- 21 Mar, 2017 2 commits
-
-
Chris Wilson authored
intel_engine_wakeup() is called by nop_request_submit() which is installed to handle third party fences completed from within irq context. As such, it needs the full irqsave/irqrestore and not the partial spin_irq_lock handling. [18942.714467] ================================= [18942.719076] [ INFO: inconsistent lock state ] [18942.723522] 4.11.0-rc2-CI-CI_DRM_2368+ #1 Tainted: G U W [18942.729970] --------------------------------- [18942.734466] inconsistent {IN-SOFTIRQ-W} -> {SOFTIRQ-ON-W} usage. [18942.740594] gem_eio/1275 [HC0[0]:SC0[0]:HE1:SE1] takes: [18942.745932] (&(&fence->lock)->rlock){+.?...}, at: [<ffffffff815ec100>] dma_fence_signal+0x100/0x 230 [18942.755331] {IN-SOFTIRQ-W} state was registered at: [18942.760356] __lock_acquire+0x5d0/0x1bb0 [18942.764444] lock_acquire+0xc9/0x220 [18942.768196] _raw_spin_lock_irqsave+0x41/0x60 [18942.772747] dma_fence_signal+0x100/0x230 [18942.776927] vgem_fence_timeout+0x9/0x10 [vgem] [18942.781701] call_timer_fn+0x92/0x380 [18942.785557] expire_timers+0x150/0x1f0 [18942.789491] run_timer_softirq+0x7c/0x160 [18942.793705] __do_softirq+0x116/0x4c0 [18942.797560] irq_exit+0xa9/0xc0 [18942.800873] smp_apic_timer_interrupt+0x38/0x50 [18942.805611] apic_timer_interrupt+0x90/0xa0 [18942.810008] cpuidle_enter_state+0x135/0x380 [18942.814503] cpuidle_enter+0x12/0x20 [18942.818250] call_cpuidle+0x1e/0x40 [18942.821906] do_idle+0x17e/0x1f0 [18942.825333] cpu_startup_entry+0x18/0x20 [18942.829463] rest_init+0x127/0x130 [18942.833025] start_kernel+0x3f1/0x3fe [18942.836908] x86_64_start_reservations+0x2a/0x2c [18942.841733] x86_64_start_kernel+0x173/0x186 [18942.846234] verify_cpu+0x0/0xfc [18942.849604] irq event stamp: 30568 [18942.853140] hardirqs last enabled at (30567): [<ffffffff8110b81f>] ktime_get+0xef/0x120 [18942.861468] hardirqs last disabled at (30568): [<ffffffff81876377>] _raw_spin_lock_irqsave+0x17/0 x60 [18942.870812] softirqs last enabled at (30462): [<ffffffff81085cd9>] __do_softirq+0x1d9/0x4c0 [18942.879443] softirqs last disabled at (30439): [<ffffffff81086139>] irq_exit+0xa9/0xc0 [18942.887616] [18942.887616] other info that might help us debug this: [18942.894279] Possible unsafe locking scenario: [18942.894279] [18942.900336] CPU0 [18942.902851] ---- [18942.905362] lock(&(&fence->lock)->rlock); [18942.909647] <Interrupt> [18942.912330] lock(&(&fence->lock)->rlock); [18942.916821] [18942.916821] *** DEADLOCK *** [18942.916821] [18942.922862] 1 lock held by gem_eio/1275: [18942.926859] #0: (&(&fence->lock)->rlock){+.?...}, at: [<ffffffff815ec100>] dma_fence_signal+0x1 00/0x230 [18942.936651] [18942.936651] stack backtrace: [18942.941142] CPU: 3 PID: 1275 Comm: gem_eio Tainted: G U W 4.11.0-rc2-CI-CI_DRM_2368+ # 1 [18942.950367] Hardware name: Gigabyte Technology Co., Ltd. Z170X-UD5/Z170X-UD5-CF, BIOS F21 01/06/2 017 [18942.959756] Call Trace: [18942.962244] dump_stack+0x67/0x92 [18942.965626] print_usage_bug.part.23+0x259/0x268 [18942.970362] mark_lock+0x12c/0x6f0 [18942.973851] ? check_usage_forwards+0x130/0x130 [18942.978487] mark_held_locks+0x6f/0xa0 [18942.982329] ? _raw_spin_unlock_irq+0x27/0x50 [18942.986797] trace_hardirqs_on_caller+0x150/0x200 [18942.991599] trace_hardirqs_on+0xd/0x10 [18942.995515] _raw_spin_unlock_irq+0x27/0x50 [18942.999796] intel_engine_wakeup+0x26/0x30 [i915] [18943.004670] intel_engine_init_global_seqno+0x131/0x1a0 [i915] [18943.010745] nop_submit_request+0x2e/0x40 [i915] [18943.015476] submit_notify+0x3f/0x5c [i915] [18943.019763] __i915_sw_fence_complete+0x176/0x220 [i915] [18943.025234] ? try_to_del_timer_sync+0x4d/0x60 [18943.029825] i915_sw_fence_complete+0x25/0x40 [i915] [18943.034887] dma_i915_sw_fence_wake+0x26/0x60 [i915] [18943.039959] dma_fence_signal+0x146/0x230 [18943.044109] vgem_fence_signal_ioctl+0x6c/0xc0 [vgem] [18943.049275] drm_ioctl+0x200/0x450 [18943.052758] ? vgem_fence_attach_ioctl+0x270/0x270 [vgem] [18943.058334] do_vfs_ioctl+0x90/0x6e0 [18943.061991] ? entry_SYSCALL_64_fastpath+0x5/0xb1 [18943.066843] ? __this_cpu_preempt_check+0x13/0x20 [18943.071643] ? trace_hardirqs_on_caller+0xe7/0x200 [18943.076532] SyS_ioctl+0x3c/0x70 [18943.079842] entry_SYSCALL_64_fastpath+0x1c/0xb1 [18943.084558] RIP: 0033:0x7f0dfcc14357 [18943.088240] RSP: 002b:00007ffeb4628da8 EFLAGS: 00000246 ORIG_RAX: 0000000000000010 [18943.095996] RAX: ffffffffffffffda RBX: ffffffff8147eb93 RCX: 00007f0dfcc14357 [18943.103311] RDX: 00007ffeb4628de0 RSI: 0000000040086442 RDI: 0000000000000005 [18943.110574] RBP: ffffc9000176ff88 R08: 0000000000000004 R09: 0000000000000000 [18943.117845] R10: 0000000000000029 R11: 0000000000000246 R12: 0000000000000001 [18943.125168] R13: 0000000000000005 R14: 0000000040086442 R15: 0000000000000000 [18943.132520] ? __this_cpu_preempt_check+0x13/0x20 Fixes: cdc3a453 ("drm/i915: No need to save/restore irq status in intel_engine_wakeup") Signed-off-by:Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Mika Kuoppala <mika.kuoppala@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170320143133.1507-1-chris@chris-wilson.co.ukReviewed-by:
Tvrtko Ursulin <tvrtko.ursulin@intel.com>
-
Tvrtko Ursulin authored
It has to be called after the global seqno has been assigned. Signed-off-by:
-
- 20 Mar, 2017 3 commits
-
-
Chris Wilson authored
The trick of using an uncached mmio read to ensure that the GGTT writes are flushed does not require us to do the forcewake dance, so avoid it in the hope of reducing the frequency that we do keep the device forced awake. Signed-off-by:
Mika Kuoppala <mika.kuoppala@intel.com>
-
Chris Wilson authored
When switching back to execlists, we also now need to restore the tasklet handler. Reported-by:
Oscar Mateo <oscar.mateo@intel.com> Fixes: 31de7350 ("drm/i915/scheduler: emulate a scheduler for guc") Signed-off-by:
Michał Winiarski <michal.winiarski@intel.com>
-
Daniel Vetter authored
Signed-off-by:Daniel Vetter <daniel.vetter@ffwll.ch>
-
- 17 Mar, 2017 13 commits
-
-
Chris Wilson authored
Use pagecache_write to avoid shmemfs clearing the pages prior to us immediately overwriting them with our data. Signed-off-by:
Matthew Auld <matthew.auld@intel.com>
-
Chris Wilson authored
i915_gem_object_create_from_data() always returns an error pointer on failure, there is no need to check against NULL. Signed-off-by:
-
Chris Wilson authored
Both object creation and backing storage page allocation do not require struct_mutex, so do not require the caller to take it. Signed-off-by:
-
Chris Wilson authored
Since obj->active_count is only updated upon retirement, if we see an active object in the batch pool, double check that is still active before deciding to allocate a new object. Signed-off-by:
Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
-
Chris Wilson authored
To improve our historical record and to simplify userspace that wants to include i915_pciids.h as its canonical breakdown of PCI IDs and their respective generations, include the gen1 ids for i810 and i815. Signed-off-by:
Jani Nikula <jani.nikula@intel.com>
-
Chris Wilson authored
Do an early read of the execlists' queue before we take the spinlock and start checking. This is safe as the first writer to the execlists queue will cause the tasklet to be run again after a memory barrier. v2: Keep guc in sync with execlists queue changes v3: Explain the mb between the tasklet running on one cpu and the execlist_first update and schedule from a second cpu. Signed-off-by:
Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com> Reviewed-by:
-
Chris Wilson authored
i915_gem_stolen_list_info() sneakily takes advantage of the obj->obj_exec_link to save itself from having to allocate. Enough of the subterfuge, just allocate an array of pointers and sort them instead of the list. Signed-off-by:
-
Chris Wilson authored
