOT can use it directly.
1.19	06-Jun-2007	pooka	Move puffs to a two clause license where it already isn't so. And as agc pointed out, even files with the third clause were already effectively two clause because of a slight bug in the language...
1.18	17-May-2007	pooka	Make it possible for the file server to specify the root vnode type and other information instead of always using VDIR. To make this possible without races, require all root node information already in puffs_mount() and nuke puffs_start2() and the associated start operation completely. requested/inspired by Tobias Nygren
1.17	01-May-2007	pooka	Fix a problem introduced when I converted puffs to use newlock2: when unmounting the file system in case of a certain timing (and possibly some other conditions), a thread would wait on a condition variable, while another thread broadcast the cv and immediately proceeded to destroy it. The result was a system frozen completely solid shorly after the process waiting for the cv woke up. So introduce reference counting to synchronize destruction of the resources in unmount. I was able to repeat the problem only on my laptop in some special cases, so I do not know how common it was. Ironically, killing the file server process violently instead of unmount() didn't have this problem because it never entered the unmount path from two directions.
1.16	16-Apr-2007	pooka	fix comment in previous
1.15	16-Apr-2007	pooka	Allow to set non-blocking mode for transport fd even if the file system is not yet mounted.
1.14	06-Apr-2007	pooka	support flushing pagecache
1.13	06-Apr-2007	pooka	actually, we don't need a separate op for flushing the whole page cache of a node, just use the range op with endoff = 0
1.12	06-Apr-2007	pooka	* enable PUFFS_INVAL_PAGECACHE_NODE_RANGE * add input parameter validation
1.11	30-Mar-2007	pooka	g/c some commented ltsleep calls accidentally left from newlock2 adaptation
1.10	29-Mar-2007	pooka	Convert spinlocks & sleep/wakeup to newlock2 locking stuff. Fix a bunch of bugs. * park structures are now always allocated from a pool instead of a mixed stack/malloc allocation * get rid of the whole adjbuf concept, always just alloc the maximal amount of memory to satisfy a request * little regression: don't allow interrupting wait from file system to userspace; this had problems already before, but now the problems really started to shine through. I'll try to make this work again some day. * fix bmap to return a sensible value in runp
1.9	20-Mar-2007	pooka	* rework the page cache interaction a bit: cache metadata in the kernel and flush it out all at once instead of continuous updating * add support for delivering notifications to the file server about when a page was written to (but disabled by default for now). the file server can use this to request flushing or invalidating the kernel page cache
1.8	16-Feb-2007	hannken	branches: 1.8.2; 1.8.6; 1.8.8; 1.8.10; Make fstrans(9) the default helper for file system suspension. Replaces the now obsolete vn_start_write()/vn_finished_write().
1.7	09-Feb-2007	ad	Merge newlock2 to head.
1.6	28-Jan-2007	pooka	don't need pi_lock for struct member access, so don't take it
1.5	26-Jan-2007	pooka	Initial attempt at suspend/snapshot support for userspace file servers. This is still pretty much on the level "if it breaks ...". It should work for single-threaded servers which handle one operation from start to finish in one go. Also, it does not yet totally correctly synchronize metadata and data in some cases. So needless to say, it needs improvement, but it is possible that will have to wait for some lock revampage.
1.4	09-Jan-2007	pooka	branches: 1.4.2; Introduce flush operations, which the fs server can use to control kernel caching. Currently supported are only flushing the name cache for a directory or flushing the name cache for the entire fs. Also, get rid of PNODE_INACTIVE status, since it was racy and essentially didn't work. All this on top of being useless in the first place ....
1.3	10-Dec-2006	pooka	branches: 1.3.2; * free puffs_instance structure in all cases when closing the descriptor * comment, rcsid & kassert police
1.2	10-Dec-2006	pooka	Fix a race condition that would cause the mountpoint to be cleaned from under someone waiting for the fs server response in puffs_unmount() if the descriptor was closed during the response wait (such as bug leading to a crash in fs implementation unmount()).
1.1	05-Dec-2006	pooka	branches: 1.1.2; shuffle functions around a bit: move the transport (/dev/puffs) to a different file from the messaging (request contents). no functional change
1.1.2.3	18-Dec-2006	yamt	sync with head.
1.1.2.2	10-Dec-2006	yamt	sync with head.
1.1.2.1	05-Dec-2006	yamt	file puffs_transport.c was added on branch yamt-splraiseipl on 2006-12-10 07:18:38 +0000
1.3.2.6	15-Nov-2007	yamt	sync with head.
1.3.2.5	27-Oct-2007	yamt	sync with head.
1.3.2.4	03-Sep-2007	yamt	sync with head.
1.3.2.3	26-Feb-2007	yamt	sync with head.
1.3.2.2	30-Dec-2006	yamt	sync with head.
1.3.2.1	10-Dec-2006	yamt	file puffs_transport.c was added on branch yamt-lazymbuf on 2006-12-30 20:50:01 +0000
1.4.2.4	01-Feb-2007	ad	Sync with head.
1.4.2.3	18-Jan-2007	christos	make things compile.
1.4.2.2	12-Jan-2007	ad	Sync with head.
1.4.2.1	09-Jan-2007	ad	file puffs_transport.c was added on branch newlock2 on 2007-01-12 01:04:05 +0000
1.8.10.1	29-Mar-2007	reinoud	Pullup to -current
1.8.8.1	11-Jul-2007	mjf	Sync with head.
1.8.6.11	12-Oct-2007	ad	Sync with head.
1.8.6.10	09-Oct-2007	ad	Sync with head.
1.8.6.9	20-Aug-2007	ad	Sync with HEAD.
1.8.6.8	15-Jul-2007	ad	Sync with head.
1.8.6.7	09-Jun-2007	ad	Sync with head.
1.8.6.6	08-Jun-2007	ad	Sync with head.
1.8.6.5	13-May-2007	ad	- Pass the error number and residual count to biodone(), and let it handle setting error indicators. Prepare to eliminate B_ERROR. - Add a flag argument to brelse() to be set into the buf's flags, instead of doing it directly. Typically used to set B_INVAL. - Add a "struct cpu_info *" argument to kthread_create(), to be used to create bound threads. Change "bool mpsafe" to "int flags". - Allow exit of LWPs in the IDL state when (l != curlwp). - More locking fixes & conversion to the new API.
1.8.6.4	10-Apr-2007	ad	Sync with head.
1.8.6.3	10-Apr-2007	ad	Nuke the deferred kthread creation stuff, as it's no longer needed. Pointed out by thorpej@.
1.8.6.2	09-Apr-2007	ad	- Add two new arguments to kthread_create1: pri_t pri, bool mpsafe. - Fork kthreads off proc0 as new LWPs, not new processes.
1.8.6.1	13-Mar-2007	ad	Pull in the initial set of changes for the vmlocking branch.
1.8.2.3	07-May-2007	yamt	sync with head.
1.8.2.2	15-Apr-2007	yamt	sync with head.
1.8.2.1	24-Mar-2007	yamt	sync with head.
1.21.2.1	15-Aug-2007	skrll	Sync with HEAD.
1.22.10.2	19-Jul-2007	pooka	define PUFFSREQSIZEOP ioctl, which can be used to fetch the maximum request size
1.22.10.1	19-Jul-2007	pooka	file puffs_transport.c was added on branch matt-mips64 on 2007-07-19 07:52:46 +0000
1.22.8.2	14-Oct-2007	yamt	sync with head.
1.22.8.1	06-Oct-2007	yamt	sync with head.
1.22.6.2	23-Mar-2008	matt	sync with HEAD
1.22.6.1	06-Nov-2007	matt	sync with HEAD
1.22.4.4	11-Nov-2007	joerg	Sync with HEAD.
1.22.4.3	26-Oct-2007	joerg	Sync with HEAD. Follow the merge of pmap.c on i386 and amd64 and move pmap_init_tmp_pgtbl into arch/x86/x86/pmap.c. Modify the ACPI wakeup code to restore CR4 before jumping back into kernel space as the large page option might cover that.
1.22.4.2	07-Oct-2007	joerg	Sync with HEAD.
1.22.4.1	02-Oct-2007	joerg	Sync with HEAD.
1.27.4.1	19-Nov-2007	mjf	Sync with HEAD.
1.27.2.1	13-Nov-2007	bouyer	Sync with HEAD
1.126	01-Apr-2021	christos	Put a copy of our existing data first in the non-error case (noticed by RVP).
1.125	27-Feb-2020	ad	branches: 1.125.6; 1.125.8; Tighten up the locking around vp->v_iflag a little more after the recent split of vmobjlock & v_interlock.
1.124	17-Jan-2020	ad	VFS_VGET(), VFS_ROOT(), VFS_FHTOVP(): give them a "int lktype" argument, to allow us to get shared locks (or no lock) on the returned vnode. Matches FreeBSD.
1.123	27-Sep-2019	christos	branches: 1.123.2; Fix copying issue that was causing errors in unit_test puffs_tstavfs by removing code.
1.122	23-Sep-2019	christos	Restore binary compatibility by using the statvfs90 structure internally.
1.121	28-May-2018	chs	branches: 1.121.2; add a genfs method to allow a file system to limit the range of pages that are given to a single GOP_WRITE() call. needed by ZFS.
1.120	01-Apr-2017	riastradh	branches: 1.120.12; KASSERT(mutex_owned(vp->v_interlock)) in vnode iterator selector.
1.119	17-Feb-2017	hannken	Add generic genfs_suspendctl() and use it for all file systems. Layered file systems need work.
1.118	20-Dec-2015	christos	branches: 1.118.2; 1.118.4; PR/50573: Andreas Gustafsson: puffs can crash kernel for lack of argument checking
1.117	16-Feb-2015	martin	Remove debug printf
1.116	15-Feb-2015	manu	Add PUFFS_KFLAG_NOFLUSH_META to prevent sending metadata flush to FUSE FUSE filesystems do not expect to get metadata updates for [amc]time and size, they updates the value on their own after operations. The PUFFS PUFFS_KFLAG_NOFLUSH_META option prevents regular metadata cache flushes to the filesystem , and libperfuse uses it to match Linux FUSE behavior. While there, fix a bug in SETATTR: do not update kernel metadata cache from SETATTR reply when the request is asynchronous, as we do not have the reply yet.
1.115	10-Nov-2014	maxv	branches: 1.115.2; Do not uselessly include <sys/malloc.h>.
1.114	28-Aug-2014	hannken	Change puffs from hashlist to vcache. - field "pa_nhashbuckets" of struct "puffs_kargs" becomes a no-op. and should be removed on the next protocol version bump.
1.113	25-May-2014	christos	branches: 1.113.2; use standard dirty vnode test.
1.112	25-May-2014	hannken	The pageflush_selector gets a vnode with v_interlock held. Remove the mutex_enter()/mutex_exit() and simplify. Hi christos...
1.111	24-May-2014	christos	Introduce a selector function to the vfs vnode iterator so that we don't need to vget() vnodes that we are not interested at, and optimize locking a bit. Iterator changes reviewed by Hannken (thanks), the rest of the bugs are mine.
1.110	16-Apr-2014	maxv	An (un)privileged user can easily make the kernel dereference a NULL pointer. The kernel allows 'data' to be NULL; it's the fs's responsibility to ensure that it isn't NULL (if the fs actually needs data). ok christos@
1.109	23-Mar-2014	hannken	branches: 1.109.2; Change all vfsops to use C99 designated initializers. No functional changes intended.
1.108	17-Mar-2014	hannken	Change pageflush() to use vfs_vnode_iterator.
1.107	16-Jan-2013	pooka	branches: 1.107.2; Do the protocol consistency check hack only when compiling ELF.
1.106	09-Aug-2012	manu	branches: 1.106.2; Backout previous bugfix attempt for unmounts. That changes did not address the real problem.
1.105	27-Jul-2012	manu	Rename slow sopreq queue into node sopreq queue, to refet the fact that is only intended for postponed node reclaims. When purging the node sopreq queue, do not call puffs_msg_sendresp(), as it makes no sense.
1.104	27-Jul-2012	manu	puffs mounts share global pools. This means that the puffs_vfsops cannot be vfs_detach'ed by module autounload before puffs_vfsop_unmount() completes and has freed ressource from the pools. By holding a reference on puffs_vfsops from each mount, we ensure that no race can occur here. Works around the crash in kern/46734
1.103	22-Jul-2012	manu	Fix hang unmount bug introduced by last commit. We introduced a slow queue for delayed reclaims, while the existing queue for unmount, flush and exist has been renamed fast queue. Both queues had timestamp for when an operation should be done, but it was useless for the fast queue, which is always used to run an operation ASAP. And the timestamp test had an error that turned ASAP into "at next tick", but nobody what there to wake the thread at next tick, hence the hang. The fix is to remove the useless and buggy timestamp test for fast queue.
1.102	21-Jul-2012	manu	- Improve PUFFS_KFLAG_CACHE_FS_TTL by reclaiming older inactive nodes. The normal kernel behavior is to retain inactive nodes in the freelist until it runs out of vnodes. This has some merit for local filesystems, where the cost of an allocation is about the same as the cost of a lookup. But that situation is not true for distributed filesystems. On the other hand, keeping inactive nodes for a long time hold memory in the file server process, and when the kernel runs out of vnodes, it produce reclaim avalanches that increase lattency for other operations. We do not reclaim inactive vnodes immediatly either, as they may be looked up again shortly. Instead we introduce a grace time and we reclaim nodes that have been inactive beyond the grace time. - Fix lookup/reclaim race condition. The above improvement undercovered a race condition between lookup and reclaim. If we reclaimed a vnode associated with a userland cookie while a lookup returning that same cookiewas inprogress, then the kernel ends up with a vnode associated with a cookie that has been reclaimed in userland. Next operation on the cookie will crash (or at least confuse) the filesystem. We fix this by introducing a lookup count in kernel and userland. On reclaim, the kernel sends the count, which enable userland to detect situation where it initiated a lookup that is not completed in kernel. In such a situation, the reclaim must be ignored, as the node is about to be looked up again.
1.101	08-Apr-2012	manu	Add name and atttribute cache with filesytem provided TTL. lookup, create, mknod, mkdir, symlink, getattr and setattr messages have been extended so that attributes and their TTL can be provided by the filesytem. lookup, create, mknod, mkdir, and symlink messages are also extended so that the filesystem can provide name TTL.
1.100	19-Oct-2011	manu	branches: 1.100.2; 1.100.6; 1.100.8; Remove #ifdef DIAGNOSTIC guards around KASSERT, as the macro contains them
1.99	18-Oct-2011	manu	Make sure pagedaemon does not sleep for memory in puffs_vnop_sleep. Add KASSERT on any sleeping memory allocation to check it cannot happen again.
1.98	07-Oct-2011	hannken	As vnalloc() always allocates with PR_WAITOK there is no longer the need to test its result for NULL.
1.97	21-Sep-2011	manu	Make sure ioflush does not sleep in PUFFS code path, waiting for a mutex, a memory allocation, or a response from the filesystem. This avoids deadlocks in the following situations: 1) when memory is low: ioflush waits the fileystem, the fielsystem waits for memory 2) when the filesystem does not respond (e.g.: network outage ona distributed filesystem)
1.96	12-Jun-2011	rmind	Welcome to 5.99.53! Merge rmind-uvmplock branch: - Reorganize locking in UVM and provide extra serialisation for pmap(9). New lock order: [vmpage-owner-lock] -> pmap-lock. - Simplify locking in some pmap(9) modules by removing P->V locking. - Use lock object on vmobjlock (and thus vnode_t::v_interlock) to share the locks amongst UVM objects where necessary (tmpfs, layerfs, unionfs). - Rewrite and optimise x86 TLB shootdown code, make it simpler and cleaner. Add TLBSTATS option for x86 to collect statistics about TLB shootdowns. - Unify /dev/mem et al in MI code and provide required locking (removes kernel-lock on some ports). Also, avoid cache-aliasing issues. Thanks to Andrew Doran and Joerg Sonnenberger, as their initial patches formed the core changes of this branch.
1.95	21-Jul-2010	hannken	branches: 1.95.6; Make holding v_interlock mandatory for callers of vget(). Announced some time ago on tech-kern.
1.94	15-Jul-2010	pooka	f_namemax is one of the static fields overridden by copy_statvfs_info(), so be sure to set it to the value coming from the file server as part of mount args. exposed, like so many other problems, by njoly's tests
1.93	06-Jul-2010	pooka	Add compat to enable running puffs in a 64bit time_t kernel against a server which runs in 32bit time_t namespace.
1.92	06-Jul-2010	pooka	ctassert size of some key structures does not change
1.91	06-Jul-2010	pooka	Make sure that pa_spare is zero-filled and does not contain any garbage which might disrupt future use.
1.90	24-Jun-2010	hannken	Clean up vnode lock operations pass 2: VOP_UNLOCK(vp, flags) -> VOP_UNLOCK(vp): Remove the unneeded flags argument. Welcome to 5.99.32. Discussed on tech-kern.
1.89	21-May-2010	pooka	Since libpuffs needs a major bump for extattr support anyway, make some changes to the user-kernel protocol. Namely, try to be a little more resilient some future changes.
1.88	21-May-2010	pooka	Support extended attributes.
1.87	17-Feb-2010	pooka	branches: 1.87.2; * add a rant about why MPSAFE isn't enabled even though puffs code is * predict_false that we are mounting when calling statvfs * KNF
1.86	14-Jan-2010	pooka	branches: 1.86.2; In case the operations thread has exited, do not queue any more operations. This prevents kernel memory leaks (one of which happened every time the file system was unmounted via PUFFSOP_UNMOUNT ... and incidentally would've been trivially caught with the old malloc(9) interface. I wonder if the message is to use a ton of pools instead of regression-attractive kmem interface).
1.85	07-Jan-2010	pooka	Rename PUFFS_SOPREQ_EXIT to PUFFS_SOPREQSYS_EXIT to better signal it comes from within the kernel instead of as a direct result of a user request. no functional change
1.84	07-Dec-2009	pooka	Process flush requests from the file server in a separate thread context. This fixes a long-standing but seldomly seen deadlock, where the kernel was holding pages busy (due to e.g. readahead request) while waiting for the server to respond, and the server made a callback into the kernel asking to invalidate those pages. ... or, well, theoretically fixes, since I didn't have any reliable way of repeating the deadlock and I think I saw it only twice.
1.83	05-Nov-2009	pooka	Kill suspend support. It was never implemented correctly: * it depended on the biglock (in a very cruel way) * it was attached to userspace transactions rather than logical fs operations (If someone wants to revisit it some day, most of the stuff can be reused from cvs history)
1.82	18-Mar-2009	cegger	Ansify function definitions w/o arguments. Generated with sed.
1.81	20-May-2008	jmcneill	branches: 1.81.6; 1.81.8; 1.81.12; 1.81.16; Add module dependency on putter.
1.80	10-May-2008	rumble	Convert file systems to dynamically attach with the new module interface. Make VFS hooks dynamic while we're here and say farewell to VFS_ATTACH and VFS_HOOKS_ATTACH linksets. As a consequence, most of the file systems can now be loaded as new style modules. Quick sanity check by ad@.
1.79	29-Apr-2008	ad	branches: 1.79.2; PR kern/38057 ffs makes assuptions about devvp file system PR kern/33406 softdeps get stuck in endless loop Introduce VFS_FSYNC() and call it when syncing a block device, if it has a mounted file system.
1.78	28-Jan-2008	dholland	branches: 1.78.6; 1.78.8; 1.78.10; Fix some race conditions in rename. Introduce a per-FS rename lock and new vfsops to manipulate it. Get this lock while renaming. Also add another relookup() in do_sys_rename, which is a hack to kludge around some of the worst deficiencies of ufs_rename. reviewed-by: pooka (and an earlier rev by ad) posted on tech-kern with no objections.
1.77	03-Jan-2008	pooka	fix vmlocking2 fallout: fstrans_mount/unmount
1.76	03-Jan-2008	pooka	valloc -> vnalloc, vfree -> vnfree Avoids collision with userland valloc(3). no functional change ad ok
1.75	02-Jan-2008	pooka	More type-punning workarounds. Curiously the kernel compilation flags cause gcc to not complain.
1.74	02-Jan-2008	ad	Merge vmlocking2 to head.
1.73	30-Dec-2007	pooka	namespace a bit: vfsops -> puffs_vfsop_x() and vops -> puffs_vnop_x()
1.72	27-Nov-2007	pooka	branches: 1.72.2; 1.72.6; Remove "puffs_cid" from the puffs interface following l-removal from the kernel vfs interfaces. puffs_cc_getcaller(pcc) can be used now should the same information be desired.
1.71	26-Nov-2007	pooka	Remove the "struct lwp *" argument from all VFS and VOP interfaces. The general trend is to remove it from all kernel interfaces and this is a start. In case the calling lwp is desired, curlwp should be used. quick consensus on tech-kern
1.70	20-Nov-2007	pooka	Retire M_PUFFS, use kmem(9) instead.
1.69	16-Nov-2007	pooka	Restructure the messaging interface a bit more: make all interfacing with the file server happen through puffs_msg_enqueue() and puffs_msg_wait() instead of having a billion different routines. Build the existing system upon these two. Most importantly though, decouple insertation into the op queue from the actual wait. This is useful for a number of reasons coming soon to a cvs repo near you.
1.68	12-Nov-2007	pooka	* split the putter header into a kernel version and a userland version + install latter to /usr/include/dev/putter * remove last dependencies to puffs from putter, it's completely independent now
1.67	12-Nov-2007	pooka	Move putter code from directly under dev/ to dev/putter/ no functional change
1.66	10-Nov-2007	pooka	Part 2/n of extensive changes to request transport to/from userspace: Rip the transport code completely out of puffs and generalize it into an independent module which will be used for multiple purposes in the future. This module is called the Pass-to-Userspace Transporter (known as "putter" among friends). This is very much work-in-progress and one dependency with puffs remains: the request framing format. The device name is still /dev/puffs, but that will change soon. Users of puffs need the following in their kernel configs now: pseudo-device putter
1.65	11-Oct-2007	pooka	branches: 1.65.2; 1.65.4; Handle suspend and flush requests from the file server.
1.64	11-Oct-2007	pooka	in case of version mismatch, print the numbers
1.63	11-Oct-2007	pooka	Part 1/n of some pretty extensive changes to how the kernel module interacts with the userspace file server: * since the kernel-user communication is not purely request-response anymore (hasn't been since 2006), try to rename some "request" to "message". more similar mangling will take place in the future. * completely rework how messages are allocated. previously most of them were borrowed from the stack (originally all of them), but now always allocate dynamically. this makes the structure of the code much cleaner. also makes it possible to fix a locking order violation. it enables plenty of future enhancements. * start generalizing the transport interface to be independent of puffs * move transport interface to read/write instead of ioctl. the old one had legacy design problems, and besides, ioctl's suck. implement a very generic version for now; this will be worked on later hopefully some day reaching "highly optimized". * implement libpuffs support behind existing library request interfaces. this will change eventually (I hate those interfaces)
1.62	11-Oct-2007	pooka	Cache vnode member variables necessary for operations after the userspace call, namely our private mount structure, in the activation record. This avoids problems in situations where the userspace file server happens to die during our upcall and the vnode is forcibly reclaimed before we roll back to the current stack frame.
