libkvm support for the new m68k pmap.

introduce a couple of new turn-off-gcc-warning variables and use them.

GCC 14 has a new annoying calloc() checker that we turn off in a bunch
of places, and there are a few more dangling-pointer issuse that come up,
but seem bogus.

Revert CC_WNO_USE_AFTER_FREE from Makefile's (thanks uwe@)

Sprinkle CC_WNO_USE_AFTER_FREE for GCC 12

All of them are blamed for idiom equivalent to:
newbuf = realloc(buf, size);
p = newbuf + (p - buf);

libkvm: suppress lint warnings about extern declarations for now

This should be properly fixed by someone who is familiar with libkvm, as
it affects <sys/cpu_data.h> as well.

mips needs _KMEMUSER for label_t in pcb.h

MKCOMPAT fixes for when compat MACHINE_CPU != normal MACHINE_CPU

branches: 1.49.2;
These directories don't handle WARNS=5

branches: 1.48.2;
add a function to report the name of the file kvm_open{,.files} is using.

Add -D_KMEMUSER to CPPFLAGS

branches: 1.46.2;
Import PAE support for kvm(3):
- add kvm_i386pae.c (used for PAE memory translations), and update Makefile
for libkvm build.
- in pdppaddr: pass a flag to indicate PAE mode. Use a bit ignored
by the MMU. Mask address with PG_FRAME to avoid side effects.

Tested with vmstat(1)/netstat(1) to debug core files of PAE and !PAE
kernels. Older kernel dumps will default to native i386 (!PAE) mode.

XXX Currently, savecore(8) will fail to dump a PAE kernel in a !PAE
environment (and reciprocally). So you need to sync and reboot
with a kernel of the same mode as the one that crashed. Once the dump
is successful, this does not matter anymore.

look for ${KVM_MACHINE_ARCH} before ${MACHINE_ARCH} and ${MACHINE_CPU}

Add new Makefile knob, USE_FORT, which extends USE_SSP by turning on the
FORTIFY_SOURCE feature of libssp, thus checking the size of arguments to
various string and memory copy and set functions (as well as a few system
calls and other miscellany) where known at function entry. RedHat has
evidently built all "core system packages" with this option for some time.

This option should be used at the top of Makefiles (or Makefile.inc where
this is used for subdirectories) but after any setting of LIB.

This is only useful for userland code, and cannot be used in libc or in
any code which includes the libc internals, because it overrides certain
libc functions with macros. Some effort has been made to make USE_FORT=yes
work correctly for a full-system build by having the bsd.sys.mk logic
disable the feature where it should not be used (libc, libssp iteself,
the kernel) but no attempt has been made to build the entire system with
USE_FORT and doing so will doubtless expose numerous bugs and misfeatures.

Adjust the system build so that all programs and libraries that are setuid,
directly handle network data (including serial comm data), perform
authentication, or appear likely to have (or have a history of having)
data-driven bugs (e.g. file(1)) are built with USE_FORT=yes by default,
with the exception of libc, which cannot use USE_FORT and thus uses
only USE_SSP by default. Tested on i386 with no ill results; USE_FORT=no
per-directory or in a system build will disable if desired.

Add ${NETBSDSRCDIR}/sys to the include path.

do the exit on lint warnings only on i386 like libc, to avoid the wrath of
the sloppy archs.

exit on lint warnings.

Add workaround for building with ${MACHINE}==sparc64, ${MACHINE_ARCH}==sparc;
symlink the appropriate machine & sparc include directories into the objdir.

XXX: this is necessary until (when?) the sparc Arch Dependent headers
(used by MACHINE_ARCH==sparc) are separated from the sparc Machine
Dependent headeres (used by MACHINE==sparc but not MACHINE==sparc64).

Merge the nathanw_sa branch.

Explicitly move setting of NOxxx and USE_SHLIBDIR to the top of the
Makefile (before including <bsd.own.mk>)

branches: 1.37.2;
* Add user-controlled mk.conf variables
- SHLIBDIR Location to install shared libraries if ${USE_SHLIBDIR}
is "yes". Defaults to "/usr/lib".

- USE_SHLIBDIR If "yes", install shared libraries in ${SHLIBDIR}
instead of ${LIBDIR}. Defaults to "no".
Sets ${_LIBSODIR} to the appropriate value.
This may be set by individual Makefiles as well.

- SHLINKDIR Location of shared linker. Defaults to "/usr/libexec".
If != "/usr/libexec", change the dynamic-linker
encoded in shared programs

* Set USE_SHLIBDIR for libraries used by /bin and /sbin:
libc libcrypt libcrypto libedit libipsec libkvm libm libmi387
libtermcap libutil libz

* If ${_LIBSODIR} != ${LIBDIR}, add symlinks from ${LIBDIR}/${LIB}.so*
to ${_LIBSODIR}/${LIB}.so* for compatibility.

* Always install /sbin/init statically (for now)


The net effect of these changes depends on how the variables are set:

1.) If nothing is set or changed, there is no change from the
current behaviour:
- Static /bin, /sbin, and bits of /usr/*
- Dynamic rest
- Shared linker is /usr/libexec/ld*so

2.) If the following make variables are set:
LDSTATIC=
SHLINKDIR=/lib
SHLIBDIR=/lib
Then the behaviour becomes:
- Dynamic tools
- .so libraries used by /bin and /sbin are installed to /lib,
with symlinks from /usr/lib/lib*so to -> /lib/lib*so
where appropriate
- Shared linker is /lib/ld*so

3.) As per 2.), but add the following variable:
USE_SHLIBDIR=yes
This forces all .so's to be instaleld in /lib (with compat
symlinks), not just those tagged by their Makefiles to be.
Again, compat symlinks are installed

Added support for sun2 kernel cores.

Rename kvm_arm32.c to kvm_arm.c, since it's shared between arm32 and arm26.
The Makefile now finds the correct file without help.

use MACIHNE_CPU to determine architecture-dependent source code directory.
is use of exists(foo) safe enough?
TODO: many other places to fix

sh3 split

Document KVM_NO_FILES and kvm_get{proc,argv,envv}2(). Better wording
kindly accepted!

XXX: Need to update distrib lists. Tomorrow...

branches: 1.31.2;
Initial commit of arm26 port

the PMAP_NEW option is gone as the new interface is no longer optional.

completely remove Mach VM support. all that is left is the all the
header files as UVM still uses (most of) these.

make this build again after the mips->mipse? transition

NetBSD/mips uses new libkvm now.

Sigh, typo.

Simplify test ... MACHINE_ARCH == "mips"

add -DPMAP_NEW if PMAP_NEW is on in mk.conf

Rewrite _kvm_kvatop() to work with the new crash dump format, and make
the i386 port use libkvm, not libkvm.old.

oops, forgot to add support for UVM here.

use CPPFLAGS instead of CFLAGS

branches: 1.20.2;
Oops, fixed edit error at previous commit.

NetBSD/x68k now use libkvm instead of libkvm.old

Switch from libkvm.old to libkvm for the arm32 port.

Use bsd.prog.mk and not bsd.{own,man}.mk

Add bsd.own.mk for bsd.man.mk.

Don't include bsd.lib.mk if you don't build a library, just use bsd.man.mk
to build the manpages.

move man pages into share/man.

typo

Build the same libkvm.* on all m68k machines. (Well... except for the
stragglers that still use libkvm.old ) The m68k machines now dispatch
based on the machine name to the appropriate libkvm implementation.

RCS id police.

Add a manual page for the kvm_dump_* functions.

specify man pages the new way.

Merge the m68k files. The code itself should be merged.

Add sun3 support (only "live" kernels for now).

branches: 1.6.2;
add support for current m68k based pmaps till they can be "fixed"

m68k-ify the hp300 kvm file; use MACHINE_ARCH to pick files up

clean up import

Add mac68k to amiga in looking for cpu040.

temporary additional lookup of cpu040 for amiga's until new kvm stuff or new
amiga 040 VM stuff.

branches: 1.1.1;
create this in preparation for later.

branches: 1.1.2;
file Version.map was initially added on branch agc-symver.

