drm(4): Mark PCI interrupt handlers as MP-safe.

drm: Simplify -- remove drm_pci_agp_destroy.

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.

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.

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

drm_dev_alloc now returns ERR_PTR, check for that.


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

delete unused drm_pci_set_unique (entire ioctl removed)

drm/drm_agpsupport.h for drm_agp_init


Author: Maya Rashish <maya@NetBSD.org>

Remove last users of drmP.h.

Remove drm_pcie_get_speed_cap_mask unused stub


Author: Maya Rashish <maya@NetBSD.org>

Move some drm PCI stuff off drmP.h


Author: Maya Rashish <maya@NetBSD.org>

drm_dev_unref -> drm_dev_put


Author: Maya Rashish <maya@NetBSD.org>

missing includes


Author: Maya Rashish <maya@NetBSD.org>

struct pci_dev::pd_drm_dev kludge is no more.

drm_pci_agp_destroy is now in drm_internal.h.

drm_pci_device_is_agp is no more; just test PCI cap.

New stub to reduce diff.

branches: 1.32.6;
Don't try to interpret the second half of a 64-bit BAR as another one.

From msaitoh@.

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.

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

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.

Make drmkms build without drmkms_pci.

Work around broken null bus_dmatag_destroy expansion.

Convert to long for format expedience.

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

Add drm_pci_set_unique.

Define dummy drm_irq_by_busid, legacy function for legacy UMS API.

struct drm_bus abstraction is no more. Cope.

struct drm_driver now has the bus-specific intr establish routine
(called request_irq/free_irq to match Linux's style).

Namespace our private fields in Linux struct pci_device.

Rename drm_core_ioremap -> drm_legacy_ioremap as upstream.

match linux side loading of headers

more bring in OLD drm code to unbreak the build
matching netbsd types etc. to get the build going


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

branches: 1.18.2; 1.18.4;
PR/52409: Avoid panic at resume.

XXX: pullup-8

branches: 1.17.2;
remove checks for failure after memory allocation calls that cannot fail:

kmem_alloc() with KM_SLEEP
kmem_zalloc() with KM_SLEEP
percpu_alloc()
pserialize_create()
psref_class_create()

all of these paths include an assertion that the allocation has not failed,
so callers should not assert that again.

drmkms_pci: Use MSI if available (Take 2)

tested by martin@ and yhardy@

Revert "drmkms_pci: use MSI if available."

Fix PR/51997.

drmkms_pci: use MSI if available.

branches: 1.13.2; 1.13.4;
Replace drm_ioremap by drm_core_ioremap, reducing diff a little.

Simplify drm shenanigans with irrelevant irq numbers.

One last round for nouveau. It links!

Adapt drmkms_pci module initialization to work as builtin, for agp.

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.

branches: 1.8.2;
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: 1.6.2; 1.6.4;
Report the same PCI domain (cf_unit) as everything else in NetBSD.

Destroy the DMA subregion tag on error in drm_pci_detach.

Now that drm_pci_attach calls the driver load function, rather than
deferring it to the caller, the DMA subregion tag can be created
here.

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 pci_intr_string

Merge riastradh-drm2 to HEAD.

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

drm: Go back to using device_unit(pciN) for pci busid.

This isn't arbitrary; it has to match existing logic in userland in
libpciaccess, and until we extend the pci(4) ioctl interface, and
find some way to handle userland compatibility, we can't change it.

drm: Use real pci segment number, not pciN device unit.

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.

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.

Make drmkms build without drmkms_pci.

Sprinkle AGP conditionals where needed.

branches: 1.4.16; 1.4.18;
Adapt drmkms_pci module initialization to work as builtin, for agp.

branches: 1.3.2;
Move all PCI/AGP code back to drmkms_pci module where it belongs.

branches: 1.2.4; 1.2.6; 1.2.8; 1.2.10;
Merge riastradh-drm2 to HEAD.

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

linux/pci.h so we can dereference struct pci_dev

const struct pci_attach_args *const pa = &dev->pdev->pd_pa;


Author: Maya Rashish <maya@NetBSD.org>

Remove last users of drmP.h.

branches: 1.3.2; 1.3.4; 1.3.20;
properly include files so that modules work.

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.

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: 1.17.4;
drm: Propagate CPPFLAGS.drmkms into all the drm drivers.

drm/vmwgfx: First draft. Passes make depend, doesn't build yet.

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.

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.

Make drmkms build without drmkms_pci.

Sprinkle AGP conditionals where needed.

Add some amdgpu goo.

Upstream nixed __OS_HAS_AGP; use CONFIG_AGP instead.

branches: 1.8.16; 1.8.18;
Hook up viadrmums and include it it i386/ALL.

Just use `drm' for DRM/UMS PCI attachments at PCI devices.

It'll make life simpler. No need to distinguish drm from drm2 here
because it's just an interface attribute, no content, and we use it
compatibly: parent provides pci_attach_args.

Make viadrm (UMS) almost build.

Missing part: something needs to issue config_found_ia("drmums_pci"),
and config(5) needs to be told that it will do that. The sensible
approach is for vga_pci to do so. But config(5) doesn't seem to
understand that *only* vga_pci will do that, not all vga.

The old scheme was to tell config(5) that vga would always attach
drm, and actually attach it in vga_pci, and hope that nobody would
ever consider trying to attach non-PCI devices but quietly leave the
token `pci' out of the name of the drm interface attribute so it
didn't look wrong to have it outside dev/pci/files.pci.

Maybe that's still the right thing, as far as anything UMS can be
right.

Factor generic parts of intelfb out into a drmfb.

Adapting radeondrmkmsfb should not be hard.

This will make nouveaufb easier too.

branches: 1.4.2; 1.4.4; 1.4.6;
Add files.nouveau to files.drmkms_pci.

First whack at radeon driver.

No hardware to test yet, but it builds.

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

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