This was originally introduced to make encoder sharing on radone
easier for userspace, but:
- It is policy and as such belongs into userspace. E.g. personally I'm
fairly annoyed that a flaky cable results in permanent changes of
the desktop layout, so I'll kick out DEs which do this. Worse if the
kernel also tries to be clever.
- It's inconsistent: We only kill disconnected outputs on setCrtc
(which userspace might also call when just changing the
framebuffer), but not when e.g. we receive a hpd event or in the
output poll worker.
- It's unexpected behaviour for the userspace driver, at least in the
intel ddx we've had tons of bugs where the driver fell over and
killed the X session becuase pageflips/vblanks suddenly stopped
working. We've had to fix this by wrapping every single setCrtc int
a big "recover kms state from the kernel again" operation.
- It's suprising for the kernel, too: It took a few mails between Rob,
Matt and me for them to notice that little dragon wreaking havoc
with the universal plane framebuffer refcounting.
- Userspace can cope with it and e.g. Gnome already kills disconnected
outputs and reconfigures the desktop automatically. And since there
have been no regression reports for the i915 change from over 1 year
ago I think all other DEs are also ready.
Note that the lines removed in this patch go back to
drm/kms: add explicit encoder disable function and detach harder.
Unfortunately the patch itself doesn't explain a hole lot about why it
was added ...
Cc: Matt Roper <matthew.d.roper@intel.com> Cc: Rob Clark <robdclark@gmail.com> Cc: Dave Airlie <airlied@redhat.com> Cc: Alex Deucher <alexdeucher@gmail.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
but due to a bit of rebase fail on my side the patch actually merged
put one hunk on the wrong side of a break statement. Fix this up.
Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Cc: Dan Carpenter <dan.carpenter@oracle.com> Cc: Dave Airlie <airlied@redhat.com> Cc: Alex Deucher <alexdeucher@gmail.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Dave Airlie <airlied@redhat.com>
Lespiau, Damien [Fri, 28 Mar 2014 12:31:05 +0000 (12:31 +0000)]
drm: Specify a bit more the DRM_CAP_CURSOR_{WIDTH, HEIGHT} caps
Earlier this week, there was a bit of confusion about those new
capabilities, to the point I think it's better to document the intention
and API contract.
The comment documents the current situation:
- the radeon driver returns the only valid size for the hw
- i915 returns the maximun cursor size
- other drivers fall back to returning 64x64
The common contract is to return a valid cursor size.
Cc: Sagar Kamble <sagar.a.kamble@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Alex Deucher <alexander.deucher@amd.com> Cc: Imre Deak <imre.deak@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
Dave Airlie [Wed, 26 Mar 2014 04:10:06 +0000 (14:10 +1000)]
drm/nouveau: don't suspend/resume display on runtime s/r
This should ensure we don't hit a locking problem when someone
wakes us up via a connector, we should never go into suspend
while the display is on anyways.
Dave Airlie [Mon, 31 Mar 2014 01:29:38 +0000 (11:29 +1000)]
Merge tag 'vmwgfx-next-2014-03-28' of git://people.freedesktop.org/~thomash/linux into drm-next
vmwgfx render-node support and drm + ttm changes it depends upon.
Pull request of 2014-03-28
* tag 'vmwgfx-next-2014-03-28' of git://people.freedesktop.org/~thomash/linux:
drm/vmwgfx: Bump driver minor and date
drm/vmwgfx: Enable render nodes
drm/vmwgfx: Tighten the security around buffer maps
drm/ttm: Add a ttm_ref_object_exists function
drm/vmwgfx: Tighten security around surface sharing v2
drm/vmwgfx: Allow prime fds in the surface reference ioctls
drm/vmwgfx: Drop authentication requirement on UNREF ioctls
drm/vmwgfx: Reinstate and tighten security around legacy master model
drm/vmwgfx: Use a per-device semaphore for reservation protection
drm: Add a function to get the ioctl flags
drm: Protect the master management with a drm_device::master_mutex v3
drm: Remove the minor master list
drm: Improve on minor type helpers v3
drm: Make control nodes master-less v3
drm: Break out ioctl permission check to a separate function v2
drm: Have the crtc code only reference master from legacy nodes v2
Thomas Hellstrom [Wed, 19 Mar 2014 09:45:11 +0000 (10:45 +0100)]
drm/vmwgfx: Tighten security around surface sharing v2
If using legacy (non-prime) surface sharing, only allow surfaces
to be shared between clients with the same master. This will block
malicious clients from peeking at contents at surfaces from other
(possibly vt-switched) masters.