Sort sections.

add a function to report the name of the file kvm_open{,.files} is using.

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22280, verified by myself.

Consistently spell "crash dump" as two separate words. From jmc@openbsd.

Merge the nathanw_sa branch.

New sentence, new line. From Robert Elz.

Use Nm instead of Xrefing ourselves.

add FILES section to refer to /dev/mem.
from openbsd.

branches: 1.5.10;
More trailing white space.

macroize BSD, NetBSD, FreeBSD and misc cleanup

use -mdoc .Pa macro for filenames

RCS id police.

branches: 1.1.1;
Initial revision

libkvm: whitespace

Get rid of usrstack/USRSTACK. Document that the old version of getargv is
broken because of ASLR.

Cast off_t to size_t to appease lint:
warning: conversion from 'long long' to 'int' may lose accuracy

Don't treat failure to map the file as fatal.

make kvm_pread use the mapped buffer if it is available.

Implement kvm_write() for dump files. We map the file privately so that
changes don't get propagated back, and then write to the private buffer.
This is not useful yes, but it is simple enough to make pread use the buffer
if it is available.

only do bounce buffering for character devices.

branches: 1.104.2;
initialise the kvm errbuf so that if an error occurs, but the
internal _kvm_*err() functions aren't called, stack garbage is
not reported as an error string.

call _kvm_err() if _kvm_kvatop() fails. now the above is not
triggered anymore :-)

branches: 1.103.2; 1.103.4;
Keep /dev/ksyms open in _kvm_open(). This way /dev/ksyms can be put into
$g_kmem without breaking the tools that need kmem+ksyms.

Discussed on tech-kern@ three weeks ago. The original issue was reported
by maya@, the patch was written by Tom Ivar Helbekkmo, ok christos@.

distinguish between partial and failed read, so that "read 0xffffffff bytes failed"
is not output..

Remove the #include <sys/user.h> from all of libkvm.
sys/user.h is a stub that just #includes sys/pcb.h.
There are no 'struct pcb' anywhere in here, so I'm extremely doubtful
any of the builds will fail.
OTOH it might be relying on a header that pcb.h includes.
In any case i386 and amd64 build.

branches: 1.100.2;
Do not fail kvm_dump_mkheader without setting an error message

branches: 1.99.2;
kill open_cloexec()

add a function to report the name of the file kvm_open{,.files} is using.

Return error on failure instead of hiding it and always returning zero.
From Henning Petersen in PR 44152.

Do not return -1 on _kvm_read when all the requested data could be read (fix bin/38648)

Change kvm_pa2off() and kvm_kvatop() prototypes (private to kvm(3)):

-int _kvm_kvatop(kvm_t *, u_long, u_long *);
-off_t _kvm_pa2off(kvm_t *, u_long);
+int _kvm_kvatop(kvm_t *, vaddr_t, paddr_t *);
+off_t _kvm_pa2off(kvm_t *, paddr_t);

Basically, use vaddr_t for VA and paddr_t for PA. In addition, for variables
representing addresses, use paddr_t or vaddr_t, depending on the context.

For most arches, vaddr_t and paddr_t are equivalent to unsigned long. However,
the change was needed for exotic situations, like i386 PAE, were unsigned long
is not suitable for PA which are 64 bits long. As this required a complete
change of the function prototypes, all arches had to be adapted accordingly.

Core files from before this commit should still work with the new code; I did
not see any direct dependency between core's structure and kvatop/pa2off.

The change was compile tested for all arches, as it impacts all of them.

See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

Move the code that tries to open _PATH_KSYMS first into a common path,
so that it is used both for the /dev/mem case and the core dump case.

Output from savecore(8) before:
savecore - - - (null): kvm_openfiles: /netbsd: No such file or directory
and after:
savecore - - - no core dump

Anoter instance of <sys/exec_aout.h> -> <sys/exec.h>.

... and make kvm_dump_inval() work on the raw device.

Handle reading from raw disk devices.

Fix a typo in the last change that broke a comparision.

Add kvm_dump_header and implement kvm_dump_wrtheader on top of it.
This function allows the caller finer control of how the writes happen
and doesn't force stdio as interface. Optimise clear_gap a bit to not
fill the gap byte-wise. Bump minor version of libkvm.

ANSIfy.

branches: 1.87.10;
#include a necessary header directly, rather than via user.h.

Coverity CID 881: Always print an error if allocation failed.
If the program is not specified use getprogname()

1. Eliminate some unnecessary to kvm_{m,re}alloc.
2. Don't malloc/free procbase/procbase2/lwpbase continuously. Keep track
of the size, and only do it if necessary.
3. Write a macro to malloc/realloc and set the size of members so that it
is done correctly. Previous open coded version in kvm_file.c always
set the length, which is incorrect.
4. Remove bogus check against INT_MAX.
5. use NULL to initialize pointers instead of 0.

as swap is now optional, ENXIO on /dev/drum is normal.

Uppercase CPU, plural is CPUs.

When using /dev/ksyms, don't keep it open more than strictly necessary.
This allows LKM manipulation while an application using libkvm, such as
vmstat or gkrellm, is running.

Patch reviewed by Anders Magnusson.

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22280, verified by myself.

Make the kvm routines use /dev/ksyms to get the kernel namelist.
If it fails, use /netbsd instead.

Merge the nathanw_sa branch.

use strlcpy to ensure string termination

Use FD_CLOEXEC instead of `1' and don't cast.

Default using the value of machdep.booted_kernel from sysctl() if it
is available and points to a file. The value of "/netbsd" is still
the ultimate fallback.

another SETFL->SETFD

actually do the close-on-exec

Set the close-on-exec bit on all file descriptors we open. Inspired by a
FreeBSD security advisory.

Reviewed by thorpej

Don't omit third argument to open(2).
PR lib/15752 from Stephen Ma.

__STDC__ is always defined on NetBSD.

branches: 1.70.2; 1.70.4;
Give initiali[sz]e all the "i"s it deserves.

Add forgotten initialization.
Fixed PR#11446.

format-string audit cleanups

<vm/vm.h> -> <uvm/uvm_extern.h>

remove redundant vm includes

branches: 1.65.4;
Add kvm interface to the new sysctls:
kvm_getproc2() -> sysctl(KERN_PROC2)
kvm_getargv2() -> sysctl(KERN_PROC_ARGS, KERN_PROC_ARGV)
kvm_getenvv2() -> sysctl(KERN_PROC_ARGS, KERN_PROC_ENV)
Add new KVM_NO_FILES flag to kvm_open*() - set up enough state to
use the above calls without having to open any kernel files.

XXX: kvm_getprocs.3 and kvm_open.3 to be updated soon.

More trailing white space.

branches: 1.63.2;
Nuke swap_pager.h.

Remove lint

bzero->memset, bcopy->memcpy, bcmp->memcmp

Remove some lint.

const poisoning.

Oops, missed a couple of pread(2) uses.

Use pread(2) and pwrite(2) rather than lseek(2)/{read,write}(2).

fix return value of kvm_read() and kvm_write()

remove obsolete register declarations

clean up WARNS on the alpha.

use <sys/cdefs.h> __RCSID() macro

Call _kvm_mdopen(kd) to do machine-dependent initialization.

Add missing argument to fprintf().

- convert unsafe strcpy(), strcat() and sprintf() to the `n' versions.
- some KNF.

Check return value of _kvm_initvtop().

Don't use a kernel database based on the file name, use only
/var/db/kvm.db, and only use that if no kernel namelist file
is named. Update the documentation to describe this.

some spaces -> tabs

add code so that if kvm_nlist() failed because __fdnlist() failed,
kvm_geterr() actually provides useful information rather than just an
empty string.

Fix the empty error message problem more completely. When the magic-check
fails on the cpu_hdr, return 0 instead of -1. This allows the caller to
distinguise between fatal errors and 'nothing there'.

Fix the mysterious empty error message when savecore is executed and no
dump is present. This was caused by the fact that kvm_dump_mkheader() was
called *before* savecore checks the dump magic and kvm_dump_mkheader() returned
-1 without setting an error message. The latter is fixed now.

