drm: Update drm2netbsd to prepare for new import.

branches: 1.5.10;
One more file by hand.

Add a couple more make variables.

Set CONFIG_DRM_VM=1, apparently needed for nouveau now.


Author: Maya Rashish <maya@NetBSD.org>

branches: 1.2.2; 1.2.6;
drm_fb_helper.c got lost in the merge, restore it.

New script to generate files.drmkms from Linux Makefile.

drm: Nix drm_agp_borrow.

This horrible kludge dates from before I understood the relation of
genfb_pci and pci drm drivers in the old and new worlds of drm.

The only user of it, in i915, was changed to use agp_i810_borrow
directly in the last drm update, so this hack can die.

Make ourselves less dependent on drmP.h, removed upstream.

This causes some trouble as CONFIG_* lines might not be appropriately
defined. A few declarations remain in drmP.h so it's not gone.


Author: Maya Rashish <maya@NetBSD.org>

drm_agp_clear got renamed to drm_agp_legacy_clear.

Include drm_internal.h by "../dist/drm/drm_internal.h".

branches: 1.3.2; 1.3.6;
if hooks init works, store the hooks in the agp. fixes start up
crash on an old radeon laptop. from @riastradh.

the console also now works, unlike previous drm2, where eg, the
problems from PR#49744 are seen.

unfortunately, i got a crash starting X, and no ddb yet, so we
can't re-enable the r100/r200 parts here yet. once this is
known to be working we can delete the old radeondrm code, and
remove the hacks to disable drm2 there.

initialize error variable

Rework PCI and AGP conditionalization.

- Push drm_pci_set_unique into driver callback.
- Eliminate drm_pci_set_unique_hook.
- Gather all drm_agp_* functions into struct drm_agp_hooks.
- Replace the nonsensical old atomic garbage by serious locking.
- Make drm_agpsupport.c its own module.
- Eliminate NDRMKMS_PCI.
- Use NAGP from "agp.h" only in drm_module.c for horrible hack.
=> See comment in file for rationale.
- Always define CONFIG_PCI=1 and CONFIG_AGP=1.
- Always go through the drm_agp_* function hooks.
- Ifdef out nouveau agp stuff that doesn't go through drm_agp_*
for reasons that I'm too frustrated to figure out tonight.
- pci_iomap no longer automagically does agp_i810_borrow.
=> Use drm_agp_borrow instead.

drm: Make drm_clflush_* noop on alpha.

As far as I know there is no need for this; the CPU has weak ordering
on loads and stores requiring ordering barriers, but I/O devices
participate in the CPU's cache protocol for access to main memory so
there's never any need for explicit cache flushes.

drm: Another pass over i915 and some supporting logic.

This makes a shim around sg_table, which essentially represents two
things:

1. an array of pages (roughly corresponding to bus_dma_segment_t[])
2. an array of DMA addresses stored in a bus_dmamap_t

Both parts are optional; different parts of i915 use sg_tables to
pass around one or both of the two parts. This helps to reduce the
ifdefs by quite a bit, although it's not always clear which part of
an sg_table any particular interface is actually using which is why I
was reluctant to do this before.

drm: Nix use of uvm pglist. Just use arrays of page pointers.

Make ourselves less dependent on drmP.h, removed upstream.

This causes some trouble as CONFIG_* lines might not be appropriately
defined. A few declarations remain in drmP.h so it's not gone.


Author: Maya Rashish <maya@NetBSD.org>

drm_clflush_* decls moved to drm_cache.h.

x86: fix several CPUID flags

- Rename: CPUID_PN -> CPUID_PSN
CPUID_CFLUSH -> CPUID_CLFSH
CPUID_SBF -> CPUID_PBE
CPUID_LZCNT -> CPUID_ABM
CPUID_P1GB -> CPUID_PAGE1GB
CPUID2_PCLMUL -> CPUID2_PCLMULQDQ
CPUID2_CID -> CPUID2_CNXTID
CPUID2_xTPR -> CPUID2_XTPR
CPUID2_AES -> CPUID2_AESNI
To match the x86 specification and the other OSes.

- Remove: CPUID_B10, CPUID_B20, CPUID_IA64. They do not exist.

branches: 1.13.4;
skip clflush on aarch64, too

Draft sparc cache flushing.

Factor mfence out. Simplify a little.

Need sync after a series of dcbf's on powerpc.

Implement drm_md_clflush_* for powerpc with dcbf.

branches: 1.8.10; 1.8.16; 1.8.18;
skip clflush on arm

machine/cpufunc.h is MD, move it to i386/amd64 block

Remove local definition of wbinvd.

Use x86_mfence explicitly, not membar_sync, just to be clear.

CLFLUSH needs to be surrounded by membars.

While here, round everything to a multiple of the cache line size.

branches: 1.3.2; 1.3.4; 1.3.6;
Make it build and boot on my test machines.

Screen blanks on boot on the Ivy Bridge system with

DRM error in cpt_serr_int_handler: PCH transcoder A FIFO underrun

But after that everything is OK. Appears to be an upstream problem.
To investigate...

I think there's a cache flushing issue somewhere -- there are little
display artefacts on my T60.

branches: 1.2.2; 1.2.4; 1.2.6;
Merge riastradh-drm2 to HEAD.

branches: 1.1.2;
file drm_cache.c was initially added on branch riastradh-drm2.

drm(4): Fix st_rdev in stat.

dminor->index already has the 64*type adjustment, as allocated in
drm_minor_alloc.

PR kern/58180

branches: 1.30.4;
uvm(9): fo_mmap caller guarantees positive size.

No functional change intended, just sprinkling assertions to make it
clearer.

amdgpu: Make amdgpu_file_to_fpriv work.

drm: Omit bogus __unused attributes.

Initialize priv->filp, lost in removing drm_fops.c


Author: Maya Rashish <maya@NetBSD.org>
Committer: Taylor R Campbell <riastradh@NetBSD.org>

Eliminate drm_file.c copypasta.

drm_agpsupport.h for drm_legacy_agp_clear.

Remove last users of drmP.h.

Make ourselves less dependent on drmP.h, removed upstream.

This causes some trouble as CONFIG_* lines might not be appropriately
defined. A few declarations remain in drmP.h so it's not gone.


Author: Maya Rashish <maya@NetBSD.org>

some declarations are now guarded under CONFIG_DRM_LEGACY.