v2:
s/legacy_client/primary_client/
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> Reviewed-by: Brian Paul <brianp@vmware.com>
Thomas Hellstrom [Tue, 18 Mar 2014 14:00:56 +0000 (15:00 +0100)]
drm/vmwgfx: Allow prime fds in the surface reference ioctls
Allow prime fds and at the same time block legacy handles for render-nodes
in the surface reference ioctls. This means these ioctls can be used
directly from prime-aware clients, and that they can be called from
render-nodes.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> Reviewed-by: Brian Paul <brianp@vmware.com>
Thomas Hellstrom [Thu, 27 Feb 2014 12:24:17 +0000 (13:24 +0100)]
drm/vmwgfx: Drop authentication requirement on UNREF ioctls
These ioctls will anyway only succeed if the client previously opened
referenced the object. Furthermore, closing the client would implicitly
execute the same action. This prevents clients from blocking on UNREF if
their master dropped, and will allow masters to UNREF after dropping
master privileges.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> Reviewed-by: Brian Paul <brianp@vmware.com>
Thomas Hellstrom [Thu, 27 Feb 2014 11:56:08 +0000 (12:56 +0100)]
drm/vmwgfx: Reinstate and tighten security around legacy master model
The following restrictions affect clients connecting using legacy nodes:
*) Masters that have dropped master privilieges are not considered
authenticated until they regain master privileges.
*) Clients whose master have dropped master privileges block interruptibly on
ioctls requiring authentication until their master regains master
privileges. If their master exits, they are killed.
This is primarily designed to prevent clients authenticated with one master to
access data from clients authenticated with another master.
(Think fast user-switching or data sniffers enabled while X is vt-switched).
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> Reviewed-by: Brian Paul <brianp@vmware.com>
Thomas Hellstrom [Thu, 27 Feb 2014 11:34:51 +0000 (12:34 +0100)]
drm/vmwgfx: Use a per-device semaphore for reservation protection
Don't use a per-master semaphore (ttm lock) for reservation protection, but
rather a per-device semaphore. This is needed since clients connecting using
render nodes aren't master aware.
The ttm lock used should probably be replaced with a reader-write semaphore
once the function down_xx_interruptible() is available.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> Reviewed-by: Brian Paul <brianp@vmware.com>
Thomas Hellstrom [Tue, 25 Feb 2014 18:57:44 +0000 (19:57 +0100)]
drm: Protect the master management with a drm_device::master_mutex v3
The master management was previously protected by the drm_device::struct_mutex.
In order to avoid locking order violations in a reworked dropped master
security check in the vmwgfx driver, break it out into a separate master_mutex.
Locking order is master_mutex -> struct_mutex.
Also remove drm_master::blocked since it's not used.
v2: Add an inline comment about what drm_device::master_mutex is protecting.
v3: Remove unneeded struct_mutex locks. Fix error returns in
drm_setmaster_ioctl().
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> Reviewed-by: Brian Paul <brianp@vmware.com> Reviewed-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Daniel Vetter <daniel@ffwll.ch>
Thomas Hellstrom [Wed, 19 Feb 2014 13:21:48 +0000 (14:21 +0100)]
drm: Make control nodes master-less v3
Like for render-nodes, there is no point in maintaining the master concept
for control nodes, so set the struct drm_file::master pointer to NULL.
At the same time, make sure DRM_MASTER | DRM_CONTROL_ALLOW ioctls are always
allowed when called through the control node. Previously the caller also
needed to be master.
v2: Adapt to refactoring of ioctl permission check.
v3: Formatting of logical expression. Use drm_is_control_client() instead of
drm_is_control().
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> Reviewed-by: Brian Paul <brianp@vmware.com> Reviewed-by: David Herrmann <dh.herrmann@gmail.com>
Dave Airlie [Fri, 28 Mar 2014 03:03:48 +0000 (13:03 +1000)]
drm/i915: fix build warning on 32-bit (v2)
/ssd/git/drm-next/drivers/gpu/drm/i915/i915_cmd_parser.c: In function ‘i915_parse_cmds’:
/ssd/git/drm-next/drivers/gpu/drm/i915/i915_cmd_parser.c:405:4: warning: format ‘%ld’ expects argument of type ‘long int’, but argument 5 has type ‘int’ [-Wformat=]
DRM_DEBUG_DRIVER("CMD: Command length exceeds batch length: 0x%08X length=%d batchlen=%ld\n",
^
That first comma looks weird as we already have ']' as a separator.
Remove it.