Allow more flexibility in the format of cpu_kcore_hdr_t and isolate
the knowledge of that struct in the machine-dependent module.

RCS id police.

Nuke the kd_live argument to kvm_dump_mkheader(). it's not needed anymore...

New libkvm/savecore implementation. With the following differences:
1) savecore will not access the dump or live-kernel directly. It
will always use the kvm-functions. Allowing it to work on kernels
that don't have a 1-1 PA-VA mapping.
2) the kvm-lib has some additional functions to accomplish 1
- kvm_dump_mkheader()
- kvm_dump_wrtheader()
- kvm_dump_inval()
3) the file formats of the dump generated by the kernel and the dump
generated by savecore have been changed. The file format now looks
like the format produced for 'normal' core dumps.
Ports not yet supporting the new kvm-format will be using libkvm.old/
savecore.old for the time being.

slight type cleanup. add some prototypes for internal functions,
some 'const' poisoning, some lvalue lossage.

Use the new ps_strings format. Find the arguments using their
addresses in the argv array, rather than guessing based on the location
of NULs.

Eliminate _kvm_uvatop(), in favor of using the machine-independent VM structures.

Move the swap page buffer into struct __kvm. Cache the page size during
kvm_open(). Fix an existing memory leak.

clean up import

field name change, and safety (minor)

update for changed flags

Add mac68k to amiga in looking for cpu040.

kill end-check bug.

queue foo

quiet compiler, too

bad cast, pointed out by Rafal Boni

Adding a ns32k definition for X_DEADKERNEL.

Sparc specific memory translations.
Some reorganization of the kernel nlist array.

temporary additional lookup of cpu040 for amiga's until new kvm stuff or new
amiga 040 VM stuff.

Always read CLBYTES chunks, and cache the last one read. Don't allocate two
large arrays statically.

get rid of the awful RSS hack

P[GD]_SHIFT --> P[GD]SHIFT

look for m68k, rather than hp300 or amiga

Forgot to initialize `cp' in setsyserr().

Remove code for old VM system. Don't use page tables in kvm_procread().
Change `#if DEBUG' to `#ifdef DEBUG'. Read some symbols from the kernel
only once.

Last change was almost correct, but we have to check that the pager actually
has the page. If not, we go up the shadow chain.

Fix an off-by-one error when scanning the vm map for a page. Also, if the
page is not in core, don't recurse into shadow objects if we've already found
a pager.

Fix a bug in virtual to physical address translation on hp300s, and add support
for Amigas.

line 860: use CLSHIFT rather than PGSHIFT.
line 927: fix bogus indentation
also make sure first chars of buffers are null char, just in
case nothing rational is read...

implement kvm_procread for the hp300. now things mostly work on
the hp300, in terms of kvm. still minor strangeness re: sendmail...

#include <machine/pte.h>, not <hp300/hp300/pte.h>. Also, make it at least
compile for non-i386 architectures, but exit with an error message (for now).

Add RCS identifiers.

branches: 1.11.2;
Read swap in CLBYTES blocks.

kernel string read function takes much less time. could be optimized further,
by making block reads kernel page aligned. (cgd will make the next optimization
in this file)

misc minor changes

update kvm.c for the new way of doing things:
(1) add support for new stack format/PS_STRINGS
(2) add support for kvm_getenv
the HP300 support in kvm_procread needs to be looked at/filled in,
from example code in kvm_getu, plus hacks

don't inlude select.h, but keep header cleanups

fix new select stuff

fix problem where kvm_getprocs(), ..., kvm_freeprocs(), kvm_getprocs()
would get no processes.

Fix memory leak.

From: pk@cs.few.eur.nl (Paul Kranenburg)

patch which allows ps(1) and w(1) to lay
their hands on process command line arguments.

after 0.2.2 "stable" patches applied

branches: 1.1.1;
Initial revision

libkvm: whitespace

Get rid of usrstack/USRSTACK. Document that the old version of getargv is
broken because of ASLR.

Catch up with aa64 is not mips change

Need register defines that were moved to armreg.h

be consistent about uint64_t and bit masks.

use uint64_t for idx_mask. fixes lookups for some PAs. from skrll.

bump copyright. we updated this a lot of this in 2018.

Fix _kvm_kvatop so that it now works.

tg1 is a 64 bit value, store it in a 64 bit variable.

compare matching TG1 bits.

aarch64 crash dump kvtop now less broken, but still not stable.

branches: 1.3.2;
Add initial support for ARMv8 (AARCH64) (by nisimura@ and ryo@)

- sys/arch/evbarm64 is gone and integrated into sys/arch/evbarm. (by skrll@)
- add support fdt. evbarm/conf/GENERIC64 fdt (bcm2837,sunxi,tegra) based generic 64bit kernel config. (by skrll@, jmcneill@)

branches: 1.2.4;
Stop including <sys/user.h> in userland libraries

This header in this context is freebsdism.

branches: 1.1.4;
Preliminary files for AARCH64 (64-bit ARM) support.
Enough for a distribution build.

libkvm: whitespace

Get rid of usrstack/USRSTACK. Document that the old version of getargv is
broken because of ASLR.

Remove the #include <sys/user.h> from all of libkvm.
sys/user.h is a stub that just #includes sys/pcb.h.
There are no 'struct pcb' anywhere in here, so I'm extremely doubtful
any of the builds will fail.
OTOH it might be relying on a header that pcb.h includes.
In any case i386 and amd64 build.

Add missing __RCSID("$NetBSD$");

branches: 1.25.6; 1.25.12;
Change kvm_pa2off() and kvm_kvatop() prototypes (private to kvm(3)):

-int _kvm_kvatop(kvm_t *, u_long, u_long *);
-off_t _kvm_pa2off(kvm_t *, u_long);
+int _kvm_kvatop(kvm_t *, vaddr_t, paddr_t *);
+off_t _kvm_pa2off(kvm_t *, paddr_t);

Basically, use vaddr_t for VA and paddr_t for PA. In addition, for variables
representing addresses, use paddr_t or vaddr_t, depending on the context.

For most arches, vaddr_t and paddr_t are equivalent to unsigned long. However,
the change was needed for exotic situations, like i386 PAE, were unsigned long
is not suitable for PA which are 64 bits long. As this required a complete
change of the function prototypes, all arches had to be adapted accordingly.

Core files from before this commit should still work with the new code; I did
not see any direct dependency between core's structure and kvatop/pa2off.

The change was compile tested for all arches, as it impacts all of them.

See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

Ansify and KNF all functions within kvm(3). No objection on current-users@.

Used as ground for u_long => vaddr_t/paddr_t replacements in kvm(3)
private functions.

Compile tested for all arches. See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

Handle reading from raw disk devices.

branches: 1.22.22;
Consistently spell "crash dump" as two separate words. From jmc@openbsd.

Just use PAGE_SHIFT from <machine/vmparam.h> instead of going through
hoops to read and cache the value from the crash dump header.

branches: 1.20.2;
Include <machine/vmparam.h> for VM_MIN_*...

Don't include <machine/pmap.h> and <machine/vmparam.h> if _KERNEL isn't
defined. Include them explicitly in the few kvm_arch.c that need them.

<vm/vm.h> -> <uvm/uvm_extern.h>

remove redundant vm includes

fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well. No functional changes.

branches: 1.15.6;
More trailing white space.

vm_size_t -> vaddr_t, thanks Matt Thomas for pointing it out.

Remove some lint.

Use pread(2) and pwrite(2) rather than lseek(2)/{read,write}(2).

Don't rely on a constant page shift value; dynamically compute it based
on the page size in the crash dump header.

Update for changes to <machine/pte.h>

adjust changes in alpha's kcore.h.

--Fix k1seg translation, which has been broken since
--10/10/97 when a WARNS "fix" was put in that broke
--kernel core dump stack backtrace and other
--things that go thru libkvm.

branches: 1.7.2;
clean up WARNS on the alpha.

Add _kvm_mdopen()

add support for real KV addresses, and clean up some.

Real alpha support, minus actual KV -> physical translation. (Does K0SEG ->
physical translation, though, which is enough for e.g. dmesg and vmstat -i.)

stub _kvm_pa2off so that new libkvm links

reminder to not forget k0seg translations. (nothing done here yet, though.)

stubs.