1.61	09-Oct-2007	pooka	g/c more unused stuff
1.60	01-Oct-2007	pooka	* better error checking: validate error values received from userland to be vaild errno values * include string describing error in PUFFS_ERR * get rid of union in puffs_req, it's nothing but trouble * pass pmp to async i/o callbacks
1.59	27-Sep-2007	pooka	Differentiate between cookie2vnode returning an error and return to caller, address unknown: no such cookie, no such node. Make the callers use this info to either create a new vnode or bail.
1.58	27-Sep-2007	pooka	Add error notifications, which are used to deliver errors from the kernel to the file server for silly things the file server did, e.g. attempting to create a file with size VSIZENOTSET. The file server can handle these as it chooses, but the default action is for it to throw its hands in the air and sing "goodbye, cruel world, it's over, walk on by".
1.57	27-Sep-2007	pooka	Fix a race in how new cookies are checked. Previously the checking was done separate of inserting the cookie into the lookup structure and without any form of interlock. This could lead to the same cookie pointing to two different nodes. Remedy the race by creating a separate "checked and ready to be inserted" cookie list which serves as an interlock without having to hold a fs-global creation lock.
1.56	05-Sep-2007	pooka	branches: 1.56.2; move static KASSERT from mount to init
1.55	04-Sep-2007	pooka	* don't allow the file server to specify a node size to be VSIZENOTSET * KASSERT that VNOVAL == VSIZENOTSET
1.54	23-Aug-2007	pooka	branches: 1.54.2; Add a third type of fh option, passthrough, where the kernel does not attempt to handle struct fid at all and passes it as such to userspace.
1.53	31-Jul-2007	pooka	branches: 1.53.2; 1.53.4; * nuke the nameidata parameter from VFS_MOUNT(). Nobody on tech-kern knew what it was supposed to be used for and wrstuden gave a go-ahead * while rototilling, convert file systems which went easily to use VFS_PROTOS() instead of manually prototyping the methods
1.52	19-Jul-2007	pooka	Make the minimum request size twice the minimum request structure size. Otherwise ops with payload would have no room for payload.
1.51	17-Jul-2007	pooka	branches: 1.51.2; Set a file server supplied file system type in the type field and set the mntfromname to be the place mounted from instead of the type.
1.50	17-Jul-2007	pooka	Make set_statvfs_info() take a parameter for the vfs name instead of always retrieving it from mp->mnt_op->vfs_name christos ok
1.49	14-Jul-2007	dsl	Remove the copyout() of the mount args from puffs_mount(), the buffer supplied is a kernel address. The puffs userspace code has been changed to do a 2nd call with MNT_GETARGS to retrieve the information.
1.48	12-Jul-2007	dsl	Change the VFS_MOUNT() interface so that the 'data' buffer passed to the fs code is a kernel buffer, pass though the length of the buffer as well. Since the length of the userspace buffer isn'it (yet) passed through the mount system call, add a field to the vfsops structure containing the default length. Split sys_mount() for calls from compat code. Ride one of the recent kernel version changes - old fs LKMs will load, but sys_mount() will reject any attempt to use them.
1.47	09-Jul-2007	ad	Merge some of the less invasive changes from the vmlocking branch: - kthread, callout, devsw API changes - select()/poll() improvements - miscellaneous MT safety improvements
1.46	01-Jul-2007	pooka	Instead of supplying a plain pid, supply an abstract struct puffs_cid *, which can currently be used to query the pid and lwpid.
1.45	21-Jun-2007	pooka	Refactor the pnode2vnode translation slightly so that VFS_ROOT can use it directly.
1.44	21-Jun-2007	pooka	Reorganize how the root vnode is fetched so that it doesn't always go through VFS_ROOT() and allow to fetch it without locking it. This allows us to call the cache flush operations also for the root vnode and most notably fixes e.g. a "No such file or directory" for a psshfs root directory ls -l when a file was locally deleted and remotely re-created. Also fix some sloppy programming in root node fetch (mostly cosmetic).
1.43	06-Jun-2007	pooka	Move puffs to a two clause license where it already isn't so. And as agc pointed out, even files with the third clause were already effectively two clause because of a slight bug in the language...
1.42	17-May-2007	pooka	Make it possible for the file server to specify the root vnode type and other information instead of always using VDIR. To make this possible without races, require all root node information already in puffs_mount() and nuke puffs_start2() and the associated start operation completely. requested/inspired by Tobias Nygren
1.41	01-May-2007	pooka	Fix a problem introduced when I converted puffs to use newlock2: when unmounting the file system in case of a certain timing (and possibly some other conditions), a thread would wait on a condition variable, while another thread broadcast the cv and immediately proceeded to destroy it. The result was a system frozen completely solid shorly after the process waiting for the cv woke up. So introduce reference counting to synchronize destruction of the resources in unmount. I was able to repeat the problem only on my laptop in some special cases, so I do not know how common it was. Ironically, killing the file server process violently instead of unmount() didn't have this problem because it never entered the unmount path from two directions.
1.40	16-Apr-2007	pooka	Sanity-check & possibly adjust number of hash buckets already before returning the mount argument structure to userspace.
1.39	16-Apr-2007	pooka	catch invalid size file handles already in the kernel
1.38	16-Apr-2007	pooka	Give the file server the ability to specify the file handle length instead of defining a static length file handle on the framework-level.
1.37	14-Apr-2007	xtraeme	size_t is unsigned, so use zu rather than zd which is for ssize_t, as Matt Thomas pointed out.
1.36	14-Apr-2007	xtraeme	Use zd to printf size_t.
1.35	13-Apr-2007	pooka	Allow file servers to request the number of hash cookie buckets for pnode -> vnode reverse lookup.
1.34	13-Apr-2007	pooka	* add fhlen to kernel argument structure * rename it to puffs_kargs instead of puffs_args
1.33	11-Apr-2007	pooka	* support VFS_FHTOVP and VFS_VPTOFH * support cookies in for VOP_READDIR nfs exporting puffs file systems works now
1.32	29-Mar-2007	pooka	convert to MALLOC_JUSTDEFINE
1.31	29-Mar-2007	pooka	Convert spinlocks & sleep/wakeup to newlock2 locking stuff. Fix a bunch of bugs. * park structures are now always allocated from a pool instead of a mixed stack/malloc allocation * get rid of the whole adjbuf concept, always just alloc the maximal amount of memory to satisfy a request * little regression: don't allow interrupting wait from file system to userspace; this had problems already before, but now the problems really started to shine through. I'll try to make this work again some day. * fix bmap to return a sensible value in runp
1.30	20-Mar-2007	pooka	* rework the page cache interaction a bit: cache metadata in the kernel and flush it out all at once instead of continuous updating * add support for delivering notifications to the file server about when a page was written to (but disabled by default for now). the file server can use this to request flushing or invalidating the kernel page cache
1.29	13-Mar-2007	ad	branches: 1.29.2; Pass an ipl argument to pool_init/POOL_INIT to be used when initializing the pool's lock.
1.28	16-Feb-2007	hannken	branches: 1.28.2; 1.28.6; 1.28.8; Make fstrans(9) the default helper for file system suspension. Replaces the now obsolete vn_start_write()/vn_finished_write().
1.27	29-Jan-2007	hannken	Change fstrans enum types to upper case. No functional change. From Antti Kantee <pooka@netbsd.org>
1.26	26-Jan-2007	pooka	Initial attempt at suspend/snapshot support for userspace file servers. This is still pretty much on the level "if it breaks ...". It should work for single-threaded servers which handle one operation from start to finish in one go. Also, it does not yet totally correctly synchronize metadata and data in some cases. So needless to say, it needs improvement, but it is possible that will have to wait for some lock revampage.
1.25	25-Jan-2007	pooka	don't hold spinlocks (except vnode interlock) when doing vget()
1.24	23-Jan-2007	pooka	fix comment (no functional change)
1.23	19-Jan-2007	hannken	New file system suspension API to replace vn_start_write and vn_finished_write. The suspension helpers are now put into file system specific operations. This means every file system not supporting these helpers cannot be suspended and therefore snapshots are no longer possible. Implemented for file systems of type ffs. The new API is enabled on a kernel option NEWVNGATE. This option is not enabled by default in any kernel config. Presented and discussed on tech-kern with much input from Bill Studenmund <wrstuden@netbsd.org> and YAMAMOTO Takashi <yamt@netbsd.org>. Welcome to 4.99.9 (new vfs op vfs_suspendctl).
1.22	15-Jan-2007	pooka	Store puffs_node's on lists hashed with the cookie value instead of just one flat list.
1.21	09-Jan-2007	pooka	In vfs_sync(), call VOP_PUTPAGES() for dirty vnodes directly instead of rolling around VOP_FSYNC(). The user server will be given the VFS_SYNC instruction and it can do its own equivalent of VOP_FSYNC() if it pleases, no need for the kernel to explicitly issue #{vnodes} FSYNCs.
1.20	09-Jan-2007	pooka	Introduce flush operations, which the fs server can use to control kernel caching. Currently supported are only flushing the name cache for a directory or flushing the name cache for the entire fs. Also, get rid of PNODE_INACTIVE status, since it was racy and essentially didn't work. All this on top of being useless in the first place ....
1.19	09-Jan-2007	pooka	in vfs_sync flush page cache only for vnodes with dirty pages, not for vnodes with pages (dirty or otherwise)
1.18	07-Jan-2007	pooka	vfs sync, flushes regular file data only (user server can take care of flushing any metadata it might have hidden away)
1.17	02-Jan-2007	pooka	* check userspace version and prevent incompatible mount * some general maintenance
1.16	10-Dec-2006	pooka	branches: 1.16.2; Fix a race condition that would cause the mountpoint to be cleaned from under someone waiting for the fs server response in puffs_unmount() if the descriptor was closed during the response wait (such as bug leading to a crash in fs implementation unmount()).
1.15	09-Dec-2006	chs	branches: 1.15.2; a smorgasbord of improvements to vnode locking and path lookup: - LOCKPARENT is no longer relevant for lookup(), relookup() or VOP_LOOKUP(). these now always return the parent vnode locked. namei() works as before. lookup() and various other paths no longer acquire vnode locks in the wrong order via vrele(). fixes PR 32535. as a nice side effect, path lookup is also up to 25% faster. - the above allows us to get rid of PDIRUNLOCK. - also get rid of WANTPARENT (just use LOCKPARENT and unlock it). - remove an assumption in layer_node_find() that all file systems implement a recursive VOP_LOCK() (unionfs doesn't). - require that all file systems supply vfs_vptofh and vfs_fhtovp routines. fill in eopnotsupp() for file systems that don't support being exported and remove the checks for NULL. (layerfs calls these without checking.) - in union_lookup1(), don't change refcounts in the ISDOTDOT case, just adjust which vnode is locked. fixes PR 33374. - apply fixes for ufs_rename() from ufs_vnops.c rev. 1.61 to ext2fs_rename().
1.14	07-Dec-2006	pooka	In case of an error, return an error. Otherwise the worst case was that dostatvfs() wrote to a recently deceased struct mount.
1.13	01-Dec-2006	pooka	branches: 1.13.2; prefix kernel flags with PUFFS_KFLAG to have a separate namespace from the library flags
1.12	01-Dec-2006	pooka	don't call the fs server for all operations, only those it has told us that it implements
1.11	18-Nov-2006	pooka	branches: 1.11.2; Always override f_iosize from stat() to DEV_BSIZE for now. Places such as vnd use the information, so until "dealing with it" is defined, it's overriden by the kernel.
1.10	18-Nov-2006	pooka	prevent value 0 for mnt_stat.f_iosize, it is sometimes used as a divider
1.9	18-Nov-2006	pooka	Require statvfs info from startreq so that we have that info available. Also, don't pass fsid to userspace and just fill it in the kernel.
1.8	17-Nov-2006	pooka	Introduce uncached operation, makes sense when the file system backend can be modified from elsewhere than the file system interface
1.7	09-Nov-2006	pooka	few renames to better differentiate between mount & start.. plus some other renaming
1.6	07-Nov-2006	pooka	attach to genfs & support page cache. most noticeable effect is mmap and therefore execution of binaries starting to work, some speed improvements with large file I/O also. caching semantics and error case handling most likely need revisiting.
1.5	06-Nov-2006	pooka	make it possible to build & load puffs as an LKM by Lubomir Kundrak, PR kern/35000
1.4	27-Oct-2006	pooka	support fifos
1.3	26-Oct-2006	pooka	support specfs
1.2	26-Oct-2006	pooka	debug print fixes
1.1	22-Oct-2006	pooka	kernel portion of puffs - the Pass-to-Userspace Framework File System. It contains the VFS attachment and userspace message-passing interface. This work was initially started and completed for Google SoC 2005 and tweaked to work a bit better in the past few weeks. While being far from complete, it is functional enough to be able and stable to host a fairly general-purpose in-memory file system in userspace. Even so, puffs should be considered experimental and no binary compatibility for interfaces or crash-freedom or zero security implications should be relied upon just yet. The GSoC project was mentored by William Studenmund and the final review for the code was done by Christos.
1.11.2.4	01-Feb-2007	ad	Sync with head.
1.11.2.3	12-Jan-2007	ad	Sync with head.
1.11.2.2	18-Nov-2006	ad	Sync with head.
1.11.2.1	18-Nov-2006	ad	file puffs_vfsops.c was added on branch newlock2 on 2006-11-18 21:39:20 +0000
1.13.2.1	17-Feb-2007	tron	Apply patch (requested by chs in ticket #422): - Fix various deadlock problems with nullfs and unionfs. - Speed up path lookups by upto 25%.
1.15.2.3	18-Dec-2006	yamt	sync with head.
1.15.2.2	10-Dec-2006	yamt	sync with head.
1.15.2.1	09-Dec-2006	yamt	file puffs_vfsops.c was added on branch yamt-splraiseipl on 2006-12-10 07:18:38 +0000
1.16.2.9	04-Feb-2008	yamt	sync with head.
1.16.2.8	21-Jan-2008	yamt	sync with head
1.16.2.7	07-Dec-2007	yamt	sync with head
1.16.2.6	15-Nov-2007	yamt	sync with head.
1.16.2.5	27-Oct-2007	yamt	sync with head.
1.16.2.4	03-Sep-2007	yamt	sync with head.
1.16.2.3	26-Feb-2007	yamt	sync with head.
1.16.2.2	30-Dec-2006	yamt	sync with head.
1.16.2.1	10-Dec-2006	yamt	file puffs_vfsops.c was added on branch yamt-lazymbuf on 2006-12-30 20:50:01 +0000
1.28.8.1	11-Jul-2007	mjf	Sync with head.
1.28.6.12	28-Oct-2007	ad	Fix up mnt_vnodelist handling.
1.28.6.11	12-Oct-2007	ad	Sync with head.
1.28.6.10	09-Oct-2007	ad	Sync with head.
1.28.6.9	20-Aug-2007	ad	Sync with HEAD.
1.28.6.8	15-Jul-2007	ad	Sync with head.
1.28.6.7	17-Jun-2007	ad	- Increase the number of thread priorities from 128 to 256. How the space is set up is to be revisited. - Implement soft interrupts as kernel threads. A generic implementation is provided, with hooks for fast-path MD code that can run the interrupt threads over the top of other threads executing in the kernel. - Split vnode::v_flag into three fields, depending on how the flag is locked (by the interlock, by the vnode lock, by the file system). - Miscellaneous locking fixes and improvements.
1.28.6.6	09-Jun-2007	ad	Sync with head.
1.28.6.5	08-Jun-2007	ad	Sync with head.
1.28.6.4	10-Apr-2007	ad	Sync with head.
1.28.6.3	05-Apr-2007	ad	Compile fixes.
1.28.6.2	13-Mar-2007	ad	Pull in the initial set of changes for the vmlocking branch.
1.28.6.1	13-Mar-2007	ad	Sync with head.
1.28.2.3	07-May-2007	yamt	sync with head.
1.28.2.2	15-Apr-2007	yamt	sync with head.
1.28.2.1	24-Mar-2007	yamt	sync with head.
1.29.2.1	29-Mar-2007	reinoud	Pullup to -current
1.51.2.3	10-Sep-2007	skrll	Sync with HEAD.
1.51.2.2	03-Sep-2007	skrll	Sync with HEAD.
1.51.2.1	15-Aug-2007	skrll	Sync with HEAD.
1.53.4.2	31-Jul-2007	pooka	* nuke the nameidata parameter from VFS_MOUNT(). Nobody on tech-kern knew what it was supposed to be used for and wrstuden gave a go-ahead * while rototilling, convert file systems which went easily to use VFS_PROTOS() instead of manually prototyping the methods
1.53.4.1	31-Jul-2007	pooka	file puffs_vfsops.c was added on branch matt-mips64 on 2007-07-31 21:14:19 +0000
1.53.2.7	27-Nov-2007	joerg	Sync with HEAD. amd64 Xen support needs testing.
1.53.2.6	21-Nov-2007	joerg	Sync with HEAD.
1.53.2.5	14-Nov-2007	joerg	Sync with HEAD.
1.53.2.4	11-Nov-2007	joerg	Sync with HEAD.
1.53.2.3	26-Oct-2007	joerg	Sync with HEAD. Follow the merge of pmap.c on i386 and amd64 and move pmap_init_tmp_pgtbl into arch/x86/x86/pmap.c. Modify the ACPI wakeup code to restore CR4 before jumping back into kernel space as the large page option might cover that.
1.53.2.2	02-Oct-2007	joerg	Sync with HEAD.
1.53.2.1	03-Sep-2007	jmcneill	Sync with HEAD.
1.54.2.3	23-Mar-2008	matt	sync with HEAD
1.54.2.2	09-Jan-2008	matt	sync with HEAD
1.54.2.1	06-Nov-2007	matt	sync with HEAD
1.56.2.2	14-Oct-2007	yamt	sync with head.
1.56.2.1	06-Oct-2007	yamt	sync with head.
1.65.4.3	18-Feb-2008	mjf	Sync with HEAD.
1.65.4.2	08-Dec-2007	mjf	Sync with HEAD.
1.65.4.1	19-Nov-2007	mjf	Sync with HEAD.
1.65.2.3	21-Nov-2007	bouyer	Sync with HEAD
1.65.2.2	18-Nov-2007	bouyer	Sync with HEAD
1.65.2.1	13-Nov-2007	bouyer	Sync with HEAD
1.72.6.2	08-Jan-2008	bouyer	Sync with HEAD
1.72.6.1	02-Jan-2008	bouyer	Sync with HEAD
1.72.2.1	04-Dec-2007	ad	Pull the vmlocking changes into a new branch.
1.78.10.4	11-Aug-2010	yamt	sync with head.
1.78.10.3	11-Mar-2010	yamt	sync with head
1.78.10.2	04-May-2009	yamt	sync with head.
1.78.10.1	16-May-2008	yamt	sync with head.
1.78.8.2	04-Jun-2008	yamt	sync with head
1.78.8.1	18-May-2008	yamt	sync with head.
1.78.6.1	02-Jun-2008	mjf	Sync with HEAD.
1.79.2.1	23-Jun-2008	wrstuden	Sync w/ -current. 34 merge conflicts to follow.
1.81.16.1	21-Apr-2010	matt	sync to netbsd-5
1.81.12.1	13-May-2009	jym	Sync with HEAD. Commit is split, to avoid a "too many arguments" protocol error.
1.81.8.4	25-Apr-2014	sborrill	Pull up the following revisions(s) (requested by maxv in ticket #1901): sys/kern/vfs_syscalls.c: revision 1.478, 1.480 via patch sys/coda/coda_vfsops.c: revision 1.81 sys/fs/adosfs/advfsops.c: revision 1.70 sys/fs/cd9660/cd9660_vfsops.c: revision 1.84 sys/fs/efs/efs_vfsops.c: revision 1.25 sys/fs/filecorefs/filecore_vfsops.c: revision 1.76 sys/fs/hfs/hfs_vfsops.c: revision 1.31 sys/fs/msdosfs/msdosfs_vfsops.c: revision 1.107 sys/fs/ntfs/ntfs_vfsops.c: revision 1.94 sys/fs/ptyfs/ptyfs_vfsops.c: revision 1.50 via patch sys/fs/puffs/puffs_vfsops.c: revision 1.110 via patch sys/fs/smbfs/smbfs_vfsops.c: revision 1.100 sys/fs/sysvbfs/sysvbfs_vfsops.c: revision 1.43 sys/fs/tmpfs/tmpfs_vfsops.c: revision 1.59 via patch sys/fs/udf/udf_vfsops.c: revision 1.67 sys/fs/union/union_vfsops.c: revision 1.72 sys/fs/unionfs/unionfs_vfsops.c: revision 1.13 sys/kern/vfs_syscalls.c: revision 1.479 sys/miscfs/nullfs/null_vfsops.c: revision 1.88 via patch sys/miscfs/overlay/overlay_vfsops.c: revision 1.61 sys/miscfs/procfs/procfs_vfsops.c: revision 1.91 sys/miscfs/umapfs/umap_vfsops.c: revision 1.92 sys/nfs/nfs_vfsops.c: revision 1.227 sys/ufs/ext2fs/ext2fs_vfsops.c: revision 1.180 sys/ufs/ffs/ffs_vfsops.c: revision 1.297 sys/ufs/lfs/lfs_vfsops.c: revision 1.321 sys/ufs/mfs/mfs_vfsops.c: revision 1.107 Due to missing checks in the mount syscall, and a wrong assumption on the file systems side, the kernel could allocate an unbounded or zero-sized memory buffer, and could dereference a NULL pointer when particular arguments are given by a user.