Put our code under CONFIG_DRM_LEGACY, too, for now.


Author: Maya Rashish <maya@NetBSD.org>

Include some headers now that drmP.h got split off.


Author: Maya Rashish <maya@NetBSD.org>

Serialize drm_read so we can back out on uiomove without reordering.

Upstream commit 9b2c0b7fb4ce79566d830d03ce7aa11cccc39f97.

drm_agp_clear got renamed to drm_agp_legacy_clear.

drm_lastclose now rightly returns void.

Include drm_internal.h by "../dist/drm/drm_internal.h".

Change the kqueue filterops::f_isfd field to filterops::f_flags, and
define a flag FILTEROP_ISFD that has the meaning of the prior f_isfd.
Field and flag name aligned with OpenBSD.

This does not constitute a functional or ABI change, as the field location
and size, and the value placed in that field, are the same as the previous
code, but we're bumping __NetBSD_Version__ so 3rd-party module source code
can adapt, as needed.

NetBSD 9.99.89

Use sel{record,remove}_knote().

branches: 1.14.4; 1.14.14;
fix various problems i've seen where cv_*wait*() return ERESTART,
which is -3 in netbsd, which we have mapped linux ERESTARTSYS to.

this has a problem because linux code often returns errors and
pointers in the same value, and pointer values between -4095 and
-1 are considered as error returns, but -3 ends up as 3 and thus
is not considered an error, and mayhem ensues.

with this in place my kabylake system seems actually stable, i
have not triggered any of my prior issues in almost 4 weeks now.


Taylor asked me to write up a description and then wrote most of
the text below for me :-)


In Linux code, we always work with ERESTARTSYS so the code meaning
start over is a positive NetBSD errno safe for PTR_ERR/ERR_PTR.
To achieve this:
1. adapt all cv_waits that return to Linux so they map ERESTART to
ERESTARTSYS, and
2. adapt all returns to userland so they convert ERESTARTSYS to
ERESTART.
Leave EINTR and all other error codes alone.

Expose nvif ioctl interface.

nvif is a variable length (nested..) ioctl, so it doesn't match the
usual drm_ioctl interface. linux uses a shim to override the ioctl
function for nouveau to allow this, do the same.

fixes 3D acceleration with nouveau.

from riastradh.

branches: 1.12.2;
Rework PCI and AGP conditionalization.

- Push drm_pci_set_unique into driver callback.
- Eliminate drm_pci_set_unique_hook.
- Gather all drm_agp_* functions into struct drm_agp_hooks.
- Replace the nonsensical old atomic garbage by serious locking.
- Make drm_agpsupport.c its own module.
- Eliminate NDRMKMS_PCI.
- Use NAGP from "agp.h" only in drm_module.c for horrible hack.
=> See comment in file for rationale.
- Always define CONFIG_PCI=1 and CONFIG_AGP=1.
- Always go through the drm_agp_* function hooks.
- Ifdef out nouveau agp stuff that doesn't go through drm_agp_*
for reasons that I'm too frustrated to figure out tonight.
- pci_iomap no longer automagically does agp_i810_borrow.
=> Use drm_agp_borrow instead.

Move decl of drm_guarantee_initialized to drmP.h.

Locks are not like negatives: doubles don't cancel out.

We don't need drm_legacy_vma_flush.

Rename drm_mmap_* -> drm_legacy_mmap_* like upstream.

Copy body of static drm_legacy_dev_reinit from drm_fops.c to drm_cdevsw.c.

open_count is protected by drm_global_mutex; count_lock is no more.

Actually convert errno as comment suggests.

Move drm_ioctl to drm_ioctl.c.

Previously it was in drm_drv.c upstream, which we overrode; then it
got moved to drm_ioctl.c and drm_drv.c got deleted, so we no longer
need to maintain a separate copy of the table in what we now call
drm_cdevsw.c.

Update DRM_IOCTL_DEF.

Make drm_lastclose match upstream.

Rename drm_drv.c to drm_cdevsw.c.

Previously we overrode upstream's drm_drv.c altogether, and used
upstream's drm_stub.c subroutines.

But upstream deleted drm_drv.c and renamed drm_stub.c to drm_drv.c,
so we have to cope with that reshuffle.

drm: Unstub drm_client slightly more.

drm: Slightly unstub this toe.

drm_client stubs

Rename drm_drv.c to drm_cdevsw.c.

Previously we overrode upstream's drm_drv.c altogether, and used
upstream's drm_stub.c subroutines.

But upstream deleted drm_drv.c and renamed drm_stub.c to drm_drv.c,
so we have to cope with that reshuffle.

directly include asm/bug.h for WARN_ON instead of fighting
to match side-loading.

include drm_legacy.h as needed


Author: coypu <coypu@sdf.org>
Committer: Taylor R Campbell <riastradh@NetBSD.org>

branches: 1.20.2; 1.20.4;
drm_stat: fix device minor calculation, ok riastradh@

add fo_name so we can identify the fileops in a simple way.

Use C99 initializer for filterops

Mostly done with spatch with touchups for indentation

@@
expression a;
identifier b,c,d;
identifier p;
@@
const struct filterops p =
- { a, b, c, d
+ {
+ .f_isfd = a,
+ .f_attach = b,
+ .f_detach = c,
+ .f_event = d,
};

branches: 1.17.8; 1.17.10;
only clear agp hook on last close when agp is in use, from riastradh@

fix __OS_HAS_AGP=0 build

Don't return events that are too large and leave them in the list.

Synchronize permission checks with upstream drm_ioctl_permit.

due to hangs seen by several folks, for now revert:
http://mail-index.netbsd.org/source-changes/2014/11/04/msg060120.html

Log Message:
This code should be MP-safe. Use IPL_SCHED in place of IPL_DRM/IPL_VM and set
D_MPSAFE flag in cdevsw.

add a new "fo_mmap" fileops method to allow use of arbitrary uvm_objects for
mappings of file objects. move vnode-specific details of mmap()ing a vnode
from uvm_mmap() to the new vnode-specific vn_mmap(). add new uvm_mmap_dev()
and uvm_mmap_anon() convenience functions for mapping character devices
and anonymous memory, and replace all other calls to uvm_mmap() with those.
use the new fileop in drm2 so that libdrm can use mmap() to map things
like on other platforms (instead of the ioctl that we have used so far).

branches: 1.11.2;
Move all PCI/AGP code back to drmkms_pci module where it belongs.