Get rid of usrstack/USRSTACK. Document that the old version of getargv is
broken because of ASLR.

Change kvm_pa2off() and kvm_kvatop() prototypes (private to kvm(3)):

-int _kvm_kvatop(kvm_t *, u_long, u_long *);
-off_t _kvm_pa2off(kvm_t *, u_long);
+int _kvm_kvatop(kvm_t *, vaddr_t, paddr_t *);
+off_t _kvm_pa2off(kvm_t *, paddr_t);

Basically, use vaddr_t for VA and paddr_t for PA. In addition, for variables
representing addresses, use paddr_t or vaddr_t, depending on the context.

For most arches, vaddr_t and paddr_t are equivalent to unsigned long. However,
the change was needed for exotic situations, like i386 PAE, were unsigned long
is not suitable for PA which are 64 bits long. As this required a complete
change of the function prototypes, all arches had to be adapted accordingly.

Core files from before this commit should still work with the new code; I did
not see any direct dependency between core's structure and kvatop/pa2off.

The change was compile tested for all arches, as it impacts all of them.

See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

Ansify and KNF all functions within kvm(3). No objection on current-users@.

Used as ground for u_long => vaddr_t/paddr_t replacements in kvm(3)
private functions.

Compile tested for all arches. See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

Handle reading from raw disk devices.

Update libkvm to be able to deal with arm kernel core dumps.

These changes allow savecore to successfully save a kernel core dump.

Final part of fixing PR cats/18026

branches: 1.2.2; 1.2.26;
Steal the variable VM_* stuff from m68k.

Rename kvm_arm32.c to kvm_arm.c, since it's shared between arm32 and arm26.
The Makefile now finds the correct file without help.

Rename kvm_arm32.c to kvm_arm.c, since it's shared between arm32 and arm26.
The Makefile now finds the correct file without help.

<vm/vm.h> -> <uvm/uvm_extern.h>

Include stdlib.h for prototype of free().

New version for libkvm rather than libkvm.old. Based on the kvm_powerpc.c
this is mainly stubs at the moment.

use <sys/cdefs.h> __RCSID() macro

Add _kvm_mdopen()

RCS id police.

Initial commit of the libkvm code for the arm32 port.

Remove 3rd and 4th clause on Leo Weppelman's license. OK leo@.

Remove redundant .br.

branches: 1.13.40;
Consistently spell "crash dump" as two separate words. From jmc@openbsd.

Use
.In header.h
instead of
.Fd #include \*[Lt]header.h\*[Gt]
Much easier to read and write, and supported by groff for ages.
Okayed by ross.

New sentence, new line. From Robert Elz.

branches: 1.10.2;
Generate <>& symbolically.

Sort sections.

Spell occurred correctly.

xref kvm_open.

More trailing white space.

there probably are bugs. however, don't make an issue of it

add LIBRARY section to man page

NetBSD 1.1A -> .Nx 1.2

Fix the empty error message problem more completely. When the magic-check
fails on the cpu_hdr, return 0 instead of -1. This allows the caller to
distinguise between fatal errors and 'nothing there'.

Add a manual page for the kvm_dump_* functions.

Remove the #include <sys/user.h> from all of libkvm.
sys/user.h is a stub that just #includes sys/pcb.h.
There are no 'struct pcb' anywhere in here, so I'm extremely doubtful
any of the builds will fail.
OTOH it might be relying on a header that pcb.h includes.
In any case i386 and amd64 build.

branches: 1.28.6; 1.28.12;
Define KREAD() inside kvm_private.h, for reusability.

Ansify and KNF all functions within kvm(3). No objection on current-users@.

Used as ground for u_long => vaddr_t/paddr_t replacements in kvm(3)
private functions.

Compile tested for all arches. See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

don't need <sys/tty.h> here.

+#define _KERNEL
+#include <sys/types.h>
+#undef _KERNEL

branches: 1.24.10; 1.24.16;
1. Eliminate some unnecessary to kvm_{m,re}alloc.
2. Don't malloc/free procbase/procbase2/lwpbase continuously. Keep track
of the size, and only do it if necessary.
3. Write a macro to malloc/realloc and set the size of members so that it
is done correctly. Previous open coded version in kvm_file.c always
set the length, which is incorrect.
4. Remove bogus check against INT_MAX.
5. use NULL to initialize pointers instead of 0.

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22280, verified by myself.

undo mallocvar.h change!

Include "sys/mallocvar.h" with "_KERNEL" defined in time to avoid
build failure because MALLOC_DECLARE() is not defined.

Fix build problem by rearranging the order of "#include" statements so
that "sys/mallocvar.h" gets include while "_KERNEL" is defined.

Merge the nathanw_sa branch.

branches: 1.18.2;
WARNS=2 fixes

<vm/vm.h> -> <uvm/uvm_extern.h>

remove redundant vm includes

Add kvm interface to the new sysctls:
kvm_getproc2() -> sysctl(KERN_PROC2)
kvm_getargv2() -> sysctl(KERN_PROC_ARGS, KERN_PROC_ARGV)
kvm_getenvv2() -> sysctl(KERN_PROC_ARGS, KERN_PROC_ENV)
Add new KVM_NO_FILES flag to kvm_open*() - set up enough state to
use the above calls without having to open any kernel files.

XXX: kvm_getprocs.3 and kvm_open.3 to be updated soon.

fix four major and one minor bug in dead kernel file-getting. Three
of the biggies crept in in rev. 1.10 (we need a lint trap!), the
others were here since the file was originally imported.

More trailing white space.

branches: 1.12.2;
Nuke swap_pager.h.

Need string.h.

Remove lint

remove obsolete register declarations

Fix compiler warnings.

use <sys/cdefs.h> __RCSID() macro
GC unused variables and fix error message typo in kvm_deadfiles()

add explicit return type for kvm_deadfiles()

RCS id police.

KERNEL -> _KERNEL

fix type of variables, add headers, protos, where necessary.

Update to match include files.

branches: 1.1.1;
Initial revision

Remove redundant .br.

branches: 1.8.40;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22280, verified by myself.

Use
.In header.h
instead of
.Fd #include \*[Lt]header.h\*[Gt]
Much easier to read and write, and supported by groff for ages.
Okayed by ross.

branches: 1.6.2;
Generate <>& symbolically.

Sort sections.

More trailing white space.

add LIBRARY section to man page

RCS id police.

branches: 1.1.1;
Initial revision

Remove redundant .br.

branches: 1.10.40;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22280, verified by myself.

Use
.In header.h
instead of
.Fd #include \*[Lt]header.h\*[Gt]
Much easier to read and write, and supported by groff for ages.
Okayed by ross.

New sentence, new line. From Robert Elz.

branches: 1.7.2;
Generate <>& symbolically.

Sort sections and SEE ALSO, fix punctuation nits.

add LIBRARY section to man page

use ".Dv NULL" rather than just "NULL", like everyplace else

RCS id police.

KERNEL -> _KERNEL

branches: 1.1.1;
Initial revision

Fix Xr.

doc for kvm_getkernelname

Reword post mortem description. It tries to get the tree averages if it can.

Clarify the difference to getloadavg(3).

branches: 1.11.44; 1.11.46;
Remove redundant .br.

branches: 1.10.40;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22280, verified by myself.

Use
.In header.h
instead of
.Fd #include \*[Lt]header.h\*[Gt]
Much easier to read and write, and supported by groff for ages.
Okayed by ross.

change DIAGNOSTICS to RETURN VALUES
since it describes return values.
from openbsd.

branches: 1.7.2;
Generate <>& symbolically.

Sort SEE ALSO.

xref kvm_open.

add LIBRARY section to man page

fix description; PR misc/3077.
also added xref to sysctl(3), which may be preferable.

RCS id police.

branches: 1.1.1;
Initial revision

Use C89 function definition

branches: 1.10.6; 1.10.8;
Define KREAD() inside kvm_private.h, for reusability.

Ansify and KNF all functions within kvm(3). No objection on current-users@.

Used as ground for u_long => vaddr_t/paddr_t replacements in kvm(3)
private functions.