1.81.8.3	02-Nov-2011	riz	branches: 1.81.8.3.2; Pull up following revision(s) (requested by manu in ticket #1679): sys/fs/puffs/puffs_vnops.c: revision 1.157 sys/fs/puffs/puffs_vnops.c: revision 1.158 sys/fs/puffs/puffs_vnops.c: revision 1.159 sys/fs/puffs/puffs_vfsops.c: revision 1.97 sys/fs/puffs/puffs_vfsops.c: revision 1.99 sys/fs/puffs/puffs_vnops.c: revision 1.160 sys/fs/puffs/puffs_vfsops.c: revision 1.100 sys/miscfs/syncfs/sync_subr.c: revision 1.47 sys/fs/puffs/puffs_node.c: revision 1.21 sys/fs/puffs/puffs_node.c: revision 1.22 sys/fs/puffs/puffs_msgif.c: revision 1.88 sys/fs/puffs/puffs_msgif.c: revision 1.89 sys/fs/puffs/puffs_vnops.c: revision 1.156 Make sure ioflush does not sleep in PUFFS code path, waiting for a mutex, a memory allocation, or a response from the filesystem. This avoids deadlocks in the following situations: 1) when memory is low: ioflush waits the fileystem, the fielsystem waits for memory 2) when the filesystem does not respond (e.g.: network outage ona distributed filesystem) Fix the build that was broken by struct lwp *updateproc reference in RUMP-visible code. Instead of checking that updateproc (aka ioflush, aka syncer) will not sleep in PUFFS code, I check for any kernel thread: after all none of them are designed to hang awaiting for a remote filesystem operation to complete. Roll back the change that forced kernel threads to not sleep in PUFFS. The change does not make consensus, since only pagedaemon should need it. Other threads will tolerate sleeping, and problems here are only symptoms that something is going wrong in memory management. The cause, not the symptoms, need to be fixed. Make sure pagedaemon does not sleep for memory in puffs_vnop_sleep. Add KASSERT on any sleeping memory allocation to check it cannot happen again. Remove #ifdef DIAGNOSTIC guards around KASSERT, as the macro contains them
1.81.8.2	17-Jul-2011	riz	Pull up following revision(s) (requested by manu in ticket #1645): lib/libc/sys/Makefile.inc 1.207 via patch lib/libc/sys/extattr_get_file.2 patch lib/libpuffs/dispatcher.c 1.34,1.36 via patch lib/libpuffs/puffs.c 1.107 via patch lib/libpuffs/puffs.h 1.115,1.118 via patch sys/fs/puffs/puffs_msgif.h 1.71,1.76 via patch sys/fs/puffs/puffs_vfsops.c 1.88 via patch sys/fs/puffs/puffs_vnops.c 1.145,1.154 via patch sys/kern/vfs_xattr.c 1.24-1.27 via patch sys/kern/vnode_if.c 1.87 via patch sys/sys/Makefile 1.133 via patch sys/sys/extattr.h 1.6 via patch sys/sys/vnode_if.h 1.81 via patch sys/ufs/ffs/ffs_vnops.c patch sys/ufs/ufs/ufs_extattr.c 1.31,1.34 via patch * support extended attributes * bump major due to structure growth * add some spare space * remove ABI sillyness Support extended attributes. Fix multiple non compliances in our Linux-like extattr API, and make it public so that it can be used. Improve a bit listxattr(2). It attemps to list both system and user extended attributes, and it faled if calling user did not have privilege for reading system EA. Now we just lise user EA and skip system EA in reading them is not allowed. Fix bug introduced in previous commuit: Do not vrele() a vnode we did not obtained. Improve UFS1 extended attributes usability - autocreate attribute backing file for new attributes - autoload attributes when issuing extattrctl start - when autoloading attributes, do not display garbage warning when looking up entries that got ENOENT Add a flag to VOP_LISTEXTATTR(9) so that the vnode interface can tell the filesystem in which format extended attribute shall be listed. There are currently two formats: - NUL-terminated strings, used for listxattr(2), this is the default. - one byte length-pprefixed, non NUL-terminated strings, used for extattr_list_file(2), which is obtanined by setting the EXTATTR_LIST_PREFIXLEN flag to VOP_LISTEXTATTR(9) This approach avoid the need for converting the list back and forth, except in libperfuse, since FUSE uses NUL-terminated strings, and the kernel may have requested EXTATTR_LIST_PREFIXLEN.
1.81.8.1	09-Jan-2010	snj	branches: 1.81.8.1.2; Pull up following revision(s) (requested by pooka in ticket #1212): sys/fs/puffs/puffs_msgif.c: revision 1.76 via patch sys/fs/puffs/puffs_sys.h: revision 1.73 via patch sys/fs/puffs/puffs_vfsops.c: revision 1.84 via patch Process flush requests from the file server in a separate thread context. This fixes a long-standing but seldomly seen deadlock, where the kernel was holding pages busy (due to e.g. readahead request) while waiting for the server to respond, and the server made a callback into the kernel asking to invalidate those pages. ... or, well, theoretically fixes, since I didn't have any reliable way of repeating the deadlock and I think I saw it only twice.
1.81.8.3.2.1	28-Apr-2014	sborrill	Pull up the following revisions(s) (requested by maxv in ticket #1901): sys/kern/vfs_syscalls.c: revision 1.478, 1.480 via patch sys/coda/coda_vfsops.c: revision 1.81 sys/fs/adosfs/advfsops.c: revision 1.70 sys/fs/cd9660/cd9660_vfsops.c: revision 1.84 sys/fs/efs/efs_vfsops.c: revision 1.25 sys/fs/filecorefs/filecore_vfsops.c: revision 1.76 sys/fs/hfs/hfs_vfsops.c: revision 1.31 sys/fs/msdosfs/msdosfs_vfsops.c: revision 1.107 sys/fs/ntfs/ntfs_vfsops.c: revision 1.94 sys/fs/ptyfs/ptyfs_vfsops.c: revision 1.50 via patch sys/fs/puffs/puffs_vfsops.c: revision 1.110 via patch sys/fs/smbfs/smbfs_vfsops.c: revision 1.100 sys/fs/sysvbfs/sysvbfs_vfsops.c: revision 1.43 sys/fs/tmpfs/tmpfs_vfsops.c: revision 1.59 via patch sys/fs/udf/udf_vfsops.c: revision 1.67 sys/fs/union/union_vfsops.c: revision 1.72 sys/fs/unionfs/unionfs_vfsops.c: revision 1.13 sys/kern/vfs_syscalls.c: revision 1.479 sys/miscfs/nullfs/null_vfsops.c: revision 1.88 via patch sys/miscfs/overlay/overlay_vfsops.c: revision 1.61 sys/miscfs/procfs/procfs_vfsops.c: revision 1.91 sys/miscfs/umapfs/umap_vfsops.c: revision 1.92 sys/nfs/nfs_vfsops.c: revision 1.227 sys/ufs/ext2fs/ext2fs_vfsops.c: revision 1.180 sys/ufs/ffs/ffs_vfsops.c: revision 1.297 sys/ufs/lfs/lfs_vfsops.c: revision 1.321 sys/ufs/mfs/mfs_vfsops.c: revision 1.107 Due to missing checks in the mount syscall, and a wrong assumption on the file systems side, the kernel could allocate an unbounded or zero-sized memory buffer, and could dereference a NULL pointer when particular arguments are given by a user.
1.81.8.1.2.1	28-Apr-2014	sborrill	Pull up the following revisions(s) (requested by maxv in ticket #1901): sys/kern/vfs_syscalls.c: revision 1.478, 1.480 via patch sys/coda/coda_vfsops.c: revision 1.81 sys/fs/adosfs/advfsops.c: revision 1.70 sys/fs/cd9660/cd9660_vfsops.c: revision 1.84 sys/fs/efs/efs_vfsops.c: revision 1.25 sys/fs/filecorefs/filecore_vfsops.c: revision 1.76 sys/fs/hfs/hfs_vfsops.c: revision 1.31 sys/fs/msdosfs/msdosfs_vfsops.c: revision 1.107 sys/fs/ntfs/ntfs_vfsops.c: revision 1.94 sys/fs/ptyfs/ptyfs_vfsops.c: revision 1.50 via patch sys/fs/puffs/puffs_vfsops.c: revision 1.110 via patch sys/fs/smbfs/smbfs_vfsops.c: revision 1.100 sys/fs/sysvbfs/sysvbfs_vfsops.c: revision 1.43 sys/fs/tmpfs/tmpfs_vfsops.c: revision 1.59 via patch sys/fs/udf/udf_vfsops.c: revision 1.67 sys/fs/union/union_vfsops.c: revision 1.72 sys/fs/unionfs/unionfs_vfsops.c: revision 1.13 sys/kern/vfs_syscalls.c: revision 1.479 sys/miscfs/nullfs/null_vfsops.c: revision 1.88 via patch sys/miscfs/overlay/overlay_vfsops.c: revision 1.61 sys/miscfs/procfs/procfs_vfsops.c: revision 1.91 sys/miscfs/umapfs/umap_vfsops.c: revision 1.92 sys/nfs/nfs_vfsops.c: revision 1.227 sys/ufs/ext2fs/ext2fs_vfsops.c: revision 1.180 sys/ufs/ffs/ffs_vfsops.c: revision 1.297 sys/ufs/lfs/lfs_vfsops.c: revision 1.321 sys/ufs/mfs/mfs_vfsops.c: revision 1.107 Due to missing checks in the mount syscall, and a wrong assumption on the file systems side, the kernel could allocate an unbounded or zero-sized memory buffer, and could dereference a NULL pointer when particular arguments are given by a user.
1.81.6.1	28-Apr-2009	skrll	Sync with HEAD.
1.86.2.2	17-Aug-2010	uebayasi	Sync with HEAD.
1.86.2.1	30-Apr-2010	uebayasi	Sync with HEAD.
1.87.2.4	05-Mar-2011	rmind	sync with head
1.87.2.3	03-Jul-2010	rmind	sync with head
1.87.2.2	30-May-2010	rmind	sync with head
1.87.2.1	16-Mar-2010	rmind	Change struct uvm_object::vmobjlock to be dynamically allocated with mutex_obj_alloc(). It allows us to share the locks among UVM objects.
1.95.6.1	23-Jun-2011	cherry	Catchup with rmind-uvmplock merge.
1.100.8.4	27-Feb-2015	martin	Pull up following revision(s) (requested by manu in ticket #1260): lib/libpuffs/puffs.3: revision 1,55,1.60 sys/fs/puffs/puffs_msgif.h: revision 1.84 lib/libperfuse/ops.c: revision 1.83 sys/fs/puffs/puffs_sys.h: revision 1.89 sys/fs/puffs/puffs_vfsops.c: revision 1.116 lib/libperfuse/perfuse.c: revision 1.36 sys/fs/puffs/puffs_vnops.c: revision 1.200-1.202 Use more markup. New sentence, new line. Bump date for previous. Add PUFFS_KFLAG_NOFLUSH_META to prevent sending metadata flush to FUSE FUSE filesystems do not expect to get metadata updates for [amc]time and size, they updates the value on their own after operations. The PUFFS PUFFS_KFLAG_NOFLUSH_META option prevents regular metadata cache flushes to the filesystem , and libperfuse uses it to match Linux FUSE behavior. While there, fix a bug in SETATTR: do not update kernel metadata cache from SETATTR reply when the request is asynchronous, as we do not have the reply yet. Update file size after write without metadata flush If we do not use metadata flush, we must make sure the size is updated in the filesystem after a write, otherwise the next GETATTR will get us a stale value and the file will be truncated.
1.100.8.3	21-Apr-2014	bouyer	Pull up following revision(s) (requested by maxv in ticket #1050): sys/ufs/chfs/chfs_vfsops.c: revision 1.11 sys/fs/unionfs/unionfs_vfsops.c: revision 1.13 sys/fs/nilfs/nilfs_vfsops.c: revision 1.16 sys/ufs/mfs/mfs_vfsops.c: revision 1.107 sys/fs/sysvbfs/sysvbfs_vfsops.c: revision 1.43 sys/ufs/ffs/ffs_vfsops.c: revision 1.297 sys/kern/vfs_syscalls.c: revision 1.478 sys/kern/vfs_syscalls.c: revision 1.479 sys/fs/puffs/puffs_vfsops.c: revision 1.110 sys/fs/cd9660/cd9660_vfsops.c: revision 1.84 sys/nfs/nfs_vfsops.c: revision 1.227 sys/fs/v7fs/v7fs_vfsops.c: revision 1.10 sys/ufs/ext2fs/ext2fs_vfsops.c: revision 1.180 sys/miscfs/umapfs/umap_vfsops.c: revision 1.92 sys/fs/filecorefs/filecore_vfsops.c: revision 1.76 sys/miscfs/nullfs/null_vfsops.c: revision 1.88 sys/fs/ptyfs/ptyfs_vfsops.c: revision 1.50 sys/coda/coda_vfsops.c: revision 1.81 sys/ufs/lfs/lfs_vfsops.c: revision 1.321 sys/fs/tmpfs/tmpfs_vfsops.c: revision 1.59 sys/fs/hfs/hfs_vfsops.c: revision 1.31 sys/miscfs/overlay/overlay_vfsops.c: revision 1.61 sys/fs/union/union_vfsops.c: revision 1.72 sys/fs/ntfs/ntfs_vfsops.c: revision 1.94 sys/kern/vfs_syscalls.c: revision 1.480 sys/fs/efs/efs_vfsops.c: revision 1.25 sys/kern/vfs_syscalls.c: revision 1.482 sys/fs/msdosfs/msdosfs_vfsops.c: revision 1.107 external/cddl/osnet/dist/uts/common/fs/zfs/zfs_vfsops.c: revision 1.12 sys/miscfs/procfs/procfs_vfsops.c: revision 1.91 sys/fs/smbfs/smbfs_vfsops.c: revision 1.100 sys/fs/adosfs/advfsops.c: revision 1.70 sys/fs/udf/udf_vfsops.c: revision 1.67 Limit check for 'data_len'. Otherwise a (un)privileged user can easily panic the system by passing a huge size. ok christos@ An (un)privileged user can easily make the kernel dereference a NULL pointer. The kernel allows 'data' to be NULL; it's the fs's responsibility to ensure that it isn't NULL (if the fs actually needs data). ok christos@ Some fs's - like kernfs - set their vfs_min_mount_data to zero. Add a check to prevent an (un)privileged user from requesting a zero-sized allocation (and thus a panic). This thing is totally buggy: 'data_len' is modified by the fs, so calling kmem_free with it while its value has changed since the kmem_alloc is far from being a good idea. If the kernel figures out that something mismatches, it will panic (typically with kernfs).
1.100.8.2	12-Aug-2012	martin	branches: 1.100.8.2.4; 1.100.8.2.6; Pull up following revision(s) (requested by manu in ticket #438): lib/libperfuse/perfuse_priv.h: revision 1.31 sys/fs/puffs/puffs_msgif.h: revision 1.80 sys/fs/puffs/puffs_vnops.c: revision 1.171 lib/libpuffs/puffs_ops.3: revision 1.31 sys/fs/puffs/puffs_vnops.c: revision 1.172 sys/fs/puffs/puffs_vnops.c: revision 1.173 sys/fs/puffs/puffs_vnops.c: revision 1.174 usr.sbin/perfused/perfused.c: revision 1.24 sys/fs/puffs/puffs_sys.h: revision 1.80 sys/fs/puffs/puffs_sys.h: revision 1.81 sys/fs/puffs/puffs_sys.h: revision 1.82 lib/libperfuse/subr.c: revision 1.19 lib/libperfuse/perfuse.c: revision 1.30 sys/fs/puffs/puffs_msgif.c: revision 1.90 sys/fs/puffs/puffs_msgif.c: revision 1.91 sys/fs/puffs/puffs_msgif.c: revision 1.92 lib/libperfuse/ops.c: revision 1.59 lib/libpuffs/puffs.3: revision 1.53 lib/libperfuse/debug.c: revision 1.12 lib/libpuffs/puffs.3: revision 1.54 sys/fs/puffs/puffs_vnops.c: revision 1.167 sys/fs/puffs/puffs_msgif.h: revision 1.79 usr.sbin/perfused/msg.c: revision 1.21 sys/fs/puffs/puffs_vfsops.c: revision 1.102 sys/fs/puffs/puffs_vfsops.c: revision 1.103 sys/fs/puffs/puffs_vfsops.c: revision 1.105 lib/libpuffs/puffs.h: revision 1.123 lib/libperfuse/perfuse_if.h: revision 1.20 lib/libperfuse/perfuse.c: revision 1.29 lib/libpuffs/dispatcher.c: revision 1.42 lib/libpuffs/dispatcher.c: revision 1.43 - Fix same vnodes associated with multiple cookies The scheme used to retreive known nodes on lookup was flawed, as it only used parent and name. This produced a different cookie for the same file if it was renamed, when looking up ../ or when dealing with multiple files associated with the same name through link(2). We therefore abandon the use of node name and introduce hashed lists of inodes. This causes a huge rewrite of reclaim code, which do not attempt to keep parents allocated until all their children are reclaimed - Fix race conditions in reclaim There are a few situations where we issue multiple FUSE operations for a PUFFS operation. On reclaim, we therefore have to wait for all FUSE operation to complete, not just the current exchanges. We do this by introducing node reference count with node_ref() and node_rele(). - Detect data loss caused by FAF VOP_PUTPAGES causes FAF writes where the kernel does not check the operation result. At least issue a warning on error. - Enjoy FAF shortcut on setattr No need to wait for the result if the kernel does not want it. There is however an exception for setattr that touch the size, we need to wait for completion because we have other operations queued for after the resize. - Fix fchmod() on write-open file fchmod() on a node open with write privilege will send setattr with both mode and size set. This confuses some FUSE filesystem. Therefore we send two FUSE operations, one for mode, and one for size. - Remove node TTL handling for netbsd-5 for simplicity sake. The code still builds on netbsd-5 but does not have the node TTL feature anymore. It works fine with kernel support on netbsd-6. - Improve PUFFS_KFLAG_CACHE_FS_TTL by reclaiming older inactive nodes. The normal kernel behavior is to retain inactive nodes in the freelist until it runs out of vnodes. This has some merit for local filesystems, where the cost of an allocation is about the same as the cost of a lookup. But that situation is not true for distributed filesystems. On the other hand, keeping inactive nodes for a long time hold memory in the file server process, and when the kernel runs out of vnodes, it produce reclaim avalanches that increase lattency for other operations. We do not reclaim inactive vnodes immediatly either, as they may be looked up again shortly. Instead we introduce a grace time and we reclaim nodes that have been inactive beyond the grace time. - Fix lookup/reclaim race condition. The above improvement undercovered a race condition between lookup and reclaim. If we reclaimed a vnode associated with a userland cookie while a lookup returning that same cookiewas inprogress, then the kernel ends up with a vnode associated with a cookie that has been reclaimed in userland. Next operation on the cookie will crash (or at least confuse) the filesystem. We fix this by introducing a lookup count in kernel and userland. On reclaim, the kernel sends the count, which enable userland to detect situation where it initiated a lookup that is not completed in kernel. In such a situation, the reclaim must be ignored, as the node is about to be looked up again. Fix hang unmount bug introduced by last commit. We introduced a slow queue for delayed reclaims, while the existing queue for unmount, flush and exist has been renamed fast queue. Both queues had timestamp for when an operation should be done, but it was useless for the fast queue, which is always used to run an operation ASAP. And the timestamp test had an error that turned ASAP into "at next tick", but nobody what there to wake the thread at next tick, hence the hang. The fix is to remove the useless and buggy timestamp test for fast queue. Rename slow sopreq queue into node sopreq queue, to refet the fact that is only intended for postponed node reclaims. When purging the node sopreq queue, do not call puffs_msg_sendresp(), as it makes no sense. Fix race condition between (create\|mknod\|mkdir\|symlino) and reclaim, just like we did it between lookup and reclaim. Missing bit in previous commit (prevent race between create\|mknod\|mkdir\|symlink and reclaim) Bump date for previous. New sentence, new line; remove trailing whitespace; fix typos; punctuation nits. Add PUFFS_KFLAG_CACHE_DOTDOT so that vnodes hold a reference on their parent, keeping them active, and allowing to lookup .. without sending a request to the filesystem. Enable the featuure for perfused, as this is how FUSE works. Missing bit in previous commit (PUFFS_KFLAG_CACHE_DOTDOT option to avoid looking up ..)
