Hi Linus,
Please pull the 'drm-fixes' branch from
ssh://master.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6.git drm-fixes
This fixes two regressions, one in radeon suspend/resume, and one in
965/GM45 suspend/resume. It also removes a major racy piece of code for
doing vblank swaps, it really was never used properly and userspace never
relied on it being there. It also stops the drm binding twice to certain
Intel PCI devices where the second head isn't a real device but a dumb
placeholder.
Intel are working on some interrupt handling issues that we might have
some fixes for quite soon, and the AGP on PAE issue is still outstanding.
Dave.
drivers/gpu/drm/drm_drv.c | 10 +-
drivers/gpu/drm/drm_irq.c | 80 --------
drivers/gpu/drm/drm_lock.c | 9 -
drivers/gpu/drm/drm_stub.c | 1 -
drivers/gpu/drm/i915/i915_dma.c | 16 +-
drivers/gpu/drm/i915/i915_drv.h | 19 +--
drivers/gpu/drm/i915/i915_gem.c | 8 +-
drivers/gpu/drm/i915/i915_irq.c | 383 ++---------------------------------
drivers/gpu/drm/i915/i915_reg.h | 3 +
drivers/gpu/drm/i915/i915_suspend.c | 9 +
drivers/gpu/drm/radeon/radeon_cp.c | 15 +-
drivers/gpu/drm/radeon/radeon_drv.h | 2 +-
include/drm/drmP.h | 5 -
include/drm/drm_pciids.h | 46 ++--
14 files changed, 84 insertions(+), 522 deletions(-)
commit 0baf823a10bd4131f70e9712d1f02de3c247f1df
Author: Keith Packard <[email protected]>
Date: Sat Nov 8 11:44:14 2008 +1000
drm/i915: Move legacy breadcrumb out of the reserved status page area
Addresses in the hardware status page below index 0x20 are reserved for use
by the hardware. The legacy breadcrumb was sitting at index 5. Move it to
index 0x21, and make sure everyone uses the defined value instead of
hard-coded constants.
Signed-off-by: Keith Packard <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
commit afa21e0584f78964c092981fad94e45d38cda249
Author: Dave Airlie <[email protected]>
Date: Tue Nov 11 18:02:12 2008 +1000
drm/i915: Filter pci devices based on PCI_CLASS_DISPLAY_VGA
This fixes hangs on 855-class hardware by avoiding double attachment of the
driver due to the stub second head device having the same pci id as the real
device.
Other DRM drivers probably want this treatment as well, but I'm applying it
just to this one for safety. But we should clean up the drm_pciids.h mess
now so that each driver has its own pci id list header in its own directory.
Lets do that in the next release.
Signed-off-by: Eric Anholt <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
commit 78538bf14995a136c2d9a22159ada49937359119
Author: Dave Airlie <[email protected]>
Date: Tue Nov 11 17:56:16 2008 +1000
drm/radeon: map registers at load time
Now that the radeon driver has suspend/resume functions, it needs to map its
registers at load time or it will likely crash if a suspend operation occurs
before the driver has been initialized.
This patch moves the register mapping code from firstopen to load and makes
the mapping into a _DRM_DRIVER one so that the core won't remove it at
lastclose time.
Fixes (at least partially) kernel bz #11891.
Signed-off-by: Jesse Barnes <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
commit 5d8e6bb7a20b6206e1fe44565efc383a941b81fa
Author: Eric Anholt <[email protected]>
Date: Tue Nov 4 18:36:29 2008 -0800
drm: Remove infrastructure for supporting i915's vblank swapping.
It's not used in any other drivers, and doesn't look like it will be from
drm.git master.
Signed-off-by: Eric Anholt <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
commit bd95e0a4a6bb9485fe35dda62719663f6ceabae1
Author: Eric Anholt <[email protected]>
Date: Tue Nov 4 12:01:24 2008 -0800
i915: Remove racy delayed vblank swap ioctl.
When userland detected that this ioctl was supported (by version number check),
it used it in a racy way -- dispatch delayed swap, wait for vblank, continue
rendering. As there was no mechanism for it to wait for the swap to finish,
sometimes it would render before the swap and garbage would be displayed on
the screen.
By removing the ioctl and returning -EINVAL, userland returns to its previous,
correct rendering path of waiting for a vblank then dispatching a swap. The
only path that could have used this ioctl correctly was page flipping, which
relied on only one client running and emitting wait-for-vblank-before-rendering
in the command stream. That path also falls back correctly, at the performance
cost of not being able to queue up rendering before the flip occurs.
Signed-off-by: Eric Anholt <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
commit d3e74d0237b102d34979015fbf6df02ca4413074
Author: Eric Anholt <[email protected]>
Date: Mon Nov 3 14:46:17 2008 -0800
i915: Don't whine when pci_enable_msi() fails.
This probably just means the chipset doesn't support MSI, which is fine.
Signed-off-by: Eric Anholt <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
commit 6a47baa6ce7e6fb5fed8d1fd0af36a96a4ad133f
Author: Owen Taylor <[email protected]>
Date: Mon Nov 3 14:38:17 2008 -0800
i915: Don't attempt to short-circuit object_wait_rendering by checking domains.
This could return early when reading after writing a buffer, if somebody
had already put it on the flushing list (write domains are 0, but still
active), leading to glReadPixels failure.
Signed-off-by: Eric Anholt <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
commit ad42ca8f4490de06462aee234ea0083cbd8b46aa
Author: Keith Packard <[email protected]>
Date: Sun Nov 2 23:38:20 2008 -0800
i915: Clean up sarea pointers on leavevt
This corresponds to the setup of the sarea pointers in DMA initialization,
though neither is exactly the point at which the sarea is set up or torn down.