Compile tested for all arches. See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22280, verified by myself.

Merge the nathanw_sa branch.

branches: 1.6.2;
use <uvm/uvm_param.h> now.

Remove lint

use <sys/cdefs.h> __RCSID() macro

include <stdlib.h> for getloadavg() prototype

RCS id police.

branches: 1.1.1;
Initial revision

Correct the signature of kvm_getlwps(3) in the man-page.

The elemsize argument is of type "size_t", not "int *".

Sponsored by <The NetBSD Foundation>

branches: 1.7.8;
Update synopses to match u_int/u_long/etc. changes. Do not bump date as
it's not a substantive change.

Remove redundant .br.

branches: 1.5.8;
Convert TNF licenses to new 2 clause variant

branches: 1.4.30;
List sys/param.h as a prerequisite of sys/sysctl.h. Bump date.

Use
.In header.h
instead of
.Fd #include \*[Lt]header.h\*[Gt]
Much easier to read and write, and supported by groff for ages.
Okayed by ross.

Merge the nathanw_sa branch.

branches: 1.1.2;
file kvm_getlwps.3 was initially added on branch nathanw_sa.

Correct the signature of kvm_getproc2(3) in the man-page.

The elemsize argument is of type "size_t", not "int".

Sponsored by <The NetBSD Foundation>

Add rest of the functions from the SYNOPSIS to the NAME section

branches: 1.15.38;
Remove redundant .br.

branches: 1.14.40;
List sys/param.h as a prerequisite of sys/sysctl.h. Bump date.

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22280, verified by myself.

Use
.In header.h
instead of
.Fd #include \*[Lt]header.h\*[Gt]
Much easier to read and write, and supported by groff for ages.
Okayed by ross.

session, not sessiod. By Adrian Mrva.

New sentence, new line. From Robert Elz.

branches: 1.9.2;
Generate <>& symbolically.

Sort sections, fix whitespace and punctuation nits.

Document KVM_NO_FILES and kvm_get{proc,argv,envv}2(). Better wording
kindly accepted!

XXX: Need to update distrib lists. Tomorrow...

branches: 1.6.2;
Document KERN_PROC_TTY_NODEV and KERN_PROC_TTY_REVOKE.

More trailing white space.

add LIBRARY section to man page

update to reflect reality.

RCS id police.

branches: 1.1.1;
Initial revision

Clean up deleted files.

new libkvm

Get rid of usrstack/USRSTACK. Document that the old version of getargv is
broken because of ASLR.

Remove the #include <sys/user.h> from all of libkvm.
sys/user.h is a stub that just #includes sys/pcb.h.
There are no 'struct pcb' anywhere in here, so I'm extremely doubtful
any of the builds will fail.
OTOH it might be relying on a header that pcb.h includes.
In any case i386 and amd64 build.

branches: 1.6.6; 1.6.12;
Change kvm_pa2off() and kvm_kvatop() prototypes (private to kvm(3)):

-int _kvm_kvatop(kvm_t *, u_long, u_long *);
-off_t _kvm_pa2off(kvm_t *, u_long);
+int _kvm_kvatop(kvm_t *, vaddr_t, paddr_t *);
+off_t _kvm_pa2off(kvm_t *, paddr_t);

Basically, use vaddr_t for VA and paddr_t for PA. In addition, for variables
representing addresses, use paddr_t or vaddr_t, depending on the context.

For most arches, vaddr_t and paddr_t are equivalent to unsigned long. However,
the change was needed for exotic situations, like i386 PAE, were unsigned long
is not suitable for PA which are 64 bits long. As this required a complete
change of the function prototypes, all arches had to be adapted accordingly.

Core files from before this commit should still work with the new code; I did
not see any direct dependency between core's structure and kvatop/pa2off.

The change was compile tested for all arches, as it impacts all of them.

See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

More KNF

ANSIfy / KNF.

Handle reading from raw disk devices.

branches: 1.2.22;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22280, verified by myself.

branches: 1.1.2;
Added hppa support (some of it incomplete) to lib/csu, lib/libc,
and lib/libkvm.

Get rid of usrstack/USRSTACK. Document that the old version of getargv is
broken because of ASLR.

Switch to the new PTE naming. The old naming is now unused, remove it.

Remove the #include <sys/user.h> from all of libkvm.
sys/user.h is a stub that just #includes sys/pcb.h.
There are no 'struct pcb' anywhere in here, so I'm extremely doubtful
any of the builds will fail.
OTOH it might be relying on a header that pcb.h includes.
In any case i386 and amd64 build.

branches: 1.29.6; 1.29.12;
Import PAE support for kvm(3):
- add kvm_i386pae.c (used for PAE memory translations), and update Makefile
for libkvm build.
- in pdppaddr: pass a flag to indicate PAE mode. Use a bit ignored
by the MMU. Mask address with PG_FRAME to avoid side effects.

Tested with vmstat(1)/netstat(1) to debug core files of PAE and !PAE
kernels. Older kernel dumps will default to native i386 (!PAE) mode.

XXX Currently, savecore(8) will fail to dump a PAE kernel in a !PAE
environment (and reciprocally). So you need to sync and reboot
with a kernel of the same mode as the one that crashed. Once the dump
is successful, this does not matter anymore.

Change kvm_pa2off() and kvm_kvatop() prototypes (private to kvm(3)):

-int _kvm_kvatop(kvm_t *, u_long, u_long *);
-off_t _kvm_pa2off(kvm_t *, u_long);
+int _kvm_kvatop(kvm_t *, vaddr_t, paddr_t *);
+off_t _kvm_pa2off(kvm_t *, paddr_t);

Basically, use vaddr_t for VA and paddr_t for PA. In addition, for variables
representing addresses, use paddr_t or vaddr_t, depending on the context.

For most arches, vaddr_t and paddr_t are equivalent to unsigned long. However,
the change was needed for exotic situations, like i386 PAE, were unsigned long
is not suitable for PA which are 64 bits long. As this required a complete
change of the function prototypes, all arches had to be adapted accordingly.

Core files from before this commit should still work with the new code; I did
not see any direct dependency between core's structure and kvatop/pa2off.

The change was compile tested for all arches, as it impacts all of them.

See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

Ansify and KNF all functions within kvm(3). No objection on current-users@.

Used as ground for u_long => vaddr_t/paddr_t replacements in kvm(3)
private functions.

Compile tested for all arches. See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

use <i386/foo.h> in a few places.

Handle reading from raw disk devices.

Add address translation support for 4MB pages, which we use to map the
kernel text.

The address translation routine needs to use the plN_pi macros (for
indexing into the PTP), not the plN_i ones (for indexing into the entire
linear page table); cf. the analogous code for amd64. Fixes ~everything
to do with kernel core dumps on i386.

sync with kernel after the merge of yamt-x86pmap branch.

branches: 1.21.16;
pdtpaddr -> pdppaddr

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22280, verified by myself.

Merge the nathanw_sa branch.

branches: 1.18.2;
Don't include <machine/pmap.h> and <machine/vmparam.h> if _KERNEL isn't
defined. Include them explicitly in the few kvm_arch.c that need them.

<vm/vm.h> -> <uvm/uvm_extern.h>

remove redundant vm includes

More trailing white space.

Remove lint

Use pread(2) and pwrite(2) rather than lseek(2)/{read,write}(2).

Rewrite _kvm_kvatop() to work with the new crash dump format, and make
the i386 port use libkvm, not libkvm.old.

use <sys/cdefs.h> __RCSID() macro

Add _kvm_mdopen()

RCS id police.

Clean up a bit.

fix a piece of sloppy coding on my part. gdb -k still doesn't work, though.

update to match reality.

Eliminate _kvm_uvatop(), in favor of using the machine-independent VM structures.

Update to match current pmap.

branches: 1.3.2;
Fix kvm_uvatop().

just punt for now; don't want to deal with this & don't want it to core

branches: 1.1.1;
Initial revision

libkvm: whitespace

Switch to the new PTE naming. The old naming is now unused, remove it.