1.100.8.1	23-Apr-2012	riz	Pull up following revision(s) (requested by manu in ticket #195): lib/libskey/skeysubr.c: revision 1.27 lib/libkvm/kvm_getloadavg.c: revision 1.11 lib/libwrap/update.c: revision 1.9 lib/liby/yyerror.c: revision 1.9 lib/libpuffs/puffs_ops.3: revision 1.30 lib/libwrap/misc.c: revision 1.10 lib/libwrap/hosts_access.c: revision 1.20 lib/libpuffs/pnode.c: revision 1.11 lib/libperfuse/subr.c: revision 1.17 lib/libpuffs/pnode.c: revision 1.12 lib/libperfuse/subr.c: revision 1.18 lib/libwrap/options.c: revision 1.15 lib/libwrap/fix_options.c: revision 1.11 lib/libperfuse/ops.c: revision 1.52 lib/libperfuse/ops.c: revision 1.53 lib/libperfuse/ops.c: revision 1.54 lib/libwrap/hosts_ctl.c: revision 1.5 lib/libintl/gettext.c: revision 1.27 lib/libwrap/shell_cmd.c: revision 1.6 lib/libpuffs/dispatcher.c: revision 1.39 lib/libperfuse/perfuse_priv.h: revision 1.27 lib/libwrap/socket.c: revision 1.19 lib/libpuffs/puffs.3: revision 1.50 lib/libperfuse/perfuse_priv.h: revision 1.28 lib/libpuffs/puffs_priv.h: revision 1.45 lib/libpuffs/puffs.3: revision 1.51 lib/libperfuse/perfuse_priv.h: revision 1.29 lib/libwrap/percent_x.c: revision 1.5 lib/libpuffs/puffs.3: revision 1.52 lib/libperfuse/debug.c: revision 1.11 sys/fs/puffs/puffs_vnops.c: revision 1.165 lib/libwrap/tcpd.h: revision 1.13 sys/fs/puffs/puffs_vnops.c: revision 1.166 lib/libwrap/eval.c: revision 1.7 sys/fs/puffs/puffs_msgif.h: revision 1.78 sys/fs/puffs/puffs_vfsops.c: revision 1.101 lib/libwrap/rfc931.c: revision 1.9 lib/libwrap/clean_exit.c: revision 1.5 lib/libpuffs/puffs.h: revision 1.120 lib/libc/stdlib/jemalloc.c: revision 1.27 lib/librmt/rmtlib.c: revision 1.26 lib/libpuffs/puffs.h: revision 1.121 sys/fs/puffs/puffs_sys.h: revision 1.79 lib/librumpclient/rumpclient.c: revision 1.48 lib/libwrap/refuse.c: revision 1.5 lib/libperfuse/perfuse.c: revision 1.26 lib/libperfuse/perfuse.c: revision 1.27 tests/fs/puffs/t_fuzz.c: revision 1.5 lib/libperfuse/perfuse.c: revision 1.28 lib/libpuffs/dispatcher.c: revision 1.40 sys/fs/puffs/puffs_node.c: revision 1.24 lib/libwrap/diag.c: revision 1.9 lib/libintl/textdomain.c: revision 1.13 Use C89 function definition Add name and atttribute cache with filesytem provided TTL. lookup, create, mknod, mkdir, symlink, getattr and setattr messages have been extended so that attributes and their TTL can be provided by the filesytem. lookup, create, mknod, mkdir, and symlink messages are also extended so that the filesystem can provide name TTL. Add PUFFS_KFLAG_CACHE_FS_TTL flag to puffs_init(3) to use name and attribute cache with filesystem provided TTL. lookup, create, mknod, mkdir, symlink, getattr and setattr messages have been extended so that attributes and their TTL can be provided by the filesytem. lookup, create, mknod, mkdir, and symlink messages are also extended so that the filesystem can provide name TTL. The filesystem updates attributes and TTL using puffs_pn_getvap(3), puffs_pn_getvattl(3), and puffs_pn_getcnttl(3) Use new PUFFS_KFLAG_CACHE_FS_TTL option to puffs_init(3) so that FUSE TTL on name and attributes are used. This save many PUFFS operations and improves performances. PUFFS_KFLAG_CACHE_FS_TTL is #ifdef'ed in many places for now so that libperfuse can still be used on netbsd-5. Split file system. Comma fixes. Remove dangling "and". Bump date for previous. - Makesure update_va does not change vnode size when it should not. For instance when doing a fault-issued VOP_GETPAGES within VOP_WRITE, changing size leads to panic: genfs_getpages: past eof. -Handle ticks wrap around for vnode name andattribute timeout - When using PUFFS_KFLAG_CACHE_FS_TTL, do not use puffs_node to carry attribute and TTL fora newly created node. Instead extend puffs_newinfo and add puffs_newinfo_setva() and puffs_newinfo_setttl() - Remove node_mk_common_final in libperfuse. It used to set uid/gid for a newly created vnode but has been made redundant along time ago since uid and gid are properly set in FUSE header. - In libperfuse, check for corner case where opc = 0 on INACTIVE and RECLAIM (how is it possible? Check for it to avoid a crash anyway) - In libperfuse, make sure we unlimit RLIMIT_AS and RLIMIT_DATA so that we do notrun out of memory because the kernel is lazy at reclaiming vnodes. - In libperfuse, cleanup style of perfuse_destroy_pn() Do not set PUFFS_KFLAG_CACHE_FS_TTL for PUFFS tests
1.100.8.2.6.1	21-Apr-2014	bouyer	Pull up following revision(s) (requested by maxv in ticket #1050): sys/ufs/chfs/chfs_vfsops.c: revision 1.11 sys/fs/unionfs/unionfs_vfsops.c: revision 1.13 sys/fs/nilfs/nilfs_vfsops.c: revision 1.16 sys/ufs/mfs/mfs_vfsops.c: revision 1.107 sys/fs/sysvbfs/sysvbfs_vfsops.c: revision 1.43 sys/ufs/ffs/ffs_vfsops.c: revision 1.297 sys/kern/vfs_syscalls.c: revision 1.478 sys/kern/vfs_syscalls.c: revision 1.479 sys/fs/puffs/puffs_vfsops.c: revision 1.110 sys/fs/cd9660/cd9660_vfsops.c: revision 1.84 sys/nfs/nfs_vfsops.c: revision 1.227 sys/fs/v7fs/v7fs_vfsops.c: revision 1.10 sys/ufs/ext2fs/ext2fs_vfsops.c: revision 1.180 sys/miscfs/umapfs/umap_vfsops.c: revision 1.92 sys/fs/filecorefs/filecore_vfsops.c: revision 1.76 sys/miscfs/nullfs/null_vfsops.c: revision 1.88 sys/fs/ptyfs/ptyfs_vfsops.c: revision 1.50 sys/coda/coda_vfsops.c: revision 1.81 sys/ufs/lfs/lfs_vfsops.c: revision 1.321 sys/fs/tmpfs/tmpfs_vfsops.c: revision 1.59 sys/fs/hfs/hfs_vfsops.c: revision 1.31 sys/miscfs/overlay/overlay_vfsops.c: revision 1.61 sys/fs/union/union_vfsops.c: revision 1.72 sys/fs/ntfs/ntfs_vfsops.c: revision 1.94 sys/kern/vfs_syscalls.c: revision 1.480 sys/fs/efs/efs_vfsops.c: revision 1.25 sys/kern/vfs_syscalls.c: revision 1.482 sys/fs/msdosfs/msdosfs_vfsops.c: revision 1.107 external/cddl/osnet/dist/uts/common/fs/zfs/zfs_vfsops.c: revision 1.12 sys/miscfs/procfs/procfs_vfsops.c: revision 1.91 sys/fs/smbfs/smbfs_vfsops.c: revision 1.100 sys/fs/adosfs/advfsops.c: revision 1.70 sys/fs/udf/udf_vfsops.c: revision 1.67 Limit check for 'data_len'. Otherwise a (un)privileged user can easily panic the system by passing a huge size. ok christos@ An (un)privileged user can easily make the kernel dereference a NULL pointer. The kernel allows 'data' to be NULL; it's the fs's responsibility to ensure that it isn't NULL (if the fs actually needs data). ok christos@ Some fs's - like kernfs - set their vfs_min_mount_data to zero. Add a check to prevent an (un)privileged user from requesting a zero-sized allocation (and thus a panic). This thing is totally buggy: 'data_len' is modified by the fs, so calling kmem_free with it while its value has changed since the kmem_alloc is far from being a good idea. If the kernel figures out that something mismatches, it will panic (typically with kernfs).
1.100.8.2.4.1	21-Apr-2014	bouyer	Pull up following revision(s) (requested by maxv in ticket #1050): sys/ufs/chfs/chfs_vfsops.c: revision 1.11 sys/fs/unionfs/unionfs_vfsops.c: revision 1.13 sys/fs/nilfs/nilfs_vfsops.c: revision 1.16 sys/ufs/mfs/mfs_vfsops.c: revision 1.107 sys/fs/sysvbfs/sysvbfs_vfsops.c: revision 1.43 sys/ufs/ffs/ffs_vfsops.c: revision 1.297 sys/kern/vfs_syscalls.c: revision 1.478 sys/kern/vfs_syscalls.c: revision 1.479 sys/fs/puffs/puffs_vfsops.c: revision 1.110 sys/fs/cd9660/cd9660_vfsops.c: revision 1.84 sys/nfs/nfs_vfsops.c: revision 1.227 sys/fs/v7fs/v7fs_vfsops.c: revision 1.10 sys/ufs/ext2fs/ext2fs_vfsops.c: revision 1.180 sys/miscfs/umapfs/umap_vfsops.c: revision 1.92 sys/fs/filecorefs/filecore_vfsops.c: revision 1.76 sys/miscfs/nullfs/null_vfsops.c: revision 1.88 sys/fs/ptyfs/ptyfs_vfsops.c: revision 1.50 sys/coda/coda_vfsops.c: revision 1.81 sys/ufs/lfs/lfs_vfsops.c: revision 1.321 sys/fs/tmpfs/tmpfs_vfsops.c: revision 1.59 sys/fs/hfs/hfs_vfsops.c: revision 1.31 sys/miscfs/overlay/overlay_vfsops.c: revision 1.61 sys/fs/union/union_vfsops.c: revision 1.72 sys/fs/ntfs/ntfs_vfsops.c: revision 1.94 sys/kern/vfs_syscalls.c: revision 1.480 sys/fs/efs/efs_vfsops.c: revision 1.25 sys/kern/vfs_syscalls.c: revision 1.482 sys/fs/msdosfs/msdosfs_vfsops.c: revision 1.107 external/cddl/osnet/dist/uts/common/fs/zfs/zfs_vfsops.c: revision 1.12 sys/miscfs/procfs/procfs_vfsops.c: revision 1.91 sys/fs/smbfs/smbfs_vfsops.c: revision 1.100 sys/fs/adosfs/advfsops.c: revision 1.70 sys/fs/udf/udf_vfsops.c: revision 1.67 Limit check for 'data_len'. Otherwise a (un)privileged user can easily panic the system by passing a huge size. ok christos@ An (un)privileged user can easily make the kernel dereference a NULL pointer. The kernel allows 'data' to be NULL; it's the fs's responsibility to ensure that it isn't NULL (if the fs actually needs data). ok christos@ Some fs's - like kernfs - set their vfs_min_mount_data to zero. Add a check to prevent an (un)privileged user from requesting a zero-sized allocation (and thus a panic). This thing is totally buggy: 'data_len' is modified by the fs, so calling kmem_free with it while its value has changed since the kmem_alloc is far from being a good idea. If the kernel figures out that something mismatches, it will panic (typically with kernfs).
1.100.6.1	29-Apr-2012	mrg	sync to latest -current.
1.100.2.4	22-May-2014	yamt	sync with head. for a reference, the tree before this commit was tagged as yamt-pagecache-tag8. this commit was splitted into small chunks to avoid a limitation of cvs. ("Protocol error: too many arguments")
1.100.2.3	23-Jan-2013	yamt	sync with head
1.100.2.2	30-Oct-2012	yamt	sync with head
1.100.2.1	17-Apr-2012	yamt	sync with head
1.106.2.3	03-Dec-2017	jdolecek	update from HEAD
1.106.2.2	20-Aug-2014	tls	Rebase to HEAD as of a few days ago.
1.106.2.1	25-Feb-2013	tls	resync with head
1.107.2.1	18-May-2014	rmind	sync with head
1.109.2.1	10-Aug-2014	tls	Rebase.
1.113.2.4	15-Mar-2015	snj	Pull up following revision(s) (requested by tron in ticket #587): sys/fs/puffs/puffs_vfsops.c: revision 1.117 Remove debug printf
1.113.2.3	27-Feb-2015	martin	Pull up following revision(s) (requested by manu in ticket #555): lib/libpuffs/puffs.3: revision 1.60 sys/fs/puffs/puffs_msgif.h: revision 1.84 lib/libperfuse/ops.c: revision 1.83 sys/fs/puffs/puffs_sys.h: revision 1.89 sys/fs/puffs/puffs_vfsops.c: revision 1.116 lib/libperfuse/perfuse.c: revision 1.36 sys/fs/puffs/puffs_vnops.c: revision 1.200-1.202 Add PUFFS_KFLAG_NOFLUSH_META to prevent sending metadata flush to FUSE FUSE filesystems do not expect to get metadata updates for [amc]time and size, they updates the value on their own after operations. The PUFFS PUFFS_KFLAG_NOFLUSH_META option prevents regular metadata cache flushes to the filesystem , and libperfuse uses it to match Linux FUSE behavior. While there, fix a bug in SETATTR: do not update kernel metadata cache from SETATTR reply when the request is asynchronous, as we do not have the reply yet. Update file size after write without metadata flush If we do not use metadata flush, we must make sure the size is updated in the filesystem after a write, otherwise the next GETATTR will get us a stale value and the file will be truncated.
1.113.2.2	17-Jan-2015	martin	Pull up following revision(s) (requested by maxv in ticket #427): sys/compat/svr4/svr4_schedctl.c: revision 1.8 sys/netinet/tcp_timer.c: revision 1.88 sys/miscfs/genfs/layer_vfsops.c: revision 1.45 sys/compat/svr4/svr4_ioctl.c: revision 1.37 sys/ufs/chfs/chfs_vfsops.c: revision 1.14 sys/miscfs/fdesc/fdesc_vfsops.c: revision 1.91 sys/compat/linux/arch/i386/linux_ptrace.c: revision 1.30 sys/compat/common/kern_time_50.c: revision 1.28 sys/netinet6/ip6_forward.c: revision 1.74 sys/miscfs/umapfs/umap_vnops.c: revision 1.57 sys/compat/svr4/svr4_fcntl.c: revision 1.74 distrib/sets/lists/comp/mi: revision 1.1931 sys/netinet6/udp6_output.c: revision 1.46 sys/fs/puffs/puffs_compat.c: revision 1.3 sys/fs/udf/udf_rename.c: revision 1.11 sys/compat/svr4/svr4_filio.c: revision 1.24 sys/fs/udf/udf_rename.c: revision 1.12 sys/netinet/tcp_usrreq.c: revision 1.202 sys/miscfs/umapfs/umap_subr.c: revision 1.29 sys/compat/linux/common/linux_fadvise64.c: revision 1.3 sys/netinet/if_atm.c: revision 1.34 sys/miscfs/procfs/procfs_subr.c: revision 1.106 sys/miscfs/genfs/layer_subr.c: revision 1.37 sys/netinet/tcp_sack.c: revision 1.30 sys/compat/freebsd/freebsd_misc.c: revision 1.33 sys/compat/freebsd/freebsd_file.c: revision 1.33 sys/ufs/chfs/chfs_vnode.c: revision 1.12 sys/compat/svr4/svr4_ttold.c: revision 1.34 sys/compat/linux/common/linux_file.c: revision 1.114 sys/compat/linux/arch/mips/linux_machdep.c: revision 1.43 sys/compat/linux/common/linux_signal.c: revision 1.76 sys/compat/common/compat_util.c: revision 1.46 sys/compat/linux/arch/arm/linux_ptrace.c: revision 1.18 sys/compat/svr4/svr4_sockio.c: revision 1.36 sys/compat/linux/arch/arm/linux_machdep.c: revision 1.32 sys/compat/svr4/svr4_signal.c: revision 1.66 sys/kern/kern_exec.c: revision 1.410 sys/fs/puffs/puffs_vfsops.c: revision 1.115 sys/compat/svr4/svr4_exec_elf64.c: revision 1.15 sys/compat/linux/arch/i386/linux_machdep.c: revision 1.159 sys/compat/linux/arch/alpha/linux_machdep.c: revision 1.50 sys/compat/linux32/common/linux32_misc.c: revision 1.24 sys/netinet/in_pcb.c: revision 1.153 sys/sys/malloc.h: revision 1.116 sys/compat/common/if_43.c: revision 1.9 share/man/man9/Makefile: revision 1.380 sys/netinet/tcp_vtw.c: revision 1.12 sys/miscfs/umapfs/umap_vfsops.c: revision 1.95 sys/ufs/ext2fs/ext2fs_vfsops.c: revision 1.186 sys/compat/common/uipc_syscalls_43.c: revision 1.46 sys/ufs/ext2fs/ext2fs_vnops.c: revision 1.115 sys/fs/puffs/puffs_msgif.c: revision 1.97 sys/compat/svr4/svr4_ipc.c: revision 1.27 sys/compat/linux/common/linux_exec.c: revision 1.117 sys/ufs/ext2fs/ext2fs_readwrite.c: revision 1.66 sys/netinet/tcp_output.c: revision 1.179 sys/compat/svr4/svr4_termios.c: revision 1.28 sys/fs/udf/udf_strat_bootstrap.c: revision 1.4 sys/fs/puffs/puffs_subr.c: revision 1.67 sys/fs/puffs/puffs_node.c: revision 1.36 sys/miscfs/overlay/overlay_vnops.c: revision 1.21 sys/fs/cd9660/cd9660_node.c: revision 1.34 sys/netinet/raw_ip.c: revision 1.146 sys/sys/mallocvar.h: revision 1.13 sys/miscfs/overlay/overlay_vfsops.c: revision 1.63 share/man/man9/malloc.9: revision 1.50 sys/netinet6/dest6.c: revision 1.18 sys/compat/linux/common/linux_uselib.c: revision 1.33 sys/compat/linux/common/linux_socket.c: revision 1.120 share/man/man9/malloc.9: revision 1.51 sys/netinet/tcp_subr.c: revision 1.257 sys/compat/linux/common/linux_socketcall.c: revision 1.45 sys/compat/linux/common/linux_fadvise64_64.c: revision 1.3 sys/compat/freebsd/freebsd_ipc.c: revision 1.17 sys/compat/linux/common/linux_misc_notalpha.c: revision 1.109 sys/compat/linux/arch/alpha/linux_pipe.c: revision 1.17 sys/netinet6/in6_pcb.c: revision 1.132 sys/netinet6/in6_ifattach.c: revision 1.94 sys/compat/svr4/svr4_exec_elf32.c: revision 1.15 sys/miscfs/nullfs/null_vfsops.c: revision 1.90 sys/fs/cd9660/cd9660_util.c: revision 1.12 sys/compat/linux/arch/powerpc/linux_machdep.c: revision 1.48 sys/compat/freebsd/freebsd_exec_elf32.c: revision 1.20 sys/miscfs/procfs/procfs_vfsops.c: revision 1.94 sys/compat/linux/arch/powerpc/linux_ptrace.c: revision 1.28 sys/compat/linux/common/linux_sched.c: revision 1.67 sys/compat/linux/common/linux_exec_aout.c: revision 1.67 sys/compat/linux/common/linux_pipe.c: revision 1.67 sys/compat/linux/common/linux_llseek.c: revision 1.34 sys/compat/linux/arch/mips/linux_ptrace.c: revision 1.10 Do not uselessly include <sys/malloc.h>. Cleanup: - remove struct kmembuckets (dead) - correctly deadify MALLOC_XX - remove MALLOC_DEFINE_LIMIT and MALLOC_JUSTDEFINE_LIMIT (dead) - remove malloc_roundup(), malloc_type_setlimit(), MALLOC_DEFINE_LIMIT() and MALLOC_JUSTDEFINE_LIMIT() from man 9 malloc New sentence, new line. Bump date for previous. Obsolete malloc_roundup(9), malloc_type_setlimit(9) and MALLOC_DEFINE_LIMIT(9) man pages.
1.113.2.1	29-Aug-2014	martin	Pull up following revision(s) (requested by hannken in ticket #67): sys/fs/puffs/puffs_sys.h: revision 1.86 sys/fs/puffs/puffs_vfsops.c: revision 1.114 sys/fs/puffs/puffs_msgif.c: revision 1.95 sys/fs/puffs/puffs_node.c: revision 1.32 sys/fs/puffs/puffs_vnops.c: revision 1.184 Change puffs from hashlist to vcache. - field "pa_nhashbuckets" of struct "puffs_kargs" becomes a no-op. and should be removed on the next protocol version bump.
1.115.2.3	28-Aug-2017	skrll	Sync with HEAD
1.115.2.2	27-Dec-2015	skrll	Sync with HEAD (as of 26th Dec)
1.115.2.1	06-Apr-2015	skrll	Sync with HEAD
1.118.4.1	21-Apr-2017	bouyer	Sync with HEAD
1.118.2.2	26-Apr-2017	pgoyette	Sync with HEAD
1.118.2.1	20-Mar-2017	pgoyette	Sync with HEAD
1.120.12.1	25-Jun-2018	pgoyette	Sync with HEAD
1.121.2.2	13-Apr-2020	martin	Mostly merge changes from HEAD upto 20200411
1.121.2.1	08-Apr-2020	martin	Merge changes from current as of 20200406
1.123.2.2	29-Feb-2020	ad	Sync with head.
1.123.2.1	17-Jan-2020	ad	Sync with head.
1.125.8.1	03-Apr-2021	thorpej	Sync with HEAD.
1.125.6.1	03-Apr-2021	thorpej	Sync with HEAD.
1.226	09-Feb-2024	andvar	fix spelling mistakes, mainly in comments and log messages.
1.225	23-Feb-2022	andvar	fix various typos in comments, mainly immediatly/immediately/, as well shared and recently fixed typos in OpenBSD code by Jonathan Grey.
1.224	05-Dec-2021	msaitoh	s/invlid/invalid/ in comment.
1.223	20-Oct-2021	thorpej	Overhaul of the EVFILT_VNODE kevent(2) filter: - Centralize vnode kevent handling in the VOP_() wrappers, rather than forcing each individual file system to deal with it (except VOP_RENAME(), because VOP_RENAME() is a mess and we currently have 2 different ways of handling it; at least it's reasonably well-centralized in the "new" way). - Add support for NOTE_OPEN, NOTE_CLOSE, NOTE_CLOSE_WRITE, and NOTE_READ, compatible with the same events in FreeBSD. - Track which kevent notifications clients are interested in receiving to avoid doing work for events no one cares about (avoiding, e.g. taking locks and traversing the klist to send a NOTE_WRITE when someone is merely watching for a file to be deleted, for example). In support of the above: - Add support in vnode_if.sh for specifying PRE- and POST-op handlers, to be invoked before and after vop_pre() and vop_post(), respectively. Basic idea from FreeBSD, but implemented differently. - Add support in vnode_if.sh for specifying CONTEXT fields in the vop__args structures. These context fields are used to convey information between the file system VOP function and the VOP wrapper, but do not occupy an argument slot in the VOP_*() call itself. These context fields are initialized and subsequently interpreted by PRE- and POST-op handlers. - Version VOP_REMOVE(), uses the a context field for the file system to report back the resulting link count of the target vnode. Return this in tmpfs, udf, nfs, chfs, ext2fs, lfs, and ufs. NetBSD 9.99.92.
1.222	24-Jul-2021	andvar	Fix all remaining typos, mainly in comments but also in few definitions and log messages, reported by me in PR kern/54889. Also fixed some additional typos in comments, found on review of same files or typos.
1.221	19-Jul-2021	dholland	Abolish all the silly indirection macros for initializing vnode ops tables. These are things of the form #define foofs_op genfs_op, or #define foofs_op genfs_eopnotsupp, or similar. They serve no purpose besides obfuscation, and have gotten cutpasted all over everywhere. Part 2; cvs randomly didn't commit these changes before, and then hid them from me until I touched the files to force it to rethink. Dunno what happened. There's probably more of these, going to have to scan the tree the hard way.
1.220	18-Jul-2021	dholland	Use macros for the canned parts of device and fifo vnode op tables. Add GENFS_SPECOP_ENTRIES and GENFS_FIFOOP_ENTRIES macros that contain the portion of the vnode ops table declaration that is (conservatively) the same in every fs. Use these in every fs that supports devices and/or fifos with separate ops tables. Note that ptyfs works differently (it has one type of vnode with open-coded dispatch to the specfs code, which I haven't changed in this commit) and rump/librump/rumpvfs/rumpfs.c has an indirect dynamic dispatch that already does more or less the same thing, which I also haven't changed. Also note that this anticipates a few bits in the next changeset here and there, and adds missing but unreachable calls in some cases (e.g. most fses weren't defining whiteout on devices and fifos, but it isn't reachable there), and it changes parsepath on devices and fifos to genfs_badop from genfs_parsepath (but it's not reachable there either). It appears that devices in kernfs were missing kqfilter, so it's possible that if you try to use kqueue on /kern/rootdev that it'll explode. And finally note that the ops declaration tables aren't order-dependent. (Other than vop_default_desc has to come first.) Otherwise this wouldn't work.
1.219	29-Jun-2021	dholland	Now remove cn_consume from struct componentname. This change requires a kernel bump. Note though that I'm not going to version the VOP_LOOKUP args structure (or any other args structure) as code that doesn't touch cn_consume doesn't need attention and code that does will fail on it without further intervention.
1.218	29-Jun-2021	dholland	- Add a new vnode op: VOP_PARSEPATH. - Move namei_getcomponent to genfs_vnops.c and call it genfs_parsepath. - Add a parsepath entry to every vnode ops table. VOP_PARSEPATH takes a directory vnode to be searched and a complete following path and chooses how much of that path to consume. To begin with, all parsepath calls are genfs_parsepath, which locates the first '/' as always. Note that the call doesn't take the whole struct componentname, only the string. The other bits of struct componentname should not be needed and there's no reason to cause potential complications by exposing them.
