- 29 Nov, 2016 15 commits
-
-
Chris Wilson authored
In order to avoid some complexity in trying to reconstruct the workqueues across reset, remember them instead. The issue comes when we have to handle a reset between request allocation and submission, the request has reserved space in the wq, but is not in any list so we fail to restore the reserved space. By keeping the execbuf client intact across the reset, we also keep the reservations. Signed-off-by:
Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by:
Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20161129121024.22650-5-chris@chris-wilson.co.uk
-
Chris Wilson authored
Set the initial value of the doorbell cookie from the client. Signed-off-by:
-
Chris Wilson authored
The client->cookie is a shadow of the doorbell->cookie value, so rename it to indicate its association with the doorbell, like the doorbell id and offset. Signed-off-by:
-
Chris Wilson authored
i915_guc_info() (part of debugfs output) tries to avoid holding struct_mutex for a long period by copying onto the stack. This causes a warning that the stack frame is massive, so stop doing that. We can even forgo holding the struct_mutex here as that doesn't serialise the values being read (and the lists used exist for the device lifetime). v2: Skip printing anything if guc->execbuf_client is disabled (avoids potential NULL dereference). Signed-off-by:
-
Libin Yang authored
This patch adds support for DP MST audio in i915. Enable audio codec when DP MST is enabled if has_audio flag is set. Disable audio codec when DP MST is disabled if has_audio flag is set. Another separated patches to support DP MST audio will be implemented in audio driver. This patch is ported from commit 3708d5e0 ("drm/i915: start adding dp mst audio") And because commit 3708d5e0 ("drm/i915: start adding dp mst audio") breaks MST multi-monitor setups on some platforms, the orignal patch is reverted by commit be754b10 ("Revert "drm/i915: start adding dp mst audio"") As the multi-monitor setups issue is fixed, let's port the patch and enable the dp mst audio. Signed-off-by:
Libin Yang <libin.yang@intel.com> Cc: Lyude <cpaul@redhat.com> Cc: Jani Nikula <jani.nikula@intel.com> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com> Reviewed-by:
Jani Nikula <jani.nikula@intel.com> Signed-off-by:
-
Libin Yang authored
Prepare for using the same code for judging ddi being audio enabled. No functional changes. Signed-off-by:
Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com> Signed-off-by:
Lyude <cpaul@redhat.com> Reviewed-by:
-
Libin Yang authored
Add the DP MST info dump in debugfs. Signed-off-by:
-
Chris Wilson authored
drivers/gpu/drm/i915/./i915_trace.h: In function ‘trace_event_raw_event_i915_gem_evict’: drivers/gpu/drm/i915/./i915_trace.h:409:24: error: ‘struct i915_address_space’ has no member named ‘dev’ __entry->dev = vm->dev->primary->index; A couple of macros missed in the s/vm->dev/vm->i915/ conversion. Fixes: 49d73912 ("drm/i915: Convert vm->dev backpointer to vm->i915") Signed-off-by:Michał Winiarski <michal.winiarski@intel.com>
-
Chris Wilson authored
Just a couple of naked 64bit divides causing link errors on 32bit builds, with: ERROR: "__udivdi3" [drivers/gpu/drm/i915/i915.ko] undefined! v2: do_div() is only u64/u32, we need a u32/u64! v3: div_u64() == u64/u32, div64_u64() == u64/u64 Reported-by:
kbuild test robot <fengguang.wu@intel.com> Fixes: d7965152 ("drm/i915: Enable i915 perf stream for Haswell OA unit") Signed-off-by:
Robert Bragg <robert@sixbynine.org>
-
Chris Wilson authored
99% of the time we access i915_address_space->dev we want the i915 device and not the drm device, so let's store the drm_i915_private backpointer instead. The only real complication here are the inlines in i915_vma.h where drm_i915_private is not yet defined and so we have to choose an alternate path for our asserts. Signed-off-by:
Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
-
Jérémy Lefaure authored
