- Dec 14, 2015
-
-
Daniel Vetter authored
Since David Herrmann's mmap vma manager rework we don't need to grab dev->struct_mutex any more to prevent races when looking up the mmap offset. Drop it and instead don't forget to use the unref_unlocked variant (since the drm core still cares). v2: Finally get rid of the copypasta from another commit in this commit message. And convert to _unlocked like we need to (Patrik). Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com> Cc: Thierry Reding <thierry.reding@gmail.com> Acked-by:
Thierry Reding <thierry.reding@gmail.com> Reviewed-by:
Patrik Jakobsson <patrik.r.jakobsson@gmail.com> Signed-off-by:
Daniel Vetter <daniel.vetter@intel.com> Signed-off-by:
Thierry Reding <treding@nvidia.com>
-
Daniel Vetter authored
For drm_gem_object_unreference callers are required to hold dev->struct_mutex, which these paths don't. Enforcing this requirement has become a bit more strict with commit ef4c6270 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Thu Oct 15 09:36:25 2015 +0200 drm/gem: Check locking in drm_gem_object_unreference Cc: Thierry Reding <thierry.reding@gmail.com> Signed-off-by:
-
Thierry Reding authored
Use the new multi-driver module helpers to get rid of some boilerplate in the module initialization and cleanup functions. Signed-off-by:
-
- Dec 04, 2015
-
-
Alex Deucher authored
commit 4dfd6486 "drm: Use vblank timestamps to guesstimate how many vblanks were missed" introduced in Linux 4.4-rc1 makes the drm core more fragile to drivers which don't update hw vblank counters and vblank timestamps in sync with firing of the vblank irq and essentially at leading edge of vblank. This exposed a problem with radeon-kms/amdgpu-kms which do not satisfy above requirements: The vblank irq fires a few scanlines before start of vblank, but programmed pageflips complete at start of vblank and vblank timestamps update at start of vblank, whereas the hw vblank counter increments only later, at start of vsync. This leads to problems like off by one errors for vblank counter updates, vblank counters apparently going backwards or vblank timestamps apparently having time going backwards. The net result is stuttering of graphics in games, or little hangs, as well as total failure of timing sensitive applications. See bug #93147 for an example of the regression on Linux 4.4-rc: https://bugs.freedesktop.org/show_bug.cgi?id=93147 This patch tries to align all above events better from the viewpoint of the drm core / of external callers to fix the problem: 1. The apparent start of vblank is shifted a few scanlines earlier, so the vblank irq now always happens after start of this extended vblank interval and thereby drm_update_vblank_count() always samples the updated vblank count and timestamp of the new vblank interval. To achieve this, the reporting of scanout positions by radeon_get_crtc_scanoutpos() now operates as if the vblank starts radeon_crtc->lb_vblank_lead_lines before the real start of the hw vblank interval. This means that the vblank timestamps which are based on these scanout positions will now update at this earlier start of vblank. 2. The driver->get_vblank_counter() function will bump the returned vblank count as read from the hw by +1 if the query happens after the shifted earlier start of the vblank, but before the real hw increment at start of vsync, so the counter appears to increment at start of vblank in sync with the timestamp update. 3. Calls from vblank irq-context and regular non-irq calls are now treated identical, always simulating the shifted vblank start, to avoid inconsistent results for queries happening from vblank irq vs. happening from drm_vblank_enable() or vblank_disable_fn(). 