- 22 Jun, 2017 1 commit
-
-
Chris Wilson authored
commit 2889caa9 ("drm/i915: Eliminate lots of iterations over the execobjects array") jiggled around the error handling and replace a test that we cleaned up properly after ourselves with an assertion. That assertion failed because in the release function (moments after the assertion) we were indeed forgetting to mark the vma as cleared. The consequence was when testing an invalid relocation address, we would try to release the vma twice (following the couple of attempts to verify the address) and on the second release notice that the first release was incomplete. Testcase: igt/gem_reloc_overflow/invalid-address Fixes: 2889caa9 ("drm/i915: Eliminate lots of iterations over the execobjects array") Signed-off-by:
Chris Wilson <chris@chris-wilson.co.uk> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170622104722.2583-1-chris@chris-wilson.co.ukReviewed-by:
Tvrtko Ursulin <tvrtko.ursulin@intel.com>
-
- 21 Jun, 2017 5 commits
-
-
Chris Wilson authored
Highly unlikely, but if the stop_machine() did suspend the tasklet, we want to make sure that when it wakes it finds there is nothing to do. Otherwise, it will loudly complain that the ELSP port tracking no longer matches the hardware, and we will be mightly confused. Signed-off-by:
-
Chris Wilson authored
As we walk the obj->vma_list in per_file_stats(), we need to hold struct_mutex to prevent alteration of that list. Fixes: 1d2ac403 ("drm: Protect dev->filelist with its own mutex") Fixes: c84455b4 ("drm/i915: Move debug only per-request pid tracking from request to ctx") Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101460Signed-off-by:
-
Chris Wilson authored
In looking at a use-after-free on Baytrail, it looks like the VMA's activity tracking is suspect. Add some asserts to catch freeing the VMA before we have decoupled all of its i915_gem_active trackers. References: https://bugs.freedesktop.org/show_bug.cgi?id=101511Signed-off-by:
Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
-
Chris Wilson authored
Since we may track unfenced access (GPU access to the vma that explicitly requires no fence), vma->last_fence may be set without any attached fence (vma->fence) and so will not be flushed when we call i915_vma_put_fence(). Since we stopped doing a full retire of the activity trackers for unbind, we need to explicitly retire each tracker. Fixes: b0decaf7 ("drm/i915: Track active vma requests") Signed-off-by:
-
Nicholas Piggin authored
kbuild test robot found a build failure when building with thin archives: http://marc.info/?l=linux-kbuild&m=149802285009737&w=2Signed-off-by:
Nicholas Piggin <npiggin@gmail.com> Signed-off-by:
Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/20170621063420.24913-1-npiggin@gmail.com
-
- 20 Jun, 2017 15 commits
-
-
Chris Wilson authored
i915_vma_move_to_active() takes the execobject flags and not a boolean! Instead of passing EXEC_OBJECT_WRITE we passed true [i.e. EXEC_OBJECT_NEEDS_FENCE] causing us to start tracking the vma->last_fence access and since we forgot to clear that on unbinding, we caused a use-after-free. [ 321.263854] BUG: KASAN: use-after-free in i915_gem_request_retire+0x1728/0x1740 [i915] [ 321.264001] Read of size 8 at addr ffff880100fc67d8 by task gem_exec_reloc/2868 [ 321.264181] CPU: 0 PID: 2868 Comm: gem_exec_reloc Not tainted 4.12.0-rc6-CI-Custom_2759+ #1 [ 321.264195] Hardware name: GIGABYTE GB-BXBT-1900/MZBAYAB-00, BIOS F6 02/17/2015 [ 321.264208] Call Trace: [ 321.264234] dump_stack+0x67/0x99 [ 321.264260] print_address_description+0x77/0x290 [ 321.264437] ? i915_gem_request_retire+0x1728/0x1740 [i915] [ 321.264459] kasan_report+0x269/0x350 [ 321.264487] __asan_report_load8_noabort+0x14/0x20 [ 321.264660] i915_gem_request_retire+0x1728/0x1740 [i915] [ 321.264841] ? intel_ring_context_pin+0x131/0x690 [i915] [ 321.265021] i915_gem_request_alloc+0x2c6/0x1220 [i915] [ 321.265044] ? _raw_spin_unlock_irqrestore+0x3d/0x60 [ 321.265226] i915_gem_do_execbuffer+0xac0/0x2a20 [i915] [ 321.265250] ? __lock_acquire+0xceb/0x5450 [ 