This code should be MP-safe. Use IPL_SCHED in place of IPL_DRM/IPL_VM and set
D_MPSAFE flag in cdevsw.

branches: 1.9.2; 1.9.4;
Let the drm driver decide what part of what object gets mmapped.

Pass the file around too so radeon/ttm can get at it to verify access.

Add drm_gem_or_legacy_mmap_object for drivers to choose the previous
behaviour, like i915.

Add d_discard to all struct cdevsw instances I could find.

All have been set to "nodiscard"; some should get a real implementation.

Make it build and boot on my test machines.

Screen blanks on boot on the Ivy Bridge system with

DRM error in cpt_serr_int_handler: PCH transcoder A FIFO underrun

But after that everything is OK. Appears to be an upstream problem.
To investigate...

I think there's a cache flushing issue somewhere -- there are little
display artefacts on my T60.

Hook up drm suspend/resume.

Destroy, don't leak, events when done with them.

branches: 1.4.2; 1.4.4;
Tweak some DRM GEM page indexing crap.

- Fix order of subtraction in drm_mmap_paddr_locked.
- Address GEM objects' pages from 0, not from the mmap cookie.
- Check page alignment earlier in mmap code paths.
- Sprinkle kasserts throughout.

Still doesn't fix the garbage that is sometimes being scribbled all
over kernel memory!

branches: 1.3.2;
Figure drm_debug into the drm_init module kludgerosity.

Merge riastradh-drm2 to HEAD.

branches: 1.1.2;
file drm_drv.c was initially added on branch riastradh-drm2.

Eliminate drm_file.c copypasta.

Make ourselves less dependent on drmP.h, removed upstream.

This causes some trouble as CONFIG_* lines might not be appropriately
defined. A few declarations remain in drmP.h so it's not gone.


Author: Maya Rashish <maya@NetBSD.org>

Include newly split off headers Don't initialize struct members that no longer exist. Make DRIVER_PRIME code unconditional, similar to upstream drm_file.c


Author: Maya Rashish <maya@NetBSD.org>

Sync drm_file.c. Prepare to convert event read lock to cv.

Rename drm_fops.c -> drm_file.c

Rename drm_fops.c -> drm_file.c

Reduce diffs by nixing DRM_SUSER macro, deleted upstream long ago.

Instead stub Linux capable(CAP_SYS_ADMIN).

branches: 1.14.6;
Draft support for drm prime.

Make sure all of struct drm_file gets initialized.

Sync drm_close_file: call drm_property_release_user_blobs.

Release the magic before releasing the master.

Releasing the master frees the idr into which we are trying to
release the magic. Use of magic that has been freed sounds like a
dangerous proposition.

Implement drm_new_set_master in drm_fops.c and use it.

More drm_fops.c updates. Hooray, code deletion!

Apply some more legacy renames. drm magic got replaced by an idr.

Fix headers for drm_fops.c.

Update always_authenticated -> authenticated.

include drm_legacy.h as needed...


Author: coypu <coypu@sdf.org>
Committer: Taylor R Campbell <riastradh@NetBSD.org>

branches: 1.4.4; 1.4.28; 1.4.30;
Make it build and boot on my test machines.

Screen blanks on boot on the Ivy Bridge system with

DRM error in cpt_serr_int_handler: PCH transcoder A FIFO underrun

But after that everything is OK. Appears to be an upstream problem.
To investigate...

I think there's a cache flushing issue somewhere -- there are little
display artefacts on my T60.

branches: 1.3.2; 1.3.4; 1.3.6;
Fix up removal of _DRM_SHM mappings and the hardware lock.

Merge riastradh-drm2 to HEAD.

branches: 1.1.2;
file drm_fops.c was initially added on branch riastradh-drm2.

avoid uninitialised variable usage in drm_gem_cma_create_internal().

in the case nothing has returned 'error', 'nsegs' and the dma info
are (potentially) uninitialised, so consider this an error.

found by GCC 12.

branches: 1.14.4;
drm: Rewrite drm_gem_cma_helper.h from scratch.

The header file was originally part of drm with no licence text,
which I took to mean BSD-licensed like the rest of the Linux drm code
base is by default -- everything in Linux under include/drm and
drivers/gpu/drm not otherwise marked.

In 2017, a tree-wide automated change was applied to Linux which
caused an SPDX GPL header to appear in drm_gem_cma_helper.h. When I
asked the drm developers about this in drm files, they explained that
it was an automated process and many of the SPDX headers were wrong,
so when excluding GPL files in drm updates, I heeded the actual GPL
vs BSD/MIT licence text and revision history rather than the SPDX
headers.

I excluded drm_gem_cma_helper.c because it was clearly originally GPL
as an exception to the default BSD licence under Linux's
drivers/gpu/drm, and Jared reimplemented the API for ues by some
(NetBSD, BSD-licensed) Arm display drivers.

During an update after that, the header file was lost in the merge
shuffle and I copied it back in without noticing the SPDX header
file. In retrospect, I see now that it's not clear what the intent
was: the .c file was obviously GPL, and maybe the .h file was
supposed to be too but it wasn't carved out as an exception among the
drm code.

To keep our story simpler, I just drafted a new .h file that provides
a compatible subset of the interface we use, derived from the .c file
Jared wrote, written according to several deterministic rules based
on KNF that yield a non-creative specification of the compatible
interface:

1. $NetBSD$ RCS id
2. standard TNF copyright statement
3. header guard derived from #include file name
4. minimal #includes, grouped and sorted according to KNF
5. forward declaration of structs, sorted lexicographically
6. definitions of structs, members sorted lexicographically by name
7. downcasting macros, written in KNF and sorted lexicographically
8. minimal extern variable declarations, sorted lexicographically
9. minimal, function prototypes, written in KNF and sorted
lexicographically

`Minimal' means added one by one, according to the errors issued by
the compiler (and in the case of #includes, according to the correct
header file for the missing types or function prototypes being used),
until the compiler is formally satisfied.

Any changes to this will, unless otherwise marked, be covered by the
standard TNF licence now on the file.

Make ourselves less dependent on drmP.h, removed upstream.

This causes some trouble as CONFIG_* lines might not be appropriately
defined. A few declarations remain in drmP.h so it's not gone.


