Disable undefined behavior sanitizer on intel_workarounds.c,
I don't see a real bug here compiling kernel ALL/i386.

Maybe a GCC 14 bug?

avoid various seemingly bogus warnings with GCC 14.

i couldn't see any real case for any of these issues. it seems like
the "maybe uninit" checker has gotten worse.

additionally, the "dangling pointer" checker also failing.

match warnings with the module build, fixes i386 with GCC 12.

adapt to ${CC_WNO_MAYBE_UNINITIALIZED}

Use ${CC_WNO_MAYBE_UNINITIALIZED} instead of
the older style more complex expressions.

Remove workarounds if they were for a specific
version of gcc < 10.

adapt to ${CC_WNO_IMPLICIT_FALLTHROUGH}

Use ${CC_WNO_IMPLICIT_FALLTHROUGH} instead of
the older style more complex expressions.

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

i915: Remove spurious local i915_perf.c override.

i915: Use new acpidisp_register_notify for ACPI VGA events.

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

i915: Draft gem fault.

i915: intel_engine_pm.c and i915_active.c

...with a little less heinous abuse of C, and an attempt to
disentangle the tentacular abstraction violations rampant in these
components

i915: Disentangle uabi engines rb/llist/list spaghetti.

i915: Odds & ends.

i915: Define trace points

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>

i915: more work

i915: more work

i915: hack hack hack

i915: more passes

i915: another whack at it

i915: More progress.

i915: Odds and ends.

i915: Adapt sw_fence.

drm: Another pass over i915. Most of i915 gem builds now.

i915_globals.c: don't build, not useful (we don't __init etc)


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

drm/i915: Undo local i915_sw_fence override.

The code is now MIT-licensed, so we no longer need to rewrite it.

drm/i915: Unhook i915_gem_client_blt.c and revert changes.

Only used by self-tests, which we're not using (for now...).

i915: Disable i915_active.c for now; needs more work.

i915: Touch up local i915_sw_fence stubs to make it compile.

i915: Disable CONFIG_DRM_I915_DEBUG_RUNTIME_PM.

No runtime_pm at all here for now.

Need to provide a value for CONFIG_DRM_I915_DEBUG_RUNTIME_PM

Maybe 1 will do.


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

i915: Disable -Wmissing-field-initializers.

i915: Fix include path for intelfb files too.

define CONFIG values to be the same as linux


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

Add include flag for intelfb.c's use of i915_drv.h header


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

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.

Merge updated i915drmkms2netbsd output.

Comment out debugfs files, we probably don't want those.


Author: Maya Rashish <maya@NetBSD.org>

Remove duplicates and non-existent files.


Author: Maya Rashish <maya@NetBSD.org>

misc drm hacks

Disable the i915_oa_* files (for now?).

The initializers with overrides got moved to i915_pci.c.

Disable i915 alpha hardware support.

i915_pci.c collides; rename ours to i915_pci_autoconf.c.

Stub out i915_gemfs.c and i915_perf.c.

i915_mm.c is of no use to us.

First pass at i915, far from complete.

HEAVILY MODIFIED IN MERGE - maya

Rename intel_dsi.c -> vlv_dsi.c

Stub i915_sw_fence.c.

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.

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: 1.40.2; 1.40.6;
apply some -Wno- to ignore several new warnings with GCC 10.


ntp: ignore truncation beyond api sizes, and ignore wrongly
guessed underflow

tmux: ignore maybe uninitialised warning for impossible case

libbfd: signed/unsigned variables assigned in the same statement

dri/gallium: -Wno-builtin-declaration-mismatch for u_atomic.c as
it implements backend functions with different in-C-machine but
same-in-real-machine types

libXfont/libXfont2: signed/unsigned variables assigned in the same
statement

i915drm: ignore impossible maybe uninitialised warnings

sysinst: ignore an invalid string truncation issue

branches: 1.39.8;
Eliminate -Wpointer-arith patches; use -Wno-pointer-arith instead.

branches: 1.38.2;
convert HAVE_GCC == 7 to HAVE_GCC >= 7.

add -Wno-error=int-in-bool-context or -Wno-error=implicit-fallthrough
for various drm files that would be best ignored or upstream fixed.
(could fix savagedrm, no one upstream is that any more.)