321.265269] ? entry_SYSCALL_64_fastpath+0x1c/0xb1 [ 321.265291] ? kvmalloc_node+0x6b/0x80 [ 321.265310] ? kvmalloc_node+0x6b/0x80 [ 321.265489] ? eb_relocate_slow+0xbe0/0xbe0 [i915] [ 321.265520] ? ___slab_alloc.constprop.28+0x2ab/0x3d0 [ 321.265549] ? debug_check_no_locks_freed+0x280/0x280 [ 321.265591] ? __might_fault+0xc6/0x1b0 [ 321.265782] i915_gem_execbuffer2+0x14a/0x3f0 [i915] [ 321.265815] drm_ioctl+0x4ba/0xaa0 [ 321.265986] ? i915_gem_execbuffer+0xde0/0xde0 [i915] [ 321.266017] ? drm_getunique+0x270/0x270 [ 321.266068] do_vfs_ioctl+0x17f/0xfa0 [ 321.266091] ? __fget+0x1ba/0x330 [ 321.266112] ? lock_acquire+0x390/0x390 [ 321.266133] ? ioctl_preallocate+0x1d0/0x1d0 [ 321.266164] ? __fget+0x1db/0x330 [ 321.266194] ? __fget_light+0x79/0x1f0 [ 321.266219] SyS_ioctl+0x3c/0x70 [ 321.266247] entry_SYSCALL_64_fastpath+0x1c/0xb1 [ 321.266265] RIP: 0033:0x7fcede207357 [ 321.266279] RSP: 002b:00007ffef0effe58 EFLAGS: 00000246 ORIG_RAX: 0000000000000010 [ 321.266307] RAX: ffffffffffffffda RBX: 0000000000000002 RCX: 00007fcede207357 [ 321.266321] RDX: 00007ffef0effef0 RSI: 0000000040406469 RDI: 0000000000000004 [ 321.266335] RBP: ffffffff812097c6 R08: 0000000000000008 R09: 0000000000000000 [ 321.266349] R10: 0000000000000008 R11: 0000000000000246 R12: ffff880116bcff98 [ 321.266363] R13: ffffffff81cb7cb3 R14: ffff880116bcff70 R15: 0000000000000000 [ 321.266385] ? __this_cpu_preempt_check+0x13/0x20 [ 321.266406] ? trace_hardirqs_off_caller+0x1d6/0x2c0 [ 321.266487] Allocated by task 2868: [ 321.266568] save_stack_trace+0x16/0x20 [ 321.266586] kasan_kmalloc+0xee/0x180 [ 321.266602] kasan_slab_alloc+0x12/0x20 [ 321.266620] kmem_cache_alloc+0xc7/0x2e0 [ 321.266795] i915_vma_instance+0x28c/0x1540 [i915] [ 321.266964] eb_lookup_vmas+0x5a7/0x2250 [i915] [ 321.267130] i915_gem_do_execbuffer+0x69a/0x2a20 [i915] [ 321.267296] i915_gem_execbuffer2+0x14a/0x3f0 [i915] [ 321.267315] drm_ioctl+0x4ba/0xaa0 [ 321.267333] do_vfs_ioctl+0x17f/0xfa0 [ 321.267350] SyS_ioctl+0x3c/0x70 [ 321.267369] entry_SYSCALL_64_fastpath+0x1c/0xb1 [ 321.267428] Freed by task 177: [ 321.267502] save_stack_trace+0x16/0x20 [ 321.267521] kasan_slab_free+0xad/0x180 [ 321.267539] kmem_cache_free+0xc5/0x340 [ 321.267710] i915_vma_unbind+0x666/0x10a0 [i915] [ 321.267880] i915_vma_close+0x23a/0x2f0 [i915] [ 321.268048] __i915_gem_free_objects+0x17d/0xc70 [i915] [ 321.268215] __i915_gem_free_work+0x49/0x70 [i915] [ 321.268234] process_one_work+0x66f/0x1410 [ 321.268252] worker_thread+0xe1/0xe90 [ 321.268269] kthread+0x304/0x410 [ 321.268285] ret_from_fork+0x27/0x40 [ 321.268346] The buggy address belongs to the object at ffff880100fc6640 which belongs to the cache i915_vma of size 656 [ 321.268550] The buggy address is located 408 bytes inside of 656-byte region [ffff880100fc6640, ffff880100fc68d0) [ 321.268741] The buggy address belongs to the page: [ 321.268837] page:ffffea000403f000 count:1 mapcount:0 mapping: (null) index:0xffff880100fc5980 compound_mapcount: 0 [ 321.269045] flags: 0x8000000000008100(slab|head) [ 321.269147] raw: 8000000000008100 0000000000000000 ffff880100fc5980 00000001001e001d [ 321.269312] raw: ffffea0004038e20 ffff880116b46240 ffff88011646c640 0000000000000000 [ 321.269484] page dumped because: kasan: bad access detected [ 321.269665] Memory state around the buggy address: [ 321.269778] ffff880100fc6680: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb [ 321.269949] ffff880100fc6700: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb [ 321.270115] >ffff880100fc6780: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb [ 321.270279] ^ [ 321.270410] ffff880100fc6800: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb [ 321.270576] ffff880100fc6880: fb fb fb fb fb fb fb fb fb fb fc fc fc fc fc fc [ 321.270740] ================================================================== [ 321.270903] Disabling lock debugging due to kernel taint Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101511 Fixes: 7dd4f672 ("drm/i915: Async GPU relocation processing") Signed-off-by: -
Michel Thierry authored