1.217	16-May-2020	christos	branches: 1.217.6; Add ACL support for FFS. From FreeBSD.
1.216	15-May-2020	maxv	hardclock_ticks -> getticks()
1.215	23-Apr-2020	ad	PR kern/54759 (vm.ubc_direct deadlock when read()/write() into mapping of itself) - Add new flag UBC_ISMAPPED which tells ubc_uiomove() the object is mmap()ed somewhere. Use it to decide whether to do direct-mapped copy, rather than poking around directly in the vnode in ubc_uiomove(), which is ugly and doesn't work for tmpfs. It would be nicer to contain all this in UVM but the filesystem provides the needed locking here (VV_MAPPED) and to reinvent that would suck more. - Rename UBC_UNMAP_FLAG() to UBC_VNODE_FLAGS(). Pass in UBC_ISMAPPED where appropriate.
1.214	23-Feb-2020	ad	branches: 1.214.4; UVM locking changes, proposed on tech-kern: - Change the lock on uvm_object, vm_amap and vm_anon to be a RW lock. - Break v_interlock and vmobjlock apart. v_interlock remains a mutex. - Do partial PV list locking in the x86 pmap. Others to follow later.
1.213	06-Nov-2018	manu	branches: 1.213.6; Fix use after RECLAIM in PUFFS filesystems From hannken@ When puffs_cookie2vnode() misses an entry and vrele() it operations puffs_vnop_reclaim() and puffs_vnop_fsync() get called with a VNON vnode. Do not notify the server in this case as the cookie is stale.
1.212	05-Nov-2018	manu	Add missing mutex pn->pn_sizemtx lock in puffs_vnop_open() puffs_vnop_open() calls flushvncache(), which calls dosetattr() if pn->pn_stat has PNODE_METACACHE_MASK. In that case, the lock on pn->pn_sizemtx is mandatory and asserted.
1.211	26-May-2017	riastradh	branches: 1.211.2; 1.211.8; 1.211.10; Make VOP_RECLAIM do the last unlock of the vnode. VOP_RECLAIM naturally has exclusive access to the vnode, so having it locked on entry is not strictly necessary -- but it means if there are any final operations that must be done on the vnode, such as ffs_update, requiring exclusive access to it, we can now kassert that the vnode is locked in those operations. We can't just have the caller release the last lock because some file systems don't use genfs_lock, and require the vnode to remain valid for VOP_UNLOCK to work, notably unionfs.
1.210	26-Apr-2017	riastradh	Change VOP_REMOVE and VOP_RMDIR to preserve lock/ref on dvp. No change to vp -- the plan is to replace the node by the componentname in the vop parameters, and let all directory vops do lookups internally. Proposed on tech-kern with no objections: https://mail-index.netbsd.org/tech-kern/2017/04/17/msg021825.html
1.209	11-Apr-2017	riastradh	Make VOP_INACTIVE preserve vnode lock on return. Discussed on tech-kern: https://mail-index.netbsd.org/tech-kern/2017/04/01/msg021751.html Ride 7.99.68, a bumpy bus of incremental vfs improvements!
1.208	08-Apr-2017	hannken	Update mtime when updating file size. PR kern/51762 (mtime not updated by open(O_TRUNC))
1.207	06-Apr-2017	christos	use ubc_zerorange
1.206	04-Apr-2017	christos	use MAX_PAGE_SIZE.
1.205	21-Jul-2016	christos	branches: 1.205.2; replace variable stack declaration with a large enough one and KASSERT.
1.204	07-Jul-2016	msaitoh	branches: 1.204.2; KNF. Remove extra spaces. No functional change.
1.203	20-Apr-2015	riastradh	Make VOP_LINK return directory still locked and referenced. Ride 7.99.10 bump.
1.202	25-Feb-2015	christos	make this compile again.
1.201	25-Feb-2015	manu	Update file size after write without metadata flush If we do not use metadata flush, we must make sure the size is updated in the filesystem after a write, otherwise the next GETATTR will get us a stale value and the file will be truncated.
1.200	15-Feb-2015	manu	Add PUFFS_KFLAG_NOFLUSH_META to prevent sending metadata flush to FUSE FUSE filesystems do not expect to get metadata updates for [amc]time and size, they updates the value on their own after operations. The PUFFS PUFFS_KFLAG_NOFLUSH_META option prevents regular metadata cache flushes to the filesystem , and libperfuse uses it to match Linux FUSE behavior. While there, fix a bug in SETATTR: do not update kernel metadata cache from SETATTR reply when the request is asynchronous, as we do not have the reply yet.
1.199	13-Jan-2015	manu	Make sure reads on empty files reach PUFFS filesystems Sending a read through the page cache will get the operation short-circuited. This is a problem with some filesystems that expect to receive the read operation in order to update atime. We fix that by bypassing the page cache when reading a file wich a size known to be zero.
1.198	04-Nov-2014	manu	branches: 1.198.2; PUFFS direct I/O cache fix There are a few situations where we must take care of the cache if direct I/O was enabled: - if we do direct I/O for write but not for read, then any write must invalidate the cache so that a reader gets the written data and not the not-updated cache. - if we used a vnode without direct I/O and it is enabled for writing, we must flush the cache before compeling the open operation, so that the cachec write are not lost. And at inactive time, we wipe direct I/O flags so that a new open without direct I/O does not inherit direct I/O.
1.197	04-Nov-2014	manu	Fix PUFFS node use-after-reclaim When puffs_cookie2vnode() misses an entry, vcache_get() creates a new node (puffs_vfsop_loadvnode being called to initialize the PUFFS part), then it discovers it is VNON, and tries to vrele() it. vrele() calls VOP_INACTIVE(), which led us in puffs_vnop_inactive() where we sent a request to the filesystem for a node that already had been reclaimed. The fix is to check for VNON nodes in puffs_vnop_inactive() and to return without doing anyting. This is suboptimal, but a better workaround would probably need to modify vcache API, with an impact on other filesystems. Let us keep it simple.
1.196	31-Oct-2014	manu	Add PUFFS support for fallocate and fdiscard operations
1.195	31-Oct-2014	manu	According to pooka@'s comment, a long time ago, VOP_STRATEGY could not fail without taking down the kernel. It seems this is not the case anymore, hence we can stop dropping errors in puffs_vnop_strategy() Approved by pooka@
1.194	07-Oct-2014	he	Do the previous correctly...
1.193	07-Oct-2014	he	As is evidenced by several of our 32-bit MIPS ports, it's wrong to print vsize_t with PRIx64 -- instead use our own PRIxVSIZE macro.
1.192	06-Oct-2014	he	Make this build again without debugging enabled; DPRINTF() can end up as empty, and in an if conditional, you then need braces if that's the only potential body.
1.191	06-Oct-2014	manu	Retore LP64 fix that was removed by mistake
1.190	06-Oct-2014	manu	Improve zero-fill of last page after shrink fix: 1) do it only if the file is open for writing, otherwise we send write requests to the FS on a file that has never been open. 2) do it inside existing if (vap->va_size != VNOVAL) block
1.189	05-Oct-2014	justin	Use PRIx64 for printing offsets
1.188	05-Oct-2014	manu	If we truncate the file, make sure we zero-fill the end of the last page, otherwise if the file is later truncated to a larger size (creating a hole), that area will not return zeroes as it should.
1.187	30-Sep-2014	hannken	Fix the puffs_sop_thread -> puffs_cookie2vnode path: - pass the cookie by reference - add missing mutex_exit() - update assertion for VNON typed vnodes
1.186	11-Sep-2014	manu	PUFFS fixes for size update ater write plus read/write sanity checks - Always update kernel metadata cache for size when writing This fixes situation where size update after appending to a file lagged - Make read/write nilpotent when called with null size, as FFS does - Return EFBIG instead of EINVAL for negative offsets, as FFS does
1.185	05-Sep-2014	manu	When changing a directory content, update the ctime/mtime in kernel cache, otherwise the updated ctime/mtime appears after the cached entry expire.
1.184	28-Aug-2014	hannken	Change puffs from hashlist to vcache. - field "pa_nhashbuckets" of struct "puffs_kargs" becomes a no-op. and should be removed on the next protocol version bump.
1.183	16-Aug-2014	manu	Add a oflags input field to open requests so that the filesystem can pass back information about the file. Implement PUFFS_OPEN_IO_DIRECT, which will force direct IO (bypassing page cache) for the file.
1.182	25-Jul-2014	dholland	branches: 1.182.2; Add VOP_FALLOCATE and VOP_FDISCARD to every vnode ops table I can find. The filesystem ones all call genfs_eopnotsupp - right now I am only implementing the plumbing and we can implement fallocate and/or fdiscard for files later. The device ones call spec_fallocate (which is also genfs_eopnotsupp) and spec_fdiscard, which dispatches to the device-level op. The fifo ones all call vn_fifo_bypass, which also ends up being EOPNOTSUPP.
1.181	24-Mar-2014	hannken	branches: 1.181.2; - Make VI_XLOCK, VI_CLEAN and VI_LOCKSHARE private to kern/vfs_*.c. - Make vwait() static. - Add vdead_check() to check a vnode for being or becoming dead. Discussed on tech-kern. Welcome to 6.99.38
1.180	07-Feb-2014	hannken	Change vnode operation lookup to return the resulting vnode *vpp unlocked. Change cache_lookup() to return an unlocked vnode. Discussed on tech-kern@ Welcome to 6.99.31
1.179	23-Jan-2014	hannken	Change vnode operations create, mknod, mkdir and symlink to return the resulting vnode *vpp unlocked. Discussed on tech-kern@ Welcome to 6.99.30
1.178	17-Jan-2014	hannken	Change vnode operations create, mknod, mkdir and symlink to keep the directory node dvp locked on return. Discussed on tech-kern@ Welcome to 6.99.29
1.177	17-Oct-2013	christos	- remove unused variables - add _NOERROR flavor macros for the case where errors are ignored.
1.176	05-Nov-2012	dholland	branches: 1.176.2; Excise struct componentname from the namecache. This uglifies the interface, because several operations need to be passed the namei flags and cache_lookup also needs for the time being to be passed cnp->cn_nameiop. Nonetheless, it's a net benefit. The glop should be able to go away eventually but requires structural cleanup elsewhere first. This change requires a kernel bump.
1.175	05-Nov-2012	dholland	Disentangle the namecache from the internals of namei. - Move the namecache's hash computation to inside the namecache code, instead of being spread out all over the place. Remove cn_hash from struct componentname and delete all uses of it. - It is no longer necessary (if it ever was) for cache_lookup and cache_lookup_raw to clear MAKEENTRY from cnp->cn_flags for the cases that cache_enter already checks for. - Rearrange the interface of cache_lookup (and cache_lookup_raw) to make it somewhat simpler, to exclude certain nonexistent error conditions, and (most importantly) to make it not require write access to cnp->cn_flags. This change requires a kernel bump.
1.174	10-Aug-2012	manu	branches: 1.174.2; Add PUFFS_KFLAG_CACHE_DOTDOT so that vnodes hold a reference on their parent, keeping them active, and allowing to lookup .. without sending a request to the filesystem. Enable the featuure for perfused, as this is how FUSE works.
1.173	10-Aug-2012	manu	Missing bit in previous commit (prevent race between create\|mknod\|mkdir\|symlink and reclaim)
1.172	10-Aug-2012	manu	Fix race condition between (create\|mknod\|mkdir\|symlino) and reclaim, just like we did it between lookup and reclaim.
1.171	27-Jul-2012	manu	Rename slow sopreq queue into node sopreq queue, to refet the fact that is only intended for postponed node reclaims. When purging the node sopreq queue, do not call puffs_msg_sendresp(), as it makes no sense.
1.170	23-Jul-2012	manu	Backout NCHNAMLEN check for cache_enter. That change collided with rmind's move of this exact check into cache_enter
1.169	23-Jul-2012	manu	Di not call cache_enter with path components bigger than NCHNAMLEN, as it panics the kernel.
1.168	22-Jul-2012	rmind	Move some the test for MAKEENTRY into the cache_enter(9). Make some variables in vfs_cache.c static, __read_mostly, etc. No objection on tech-kern@.
1.167	21-Jul-2012	manu	- Improve PUFFS_KFLAG_CACHE_FS_TTL by reclaiming older inactive nodes. The normal kernel behavior is to retain inactive nodes in the freelist until it runs out of vnodes. This has some merit for local filesystems, where the cost of an allocation is about the same as the cost of a lookup. But that situation is not true for distributed filesystems. On the other hand, keeping inactive nodes for a long time hold memory in the file server process, and when the kernel runs out of vnodes, it produce reclaim avalanches that increase lattency for other operations. We do not reclaim inactive vnodes immediatly either, as they may be looked up again shortly. Instead we introduce a grace time and we reclaim nodes that have been inactive beyond the grace time. - Fix lookup/reclaim race condition. The above improvement undercovered a race condition between lookup and reclaim. If we reclaimed a vnode associated with a userland cookie while a lookup returning that same cookiewas inprogress, then the kernel ends up with a vnode associated with a cookie that has been reclaimed in userland. Next operation on the cookie will crash (or at least confuse) the filesystem. We fix this by introducing a lookup count in kernel and userland. On reclaim, the kernel sends the count, which enable userland to detect situation where it initiated a lookup that is not completed in kernel. In such a situation, the reclaim must be ignored, as the node is about to be looked up again.
1.166	18-Apr-2012	manu	- Makesure update_va does not change vnode size when it should not. For instance when doing a fault-issued VOP_GETPAGES within VOP_WRITE, changing size leads to panic: genfs_getpages: past eof. -Handle ticks wrap around for vnode name andattribute timeout
1.165	08-Apr-2012	manu	Add name and atttribute cache with filesytem provided TTL. lookup, create, mknod, mkdir, symlink, getattr and setattr messages have been extended so that attributes and their TTL can be provided by the filesytem. lookup, create, mknod, mkdir, and symlink messages are also extended so that the filesystem can provide name TTL.
1.164	16-Mar-2012	jakllsch	Prevent access beyond end of PUFFS file on read, similar to as is done for NFS.
1.163	17-Jan-2012	martin	branches: 1.163.2; Add a few KASSERT() - I have a crash that likely will cause one of them to fire...
1.162	18-Nov-2011	christos	branches: 1.162.4; Obey MNT_RELATIME, the only addition is that mkdir in ufs sets IN_ACCESS too.
1.161	30-Oct-2011	hannken	branches: 1.161.2; Add a comment that pn_sizemtx should be useless as VOP_GETATTR now needs a shared lock at least.
1.160	19-Oct-2011	manu	Remove #ifdef DIAGNOSTIC guards around KASSERT, as the macro contains them
1.159	18-Oct-2011	manu	Make sure pagedaemon does not sleep for memory in puffs_vnop_sleep. Add KASSERT on any sleeping memory allocation to check it cannot happen again.
1.158	17-Oct-2011	manu	Roll back the change that forced kernel threads to not sleep in PUFFS. The change does not make consensus, since only pagedaemon should need it. Other threads will tolerate sleeping, and problems here are only symptoms that something is going wrong in memory management. The cause, not the symptoms, need to be fixed.
1.157	23-Sep-2011	manu	Fix the build that was broken by struct lwp *updateproc reference in RUMP-visible code. Instead of checking that updateproc (aka ioflush, aka syncer) will not sleep in PUFFS code, I check for any kernel thread: after all none of them are designed to hang awaiting for a remote filesystem operation to complete.
1.156	21-Sep-2011	manu	Make sure ioflush does not sleep in PUFFS code path, waiting for a mutex, a memory allocation, or a response from the filesystem. This avoids deadlocks in the following situations: 1) when memory is low: ioflush waits the fileystem, the fielsystem waits for memory 2) when the filesystem does not respond (e.g.: network outage ona distributed filesystem)
1.155	29-Aug-2011	manu	Add a mutex for operations that touch size (setattr, getattr, write, fsync). This is required to avoid data corruption bugs, where a getattr slices itself within a setattr operation, and sets the size to the stall value it got from the filesystem. That value is smaller than the one set by setattr, and the call to uvm_vnp_setsize() trigged a spurious truncate. The result is a chunk of zeroed data in the file. Such a situation can easily happen when the ioflush thread issue a VOP_FSYNC/puffs_vnop_sync/flushvncache/dosetattrn while andother process do a sys_stat/VOP_GETATTR/puffs_vnop_getattr. This mutex on size operation can be removed the day we decide VOP_GETATTR has to operated on a locked vnode, since the other operations that touch size already require that.
1.154	04-Jul-2011	manu	Add a flag to VOP_LISTEXTATTR(9) so that the vnode interface can tell the filesystem in which format extended attribute shall be listed. There are currently two formats: - NUL-terminated strings, used for listxattr(2), this is the default. - one byte length-pprefixed, non NUL-terminated strings, used for extattr_list_file(2), which is obtanined by setting the EXTATTR_LIST_PREFIXLEN flag to VOP_LISTEXTATTR(9) This approach avoid the need for converting the list back and forth, except in libperfuse, since FUSE uses NUL-terminated strings, and the kernel may have requested EXTATTR_LIST_PREFIXLEN.
1.153	12-Jun-2011	rmind	Welcome to 5.99.53! Merge rmind-uvmplock branch: - Reorganize locking in UVM and provide extra serialisation for pmap(9). New lock order: [vmpage-owner-lock] -> pmap-lock. - Simplify locking in some pmap(9) modules by removing P->V locking. - Use lock object on vmobjlock (and thus vnode_t::v_interlock) to share the locks amongst UVM objects where necessary (tmpfs, layerfs, unionfs). - Rewrite and optimise x86 TLB shootdown code, make it simpler and cleaner. Add TLBSTATS option for x86 to collect statistics about TLB shootdowns. - Unify /dev/mem et al in MI code and provide required locking (removes kernel-lock on some ports). Also, avoid cache-aliasing issues. Thanks to Andrew Doran and Joerg Sonnenberger, as their initial patches formed the core changes of this branch.
1.152	19-May-2011	rmind	branches: 1.152.2; Remove cache_purge(9) calls from reclamation routines in the file systems, as vclean(9) performs it for us since Lite2 merge.
1.151	03-May-2011	manu	Call advlock method if supplied
1.150	11-Jan-2011	kefren	branches: 1.150.2; add advlock to puffs. ok pooka@ should fix kern/43321
1.149	30-Nov-2010	dholland	Abolish the SAVENAME and HASBUF flags. There is now always a buffer, so the path in a struct componentname is now always valid during VOP calls.
1.148	30-Nov-2010	dholland	Abolish struct componentname's cn_pnbuf. Use the path buffer in the pathbuf object passed to namei as work space instead. (For now a pnbuf pointer appears in struct nameidata, to support certain unclean things that haven't been fixed yet, but it will be going away in the future.) This removes the need for the SAVENAME and HASBUF namei flags.
1.147	14-Jul-2010	pooka	RENAME lookup semantics say return EISDIR if dvp = *vpp for the last component .... obviously(!!)
1.146	24-Jun-2010	hannken	Clean up vnode lock operations pass 2: VOP_UNLOCK(vp, flags) -> VOP_UNLOCK(vp): Remove the unneeded flags argument. Welcome to 5.99.32. Discussed on tech-kern.
1.145	21-May-2010	pooka	Support extended attributes.
1.144	29-Mar-2010	pooka	Stop exposing fifofs internals and leave only fifo_vnodeop_p visible.
1.143	27-Mar-2010	pooka	\n, police!
1.142	14-Jan-2010	pooka	branches: 1.142.2; 1.142.4; Since VOP_GETATTR() does not require a locked vnode, resolve and reference the puffs_node before sending the request to the file server. This diminishes the window where the inode can be reclaimed and be invalidated before it is accessed (but does not completely eliminate the race, as that is a caller problem which we cannot fix here).
1.141	04-Dec-2009	pooka	Push all information cached in the vnode to the file server before issuing INACTIVE. PR kern/42194. Also, send setattr in fsync asynchronously if FSYNC_WAIT is not set.
1.140	19-Nov-2009	pooka	Send VOP_ABORTOP() in case attempting cross-dev rename, part of PR kern/42210. Also, fix a memory management error in said case.
1.139	19-Nov-2009	pooka	Send VOP_ABORTOP() as a FAF -- we don't care about the return value.
1.138	05-Nov-2009	pooka	Kill suspend support. It was never implemented correctly: * it depended on the biglock (in a very cruel way) * it was attached to userspace transactions rather than logical fs operations (If someone wants to revisit it some day, most of the stuff can be reused from cvs history)
1.137	05-Nov-2009	pooka	Reinstante PNODE_DYING. vmlocking had a brief hiatus when it was not a valid optimization, but that's long gone and once VOP_INACTIVE is called and the file server says that the vnode is going to be recycled, it really is going to be recycled extra references gained or not.
1.136	17-Oct-2009	pooka	Transmit VOP_ABORTOP() to the server.
1.135	30-Sep-2009	pooka	remove leading whitespace. no functional change.
1.134	30-Sep-2009	pooka	* fix a race i introduced almost two years ago in rev 1.116: operations creating a node cannot be considered outgoing operations, since after return from userspace they modify file system state by creating a new node. if we do not protect the file system by holding the directory lock, a lookup operation might race us into the kernel and create the node earlier. * remove pnode from hashlish before sending the reclaim faf off to userspace. also, hold pmp_lock while frobbing the list.
1.133	19-Sep-2009	pooka	Set SAVENAME for rmdir and remove. Addresses an easy part of PR kern/38188
1.132	12-Sep-2009	tsutsui	Fix typo: - pcinfo = kmem_zalloc(sizeof_puffs_cacheinfo) + runsize, + pcinfo = kmem_zalloc(sizeof(struct puffs_cacheinfo) + runsize, in #ifdef'ed out code, per paired kmem_free() in the same function. Closes PR kern/41840.
1.131	26-Nov-2008	pooka	Rototill all remaining file systems to use ubc_uiomove() instead of the ubc_alloc() - uiomove() - ubc_release() dance.
1.130	16-Nov-2008	pooka	more <sys/buf.h> police
1.129	10-Sep-2008	christos	branches: 1.129.2; 1.129.4; 1.129.8; replace 0xa0 with space from Andy Shevchenko
1.128	30-Jan-2008	ad	branches: 1.128.6; 1.128.10; 1.128.12; 1.128.16; Replace struct lock on vnodes with a simpler lock object built on krwlock_t. This is a step towards removing lockmgr and simplifying vnode locking. Discussed on tech-kern.
1.127	28-Jan-2008	pooka	For code clarity typedef void *puffs_cookie_t. No functional change.
1.126	25-Jan-2008	ad	Remove VOP_LEASE. Discussed on tech-kern.