Fix merge botch.

Merge the [pgoyette-compat] branch

i915drmkms requires firmload. PR 53892

Appease GCC with a kernel sanitizer

Add -Wno-maybe-uninitialized for intel_sprite.c that is a false positive.

Tag files.* with newer config version where needed.

Draft support for drm prime.

Put override directory first in CPPFLAGS.

Move CPPFLAGS to top, CWARNFLAGS below.

Omit commented warn flag, evidently needless.

Stub out intel_dsi_init.

Ifdef out Linux audio API stuff and re-enable intel_audio.c.

Override i915_gem_userptr.c with stubs for now.

Restore intel_fbdev.c.

Disable mipi dsi stuff until I can figure it out.

Just do all of i915drmkms with -Wno-shadow.

Disable intel_audio.c until we figure out how to make it talk to hdaudio.

i915_gem_gtt.c shadows.

Override i915_sysfs.c.

Allow shadowing in i915_vgpu.c.

i915_gpu_error.c shadows, don't worry about it.

Don't worry about shadowing here.

Think we're gonna need to override i915_gem_userptr.c.

bring in old drm changes

ifdef out some fb_info usage which we don't have and some of it
ifdef'd out in old drm

strangely reg_offset isn't used in the i2c whatever-thing

we need one more member to drm_fb_helper

define CONFIG_DRM_FBDEV_EMULATION so hopefully fewer redefinitions


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

regen files.i915drmkms

Move local i915 files earlier to make them more obvious.

branches: 1.11.16; 1.11.18;
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.

branches: 1.9.2;
Fix build with clang.

i915drmkms(4): Enable CONFIG_ACPI if NACPICA > 0.
Now brightness can be adjusted via hotkey on Mouse Computer LB-J300X (Clevo W330SU2).

We need the i810 agp here, intel_gtt has it hard-coded.

branches: 1.6.2;
Restore Intel opregion stuff.

branches: 1.5.2;
Move CONFIG_X86 from files.i915drmkms to drm_os_netbsd.h.

Rework Intel framebuffer attachment.

Turns out genfb_softc must be at the start of its own device_private,
so we can't have it in the middle of a structure, and certainly can't
have potentially multiple genfbs per graphics device. Failing to do
this is why entering ddb didn't work with an i915drmkms console.

Instead of putting a genfb_softc into each drm_fb_helper, put a
device_t into each drm_fb_helper and create a new intelfb(4) device
for each Intel framebuffer. This will be more flexible, too, in case
we want accelerated framebuffers later on. (XXX Need to adapt the
radeon code, which I'm disabling until that happens to avoid breaking
the build.)

While here, defer disabling VGA, vga_cndetach, and genfb_attach until
we're actually ready to do them all together along with the mode
switch. This should reduce the amount of time during which the
screen is blanked, in case it hangs.

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 files.i915drmkms was initially added on branch riastradh-drm2.

branches: 1.1.2;
file i915_gem.c was initially added on branch riastradh-drm2.

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.

Rework gen6 aperture/gtt size detection. Mark variables __diagused.

Work around broken GTT size detection in agp_i810 code.

This will do until I work out the twisty maze of registers, all
different, to reliably determine the size of the GTT (and hence the
GPU's virtual address space) separately from the size of the AGP
aperture.

Fix bogus GTT total size calculation. Omit gtt_bsh for `agp' gtt.

Fix GTT PTE flag bits.

branches: 1.11.2;
Miscellaneous little fixes for harmless issues:

- Fix sense of subtraction in i915_gem_restore_gtt_mappings (not
currently used, so this was harmless).

- Program gtt->gtt_scratch_map->dm_segs[0].ds_addr rather than
gtt->gtt_scratch_seg.ds_addr into the GTT for consistency with the
bus_dma API. These are currently the same, but if we ever start
using x86 iommu perhaps that may change.