4. The radeon_flip_work_func will delay mmio programming a pageflip until the start of the real vblank iff it happens to execute inside the shifted earlier start of the vblank, so pageflips now also appear to execute at start of the shifted vblank, in sync with vblank counter and timestamp updates. This to avoid some races between updates of vblank count and timestamps that are used for swap scheduling and pageflip execution which could cause pageflips to execute before the scheduled target vblank. The lb_vblank_lead_lines "fudge" value is calculated as the size of the display controllers line buffer in scanlines for the given video mode: Vblank irq's are triggered by the line buffer logic when the line buffer refill for a video frame ends, ie. when the line buffer source read position enters the hw vblank. This means that a vblank irq could fire at most as many scanlines before the current reported scanout position of the crtc timing generator as the number of scanlines the line buffer can maximally hold for a given video mode. This patch has been successfully tested on a RV730 card with DCE-3 display engine and on a evergreen card with DCE-4 display engine, in single-display and dual-display configuration, with different video modes. A similar patch is needed for amdgpu-kms to fix the same problem. Limitations: - Maybe replace the udelay() in the flip_work_func() by a suitable usleep_range() for a bit better efficiency? Will try that. - Line buffer sizes in pixels are hard-coded on < DCE-4 to a value i just guessed to be high enough to work ok, lacking info on the true sizes atm. Probably fixes: fdo#93147 Port of Mario's radeon fix to amdgpu. Signed-off-by:
Alex Deucher <alexander.deucher@amd.com> (v1) Reviewed-by:
Mario Kleiner <mario.kleiner.de@gmail.com> (v2) Refine amdgpu_flip_work_func() for better efficiency. In amdgpu_flip_work_func, replace the busy waiting udelay(5) with event lock held by a more performance and energy efficient usleep_range() until at least predicted true start of hw vblank, with some slack for scheduler happiness. Release the event lock during waits to not delay other outputs in doing their stuff, as the waiting can last up to 200 usecs in some cases. Also small fix to code comment and formatting in that function. (v2) Signed-off-by:
-
Mario Kleiner authored
commit 4dfd6486 "drm: Use vblank timestamps to guesstimate how many vblanks were missed" introduced in Linux 4.4-rc1 makes the drm core more fragile to drivers which don't update hw vblank counters and vblank timestamps in sync with firing of the vblank irq and essentially at leading edge of vblank. This exposed a problem with radeon-kms/amdgpu-kms which do not satisfy above requirements: The vblank irq fires a few scanlines before start of vblank, but programmed pageflips complete at start of vblank and vblank timestamps update at start of vblank, whereas the hw vblank counter increments only later, at start of vsync. This leads to problems like off by one errors for vblank counter updates, vblank counters apparently going backwards or vblank timestamps apparently having time going backwards. The net result is stuttering of graphics in games, or little hangs, as well as total failure of timing sensitive applications. See bug #93147 for an example of the regression on Linux 4.4-rc: https://bugs.freedesktop.org/show_bug.cgi?id=93147 This patch tries to align all above events better from the viewpoint of the drm core / of external callers to fix the problem: 1. The apparent start of vblank is shifted a few scanlines earlier, so the vblank irq now always happens after start of this extended vblank interval and thereby drm_update_vblank_count() always samples the updated vblank count and timestamp of the new vblank interval. To achieve this, the reporting of scanout positions by radeon_get_crtc_scanoutpos() now operates as if the vblank starts radeon_crtc->lb_vblank_lead_lines before the real start of the hw vblank interval. This means that the vblank timestamps which are based on these scanout positions will now update at this earlier start of vblank. 2. The driver->get_vblank_counter() function will bump the returned vblank count as read from the hw by +1 if the query happens after the shifted earlier start of the vblank, but before the real hw increment at start of vsync, so the counter appears to increment at start of vblank in sync with the timestamp update. 3. Calls from vblank irq-context and regular non-irq calls are now treated identical, always simulating the shifted vblank start, to avoid inconsistent results for queries happening from vblank irq vs. happening from drm_vblank_enable() or vblank_disable_fn(). 