Remove the #include <sys/user.h> from all of libkvm.
sys/user.h is a stub that just #includes sys/pcb.h.
There are no 'struct pcb' anywhere in here, so I'm extremely doubtful
any of the builds will fail.
OTOH it might be relying on a header that pcb.h includes.
In any case i386 and amd64 build.

branches: 1.1.6; 1.1.12;
Import PAE support for kvm(3):
- add kvm_i386pae.c (used for PAE memory translations), and update Makefile
for libkvm build.
- in pdppaddr: pass a flag to indicate PAE mode. Use a bit ignored
by the MMU. Mask address with PG_FRAME to avoid side effects.

Tested with vmstat(1)/netstat(1) to debug core files of PAE and !PAE
kernels. Older kernel dumps will default to native i386 (!PAE) mode.

XXX Currently, savecore(8) will fail to dump a PAE kernel in a !PAE
environment (and reciprocally). So you need to sync and reboot
with a kernel of the same mode as the one that crashed. Once the dump
is successful, this does not matter anymore.

remove clause 3 from all my licenses that aren't conflicting with
another copyright claim line. again. (i did this in 2008 and then
did not update all of my personal templates.)

various ia64 updates:
- add a compat vfork because of stupid
- add a weak sbrk
- add a shmat syscall
- add an empty kvm implementation that links
- add missing fp stuff

Dummy support for ia64

libkvm support for the new m68k pmap.

Get rid of usrstack/USRSTACK. Document that the old version of getargv is
broken because of ASLR.

Add missing __RCSID("$NetBSD$");

branches: 1.18.6; 1.18.12;
Change kvm_pa2off() and kvm_kvatop() prototypes (private to kvm(3)):

-int _kvm_kvatop(kvm_t *, u_long, u_long *);
-off_t _kvm_pa2off(kvm_t *, u_long);
+int _kvm_kvatop(kvm_t *, vaddr_t, paddr_t *);
+off_t _kvm_pa2off(kvm_t *, paddr_t);

Basically, use vaddr_t for VA and paddr_t for PA. In addition, for variables
representing addresses, use paddr_t or vaddr_t, depending on the context.

For most arches, vaddr_t and paddr_t are equivalent to unsigned long. However,
the change was needed for exotic situations, like i386 PAE, were unsigned long
is not suitable for PA which are 64 bits long. As this required a complete
change of the function prototypes, all arches had to be adapted accordingly.

Core files from before this commit should still work with the new code; I did
not see any direct dependency between core's structure and kvatop/pa2off.

The change was compile tested for all arches, as it impacts all of them.

See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

Ansify and KNF all functions within kvm(3). No objection on current-users@.

Used as ground for u_long => vaddr_t/paddr_t replacements in kvm(3)
private functions.

Compile tested for all arches. See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

Remove clause 3 and 4 from TNF licenses

branches: 1.15.2; 1.15.34;
Added support for sun2 kernel cores.

Never ever commit changes without comiling it, even if you only change a
comment.

Use <m68k/kcore.h> rather than <machine/kcore.h> to ensure
machine independence.

Add _kvm_mdopen()

Update for new m68k/kcore.h. We use dispatch information provided
by the crash dump itself to facilitate cross-analysis. The m68k_cmn
for Utah pmaps reads MMU configuration from the crash dump to that
this works on 4k and 8k h/w page size systems.

Build the same libkvm.* on all m68k machines. (Well... except for the
stragglers that still use libkvm.old ) The m68k machines now dispatch
based on the machine name to the appropriate libkvm implementation.

Fix small typo in last commit.

Allow more flexibility in the format of cpu_kcore_hdr_t and isolate
the knowledge of that struct in the machine-dependent module.

RCS id police.

New libkvm/savecore implementation. With the following differences:
1) savecore will not access the dump or live-kernel directly. It
will always use the kvm-functions. Allowing it to work on kernels
that don't have a 1-1 PA-VA mapping.
2) the kvm-lib has some additional functions to accomplish 1
- kvm_dump_mkheader()
- kvm_dump_wrtheader()
- kvm_dump_inval()
3) the file formats of the dump generated by the kernel and the dump
generated by savecore have been changed. The file format now looks
like the format produced for 'normal' core dumps.
Ports not yet supporting the new kvm-format will be using libkvm.old/
savecore.old for the time being.

Take out cpu040 case for mac68k.

should work without specialization now.

Eliminate _kvm_uvatop(), in favor of using the machine-independent VM structures.

Merge the m68k files. The code itself should be merged.

m68k-ify the hp300 kvm file; use MACHINE_ARCH to pick files up

libkvm support for the new m68k pmap.

Change kvm_pa2off() and kvm_kvatop() prototypes (private to kvm(3)):

-int _kvm_kvatop(kvm_t *, u_long, u_long *);
-off_t _kvm_pa2off(kvm_t *, u_long);
+int _kvm_kvatop(kvm_t *, vaddr_t, paddr_t *);
+off_t _kvm_pa2off(kvm_t *, paddr_t);

Basically, use vaddr_t for VA and paddr_t for PA. In addition, for variables
representing addresses, use paddr_t or vaddr_t, depending on the context.

For most arches, vaddr_t and paddr_t are equivalent to unsigned long. However,
the change was needed for exotic situations, like i386 PAE, were unsigned long
is not suitable for PA which are 64 bits long. As this required a complete
change of the function prototypes, all arches had to be adapted accordingly.

Core files from before this commit should still work with the new code; I did
not see any direct dependency between core's structure and kvatop/pa2off.

The change was compile tested for all arches, as it impacts all of them.

See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

Ansify and KNF all functions within kvm(3). No objection on current-users@.

Used as ground for u_long => vaddr_t/paddr_t replacements in kvm(3)
private functions.

Compile tested for all arches. See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

Remove clause 3 and 4 from TNF licenses

branches: 1.3.2; 1.3.34;
Added support for sun2 kernel cores.

Update for new m68k/kcore.h. We use dispatch information provided
by the crash dump itself to facilitate cross-analysis. The m68k_cmn
for Utah pmaps reads MMU configuration from the crash dump to that
this works on 4k and 8k h/w page size systems.

Build the same libkvm.* on all m68k machines. (Well... except for the
stragglers that still use libkvm.old ) The m68k machines now dispatch
based on the machine name to the appropriate libkvm implementation.

#include <m68k/m68k.h>

branches: 1.17.2; 1.17.8;
Include <sys/param.h> for MAXPATHLEN

Define KREAD() inside kvm_private.h, for reusability.

Change kvm_pa2off() and kvm_kvatop() prototypes (private to kvm(3)):

-int _kvm_kvatop(kvm_t *, u_long, u_long *);
-off_t _kvm_pa2off(kvm_t *, u_long);
+int _kvm_kvatop(kvm_t *, vaddr_t, paddr_t *);
+off_t _kvm_pa2off(kvm_t *, paddr_t);

Basically, use vaddr_t for VA and paddr_t for PA. In addition, for variables
representing addresses, use paddr_t or vaddr_t, depending on the context.

For most arches, vaddr_t and paddr_t are equivalent to unsigned long. However,
the change was needed for exotic situations, like i386 PAE, were unsigned long
is not suitable for PA which are 64 bits long. As this required a complete
change of the function prototypes, all arches had to be adapted accordingly.

Core files from before this commit should still work with the new code; I did
not see any direct dependency between core's structure and kvatop/pa2off.

The change was compile tested for all arches, as it impacts all of them.

See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

Ansify and KNF all functions within kvm(3). No objection on current-users@.

Used as ground for u_long => vaddr_t/paddr_t replacements in kvm(3)
private functions.

Compile tested for all arches. See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

Handle reading from raw disk devices.

branches: 1.12.22;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22280, verified by myself.

Consistently spell "crash dump" as two separate words. From jmc@openbsd.

Use %lx instead of %x with _kvm_err when arg is u_long.

branches: 1.9.8;
More trailing white space.

branches: 1.8.2;
Use pread(2) and pwrite(2) rather than lseek(2)/{read,write}(2).

Fix a couple of warnings.

Put spaces around / in comment so 'machine/*.h' does not generate '/* inside
comment' warning.

