aboutsummaryrefslogtreecommitdiffstats
path: root/drivers/gpu
Commit message (Collapse)AuthorAgeFilesLines
* i915: suspend/resume interrupt stateJesse Barnes2009-02-231-3/+7
| | | | | | | | | In the KMS case, enter/leavevt won't fix up the interrupt handler for us, so we need to do it at suspend/resume time. Make sure we don't fail the resume if the chip is hung either. Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* Fix an oops in i915_gem_retire_requests()Karsten Wiese2009-02-231-0/+3
| | | | | | | | dev_priv->hw_status_page can be NULL, if i915_gem_retire_requests() is called from i915_gem_busy_ioctl(). Signed-off-by Karsten Wiese <fzu@wemgehoertderstaat.de> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* drm/i915: Add missing mutex_lock(&dev->struct_mutex)Pierre Willenbrock2009-02-231-0/+1
| | | | | | | there might be a nicer way to fix this but this is the simplest for now. Signed-off-by: Pierre Willenbrock <pierre@pirsoft.de> Signed-off-by: Dave Airlie <airlied@redhat.com>
* drm/i915: fix WC mapping in non-GEM i915 code.Jesse Barnes2009-02-231-1/+1
| | | | | | [airlied - taken from mailing list posting] Signed-off-by: Dave Airlie <airlied@redhat.com>
* drm/i915: Fix regression in 95ca9dChris Wilson2009-02-231-2/+6
| | | | | | | | | | The object is dereferenced before the NULL check. Oops. Fixes http://bugs.freedesktop.org/show_bug.cgi?id=20235 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@redhat.com>
* drm/i915: Retire requests from i915_gem_busy_ioctl.Eric Anholt2009-02-231-0/+7
| | | | | | | | | This ensures that the user gets the latest information from the hardware on whether the buffer is busy, potentially reducing the working set of objects that the user chooses. Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@redhat.com>
* drm/i915: suspend/resume GEM when KMS is activeJesse Barnes2009-02-233-2/+24
| | | | | | | | | | In the KMS case, we need to suspend/resume GEM as well. So on suspend, make sure we idle GEM and stop any new rendering from coming in, and on resume, re-init the framebuffer and clear the suspended flag. Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@redhat.com>
* drm/i915: Don't let a device flush to prepare buffers clear new write_domains.Eric Anholt2009-02-231-2/+14
| | | | | | | | | | | | | | | | | | The problem was that object_set_to_gpu_domain would set the new write_domains that are getting set by this batchbuffer, then the accumulated flushes required for all the objects in preparation for this batchbuffer were posted, and the brand new write domain would get cleared by the flush being posted. Instead, hang on to the new (or old if we're not changing it) value and set it after the flush is queued. Results from this noticably included conformance test failures from reads shortly after writes (where the new write domain had been lost and thus not flushed and waited on), but is a suspected cause of hangs in some apps when a write domain is lost on a buffer that gets reused for instruction or commmand state. Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@redhat.com>
* drm/i915: Cut two args to set_to_gpu_domain that confused this tricky path.Eric Anholt2009-02-231-22/+16
| | | | | | | | While not strictly required, it helped while thinking about the following change. This change should be invariant. Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@redhat.com>
* drm/radeon: update sarea copies of last_ variables on resume.etienne2009-02-201-6/+15
| | | | | | | | | | This fixes a regression reported in bug #12613. [airlied: not I tweaked the patch slightly and fixed it by etienne did all the hardwork so gets authorship] Signed-off-by: etienne <etienne.basset@numericable.fr> Signed-off-by: Dave Airlie <airlied@redhat.com>
* drm/i915: Keep refs on the object over the lifetime of vmas for GTT mmap.Jesse Barnes2009-02-203-18/+55
| | | | | | | | | | | This fixes potential fault at fault time if the object was unreferenced while the mapping still existed. Now, while the mmap_offset only lives for the lifetime of the object, the object also stays alive while a vma exists that needs it. Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@redhat.com>
* drm/i915: take struct mutex around fb unrefJesse Barnes2009-02-201-0/+2
| | | | | | | | Need to do this in case the unref ends up doing a free. Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@redhat.com>
* drm: Use spread spectrum when the bios tells us it's ok.Kristian Høgsberg2009-02-203-6/+24
| | | | | | | | Lifted from the DDX modesetting. Signed-off-by: Kristian Høgsberg <krh@redhat.com> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@redhat.com>
* drm: Collapse identical i8xx_clock() and i9xx_clock().Kristian Høgsberg2009-02-201-27/+6