- Kassert that the scratch PTE decodes into the scratch address.

Remove (bad) debugging message that crept into i915_gem_gtt_init.

Fix page/byte sense of AGP GTT parameter calculations.

Use correct value for gtt->gtt_bsh in agp case (unused, so harmless).

Use the scratch page, not zero, to clear ggtt entries for gen<6.

branches: 1.6.2;
Tweak i915 gen6_gtt_init calculations for clarity.

Cast from uint32_t to bus_addr_t earlier, since it may exceed 32 bits.

Clear only the requested range in gen6_ggtt_clear_range.

Sprinkle kasserts throughout i915_gem_gtt.c.

branches: 1.3.2;
Miscellaneous fixes for drm2 on i386.

Merge riastradh-drm2 to HEAD.

branches: 1.1.2;
file i915_gem_gtt.c was initially added on branch riastradh-drm2.

i915: Initialize userptr wq for now. Should wire this up for real!

drm: Another pass over i915. Most of i915 gem builds now.

../dist/drm/i915/gem/i915_gem_ioctls.h for i915_gem_userptr_ioctl prototype


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

Update i915_gem_init_userptr signature.

branches: 1.1.2; 1.1.6;
Override i915_gem_userptr.c with stubs for now.

Change local drm files include path to match upstream changes.


Author: Maya Rashish <maya@NetBSD.org>

Stub out i915_gemfs.c and i915_perf.c.

branches: 1.1.2;
file i915_irq.c was initially added on branch riastradh-drm2.

i915: Omit needless i915_global_buddy_init declaration.

i915: Use new acpidisp_register_notify for ACPI VGA events.

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>

i915: Odds and ends.

i915: Adapt sw_fence.

drm: Another pass over i915. Most of i915 gem builds now.

drm/i915: Unhook i915_gem_client_blt.c and revert changes.

Only used by self-tests, which we're not using (for now...).

i915: machete

i915: Call i915_global_buddy_init from i915 module init.

Remove last users of drmP.h.

No more load-time hacking of the struct drm_driver structure, yay!

Move decl of drm_guarantee_initialized to drmP.h.

Eliminate drm_pci_init/exit. Empty for a while, no need.

Fix definition and initialization of mchdev_lock.

branches: 1.5.18; 1.5.20;
prettify and add to all the modules that have it.

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

branches: 1.3.2; 1.3.4;
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 i915_module.c was initially added on branch riastradh-drm2.

i915_pci.c collides; rename ours to i915_pci_autoconf.c.

Call i915_drm_suspend_late too.

Call linux_pci_dev_destroy on detach.

Pass the parent explicitly to linux_pci_dev_init.

This will allow us to use it in match routines.

Move linux_pci_dev_init into callers of drm_pci_attach.

The order in nouveau has to be

- linux_pci_dev_init
- nvkm_device_pci_new
- rest of drm_pci_attach

i915 needs firmware now too, so defer to mountroot.

Attempt to hook up i915 suspend/resume callbacks.

branches: 1.15.4; 1.15.28; 1.15.30;
Tweak i915drmkms_detach to be more robust.

Rework Intel framebuffer attachment.

Turns out genfb_softc must be at the start of its own device_private,
so we can't have it in the middle of a structure, and certainly can't
have potentially multiple genfbs per graphics device. Failing to do
this is why entering ddb didn't work with an i915drmkms console.

Instead of putting a genfb_softc into each drm_fb_helper, put a
device_t into each drm_fb_helper and create a new intelfb(4) device
for each Intel framebuffer. This will be more flexible, too, in case
we want accelerated framebuffers later on. (XXX Need to adapt the
radeon code, which I'm disabling until that happens to avoid breaking
the build.)

While here, defer disabling VGA, vga_cndetach, and genfb_attach until
we're actually ready to do them all together along with the mode
switch. This should reduce the amount of time during which the
screen is blanked, in case it hangs.

Omit redundant vga_is_console test.