Author: Maya Rashish <maya@NetBSD.org>

Catch up with function renames

reference -> get, unreference -> put


Author: Maya Rashish <maya@NetBSD.org>

drm_gem_cma_dumb_map_offset is no more.

do not wait for memory in pgo_fault methods, just return ENOMEM
and let the uvm_fault code wait if it is appropriate.

branches: 1.9.2;
Add support for DRM GEM/CMA helpers using a driver provided vmem arena.

Map buffer objects with BUS_DMA_PREFETCHABLE.

branches: 1.7.4;
drm_gem_cma_fault() is a UVM fault function. return netbsd errnos.

Use dmat, not bus_dmat, to respect drm_limit_dma_space.

drm_gem_cma_create_internal can fail.

Rework drm prime guts.

- Our faux struct sg_table is now just an array of _pages_, not of
bus dma segments.

- We handle bus dma segments only in the presence of an adult bus dma
tag.

- All bus dma hacks are once again confined to bus_dma_hacks.h, and
more neatly organized:
. bus_dmamem_export_pages translates bus dma segs to vm_page array
. bus_dmamem_import_pages translates vm_page array to bus dma segs

Implement remainder of drm_gem_cma_helper API.

Fix up drm_gem_cma_helper.

branches: 1.1.2; 1.1.4;
Implement the DRM GEM/CMA helpers. The implementation has been extracted
from our tegra DRM driver, but generalized for use with other drivers.

drm/drm_print.h for struct drm_printer drm/drm_fourcc.h for struct drm_format_info


Author: Maya Rashish <maya@NetBSD.org>

Add includes, fix reference path.

Local reimplementation of GPL drm_gem_framebuffer_helper.c.

uvm(9): fo_mmap caller guarantees positive size.

No functional change intended, just sprinkling assertions to make it
clearer.

Make ourselves less dependent on drmP.h, removed upstream.

This causes some trouble as CONFIG_* lines might not be appropriately
defined. A few declarations remain in drmP.h so it's not gone.


Author: Maya Rashish <maya@NetBSD.org>

drm/drm_drv.h for struct drm_device


Author: Maya Rashish <maya@NetBSD.org>

Catch up with function renames

reference -> get, unreference -> put


Author: Maya Rashish <maya@NetBSD.org>

Need <sys/file.h> to dereference struct file.

Pass drm_file, not file, pointer to drm_vma_node_is_allowed.

Rename drm_mmap_* -> drm_legacy_mmap_* like upstream.

Need drm_legacy.h for legacy mmap stuff.

Need to use new <drm/drm_gem.h> for GEM stuff.

branches: 1.6.16; 1.6.18;
Apply access control to gem mmap.

branches: 1.5.2; 1.5.4; 1.5.6;
Let the drm driver decide what part of what object gets mmapped.

Pass the file around too so radeon/ttm can get at it to verify access.

Add drm_gem_or_legacy_mmap_object for drivers to choose the previous
behaviour, like i915.

Make it build and boot on my test machines.

Screen blanks on boot on the Ivy Bridge system with

DRM error in cpt_serr_int_handler: PCH transcoder A FIFO underrun

But after that everything is OK. Appears to be an upstream problem.
To investigate...

I think there's a cache flushing issue somewhere -- there are little
display artefacts on my T60.

branches: 1.3.2; 1.3.4;
Tweak some DRM GEM page indexing crap.

- Fix order of subtraction in drm_mmap_paddr_locked.
- Address GEM objects' pages from 0, not from the mmap cookie.
- Check page alignment earlier in mmap code paths.
- Sprinkle kasserts throughout.

Still doesn't fix the garbage that is sometimes being scribbled all
over kernel memory!

branches: 1.2.2;
Merge riastradh-drm2 to HEAD.

branches: 1.1.2;
file drm_gem_vm.c was initially added on branch riastradh-drm2.

drm: Stub out drm_hdcp.c

stub drm_lease.c the same way as openbsd


Author: Maya Rashish <maya@NetBSD.org>
Committer: Taylor R Campbell <riastradh@NetBSD.org>

drm_lease_destroy stub

Fix typo: list lessees, not list leases.

Sprinkle some necessary includes and forward decls.

Unfinished local implementation of GPL drm_lease.c.

drm: Make some legacy UMS crap build.

Make ourselves less dependent on drmP.h, removed upstream.

This causes some trouble as CONFIG_* lines might not be appropriately
defined. A few declarations remain in drmP.h so it's not gone.


Author: Maya Rashish <maya@NetBSD.org>

Make our code follow CONFIG_DRM_LEGACY.

We might want to delete all this code later, but no rush.


Author: Maya Rashish <maya@NetBSD.org>

Sync drm_lock.c. (Needs review, but only used by legacy drivers.)

Include drm_internal.h by "../dist/drm/drm_internal.h".

Move proc_lock into the data segment. It was dynamically allocated because
at the time we had mutex_obj_alloc() but not __cacheline_aligned.

Need <drm/drm_internal.h> for drm_global_mutex.

Rename some drm_* to drm_legacy_*.

include drm_legacy.h as needed...


Author: coypu <coypu@sdf.org>
Committer: Taylor R Campbell <riastradh@NetBSD.org>

branches: 1.4.16; 1.4.18;
Take a stab at implementing drm_idlelock_take/release.

Evidently needed by VIA DRM/UMS. Noted and tested by medfly/coypu.

branches: 1.3.4; 1.3.6;
Make it build and boot on my test machines.

Screen blanks on boot on the Ivy Bridge system with

DRM error in cpt_serr_int_handler: PCH transcoder A FIFO underrun

But after that everything is OK. Appears to be an upstream problem.
To investigate...

I think there's a cache flushing issue somewhere -- there are little
display artefacts on my T60.

branches: 1.2.2; 1.2.4; 1.2.6;
Merge riastradh-drm2 to HEAD.

branches: 1.1.2;
file drm_lock.c was initially added on branch riastradh-drm2.

Misc stubs.

<drm/drm_print.h> for DRM_DEBUG

Remove last users of drmP.h.

Make ourselves less dependent on drmP.h, removed upstream.

This causes some trouble as CONFIG_* lines might not be appropriately
defined. A few declarations remain in drmP.h so it's not gone.


Author: Maya Rashish <maya@NetBSD.org>

Tread carefully around DMA subregions.