| | | | | | | | They used to be different. Now they're identical. Signed-off-by: Kristian Høgsberg <krh@redhat.com> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@redhat.com>
* drm: Bring PLL limits in sync with DDX values.Kristian Høgsberg2009-02-201-3/+3
| | | | | | Signed-off-by: Kristian Høgsberg <krh@redhat.com> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@redhat.com>
* drm: Add locking around cursor gem operations.Kristian Høgsberg2009-02-201-4/+5
| | | | | | | | | We need to hold the struct_mutex around pinning and the phys object operations. Signed-off-by: Kristian Høgsberg <krh@redhat.com> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@redhat.com>
* drm: Propagate failure from setting crtc base.Chris Wilson2009-02-202-35/+64
| | | | | | | | | Check the error paths within intel_pipe_set_base() to first cleanup and then report back the error. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm: Check for a NULL encoder when reverting on error pathChris Wilson2009-02-201-1/+5
| | | | | | | | | We need to skip the connectors with a NULL encoder to match the success path and avoid an OOPS. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: Cleanup the hws on ringbuffer constrution failure.Chris Wilson2009-02-201-14/+25
| | | | | | | | | If we fail to create the ringbuffer, then we need to cleanup the allocated hws. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: Don't add panel_fixed_mode to the probed modes list at LVDS init.Steve Aarnio2009-02-201-2/+0
| | | | | | | | | | | | | | | | | | | In the case where no EDID data is read from the device, adding the panel_fixed_mode pointer to the probed modes list causes data corruption. If the panel_fixed_mode pointer is added to the probed modes list at init time, a copy of the mode is added again at drm_get_modes() request time. Then, the panel_fixed_mode pointer is freed because it is seen as a duplicate mode. Unfortunately, this pointer is still stored and used in mode_fixup(). Because the panel_fixed_mode data is copied and returned at drm_get_modes() time, it is unnecessary to add this information at init time. Signed-off-by: Steve Aarnio <steve.j.aarnio@intel.com> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm: Release user fbs in drm_releaseKristian Høgsberg2009-02-202-2/+4
| | | | | | | | | Avoids leaking fbs and associated buffers on release. Signed-off-by: Kristian Høgsberg <krh@redhat.com> Tested-by: Tested-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: Unpin the fb on error during construction.Chris Wilson2009-02-201-3/+5
| | | | | | | | If we fail whilst constructing the fb, then we need to unpin it as well. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: Unpin the hws if we fail to kmap.Chris Wilson2009-02-201-0/+1
| | | | | | | | A missing unpin on the error path. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: Unpin the ringbuffer if we fail to ioremap it.Chris Wilson2009-02-201-0/+1
| | | | | | | | A missing unpin on the error path. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: unpin for an invalid memory domain.Chris Wilson2009-02-201-0/+2
| | | | | | | | | A missing unreference and unpin after rejecting the relocation for an invalid memory domain. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: Release and unlock on mmap_gtt error path.Chris Wilson2009-02-201-1/+4
| | | | | | | | We failed to unlock the mutex after failing to create the mmap offset. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: Set framebuffer alignment based upon the fence constraints.Chris Wilson2009-02-201-4/+2
| | | | | | | | | | | | | | Set the request alignment to 0, and leave it up to i915_gem_object_pin() to set the appropriate alignment to match the fence covering the object. Eric Anholt mentioned that the pinning code is meant to choose the maximum of the request alignment and that of the fence covering the object... However currently, the pinning code will only apply the fence constraints if the supplied alignment is 0. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm: Do not leak a new reference for flink() on an existing nameChris Wilson2009-02-201-17/+16
| | | | | | | | | The name table should only hold a single reference, so avoid leaking additional references for secondary calls to flink(). Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: Fix potential AB-BA deadlock in i915_gem_execbuffer()Roland Dreier2009-02-201-9/+12