This feature is made available only from Gen8, for previous gen devices driver uses legacy full gpu reset. Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Signed-off-by:
Tomas Elf <tomas.elf@intel.com> Signed-off-by:
Arun Siluvery <arun.siluvery@linux.intel.com> Signed-off-by:
Michel Thierry <michel.thierry@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170615201828.23144-10-michel.thierry@intel.comReviewed-by:
-
Michel Thierry authored
Check that we can reset specific engines, also check the fallback to full reset if something didn't work. v2: rebase. v3: use RESET_ENGINE_IN_PROGRESS flag. v4: use I915_RESET_ENGINE flag. Signed-off-by:
-
Michel Thierry authored
A new variable is added to export the reset counts to debugfs, this includes full gpu reset and engine reset count. This is useful for tests where they are expected to trigger reset; these counts are checked before and after the test to ensure the same. v2: Include reset engine count in i915_engine_info too (Chris). Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Signed-off-by:
-
Michel Thierry authored
Driver maintains count of how many times a given engine is reset, useful to capture this in error state also. It gives an idea of how engine is coping up with the workloads it is executing before this error state. A follow-up patch will provide this information in debugfs. v2: s/engine_reset/reset_engine/ (Chris) Define count as unsigned int (Tvrtko) Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Signed-off-by: -
Michel Thierry authored
This change implements support for per-engine reset as an initial, less intrusive hang recovery option to be attempted before falling back to the legacy full GPU reset recovery mode if necessary. This is only supported from Gen8 onwards. Hangchecker determines which engines are hung and invokes error handler to recover from it. Error handler schedules recovery for each of those engines that are hung. The recovery procedure is as follows, - identifies the request that caused the hang and it is dropped - force engine to idle: this is done by issuing a reset request - reset the engine - re-init the engine to resume submissions. If engine reset fails then we fall back to heavy weight full gpu reset which resets all engines and reinitiazes complete state of HW and SW. v2: Rebase. v3: s/*engine_reset*/*reset_engine*/; freeze engine and irqs before calling i915_gem_reset_engine (Chris). v4: Rebase, modify i915_gem_reset_prepare to use a ring mask and reuse the function for reset_engine. v5: intel_reset_engine_start/cancel instead of request/unrequest_reset. v6: Clean up reset_engine function to not require mutex, i.e. no need to call revoke/restore_fences and _retire_requests (Chris). v7: Remove leftovers from v5, i.e. no need to disable irq, hold forcewake or wakeup the handoff bit (Chris). v8: engine_retire_requests should be (and it was) static; explain that we have to re-init the engine after reset, which is why the init_hw call is needed; check reset-in-progress flag (Chris). v9: Rebase, include code to pass the active request to gem_reset_engine (as it is already done in full reset). Remove unnecessary intel_reset_engine_start/cancel, these are executed as part of the reset. v10: Rebase, use the right I915_RESET_ENGINE flag. v11: Fixup to call reset_finish_engine even on error. Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Signed-off-by:
-
Michel Thierry authored
This is a preparatory patch which modifies error handler to do per engine hang recovery. The actual patch which implements this sequence follows later in the series. The aim is to prepare existing recovery function to adapt to this new function where applicable (which fails at this point because core implementation is lacking) and continue recovery using legacy full gpu reset. A helper function is also added to query the availability of engine reset. A subsequent patch will add the capability to query which type of reset is present (engine -> full -> no-reset) via the get-param ioctl. It has been decided that the error events that