Not all platforms support it.

- If we want to limit to 32-bit DMA, just use a known 32-bit DMA tag.
- If we want to limit to >32-bit DMA, but the host only has a known
32-bit DMA tag, then just use that anyway.
- If we want to limit to >32-bit DMA, but the host doesn't support
bus_dmatag_subregion, then just use the 32-bit DMA tag anyway.

Gives powerpc, sparc a chance of working.

Get the legacy drm declarations from drm_legacy.h.

Rename drm_core_ioremap -> drm_legacy_ioremap as upstream.

branches: 1.10.16; 1.10.18;
Non-x86 doesn't have AGP. This should really be handled by config logic.

Make agp_i810 also depend on agp so that a 'no agp' build is possible.

build on platforms without agp_i810

Replace drm_ioremap by drm_core_ioremap, reducing diff a little.

branches: 1.6.2; 1.6.4; 1.6.6;
Make it build and boot on my test machines.

Screen blanks on boot on the Ivy Bridge system with

DRM error in cpt_serr_int_handler: PCH transcoder A FIFO underrun

But after that everything is OK. Appears to be an upstream problem.
To investigate...

I think there's a cache flushing issue somewhere -- there are little
display artefacts on my T60.

Honour write-combining flag in drm_ioremap.

Check bounds in agp_i810_borrow.

Out of paranoia, do a bus_space_subregion in case the old drm code
tries sizes that the agp_i810 code doesn't agree with.

branches: 1.3.2; 1.3.4;
Fix error branch in drm_limit_dma_space.

branches: 1.2.2;
Merge riastradh-drm2 to HEAD.

branches: 1.1.2;
file drm_memory.c was initially added on branch riastradh-drm2.

drm: Fix conditionals around drmkms_pci and agp.

Kernel should build now with all pci drm drivers stripped out but
DRM_LEGACY still enabled. (Might not be very useful, but it'll
build. Maybe we should also have DRM_LEGACY_PCI so those drivers can
be modloaded later.)

branches: 1.31.4;
drm: Note dependency on sysmon_power.

drm uses this for pswitch events to notify of hotplug.

drm: Dust off module build a little.

drm: Reduce default debug level to driver-specific messages.

Let's make a liiiiittle less log spew by default with `boot -x'.

rkdrm: Convert to atomic modesetting, as needed for bridges.


Author: Jared McNeill <jmcneill@invisible.ca>
Committer: Taylor R Campbell <riastradh@NetBSD.org>

Make intel_psr.c build
netbsd-ify
Stub out intel_lpe_audio.c (for now?).
Hammer intel_opregion.c into shape.
Hack up intel_lrc.c
Adapt intel_reset.c
Stub out intel_region_lmem.c
Adapt intel_ring_submission.c
Adapt intel_rps.c
Adapt intel_sdvo.c
Adapt intel_runtime_pm.c
Adapt intel_sideband.c
Adapt intel_sprite.c
Adapt intel_tv.c
Adapt intel_tc.c
Adapt intel_sseu.c
Adapt intel_timeline.c
Adapt intel_uc_fw.c and fix typo
Fix up intel_uncore.c
intel_vga.c
intel_wakeref.c
intel_vdsc.c
intel_wopcm.c
intel_workarounds.c
i915_globals.c
Fix up intelfb.c
Stub out intel_guc_log.c
Stub out intel_dsi_dcs_backlight.c
i915_perf.c
intel_hdmi.c
intel_hdcp.c


Author: Maya Rashish <maya@NetBSD.org>
Committer: Taylor R Campbell <riastradh@NetBSD.org>

Initialize drm_core_init_complete in drm_module.c

Linux achieves this with a module_init(drm_core_init);


Author: Maya Rashish <maya@NetBSD.org>
Committer: Taylor R Campbell <riastradh@NetBSD.org>

drm_global_init/release is no more

Remove last users of drmP.h.

Make ourselves less dependent on drmP.h, removed upstream.

This causes some trouble as CONFIG_* lines might not be appropriately
defined. A few declarations remain in drmP.h so it's not gone.


Author: Maya Rashish <maya@NetBSD.org>

Missing includes, conditionalize on drm legacy

Might want to remove drm legacy code entirely later.


Author: Maya Rashish <maya@NetBSD.org>

Initialize kernel_fb_helper_lock in drm module load.

Declare and initialize drm_unplug_srcu our way.

Need drm_crtc_internal.h for drm_connector_ida_init/fini.

Include drm_internal.h by "../dist/drm/drm_internal.h".

Initialize drm_bridge and drm_panel locks.

when "boot -x", don't turn on *all* drm debug, just the core, driver
and kms messages, eliding the vblank, atomic and prime messages,
which are the truly noisy ones (and may result in impossibly slow
to use systems.)

XXX: pullup-all.

branches: 1.15.4;
Rework PCI and AGP conditionalization.

- Push drm_pci_set_unique into driver callback.
- Eliminate drm_pci_set_unique_hook.
- Gather all drm_agp_* functions into struct drm_agp_hooks.
- Replace the nonsensical old atomic garbage by serious locking.
- Make drm_agpsupport.c its own module.
- Eliminate NDRMKMS_PCI.
- Use NAGP from "agp.h" only in drm_module.c for horrible hack.
=> See comment in file for rationale.
- Always define CONFIG_PCI=1 and CONFIG_AGP=1.
- Always go through the drm_agp_* function hooks.
- Ifdef out nouveau agp stuff that doesn't go through drm_agp_*
for reasons that I'm too frustrated to figure out tonight.
- pci_iomap no longer automagically does agp_i810_borrow.
=> Use drm_agp_borrow instead.

Move decl of drm_guarantee_initialized to drmP.h.

Make sure to create and destroy the lock too...

Make drmkms build without drmkms_pci.

branches: 1.11.16; 1.11.18;
Update module dependencies:

drmkms does not depend on iic
drmkms_linux depends on i2cexec

Rewrite drm_encoder_slave.c, as nouveau will need.

branches: 1.9.2;
prettify and add to all the modules that have it.

Fix module build of drmkms.

branches: 1.7.2; 1.7.4;
Initialize the drm_global stuff.

Make it build and boot on my test machines.

Screen blanks on boot on the Ivy Bridge system with

DRM error in cpt_serr_int_handler: PCH transcoder A FIFO underrun

But after that everything is OK. Appears to be an upstream problem.
To investigate...