This was a copypasta error -- before it tested iot then memt -- but I
see no evidence that passing memt to vga_is_console is ever sensible.

Get suspend/resume sorta working for i915drm.

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.

The vga driver now detaches wscons.

Implement i915drmkms wsdisplay blank/backlight/brightness controls.

branches: 1.8.2; 1.8.4;
Take the console only from vga or generic genfb.

Adapted from a patch by nonaka@ in PR 48705.

While here, attach i915 genfb earlier as in PR 48706.

Restore is_console=1 for i915, even if I keep it off for debugging.

Add missing declaration from last commit.

branches: 1.5.2;
Tweak i915 framebuffer attachment in preparation for detach/reconfig.

Can't do this yet because genfb can't detach or reconfigure...

Miscellaneous fixes for drm2 on i386.

Make i915drmkms grab the console by default.

Merge riastradh-drm2 to HEAD.

branches: 1.1.2;
file i915_pci.c was initially added on branch riastradh-drm2.

drm: Simplify async framebuffer child task attachment logic.

No need for scheduling the task to fail. Nix needless error
branches.

branches: 1.1.2; 1.1.6;
Rework Intel framebuffer attachment.

Turns out genfb_softc must be at the start of its own device_private,
so we can't have it in the middle of a structure, and certainly can't
have potentially multiple genfbs per graphics device. Failing to do
this is why entering ddb didn't work with an i915drmkms console.

Instead of putting a genfb_softc into each drm_fb_helper, put a
device_t into each drm_fb_helper and create a new intelfb(4) device
for each Intel framebuffer. This will be more flexible, too, in case
we want accelerated framebuffers later on. (XXX Need to adapt the
radeon code, which I'm disabling until that happens to avoid breaking
the build.)

While here, defer disabling VGA, vga_cndetach, and genfb_attach until
we're actually ready to do them all together along with the mode
switch. This should reduce the amount of time during which the
screen is blanked, in case it hangs.

i915: Suspend ioctls while device is suspended.

i915: Make sure to call i915_drm_prepare too on suspend.

drm: Simplify async framebuffer child task attachment logic.

No need for scheduling the task to fail. Nix needless error
branches.

i915: Mark a KASSERT-only variable __diagused.

Minor KNF fix while here.

drm: Rework attach/detach and deferred task logic.

- Reduce the number of states the softc can be in.
- Fix races between attach and other threads.

i915: Omit needless reachover of i915_drm_driver.

i915: Nix drm_dev_alloc, and do drm_pci_attach in i915_driver_probe.

i915 uses a struct drm_device embedded in struct drm_i915_private
now, so we can't allocate it separately.

drm: Omit needless pci_attach_args from our drm_pci_attach.

Can already get it from the struct pci_dev, and this way we can more
easily call it from patched upstream code as i915 will need to do.

Return pci_device_t from match so we can call i915_driver_probe


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

drm: Another pass over i915. Most of i915 gem builds now.

Split drm_dev_alloc/register out of drm_pci_attach.

Needed by nouveau which needs to do

drm_dev_alloc
drm_pci_attach
nouveau_drm_device_init
drm_dev_register

Remove last users of drmP.h.

Reflect rename IS_PRELIMINARY_HW -> IS_ALPHA_SUPPORT.

i915_pci.c collides; rename ours to i915_pci_autoconf.c.

i915: Remove spurious local i915_perf.c override.

Fix copypasto.

Stub out i915_gemfs.c and i915_perf.c.

drm/i915: Undo local i915_sw_fence override.

The code is now MIT-licensed, so we no longer need to rewrite it.

Stub i915_sw_fence.c.

drm/i915: Undo local i915_sw_fence override.

The code is now MIT-licensed, so we no longer need to rewrite it.

i915: Touch up local i915_sw_fence stubs to make it compile.

Add some more stub i915_sw_fence prototypes.

Add i915_sw_fence_await_sw_fence_gfp stub declaration.

Stub i915_sw_fence.c.

i915: More progress.

i915: More progress.