Before rc6 is initialised (after driver load or resume), the value inside VLV_COUNTER_CONTROL is undefined so we cannot make an assertion that is in HIGH_RANGE mode. Fixes: 6b7f6aa7 ("drm/i915: Use coarse grained residency counter with byt") Testcase: igt/drv_suspend/debugfs-reader Signed-off-by:
-
Ander Conselvan de Oliveira authored
Geminilake also supports pooled EUs. Enable it. It is unclear if the recommendation to disable it for 2x6 configurations from commit e015dd69 ("drm/i915/bxt: Add WaEnablePooledEuFor2x6") should also apply to GLK, but it is applied anyway to be on the safe side. That restriction can be lifted later if determined not to impact performance. The extra restriction should not impact user space either. The only user space that uses this feature is Beignet, and it only does so for 3x6 devices. See See Beignet's commit 6901899ec90a ("Runtime: set the sub slice according to kernel pooled EU configure."). v2: Improve commit message. (Mika, Roy) Cc: Arun Siluvery <arun.siluvery@intel.com> Cc: Mika Kuoppala <mika.kuoppala@intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Yang Rong <rong.r.yang@intel.com> Signed-off-by:
Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com> Reviewed-by:
-
Chris Wilson authored
The only time we need to emit a flush inside request emission is after an execbuffer, for which we can use the full __i915_add_request(). All other instances want the simpler i915_add_request() without flushing, so remove the useless helper. Signed-off-by:
-
Tvrtko Ursulin authored
If we avoid initializing forcewake domains when running as a guest, and also use gen2 mmio accessors in that case, we can avoid the timer traffic and any looping through the forcewake code which is currently just so it can end up in the no-op forcewake implementation. Signed-off-by:
Terrence Xu <terrence.xu@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170310095747.12258-1-tvrtko.ursulin@linux.intel.com [tursulin: commit spelling fix]
-
Zhenyu Wang authored
From commit a6508ded ("drm/i915: Use page coloring to provide the guard page at the end of the GTT"), we no longer explicitly subtract guard page at end for GGTT address space init, so shouldn't subtract that for vGPU balloon too, as that will leave that end page to be available for vGPU. Change balloon to cover full range too. This fixes to use recent drm-intel tip kernel for guest OS. Found by GVT-g cmd parser that guest kernel uses end page as scratch then try to run MI_STORE_REG_MEM onto it. v2: remove old comments Cc: Terrence Xu <terrence.xu@intel.com> Signed-off-by:
Zhenyu Wang <zhenyuw@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170310022238.3191-1-zhenyuw@linux.intel.comReviewed-by:
-
Chris Wilson authored
trace_i915_gem_request_out may be used after the request is completed, and so the request may have been retired on another thread, invalidating the rq->ctx. Avoid dereferencing rq->ctx in the tracepoint by switching to the fence context id instead, updating all tracepoints to match. Signed-off-by:
-
- 16 Mar, 2017 18 commits
-
-
Chris Wilson authored
This should be impossible, but let's assert that we do not pin a context 4 billion times before retiring! v2: Fix the assertion -- the patch had just one job to do! Signed-off-by:
-
Chris Wilson authored
Provide some serialisation between user operations by waiting for the reset initiated by setting i915_wedged to complete. The automatic wait here makes echo 1 > i915_wedged; cat i915_error_state do the right thing, and not risk reporting "No error collected". Signed-off-by: -
Chris Wilson authored
When we wedge the device, we override engine->submit_request with a nop to ensure that all in-flight requests are marked in error. However, igt would like to unwedge the device to test -EIO handling. This requires us to flush those in-flight requests and restore the original engine->submit_request. v2: Use a vfunc to unify enabling request submission to engines v3: Split new vfunc to a separate patch. v4: Make the wait interruptible -- the third party fences we wait upon may be indefinitely broken, so allow the reset to be aborted. Fixes: 821ed7df ("drm/i915: Update reset path to fix incomplete requests") Testcase: igt/gem_eio Signed-off-by:
-
Chris Wilson authored
It turns out that we may want to restore the original engine->submit_request (and engine->schedule) callbacks from more than just the guc <-> execlists transition. Move this to a vfunc so we can have a common interface. v2: Move initial selection to intel_engines_init_common(), repaint vfunc with engine->set_default_submission (and a similar colour for the helper). Signed-off-by:
-
Chris Wilson authored
I915_RESET_IN_PROGRESS is being used for both signaling the requirement to i915_mutex_lock_interruptible() to avoid taking the struct_mutex and to instruct a waiter (already holding the struct_mutex) to perform the reset. To allow for a little more coordination, split these two meaning into a couple of distinct flags. I915_RESET_BACKOFF tells i915_mutex_lock_interruptible() not to acquire the mutex and I915_RESET_HANDOFF tells the waiter to call i915_reset(). Signed-off-by:
Michel Thierry <michel.thierry@intel.com> Reviewed-by:
-
Changbin Du authored
GVTg has introduced the context status notifier to schedule the GVTg workload. At that time, the notifier is bound to GVTg context only, so GVTg is not aware of host workloads. Now we are going to improve GVTg's guest workload scheduler policy, and add Guc emulation support for new Gen graphics. Both these two features require acknowledgment for all contexts running on hardware. (But will not alter host workload.) So here try to make some change. The change is simple: 1. Move the context status notifier head from i915_gem_context to intel_engine_cs. Which means there is a notifier head per engine instead of per context. Execlist driver still call notifier for each context sched-in/out events of current engine. 2. At GVTg side, it binds a notifier_block for each physical engine at GVTg initialization period. Then GVTg can hear all context status events. In this patch, GVTg do nothing for host context event, but later will add a function there. But in any case, the notifier callback is a noop if this is no active vGPU. Since intel_gvt_init() is called at early initialization stage and require the status notifier head has been initiated, I initiate it in intel_engine_setup(). v2: remove a redundant newline. (chris) Fixes: 3c7ba635 ("drm/i915: Introduce execlist context status change notification") Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100232Signed-off-by:Changbin Du <changbin.du@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com> Cc: Zhi Wang <zhi.a.wang@intel.com> Reviewed-by:
-
Chris Wilson authored
This emulates execlists on top of the GuC in order to defer submission of requests to the hardware. This deferral allows time for high priority requests to gazump their way to the head of the queue, however it nerfs the GuC by converting it back into a simple execlist (where the CPU has to wake up after every request to feed new commands into the GuC). v2: Drop hack status - though iirc there is still a lockdep inversion between fence and engine->timeline->lock (which is impossible as the nesting only occurs on different fences - hopefully just requires some judicious lockdep annotation) v3: Apply lockdep nesting to enabling signaling on the request, using the pattern we already have in __i915_gem_request_submit(); v4: Replaying requests after a hang also now needs the timeline spinlock, to disable the interrupts at least v5: Hold wq lock for completeness, and emit a tracepoint for enabling signal v6: Reorder interrupt checking for a happier gcc. v7: Only signal the tasklet after a user-interrupt if using guc scheduling v8: Restore lost update of rq through the i915_guc_irq_handler (Tvrtko) v9: Avoid re-initialising the engine->irq_tasklet from inside a reset v10: Hook up the execlists-style tracepoints v11: Clear the execlists irq_posted bit after taking over the interrupt/tasklet Signed-off-by:
-
Chris Wilson authored
When manually overwriting the HWS, rather than assume irq_seqno_barrier does the right thing, we can explicitly flush the cacheline instead. This avoids us calling the engine->irq_seqno_barrier() from an illegal context: [ 1472.651797] BUG: scheduling while atomic: migration/0/11/0x00000002 [ 1472.651807] Modules linked in: ctr ccm arc4 snd_hda_codec_hdmi bnep rfcomm iwldvm snd_hda_codec_conexant snd_hda_codec_generic snd_hda_intel mac80211 snd_hda_codec snd_hda_core snd_pcm dm_multipath snd_hwdep intel_powerclamp coretemp snd_seq_midi crct10dif_pclmul snd_seq_midi_event crc32_pclmul iwlwifi ghash_clmulni_intel btusb snd_rawmidi btrtl aesni_intel btbcm aes_x86_64 crypto_simd btintel cryptd glue_helper bluetooth snd_seq cfg80211 snd_timer snd_seq_device intel_ips binfmt_misc snd mei_me soundcore mei dm_mirror dm_region_hash dm_log i915 intel_gtt i2c_algo_bit drm_kms_helper cfbfillrect syscopyarea cfbimgblt sysfillrect sysimgblt fb_sys_fops cfbcopyarea prime_numbers e1000e drm ahci libahci [ 1472.651897] CPU: 0 PID: 11 Comm: migration/0 Tainted: G U 4.11.0-rc1+ #203 [ 1472.651899] Hardware name: LENOVO 514328U/514328U, BIOS 6QET44WW (1.14 ) 04/20/2010 [ 1472.651900] Call Trace: [ 1472.651913] dump_stack+0x63/0x90 [ 1472.651922] __schedule_bug+0x5d/0x6b [ 1472.651930] __schedule+0x46a/0x5f0 [ 1472.651934] schedule+0x38/0x90 [ 1472.651938] schedule_hrtimeout_range_clock+0x85/0x110 [ 1472.651945] ? hrtimer_init+0x10/0x10 [ 1472.651949] schedule_hrtimeout_range+0xe/0x10 [ 1472.651952] usleep_range+0x4d/0x60 [ 1472.652037] gen5_seqno_barrier+0x13/0x20 [i915] [ 1472.652101] intel_engine_init_global_seqno+0xd7/0x160 [i915] [ 1472.652160] __i915_gem_set_wedged_BKL+0xa0/0x180 [i915] [ 1472.652166] multi_cpu_stop+0xbb/0xe0 [ 1472.652170] ? cpu_stop_queue_work+0x90/0x90 [ 1472.652174] cpu_stopper_thread+0x82/0x110 [ 1472.652179] smpboot_thread_fn+0x137/0x190 [ 1472.652184] kthread+0xf7/0x130 [ 1472.652187] ? sort_range+0x20/0x20 [ 1472.652191] ? kthread_park+0x90/0x90 [ 1472.652195] ret_from_fork+0x2c/0x40 Testcase: igt/gem_eio #ilk Signed-off-by:
-
Mika Kuoppala authored
Set byt rc residency counters high level as chv does by default. We lose some accuracy on byt but we can do the calculation without extra hw read on both platforms, as now they behave identically in this respect. v2: use ktime v3: keep comparison u32 (Chris) Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by:
-
Mika Kuoppala authored
We have used cz timestamp register to gain a reference time wrt to residency calculations. The residency counts are in cz clk ticks (333Mhz clock) but for some reason the cz timestamp register gives 100us units. Perhaps for some other usage, the base-ten based values are easier, but in residency calculations raw units would have been the easiest. As there is not much advantage of using base-ten clock through a more costly punit access, take our reference times directly from kernel clock. v2: use ktime (Chris, Ville) Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by:
-
Mika Kuoppala authored
Use intel_rc6_residency to get benefit for increased resolution in byt/chv. v2: output raw and time (Chris) Signed-off-by:
-
Mika Kuoppala authored
Vlv and chv residency counters are 40 bits in width. With a control bit, we can choose between upper or lower 32 bit window into this counter. Lets toggle this bit on and off on and read both parts. As a result we can push the wrap from 13 seconds to 54 minutes. v2: commit msg, loop readability, goto elimination (Chris) v3: bug ref, divide outside runtime pm lock (Chris) References: https://bugs.freedesktop.org/show_bug.cgi?id=94852Reported-by:
Len Brown <len.brown@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by:
-
Mika Kuoppala authored
Change the granularity from milliseconds to microseconds when returning rc6 residencies. This is in preparation for increased resolution on some platforms. v2: use 64bit div macro (Chris) Cc: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by:
-
Mika Kuoppala authored
Plan is to make generic residency calculation utility function for usage outside of sysfs. As a first step move residency calculation into intel_pm.c Signed-off-by:
-
Chris Wilson authored
lockdep doesn't like us taking the mm->mmap_sem inside the get_pages callback for a couple of reasons. The straightforward deadlock: [13755.434059] ============================================= [13755.434061] [ INFO: possible recursive locking detected ] [13755.434064] 4.11.0-rc1-CI-CI_DRM_297+ #1 Tainted: G U [13755.434066] --------------------------------------------- [13755.434068] gem_userptr_bli/8398 is trying to acquire lock: [13755.434070] (&mm->mmap_sem){++++++}, at: [<ffffffffa00c988a>] i915_gem_userptr_get_pages+0x5a/0x2e0 [i915] [13755.434096] but task is already holding lock: [13755.434098] (&mm->mmap_sem){++++++}, at: [<ffffffff8104d485>] __do_page_fault+0x105/0x560 [13755.434105] other info that might help us debug this: [13755.434108] Possible unsafe locking scenario: [13755.434110] CPU0 [13755.434111] ---- [13755.434112] lock(&mm->mmap_sem); [13755.434115] lock(&mm->mmap_sem); [13755.434117] *** DEADLOCK *** [13755.434121] May be due to missing lock nesting notation [13755.434126] 2 locks held by gem_userptr_bli/8398: [13755.434128] #0: (&mm->mmap_sem){++++++}, at: [<ffffffff8104d485>] __do_page_fault+0x105/0x560 [13755.434135] #1: (&obj->mm.lock){+.+.+.}, at: [<ffffffffa00b887d>] __i915_gem_object_get_pages+0x1d/0x70 [i915] [13755.434156] stack backtrace: [13755.434161] CPU: 3 PID: 8398 Comm: gem_userptr_bli Tainted: G U 4.11.0-rc1-CI-CI_DRM_297+ #1 [13755.434165] Hardware name: GIGABYTE GB-BKi7(H)A-7500/MFLP7AP-00, BIOS F4 02/20/2017 [13755.434169] Call Trace: [13755.434174] dump_stack+0x67/0x92 [13755.434178] __lock_acquire+0x133a/0x1b50 [13755.434182] lock_acquire+0xc9/0x220 [13755.434200] ? i915_gem_userptr_get_pages+0x5a/0x2e0 [i915] [13755.434204] down_read+0x42/0x70 [13755.434221] ? i915_gem_userptr_get_pages+0x5a/0x2e0 [i915] [13755.434238] i915_gem_userptr_get_pages+0x5a/0x2e0 [i915] [13755.434255] ____i915_gem_object_get_pages+0x25/0x60 [i915] [13755.434272] __i915_gem_object_get_pages+0x59/0x70 [i915] [13755.434288] i915_gem_fault+0x397/0x6a0 [i915] [13755.434304] ? i915_gem_fault+0x1a1/0x6a0 [i915] [13755.434308] ? __lock_acquire+0x449/0x1b50 [13755.434311] ? __lock_acquire+0x449/0x1b50 [13755.434315] ? vm_mmap_pgoff+0xa9/0xd0 [13755.434318] __do_fault+0x19/0x70 [13755.434321] __handle_mm_fault+0x863/0xe50 [13755.434325] handle_mm_fault+0x17f/0x370 [13755.434329] ? handle_mm_fault+0x40/0x370 [13755.434332] __do_page_fault+0x279/0x560 [13755.434336] do_page_fault+0xc/0x10 [13755.434339] page_fault+0x22/0x30 [13755.434342] RIP: 0033:0x7f5ab91b5880 [13755.434345] RSP: 002b:00007fff62922218 EFLAGS: 00010216 [13755.434348] RAX: 0000000000b74500 RBX: 00007f5ab7f81000 RCX: 0000000000000000 [13755.434352] RDX: 0000000000100000 RSI: 00007f5ab7f81000 RDI: 00007f5aba61c000 [13755.434355] RBP: 00007f5aba61c000 R08: 0000000000000007 R09: 0000000100000000 [13755.434359] R10: 000000000000037d R11: 00007f5ab91b5840 R12: 0000000000000001 [13755.434362] R13: 0000000000000005 R14: 0000000000000001 R15: 0000000000000000 and cyclic deadlocks: [ 2566.458979] ====================================================== [ 2566.459054] [ INFO: possible circular locking dependency detected ] [ 2566.459127] 4.11.0-rc1+ #26 Not tainted [ 2566.459194] ------------------------------------------------------- [ 2566.459266] gem_streaming_w/759 is trying to acquire lock: [ 2566.459334] (&obj->mm.lock){+.+.+.