1.125	02-Jan-2008	pooka	More type-punning workarounds. Curiously the kernel compilation flags cause gcc to not complain.
1.124	02-Jan-2008	ad	Merge vmlocking2 to head.
1.123	30-Dec-2007	pooka	namespace a bit: vfsops -> puffs_vfsop_x() and vops -> puffs_vnop_x()
1.122	08-Dec-2007	pooka	branches: 1.122.4; Now that "l" is gone both as an argument to operations and from componentname, remove all vestiges of puffs_cid.
1.121	27-Nov-2007	pooka	branches: 1.121.2; Remove "puffs_cid" from the puffs interface following l-removal from the kernel vfs interfaces. puffs_cc_getcaller(pcc) can be used now should the same information be desired.
1.120	26-Nov-2007	pooka	Remove the "struct lwp *" argument from all VFS and VOP interfaces. The general trend is to remove it from all kernel interfaces and this is a start. In case the calling lwp is desired, curlwp should be used. quick consensus on tech-kern
1.119	21-Nov-2007	pooka	use BUF_ISREAD/WRITE instead of homegrown variants
1.118	20-Nov-2007	pooka	Retire M_PUFFS, use kmem(9) instead.
1.117	17-Nov-2007	pooka	Make puffs_updatenode() take a puffs_node instead of a vnode. This way we don't need to worry if a vnode has been reclaimed from under us.
1.116	17-Nov-2007	pooka	Start playing around with vnode locks. For now, do the very easy thing and release locks before the userspace wait for operations which release the lock before exit from the method in any case. However, releasing the lock after inserting the request on the operation queue gives us proper ordering possibilities in userspace (at least if that bit were implemented, but I don't think there any file system in userspace that depends on kernel locking and probably there never should be one). inspired by a conversation with Nacho Navarro
1.115	17-Nov-2007	pooka	Implement a biodone callback for async writes similar to reads and use that when possible.
1.114	16-Nov-2007	pooka	Restructure the messaging interface a bit more: make all interfacing with the file server happen through puffs_msg_enqueue() and puffs_msg_wait() instead of having a billion different routines. Build the existing system upon these two. Most importantly though, decouple insertation into the op queue from the actual wait. This is useful for a number of reasons coming soon to a cvs repo near you.
1.113	26-Oct-2007	pooka	branches: 1.113.2; Read/write can reuse message memory if operating uncached. This will change evetually, but for now just appease a KASSERT by resetting the message header to 0 after each loop.
1.112	23-Oct-2007	pooka	The kernel (genfs, uvm) can't deal with strategy returning an error when vclean()ing. Pending an adventure to the genfs/vm labyrinth to fix this properly, compensate here by not allowing unstrategic (no pun) return values. They are always due to the userspace server crashing anyway, so it's no big deal if we lie about the final resting place of the pages.
1.111	21-Oct-2007	pooka	* release pathname buffer in link * some variable massage
1.110	19-Oct-2007	pooka	When doing a read operation, don't copy the whole kernel buffer to userspace, since it doesn't contain any information yet. I should still rework this more so this is just a quickie to get the read/write style interface more up to speed with the ioctl version.
1.109	19-Oct-2007	pooka	comment polish
1.108	18-Oct-2007	pooka	Fix wrong argument order which just happened to work by luck.
1.107	11-Oct-2007	pooka	branches: 1.107.2; Part 1/n of some pretty extensive changes to how the kernel module interacts with the userspace file server: * since the kernel-user communication is not purely request-response anymore (hasn't been since 2006), try to rename some "request" to "message". more similar mangling will take place in the future. * completely rework how messages are allocated. previously most of them were borrowed from the stack (originally all of them), but now always allocate dynamically. this makes the structure of the code much cleaner. also makes it possible to fix a locking order violation. it enables plenty of future enhancements. * start generalizing the transport interface to be independent of puffs * move transport interface to read/write instead of ioctl. the old one had legacy design problems, and besides, ioctl's suck. implement a very generic version for now; this will be worked on later hopefully some day reaching "highly optimized". * implement libpuffs support behind existing library request interfaces. this will change eventually (I hate those interfaces)
1.106	11-Oct-2007	pooka	Cache vnode member variables necessary for operations after the userspace call, namely our private mount structure, in the activation record. This avoids problems in situations where the userspace file server happens to die during our upcall and the vnode is forcibly reclaimed before we roll back to the current stack frame.
1.105	10-Oct-2007	ad	Merge from vmlocking: - Split vnode::v_flag into three fields, depending on field locking. - simple_lock -> kmutex in a few places. - Fix some simple locking problems.
1.104	04-Oct-2007	pooka	g/c the "sizeop" code previous used for ioctl/fcntl. It was already commented out and has bitrotted beyond all recognition, so it needs complete rethinking.
1.103	02-Oct-2007	pooka	If kernel resource allocation fails after the file server has committed something, issue an abort. The abort is done through the regular op channel, e.g. failed mkdir leads to regular rmdir, inactive and reclaim. No internal interface is planned currently for the one file system out of a million which would implement it to benefit from the one case in a billion where kernel resource allocation actually does fail and out of that one case in a trillion where internal vs. external would make a difference.
1.102	01-Oct-2007	pooka	* better error checking: validate error values received from userland to be vaild errno values * include string describing error in PUFFS_ERR * get rid of union in puffs_req, it's nothing but trouble * pass pmp to async i/o callbacks
1.101	27-Sep-2007	pooka	Differentiate between cookie2vnode returning an error and return to caller, address unknown: no such cookie, no such node. Make the callers use this info to either create a new vnode or bail.
1.100	27-Sep-2007	pooka	Add error notifications, which are used to deliver errors from the kernel to the file server for silly things the file server did, e.g. attempting to create a file with size VSIZENOTSET. The file server can handle these as it chooses, but the default action is for it to throw its hands in the air and sing "goodbye, cruel world, it's over, walk on by".
1.99	27-Sep-2007	pooka	Fix a race in how new cookies are checked. Previously the checking was done separate of inserting the cookie into the lookup structure and without any form of interlock. This could lead to the same cookie pointing to two different nodes. Remedy the race by creating a separate "checked and ready to be inserted" cookie list which serves as an interlock without having to hold a fs-global creation lock.
1.98	22-Aug-2007	pooka	branches: 1.98.2; 1.98.4; Mimic namei structure changes for puffs. bump both kernel & lib version.
1.97	13-Aug-2007	pooka	* don't call VOP_ACCESS in lookup, that's the file system's problem * be more careful with r/o fs to catch EEXIST in lookup CREATE * some comment polish
1.96	12-Aug-2007	pooka	enforce MNT_RDONLY
1.95	30-Jul-2007	pooka	branches: 1.95.4; 1.95.6; properly setup ubcflags
1.94	29-Jul-2007	ad	It's not a good idea for device drivers to modify b_flags, as they don't need to understand the locking around that field. Instead of setting B_ERROR, set b_error instead. b_error is 'owned' by whoever completes the I/O request.
1.93	27-Jul-2007	yamt	ubc_uiomove: add an "advice" argument rather than using UVM_ADV_RANDOM blindly.
1.92	27-Jul-2007	pooka	Change unused fflags parameter in VOP_MMAP to prot and pass in desired vm protection.
1.91	22-Jul-2007	pooka	use NULL, not 0, to pass a pointer
1.90	22-Jul-2007	pooka	Keep track of the maximum size we have supplied the file server (or it has supplied us). If we fault pages which are at offset >= server size, but less than the in-kernel vnode size, inform the file server of the latest developments in file size before issueing the fault. The avoids confusion with files which are not written start to finish. fixes kern/36429 by yamt
1.89	19-Jul-2007	pooka	don't request more than the maximum request size in readdir
1.88	09-Jul-2007	ad	branches: 1.88.2; s/pagedaemon_lwp/pagedaemon_proc/
1.87	09-Jul-2007	ad	Merge some of the less invasive changes from the vmlocking branch: - kthread, callout, devsw API changes - select()/poll() improvements - miscellaneous MT safety improvements
1.86	02-Jul-2007	pooka	support turning REQUIREDIR off and extra consume in lookup
1.85	02-Jul-2007	pooka	Get rid of the "int *refs" parameter to inactive: the same can be accomplished now with puffs_setbacks.
1.84	01-Jul-2007	pooka	loosen KASSERT: we can also fail due to ENOMEM
1.83	01-Jul-2007	pooka	Give the file server to ability to request the entire pathname buffer under lookup by using PUFFS_KFLAG_LOOKUP_FULLPNBUF instead just the current component.
1.82	01-Jul-2007	pooka	Instead of supplying a plain pid, supply an abstract struct puffs_cid *, which can currently be used to query the pid and lwpid.
1.81	01-Jul-2007	pooka	make puffs_cred an opaque type
1.80	30-Jun-2007	pooka	Fix logic flaw in KASSERT. Seems like my lkm wasn't compiled with DIAGNOSTIC ...
1.79	26-Jun-2007	pooka	Simplify code, mainly vop_strategy. No functional change
1.78	24-Jun-2007	pooka	Split the NOCACHE option in twain: NOCACHE_NAME & NOCACHE_PAGE.
1.77	21-Jun-2007	pooka	Refactor the pnode2vnode translation slightly so that VFS_ROOT can use it directly.
1.76	06-Jun-2007	pooka	Move puffs to a two clause license where it already isn't so. And as agc pointed out, even files with the third clause were already effectively two clause because of a slight bug in the language...
1.75	06-Jun-2007	pooka	In very verbose debug mode, print also return values for operations (well, at least for those that go through checkop()).
1.74	05-Jun-2007	yamt	improve post-ubc file overwrite performance in common cases. ie. when it's safe, actually overwrite blocks rather than doing read-modify-write. also fixes PR/33152 and PR/36303.
1.73	01-Jun-2007	yamt	\xa0 -> space.
1.72	19-May-2007	pooka	Actually, we do need separate "no references in file server" and "noref + inactive" flags if we wish to correctly support unix open file semantics and optimize away pre-reclaim cache flushes. So, add PNODE_DYING which stands for norefs + inactive.
1.71	18-May-2007	pooka	Introduce noref setbacks, which the file server can use to signal the kernel it has 0 references to the node in question. In other words, this can be used to avoid inactive(), or, if the file server does not implement inactive, prompt reclaim for removed nodes.
1.70	18-May-2007	pooka	selrecord() before calling userspace to avoid (very theoretical) race where selinfo contains uninitialized garbage
1.69	18-May-2007	pooka	Support VOP_POLL. This requires some acrobatics on the puffs_node, as we give a reference to userspace for the puffs_node for the duration of the poll call. So reference count puffs_node separately from the parent vnode. vref()/vrele() is not possible due to a possible surprise visit from VOP_INACTIVE.
1.68	15-May-2007	pooka	In case strategy memory allocation for B_ASYNC\|B_READ fails, make sure to release the buf.
1.67	08-May-2007	pooka	Adventures in file systems, part (u_quad_t)-1: we can't use the file system value for the size of device special files, as that comes from specfs instead of the "host" file system. Therefore, take care that getattr doesn't override the value of vp->v_size.
1.66	07-May-2007	pooka	Introduce puffs "setbacks", which can be used to set certain flags for nodes upon return from the userspace. Currently it can be used to indicate that the file server should be notified of "inactive" in case the file server has opted to not receive inactive every time the reference count for a vnode drops to zero. (inactive is a common event, almost never requires any action and must be executed sychronously, so it is wasteful). While doing this, cleanup the release-relock nonsense from the vntouser*() arguments. It was never enabled and the whole LOCKEDVP() concept was very broken to begin with.
1.65	06-May-2007	pooka	If setattr is called explicitly, use that as the sign to flush out all metadata info cached in the kernel while we're setattr'ing in any case. Solves problems such as truncate (via extend-by-write) + chmod resulting in EPERM because the file was already read-only when the actual truncate was flushed out of the kernel in fsync.
1.64	24-Apr-2007	pooka	If ubc style write fails, do not extend the file by zero-padding it. It might be that the file server is either crashing or just returning consistent errors. uiomove() would handle the error, but if the pages weren't faulted in, memset() to the unfaultable ubc window would cause a kernel page fault.
1.63	22-Apr-2007	pooka	Issue close to the file server asynchronously. We're not interested in the return value.
1.62	22-Apr-2007	pooka	define PUFFS_KFLAG_WTCACHE, which makes the page cache write-through
1.61	20-Apr-2007	pooka	* in readdir, don't copy extra memory back and forth to userspace * consistent usage of the variable argsize with the rest of the module
1.60	20-Apr-2007	pooka	Size of a readdir cookie is sizeof(*ap->a_cookies), not sizeof(ap->a_cookies). Fixes nfs readdir in the case that a directory had lots of entries with short names.
1.59	16-Apr-2007	pooka	Give the file server the ability to specify the file handle length instead of defining a static length file handle on the framework-level.
1.58	11-Apr-2007	pooka	* support VFS_FHTOVP and VFS_VPTOFH * support cookies in for VOP_READDIR nfs exporting puffs file systems works now
1.57	04-Apr-2007	pooka	Make it possible to interrupt waiters for fs operation completion again. This is useful until locking is further developed and basically any deadlocks can be solved by killing appropriate processes. Thanks especially to Tommi Kyntola and Antti Louko for sitting down with me and discussing resource ownership and locking strategies in implementing this.
1.56	30-Mar-2007	pooka	* abstract ASYNCBIOREAD and let callers freely issue a callback called from putop. even though there's only one user currently, makes code more readable * move "delta" to a standard parameter in vntouser and get rid of the specialcase vntouser_delta
1.55	29-Mar-2007	pooka	Convert spinlocks & sleep/wakeup to newlock2 locking stuff. Fix a bunch of bugs. * park structures are now always allocated from a pool instead of a mixed stack/malloc allocation * get rid of the whole adjbuf concept, always just alloc the maximal amount of memory to satisfy a request * little regression: don't allow interrupting wait from file system to userspace; this had problems already before, but now the problems really started to shine through. I'll try to make this work again some day. * fix bmap to return a sensible value in runp
1.54	20-Mar-2007	pooka	* rework the page cache interaction a bit: cache metadata in the kernel and flush it out all at once instead of continuous updating * add support for delivering notifications to the file server about when a page was written to (but disabled by default for now). the file server can use this to request flushing or invalidating the kernel page cache
1.53	14-Mar-2007	pooka	branches: 1.53.2; Support B_READ\|B_ASYNC in strategy by calling biodone() directly when the file server puts the result.
1.52	20-Feb-2007	pooka	branches: 1.52.4; 1.52.6; Properly fix rev 1.44: limit error values from the file server to positive values of errno and 0. Otherwise it can return internal values such as EJUSTRETURN and screw things up. thanks to Bill for reminding me to revisit this
1.51	15-Feb-2007	pooka	branches: 1.51.2; Sanity-check linklen returned from file server in READLINK.
1.50	10-Feb-2007	pooka	* in write, do sync pageflush for the ubc case every 64k, otherwise the user file server can't really keep up and just writing and writing may result in kernel memory exhaustion. this lossage is also partially due to the stupid way mtime + size info is handled currently, but that should change soon (knock knock ;) * score a few debug printfs
1.49	09-Feb-2007	pooka	honor B_ASYNC
1.48	09-Feb-2007	pooka	assign value for strategy output parameter b_resid instead of decreasing it
1.47	08-Feb-2007	pooka	If the file server doesn't support write, don't use genfs_null_putpages for putpages, as it assumes a vnode doesn't have any pages. For mounts using the page cache this is simply not true. Rather, prevent opening a regular file in write-mode. That way a vnode can never have dirty pages which would need to be flushed (i.e. written).
1.46	08-Feb-2007	pooka	chuq shone arcane wisdom on me: b_bcount comes in, b_resid goes out
1.45	08-Feb-2007	pooka	Don't block and wait for file server response in case strategy is run in pagedaemon context: it gives the file server way more control over the fate of the entire kernel than what we're comfortable with.
1.44	06-Feb-2007	pooka	Limit errors from puffs_lookup to 0, EJUSTRETURN and ENOENT, as that's what namei/lookup expects.
1.43	29-Jan-2007	hannken	Change fstrans enum types to upper case. No functional change. From Antti Kantee <pooka@netbsd.org>
1.42	26-Jan-2007	pooka	We don't handle fsync in checkop anymore, so direct the fifoop fsync also to a place less panicy, namely fifo_fsync (because currently the metadata information is update when the node is changed. This will probably change soon, though).
1.41	26-Jan-2007	pooka	Initial attempt at suspend/snapshot support for userspace file servers. This is still pretty much on the level "if it breaks ...". It should work for single-threaded servers which handle one operation from start to finish in one go. Also, it does not yet totally correctly synchronize metadata and data in some cases. So needless to say, it needs improvement, but it is possible that will have to wait for some lock revampage.
1.40	25-Jan-2007	pooka	if strategy fails, set bp->b_error and B_ERROR
1.39	25-Jan-2007	pooka	don't hold spinlocks (except vnode interlock) when doing vget()
1.38	21-Jan-2007	pooka	optimize a bit: don't flush pages for vnodes which have no references in the kernel or links in the backend
1.37	21-Jan-2007	pooka	remove diagnostic printf
1.36	19-Jan-2007	pooka	hannken noted that the latest gcc (?) complains about uninitialized variable use in puffs_strategy() for "dowritefaf" (incorrectly) and "error" (correctly, although the function is practically of type void)
1.35	19-Jan-2007	pooka	In case the fs server is in the kernel doing an operation on a completely different file system, we still might re-enter the same puffs fs in case we execute something on the other file system, which wants to get a new vnode and ends up recycling a puffs vnode for the purpose. In this case the fs server will sleep in the kernel until it itself handles the operation .... which of course is a slightly unlikely event. After analyzing the path from getcleanvnode() to the vnode cemetary, identify that fsync and putpages (strategy) are the ones in danger of striking a deadlock deal. Abuse the vnode flag VXLOCK to tell them "this vnode is irreversably going to meet its maker, don't care about user server return values" (failure is not acceptable down the vgonel() path) and issue the respective operations as Fire-And-Forget (FAF) operations. no wait -> no deadlock. This of course is a "fix" skating on thin ice. A better, more generic solution is already in sight, but will take more effort to implement.
1.34	16-Jan-2007	pooka	* don't wait for the answer of VOP_RECLAIM, just fire-and-forget * revoke puffs_revoke. we can deal with it just by calling genfs_revoke
1.33	15-Jan-2007	pooka	Store puffs_node's on lists hashed with the cookie value instead of just one flat list.
1.32	15-Jan-2007	pooka	* do not accept the directory cookie as the result of a lookup (otherwise we'd be locking against ourselves) * do not accept duplicate cookies when creating new nodes
1.31	11-Jan-2007	pooka	Since fsync is really putpages + fsync, check for both separately instead of using just putpages to decide the op's faith. And the real beef in this commit is of course a tyop fix in a comment.
1.30	09-Jan-2007	pooka	Introduce flush operations, which the fs server can use to control kernel caching. Currently supported are only flushing the name cache for a directory or flushing the name cache for the entire fs. Also, get rid of PNODE_INACTIVE status, since it was racy and essentially didn't work. All this on top of being useless in the first place ....
1.29	07-Jan-2007	pooka	getcwd wants eofflag - set eofflag in readdir if amount of data is 0
1.28	02-Jan-2007	pooka	In rename, tdvp == tvp holds if we are renaming a directory to "." (XXX: for all the sense that makes). Deal with it gracefully here for now.
1.27	01-Jan-2007	pooka	remove r/o mount check done also in vfs lookup()
1.26	01-Jan-2007	pooka	async update node metadata for spec- and fifoops
1.25	01-Jan-2007	pooka	properly handle VOP_REMOVE case where vp == dvp
1.24	01-Jan-2007	pooka	explicitly disable ioctl and fcntl for now - support has bitrotted
1.23	30-Dec-2006	pooka	branches: 1.23.2; * use PUFFS_KFLAG_NOCACHE to also signal that we don't want the namecache * enter files into the namecache immediately when new nodes are created (if it's a caching mount, of course)
1.22	09-Dec-2006	chs	branches: 1.22.2; a smorgasbord of improvements to vnode locking and path lookup: - LOCKPARENT is no longer relevant for lookup(), relookup() or VOP_LOOKUP(). these now always return the parent vnode locked. namei() works as before. lookup() and various other paths no longer acquire vnode locks in the wrong order via vrele(). fixes PR 32535. as a nice side effect, path lookup is also up to 25% faster. - the above allows us to get rid of PDIRUNLOCK. - also get rid of WANTPARENT (just use LOCKPARENT and unlock it). - remove an assumption in layer_node_find() that all file systems implement a recursive VOP_LOCK() (unionfs doesn't). - require that all file systems supply vfs_vptofh and vfs_fhtovp routines. fill in eopnotsupp() for file systems that don't support being exported and remove the checks for NULL. (layerfs calls these without checking.) - in union_lookup1(), don't change refcounts in the ISDOTDOT case, just adjust which vnode is locked. fixes PR 33374. - apply fixes for ufs_rename() from ufs_vnops.c rev. 1.61 to ext2fs_rename().
1.21	07-Dec-2006	pooka	let implementation ultimately decide if mmap is supported - pass VOP_MMAP to fs server
1.20	05-Dec-2006	pooka	adjust file size in write only if file grows. but since this change is in the "never use ubc" branch, I don't think it matters except for cosmetics.
1.19	05-Dec-2006	pooka	Allow multiple requests to be transferred in each GET/PUTOP. For a single request, the performance is still the same.
1.18	01-Dec-2006	pooka	branches: 1.18.2; prefix kernel flags with PUFFS_KFLAG to have a separate namespace from the library flags
1.17	01-Dec-2006	pooka	don't call the fs server for all operations, only those it has told us that it implements
1.16	28-Nov-2006	pooka	don't allow mmap if operating uncached
1.15	18-Nov-2006	pooka	Actually, for NOCACHE, use direct read/write instead of going through page cache at all and invalidating. XXX: mmap
1.14	18-Nov-2006	pooka	branches: 1.14.2; make puffs_strategy more robust
1.13	18-Nov-2006	pooka	Require statvfs info from startreq so that we have that info available. Also, don't pass fsid to userspace and just fill it in the kernel.
1.12	18-Nov-2006	pooka	As a first generation best-effort hack, use NOCACHE to mean "file size can change without the kernel knowing" and therefore query the file size before invoking read or write operations.
1.11	17-Nov-2006	pooka	Introduce uncached operation, makes sense when the file system backend can be modified from elsewhere than the file system interface
1.10	13-Nov-2006	pooka	No need to return a special value for CREATE/RENAME lookup, so just handle ENOENT. If there's a real error, userspace will return something else.
1.9	08-Nov-2006	pooka	update struct buf resid in strategy according to what was transferred. seems like only nestiobuf complains when it wasn't updated ...
1.8	07-Nov-2006	pooka	attach to genfs & support page cache. most noticeable effect is mmap and therefore execution of binaries starting to work, some speed improvements with large file I/O also. caching semantics and error case handling most likely need revisiting.