I think there's a cache flushing issue somewhere -- there are little
display artefacts on my T60.

branches: 1.5.2; 1.5.4; 1.5.6;
Figure drm_debug into the drm_init module kludgerosity.

Need <sys/reboot.h> for AB_DEBUG.

Enable drm debug output iff boothowto has AB_DEBUG set.

Merge riastradh-drm2 to HEAD.

branches: 1.1.2;
file drm_module.c was initially added on branch riastradh-drm2.

drm: Make some legacy UMS crap build.

Remove last users of drmP.h.

Missing includes, conditionalize on drm legacy

Might want to remove drm legacy code entirely later.


Author: Maya Rashish <maya@NetBSD.org>

Include drm_internal.h by "../dist/drm/drm_internal.h".

drm_sg_* legacy

branches: 1.4.2; 1.4.4;
XXX: add a NULL init to avoid a GCC 6 maybe uninit warning.

branches: 1.3.4; 1.3.22;
Make it build and boot on my test machines.

Screen blanks on boot on the Ivy Bridge system with

DRM error in cpt_serr_int_handler: PCH transcoder A FIFO underrun

But after that everything is OK. Appears to be an upstream problem.
To investigate...

I think there's a cache flushing issue somewhere -- there are little
display artefacts on my T60.

branches: 1.2.2; 1.2.4; 1.2.6;
Merge riastradh-drm2 to HEAD.

branches: 1.1.2;
file drm_scatter.c was initially added on branch riastradh-drm2.

branches: 1.1.8;
Add a stub for drm_ioctl() so that compat32 links with kernels without drmkms.

i915: Adapt i915_perf.c.

Teach drm_sysctl about uint, kludgily.

branches: 1.6.16; 1.6.18; 1.6.20;
Add $NetBSD$ tag.

Add string Linux module parameters.

Needed by nouveau.

From chs@.

branches: 1.4.2; 1.4.4;
Trivial build fix for no options SYSCTL_INCLUDE_DESCR case, to
avoid -Werror.
OK christos@ and riastradh@

fix description setting.

prettify and add to all the modules that have it.

Add __link_set based code to automatically convert the linux module parameters
into sysctls.

drm: Fill in fake connector->kdev pointing at the drm device.

Hope this will make progress -- hope it won't screw up too badly if
two connectors have the same kdev!

Restore drm_sysfs_connector_add/remove stubs.

Update drm_sysfs stubs.

Use drmP.h less in our local files.


Author: Maya Rashish <maya@NetBSD.org>

Include drm_internal.h by "../dist/drm/drm_internal.h".

Internal prototypes from drm_internal.h. Reflect upstream deletions.

branches: 1.3.4; 1.3.28; 1.3.30;
Make it build and boot on my test machines.

Screen blanks on boot on the Ivy Bridge system with

DRM error in cpt_serr_int_handler: PCH transcoder A FIFO underrun

But after that everything is OK. Appears to be an upstream problem.
To investigate...

I think there's a cache flushing issue somewhere -- there are little
display artefacts on my T60.

branches: 1.2.2; 1.2.4; 1.2.6;
Merge riastradh-drm2 to HEAD.

branches: 1.1.2;
file drm_sysfs.c was initially added on branch riastradh-drm2.

uvm(9): fo_mmap caller guarantees positive size.

No functional change intended, just sprinkling assertions to make it
clearer.

Use drmP.h less in our local files.


Author: Maya Rashish <maya@NetBSD.org>

Reduce diffs by nixing DRM_SUSER macro, deleted upstream long ago.

Instead stub Linux capable(CAP_SYS_ADMIN).

branches: 1.10.6;
Need <uvm/uvm.h> for struct uvm_object.

Rename drm_mmap_* -> drm_legacy_mmap_* like upstream.

Get the legacy drm declarations from drm_legacy.h.

branches: 1.7.16; 1.7.18;
fix a confusion between dev_t and devmajor_t, and include the primary
minor when calling makedev().

add a new "fo_mmap" fileops method to allow use of arbitrary uvm_objects for
mappings of file objects. move vnode-specific details of mmap()ing a vnode
from uvm_mmap() to the new vnode-specific vn_mmap(). add new uvm_mmap_dev()
and uvm_mmap_anon() convenience functions for mapping character devices
and anonymous memory, and replace all other calls to uvm_mmap() with those.
use the new fileop in drm2 so that libdrm can use mmap() to map things
like on other platforms (instead of the ioctl that we have used so far).

branches: 1.5.2; 1.5.4; 1.5.6;
Let the drm driver decide what part of what object gets mmapped.

Pass the file around too so radeon/ttm can get at it to verify access.

Add drm_gem_or_legacy_mmap_object for drivers to choose the previous
behaviour, like i915.

Make it build and boot on my test machines.

Screen blanks on boot on the Ivy Bridge system with

DRM error in cpt_serr_int_handler: PCH transcoder A FIFO underrun

But after that everything is OK. Appears to be an upstream problem.
To investigate...

I think there's a cache flushing issue somewhere -- there are little
display artefacts on my T60.

branches: 1.3.2; 1.3.4;
Tweak some DRM GEM page indexing crap.

- Fix order of subtraction in drm_mmap_paddr_locked.
- Address GEM objects' pages from 0, not from the mmap cookie.
- Check page alignment earlier in mmap code paths.
- Sprinkle kasserts throughout.

Still doesn't fix the garbage that is sometimes being scribbled all
over kernel memory!

branches: 1.2.2;
Merge riastradh-drm2 to HEAD.

branches: 1.1.2;
file drm_vm.c was initially added on branch riastradh-drm2.

drm: Add drm_vma_offset_exact_lookup_locked.

drm: Fix sign error in comparison for vma lookup.

drm_vma_manager deals in struct drm_file, not struct file, now.

Kinda inconvenient because it means we'll need to patch
filp->private_data with filp->f_data in various places...

Lock the manager and not just the node for inserting/removing nodes

should fix/help PR kern/50349: radeondrmkms vt-switching crash

ok riastradh

branches: 1.4.2;
Make sure rbtrees are empty on desctruction.

If related to PR kern/51076, might help catch the bug a bit earlier.

plug memory leak in drm_vma_node_revoke().

Don't sleep in the drm_vma allocator. Fail with ENOSPC, not ENOMEM.