Update i915_setup/teardown_sysfs prototypes.

branches: 1.2.4; 1.2.6; 1.2.10;
Merge riastradh-drm2 to HEAD.

branches: 1.1.2;
file i915_sysfs.c was initially added on branch riastradh-drm2.

i915: Define trace points

i915: Update i915drmkms2netbsd to prepare for new import.

branches: 1.3.10;
Break down and use gmake for i915drmkms2netbsd too.


Author: Maya Rashish <maya@NetBSD.org>

branches: 1.2.2; 1.2.6;
Restore intel_fbdev.c.

Scripts to generate radeon and intel file lists.

i915: Add missing stub file.

Rename intel_dsi.c -> vlv_dsi.c

branches: 1.2.2; 1.2.6;
Use dtrace probes to implement i915_trace.h.

Stub out intel_dsi_init.

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>

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 intel_fb.c was initially added on branch riastradh-drm2.

i915: more work

i915: ggtt

Update intel gtt API.

TODO: Find where the stolen memory base is in agp_i810.

Convert to uintmax_t if we're going to use %jx.

Linux uses uint64 for these parameters now, so follow suit.

Cope with multi-page segments.

Call it intel_gtt_insert_sg_entries for expedience.

It uses a bus_dmamap, not a Linux scatterlist a.k.a. sg, but
otherwise it functions the same way and this avoids patching
upstream.

Restore gtt page table control register on resume.

Additional subroutine agp_i810_reset is used by i915drmkms to make
sure it happens early enough, since i915drmkms resumes before agp.

XXX pullup-7
XXX pullup-8

Flush chipset writes after GGTT update.

Echoes Linux commit:

commit 8516673a996870ea0ceb337ee4f83c33c5ec3111
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date: Fri Dec 8 21:46:16 2017 +0000

agp/intel: Flush all chipset writes after updating the GGTT

Before accessing the GGTT we must flush the PTE writes and make them
visible to the chipset, or else the indirect access may end up in the
wrong page. In commit 3497971a71d8 ("agp/intel: Flush chipset writes
after updating a single PTE"), we noticed corruption of the uploads for
pwrite and for capturing GPU error states, but it was presumed that the
explicit calls to intel_gtt_chipset_flush() were sufficient for the
execbuffer path. However, we have not been flushing the chipset between
the PTE writes and access via the GTT itself.

For simplicity, do the flush after any PTE update rather than try and
batch the flushes on a just-in-time basis.

References: 3497971a71d8 ("agp/intel: Flush chipset writes after updating a single PTE")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Cc: drm-intel-fixes@lists.freedesktop.org
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171208214616.30147-1-chris@chris-wilson.co.uk

XXX pullup

branches: 1.5.10; 1.5.16; 1.5.18;
Pass cache-related flags through to the GTT on pre-SNB devices.

I had assumed for ages this would increase the amount of caching and
thereby increase the chance of stale caches leading to rendering
glitches. But apparently I was wrong, and failing to pass these
through was causing all sorts of problems!

branches: 1.4.2; 1.4.4; 1.4.6;
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.

Implement intel_gtt_chipset_flush correctly.

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

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

i915: Make intel_gtt_get prototype match Linux.

i915: Figure out the graphics stolen memory base.

i915: more work

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>

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>

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

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>

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

Use the offset in mmap calculations.

It fixes wsfb(4) when used with intelfb(4).

Ok riastradh@

XXX pullup-10

branches: 1.24.4;
drm: Simplify async framebuffer child task attachment logic.

No need for scheduling the task to fail. Nix needless error
branches.

drm: Fix asynchronous attach logic for drm framebuffer devices.

The driver-specific sc_scheduled boolean flag was racy. It is
necessary to use config_pending_incr/decr to hold up console
selection and userland start (and block detach, not that that works
yet) until the drmfb attach has completed.

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>

Have i915_vma_pin_iomap map us the framebuffer

Rather than passing bus space arguments around. Also fixes build.