If anyone sees this commit message and also thinks that auth=1 isn't the
most useful info to have here, let's just say I'd happily review a patch
removing it. If I don't get annoyed enough to submit a patch, that is.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
Lespiau, Damien [Mon, 24 Mar 2014 15:53:16 +0000 (15:53 +0000)]
drm: drm_ut_debug_printk() isn't called with NULL anywmore
The DRM_LOG* macros where the only sites where drm_ut_debug_printk was
called with NULL arguments for prefix and function_name. Now that they
are gone, we can remove that case.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
Lespiau, Damien [Mon, 24 Mar 2014 15:53:15 +0000 (15:53 +0000)]
drm: Pull the test on drm_debug in the logging macros
In the logging code, we are currently checking is we need to output in
drm_ut_debug_printk(). This is too late. The problem is that when we write
something like:
DRM_DEBUG_DRIVER("ELD on [CONNECTOR:%d:%s], [ENCODER:%d:%s]\n",
connector->base.id,
drm_get_connector_name(connector),
connector->encoder->base.id,
drm_get_encoder_name(connector->encoder));
We start by evaluating the arguments (so call drm_get_connector_name() and
drm_get_connector_name()) before ending up in drm_ut_debug_printk() which will
then does nothing.
This means we execute a lot of instructions (drm_get_connector_name(), in turn,
calls snprintf() for example) to happily discard them in the normal case,
drm.debug=0.
So, let's put the test on drm_debug earlier, in the macros themselves.
Sprinkle an unlikely() as well for good measure.
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
Lespiau, Damien [Mon, 24 Mar 2014 15:53:13 +0000 (15:53 +0000)]
staging: imx-drm: Replace DRM_LOG_KMS() by DRM_DEBUG_KMS()
There are only a few users of the DRM_LOG_KMS() macro. We can simplify
the DRM code a bit by replacing them by DRM_DEBUG_KMS().
Cc: Philipp Zabel <p.zabel@pengutronix.de> Cc: Lucas Stach <l.stach@pengutronix.de> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Acked-by: Philipp Zabel <p.zabel@pengutronix.de> Signed-off-by: Dave Airlie <airlied@redhat.com>
Lespiau, Damien [Mon, 24 Mar 2014 15:53:08 +0000 (15:53 +0000)]
drm: Refresh the explanation of debug categories
That comment wasn't super-readable, so I tried to improve it:
- Put the comment before the values it's documenting
- Add a mention to PRIME
- Reword things a bit to be a lighter read
- Add a note about the option to set the debug value at run-time
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
Arnd Bergmann [Tue, 25 Mar 2014 11:06:46 +0000 (12:06 +0100)]
drm/bridge: PTN3460 needs DRM_KMS_HELPER
The recently added PTN3460 device driver uses interfaces that
are provided by the KMS helper infrastructure, so we should
explicitly select that to avoid this linker error:
Dave Airlie [Wed, 26 Mar 2014 04:34:16 +0000 (14:34 +1000)]
Merge branch 'drm-nouveau-next' of git://anongit.freedesktop.org/git/nouveau/linux-2.6 into drm-next
- first stage of (ongoing) gpu fault recovery work
- initial support for maxwell (binary driver fw needed)
- various random fixes across the board
* 'drm-nouveau-next' of git://anongit.freedesktop.org/git/nouveau/linux-2.6: (87 commits)
drm/nouveau: fix missing newline
drm/nouveau/bios: fetch the vbios from PROM using only aligned 32-bit accesses
drm/nouveau/therm: let the vbios decide on the automatic fan management mode
drm/nvd7/therm: handle another kind of PWM fans
drm/nouveau/pm/fan: drop the fan lock in fan_update() before rescheduling
drm/nouveau: fix small thinko in vblank timestamping.
drm/nouveau/therm: check for sensor presence with requested mode, not current
drm/nouveau/disp/dp: allow 540MHz data rate
drm/nouveau: recognise higher link rate for available dp bw calculations
drm/nouveau/disp: limit dp capabilities as per dcb
drm/nva3/fbram: restrict training pattern setup to GT218
drm/nva3/devinit: restrict script access to some PFB regs
drm/nouveau/devinit: add interface to check if a mmio access by scripts is ok
drm/nouveau/bios: have strap reads show on devinit spam debug level
drm/nv50/gpio: fixup reset for gpios >= 16
drm/nv50/gpio: exclude sense value from mask when changing registers
drm/gk104/gr: therm magic needed on some kepler boards
drm/gm107/gr: initial support
drm/gf100-/gf: fix a stupid typo, waiting on wrong signal for mmctx
drm/nouveau/bios: parsing of some random table needed to bring up gr
...