Two warnings are produced by gcc (tested with gcc 6.2.1): drivers/gpu/drm/i915/intel_csr.c: In function ‘csr_load_work_fn’: drivers/gpu/drm/i915/intel_csr.c:400:5: error: ‘fw’ is used uninitialized in this function [-Werror=uninitialized] if (fw) ^ and In file included from drivers/gpu/drm/i915/i915_drv.h:47:0, from drivers/gpu/drm/i915/intel_guc_loader.c:30: drivers/gpu/drm/i915/intel_guc_loader.c: In function ‘intel_guc_init’: ./include/drm/drmP.h:228:2: error: ‘fw’ may be used uninitialized in this function -Werror=maybe-uninitialized] drm_printk(KERN_DEBUG, DRM_UT_DRIVER, fmt, ##__VA_ARGS__) ^~~~~~~~~~ drivers/gpu/drm/i915/intel_guc_loader.c:595:25: note: ‘fw’ was declared here const struct firmware *fw; ^~ When CONFIG_DRM_I915_WERROR is set, those warnings break the build. Initializing fw pointer to NULL in both cases removes the warnings. Signed-off-by:Jérémy Lefaure <jeremy.lefaure@lse.epita.fr> Signed-off-by:
Daniel Vetter <daniel.vetter@ffwll.ch> Link: http://patchwork.freedesktop.org/patch/msgid/20161128234319.20800-1-jeremy.lefaure@lse.epita.fr
-
Zhi Wang authored
a PT page will be released if it doesn't contain any meaningful mappings during PPGTT page table shrinking. The PT entry in the upper level will be set to a scratch entry. Normally this works nicely, but in virtualization world, the PPGTT page table is tracked by hypervisor. Releasing the PT page before modifying the upper level PT entry would cause extra efforts. As the tracked page has been returned to OS before losing track from hypervisor, it could be written in any pattern. Hypervisor has to recognize if a page is still being used as a PT page by validating these writing patterns. It's complicated. Better let the guest modify the PT entry in upper level PT first, then release the PT page. Reviewed-by:
Zhi Wang <zhi.a.wang@intel.com> Link: https://patchwork.freedesktop.org/patch/122697/msgid/1479728666-25333-1-git-send-email-zhi.a.wang@intel.comSigned-off-by:
-
Matthew Auld authored
We grab the struct_mutex in intel_crtc_page_flip, but if we are wedged or a reset is in progress we bail early but never seem to actually release the lock. Fixes: 7f1847eb ("drm/i915: Simplify checking of GPU reset_counter in display pageflips") Cc: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by:
Matthew Auld <matthew.auld@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20161128103648.9235-1-matthew.auld@intel.comReviewed-by:
-
Chris Wilson authored
This reverts commit 27745e82 ("drm/i915/execlists: Use a local lock for dfs_link access") as the struct_mutex was required to prevent concurrent retiring and freeing, now restored in the previous patch. Signed-off-by:
-
Chris Wilson authored
David found another issue with priority bumping from mmioflips, where we are accessing the requests concurrently to them being retired and freed. Whilst we are skipping the dependency if has been submitted, that is not sufficient to stop the dependency from disappearing if another thread retires that request. To prevent we can either employ the struct_mutex (or a request mutex in the future) to serialise retiring before it is freed. Alternatively, we need to keep the dependencies alive using RCU whilst they are being accessed via the DFS. [ 1746.698111] general protection fault: 0000 [#1] PREEMPT SMP [ 1746.698305] Modules linked in: snd_hda_intel snd_hda_codec snd_hwdep x86_pkg_temp_thermal snd_hda_core coretemp crct10dif_pclmul crc32_pclmul snd_pcm ghash_clmulni_intel mei_me mei i915 e1000e ptp pps_core i2c_hid [ 1746.698750] CPU: 1 PID: 6716 Comm: kworker/u8:2 Not tainted 4.9.0-rc6-CI-Nightly_816+ #1 [ 1746.698871] Hardware name: GIGABYTE GB-BKi7A-7500/MFLP7AP-00, BIOS F1 07/27/2016 [ 1746.699125] Workqueue: events_unbound intel_mmio_flip_work_func [i915] [ 1746.699266] task: ffff880260a5e800 task.stack: ffffc90000f6c000 [ 1746.699361] RIP: 0010:[<ffffffffa006595d>] [<ffffffffa006595d>] execlists_schedule+0x8d/0x300 [i915] [ 1746.699632] RSP: 0018:ffffc90000f6fcd8 EFLAGS: 00010206 [ 1746.699724] RAX: dead0000000000f8 RBX: ffff8801f64b2bf0 RCX: ffff8801f64b2c10 [ 1746.699842] RDX: dead000000000100 RSI: 0000000000000000 RDI: ffff8801f64b0458 [ 1746.699972] RBP: ffffc90000f6fd68 R08: ffff88026488dc00 R09: 0000000000000002 [ 1746.700090] R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000400 [ 1746.700195] R13: ffffc90000f6fcf0 R14: ffff88020955aa40 R15: ffff88020955aa68 [ 1746.700307] FS: 0000000000000000(0000) GS:ffff88026dc80000(0000) knlGS:0000000000000000 [ 1746.700435] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 1746.700532] CR2: 0000000002a69e90 CR3: 0000000002c07000 CR4: 00000000003406e0 [ 1746.700635] Stack: [ 1746.700682] ffff880260a5e880 ffffc90000f6fd50 ffffffff810af69a ffffc90000f6fd28 [ 1746.700827] ffff88020955a628 ffff8801e1eaebf0 0000000000000020 0000000000000000 [ 1746.700947] 00000196af1edc96 ffff88025dfa4000 ffff8801f0b030a8 ffffc90000f6fcf0 [ 1746.701071] Call Trace: [ 1746.701117] [<ffffffff810af69a>] ? dequeue_entity+0x25a/0xb50 [ 1746.701260] [<ffffffffa00516be>] fence_set_priority+0x7e/0x80 [i915] [ 1746.701406] [<ffffffffa0051a15>] i915_gem_object_wait_priority+0x85/0x160 [i915] [ 1746.701599] [<ffffffffa008ccd7>] intel_mmio_flip_work_func+0x47/0x2b0 [i915] [ 1746.701717] [<ffffffff81094c4d>] process_one_work+0x14d/0x470 [ 1746.701809] [<ffffffff81094fb3>] worker_thread+0x43/0x4e0 [ 1746.701888] [<ffffffff81094f70>] ? process_one_work+0x470/0x470 [ 1746.701969] [<ffffffff81094f70>] ? process_one_work+0x470/0x470 [ 1746.702072] [<ffffffff8109a4d5>] kthread+0xc5/0xe0 [ 1746.702152] [<ffffffff81771c59>] ? _raw_spin_unlock_irq+0x9/0x10 [ 1746.702234] [<ffffffff8109a410>] ? kthread_park+0x60/0x60 [ 1746.702318] [<ffffffff81772272>] ret_from_fork+0x22/0x30 [ 1746.702387] Code: 89 42 08 48 8b 45 88 48 89 55 c0 4c 89 6d c8 4c 8d 70 d8 4d 8d 7e 28 4d 39 ef 74 72 49 8b 1e 48 8b 13 48 39 d3 48 8d 42 f8 74 3e <48> 8b 10 8b 52 38 41 39 d4 7e 26 48 8b 50 30 48 8b 78 28 48 8d [ 1746.702921] RIP [<ffffffffa006595d>] execlists_schedule+0x8d/0x300 [i915] Nov 25 21:42:54 kbl-gbbki7 kernel: [ 1746.703027] RSP <ffffc90000f6fcd8> Fixes: 27745e82 ("drm/i915/execlists: Use a local lock for dfs_link access") Fixes: 9a151987 ("drm/i915: Add execution priority boosting for mmioflips") Signed-off-by:
-
- 25 Nov, 2016 13 commits
-
-
Arkadiusz Hiler authored
Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Michal Winiarski <michal.winiarski@intel.com> Signed-off-by:
Arkadiusz Hiler <arkadiusz.hiler@intel.com> Reviewed-by:
-
Arkadiusz Hiler authored
send_mutex is used to serialise communication with GuC via intel_guc_send(). Since functions that utilize it are no longer limited to submission, initialization should be handled as a part of general setup. v2: move initialization to *_early() Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Michal Winiarski <michal.winiarski@intel.com> Signed-off-by:
-
Arkadiusz Hiler authored
guc_send(), guc_recv() and related functions were introduced in the i915_guc_submission.c and their scope was limited only to that file. Those are not submission specific though. This patch moves moves them to intel_uc.c with intel_ prefix added. v2: rename intel_guc_log_* functions and clean up intel_guc_send usages Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Michal Winiarski <michal.winiarski@intel.com> Signed-off-by:
-
Arkadiusz Hiler authored
To facilitate code reorganization we are renaming everything that contains guc2host or host2guc. host2guc_action() and host2guc_action_response() become guc_send() and guc_recv() respectively. Other host2guc_*() functions become simply guc_*(). Other entities are renamed basing on context they appear in: - HOST2GUC_ACTIONS_& become INTEL_GUC_ACTION_* - HOST2GUC_{INTERRUPT,TRIGGER} become GUC_SEND_{INTERRUPT,TRIGGER} - GUC2HOST_STATUS_* become INTEL_GUC_STATUS_* - GUC2HOST_MSG_* become INTEL_GUC_RECV_MSG_* - action_lock becomes send_mutex v2: drop unnecessary backslashes and use BIT() instead of '<<' v3: shortened enum names and INTEL_GUC_STATUS_* Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Michal Winiarski <michal.winiarski@intel.com> Signed-off-by: -
Arkadiusz Hiler authored
GuC is not the only one micro controller we have. There are also HuC and DMC. Making the file more general will help with code organization. Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Michal Winiarski <michal.winiarski@intel.com> Signed-off-by:
-
Ander Conselvan de Oliveira authored