4. The radeon_flip_work_func will delay mmio programming a pageflip until the start of the real vblank iff it happens to execute inside the shifted earlier start of the vblank, so pageflips now also appear to execute at start of the shifted vblank, in sync with vblank counter and timestamp updates. This to avoid some races between updates of vblank count and timestamps that are used for swap scheduling and pageflip execution which could cause pageflips to execute before the scheduled target vblank. The lb_vblank_lead_lines "fudge" value is calculated as the size of the display controllers line buffer in scanlines for the given video mode: Vblank irq's are triggered by the line buffer logic when the line buffer refill for a video frame ends, ie. when the line buffer source read position enters the hw vblank. This means that a vblank irq could fire at most as many scanlines before the current reported scanout position of the crtc timing generator as the number of scanlines the line buffer can maximally hold for a given video mode. This patch has been successfully tested on a RV730 card with DCE-3 display engine and on a evergreen card with DCE-4 display engine, in single-display and dual-display configuration, with different video modes. A similar patch is needed for amdgpu-kms to fix the same problem. Limitations: - Line buffer sizes in pixels are hard-coded on < DCE-4 to a value i just guessed to be high enough to work ok, lacking info on the true sizes atm. Fixes: fdo#93147 Signed-off-by:
Dave Witbrodt <dawitbro@sbcglobal.net> (v2) Refine radeon_flip_work_func() for better efficiency: In radeon_flip_work_func, replace the busy waiting udelay(5) with event lock held by a more performance and energy efficient usleep_range() until at least predicted true start of hw vblank, with some slack for scheduler happiness. Release the event lock during waits to not delay other outputs in doing their stuff, as the waiting can last up to 200 usecs in some cases. Retested on DCE-3 and DCE-4 to verify it still works nicely. (v2) Signed-off-by:
-
Lyude authored
HPD signals on DVI ports can be fired off before the pins required for DDC probing actually make contact, due to the pins for HPD making contact first. This results in a HPD signal being asserted but DDC probing failing, resulting in hotplugging occasionally failing. This is somewhat rare on most cards (depending on what angle you plug the DVI connector in), but on some cards it happens constantly. The Radeon R5 on the machine used for testing this patch for instance, runs into this issue just about every time I try to hotplug a DVI monitor and as a result hotplugging almost never works. Rescheduling the hotplug work for a second when we run into an HPD signal with a failing DDC probe usually gives enough time for the rest of the connector's pins to make contact, and fixes this issue. Reviewed-by:
Christian König <christian.koenig@amd.com> Signed-off-by:
Lyude <cpaul@redhat.com> Signed-off-by:
-
jimqu authored
there is a protection fault about freed list when OCL test. add a spin lock to protect it. v2: drop changes in vm_fini Signed-off-by:
JimQu <jim.qu@amd.com> Reviewed-by:
-
Christian König authored
The gtt_end is already inclusive, we don't need to subtract one here. v2 (chk): keep the fix for the VM code, cause here it really applies. Signed-off-by:
Anatoli Antonovitch <anatoli.antonovitch@amd.com> Reviewed-by:
-
Christian König authored
No need for a GEM reference here. Reviewed-by:
Michel Dänzer <michel.daenzer@amd.com> Signed-off-by:
-
Christian König authored
No need for the GEM reference here. Reviewed-by:
-
Christian König authored
Not necessary for VRAM. v2: no need to check if ttm is NULL. Signed-off-by:
-
Jani Nikula authored
We've had human readable connector status change debug logging since commit ed7951dc Author: Lespiau, Damien <damien.lespiau@intel.com> Date: Fri May 10 12:36:42 2013 +0000 drm: Make the HPD status updates debug logs more readable but commit 162b6a57 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Wed Jan 21 08:45:21 2015 +0100 drm/probe-helper: don't lose hotplug event added a new one with just the numbers. Fix it. Signed-off-by:
Jani Nikula <jani.nikula@intel.com> Signed-off-by:
Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/1449144003-2877-1-git-send-email-jani.nikula@intel.com
-
Daniel Vetter authored