Martin Peres [Wed, 19 Feb 2014 00:04:56 +0000 (01:04 +0100)]
drm/nouveau/therm: let the vbios decide on the automatic fan management mode
This should fix automatic fan management on fermi cards who do not have
0x46 entries in the thermal table.
On my nve6, the blob sets the default linear range from 40°C to 100°C
but my nvcf's default values are 40°C to 85°C. Let's keep 85 as a default
for everyone.
Signed-off-by: Martin Peres <martin.peres@labri.fr> Tested-by: Timothée Ravier <tim@siosm.fr> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Martin Peres [Thu, 13 Mar 2014 23:26:52 +0000 (00:26 +0100)]
drm/nouveau/pm/fan: drop the fan lock in fan_update() before rescheduling
This should fix a deadlock that has been reported to us where fan_update()
would hold the fan lock and try to grab the alarm_program_lock to reschedule
an update. On an other CPU, the alarm_program_lock would have been taken
before calling fan_update(), leading to a deadlock.
We should Cc: <stable@vger.kernel.org> # 3.9+
Reported-by: Marcin Slusarz <marcin.slusarz@gmail.com> Tested-by: Timothée Ravier <tim@siosm.fr> Tested-by: Boris Fersing (IRC nick fersingb, no public email address) Signed-off-by: Martin Peres <martin.peres@free.fr> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Mario Kleiner [Wed, 19 Mar 2014 09:02:11 +0000 (10:02 +0100)]
drm/nouveau: fix small thinko in vblank timestamping.
*hpos horizontal scanout position doesn't need to be corrected
to count the pixels between hactive end and htotal negative.
That is only needed for *vpos to count lines until end of
vblank for the vblank timestamping.
Use hpos as is without correction.
Removes occassional spikes in timestamps of up to 1 scanline
duration, thereby improves accuracy to about +/- 2 usecs instead
of +/- 12 usecs, wrt. true onset time as measured with high
precision equipment on NV-A5.
Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Ilia Mirkin [Wed, 19 Mar 2014 14:45:55 +0000 (10:45 -0400)]
drm/nouveau/disp/dp: allow 540MHz data rate
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=76319 Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Ben Skeggs [Mon, 3 Mar 2014 06:18:56 +0000 (16:18 +1000)]
drm/nouveau: fix fbcon not being accelerated after suspend
This does *not* (and is not intended to) fix the issue reported by
Christoph Rudorff on the nouveau mailinglist.
The patch proposed (which is similar to this one, but also reorders
whether we disable accel or call fb_set_suspend first), papers over
another problem entirely by avoiding touching the framebuffer.
Ben Skeggs [Wed, 26 Feb 2014 00:14:15 +0000 (10:14 +1000)]
drm/nouveau/bios: add more checks to PRAMIN image fetching
Prevents an attempt to access VRAM on an un-posted board, which, on a
particular system with a GRID K1 installed, causes a MCE and chokes
the entire system.
Upcoming mobile Kepler GPUs (such as GK20A) use the platform bus instead
of PCI to which Nouveau is tightly dependent. This patch allows Nouveau
to handle platform devices by:
- abstracting PCI-dependent functions that were typically used for
resource querying and page mapping,
- introducing a nv_device_is_pci() function that allows to make
PCI-dependent code conditional,
- providing a nouveau_drm_platform_probe() function that takes a GPU
platform device to be probed.
Core code as well as engine/subdev drivers are updated wherever possible
to make use of these functions. Some older drivers are too dependent on
PCI to be properly updated, but all newer code on which future chips may
depend should at least be runnable with platform devices.
Signed-off-by: Alexandre Courbot <acourbot@nvidia.com> Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Dave Airlie [Wed, 26 Mar 2014 04:00:53 +0000 (14:00 +1000)]
Merge branch 'drm-next-3.15' of git://people.freedesktop.org/~deathsimple/linux into drm-next
this is the third pull request for 3.15 radeon changes. Highlights this time:
- More DP work from Alex, especially making use of the new DP aux helpers
- Marek's 1D and linear tiling fixes for CIK
* 'drm-next-3.15' of git://people.freedesktop.org/~deathsimple/linux:
drm/radeon: set PIPE_CONFIG for 1D and linear tiling modes on CIK
drm/radeon: use drm_dp_dpcd_read_link_status()
drm/radeon: use the new drm helpers for dp aux
drm/dp: make aux retries less chatty
drm/radeon: clarify special handling in i2c over aux
drm/radeon/atom: rework encoder enable/disable sequence
drm/radeon/dp: move sink power control to a separate function
drm/radeon/dp: use i2c_get_adapdata rather than casting