- 03 Mar, 2017 34 commits
-
-
Chris Wilson authored
Being inside a spinlock signaling that the hardware just completed a request doesn't prevent a second thread already spotting that the request is complete, freeing it and reallocating it! The code currently tries to prevent this using RCU -- but that only prevents the request from being freed, it doesn't prevent us from reallocating it - that requires us to take a reference. [ 206.922985] BUG: spinlock already unlocked on CPU#4, gem_exec_parall/7796 [ 206.922994] lock: 0xffff8801c6047120, .magic: dead4ead, .owner: <none>/-1, .owner_cpu: -1 [ 206.923000] CPU: 4 PID: 7796 Comm: gem_exec_parall Not tainted 4.10.0-CI-Patchwork_4008+ #1 [ 206.923006] Hardware name: System manufacturer System Product Name/Z170M-PLUS, BIOS 1805 06/20/2016 [ 206.923012] Call Trace: [ 206.923014] <IRQ> [ 206.923019] dump_stack+0x67/0x92 [ 206.923023] spin_dump+0x73/0xc0 [ 206.923027] do_raw_spin_unlock+0x79/0xb0 [ 206.923031] _raw_spin_unlock_irqrestore+0x27/0x60 [ 206.923042] dma_fence_signal+0x160/0x230 [ 206.923060] notify_ring+0xae/0x2e0 [i915] [ 206.923073] ? ibx_hpd_irq_handler+0xc0/0xc0 [i915] [ 206.923086] gen8_gt_irq_handler+0x219/0x290 [i915] [ 206.923100] gen8_irq_handler+0x8e/0x6b0 [i915] [ 206.923105] __handle_irq_event_percpu+0x58/0x370 [ 206.923109] handle_irq_event_percpu+0x1e/0x50 [ 206.923113] handle_irq_event+0x34/0x60 [ 206.923117] handle_edge_irq+0xbe/0x150 [ 206.923122] handle_irq+0x15/0x20 [ 206.923126] do_IRQ+0x63/0x130 [ 206.923142] ? i915_mutex_lock_interruptible+0x39/0x140 [i915] [ 206.923148] common_interrupt+0x90/0x90 [ 206.923153] RIP: 0010:osq_lock+0x77/0x110 [ 206.923157] RSP: 0018:ffffc90001cabaa0 EFLAGS: 00000246 ORIG_RAX: ffffffffffffff6e [ 206.923164] RAX: 0000000000000000 RBX: ffff880236d1abc0 RCX: ffff8801ef642fc0 [ 206.923169] RDX: ffff8801ef6427c0 RSI: ffffffff81c6e7fd RDI: ffffffff81c7c848 [ 206.923175] RBP: ffffc90001cabab8 R08: 00000000692bb19b R09: 08c1493200000000 [ 206.923180] R10: 0000000000000001 R11: 0000000000000001 R12: ffff880236cdabc0 [ 206.923185] R13: ffff8802207f00b0 R14: ffffffffa00b7cd9 R15: ffff8802207f0070 [ 206.923191] </IRQ> [ 206.923206] ? i915_mutex_lock_interruptible+0x39/0x140 [i915] [ 206.923213] __mutex_lock+0x649/0x990 [ 206.923217] ? __mutex_lock+0xb0/0x990 [ 206.923221] ? _raw_spin_unlock+0x2c/0x50 [ 206.923226] ? __pm_runtime_resume+0x56/0x80 [ 206.923242] ? i915_mutex_lock_interruptible+0x39/0x140 [i915] [ 206.923249] mutex_lock_interruptible_nested+0x16/0x20 [ 206.923264] i915_mutex_lock_interruptible+0x39/0x140 [i915] [ 206.923270] ? __pm_runtime_resume+0x56/0x80 [ 206.923285] i915_gem_do_execbuffer.isra.15+0x442/0x1d10 [i915] [ 206.923291] ? __lock_acquire+0x449/0x1b50 [ 206.923296] ? __might_fault+0x3e/0x90 [ 206.923301] ? __might_fault+0x87/0x90 [ 206.923305] ? __might_fault+0x3e/0x90 [ 206.923320] i915_gem_execbuffer2+0xb5/0x220 [i915] [ 206.923327] drm_ioctl+0x200/0x450 [ 206.923341] ? i915_gem_execbuffer+0x330/0x330 [i915] [ 206.923348] do_vfs_ioctl+0x90/0x6e0 [ 206.923352] ? __fget+0x108/0x200 [ 206.923356] ? expand_files+0x2b0/0x2b0 [ 206.923361] SyS_ioctl+0x3c/0x70 [ 206.923365] entry_SYSCALL_64_fastpath+0x1c/0xb1 [ 206.923369] RIP: 0033:0x7fdd75fc6357 [ 206.923373] RSP: 002b:00007fdd20e59bf8 EFLAGS: 00000246 ORIG_RAX: 0000000000000010 [ 206.923380] RAX: ffffffffffffffda RBX: ffffffff81481ff3 RCX: 00007fdd75fc6357 [ 206.923385] RDX: 00007fdd20e59c70 RSI: 0000000040406469 RDI: 0000000000000003 [ 206.923390] RBP: ffffc90001cabf88 R08: 0000000000000040 R09: 00000000000003f7 [ 206.923396] R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000 [ 206.923401] R13: 0000000000000003 R14: 0000000040406469 R15: 0000000001cf9cb0 [ 206.923408] ? __this_cpu_preempt_check+0x13/0x20 Fixes: 56299fb7 ("drm/i915: Signal first fence from irq handler if complete") Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100051Signed-off-by:
Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170303144557.4815-1-chris@chris-wilson.co.ukReviewed-by:
Mika Kuoppala <mika.kuoppala@linux.intel.com>
-
Ville Syrjälä authored
Add tracepoints for display FIFO underruns. Makes it more convenient to correlate the underruns with other display tracepoints. v2: s/i915/intel/ in the tracepoint name Signed-off-by:
Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by:
Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20170302171508.1666-19-ville.syrjala@linux.intel.com
-
Ville Syrjälä authored
Add a tracepoint for observing changes in the cxsr state. The tracepoint will dump out the frame and scanline counters for each pipe so that the information can be compared with eg. plane update tracepoints. Signed-off-by:
-
Ville Syrjälä authored
Add tracepoints for observing the WM/FIFO programming on VLV/CHV. When compared with the plane and pipe update tracepoints this can be used to verify that everything is performed in the right sequence. Signed-off-by:
-
Ville Syrjälä authored
Add tracepoints for plane programming. The tracepoints will dump the frame and scanline counters, so this can be used to verify eg. that the plane gets reprogrammed at the right time with respect to watermark programming (if we have appropriate tracepoints for that as well). v2: Rebase due to legacy cursor changes Signed-off-by:
-
Ville Syrjälä authored
We now compute the watermarks correctly, so just return an error if we can't support the configuration. Signed-off-by:
-
Ville Syrjälä authored
On VLV/CHV enabling sprite0 when sprite1 has already been enabled may lead to an underrun. This only happens when sprite0 FIFO size is zero prior to enabling it. Hence an effective workaround is to always allocate at least one cacheline for sprite0 when sprite1 is active. I've not observed this sort of failure during any other type of plane enable/disable sequence. v2: s/noninverted/raw/ for consistency with other platforms Testcase: igt/kms_plane_blinker Signed-off-by:
-
Ville Syrjälä authored
Clear out the watermark for all disabled planes to 0. This is required to avoid falsely thinking that the inherited watermarks are bogus in case the watermark is actually higher than the FIFO size. v2: s/noninverted/raw/ for consistency with other platforms Signed-off-by:
-
Ville Syrjälä authored
Now that vlv/chv have more proper wm programming support, let's reduce the the update_wm_{pre,post} flags to only cover the pre-ilk platforms. When we finally convert those as well we can drop these flags entirely. Signed-off-by: -
Ville Syrjälä authored
Remove crtc->wm.cxsr_allowed and just rely on crtc_state->disable_cxsr instead. This was used only by vlv/chv to indicate whether to enable cxsr in the wm computation. That doesn't really work anymore, and as far as the optimal watermarks go we'll just consider the number of planes and the current pipe, and for the intermediate watermarks we'll also start to consider disable_cxsr which is set appropriately when planes are being enabled/disabled. We'll also flip over the crtc_state->wm.need_postvbl_update setup so that it's the wm code that will set it. Previously the generic code set it up, and then the wm code cleared it again if it thought it's not needed after all. Signed-off-by:
-
Ville Syrjälä authored
Since the watermark registers arent double buffered on VLV/CHV, we'll need to play around with intermediate watermarks same was as we do on ILK-BDW. The watermark registers on VLV/CHV contain inverted values, so to find the intermediate watermark value we just take the minimum of the active and optimal values. This also means that, unlike ILK-BDW, there's no chance that we'd fail to find a working intermediate watermarks. As long as both the active and optimal watermarks are valid the intermediate watermarks will come out valid as well. Signed-off-by:
-
Ville Syrjälä authored
Check whether anything relevant has actually change when we compute new watermarks for each plane in the state. If the watermarks for no primary/sprite planes changed we don't have to recompute the FIFO split or reprogram the DSBARB registers. And even the cursor watermarks didn't change we can skip the merge+invert step between all the planes on the pipe as well. v2: s/noninverted/raw/ for consistency with other platforms v3: Drop duplicated vlv_get_fifo_size() call during init Signed-off-by:
-
Ville Syrjälä authored
Start computing the vlv/chv watermarks the atomic way, from the .compute_pipe_wm() hook. We'll recompute the actual watermarks for only planes that are part of the state, the other planes will keep their watermark from the last time it was computed. And the actual watermark programming will happen from the .initial_watermarks() hook. For now we'll just compute the optimal watermarks, and we'll hook up the intermediate watermarks properly later. The DSPARB registers responsible for the FIFO paritioning are double buffered, so they will be programming from intel_begin_crtc_commit(). v2: s/noninverted/raw/ for consistency with other platforms s/vlv_plane_wm_set/vlv_raw_plane_wm_set/ for clarity Signed-off-by: -
Ville Syrjälä authored
Let's compute the watermarks first and the FIFO size second. This way we can make sure the FIFO split is the most accommodating to the watermarks. Previously we could have potentially computed a FIFO split that couldn't accommodate the PM2 watermarks simply due to a bad split even if the total FIFO size would have been sufficient. It'll also allow us to avoid recomputing the wms for all planes whenever the FIFO split would change. Thus we don't have to add any extra planes to the state when the FIFO needs to be repartitioned. To help with this we'll keep around copies of the non-inverted watermarks in the crtc state. For now that doesn't help too much, but once we start to do the watermark computation only for the planes that change we'll need the non-inverted values around for the other planes. v2: s/noninverted/raw/ for consistency with other platforms Fix the memset() of the "raw" watermarks Signed-off-by: -
Ville Syrjälä authored
Move the vlv/chv FIFO size tracking into the crtc_state. As with the wms for now this just acts as temporary storage. Signed-off-by:
-
Ville Syrjälä authored
Relocate the vlv/chv wm state to live under intel_crtc_state. Note that for now this just behaves as a temporary storage. But it'll be easier to conver the thing over to properly pre-computing the state when it's already in the right place. Signed-off-by:
-
Ville Syrjälä authored
In an effort to make the vlv/chv wm code look and behave more like the ilk+ code, let's move the current active wms next to the corresponding ilk wms. Signed-off-by:
-
Ville Syrjälä authored
Track the plane fifo sizes under intel_crtc instead of under each intel_plane. Avoids looping over the planes in a bunch of places, and later we'll move this tracking into the crtc state properly. v2: Nuke intel_plane_wm_parameters (Maarten) Signed-off-by:
-
Ville Syrjälä authored
In a lot of place we wish to know which planes on the crtc are actually visible, or how many of them there are. Let's start tracking that in a bitmask in the crtc state. We already track enabled planes (ie. ones with an fb and crtc specified by the user) but that's not quite the same thing as enabled planes may still end up being invisible due to clipping and whatnot. Signed-off-by:
-
Mika Kuoppala authored
If we manage to tangle errorpaths and get call to callbacks, it is better to defensively keep them as null until object init is finished so that we get clean null deref on callsite, instead of more cryptic wreckage with partly initialized vm objects. Signed-off-by:
Mika Kuoppala <mika.kuoppala@intel.com> Reviewed-by:
-
Mika Kuoppala authored
The term legacy is subjective. Use 3lvl and 4lvl where appropriate. Signed-off-by:
-
Mika Kuoppala authored
If we setup the vm size early, we can use the newly introduced i915_vm_is_48bit() in majority of callsites wanting to know the vm size. As we operate either with 3lvl or 4lvl page table structure, wrap the vm size query inside a function which tells us if 4lvl setup is needed for particular vm, as the following code uses the function names where level is noted. v2: use_4lvl (Chris) Signed-off-by:
-
Mika Kuoppala authored
Don't mark pdps clear if never do the necessary actions with the hardware to make them clear. v2: totally get rid of confusing ppgtt bool (Chris) Signed-off-by:
-
Mika Kuoppala authored
The macro takes a vm pointer at some sites, and dev_priv on others We were saved as the internal macro never deferences the pointer given. As the number of pdpes depend on vm configuration, make it as a inline function that accepts vm pointer. Cc: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by:
Chris Wilson <chris@chris-wsilon.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1488295691-9404-1-git-send-email-mika.kuoppala@intel.com
-
Chris Wilson authored
The code to check for execlists completion is generic, so move it to intel_engine_cs.c, where we can reuse the new intel_engine_is_idle(). Signed-off-by:
-
Chris Wilson authored
During reset_all_global_seqno() on seqno rollover, we have to update the HWS. This causes all in flight requests to be completed, so first we wait. However, we were only waiting for the requests themselves to be completed and clearing out the waiter rbtrees - what I had missed was the extra reference in execlists->port[]. Since commit fe9ae7a3 ("drm/i915/execlists: Detect an out-of-order context switch") we can detect when the request is retired before the context switch interrupt is completed. The impact should be neglible outside of debugging. Testcase: igt/gem_exec_whisper Signed-off-by:
-
Ander Conselvan de Oliveira authored
The logic to enable a DDI in intel_mst_pre_enable_dp() is essentially the same as in intel_ddi_pre_enable_dp(). So reuse the latter function by calling the post_disable hook on the intel_dig_port instead of duplicating that code. v2: Don't oops because of a NULL encoder->crtc. (Ville) v3: Warn for MST + PORT_E too. (Ville) Cc: Imre Deak <imre.deak@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by:
Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com> Reviewed-by:
-
Ander Conselvan de Oliveira authored
Remove direct usages of intel_crtc->config from the DDI code. Functions that didn't yet take a pipe_config as an argument were coverted to do so. v2: s/pipe_config/const crtc_state/ (Ville) - take crtc from crtc_state. (Ville) Signed-off-by:
-
Ander Conselvan de Oliveira authored
Pass intel_crtc to functions intel_ddi_enable_transcoder_func(), intel_ddi_set_pipe_settings() and intel_ddi_set_vc_payload_alloc(), instead of the generic crtc type. By changing the functions intel_ddi_get_crtc_encoder() so that it receives an intel_crtc parameter, there is no need for the drm_crtc in the callers. Signed-off-by:
-
Ander Conselvan de Oliveira authored
It is preferred to pass pipe_config to functions instead of accessing crtc->config directly. Follow suit and pass pipe_config to the fdi link train functions. v2: Add const; s/pipe_config/crtc_state/ (Ville) Signed-off-by:
-
Ander Conselvan de Oliveira authored
Using crtc->config directly is being removed in favor of passing a pipe_config. Follow the trend and pass pipe_config to pch_enable() functions. v2: s/pipe_config/crtc_state/ (Ville) - constify crtc_state. (Ville) - take crtc from crtc_state. (Ville) Signed-off-by:
-
Ander Conselvan de Oliveira authored
The function intel_lpt_pch_enable() needs an intel_crtc so pass that instead of the generic crtc type. Signed-off-by:
-
Ander Conselvan de Oliveira authored
The implementation of the fdi_link_train() hooks need an intel_crtc so just pass that instead of the generic crtc type. Signed-off-by:
-
Chris Wilson authored
Check timer_pending() as well as work_pending() to see if the timer for the hangcheck has already expired and the work is pending execution on some list somewhere. v2: Use a more compact if-chain Signed-off-by:
-
- 02 Mar, 2017 6 commits
-
-
Chris Wilson authored
Adding to the tail of the client request list as the only other user is in the throttle ioctl that iterates forwards over the list. It only needs protection against deletion of a request as it reads it, it simply won't see a new request added to the end of the list, or it would be too early and rejected. We can further reduce the number of spinlocks required when throttling by removing stale requests from the client_list as we throttle. Signed-off-by:
-
Ville Syrjälä authored
Currently ILK-BDW explicitly disable LP1+ watermarks from their .init_clock_gating() hooks. Unfortunately that hook gets called way too late since by that time we've already initialized all the watermark state tracking which then gets out of sync with the hardware state. We may eventually want to consider killing off the explicit LP1+ disable from .init_clock_gating(). In the meantime however, we can avoid the problem by reordering the init sequence such that intel_modeset_init_hw()->intel_init_clock_gating() gets called prior to the hardware state takeover. I suppose prior to the two stage watermark programming we were magically saved by something that forced the watermarks to be reprogrammed fully after .init_clock_gating() got called. But now that no longer happens. Note that the diff might look a bit odd as it kills off one call of intel_update_cdclk(), but that's fine because intel_modeset_init_hw() does the exact same thing. Previously we just did it twice. Actually even this new init sequence is pretty bogus as .init_clock_gating() really should be called before any gem hardware init since it can configure various clock gating workarounds and whatnot that affect the GT side as well. Also intel_modeset_init() really should get split up into better defined init stages. Another "fun" detail is that intel_modeset_gem_init() is where RPS/RC6 gets configured. Why that is done from the display code is beyond me. I've decided to leave all this be for now, and just try to fix the init sequence enough for watermarks to work. Cc: stable@vger.kernel.org Cc: Gabriele Mazzotta <gabriele.mzt@gmail.com> Cc: David Purton <dcpurton@marshwiggle.net> Cc: Matt Roper <matthew.d.roper@intel.com> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Reported-by:
Gabriele Mazzotta <gabriele.mzt@gmail.com> Reported-by:
David Purton <dcpurton@marshwiggle.net> Tested-by:
Daniel Vetter <daniel.vetter@ffwll.ch>
-
Chris Wilson authored
Useful for double checking that the device is powered up when it hung, include both the status of the power management and our rpm wakelock. Signed-off-by:
-
Chris Wilson authored
Whilst investigating some mysterious failures with hangcheck not running during gem_busy/basic-hang-default, the question is why did we decide to cancel the retire_work (which queues the hangcheck)? That decision is based around GT activity, so include that information in the debug report. v2: Include the GT awake status in the error state Signed-off-by:
-
Chris Wilson authored
__i915_guc_submit may be, despite my assertion, called from outside of an irq-safe spinlock so we need to use a full spin_lock_irqsave and not cheat using a spin_lock. (The initial notify callback from the completed fence is called before the spinlock is taken to wake up all waiters and call their callbacks.) [ 48.166581] kernel BUG at drivers/gpu/drm/i915/i915_guc_submission.c:527! [ 48.166617] invalid opcode: 0000 [#1] PREEMPT SMP [ 48.166644] Modules linked in: i915 prime_numbers x86_pkg_temp_thermal intel_powerclamp coretemp crct10dif_pclmul crc32_pclmul ghash_clmulni_intel mei_me mei i2c_i801 netconsole i2c_hid [last unloaded: i915] [ 48.166733] CPU: 2 PID: 5 Comm: kworker/u8:0 Tainted: G U 4.10.0nightly-170302-guc_scrub+ #19 [ 48.166778] Hardware name: /NUC6i5SYB, BIOS SYSKLi35.86A.0054.2016.0930.1102 09/30/2016 [ 48.166835] Workqueue: i915 __intel_autoenable_gt_powersave [i915] [ 48.166865] task: ffff88084ab7cf40 task.stack: ffffc90000064000 [ 48.166921] RIP: 0010:__i915_guc_submit+0x1e6/0x2a0 [i915] [ 48.166953] RSP: 0018:ffffc90000067c80 EFLAGS: 00010202 [ 48.166979] RAX: 0000000000000202 RBX: ffff8808465e0c68 RCX: 0000000000000201 [ 48.167016] RDX: 0000000080000201 RSI: ffff88084ab7d798 RDI: ffff88082b8a8040 [ 48.167054] RBP: ffffc90000067cd8 R08: 0000000000000001 R09: 0000000000000000 [ 48.167085] R10: 0000000000000000 R11: 0000000000000000 R12: ffff88082b8a8148 [ 48.167126] R13: 0000000000000000 R14: ffff88082f440000 R15: ffff88082e85e660 [ 48.167156] FS: 0000000000000000(0000) GS:ffff88086ed00000(0000) knlGS:0000000000000000 [ 48.167195] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 48.167226] CR2: 000055862ffcdc2c CR3: 0000000001e0f000 CR4: 00000000003406e0 [ 48.167257] Call Trace: [ 48.168112] ? trace_hardirqs_on+0xd/0x10 [ 48.168966] ? _raw_spin_unlock_irqrestore+0x4a/0x80 [ 48.169831] i915_guc_submit+0x1a/0x20 [i915] [ 48.170680] submit_notify+0x89/0xc0 [i915] [ 48.171512] __i915_sw_fence_complete+0x175/0x220 [i915] [ 48.172340] i915_sw_fence_complete+0x2a/0x50 [i915] [ 48.173158] i915_sw_fence_commit+0x21/0x30 [i915] [ 48.173968] __i915_add_request+0x238/0x530 [i915] [ 48.174764] __intel_autoenable_gt_powersave+0x8b/0xb0 [i915] [ 48.175549] process_one_work+0x218/0x690 [ 48.176318] ? process_one_work+0x197/0x690 [ 48.177183] worker_thread+0x4e/0x4a0 [ 48.178039] kthread+0x10c/0x140 [ 48.178878] ? process_one_work+0x690/0x690 [ 48.179718] ? kthread_create_on_node+0x40/0x40 [ 48.180568] ret_from_fork+0x31/0x40 [ 48.181423] Code: 02 00 00 43 89 84 ae 50 11 00 00 e8 75 01 62 e1 48 83 c4 30 5b 41 5c 41 5d 41 5e 41 5f 5d c3 48 c1 e0 20 48 09 c2 49 89 d0 eb 82 <0f> 0b 0f 0b 0f 0b 0f 0b 0f 0b 0f 0b 49 c1 e8 20 44 89 43 34 4a [ 48.183336] RIP: __i915_guc_submit+0x1e6/0x2a0 [i915] RSP: ffffc90000067c80 Reported-by:
Arkadiusz Hiler <arkadiusz.hiler@intel.com> Fixes: 349ab919 ("drm/i915/guc: Make wq_lock irq-safe") Fixes: 67b807a8 ("drm/i915: Delay disabling the user interrupt for breadcrumbs") Signed-off-by:
Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by:
-
Chris Wilson authored
assert_spin_locked() becomes an unconditionally compiled BUG_ON(), adding debug code right into the heart of critical routines like interrupt handlers. text data bss dec hex 1296480 19944 2272 1318696 141f28 before (lockdep disabled) 1295984 19944 2272 1318200 141d38 after 1336261 21139 3208 1360608 14c2e0 before (lockdep enabled) 1339920 21139 3208 1364267 14d12b after Small saving for release; hopefully more instructive in debug. Signed-off-by:
Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
-