}, at: [<ffffffffa034bc80>] i915_gem_object_pin_pages+0x0/0xc0 [i915] [ 2566.459605] [ 2566.459605] but task is already holding lock: [ 2566.459699] (&mm->mmap_sem){++++++}, at: [<ffffffff8106fd11>] __do_page_fault+0x121/0x500 [ 2566.459814] [ 2566.459814] which lock already depends on the new lock. [ 2566.459814] [ 2566.459934] [ 2566.459934] the existing dependency chain (in reverse order) is: [ 2566.460030] [ 2566.460030] -> #1 (&mm->mmap_sem){++++++}: [ 2566.460139] lock_acquire+0xfe/0x220 [ 2566.460214] down_read+0x4e/0x90 [ 2566.460444] i915_gem_userptr_get_pages+0x6e/0x340 [i915] [ 2566.460669] ____i915_gem_object_get_pages+0x8b/0xd0 [i915] [ 2566.460900] __i915_gem_object_get_pages+0x6a/0x80 [i915] [ 2566.461132] __i915_vma_do_pin+0x7fa/0x930 [i915] [ 2566.461352] eb_add_vma+0x67b/0x830 [i915] [ 2566.461572] eb_lookup_vmas+0xafe/0x1010 [i915] [ 2566.461792] i915_gem_do_execbuffer+0x715/0x2870 [i915] [ 2566.462012] i915_gem_execbuffer2+0x106/0x2b0 [i915] [ 2566.462152] drm_ioctl+0x36c/0x670 [drm] [ 2566.462236] do_vfs_ioctl+0x12c/0xa60 [ 2566.462317] SyS_ioctl+0x41/0x70 [ 2566.462399] entry_SYSCALL_64_fastpath+0x1c/0xb1 [ 2566.462477] [ 2566.462477] -> #0 (&obj->mm.lock){+.+.+.}: [ 2566.462587] __lock_acquire+0x1602/0x1790 [ 2566.462661] lock_acquire+0xfe/0x220 [ 2566.462893] i915_gem_object_pin_pages+0x4c/0xc0 [i915] [ 2566.463116] i915_gem_fault+0x2c2/0x8c0 [i915] [ 2566.463197] __do_fault+0x42/0x130 [ 2566.463276] __handle_mm_fault+0x92c/0x1280 [ 2566.463356] handle_mm_fault+0x1e2/0x440 [ 2566.463443] __do_page_fault+0x1c4/0x500 [ 2566.463529] do_page_fault+0xc/0x10 [ 2566.463613] page_fault+0x1f/0x30 [ 2566.463693] [ 2566.463693] other info that might help us debug this: [ 2566.463693] [ 2566.463820] Possible unsafe locking scenario: [ 2566.463820] [ 2566.463918] CPU0 CPU1 [ 2566.463988] ---- ---- [ 2566.464068] lock(&mm->mmap_sem); [ 2566.464143] lock(&obj->mm.lock); [ 2566.464226] lock(&mm->mmap_sem); [ 2566.464304] lock(&obj->mm.lock); [ 2566.464378] [ 2566.464378] *** DEADLOCK *** [ 2566.464378] [ 2566.464504] 1 lock held by gem_streaming_w/759: [ 2566.464576] #0: (&mm->mmap_sem){++++++}, at: [<ffffffff8106fd11>] __do_page_fault+0x121/0x500 [ 2566.464699] [ 2566.464699] stack backtrace: [ 2566.464801] CPU: 0 PID: 759 Comm: gem_streaming_w Not tainted 4.11.0-rc1+ #26 [ 2566.464881] Hardware name: GIGABYTE GB-BXBT-1900/MZBAYAB-00, BIOS F8 03/02/2016 [ 2566.464983] Call Trace: [ 2566.465061] dump_stack+0x68/0x9f [ 2566.465144] print_circular_bug+0x20b/0x260 [ 2566.465234] __lock_acquire+0x1602/0x1790 [ 2566.465323] ? debug_check_no_locks_freed+0x1a0/0x1a0 [ 2566.465564] ? i915_gem_object_wait+0x238/0x650 [i915] [ 2566.465657] ? debug_lockdep_rcu_enabled.part.4+0x1a/0x30 [ 2566.465749] lock_acquire+0xfe/0x220 [ 2566.465985] ? i915_sg_trim+0x1b0/0x1b0 [i915] [ 2566.466223] i915_gem_object_pin_pages+0x4c/0xc0 [i915] [ 2566.466461] ? i915_sg_trim+0x1b0/0x1b0 [i915] [ 2566.466699] i915_gem_fault+0x2c2/0x8c0 [i915] [ 2566.466939] ? i915_gem_pwrite_ioctl+0xce0/0xce0 [i915] [ 2566.467030] ? __lock_acquire+0x642/0x1790 [ 2566.467122] ? __lock_acquire+0x642/0x1790 [ 2566.467209] ? debug_lockdep_rcu_enabled+0x35/0x40 [ 2566.467299] ? get_unmapped_area+0x1b4/0x1d0 [ 2566.467387] __do_fault+0x42/0x130 [ 2566.467474] __handle_mm_fault+0x92c/0x1280 [ 2566.467564] ? __pmd_alloc+0x1e0/0x1e0 [ 2566.467651] ? vm_mmap_pgoff+0x160/0x190 [ 2566.467740] ? handle_mm_fault+0x111/0x440 [ 2566.467827] handle_mm_fault+0x1e2/0x440 [ 2566.467914] ? handle_mm_fault+0x5d/0x440 [ 2566.468002] __do_page_fault+0x1c4/0x500 [ 2566.468090] do_page_fault+0xc/0x10 [ 2566.468180] page_fault+0x1f/0x30 [ 2566.468263] RIP: 0033:0x557895ced32a [ 2566.468337] RSP: 002b:00007fffd6dd8a10 EFLAGS: 00010202 [ 2566.468419] RAX: 00007f659a4db000 RBX: 0000000000000003 RCX: 00007f659ad032da [ 2566.468501] RDX: 0000000000000000 RSI: 0000000000100000 RDI: 0000000000000000 [ 2566.468586] RBP: 