| | | | | | | | | | | | | | | | | | | | | | | Lockdep warns that i915_gem_execbuffer() can trigger a page fault (which takes mmap_sem) while holding dev->struct_mutex, while drm_vm_open() (which is called with mmap_sem already held) takes dev->struct_mutex. So this is a potential AB-BA deadlock. The way that i915_gem_execbuffer() triggers a page fault is by doing copy_to_user() when returning new buffer offsets back to userspace; however there is no reason to hold the struct_mutex when doing this copy, since what is being copied is the contents of an array private to i915_gem_execbuffer() anyway. So we can fix the potential deadlock (and get rid of the lockdep warning) by simply moving the copy_to_user() outside of where struct_mutex is held. This fixes <http://bugzilla.kernel.org/show_bug.cgi?id=12491>. Reported-by: Jesse Brandeburg <jesse.brandeburg@intel.com> Tested-by: Jesse Brandeburg <jesse.brandeburg@intel.com> Signed-off-by: Roland Dreier <rolandd@cisco.com> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: refleak along pin() error path.Chris Wilson2009-02-201-0/+1
| | | | | | | | | A missing unreference if the user calls pin() a second time on a pinned buffer. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: hold mutex for unreference() in i915_gem_tiling.cChris Wilson2009-02-201-4/+2
| | | | | | Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: Cleanup trivial leak on execbuffer error path.Chris Wilson2009-02-201-2/+4
| | | | | | | | Also spotted by Owain Ainsworth. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm: Free the object ref on error.Chris Wilson2009-02-201-8/+12
| | | | | | | | | Ensure that the object is unreferenced if we fail to allocate during drm_gem_flink_ioctl(). Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm: Potential use-after-free on error path.Chris Wilson2009-02-201-1/+1
| | | | | | | | Remove the member from the hash table before we free the structure! Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: Storage class should be before const qualifierTobias Klauser2009-02-202-2/+2
| | | | | | | | | | | | The C99 specification states in section 6.11.5: The placement of a storage-class specifier other than at the beginning of the declaration specifiers in a declaration is an obsolescent feature. Signed-off-by: Tobias Klauser <tklauser@distanz.ch> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* fbdev/drm: fix Kconfig submenu mess in "Graphics support"Krzysztof Helt2009-02-181-7/+6
| | | | | | | | | | | | | | | | | | | | | | | Submenus of the graphics support "Support for frame buffer devices" and "Direct Rendering Manager (XFree86 4.1.0 and higher DRI support)" are broken in half after latest changes for Intel 915 mode setting support. The DRM subsection is broken because one option is put outside the choice section it depends on. The frame buffers part is broken then due to circular dependency. Fix this by make Intel frame buffers depend on CONFIG_INTEL_AGP. Kconfigs are broken by d2f59357700487a8b944f4f7777d1e97cf5ea2ed ("drm/i915: select framebuffer support automatically"). This is probably not only way to fix this. Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl> Cc: Ingo Molnar <mingo@elte.hu> Cc: Dave Airlie <airlied@linux.ie> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* i915: Fix more size_t format string warningsLinus Torvalds2009-02-091-1/+1
| | | | | | | The DRI people seem to have a hard time getting these right (see also commit aeb565dfc3ac4c8b47c5049085b4c7bfb2c7d5d7). Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* drm/i915: select framebuffer support automaticallyIngo Molnar2009-02-081-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Migration helper. The i915 driver recently added a 'depends on FB' rule to its Kconfig entry - which silently turns off DRM_I915 if someone has a working config but no CONFIG_FB selected, and upgrades to the latest upstream kernel. Norbert Preining reported this problem: Bug-Entry : http://bugzilla.kernel.org/show_bug.cgi?id=12599 Subject : dri /dev node disappeared with 2.6.29-rc1 So change it to "select FB", which auto-selects framebuffer support. This way the driver keeps working, regardless of whether FB was enabled before or not. Kconfig select's of interactive options can be problematic to dependencies and can cause build breakages - but in this case it's safe because it's a leaf entry with no dependencies of its own. ( There is some minor circular dependency fallout as FB_I810 and FB_INTEL also used 'depends on FB' constructs - update those to "select FB" too. ) Reported-by: Norbert Preining <preining@logic.at> Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: add get_vblank_counter function for GM45Jesse Barnes2009-02-085-1/+24
| | | | | | | | | | | As discussed in the long thread about vblank related timeouts, it turns out GM45 has different frame count registers than previous chips. This patch adds support for them, which prevents us from waiting on really stale sequence values in drm_wait_vblank (which rather than returning immediately ends up timing out or getting interrupted). Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: capture last_vblank count at IRQ uninstall time tooJesse Barnes2009-02-081-0/+1