Apparently pre-nv50 pageflip events happen before the actual vblank period. Therefore that functionality got semi-disabled in commit af4870e4 Author: Mario Kleiner <mario.kleiner.de@gmail.com> Date: Tue May 13 00:42:08 2014 +0200 drm/nouveau/kms/nv04-nv40: fix pageflip events via special case. Unfortunately that hack got uprooted in commit cc1ef118 Author: Thierry Reding <treding@nvidia.com> Date: Wed Aug 12 17:00:31 2015 +0200 drm/irq: Make pipe unsigned and name consistent Triggering a warning when trying to sample the vblank timestamp for a non-existing pipe. There's a few ways to fix this: - Open-code the old behaviour, which just enshrines this slight breakage of the userspace ABI. - Revert Mario's commit and again inflict broken timestamps, again not pretty. - Fix this for real by delaying the pageflip TS until the next vblank interrupt, thereby making it accurate. This patch implements the third option. Since having a page flip interrupt that happens when the pageflip gets armed and not when it completes in the next vblank seems to be fairly common (older i915 hw works very similarly) create a new helper to arm vblank events for such drivers. v2 (Mario Kleiner): - Fix function prototypes in drmP.h - Add missing vblank_put() for pageflip completion without pageflip event. - Initialize sequence number for queued pageflip event to avoid trouble in drm_handle_vblank_events(). - Remove dead code and spelling fix. v3 (Mario Kleiner): - Add a signed-off-by and cc stable tag per Ilja's advice. v4 (Thierry Reding): - Fix kerneldoc typo, discovered by Michel Dänzer - Rearrange tags and changelog Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=106431 Cc: Thierry Reding <treding@nvidia.com> Cc: Mario Kleiner <mario.kleiner.de@gmail.com> Acked-by:
Ben Skeggs <bskeggs@redhat.com> Cc: Ilia Mirkin <imirkin@alum.mit.edu> Signed-off-by:
Dave Airlie <airlied@redhat.com>
-
Thomas Hellstrom authored
A client calling drmSetMaster() using a file descriptor that was opened when another client was master would inherit the latter client's master object and all its authenticated clients. This is unwanted behaviour, and when this happens, instead allocate a brand new master object for the client calling drmSetMaster(). Fixes a BUG() throw in vmw_master_set(). Cc: <stable@vger.kernel.org> Signed-off-by:
Thomas Hellstrom <thellstrom@vmware.com> Signed-off-by:
-
- Dec 03, 2015
-
-
Daniel Vetter authored
This can happen when we run out of encoders for a multi-crtc modeset, or also when userspace is silly and tries to clone multiple connectors that need the same encoder on the same crtc. Reported-and-Tested-and-Reviewed-by:
Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Signed-off-by:
-
- Dec 02, 2015
-
-
Thierry Reding authored
Provide subsystem-level suspend and resume helpers that can be used to implement suspend/resume on atomic mode-setting enabled drivers. v2: simplify locking, enhance kerneldoc comments v3: pass lock acquisition context by parameter, improve kerneldoc v4: - remove redundant code (already provided by atomic helpers) (Maarten Lankhorst) - move backoff dance from drm_modeset_lock_all_ctx() into suspend helper (Daniel Vetter) v5: handle potential EDEADLK from drm_atomic_helper_duplicate_state() and drm_atomic_helper_disable_all() (Daniel Vetter) Signed-off-by: -
Thierry Reding authored
This function is like drm_modeset_lock_all(), but it takes the lock acquisition context as a parameter rather than storing it in the DRM device's mode_config structure. Implement drm_modeset_{,un}lock_all() in terms of the new function for better code reuse, and add a note to the kerneldoc that new code should use the new functions. v2: improve kerneldoc v4: rename drm_modeset_lock_all_crtcs() to drm_modeset_lock_all_ctx() and take mode_config's .connection_mutex instead of .mutex lock to avoid lock inversion (Daniel Vetter), use drm_modeset_drop_locks() which is now the equivalent of drm_modeset_unlock_all_ctx() v5: do not take the dev->mode_config.connection_mutex in drm_atomic_legacy_backoff() since drm_modeset_lock_all_ctx() already keeps it, enhance kerneldoc for drm_modeset_lock_all_ctx() (Daniel Vetter) Signed-off-by: -
Nicolai Hähnle authored