The check in __intel_uncore_early_sanitize() to disable decoupled mmio would disable it for every platform that is not broxton. While that's not a problem now since only broxton supports that, simply setting .has_decoupled_mmio in a new platform's device info wouldn't suffice. So avoid future confusion and change the workaround to only change the value of has_decoupled_mmio for broxton. v2: git add compile fix. (Ander) Cc: Praveen Paneri <praveen.paneri@intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Imre Deak <imre.deak@intel.com> Signed-off-by:
Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com> Reviewed-by:
Imre Deak <imre.deak@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1479993807-29353-1-git-send-email-ander.conselvan.de.oliveira@intel.com
-
Ander Conselvan de Oliveira authored
Pass dev_priv to intel_setup_outputs() and functions called by it, since those are all intel i915 specific functions. Also, in the majority of the functions dev_priv is used more often than dev. In the rare cases where there are a few calls back into drm core, a local dev variable was added. v2: Don't convert dev to &dev_priv->drm in intel_dsi_init. (Ville) Signed-off-by:
Ville Syrjälä <ville.syrjala@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1479910904-11005-1-git-send-email-ander.conselvan.de.oliveira@intel.com
-
Chris Wilson authored
Add the tracking required to enable debugobjects for fences to improve error detection in BAT. The debugobject interface lets us track the lifetime and phases of the fences even while being embedded into larger structs, i.e. to check they are not used after they have been released. v2: Don't populate the stubs, debugobjects checks for a NULL pointer and treats it equivalently. Signed-off-by:
-
Chris Wilson authored
Currently, we have an active reference for the request until it is retired. Though it cannot be retired before it has been executed by hardware, the request may be completed before we have finished processing the execute fence, i.e. we may continue to process that fence as we free the request. Fixes: 5590af3e ("drm/i915: Drive request submission through fence callbacks") Fixes: 23902e49 ("drm/i915: Split request submit/execute phase into two") Signed-off-by:
-
Chris Wilson authored
Before we return the request back to the kmem_cache after a failed i915_gem_request_alloc(), we should assert that it has not been added to any global state tracking. Signed-off-by:
-
Chris Wilson authored
While we will check that the request is completed prior to being retired, by placing an assert that the request is complete at the entrypoint of the function we can more clearly document the function's preconditions. Signed-off-by:
-
Chris Wilson authored
Show the last submitted seqno to the engine, not the overall next seqno, as this is more pertinent information when inspecting the pageflip and whether the CS or display engine stalled. Signed-off-by:
-
Joonas Lahtinen authored
Rename i915_gem_timeline member 'next_seqno' into 'seqno' as the variable is pre-increment. We've already had two bugs due to the confusing name, second is fixed as follow-up patch. Cc: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by:
-
- 24 Nov, 2016 6 commits
-
-
Libin Yang authored
Some monitors will have noise or even no sound after applying the patch 6014ac12. In patch 6014ac12, it will reset the cts value to 0 for HDMI. However, we need to disable Enable CTS or M Prog bit. This is the initial setting after HW reset. Fixes: 6014ac12 ("drm/i915/audio: set proper N/M in modeset") Signed-off-by:
-
Chris Wilson authored
As i915.enable_cmd_parser is an unsafe option, make it read-only at runtime. Now that it is constant, we can use the value determined during initialisation as to whether we need the cmdparser at execbuffer time. v2: Remove the inline for its single user, it is clear enough (and shorter) without! Signed-off-by:
-
Chris Wilson authored