/* Tyop. */

use <sys/cdefs.h> __RCSID() macro

In kvm_cmn_pa2off(), make sure to use the correct RAM segment start/size
values when adjusting the PA/offset. (Only broken in the case where we
had multiple RAM segments.)

Update for new m68k/kcore.h. We use dispatch information provided
by the crash dump itself to facilitate cross-analysis. The m68k_cmn
for Utah pmaps reads MMU configuration from the crash dump to that
this works on 4k and 8k h/w page size systems.

Build the same libkvm.* on all m68k machines. (Well... except for the
stragglers that still use libkvm.old ) The m68k machines now dispatch
based on the machine name to the appropriate libkvm implementation.

libkvm support for the new m68k pmap.

Get rid of usrstack/USRSTACK. Document that the old version of getargv is
broken because of ASLR.

Remove the #include <sys/user.h> from all of libkvm.
sys/user.h is a stub that just #includes sys/pcb.h.
There are no 'struct pcb' anywhere in here, so I'm extremely doubtful
any of the builds will fail.
OTOH it might be relying on a header that pcb.h includes.
In any case i386 and amd64 build.

branches: 1.21.4; 1.21.10;
Teach libkvm how kernel virtual memory is laided on 64-bit MIPS kernels.

branches: 1.20.2;
Change kvm_pa2off() and kvm_kvatop() prototypes (private to kvm(3)):

-int _kvm_kvatop(kvm_t *, u_long, u_long *);
-off_t _kvm_pa2off(kvm_t *, u_long);
+int _kvm_kvatop(kvm_t *, vaddr_t, paddr_t *);
+off_t _kvm_pa2off(kvm_t *, paddr_t);

Basically, use vaddr_t for VA and paddr_t for PA. In addition, for variables
representing addresses, use paddr_t or vaddr_t, depending on the context.

For most arches, vaddr_t and paddr_t are equivalent to unsigned long. However,
the change was needed for exotic situations, like i386 PAE, were unsigned long
is not suitable for PA which are 64 bits long. As this required a complete
change of the function prototypes, all arches had to be adapted accordingly.

Core files from before this commit should still work with the new code; I did
not see any direct dependency between core's structure and kvatop/pa2off.

The change was compile tested for all arches, as it impacts all of them.

See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

Ansify and KNF all functions within kvm(3). No objection on current-users@.

Used as ground for u_long => vaddr_t/paddr_t replacements in kvm(3)
private functions.

Compile tested for all arches. See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

branches: 1.18.16;
Handle reading from raw disk devices.

branches: 1.17.22;
Consistently spell "crash dump" as two separate words. From jmc@openbsd.

branches: 1.16.2;
Don't include <machine/pmap.h> and <machine/vmparam.h> if _KERNEL isn't
defined. Include them explicitly in the few kvm_arch.c that need them.

<vm/vm.h> -> <uvm/uvm_extern.h>

remove redundant vm includes

fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well. No functional changes.

branches: 1.12.6;
More trailing white space.

Use pread(2) and pwrite(2) rather than lseek(2)/{read,write}(2).

Rewrite NetBSD/mips crash dump address translation, based on NetBSD/alpha's,
using new crash dump format.

remove obsolete register declarations

kill lint.

use <sys/cdefs.h> __RCSID() macro

Add _kvm_mdopen()

MACH_CACHED_TO_PHYS() -> MIPS_KSEG0_TO_PHYS().

Include mips1_pte.h and mips3_pte.h. Read cpu_arch from the kernel
and set local variables for MIPSx_PG_V, MIPSx_PG_SHIFT_,
MIPSx_PG_FRAME accordingly.

(defining both MIPS3 and MIPS1 and using pte.h. doesn't yet work in userland.)

RCS id police.

Eliminate _kvm_uvatop(), in favor of using the machine-independent VM structures.

branches: 1.1.1;
Initial revision

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22280, verified by myself.

Bump date for last; remove end-of-line spaces.

Update after recent changes.

Use
.In header.h
instead of
.Fd #include \*[Lt]header.h\*[Gt]
Much easier to read and write, and supported by groff for ages.
Okayed by ross.

New sentence, new line. From Robert Elz.

branches: 1.6.2;
Generate <>& symbolically.

add LIBRARY section to man page

Don't use a kernel database based on the file name, use only
/var/db/kvm.db, and only use that if no kernel namelist file
is named. Update the documentation to describe this.

RCS id police.

vmunix -> netbsd (PR #375)

branches: 1.1.1;
Initial revision

Remove support for NetBSD/pc532.

branches: 1.17.22;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22280, verified by myself.

Consistently spell "crash dump" as two separate words. From jmc@openbsd.

branches: 1.15.2;
Need <machine/pmap.h> for the pdei and ptei macros.

Don't include <machine/pmap.h> and <machine/vmparam.h> if _KERNEL isn't
defined. Include them explicitly in the few kvm_arch.c that need them.

Use %lx instead of %x with _kvm_err when arg is u_long.

<vm/vm.h> -> <uvm/uvm_extern.h>

remove redundant vm includes

branches: 1.10.8;
More trailing white space.

branches: 1.9.2;
Use pread(2) and pwrite(2) rather than lseek(2)/{read,write}(2).

include stdlib.h to define free().

branches: 1.7.2;
use <sys/cdefs.h> __RCSID() macro

Add _kvm_mdopen()

The pc532 port now uses the new libkvm.

RCS id police.

Update for recent changes in the kernel.

Eliminate _kvm_uvatop(), in favor of using the machine-independent VM structures.

pc532 support in libkvm

Clean up deleted files.

add support for current m68k based pmaps till they can be "fixed"

add a function to report the name of the file kvm_open{,.files} is using.

Use .In instead of .Ar Pa for headers.

Make the documentation match reality. If no explicit kernel name is
provided, kvm_open() and kvm_openfiles() will try _PATH_KSYMS first,
then either the path from sysctl machdep.booted_kernel (if that's
available) or _PATH_UNIX (if the sysctl is not available).

Remove redundant .br.

branches: 1.14.40;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22280, verified by myself.

Pa Aq -> Aq Pa so that both work.

Use
.In header.h
instead of
.Fd #include \*[Lt]header.h\*[Gt]
Much easier to read and write, and supported by groff for ages.
Okayed by ross.

New sentence, new line. From Robert Elz.

"initialse" looks nice, but is not correct.

branches: 1.9.2;
Generate <>& symbolically.

Sort sections, fix punctuation nits.

Document KVM_NO_FILES and kvm_get{proc,argv,envv}2(). Better wording
kindly accepted!

XXX: Need to update distrib lists. Tomorrow...

branches: 1.6.6;
More trailing white space.

macroize BSD, NetBSD, FreeBSD and misc cleanup

add LIBRARY section to man page

use -mdoc .Pa macro for filenames

RCS id police.

branches: 1.1.1;
Initial revision

libkvm: whitespace

Get rid of usrstack/USRSTACK. Document that the old version of getargv is
broken because of ASLR.

New files for OR1K support

libkvm: whitespace

Get rid of usrstack/USRSTACK. Document that the old version of getargv is
broken because of ASLR.

Add missing __RCSID("$NetBSD$");

branches: 1.12.6; 1.12.12;
Change kvm_pa2off() and kvm_kvatop() prototypes (private to kvm(3)):

-int _kvm_kvatop(kvm_t *, u_long, u_long *);
-off_t _kvm_pa2off(kvm_t *, u_long);
+int _kvm_kvatop(kvm_t *, vaddr_t, paddr_t *);
+off_t _kvm_pa2off(kvm_t *, paddr_t);

Basically, use vaddr_t for VA and paddr_t for PA. In addition, for variables
representing addresses, use paddr_t or vaddr_t, depending on the context.

For most arches, vaddr_t and paddr_t are equivalent to unsigned long. However,
the change was needed for exotic situations, like i386 PAE, were unsigned long
is not suitable for PA which are 64 bits long. As this required a complete
change of the function prototypes, all arches had to be adapted accordingly.

Core files from before this commit should still work with the new code; I did
not see any direct dependency between core's structure and kvatop/pa2off.

The change was compile tested for all arches, as it impacts all of them.