As soon as we leave the spinlock after the job has been added to the job queue, we can no longer rely on the job's data to be available. I have seen a null-pointer dereference due to sched == NULL in amd_sched_wakeup via amd_sched_entity_push_job and amd_sched_ib_submit_kernel_helper. Since the latter initializes sched_job->sched with the address of the ring scheduler, which is guaranteed to be non-NULL, this race appears to be a likely culprit. Signed-off-by:
Nicolai Hähnle <nicolai.haehnle@amd.com> Bugzilla: https://bugs.freedesktop.org/attachment.cgi?bugid=93079 Reviewed-by:
-
Chunming Zhou authored
Missing error check if the operation failed. Signed-off-by:
Chunming Zhou <David1.Zhou@amd.com> Reviewed-by:
-
Imre Deak authored
There are platforms that don't need the full GMBUS power domain (BXT) while others do (PCH, VLV/CHV). For optimizing this we would need to add a new power domain, but it's not clear how much we would benefit given the short time we hold the reference. So for now let's keep things simple. This fixes a regression introduced in commit 237ed86c Author: Sonika Jindal <sonika.jindal@intel.com> Date: Tue Sep 15 09:44:20 2015 +0530 drm/i915: Check live status before reading edid v2: - fix commit message, PCH won't take any redundant power resource after this change (Ville) Reviewed-by:
Ville Syrjälä <ville.syrjala@linux.intel.com> [fix commit message in v2 (Imre)] [Cherry-picked from drm-intel-next-queued 29bb94bb (Imre)] Signed-off-by:
Imre Deak <imre.deak@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1448643329-18675-6-git-send-email-imre.deak@intel.com Signed-off-by:
-
Imre Deak authored
MISSING_CASE() would have been useful to track down a recent problem in intel_display_port_aux_power_domain(), so add it there and a few related helpers. This was also suggested by Ville in his review of the latest DMC/DC changes, we forgot to address that. Reviewed-by:
Patrik Jakobsson <patrik.jakobsson@linux.intel.com> [Cherry-picked from drm-intel-next-queued b9fec167 (Imre)] Signed-off-by:
-
Imre Deak authored
Due to the current sharing of the DDI encoder between DP and HDMI connectors we can run the DP detection after the HDMI detection has already set the shared encoder's type. For now solve this keeping the current behavior and running the detection in this case too. For a proper solution Ville suggested to split the encoder into an HDMI and DP one, that can be done as a follow-up. This issue triggers the WARN in intel_display_port_aux_power_domain() and was introduced in: commit 25f78f58 Author: Ville Syrjälä <ville.syrjala@linux.intel.com> Date: Mon Nov 16 15:01:04 2015 +0100 drm/i915: Clean up AUX power domain handling CC: Patrik Jakobsson <patrik.jakobsson@linux.intel.com> CC: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by:
-
Ville Syrjälä authored
Currently the gmbus code uses intel_aux_display_runtime_get/put in an effort to make sure the hardware is powered up sufficiently for gmbus. That function only takes the runtime PM reference which on VLV/CHV/BXT is not enough. We need the disp2d/pipe-a well on VLV/CHV and power well 2 on BXT. So add a new power domnain for gmbus and kill off the now unused intel_aux_display_runtime_get/put. And change intel_hdmi_set_edid() to use the gmbus power domain too since that's all we need there. Also toss in a BUILD_BUG_ON() to catch problems if we run out of bits for power domains. We're already really close to the limit... [Patrik: Add gmbus string to debugfs output] Signed-off-by:
-
Ville Syrjälä authored
Introduce intel_display_port_aux_power_domain() which simply returns the appropriate AUX power domain for a specific port, and then replace the intel_display_port_power_domain() with calls to the new function in the DP code. As long as we're not actually enabling the port we don't need the lane power domains, and those are handled now purely from modeset_update_crtc_power_domains(). My initial motivation for this was to see if I could keep the DPIO power wells powered down while doing AUX on CHV, but turns out I can't so this doesn't change anything for CHV at least. But I think it's still a worthwile change. v2: Add case for PORT E. Default to POWER_DOMAIN_AUX_D for now. (Ville) Signed-off-by:
-
Daniel Stone authored