are used to notify user of reset will only be sent in case if full chip reset. In case of just single (or multiple) engine resets, userspace won't be notified by these events. Note that this implementation of engine reset is for i915 directly submitting to the ELSP, where the driver manages the hang detection, recovery and resubmission. With GuC submission these tasks are shared between driver and firmware; i915 will still responsible for detecting a hang, and when it does it will have to request GuC to reset that Engine and remind the firmware about the outstanding submissions. This will be added in different patch. v2: rebase, advertise engine reset availability in platform definition, add note about GuC submission. v3: s/*engine_reset*/*reset_engine*/. (Chris) Handle reset as 2 level resets, by first going to engine only and fall backing to full/chip reset as needed, i.e. reset_engine will need the struct_mutex. v4: Pass the engine mask to i915_reset. (Chris) v5: Rebase, update selftests. v6: Rebase, prepare for mutex-less reset engine. v7: Pass reset_engine mask as a function parameter, and iterate over the engine mask for reset_engine. (Chris) v8: Use i915.reset >=2 in has_reset_engine; remove redundant reset logging; add a reset-engine-in-progress flag to prevent concurrent resets, and avoid dual purposing of reset-backoff. (Chris) v9: Support reset of different engines in parallel (Chris) v10: Handle reset-engine flag locking better (Chris) v11: Squash in reporting of per-engine-reset availability. Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Signed-off-by:
Ian Lister <ian.lister@intel.com> Signed-off-by:
-
Michel Thierry authored
In preparation for engine reset work update this parameter to handle more than one type of reset. Default at the moment is still full gpu reset. Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Signed-off-by:
-
Michel Thierry authored
And store the active request so that we only search for it once. v2: Check for request completion inside _prepare_engine, don't use ECANCELED, remove unnecessary null checks (Chris). v3: Capture active requests during reset_prepare and store it the engine hangcheck obj. v4: Rename commit, change i915_gem_reset_request to just confirm the active_request is still incomplete, instead of engine_stalled (Chris). v5: With style; pass the active request to gem_reset_engine, keep single return in reset_prepare_engine (Chris). v6: Moved before reset-engine code appears (Chris) Suggested-by:
-
Chris Wilson authored
If we enter i915_handle_error() a second time and a global reset is already in progress, we can simply wait for completion of the first reset. Currently we exit early prior to the actual reset being performed -- the worst of both worlds! v2: Plug into the existing reset_queue, and remember that kselftests is playing games with I915_RESET_BACKOFF to prevent hangcheck from screwing up. v3: Rename to i915_reset_device to fit in better with i915_reset_engine Signed-off-by:
-
Chris Wilson authored
Whilst the contents of the context is still protected by the big struct_mutex, this is not much of an improvement. It is just one tiny step towards reducing our BKL. Signed-off-by:
Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170620110547.15947-3-chris@chris-wilson.co.uk
-
Chris Wilson authored
If we move the actual cleanup of the context to a worker, we can allow the final free to be called from any context and avoid undue latency in the caller. v2: Negotiate handling the delayed contexts free by flushing the workqueue before calling i915_gem_context_fini() and performing the final free of the kernel context directly v3: Flush deferred frees before new context allocations Signed-off-by:
-
Chris Wilson authored
Create a substruct to hold all the global context state under drm_i915_private. Signed-off-by:
-
Tvrtko Ursulin authored
Since bb8f0f5a ("drm/i915: Split intel_engine allocation and initialisation") intel_info->num_rings is set early in the load sequence and so available to be used direclty in the 2nd load phase. Signed-off-by:
-
Tvrtko Ursulin authored
We do not want to carry on over missing constructors and don't need a duplicated engine mask checking which is already done in the setup phase. Signed-off-by:
-
- 19 Jun, 2017 3 commits
-
-
Rodrigo Vivi authored
Paulo noticed that we were missing few bits clear before writing values back to the register on these RMW MMIO operations. v2: Remove "POST_" from CURSOR_COEFF_MASK. (Paulo). v3: Remove unnecessary braces. (Jani). Fixes: cf54ca8b ("drm/i915/cnl: Implement voltage swing sequence.") Cc: Paulo Zanoni <paulo.r.zanoni@intel.com> Cc: Manasi Navare <manasi.d.navare@intel.com> Cc: Jani Nikula <jani.nikula@intel.com> Signed-off-by:
Rodrigo Vivi <rodrigo.vivi@intel.com> Reviewed-by:
Paulo Zanoni <paulo.r.zanoni@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1497897572-22520-1-git-send-email-rodrigo.vivi@intel.com
-
Ville Syrjälä authored
Add the missing port A handling to intel_digital_port_connected() and also separate SPT from the CPT/LPT code a bit. Cc: Manasi Navare <manasi.d.navare@intel.com> Signed-off-by:
Ville Syrjälä <ville.syrjala@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170615171252.11921-1-ville.syrjala@linux.intel.comReviewed-by:
Manasi Navare <manasi.d.navare@intel.com>
-
Daniel Vetter authored
Signed-off-by:
-
- 16 Jun, 2017 16 commits
-
-
Rodrigo Vivi authored
Coffee Lake inherit most of Kabylake production workarounds. v2: Fix typo on commit message and remove WaDisableKillLogic and GEN9_DISABLE_OCL_OOB_SUPPRESS_LOGIC, since as Mika pointed out they shouldn't be here for cfl according to BSpec. Cc: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com> Signed-off-by:Mika Kuoppala <mika.kuoppala@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1497653398-15722-1-git-send-email-rodrigo.vivi@intel.com
-
Dhinakaran Pandiyan authored
Although we use 9 bits of Device ID for identifying PCH, only 8 bits are stored in dev_priv->pch_id. This makes HAS_PCH_CNP_LP() and HAS_PCH_SPT_LP() incorrect. Fix this by storing all the 9 bits for the platforms with LP PCH. v2: Drop PCH_LPT_LP change (Imre) Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: Jani Nikula <jani.nikula@intel.com> Cc: Imre Deak <imre.deak@intel.com> Fixes: commit ec7e0bb3 ("drm/i915/cnp: Add PCI ID for Cannonpoint LP PCH") Reported-by:
Imre Deak <imre.deak@intel.com> Reviewed-by:
Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com> Signed-off-by:
-
Chris Wilson authored
During execbuf, a mandatory step is that we add this request (this fence) to each object's reservation_object. Inside execbuf, we track the vma, and to add the fence to the reservation_object then means having to first chase the obj, incurring another cache miss. We can reduce the number of cache misses by stashing a pointer to the reservation_object in the vma itself. Signed-off-by:
-
Chris Wilson authored
If the user requires patching of their batch or auxiliary buffers, we currently make the alterations on the cpu. If they are active on the GPU at the time, we wait under the struct_mutex for them to finish executing before we rewrite the contents. This happens if shared relocation trees are used between different contexts with separate address space (and the buffers then have different addresses in each), the 3D state will need to be adjusted between execution on each context. However, we don't need to use the CPU to do the relocation patching, as we could queue commands to the GPU to perform it and use fences to serialise the operation with the current activity and future - so the operation on the GPU appears just as atomic as performing it immediately. Performing the relocation rewrites on the GPU is not free, in terms of pure throughput, the number of relocations/s is about halved - but more importantly so is the time under the struct_mutex. v2: Break out the request/batch allocation for clearer error flow. v3: A few asserts to ensure rq ordering is maintained Signed-off-by:
-
Chris Wilson authored