Caller has dev->struct_mutex locked and will handle ENOSPC anyway.

branches: 1.1.2; 1.1.4; 1.1.6; 1.1.8;
Make it build and boot on my test machines.

Screen blanks on boot on the Ivy Bridge system with

DRM error in cpt_serr_int_handler: PCH transcoder A FIFO underrun

But after that everything is OK. Appears to be an upstream problem.
To investigate...

I think there's a cache flushing issue somewhere -- there are little
display artefacts on my T60.

drm_writeback stubs

drmfb: Rework console detection and takeover.

While here, simplify logic by removing an unreachable error branch so
we never have to contemplate unwinding claiming of the console.

drmfb: Take is_console property from parent, not from self.

The parent is the pci or platform device that firmware knows about.
The `drmfb' device is more of a fictitious invention of the drm stack
that exists for the convenience of the genfb abstraction which needs
device_private to be a struct genfb_softc.

Let's see if this does any better than the last attempt to rework the
logic here.

revert "drm: Respect is_console property on main drm device, not fb child."

breaks booting on some machines

ok riastradh@

drm: Respect is_console property on main drm device, not fb child.

The MD device enumeration sets the property on, e.g., PCI devices --
not on drm framebuffer children, which are mostly a software
abstraction. Not sure examining the nouveaufb, intelfb, &c., device
properties is worthwhile at all, but in case it breaks something
let's just leave it in while we're fixing other things.

drm: add missing KERNEL_LOCK around calls to config_found().

drm: Simplify use of drm API for display on/off logic in drmfb.

Remove last users of drmP.h.

Iterate over connectors differently (i.e. in a way that compiles)


Author: Maya Rashish <maya@NetBSD.org>

prop_data_create_data+prop_dictionary_set+prop_object_release -> prop_dictionary_set_data

Check property length, not array (fixes clang build)

Add missing prop_object_release(), from thorpej

Make the first EDID we find available to genfb

Remove temporary drm_fb_helper_set_config stub from old code.

branches: 1.3.14; 1.3.16; 1.3.18;
add da_fb_linebytes to drmfb_attach_args and use it to pass linebytes
from nouveau code to drmfb. keep the same linebytes logic for i915.

nvidia hardware needs 256 byte alignment, so aligning to just 64 was
not enough.

fixes broken console with a width of 1440px (PR kern/51181)
ok riastradh

branches: 1.2.2;
If the "is_console" property is already specified, do not overwrite it.

branches: 1.1.2; 1.1.4;
Factor generic parts of intelfb out into a drmfb.

Adapting radeondrmkmsfb should not be hard.

This will make nouveaufb easier too.

drm: New option DRM_LEGACY.

This brings in the `drmums' components, which are the parts of the
current drm code base (sys/external/bsd/drm2) needed by legacy
user-mode-setting drivers (currently just viadrmums), in contrast to
the old drm code base (sys/external/bsd/drm).

This way, viadrmums.kmod can be loaded into GENERIC i386 kernels, but
platforms without any relevant legacy UMS drivers don't get the
legacy goo statically baked in.

drmkms_sched: Remove duplicate module definition.

drm: Fix CPPFLAGS for drm scheduler code.

drm: Split scheduler into separate module.

Used only by amdgpu, and it uses something that is currently broken
from the gcc update on aarch64, so let's make it a little more
optional.

drm: Keep drm include flags in CPPFLAGS.drmkms after all.

Failures from this are much more obvious to debug than files with
missing #ifdefs, which sometimes lead to build failures but could
also just lead to silent miscompilation.

drm: Restore pollution of CPPFLAGS with include directories.

Too much trouble to make this fine-grained.

drm: Undefine `alpha' in CPPFLAGS. Causes lotsa trouble!

But don't undefine it outside drmkms; `#ifdef alpha' or equivalent is
used elsewhere in-tree. (Maybe it should be replaced by __alpha__.)

drm: Narrow scope of drm CPPLAGS just to drm files.

No need for various Linux compat definitions elsewhere in the kernel.

drm: Propagate CPPFLAGS.drmkms into all the drm drivers.

drm: Note dependency on sysmon_power.

drm uses this for pswitch events to notify of hotplug.

rename CONFIG_AARCH64 to CONFIG_ARM64 which is what upstream code expects

drm: sprinkle some -Wno-unused-function for clang

915_sw_fence.c: for debug_fence_init_onstack
drm_mm.c: for rb_hole_size_to_node
intel_hdmi.c: for intel_hdmi_hdcp2_protocol
intel_guc_submission.c: for rq_prio
nouveau_nvkm_subdev_instmem_base.c: for fake_ioread32_native
nouveau_nvkm_subdev_mmu_vmmnv04.c: for nv04_vmm_pgt_pte

drm: Make scheduler mandatory to make amdgpu module simpler.

drm: Make some legacy UMS crap build.

amdgpu: Stub out irq domain stuff with a finer grain.

drm: Place drm_hdcp.c like other overrides in files.drmkms.

drm: Stub out drm_hdcp.c

drm: Remove local diffs for drm_pci_set_busid.

Instead, move it to its own file that can be safely used without
bringing in a link/run-time dependency on pci.

Get drm_mm.c building.

Just replace the interval tree lookup by a linear search for now.

Revert part of previous: no separate CPPFLAGS.drmkms.

There are too many drms users outside sys/external/bsd/drm2 for now;
will hold off on this until we find a better way to keep the drm
includes modularized.

Narrow the CPPFLAGS for drm subsystems.

No need for the whole kernel to be built with -I flags for the drm
header files.

drm_client stubs

Regen and remerge.

drm_writeback stubs

Eliminate drm_file.c copypasta.

Rerun drm2netbsd and merge.

New tag drmums for the legacy stuff; to deal with later for via.

Empty out drm_os_netbsd.h.

Disable missing files for now.

Merge updated drm2netbsd output.

Remove these two files from the list.

We have our own copies of the same name. Since we don't support
building two files with the same name, it's pretty clear the
intention was to build just our copy of the file.


Author: Maya Rashish <maya@NetBSD.org>

upstream shadows, disable warning


Author: Maya Rashish <maya@NetBSD.org>

We have local versions of these files, don't use upstream ones


Author: Maya Rashish <maya@NetBSD.org>

Upstream uses shadowing, disable warning


Author: Maya Rashish <maya@NetBSD.org>

Disable warnings about shadowing.