0000000000000007 R08: 0000000000000003 R09: 0000000100000000 [ 2566.468667] R10: 0000000000000001 R11: 0000000000000246 R12: 0000557895ceda60 [ 2566.468749] R13: 0000000000000001 R14: 00007fffd6dd8ac0 R15: 00007f659a4db000 By checking the status of the gup worker (serialized by the obj->mm.lock) we can determine whether it is still active, has failed or has succeeded. If the worker is still active (or failed), we know that it cannot be bound and so we can skip taking struct_mutex (risking potential recursion). As we check the worker status, we mark it to discard any partial results, forcing us to restart on the next get_pages. Reported-by:Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Fixes: 1c8782dd ("drm/i915/userptr: Disallow wrapping GTT into a userptr") Testcase: igt/gem_userptr_blits/map-fixed-invalidate-gup Testcase: igt/gem_userptr_blits/dmabuf-sync Signed-off-by:
-
Michal Wajdeczko authored
After negative guc fw selection we could leave guc submission flag still turned on. Reorder some checks to cover this case. While here, fix info message and return early if there is no Guc. Signed-off-by:
Michal Wajdeczko <michal.wajdeczko@intel.com> Cc: Arkadiusz Hiler <arkadiusz.hiler@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Reviewed-by:
Arkadiusz Hiler <arkadiusz.hiler@intel.com> [tursulin: fixup bad alignment] Signed-off-by:
-
Arkadiusz Hiler authored
This field is used to determine which kind of firmware the struct describes (GuC/HuC) - the name does not reflect. The enum used here have "type" in the name, so let's go with that. Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Oscar Mateo <oscar.mateo@intel.com> Signed-off-by:
-
Chris Wilson authored
Tvrtko spotted a stale reference to b->lock (now b->rb_lock) so review the comments and try to improve them in passing. Signed-off-by:
-
- 15 Mar, 2017 4 commits
-
-
Chris Wilson authored
Check that request has not been signaled before acquiring a reference to the request for signaling later in the interrupt handler. The loading of the cacheline (for request->fence.flags) should be "free" when followed by the locked increment of the request->fence.refcount (which then sets the cacheline to exclusive mode), i.e. the cost of test_bit prior to an atomic_inc should be negligible. This should benefit us when we have a pile of bare breadcrumbs (interrupted execbuf) where we may get interrupts faster than we can get rid of the intel_wait, or if the device is too slow to run the bottom-half between interrupts. Signed-off-by:
-
Chris Wilson authored
We need to ensure that we always serialize updates to the bottom-half using the breadcrumbs.irq_lock so that we don't race with a concurrent interrupt handler. This is most important just prior to leaving the waiter (when the intel_wait will be overwritten), so make sure we are not the current bottom-half when skipping the irq locks. Signed-off-by:
-
Chris Wilson authored
Before walking the rbtree of waiters (marking them as complete and waking them), decouple the interrupt handler. This prevents a race between the missed waiter waking up and removing its intel_wait (which skips checking the lock) and the interrupt handler dereferencing the intel_wait. (Though we do not expect to encounter waiters during idle!) Fixes: e1c0c91b ("drm/i915: Wake up all waiters before idling") Signed-off-by:
-
Chris Wilson authored
When adding a new request to the breadcrumb rbtree, we mark all those requests inside the rbtree that are already completed as complete. This wakes those waiters up and allows them to skip the spinlock before returning to userspace. If one of those is the current bottom-half and allocated its intel_wait on the stack, it may then overwrite the b->irq_wait upon exiting i915_wait_request() just as the interrupt handler dereferences it. Fixes: 56299fb7 ("drm/i915: Signal first fence from irq handler if complete") Signed-off-by:
-