1.7	27-Oct-2006	pooka	Use spec_fsync for specops vop_fsync: it knows about vflushbuf(), which is more than what puffs currently knows. makes e.g. ffs unmount for a puffs-based device node work.
1.6	27-Oct-2006	pooka	support fifos
1.5	26-Oct-2006	pooka	support specfs
1.4	26-Oct-2006	pooka	Fix operations creating new nodes to honor the vnode locking protocol if the userspace server returns an error. Fixes lockups if any of the following operations failed: create, mknod, mkdir, symlink
1.3	25-Oct-2006	pooka	pass VOP_INACTIVE() to userspace
1.2	23-Oct-2006	pooka	fix print in VOP_PRINT also make it compile on amd64. problem noticed by Blair Sadewitz on current-users
1.1	22-Oct-2006	pooka	kernel portion of puffs - the Pass-to-Userspace Framework File System. It contains the VFS attachment and userspace message-passing interface. This work was initially started and completed for Google SoC 2005 and tweaked to work a bit better in the past few weeks. While being far from complete, it is functional enough to be able and stable to host a fairly general-purpose in-memory file system in userspace. Even so, puffs should be considered experimental and no binary compatibility for interfaces or crash-freedom or zero security implications should be relied upon just yet. The GSoC project was mentored by William Studenmund and the final review for the code was done by Christos.
1.14.2.5	09-Feb-2007	ad	Sync with HEAD.
1.14.2.4	01-Feb-2007	ad	Sync with head.
1.14.2.3	12-Jan-2007	ad	Sync with head.
1.14.2.2	18-Nov-2006	ad	Sync with head.
1.14.2.1	18-Nov-2006	ad	file puffs_vnops.c was added on branch newlock2 on 2006-11-18 21:39:20 +0000
1.18.2.1	17-Feb-2007	tron	Apply patch (requested by chs in ticket #422): - Fix various deadlock problems with nullfs and unionfs. - Speed up path lookups by upto 25%.
1.22.2.2	10-Dec-2006	yamt	sync with head.
1.22.2.1	09-Dec-2006	yamt	file puffs_vnops.c was added on branch yamt-splraiseipl on 2006-12-10 07:18:38 +0000
1.23.2.8	04-Feb-2008	yamt	sync with head.
1.23.2.7	21-Jan-2008	yamt	sync with head
1.23.2.6	07-Dec-2007	yamt	sync with head
1.23.2.5	27-Oct-2007	yamt	sync with head.
1.23.2.4	03-Sep-2007	yamt	sync with head.
1.23.2.3	26-Feb-2007	yamt	sync with head.
1.23.2.2	30-Dec-2006	yamt	sync with head.
1.23.2.1	30-Dec-2006	yamt	file puffs_vnops.c was added on branch yamt-lazymbuf on 2006-12-30 20:50:01 +0000
1.51.2.5	17-May-2007	yamt	sync with head.
1.51.2.4	07-May-2007	yamt	sync with head.
1.51.2.3	15-Apr-2007	yamt	sync with head.
1.51.2.2	24-Mar-2007	yamt	sync with head.
1.51.2.1	27-Feb-2007	yamt	- sync with head. - move sched_changepri back to kern_synch.c as it doesn't know PPQ anymore.
1.52.6.1	11-Jul-2007	mjf	Sync with head.
1.52.4.14	23-Oct-2007	ad	Sync with head.
1.52.4.13	12-Oct-2007	ad	Sync with head.
1.52.4.12	09-Oct-2007	ad	Sync with head.
1.52.4.11	09-Oct-2007	ad	Sync with head.
1.52.4.10	16-Sep-2007	ad	Checkpoint work in progress on the vnode lifecycle and reference counting stuff. This makes it work properly without kernel_lock and fixes a few quite old bugs. See vfs_subr.c 1.283.2.17 for details.
1.52.4.9	20-Aug-2007	ad	Sync with HEAD.
1.52.4.8	19-Aug-2007	ad	- Back out the biodone() changes. - Eliminate B_ERROR (from HEAD).
1.52.4.7	15-Jul-2007	ad	Sync with head.
1.52.4.6	17-Jun-2007	ad	- Increase the number of thread priorities from 128 to 256. How the space is set up is to be revisited. - Implement soft interrupts as kernel threads. A generic implementation is provided, with hooks for fast-path MD code that can run the interrupt threads over the top of other threads executing in the kernel. - Split vnode::v_flag into three fields, depending on how the flag is locked (by the interlock, by the vnode lock, by the file system). - Miscellaneous locking fixes and improvements.
1.52.4.5	09-Jun-2007	ad	Sync with head.
1.52.4.4	08-Jun-2007	ad	Sync with head.
1.52.4.3	10-Apr-2007	ad	Sync with head.
1.52.4.2	09-Apr-2007	ad	- Add two new arguments to kthread_create1: pri_t pri, bool mpsafe. - Fork kthreads off proc0 as new LWPs, not new processes.
1.52.4.1	05-Apr-2007	ad	Compile fixes.
1.53.2.1	29-Mar-2007	reinoud	Pullup to -current
1.88.2.2	03-Sep-2007	skrll	Sync with HEAD.
1.88.2.1	15-Aug-2007	skrll	Sync with HEAD.
1.95.6.2	30-Jul-2007	pooka	properly setup ubcflags
1.95.6.1	30-Jul-2007	pooka	file puffs_vnops.c was added on branch matt-mips64 on 2007-07-30 14:49:02 +0000
1.95.4.9	09-Dec-2007	jmcneill	Sync with HEAD.
1.95.4.8	27-Nov-2007	joerg	Sync with HEAD. amd64 Xen support needs testing.
1.95.4.7	21-Nov-2007	joerg	Sync with HEAD.
1.95.4.6	28-Oct-2007	joerg	Sync with HEAD.
1.95.4.5	26-Oct-2007	joerg	Sync with HEAD. Follow the merge of pmap.c on i386 and amd64 and move pmap_init_tmp_pgtbl into arch/x86/x86/pmap.c. Modify the ACPI wakeup code to restore CR4 before jumping back into kernel space as the large page option might cover that.
1.95.4.4	07-Oct-2007	joerg	Sync with HEAD.
1.95.4.3	02-Oct-2007	joerg	Sync with HEAD.
1.95.4.2	03-Sep-2007	jmcneill	Sync with HEAD.
1.95.4.1	16-Aug-2007	jmcneill	Sync with HEAD.
1.98.4.2	14-Oct-2007	yamt	sync with head.
1.98.4.1	06-Oct-2007	yamt	sync with head.
1.98.2.3	23-Mar-2008	matt	sync with HEAD
1.98.2.2	09-Jan-2008	matt	sync with HEAD
1.98.2.1	06-Nov-2007	matt	sync with HEAD
1.107.2.4	21-Nov-2007	bouyer	Sync with HEAD
1.107.2.3	18-Nov-2007	bouyer	Sync with HEAD
1.107.2.2	13-Nov-2007	bouyer	Sync with HEAD
1.107.2.1	25-Oct-2007	bouyer	Sync with HEAD.
1.113.2.4	18-Feb-2008	mjf	Sync with HEAD.
1.113.2.3	27-Dec-2007	mjf	Sync with HEAD.
1.113.2.2	08-Dec-2007	mjf	Sync with HEAD.
1.113.2.1	19-Nov-2007	mjf	Sync with HEAD.
1.121.2.2	26-Dec-2007	ad	Sync with head.
1.121.2.1	04-Dec-2007	ad	Pull the vmlocking changes into a new branch.
1.122.4.2	08-Jan-2008	bouyer	Sync with HEAD
1.122.4.1	02-Jan-2008	bouyer	Sync with HEAD
1.128.16.2	13-Dec-2008	haad	Update haad-dm branch to haad-dm-base2.
1.128.16.1	19-Oct-2008	haad	Sync with HEAD.
1.128.12.1	24-Sep-2008	wrstuden	Merge in changes between wrstuden-revivesa-base-2 and wrstuden-revivesa-base-3.
1.128.10.4	11-Aug-2010	yamt	sync with head.
1.128.10.3	11-Mar-2010	yamt	sync with head
1.128.10.2	16-Sep-2009	yamt	sync with head
1.128.10.1	04-May-2009	yamt	sync with head.
1.128.6.2	17-Jan-2009	mjf	Sync with HEAD.
1.128.6.1	28-Sep-2008	mjf	Sync with HEAD.
1.129.8.1	21-Apr-2010	matt	sync to netbsd-5
1.129.4.11	02-Nov-2011	riz	Pull up following revision(s) (requested by manu in ticket #1679): sys/fs/puffs/puffs_vnops.c: revision 1.157 sys/fs/puffs/puffs_vnops.c: revision 1.158 sys/fs/puffs/puffs_vnops.c: revision 1.159 sys/fs/puffs/puffs_vfsops.c: revision 1.97 sys/fs/puffs/puffs_vfsops.c: revision 1.99 sys/fs/puffs/puffs_vnops.c: revision 1.160 sys/fs/puffs/puffs_vfsops.c: revision 1.100 sys/miscfs/syncfs/sync_subr.c: revision 1.47 sys/fs/puffs/puffs_node.c: revision 1.21 sys/fs/puffs/puffs_node.c: revision 1.22 sys/fs/puffs/puffs_msgif.c: revision 1.88 sys/fs/puffs/puffs_msgif.c: revision 1.89 sys/fs/puffs/puffs_vnops.c: revision 1.156 Make sure ioflush does not sleep in PUFFS code path, waiting for a mutex, a memory allocation, or a response from the filesystem. This avoids deadlocks in the following situations: 1) when memory is low: ioflush waits the fileystem, the fielsystem waits for memory 2) when the filesystem does not respond (e.g.: network outage ona distributed filesystem) Fix the build that was broken by struct lwp *updateproc reference in RUMP-visible code. Instead of checking that updateproc (aka ioflush, aka syncer) will not sleep in PUFFS code, I check for any kernel thread: after all none of them are designed to hang awaiting for a remote filesystem operation to complete. Roll back the change that forced kernel threads to not sleep in PUFFS. The change does not make consensus, since only pagedaemon should need it. Other threads will tolerate sleeping, and problems here are only symptoms that something is going wrong in memory management. The cause, not the symptoms, need to be fixed. Make sure pagedaemon does not sleep for memory in puffs_vnop_sleep. Add KASSERT on any sleeping memory allocation to check it cannot happen again. Remove #ifdef DIAGNOSTIC guards around KASSERT, as the macro contains them
1.129.4.10	17-Sep-2011	bouyer	Pull up following revision(s) (requested by manu in ticket #1666): sys/fs/puffs/puffs_sys.h: revision 1.78 via patch sys/fs/puffs/puffs_node.c: revision 1.20 via patch sys/fs/puffs/puffs_vnops.c: revision 1.155 via patch Add a mutex for operations that touch size (setattr, getattr, write, fsync). This is required to avoid data corruption bugs, where a getattr slices itself within a setattr operation, and sets the size to the stall value it got from the filesystem. That value is smaller than the one set by setattr, and the call to uvm_vnp_setsize() trigged a spurious truncate. The result is a chunk of zeroed data in the file. Such a situation can easily happen when the ioflush thread issue a VOP_FSYNC/puffs_vnop_sync/flushvncache/dosetattrn while andother process do a sys_stat/VOP_GETATTR/puffs_vnop_getattr. This mutex on size operation can be removed the day we decide VOP_GETATTR has to operated on a locked vnode, since the other operations that touch size already require that.
1.129.4.9	17-Jul-2011	riz	Pull up following revision(s) (requested by manu in ticket #1645): lib/libc/sys/Makefile.inc 1.207 via patch lib/libc/sys/extattr_get_file.2 patch lib/libpuffs/dispatcher.c 1.34,1.36 via patch lib/libpuffs/puffs.c 1.107 via patch lib/libpuffs/puffs.h 1.115,1.118 via patch sys/fs/puffs/puffs_msgif.h 1.71,1.76 via patch sys/fs/puffs/puffs_vfsops.c 1.88 via patch sys/fs/puffs/puffs_vnops.c 1.145,1.154 via patch sys/kern/vfs_xattr.c 1.24-1.27 via patch sys/kern/vnode_if.c 1.87 via patch sys/sys/Makefile 1.133 via patch sys/sys/extattr.h 1.6 via patch sys/sys/vnode_if.h 1.81 via patch sys/ufs/ffs/ffs_vnops.c patch sys/ufs/ufs/ufs_extattr.c 1.31,1.34 via patch * support extended attributes * bump major due to structure growth * add some spare space * remove ABI sillyness Support extended attributes. Fix multiple non compliances in our Linux-like extattr API, and make it public so that it can be used. Improve a bit listxattr(2). It attemps to list both system and user extended attributes, and it faled if calling user did not have privilege for reading system EA. Now we just lise user EA and skip system EA in reading them is not allowed. Fix bug introduced in previous commuit: Do not vrele() a vnode we did not obtained. Improve UFS1 extended attributes usability - autocreate attribute backing file for new attributes - autoload attributes when issuing extattrctl start - when autoloading attributes, do not display garbage warning when looking up entries that got ENOENT Add a flag to VOP_LISTEXTATTR(9) so that the vnode interface can tell the filesystem in which format extended attribute shall be listed. There are currently two formats: - NUL-terminated strings, used for listxattr(2), this is the default. - one byte length-pprefixed, non NUL-terminated strings, used for extattr_list_file(2), which is obtanined by setting the EXTATTR_LIST_PREFIXLEN flag to VOP_LISTEXTATTR(9) This approach avoid the need for converting the list back and forth, except in libperfuse, since FUSE uses NUL-terminated strings, and the kernel may have requested EXTATTR_LIST_PREFIXLEN.
1.129.4.8	18-Jun-2011	bouyer	Pull up following revision(s) (requested by manu in ticket #1623): lib/libpuffs/puffs.c: revision 1.116 via patch sys/fs/puffs/puffs_vnops.c: revision 1.151 via patch Call advlock method if supplied
1.129.4.7	16-Jan-2010	bouyer	Pull up following revision(s) (requested by pooka in ticket #1244): sys/fs/puffs/puffs_vnops.c: revision 1.142 Since VOP_GETATTR() does not require a locked vnode, resolve and reference the puffs_node before sending the request to the file server. This diminishes the window where the inode can be reclaimed and be invalidated before it is accessed (but does not completely eliminate the race, as that is a caller problem which we cannot fix here).
1.129.4.6	18-Dec-2009	snj	Pull up following revision(s) (requested by pooka in ticket #1184): sys/fs/puffs/puffs_vnops.c: revision 1.141 via patch Push all information cached in the vnode to the file server before issuing INACTIVE. PR kern/42194. Also, send setattr in fsync asynchronously if FSYNC_WAIT is not set.
1.129.4.5	28-Nov-2009	bouyer	Pull up following revision(s) (requested by pooka in ticket #1154): sys/fs/puffs/puffs_vnops.c: revision 1.140 Send VOP_ABORTOP() in case attempting cross-dev rename, part of PR kern/42210. Also, fix a memory management error in said case.
1.129.4.4	28-Nov-2009	bouyer	Pull up following revision(s) (requested by pooka in ticket #1153): sys/fs/puffs/puffs_vnops.c: revision 1.139 Send VOP_ABORTOP() as a FAF -- we don't care about the return value.
1.129.4.3	18-Oct-2009	sborrill	Pull up the following revisions(s) (requested by pooka in ticket #1100): lib/libpuffs/dispatcher.c: revision 1.33 lib/libpuffs/puffs.c: revision 1.99 lib/libpuffs/puffs.h: revision 1.111 sys/fs/puffs/puffs_msgif.h: revision 1.67 via patch sys/fs/puffs/puffs_vnops.c: revision 1.136 Support VOP_ABORTOP() in puffs.
1.129.4.2	03-Oct-2009	snj	Pull up following revision(s) (requested by pooka in ticket #1042): sys/fs/puffs/puffs_node.c: revision 1.14 sys/fs/puffs/puffs_vnops.c: revision 1.134 * fix a race i introduced almost two years ago in rev 1.116: operations creating a node cannot be considered outgoing operations, since after return from userspace they modify file system state by creating a new node. if we do not protect the file system by holding the directory lock, a lookup operation might race us into the kernel and create the node earlier. * remove pnode from hashlish before sending the reclaim faf off to userspace. also, hold pmp_lock while frobbing the list.
1.129.4.1	26-Sep-2009	snj	Pull up following revision(s) (requested by pooka in ticket #1014): sys/fs/puffs/puffs_vnops.c: revision 1.133 Set SAVENAME for rmdir and remove. Addresses an easy part of PR kern/38188
1.129.2.1	19-Jan-2009	skrll	Sync with HEAD.
1.142.4.5	31-May-2011	rmind	sync with head
1.142.4.4	05-Mar-2011	rmind	sync with head
1.142.4.3	03-Jul-2010	rmind	sync with head
1.142.4.2	30-May-2010	rmind	sync with head
1.142.4.1	16-Mar-2010	rmind	Change struct uvm_object::vmobjlock to be dynamically allocated with mutex_obj_alloc(). It allows us to share the locks among UVM objects.
1.142.2.2	17-Aug-2010	uebayasi	Sync with HEAD.
1.142.2.1	30-Apr-2010	uebayasi	Sync with HEAD.
1.150.2.1	06-Jun-2011	jruoho	Sync with HEAD.
1.152.2.1	23-Jun-2011	cherry	Catchup with rmind-uvmplock merge.
1.161.2.5	22-May-2014	yamt	sync with head. for a reference, the tree before this commit was tagged as yamt-pagecache-tag8. this commit was splitted into small chunks to avoid a limitation of cvs. ("Protocol error: too many arguments")
1.161.2.4	16-Jan-2013	yamt	sync with (a bit old) head
1.161.2.3	30-Oct-2012	yamt	sync with head
1.161.2.2	23-May-2012	yamt	sync with head.
1.161.2.1	17-Apr-2012	yamt	sync with head
1.162.4.3	29-Apr-2012	mrg	sync to latest -current.
1.162.4.2	05-Apr-2012	mrg	sync to latest -current.
1.162.4.1	18-Feb-2012	mrg	merge to -current.
1.163.2.12	27-Feb-2015	martin	Pull up following revision(s) (requested by manu in ticket #1260): lib/libpuffs/puffs.3: revision 1,55,1.60 sys/fs/puffs/puffs_msgif.h: revision 1.84 lib/libperfuse/ops.c: revision 1.83 sys/fs/puffs/puffs_sys.h: revision 1.89 sys/fs/puffs/puffs_vfsops.c: revision 1.116 lib/libperfuse/perfuse.c: revision 1.36 sys/fs/puffs/puffs_vnops.c: revision 1.200-1.202 Use more markup. New sentence, new line. Bump date for previous. Add PUFFS_KFLAG_NOFLUSH_META to prevent sending metadata flush to FUSE FUSE filesystems do not expect to get metadata updates for [amc]time and size, they updates the value on their own after operations. The PUFFS PUFFS_KFLAG_NOFLUSH_META option prevents regular metadata cache flushes to the filesystem , and libperfuse uses it to match Linux FUSE behavior. While there, fix a bug in SETATTR: do not update kernel metadata cache from SETATTR reply when the request is asynchronous, as we do not have the reply yet. Update file size after write without metadata flush If we do not use metadata flush, we must make sure the size is updated in the filesystem after a write, otherwise the next GETATTR will get us a stale value and the file will be truncated.
1.163.2.11	16-Jan-2015	martin	Pull up following revision(s) (requested by manu in ticket #1236): sys/fs/puffs/puffs_vnops.c: revision 1.199 Make sure reads on empty files reach PUFFS filesystems Sending a read through the page cache will get the operation short-circuited. This is a problem with some filesystems that expect to receive the read operation in order to update atime. We fix that by bypassing the page cache when reading a file wich a size known to be zero.
1.163.2.10	09-Nov-2014	msaitoh	Pull up following revision(s) (requested by manu in ticket #1187): sys/fs/puffs/puffs_vnops.c: revision 1.198 PUFFS direct I/O cache fix There are a few situations where we must take care of the cache if direct I/O was enabled: - if we do direct I/O for write but not for read, then any write must invalidate the cache so that a reader gets the written data and not the not-updated cache. - if we used a vnode without direct I/O and it is enabled for writing, we must flush the cache before compeling the open operation, so that the cachec write are not lost. And at inactive time, we wipe direct I/O flags so that a new open without direct I/O does not inherit direct I/O.
1.163.2.9	09-Nov-2014	msaitoh	Pull up following revision(s) (requested by manu in ticket #1184): sys/fs/puffs/puffs_vnops.c: revision 1.195 According to pooka@'s comment, a long time ago, VOP_STRATEGY could not fail without taking down the kernel. It seems this is not the case anymore, hence we can stop dropping errors in puffs_vnop_strategy() Approved by pooka@
1.163.2.8	09-Nov-2014	msaitoh	Pull up following revision(s) (requested by manu in ticket #1166): sys/fs/puffs/puffs_vnops.c: revision 1.188-1.194 - If we truncate the file, make sure we zero-fill the end of the last page, otherwise if the file is later truncated to a larger size (creating a hole), that area will not return zeroes as it should. - Use PRIx64 for printing offsets - Improve zero-fill of last page after shrink fix: 1) do it only if the file is open for writing, otherwise we send write requests to the FS on a file that has never been open. 2) do it inside existing if (vap->va_size != VNOVAL) block - Retore LP64 fix that was removed by mistake - Make this build again without debugging enabled; DPRINTF() can end up as empty, and in an if conditional, you then need braces if that's the only potential body. - As is evidenced by several of our 32-bit MIPS ports, it's wrong to print vsize_t with PRIx64 -- instead use our own PRIxVSIZE macro. - Do the previous correctly...
1.163.2.7	03-Nov-2014	msaitoh	Pull up following revision(s) (requested by manu in ticket #1152): sys/fs/puffs/puffs_vnops.c: revision 1.186 PUFFS fixes for size update ater write plus read/write sanity checks - Always update kernel metadata cache for size when writing This fixes situation where size update after appending to a file lagged - Make read/write nilpotent when called with null size, as FFS does - Return EFBIG instead of EINVAL for negative offsets, as FFS does
1.163.2.6	03-Nov-2014	msaitoh	Pull up following revision(s) (requested by manu in ticket #1149): sys/fs/puffs/puffs_node.c: revision 1.33 sys/fs/puffs/puffs_vnops.c: revision 1.185 When changing a directory content, update the ctime/mtime in kernel cache, otherwise the updated ctime/mtime appears after the cached entry expire.