Currently, the last object in the execlist is the always the batch. However, when building the batch buffer we often know the batch object first and if we can use the first slot in the execlist we can emit relocation instructions relative to it immediately and avoid a separate pass to adjust the relocations to point to the last execlist slot. Signed-off-by:
-
Chris Wilson authored
This simply hides the EAGAIN caused by userptr when userspace causes resource contention. However, it is quite beneficial with highly contended userptr users as we avoid repeating the setup costs and kernel-user context switches. Signed-off-by:
Michał Winiarski <michal.winiarski@intel.com>
-
Chris Wilson authored
When choosing a slot for an execbuffer, we ideally want to use the same address as last time (so that we don't have to rebind it) and the same address as expected by the user (so that we don't have to fixup any relocations pointing to it). If we first try to bind the incoming execbuffer->offset from the user, or the currently bound offset that should hopefully achieve the goal of avoiding the rebind cost and the relocation penalty. However, if the object is not currently bound there we don't want to arbitrarily unbind an object in our chosen position and so choose to rebind/relocate the incoming object instead. After we report the new position back to the user, on the next pass the relocations should have settled down. Signed-off-by:
Joonas Lahtinen <joonas.lahtien@linux.intel.com>
-
Chris Wilson authored
If we take a reference to the object/vma when it is first used in an execbuf, we can keep that reference until the object's file-local handle is closed. Thereby saving a frequent ref/unref pair. Signed-off-by:
-
Chris Wilson authored
The major scaling bottleneck in execbuffer is the processing of the execobjects. Creating an auxiliary list is inefficient when compared to using the execobject array we already have allocated. Reservation is then split into phases. As we lookup up the VMA, we try and bind it back into active location. Only if that fails, do we add it to the unbound list for phase 2. In phase 2, we try and add all those objects that could not fit into their previous location, with fallback to retrying all objects and evicting the VM in case of severe fragmentation. (This is the same as before, except that phase 1 is now done inline with looking up the VMA to avoid an iteration over the execobject array. In the ideal case, we eliminate the separate reservation phase). During the reservation phase, we only evict from the VM between passes (rather than currently as we try to fit every new VMA). In testing with Unreal Engine's Atlantis demo which stresses the eviction logic on gen7 class hardware, this speed up the framerate by a factor of 2. The second loop amalgamation is between move_to_gpu and move_to_active. As we always submit the request, even if incomplete, we can use the current request to track active VMA as we perform the flushes and synchronisation required. The next big advancement is to avoid copying back to the user any execobjects and relocations that are not changed. v2: Add a Theory of Operation spiel. v3: Fall back to slow relocations in preparation for flushing userptrs. v4: Document struct members, factor out eb_validate_vma(), add a few more comments to explain some magic and hide other magic behind macros. Signed-off-by:
-
Chris Wilson authored
If we write a relocation into the buffer, we require our own implicit synchronisation added after the start of the execbuf, outside of the user's control. As we may end up clflushing, or doing the patch itself on the GPU, asynchronously we need to look at the implicit serialisation on obj->resv and hence need to disable EXEC_OBJECT_ASYNC for this object. If the user does trigger a stall for relocations, we make sure the stall is complete enough so that the batch is not submitted before we complete those relocations. Fixes: 77ae9957 ("drm/i915: Enable userspace to opt-out of implicit fencing") Signed-off-by:
-
Chris Wilson authored