Upstream does shadowing here, I'm feeling uncomfortable changing their
code.


Author: Maya Rashish <maya@NetBSD.org>

skip upstream drm_cache.c, we have our own


Author: Maya Rashish <maya@NetBSD.org>

Remove duplicates and non-existent files.


Author: Maya Rashish <maya@NetBSD.org>

misc drm hacks

Unhook drm_kms_helper_common.c from build.

Define CONFIG_DRM_FBDEV_OVERALLOC=100 to match upstream.

Rename drm_fops.c -> drm_file.c

Use -Wno-missing-field-initializers for drm to reduce diffs.

Move IS_REACHABLE to <linux/kernel.h> to avoid double definition.

Define IS_ENABLED and IS_REACHABLE on the command line.

Header file is too late for it, apparently.

XXX Find a way to put this into a header file; command-line -D
arguments are ugly.

No more hacks for dist/uapi -- got moved to dist/include/uapi.

Disable *_trace_points.c.

I accidentally deleted these in the import, because the *_trace.h
files got in the way and I forgot the *_trace_points.c files were
fine.

There's nothing important in them anyway, so nothing of value is
lost. Maybe next time I will remember to delete the *_trace.h files
on import but not the *_trace_points.c to reduce diffs, but that's
too much trouble at this point.

Unfinished local implementation of GPL drm_lease.c.

Local reimplementation of GPL drm_gem_framebuffer_helper.c.

Generate files.* using the 2netbsd scripts.


Author: Maya Rashish <maya@NetBSD.org>

drm2: build with -fwrapv, since the Linux kernel does and we don't
want to spend too much time fixing their code. silences a lot of
KUBSAN warnings.

we need to use COPTS to have per-option compiler flags. help from
Riastradh.

branches: 1.34.10; 1.34.14;
Eliminate -Wpointer-arith patches; use -Wno-pointer-arith instead.

branches: 1.33.2;
build drm_panel.c w/ drmkms

don't build stub if legacy drm is present too.

Add a stub for drm_ioctl() so that compat32 links with kernels without drmkms.

branches: 1.30.4;
Rework PCI and AGP conditionalization.

- Push drm_pci_set_unique into driver callback.
- Eliminate drm_pci_set_unique_hook.
- Gather all drm_agp_* functions into struct drm_agp_hooks.
- Replace the nonsensical old atomic garbage by serious locking.
- Make drm_agpsupport.c its own module.
- Eliminate NDRMKMS_PCI.
- Use NAGP from "agp.h" only in drm_module.c for horrible hack.
=> See comment in file for rationale.
- Always define CONFIG_PCI=1 and CONFIG_AGP=1.
- Always go through the drm_agp_* function hooks.
- Ifdef out nouveau agp stuff that doesn't go through drm_agp_*
for reasons that I'm too frustrated to figure out tonight.
- pci_iomap no longer automagically does agp_i810_borrow.
=> Use drm_agp_borrow instead.

Unhook upstream drm_pci.c, apparently overridden by files.drmkms_pci.

Draft support for drm prime.

Define CONFIG_DRM_FBDEV_EMULATION=1 in generic drmkms code.

This was accidentally pulled in by i915drmkms, but we need it
generically or else we need to patch upstream differently.

Sort. Put CWARNFLAGS next to source lists.

drm_fb_helper.c got lost in the merge, restore it.

Move drm_ioctl to drm_ioctl.c.

Previously it was in drm_drv.c upstream, which we overrode; then it
got moved to drm_ioctl.c and drm_drv.c got deleted, so we no longer
need to maintain a separate copy of the table in what we now call
drm_cdevsw.c.

Patch up drm_atomic_helper.c.

- Mark unused variables as such.
. Mistakes in local patches are likely to lead to unused variables,
so keep this warning enabled.
- Build with -Wno-shadow.
. Local patches aren't likely to cause this, so just suppress it.
- static plane_crtc_active

We have a local override for drm_encoder_slave.c.

Reorder header files to avoid drm_legacy.h lossage.

Restore more locally overridden files.

Rename drm_drv.c to drm_cdevsw.c.

Previously we overrode upstream's drm_drv.c altogether, and used
upstream's drm_stub.c subroutines.

But upstream deleted drm_drv.c and renamed drm_stub.c to drm_drv.c,
so we have to cope with that reshuffle.

Apply locally overridden files.

Omit PCI files from files.drmkms.

regen files.drmkms

Move local files to the top to make them more visible.

branches: 1.14.2; 1.14.4;
Implement the DRM GEM/CMA helpers. The implementation has been extracted
from our tegra DRM driver, but generalized for use with other drivers.

Move linux_work.c and workqueue.h from sys/external/bsd/drm2 to
sys/external/common so that they can be used by others.

LGTM from riastradh@

These should really live outside sys/external, but that can be decided
later

Tweak IS_ENABLED to allow CONFIG_FOO values.

No functional change to drm2.

LGTM from riastradh@

Factor generic parts of intelfb out into a drmfb.

Adapting radeondrmkmsfb should not be hard.

This will make nouveaufb easier too.

Rewrite drm_encoder_slave.c, as nouveau will need.

branches: 1.9.2;
Add __link_set based code to automatically convert the linux module parameters
into sysctls.

Split drm ttm config into a separate file.

branches: 1.7.2; 1.7.4;
First whack at radeon driver.

No hardware to test yet, but it builds.

Make it build and boot on my test machines.

Screen blanks on boot on the Ivy Bridge system with

DRM error in cpt_serr_int_handler: PCH transcoder A FIFO underrun

But after that everything is OK. Appears to be an upstream problem.
To investigate...

I think there's a cache flushing issue somewhere -- there are little
display artefacts on my T60.

branches: 1.5.2; 1.5.4; 1.5.6;
Make drmkms find header files from sys/external/bsd/common.

Move the drmkms i2c dependencies to where they belong.

add missing dependencies (Takahiro HAYASHI)

Merge riastradh-drm2 to HEAD.

branches: 1.1.2;
file files.drmkms was initially added on branch riastradh-drm2.

drmkms_sched: Remove duplicate module definition.

drm: Split scheduler into separate module.

Used only by amdgpu, and it uses something that is currently broken
from the gcc update on aarch64, so let's make it a little more
optional.