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

No more intel_drv.h, probably meant i915_drv.h if not used


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

Remove last users of drmP.h.

ipmi(4), intelfb(4): Use config_pending_incr/decr.

Don't abuse autoconf's private (and incoherent) dv_flags.

branches: 1.17.12;
kern/52229

Correct linebytes (stride) for intelfb(4), which fixes screen
corruption on the following machines:

- MacBookAir5,2 (Ivy Bridge, 13-inch, Mid 2012)
- MacBookAir6,1 (Haswell, 11-inch, Early 2014)

Now, wsdisplay(4) console as well as intel(4) driver of Xorg
works fine on these machines.

Also there's no side effects found on another machine:

- ThinkPad T480s (Kaby Lake R)

XXX
pullup to netbsd-9

branches: 1.16.4;
Use DVF_ATTACH_INPROGRESS.

Use dtrace probes to implement i915_trace.h.

branches: 1.14.14; 1.14.16;
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: 1.13.2;
pmf_device_register returns false on failure, not true

Factor out intelfb_is_vga_console to drmfb_pci.

Factor generic parts of intelfb out into a drmfb.

Adapting radeondrmkmsfb should not be hard.

This will make nouveaufb easier too.

branches: 1.10.2;
Disable VGA plane after vga_cndetach.

Disabling the VGA plane requires access to the VGA registers, which
don't become available until vga_cndetach. Oops!

branches: 1.9.2; 1.9.4; 1.9.6;
no need to call drm_fb_helper_set_config after genfb attaches, as the setmode callback is called by genfb_attach

add a pmf shutdown handler to switch genfb into polling mode

provide a setmode callback for genfb; with this change, vt switching in and out of X works

Use surface_width/height, not fb_width/height.

This matches the sizes the upstream code uses.

Prettify intelfb/radeondrmkmsfb autoconf output.

Remove leftover debugging kludge.

Make ddb enter/exit work while X is running with i915drmkms.

Fix WSDISPLAYIO_GET_BUSID on new intelfb(4).

Need to pass the PCI device, not one of its children (which intelfb
is), to wsdisplayio_busid_pci.

Fixes starting X.

Rework Intel framebuffer attachment.

Turns out genfb_softc must be at the start of its own device_private,
so we can't have it in the middle of a structure, and certainly can't
have potentially multiple genfbs per graphics device. Failing to do
this is why entering ddb didn't work with an i915drmkms console.

Instead of putting a genfb_softc into each drm_fb_helper, put a
device_t into each drm_fb_helper and create a new intelfb(4) device
for each Intel framebuffer. This will be more flexible, too, in case
we want accelerated framebuffers later on. (XXX Need to adapt the
radeon code, which I'm disabling until that happens to avoid breaking
the build.)

While here, defer disabling VGA, vga_cndetach, and genfb_attach until
we're actually ready to do them all together along with the mode
switch. This should reduce the amount of time during which the
screen is blanked, in case it hangs.

i915: Tweak style.

Have i915_vma_pin_iomap map us the framebuffer

Rather than passing bus space arguments around. Also fixes build.


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

Remove last users of drmP.h.

branches: 1.1.2; 1.1.6;
Rework Intel framebuffer attachment.

Turns out genfb_softc must be at the start of its own device_private,
so we can't have it in the middle of a structure, and certainly can't
have potentially multiple genfbs per graphics device. Failing to do
this is why entering ddb didn't work with an i915drmkms console.

Instead of putting a genfb_softc into each drm_fb_helper, put a
device_t into each drm_fb_helper and create a new intelfb(4) device
for each Intel framebuffer. This will be more flexible, too, in case
we want accelerated framebuffers later on. (XXX Need to adapt the
radeon code, which I'm disabling until that happens to avoid breaking
the build.)

While here, defer disabling VGA, vga_cndetach, and genfb_attach until
we're actually ready to do them all together along with the mode
switch. This should reduce the amount of time during which the
screen is blanked, in case it hangs.

i915: Odds & ends.

i915: another whack at it

Rename intel_dsi.c -> vlv_dsi.c