Passing -1 as the pipe for vblank events now triggers a WARN_ON, but had previously made multi-screen unusable anyway. Pass the correct pipe to the event-send function, and use the new API to make this a bit easier for us. Fixes WARN present since cc1ef118 for every pageflip event sent: [ 209.549969] ------------[ cut here ]------------ [ 209.554592] WARNING: CPU: 3 PID: 238 at drivers/gpu/drm/drm_irq.c:924 drm_vblank_count_and_time+0x80/0x88 [drm]() [ 209.564832] Modules linked in: [...] [ 209.612401] CPU: 3 PID: 238 Comm: irq/41-ff940000 Tainted: G W 4.3.0-rc6+ #71 [ 209.620647] Hardware name: Rockchip (Device Tree) [ 209.625348] [<c001bb80>] (unwind_backtrace) from [<c001615c>] (show_stack+0x20/0x24) [ 209.633079] [<c001615c>] (show_stack) from [<c02b2c50>] (dump_stack+0x8c/0x9c) [ 209.640289] [<c02b2c50>] (dump_stack) from [<c0052e88>] (warn_slowpath_common+0x94/0xc4) [ 209.648364] [<c0052e88>] (warn_slowpath_common) from [<c0052f74>] (warn_slowpath_null+0x2c/0x34) [ 209.657139] [<c0052f74>] (warn_slowpath_null) from [<bf17dc30>] (drm_vblank_count_and_time+0x80/0x88 [drm]) [ 209.666875] [<bf17dc30>] (drm_vblank_count_and_time [drm]) from [<bf17e484>] (drm_send_vblank_event+0x74/0x7c [drm]) [ 209.677385] [<bf17e484>] (drm_send_vblank_event [drm]) from [<bf4c1144>] (vop_win_state_complete+0x4c/0x70 [rockchip_drm_vop]) [ 209.688757] [<bf4c1144>] (vop_win_state_complete [rockchip_drm_vop]) from [<bf4c3bdc>] (vop_isr_thread+0x170/0x1d4 [rockchip_drm_vop]) [ 209.700822] [<bf4c3bdc>] (vop_isr_thread [rockchip_drm_vop]) from [<c00ab93c>] (irq_thread_fn+0x2c/0x50) [ 209.710284] [<c00ab93c>] (irq_thread_fn) from [<c00abcac>] (irq_thread+0x13c/0x188) [ 209.717927] [<c00abcac>] (irq_thread) from [<c00723c8>] (kthread+0xec/0x104) [ 209.724965] [<c00723c8>] (kthread) from [<c0011638>] (ret_from_fork+0x14/0x3c) [ 209.732171] ---[ end trace 0690bc604f5d535d ]--- Signed-off-by:
Daniel Stone <daniels@collabora.com> Cc: Sjoerd Simons <sjoerd.simons@collabora.co.uk> Cc: Thierry Reding <treding@nvidia.com> Cc: Heiko Stuebner <heiko@sntech.de> Tested-By:
Sjoerd Simons <sjoerd.simons@collabora.co.uk> Tested-by:
Heiko Stuebner <heiko@sntech.de> Reviewed-by:
-
Luis de Bethencourt authored
This platform driver has a OF device ID table but the OF module alias information is not created so module autoloading won't work. Signed-off-by:Luis de Bethencourt <luisbg@osg.samsung.com>
-
Dominik Behr authored
VOP_WINx_DSP_ST does not require subtracting 1 from the values written to it. It actually causes the screen to be shifted by one pixel. Signed-off-by:
Mark Yao <mark.yao@rock-chips.com> Tested-by:
Yakir Yang <ykk@rock-chips.com> Reviewed-by:
Dominik Behr <dbehr@chromium.org> Signed-off-by:
-
Heiko Stuebner authored
Commit 371f0f08 ("ARM: 8426/1: dma-mapping: add missing range check in dma_mmap()") introduced offset-checking for mappings, which collides with the fake-offset the drm sets for gems. Other drm-drivers set this offset to 0 before doing the mapping, so this looks like the correct way to go for rockchip as well. Fixes: 371f0f08 ("ARM: 8426/1: dma-mapping: add missing range check in dma_mmap()") Signed-off-by:
-
- Dec 01, 2015
-
-
Sjoerd Simons authored
When doing the initial setup both the hclk and the aclk need to be enabled otherwise the board will simply hang. This only occurs when building the vop driver as a module, when its built-in the initial setup happens to run before the clock framework shuts of unused clocks (including the aclk). While there also switch to doing prepare and enable in one step rather then separate steps to reduce the amount of code required. Signed-off-by:
Romain Perier <romain.perier@gmail.com>
-
Daniel Vetter authored
There's currently two places where the gma500 fault handler relies upon dev->struct_mutex: - To protect r->mappping - To make sure vm_insert_pfn isn't called concurrently (in which case the 2nd thread would get an error code). Everything else (specifically psb_gtt_pin) is already protected by some other locks. Hence just create a new driver-private mmap_mutex just for this function. With this gma500 is complete dev->struct_mutex free! Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com> Acked-by:
-
Daniel Vetter authored