1.163.2.5	03-Nov-2014	msaitoh	Pull up following revision(s) (requested by manu in ticket #1140): lib/libperfuse/ops.c 1.63-1.69 lib/libperfuse/perfuse.c 1.32-1.33 lib/libperfuse/perfuse_priv.h 1.32-1.34 lib/libperfuse/subr.c 1.20 lib/libpuffs/creds.c 1.16 lib/libpuffs/dispatcher.c 1.47 lib/libpuffs/puffs.h 1.125 lib/libpuffs/puffs_ops.3 1.37-1.38 lib/libpuffs/requests.c 1.24 sys/fs/puffs/puffs_msgif.h 1.81 sys/fs/puffs/puffs_sys.h 1.85 sys/fs/puffs/puffs_vnops.c 1.183 usr.sbin/perfused/msg.c 1.22 Bring libpuffs, libperfuse and perfused on par with -current: - implement FUSE direct I/O - remove useless code and warnings - fix missing GETATTR bugs - fix exended attribute get and list operations
1.163.2.4	12-Aug-2012	martin	Pull up following revision(s) (requested by manu in ticket #438): lib/libperfuse/perfuse_priv.h: revision 1.31 sys/fs/puffs/puffs_msgif.h: revision 1.80 sys/fs/puffs/puffs_vnops.c: revision 1.171 lib/libpuffs/puffs_ops.3: revision 1.31 sys/fs/puffs/puffs_vnops.c: revision 1.172 sys/fs/puffs/puffs_vnops.c: revision 1.173 sys/fs/puffs/puffs_vnops.c: revision 1.174 usr.sbin/perfused/perfused.c: revision 1.24 sys/fs/puffs/puffs_sys.h: revision 1.80 sys/fs/puffs/puffs_sys.h: revision 1.81 sys/fs/puffs/puffs_sys.h: revision 1.82 lib/libperfuse/subr.c: revision 1.19 lib/libperfuse/perfuse.c: revision 1.30 sys/fs/puffs/puffs_msgif.c: revision 1.90 sys/fs/puffs/puffs_msgif.c: revision 1.91 sys/fs/puffs/puffs_msgif.c: revision 1.92 lib/libperfuse/ops.c: revision 1.59 lib/libpuffs/puffs.3: revision 1.53 lib/libperfuse/debug.c: revision 1.12 lib/libpuffs/puffs.3: revision 1.54 sys/fs/puffs/puffs_vnops.c: revision 1.167 sys/fs/puffs/puffs_msgif.h: revision 1.79 usr.sbin/perfused/msg.c: revision 1.21 sys/fs/puffs/puffs_vfsops.c: revision 1.102 sys/fs/puffs/puffs_vfsops.c: revision 1.103 sys/fs/puffs/puffs_vfsops.c: revision 1.105 lib/libpuffs/puffs.h: revision 1.123 lib/libperfuse/perfuse_if.h: revision 1.20 lib/libperfuse/perfuse.c: revision 1.29 lib/libpuffs/dispatcher.c: revision 1.42 lib/libpuffs/dispatcher.c: revision 1.43 - Fix same vnodes associated with multiple cookies The scheme used to retreive known nodes on lookup was flawed, as it only used parent and name. This produced a different cookie for the same file if it was renamed, when looking up ../ or when dealing with multiple files associated with the same name through link(2). We therefore abandon the use of node name and introduce hashed lists of inodes. This causes a huge rewrite of reclaim code, which do not attempt to keep parents allocated until all their children are reclaimed - Fix race conditions in reclaim There are a few situations where we issue multiple FUSE operations for a PUFFS operation. On reclaim, we therefore have to wait for all FUSE operation to complete, not just the current exchanges. We do this by introducing node reference count with node_ref() and node_rele(). - Detect data loss caused by FAF VOP_PUTPAGES causes FAF writes where the kernel does not check the operation result. At least issue a warning on error. - Enjoy FAF shortcut on setattr No need to wait for the result if the kernel does not want it. There is however an exception for setattr that touch the size, we need to wait for completion because we have other operations queued for after the resize. - Fix fchmod() on write-open file fchmod() on a node open with write privilege will send setattr with both mode and size set. This confuses some FUSE filesystem. Therefore we send two FUSE operations, one for mode, and one for size. - Remove node TTL handling for netbsd-5 for simplicity sake. The code still builds on netbsd-5 but does not have the node TTL feature anymore. It works fine with kernel support on netbsd-6. - Improve PUFFS_KFLAG_CACHE_FS_TTL by reclaiming older inactive nodes. The normal kernel behavior is to retain inactive nodes in the freelist until it runs out of vnodes. This has some merit for local filesystems, where the cost of an allocation is about the same as the cost of a lookup. But that situation is not true for distributed filesystems. On the other hand, keeping inactive nodes for a long time hold memory in the file server process, and when the kernel runs out of vnodes, it produce reclaim avalanches that increase lattency for other operations. We do not reclaim inactive vnodes immediatly either, as they may be looked up again shortly. Instead we introduce a grace time and we reclaim nodes that have been inactive beyond the grace time. - Fix lookup/reclaim race condition. The above improvement undercovered a race condition between lookup and reclaim. If we reclaimed a vnode associated with a userland cookie while a lookup returning that same cookiewas inprogress, then the kernel ends up with a vnode associated with a cookie that has been reclaimed in userland. Next operation on the cookie will crash (or at least confuse) the filesystem. We fix this by introducing a lookup count in kernel and userland. On reclaim, the kernel sends the count, which enable userland to detect situation where it initiated a lookup that is not completed in kernel. In such a situation, the reclaim must be ignored, as the node is about to be looked up again. Fix hang unmount bug introduced by last commit. We introduced a slow queue for delayed reclaims, while the existing queue for unmount, flush and exist has been renamed fast queue. Both queues had timestamp for when an operation should be done, but it was useless for the fast queue, which is always used to run an operation ASAP. And the timestamp test had an error that turned ASAP into "at next tick", but nobody what there to wake the thread at next tick, hence the hang. The fix is to remove the useless and buggy timestamp test for fast queue. Rename slow sopreq queue into node sopreq queue, to refet the fact that is only intended for postponed node reclaims. When purging the node sopreq queue, do not call puffs_msg_sendresp(), as it makes no sense. Fix race condition between (create\|mknod\|mkdir\|symlino) and reclaim, just like we did it between lookup and reclaim. Missing bit in previous commit (prevent race between create\|mknod\|mkdir\|symlink and reclaim) Bump date for previous. New sentence, new line; remove trailing whitespace; fix typos; punctuation nits. Add PUFFS_KFLAG_CACHE_DOTDOT so that vnodes hold a reference on their parent, keeping them active, and allowing to lookup .. without sending a request to the filesystem. Enable the featuure for perfused, as this is how FUSE works. Missing bit in previous commit (PUFFS_KFLAG_CACHE_DOTDOT option to avoid looking up ..)
1.163.2.3	12-Aug-2012	martin	Pull up following revision(s) (requested by manu in ticket #484): sys/fs/nilfs/nilfs_vnops.c: revision 1.18 sys/ufs/ufs/ufs_lookup.c: revision 1.117 sys/nfs/nfs_vnops.c: revision 1.295 sys/ufs/chfs/chfs_vnops.c: revision 1.8 sys/ufs/ext2fs/ext2fs_lookup.c: revision 1.70 sys/fs/unionfs/unionfs_vnops.c: revision 1.6 sys/kern/vfs_cache.c: revision 1.89 sys/fs/efs/efs_vnops.c: revision 1.26 sys/fs/hfs/hfs_vnops.c: revision 1.26 sys/fs/adosfs/adlookup.c: revision 1.16 sys/fs/puffs/puffs_vnops.c: revision 1.168 sys/fs/tmpfs/tmpfs_vnops.c: revision 1.98 sys/fs/ntfs/ntfs_vnops.c: revision 1.52 sys/fs/cd9660/cd9660_lookup.c: revision 1.20 sys/fs/msdosfs/msdosfs_lookup.c: revision 1.24 sys/fs/smbfs/smbfs_vnops.c: revision 1.80 sys/fs/udf/udf_vnops.c: revision 1.72 sys/fs/filecorefs/filecore_lookup.c: revision 1.14 sys/fs/puffs/puffs_node.c: revision 1.25 Move some the test for MAKEENTRY into the cache_enter(9). Make some variables in vfs_cache.c static, __read_mostly, etc. No objection on tech-kern@.
1.163.2.2	23-Apr-2012	riz	Pull up following revision(s) (requested by manu in ticket #195): lib/libskey/skeysubr.c: revision 1.27 lib/libkvm/kvm_getloadavg.c: revision 1.11 lib/libwrap/update.c: revision 1.9 lib/liby/yyerror.c: revision 1.9 lib/libpuffs/puffs_ops.3: revision 1.30 lib/libwrap/misc.c: revision 1.10 lib/libwrap/hosts_access.c: revision 1.20 lib/libpuffs/pnode.c: revision 1.11 lib/libperfuse/subr.c: revision 1.17 lib/libpuffs/pnode.c: revision 1.12 lib/libperfuse/subr.c: revision 1.18 lib/libwrap/options.c: revision 1.15 lib/libwrap/fix_options.c: revision 1.11 lib/libperfuse/ops.c: revision 1.52 lib/libperfuse/ops.c: revision 1.53 lib/libperfuse/ops.c: revision 1.54 lib/libwrap/hosts_ctl.c: revision 1.5 lib/libintl/gettext.c: revision 1.27 lib/libwrap/shell_cmd.c: revision 1.6 lib/libpuffs/dispatcher.c: revision 1.39 lib/libperfuse/perfuse_priv.h: revision 1.27 lib/libwrap/socket.c: revision 1.19 lib/libpuffs/puffs.3: revision 1.50 lib/libperfuse/perfuse_priv.h: revision 1.28 lib/libpuffs/puffs_priv.h: revision 1.45 lib/libpuffs/puffs.3: revision 1.51 lib/libperfuse/perfuse_priv.h: revision 1.29 lib/libwrap/percent_x.c: revision 1.5 lib/libpuffs/puffs.3: revision 1.52 lib/libperfuse/debug.c: revision 1.11 sys/fs/puffs/puffs_vnops.c: revision 1.165 lib/libwrap/tcpd.h: revision 1.13 sys/fs/puffs/puffs_vnops.c: revision 1.166 lib/libwrap/eval.c: revision 1.7 sys/fs/puffs/puffs_msgif.h: revision 1.78 sys/fs/puffs/puffs_vfsops.c: revision 1.101 lib/libwrap/rfc931.c: revision 1.9 lib/libwrap/clean_exit.c: revision 1.5 lib/libpuffs/puffs.h: revision 1.120 lib/libc/stdlib/jemalloc.c: revision 1.27 lib/librmt/rmtlib.c: revision 1.26 lib/libpuffs/puffs.h: revision 1.121 sys/fs/puffs/puffs_sys.h: revision 1.79 lib/librumpclient/rumpclient.c: revision 1.48 lib/libwrap/refuse.c: revision 1.5 lib/libperfuse/perfuse.c: revision 1.26 lib/libperfuse/perfuse.c: revision 1.27 tests/fs/puffs/t_fuzz.c: revision 1.5 lib/libperfuse/perfuse.c: revision 1.28 lib/libpuffs/dispatcher.c: revision 1.40 sys/fs/puffs/puffs_node.c: revision 1.24 lib/libwrap/diag.c: revision 1.9 lib/libintl/textdomain.c: revision 1.13 Use C89 function definition Add name and atttribute cache with filesytem provided TTL. lookup, create, mknod, mkdir, symlink, getattr and setattr messages have been extended so that attributes and their TTL can be provided by the filesytem. lookup, create, mknod, mkdir, and symlink messages are also extended so that the filesystem can provide name TTL. Add PUFFS_KFLAG_CACHE_FS_TTL flag to puffs_init(3) to use name and attribute cache with filesystem provided TTL. lookup, create, mknod, mkdir, symlink, getattr and setattr messages have been extended so that attributes and their TTL can be provided by the filesytem. lookup, create, mknod, mkdir, and symlink messages are also extended so that the filesystem can provide name TTL. The filesystem updates attributes and TTL using puffs_pn_getvap(3), puffs_pn_getvattl(3), and puffs_pn_getcnttl(3) Use new PUFFS_KFLAG_CACHE_FS_TTL option to puffs_init(3) so that FUSE TTL on name and attributes are used. This save many PUFFS operations and improves performances. PUFFS_KFLAG_CACHE_FS_TTL is #ifdef'ed in many places for now so that libperfuse can still be used on netbsd-5. Split file system. Comma fixes. Remove dangling "and". Bump date for previous. - Makesure update_va does not change vnode size when it should not. For instance when doing a fault-issued VOP_GETPAGES within VOP_WRITE, changing size leads to panic: genfs_getpages: past eof. -Handle ticks wrap around for vnode name andattribute timeout - When using PUFFS_KFLAG_CACHE_FS_TTL, do not use puffs_node to carry attribute and TTL fora newly created node. Instead extend puffs_newinfo and add puffs_newinfo_setva() and puffs_newinfo_setttl() - Remove node_mk_common_final in libperfuse. It used to set uid/gid for a newly created vnode but has been made redundant along time ago since uid and gid are properly set in FUSE header. - In libperfuse, check for corner case where opc = 0 on INACTIVE and RECLAIM (how is it possible? Check for it to avoid a crash anyway) - In libperfuse, make sure we unlimit RLIMIT_AS and RLIMIT_DATA so that we do notrun out of memory because the kernel is lazy at reclaiming vnodes. - In libperfuse, cleanup style of perfuse_destroy_pn() Do not set PUFFS_KFLAG_CACHE_FS_TTL for PUFFS tests
1.163.2.1	03-Apr-2012	riz	Pull up following revision(s) (requested by jakllsch in ticket #154): sys/fs/puffs/puffs_vnops.c: revision 1.164 Prevent access beyond end of PUFFS file on read, similar to as is done for NFS.
1.174.2.3	03-Dec-2017	jdolecek	update from HEAD
1.174.2.2	20-Aug-2014	tls	Rebase to HEAD as of a few days ago.
1.174.2.1	20-Nov-2012	tls	Resync to 2012-11-19 00:00:00 UTC
1.176.2.1	18-May-2014	rmind	sync with head
1.181.2.1	10-Aug-2014	tls	Rebase.
1.182.2.13	27-Feb-2015	martin	Pull up following revision(s) (requested by manu in ticket #555): lib/libpuffs/puffs.3: revision 1.60 sys/fs/puffs/puffs_msgif.h: revision 1.84 lib/libperfuse/ops.c: revision 1.83 sys/fs/puffs/puffs_sys.h: revision 1.89 sys/fs/puffs/puffs_vfsops.c: revision 1.116 lib/libperfuse/perfuse.c: revision 1.36 sys/fs/puffs/puffs_vnops.c: revision 1.200-1.202 Add PUFFS_KFLAG_NOFLUSH_META to prevent sending metadata flush to FUSE FUSE filesystems do not expect to get metadata updates for [amc]time and size, they updates the value on their own after operations. The PUFFS PUFFS_KFLAG_NOFLUSH_META option prevents regular metadata cache flushes to the filesystem , and libperfuse uses it to match Linux FUSE behavior. While there, fix a bug in SETATTR: do not update kernel metadata cache from SETATTR reply when the request is asynchronous, as we do not have the reply yet. Update file size after write without metadata flush If we do not use metadata flush, we must make sure the size is updated in the filesystem after a write, otherwise the next GETATTR will get us a stale value and the file will be truncated.
1.182.2.12	17-Jan-2015	martin	Pull up following revision(s) (requested by manu in ticket #423): sys/fs/puffs/puffs_vnops.c: revision 1.199 Make sure reads on empty files reach PUFFS filesystems Sending a read through the page cache will get the operation short-circuited. This is a problem with some filesystems that expect to receive the read operation in order to update atime. We fix that by bypassing the page cache when reading a file wich a size known to be zero.
1.182.2.11	09-Nov-2014	msaitoh	Pull up following revision(s) (requested by manu in ticket #194): sys/fs/puffs/puffs_vnops.c: revision 1.197 sys/fs/puffs/puffs_node.c: revision 1.35 Fix PUFFS node use-after-reclaim When puffs_cookie2vnode() misses an entry, vcache_get() creates a new node (puffs_vfsop_loadvnode being called to initialize the PUFFS part), then it discovers it is VNON, and tries to vrele() it. vrele() calls VOP_INACTIVE(), which led us in puffs_vnop_inactive() where we sent a request to the filesystem for a node that already had been reclaimed. The fix is to check for VNON nodes in puffs_vnop_inactive() and to return without doing anyting. This is suboptimal, but a better workaround would probably need to modify vcache API, with an impact on other filesystems. Let us keep it simple.
1.182.2.10	09-Nov-2014	msaitoh	Pull up following revision(s) (requested by manu in ticket #193): sys/fs/puffs/puffs_vnops.c: revision 1.198 PUFFS direct I/O cache fix There are a few situations where we must take care of the cache if direct I/O was enabled: - if we do direct I/O for write but not for read, then any write must invalidate the cache so that a reader gets the written data and not the not-updated cache. - if we used a vnode without direct I/O and it is enabled for writing, we must flush the cache before compeling the open operation, so that the cachec write are not lost. And at inactive time, we wipe direct I/O flags so that a new open without direct I/O does not inherit direct I/O.
1.182.2.9	05-Nov-2014	snj	Pull up following revision(s) (requested by manu in ticket #182): sys/fs/puffs/puffs_vnops.c: revision 1.195 According to pooka@'s comment, a long time ago, VOP_STRATEGY could not fail without taking down the kernel. It seems this is not the case anymore, hence we can stop dropping errors in puffs_vnop_strategy() Approved by pooka@
1.182.2.8	05-Nov-2014	snj	Pull up following revision(s) (requested by manu in ticket #181): lib/libperfuse/fuse.h: revision 1.6 lib/libperfuse/ops.c: revision 1.78 lib/libperfuse/perfuse.c: revision 1.35 lib/libperfuse/perfuse_priv.h: revision 1.36 lib/libpuffs/dispatcher.c: revision 1.48 lib/libpuffs/opdump.c: revision 1.37 lib/libpuffs/puffs.c: revision 1.118 lib/libpuffs/puffs.h: revision 1.126 lib/libpuffs/puffs_ops.3: revisions 1.40-1.41 sys/fs/puffs/puffs_msgif.h: revision 1.82-1.83 sys/fs/puffs/puffs_msgif.h: revision 1.82 sys/fs/puffs/puffs_vnops.c: revision 1.196 Add PUFFS support for fallocate and fdiscard operations -- libpuffs support for fallocate and fdiscard operations -- Add PUFFS_HAVE_FALLOCATE in puffs_msgif.h so that filesystem can decide at build time wether fallocate is usable -- FUSE fallocate support There seems to be no fdiscard FUSE operation at the moment, hence that one is left unused.
1.182.2.7	14-Oct-2014	martin	Pull up revisions 1.192-1.194: fix debug printf formatting and make it compile without debugging enabled.
1.182.2.6	13-Oct-2014	martin	Pull up following revision(s) (requested by manu in ticket #136): sys/fs/puffs/puffs_vnops.c: revision 1.189-1.191 If we truncate a file open for writing, make sure we zero-fill the end of the last page, otherwise if the file is later truncated to a larger size (creating a hole), that area will not return zeroes as it should.
1.182.2.5	30-Sep-2014	martin	Pull up following revision(s) (requested by hannken in ticket #67): sys/fs/puffs/puffs_node.c: revision 1.34 sys/fs/puffs/puffs_vnops.c: revision 1.187 Fix the puffs_sop_thread -> puffs_cookie2vnode path: - pass the cookie by reference - add missing mutex_exit() - update assertion for VNON typed vnodes
1.182.2.4	11-Sep-2014	martin	Pull up following revision(s) (requested by manu in ticket #93): sys/fs/puffs/puffs_vnops.c: revision 1.186 PUFFS fixes for size update ater write plus read/write sanity checks - Always update kernel metadata cache for size when writing This fixes situation where size update after appending to a file lagged - Make read/write nilpotent when called with null size, as FFS does - Return EFBIG instead of EINVAL for negative offsets, as FFS does
1.182.2.3	10-Sep-2014	martin	Pull up following revision(s) (requested by manu in ticket #79): sys/fs/puffs/puffs_node.c: revision 1.33 sys/fs/puffs/puffs_vnops.c: revision 1.185 When changing a directory content, update the ctime/mtime in kernel cache, otherwise the updated ctime/mtime appears after the cached entry expire.
1.182.2.2	29-Aug-2014	martin	Pull up following revision(s) (requested by hannken in ticket #67): sys/fs/puffs/puffs_sys.h: revision 1.86 sys/fs/puffs/puffs_vfsops.c: revision 1.114 sys/fs/puffs/puffs_msgif.c: revision 1.95 sys/fs/puffs/puffs_node.c: revision 1.32 sys/fs/puffs/puffs_vnops.c: revision 1.184 Change puffs from hashlist to vcache. - field "pa_nhashbuckets" of struct "puffs_kargs" becomes a no-op. and should be removed on the next protocol version bump.
1.182.2.1	26-Aug-2014	riz	Pull up following revision(s) (requested by manu in ticket #52): sys/fs/puffs/puffs_msgif.h: revision 1.81 sys/fs/puffs/puffs_sys.h: revision 1.85 sys/fs/puffs/puffs_vnops.c: revision 1.183 Add a oflags input field to open requests so that the filesystem can pass back information about the file. Implement PUFFS_OPEN_IO_DIRECT, which will force direct IO (bypassing page cache) for the file.
1.198.2.5	28-Aug-2017	skrll	Sync with HEAD
1.198.2.4	05-Oct-2016	skrll	Sync with HEAD
1.198.2.3	09-Jul-2016	skrll	Sync with HEAD
1.198.2.2	06-Jun-2015	skrll	Sync with HEAD
1.198.2.1	06-Apr-2015	skrll	Sync with HEAD
1.204.2.2	26-Apr-2017	pgoyette	Sync with HEAD
1.204.2.1	26-Jul-2016	pgoyette	Sync with HEAD
1.205.2.1	21-Apr-2017	bouyer	Sync with HEAD
1.211.10.2	08-Apr-2020	martin	Merge changes from current as of 20200406
1.211.10.1	10-Jun-2019	christos	Sync with HEAD
1.211.8.1	26-Nov-2018	pgoyette	Sync with HEAD, resolve a couple of conflicts
1.211.2.1	06-Nov-2018	martin	Pull up following revision(s) (requested by manu in ticket #1082): sys/fs/puffs/puffs_vnops.c: revision 1.213 Fix use after RECLAIM in PUFFS filesystems From hannken@ When puffs_cookie2vnode() misses an entry and vrele() it operations puffs_vnop_reclaim() and puffs_vnop_fsync() get called with a VNON vnode. Do not notify the server in this case as the cookie is stale.
1.213.6.1	29-Feb-2020	ad	Sync with head.
1.214.4.1	25-Apr-2020	bouyer	Sync with bouyer-xenpvh-base2 (HEAD)
1.217.6.1	01-Aug-2021	thorpej	Sync with HEAD.

OpenGrok