Signed-off-by: Keith Packard <[email protected]>
Signed-off-by: Eric Anholt <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
commit 881ee9889c8b98671c5491e43666bf5d4f78a180
Author: Keith Packard <[email protected]>
Date: Sun Nov 2 23:08:44 2008 -0800
i915: Save/restore MCHBAR_RENDER_STANDBY on GM965/GM45
This register is set by the 2D driver to prevent lockups, and so it needs to
be preserved across suspend/resume too. This makes my X200s work.
Signed-off-by: Keith Packard <[email protected]>
Signed-off-by: Eric Anholt <[email protected]>
Signed-off-by: Dave Airlie <[email protected]>
On Tue, 11 Nov 2008 08:15:26 +0000 (GMT) Dave Airlie <[email protected]> wrote:
> commit 78538bf14995a136c2d9a22159ada49937359119
> Author: Dave Airlie <[email protected]>
> Date: Tue Nov 11 17:56:16 2008 +1000
>
> drm/radeon: map registers at load time
>
> Now that the radeon driver has suspend/resume functions, it needs to map its
> registers at load time or it will likely crash if a suspend operation occurs
> before the driver has been initialized.
>
> This patch moves the register mapping code from firstopen to load and makes
> the mapping into a _DRM_DRIVER one so that the core won't remove it at
> lastclose time.
Does this make the below patch obsolete?
> Fixes (at least partially) kernel bz #11891.
A little thing: there are (or used to be) people who troll commits for
bugzilla reports to close off. I've adopted the convention of
indicating bugzilla reports via their full URL to make those efforts
easier and to increase their accuracy.
From: Jiri Slaby <[email protected]>
When the driver is bound to a device and nobody opens the device node, it
will oops on suspend and resume, since it's not mapped and dev_priv->mmio
is NULL.
Signed-off-by: Jiri Slaby <[email protected]>
Cc: David Airlie <[email protected]>
Cc: Rafael J. Wysocki <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
---
drivers/gpu/drm/radeon/radeon_drv.c | 6 ++++++
1 file changed, 6 insertions(+)
diff -puN drivers/gpu/drm/radeon/radeon_drv.c~drm-fix-radeon-suspend-resume-oops drivers/gpu/drm/radeon/radeon_drv.c
--- a/drivers/gpu/drm/radeon/radeon_drv.c~drm-fix-radeon-suspend-resume-oops
+++ a/drivers/gpu/drm/radeon/radeon_drv.c
@@ -56,6 +56,9 @@ static int radeon_suspend(struct drm_dev
{
drm_radeon_private_t *dev_priv = dev->dev_private;
+ if (!dev_priv->mmio)
+ return 0;
+
/* Disable *all* interrupts */
if ((dev_priv->flags & RADEON_FAMILY_MASK) >= CHIP_RS690)
RADEON_WRITE(R500_DxMODE_INT_MASK, 0);
@@ -67,6 +70,9 @@ static int radeon_resume(struct drm_devi
{
drm_radeon_private_t *dev_priv = dev->dev_private;
+ if (!dev_priv->mmio)
+ return 0;
+
/* Restore interrupt registers */
if ((dev_priv->flags & RADEON_FAMILY_MASK) >= CHIP_RS690)
RADEON_WRITE(R500_DxMODE_INT_MASK, dev_priv->r500_disp_irq_reg);
_
On Tue, Nov 11, 2008 at 6:29 PM, Andrew Morton
<[email protected]> wrote:
> On Tue, 11 Nov 2008 08:15:26 +0000 (GMT) Dave Airlie <[email protected]> wrote:
>
>> commit 78538bf14995a136c2d9a22159ada49937359119
>> Author: Dave Airlie <[email protected]>
>> Date: Tue Nov 11 17:56:16 2008 +1000
>>
>> drm/radeon: map registers at load time
>>
>> Now that the radeon driver has suspend/resume functions, it needs to map its
>> registers at load time or it will likely crash if a suspend operation occurs
>> before the driver has been initialized.
>>
>> This patch moves the register mapping code from firstopen to load and makes
>> the mapping into a _DRM_DRIVER one so that the core won't remove it at
>> lastclose time.
>
> Does this make the below patch obsolete?
Yes please drop this, the correct fix is the one Jesse posted + a fix
on top of it I made in the same patch.
>> Fixes (at least partially) kernel bz #11891.
>
> A little thing: there are (or used to be) people who troll commits for
> bugzilla reports to close off. I've adopted the convention of
> indicating bugzilla reports via their full URL to make those efforts
> easier and to increase their accuracy.
>
Cool will try that in future.
Dave.
On Tuesday, November 11, 2008 12:37 am Dave Airlie wrote:
> On Tue, Nov 11, 2008 at 6:29 PM, Andrew Morton
>
> <[email protected]> wrote:
> > On Tue, 11 Nov 2008 08:15:26 +0000 (GMT) Dave Airlie <[email protected]>
wrote:
> >> commit 78538bf14995a136c2d9a22159ada49937359119
> >> Author: Dave Airlie <[email protected]>
> >> Date: Tue Nov 11 17:56:16 2008 +1000
> >>
> >> drm/radeon: map registers at load time
> >>
> >> Now that the radeon driver has suspend/resume functions, it needs to
> >> map its registers at load time or it will likely crash if a suspend
> >> operation occurs before the driver has been initialized.
> >>
> >> This patch moves the register mapping code from firstopen to load
> >> and makes the mapping into a _DRM_DRIVER one so that the core won't
> >> remove it at lastclose time.
> >
> > Does this make the below patch obsolete?
>
> Yes please drop this, the correct fix is the one Jesse posted + a fix
> on top of it I made in the same patch.
Oh geez, yeah that's ugly, thanks for catching it.
Jesse