Simply forgotten about this when I was doing my general cleansing of simple gem mmap offset functions. There's nothing but core functions called here, and they all have their own protection already. Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com> Acked-by:
-
Daniel Vetter authored
This is init/teardown code, locking is just to appease locking checks. And since gem create/free doesn't need this any more there's really no reason for grabbing dev->struct_mutex. Again important to switch obj_unref to _unlocked variants. Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com> Acked-by:
-
Daniel Vetter authored
It's either init code or already protected by other means. Note that psb_gtt_pin/unpin has it's own lock, and that's really the only piece of driver private state - all the modeset state is protected with modeset locks already. The only important bit is to switch all gem_obj_unref calls to the _unlocked variant. Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com> Acked-by:
-
Daniel Vetter authored
This is called without dev->struct_mutex held, we need to use the _unlocked variant. Never caught in the wild since you'd need an evil userspace which races a gem_close ioctl call with the in-progress open. Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com> Acked-by:
-
Chris Wilson authored
We have relied upon the sole caller (wait_ioctl) validating the timeout argument. However, when waiting for multiple requests I forgot to ensure that the timeout was still positive on the later requests. This is more simply done inside __i915_wait_request. Fixes regression introduced in commit b4716185 Author: Chris Wilson <chris@chris-wilson.co.uk> Date: Mon Apr 27 13:41:17 2015 +0100 drm/i915: Implement inter-engine read-read optimisations The impact of the regression is 1 jiffie for each extra active ring for a wait_ioctl with a timeout -- I don't think anyone has noticed. Signed-off-by:
Chris Wilson <chris@chris-wilson.co.uk> Cc: Lionel Landwerlin <lionel.g.landwerlin@linux.intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by:
-
Ville Syrjälä authored
Rather than using drm_match_cea_mode() to see if the EDID detailed timings are supposed to represent one of the CEA/HDMI modes, add a special version of that function that takes in an explicit clock tolerance value (in kHz). When looking at the detailed timings specify the tolerance as 5kHz due to the 10kHz clock resolution limit inherent in detailed timings. drm_match_cea_mode() uses the normal KHZ2PICOS() matching of clocks, which only allows smaller errors for lower clocks (eg. for 25200 it won't allow any error) and a bigger error for higher clocks (eg. for 297000 it actually matches 296913-297000). So it doesn't really match what we want for the fixup. Using the explicit +-5kHz is much better for this use case. Not sure if we should change the normal mode matching to also use something else besides KHZ2PICOS() since it allows a different proportion of error depending on the clock. I believe VESA CVT allows a maximum deviation of .5%, so using that for normal mode matching might be a good idea? Cc: Adam Jackson <ajax@redhat.com> Tested-by:
<nathan.d.ciobanu@linux.intel.com> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=92217 Fixes: fa3a7340 ("drm/edid: Fix up clock for CEA/HDMI modes specified via detailed timings") Signed-off-by:
Adam Jackson <ajax@redhat.com> Signed-off-by:
-
Marc-André Lureau authored
Avoid the sticky preferred mode bit by using the no-merge version of the function (this allows gnome-shell to resize to lower resolutions than the default resolution) Signed-off-by:
Marc-André Lureau <marcandre.lureau@redhat.com> Cc: stable@vger.kernel.org Signed-off-by:
-
- Nov 30, 2015
-
-
Luis de Bethencourt authored
This platform driver has a OF device ID table but the OF module alias information is not created so module autoloading won't work. Signed-off-by:
Philipp Zabel <p.zabel@pengutronix.de>
-
Pavel Machek authored
Thinkpad T40p needs agpmode 1. Signed-off-by:
Pavel Machek <pavel@ucw.cz> Signed-off-by:
-
Christian König authored
We use the reservation object of the page directory for the page tables as well, because of this the page directory should be freed last. Ensure that by keeping a reference from the page tables to the directory. Signed-off-by:
-