We can simplify our tracking of pending writes in an execbuf to the single bit in the vma->exec_entry->flags, but that requires the relocation function knowing the object's vma. Pass it along. Note we have only been using a single bit to track flushing since commit cc889e0f Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Wed Jun 13 20:45:19 2012 +0200 drm/i915: disable flushing_list/gpu_write_list unconditionally flushed all render caches before the breadcrumb and commit 6ac42f41 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Sat Jul 21 12:25:01 2012 +0200 drm/i915: Replace the complex flushing logic with simple invalidate/flush all did away with the explicit GPU domain tracking. This was then codified into the ABI with NO_RELOC in commit ed5982e6 Author: Daniel Vetter <daniel.vetter@ffwll.ch> # Oi! Patch stealer! Date: Thu Jan 17 22:23:36 2013 +0100 drm/i915: Allow userspace to hint that the relocations were known Signed-off-by:
-
Chris Wilson authored
The advent of full-ppgtt lead to an extra indirection between the object and its binding. That extra indirection has a noticeable impact on how fast we can convert from the user handles to our internal vma for execbuffer. In order to bypass the extra indirection, we use a resizable hashtable to jump from the object to the per-ctx vma. rhashtable was considered but we don't need the online resizing feature and the extra complexity proved to undermine its usefulness. Instead, we simply reallocate the hastable on demand in a background task and serialize it before iterating. In non-full-ppgtt modes, multiple files and multiple contexts can share the same vma. This leads to having multiple possible handle->vma links, so we only use the first to establish the fast path. The majority of buffers are not shared and so we should still be able to realise speedups with multiple clients. v2: Prettier names, more magic. v3: Many style tweaks, most notably hiding the misuse of execobj[].rsvd2 Signed-off-by:
-
Chris Wilson authored
The default context is always supported (as it contains the global hangcheck stats) and the contexts for hangcheck are not limited to any ring. This was dropped in 2013 because it was supposed to have been included with Ben's full-ppgtt patch set. It never landed and the bug remains. References: https://bugs.freedesktop.org/show_bug.cgi?id=65845 Link: http://patchwork.freedesktop.org/patch/msgid/1372175222-27622-1-git-send-email-mika.kuoppala@intel.comSigned-off-by:
-
Chris Wilson authored
For ease of use (i.e. avoiding a few checks and function calls), store the object's cache coherency next to the cache is dirty bit. Specifically this patch aims to reduce the frequency of no-op calls to i915_gem_object_clflush() to counter-act the increase of such calls for GPU only objects in the previous patch. v2: Replace cache_dirty & ~cache_coherent with cache_dirty && !cache_coherent as gcc generates much better code for the latter (Tvrtko) Signed-off-by:
Dongwon Kim <dongwon.kim@intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170616105455.16977-1-chris@chris-wilson.co.ukReviewed-by:
-
Chris Wilson authored
Currently, we only mark the CPU cache as dirty if we skip a clflush. This leads to some confusion where we have to ask if the object is in the write domain or missed a clflush. If we always mark the cache as dirty, this becomes a much simply question to answer. The goal remains to do as few clflushes as required and to do them as late as possible, in the hope of deferring the work to a kthread and not block the caller (e.g. execbuf, flips). v2: Always call clflush before GPU execution when the cache_dirty flag is set. This may cause some extra work on llc systems that migrate dirty buffers back and forth - but we do try to limit that by only setting cache_dirty at the end of the gpu sequence. v3: Always mark the cache as dirty upon a level change, as we need to invalidate any stale cachelines due to external writes. Reported-by:
-
Chris Wilson authored
i915_vma_destroy() is now not used outside of i915_vma.c so we can remove the export and make the function static. Signed-off-by:
-