| | | | | | | | | | | | | | | | | | | | | In dc1336ff4fe08ae7cfe8301bfd7f0b2cfd31d20a (set vblank enable flag correctly across IRQ uninstall), we made sure drivers that uninstall their interrupt handler set the vblank enabled flag correctly, so that when interrupts are re-enabled, vblank interrupts & counts work as expected. However I missed the last_vblank field: it needs to be updated as well, otherwise, at the next drm_update_vblank_count we'll end up comparing a current count to a stale one (the last one captured by the disable function), which may trigger the wraparound handling, leading to a jumpy counter and hangs in drm_wait_vblank. The jumpy counter can prevent the DRM_WAIT_ON from returning success if the difference between the current count and the requested count is greater than 2^23, leading to timeouts or hangs, if the ioctl is restarted in a loop (as is the case in libdrm < 2.4.4). Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> Acked-by: Michel Dänzer <michel@daenzer.net> Tested-by: Timo Aaltonen <tjaalton@cc.hut.fi> Signed-off-by: Dave Airlie <airlied@redhat.com>
* drm/i915: Unlock mutex on i915_gem_fault() error pathChris Wilson2009-02-081-1/+3
| | | | | | | | | If we failed to allocate a new fence register we would return VM_FAULT_SIGBUS without relinquishing the lock. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: Quiet the message on get/setparam ioctl with an unknown value.Eric Anholt2009-02-081-2/+2
| | | | | | | | | | Getting an unknown get/setparam used to be more significant back when they didn't change much. However, now that we're in the git world we're using them instead of a monotonic version number to signal feature availability, so clients ask about unknown params on older kernels more often. Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: skip LVDS initialization on Apple Mac MiniPaul Collins2009-02-081-34/+13
| | | | | | | | | | The Apple Mac Mini falsely reports LVDS. Use DMI to check whether we are running on a Mac Mini, and skip LVDS initialization if that proves to be the case. Signed-off-by: Paul Collins <paul@ondioline.org> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: sync SDVO code with stable userland modesetting driverJesse Barnes2009-02-084-104/+1173
| | | | | | | | | Pull in an update from the 2D driver (hopefully the last one, future work should be done here and pulled back into xf86-video-intel as needed). Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: Unref the object after failing to set tiling mode.Chris Wilson2009-02-081-1/+4
| | | | | | | | Cleanup the object reference on the error paths. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: add fence register management to execbufJesse Barnes2009-02-085-20/+144
| | | | | | | | | | | | Adds code to set up fence registers at execbuf time on pre-965 chips as necessary. Also fixes up a few bugs in the pre-965 tile register support (get_order != ffs). The number of fences available to the kernel defaults to the hw limit minus 3 (for legacy X front/back/depth), but a new parameter allows userspace to override that as needed. Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: Return error from i915_gem_object_get_fence_reg() when failing.Eric Anholt2009-02-081-7/+13
| | | | | | | | | | Previously, the caller would continue along without knowing that the function failed, resulting in potential mis-rendering. Right now vm_fault just returns SIGBUS in that case, and we may need to disable signal handling to avoid that happening. Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: Set up an MTRR covering the GTT at driver load.Eric Anholt2009-02-083-11/+25
| | | | | | | | | We'd love to just be using PAT, but even on chips with PAT it gets disabled sometimes due to an errata. It would probably be better to have pat_enabled exported and only bother with this when !pat_enabled. Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: Skip SDVO/HDMI init when the chipset tells us it's not present.Eric Anholt2009-02-081-7/+11
| | | | | | | This saves startup time from probing SDVO, and saves setting up HDMI outputs on G4X devices that don't have them. Signed-off-by: Dave Airlie <airlied@linux.ie>
* drm/i915: Suppress GEM teardown on X Server exit in KMS mode.Eric Anholt2009-02-081-0/+3
| | | | | | | Fixes hangs when starting X for the second time. Signed-off-by: Eric Anholt <eric@anholt.net> Signed-off-by: Dave Airlie <airlied@linux.ie>
generated by cgit v1.1 at 2024-12-26 07:40:20 +0000