See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

KNF

Ansify and KNF all functions within kvm(3). No objection on current-users@.

Used as ground for u_long => vaddr_t/paddr_t replacements in kvm(3)
private functions.

Compile tested for all arches. See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

Adapt to <powerpc/spr.h> breakup.

branches: 1.8.20;
Handle reading from raw disk devices.

branches: 1.7.16;
Supply _kvm_kvatop() and _kvm_pa2off() for OEA-type CPUs. This does
not yet support other PPC CPU types.
Contributed by Wasabi Systems Inc.

branches: 1.6.2;
Don't include <machine/pmap.h> and <machine/vmparam.h> if _KERNEL isn't
defined. Include them explicitly in the few kvm_arch.c that need them.

<vm/vm.h> -> <uvm/uvm_extern.h>

Need stdlib.h for prototypes.

Need <vm/vm.h>

Add _kvm_mdopen()

kvm routines for PowerPC; just stubs, for now. From Wolfgang Solfrank
<ws@tools.de>

libkvm: whitespace

Get rid of usrstack/USRSTACK. Document that the old version of getargv is
broken because of ASLR.

#if 0 some code not in use yet

branches: 1.5.4;
Add missing __RCSID("$NetBSD$");

branches: 1.4.6; 1.4.12;
Change kvm_pa2off() and kvm_kvatop() prototypes (private to kvm(3)):

-int _kvm_kvatop(kvm_t *, u_long, u_long *);
-off_t _kvm_pa2off(kvm_t *, u_long);
+int _kvm_kvatop(kvm_t *, vaddr_t, paddr_t *);
+off_t _kvm_pa2off(kvm_t *, paddr_t);

Basically, use vaddr_t for VA and paddr_t for PA. In addition, for variables
representing addresses, use paddr_t or vaddr_t, depending on the context.

For most arches, vaddr_t and paddr_t are equivalent to unsigned long. However,
the change was needed for exotic situations, like i386 PAE, were unsigned long
is not suitable for PA which are 64 bits long. As this required a complete
change of the function prototypes, all arches had to be adapted accordingly.

Core files from before this commit should still work with the new code; I did
not see any direct dependency between core's structure and kvatop/pa2off.

The change was compile tested for all arches, as it impacts all of them.

See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

Ansify and KNF all functions within kvm(3). No objection on current-users@.

Used as ground for u_long => vaddr_t/paddr_t replacements in kvm(3)
private functions.

Compile tested for all arches. See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

disable via radical lobotomy
this can't be properly converted now until kernel includes are available

ppc64 outline, needs more work

lib: remove CONSTCOND comment

Since 2021-01-31, lint doesn't need it anymore for the common pattern of
'do ... while (0)'.

Get rid of usrstack/USRSTACK. Document that the old version of getargv is
broken because of ASLR.

Implement kvm_write() for dump files. We map the file privately so that
changes don't get propagated back, and then write to the private buffer.
This is not useful yes, but it is simple enough to make pread use the buffer
if it is available.

add a function to report the name of the file kvm_open{,.files} is using.

Define KREAD() inside kvm_private.h, for reusability.

Change kvm_pa2off() and kvm_kvatop() prototypes (private to kvm(3)):

-int _kvm_kvatop(kvm_t *, u_long, u_long *);
-off_t _kvm_pa2off(kvm_t *, u_long);
+int _kvm_kvatop(kvm_t *, vaddr_t, paddr_t *);
+off_t _kvm_pa2off(kvm_t *, paddr_t);

Basically, use vaddr_t for VA and paddr_t for PA. In addition, for variables
representing addresses, use paddr_t or vaddr_t, depending on the context.

For most arches, vaddr_t and paddr_t are equivalent to unsigned long. However,
the change was needed for exotic situations, like i386 PAE, were unsigned long
is not suitable for PA which are 64 bits long. As this required a complete
change of the function prototypes, all arches had to be adapted accordingly.

Core files from before this commit should still work with the new code; I did
not see any direct dependency between core's structure and kvatop/pa2off.

The change was compile tested for all arches, as it impacts all of them.

See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

Ansify and KNF all functions within kvm(3). No objection on current-users@.

Used as ground for u_long => vaddr_t/paddr_t replacements in kvm(3)
private functions.

Compile tested for all arches. See also:

http://mail-index.netbsd.org/current-users/2010/09/07/msg014249.html

Handle reading from raw disk devices.

branches: 1.15.10;
1. Eliminate some unnecessary to kvm_{m,re}alloc.
2. Don't malloc/free procbase/procbase2/lwpbase continuously. Keep track
of the size, and only do it if necessary.
3. Write a macro to malloc/realloc and set the size of members so that it
is done correctly. Previous open coded version in kvm_file.c always
set the length, which is incorrect.
4. Remove bogus check against INT_MAX.
5. use NULL to initialize pointers instead of 0.

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22280, verified by myself.

Consistently spell "crash dump" as two separate words. From jmc@openbsd.

Make the kvm routines use /dev/ksyms to get the kernel namelist.
If it fails, use /netbsd instead.

Merge the nathanw_sa branch.

branches: 1.10.2;
format-string audit cleanups

Add kvm interface to the new sysctls:
kvm_getproc2() -> sysctl(KERN_PROC2)
kvm_getargv2() -> sysctl(KERN_PROC_ARGS, KERN_PROC_ARGV)
kvm_getenvv2() -> sysctl(KERN_PROC_ARGS, KERN_PROC_ENV)
Add new KVM_NO_FILES flag to kvm_open*() - set up enough state to
use the above calls without having to open any kernel files.

XXX: kvm_getprocs.3 and kvm_open.3 to be updated soon.

branches: 1.8.4;
Add the members: usrstack, min_uva, max_uva to struct __kvm
and declare the function _kvm_mdopen() that initializes them.

Allow more flexibility in the format of cpu_kcore_hdr_t and isolate
the knowledge of that struct in the machine-dependent module.

RCS id police.

New libkvm/savecore implementation. With the following differences:
1) savecore will not access the dump or live-kernel directly. It
will always use the kvm-functions. Allowing it to work on kernels
that don't have a 1-1 PA-VA mapping.
2) the kvm-lib has some additional functions to accomplish 1
- kvm_dump_mkheader()
- kvm_dump_wrtheader()
- kvm_dump_inval()
3) the file formats of the dump generated by the kernel and the dump
generated by savecore have been changed. The file format now looks
like the format produced for 'normal' core dumps.
Ports not yet supporting the new kvm-format will be using libkvm.old/
savecore.old for the time being.

Use the new ps_strings format. Find the arguments using their
addresses in the argv array, rather than guessing based on the location
of NULs.

Eliminate _kvm_uvatop(), in favor of using the machine-independent VM structures.

Move the swap page buffer into struct __kvm. Cache the page size during
kvm_open(). Fix an existing memory leak.

branches: 1.1.1;
Initial revision

remove unused.

branches: 1.99.2;
avoid various use-after-free issues.

create a ptrdiff_t offset between the start of an allocation region and
some interesting pointer, so it can be adjusted with this offset after
realloc() returns.

found by GCC 12.

lib: remove CONSTCOND comment

Since 2021-01-31, lint doesn't need it anymore for the common pattern of
'do ... while (0)'.

read the ps arguments from the correct place.

Get rid of usrstack/USRSTACK. Document that the old version of getargv is
broken because of ASLR.

Match the declaration in <sys/kauth.h>

Redo previous more conventionally. Requested by kre@.

Mask out always zero bits off pg->phys_addr.

branches: 1.92.16;
some ports need <sys/wait.h>

catch up with p_xstat split.

Remove the #include <sys/user.h> from all of libkvm.
sys/user.h is a stub that just #includes sys/pcb.h.
There are no 'struct pcb' anywhere in here, so I'm extremely doubtful
any of the builds will fail.
OTOH it might be relying on a header that pcb.h includes.
In any case i386 and amd64 build.

branches: 1.89.2;
Measure kinfo_proc2::p_vm_vsize in pages, as it was always documented.
This value seems to never have been used anywhere.
This makes it consistent with it's cousin p_vm_msize (which is in pages as
well and has several uses).