The i915_next_seqno read value is to be the next seqno used by the kernel. However, in the conversion to atomics ops for gt.next_seqno, in commit 28176ef4 ("drm/i915: Reserve space in the global seqno during request allocation"), this was changed from a post-increment to a pre-increment. This increment was missed from the value reported by debugfs, so in effect it was reporting the current seqno (last assigned), not the next seqno. Fixes: 28176ef4 ("drm/i915: Reserve space in the global seqno during request allocation") Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=81209Signed-off-by:
-
Matthew Auld authored
No sense in keeping the cmd_descriptor and cmd_table structs in i915_drv.h, now that they are no longer referenced externally. Cc: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by:
-
Matthew Auld authored
Doing cmd_header >> 29 to extract our 3-bit client value where we know cmd_header is a u32 shouldn't then also require the use of a mask. So remove the redundant operation and get rid of INSTR_CLIENT_MASK now that there are no longer any users. Cc: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by:
-
Chris Wilson authored
i915_hws_info() has not been kept upto date (missing new engines) and so I consider it to be unused. HWS is included in the error state, which would be an avenue to retrieving it if required in future (possibly via i915_engine_info). As it is currently oopsing with an rpm testcase, just remove it. Fixes: 3b3f1650 ("drm/i915: Allocate intel_engine_cs structure only for the enabled engines") Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=98838Signed-off-by:
-
- 23 Nov, 2016 6 commits
-
-
Maarten Lankhorst authored
In all cases we can now obtain the relevant crtc_state/conn_state from the relevant callbacks, which means all the ->config accesses can be removed and the code cleaned up. Changes since v1: - cstate -> crtc_state Signed-off-by:
Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/8b02a6b4-606a-e43a-b357-ad17f491525b@linux.intel.com [mlankhorst: Reinstate missing comment] Reviewed-by:
-
Ville Syrjälä authored
No need for the extra break statements and whatnot, just return the error directly. And tighten the scope of the local variables while at it. Signed-off-by:
-
Ville Syrjälä authored
A modeset on one pipe can update dev_priv->atomic_cdclk_freq without actually touching the hardware, in which case we won't force a modeset on all the pipes, and thus won't lock any of the other pipes either. That means a parallel plane update on another pipe could be looking at a stale dev_priv->atomic_cdcdlk_freq and thus fail to notice when the plane configuration is invalid, or potentially reject a valid update. To overcome this we must protect writes to atomic_cdclk_freq with all the crtc locks, and thus for reads any single crtc lock will be sufficient protection. Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Signed-off-by:
-
Ville Syrjälä authored
When we end up not recomputing the cdclk, we need to populate intel_state->cdclk with the "atomic_cdclk_freq" instead of the current cdclk_freq. When no pipes are active, the actual cdclk_freq may be lower than what the configuration of the planes and pipes would require from the point of view of the software state. This fixes bogus WARNS from skl_max_scale() which is trying to check the plane software state against the cdclk frequency. So any time it got called during DPMS off for instance, we might have tripped the warn if the current mode would have required a higher than minimum cdclk. v2: Drop the dev_cdclk stuff (Maarten) Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Mika Kahola <mika.kahola@intel.com> Cc: bruno.pagani@ens-lyon.org Cc: Daniel J Blueman <daniel.blueman@gmail.com> Cc: Paul Bolle <pebolle@tiscali.nl> Cc: Joseph Yasi <joe.yasi@gmail.com> Tested-by:
Paul Bolle <pebolle@tiscali.nl> Tested-by: Joseph Yasi <joe.yasi@gmail.com> (v1) Cc: stable@vger.kernel.org Fixes: 1a617b77 ("drm/i915: Keep track of the cdclk as if all crtc's were active.") Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=98214Signed-off-by:
-
Bob Paauwe authored
For BXT, there is only one bit that enables/disables dual-link mode and not different bits depending on which pipe is being used. Signed-off-by:
Bob Paauwe <bob.j.paauwe@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1479767046-3964-1-git-send-email-bob.j.paauwe@intel.comSigned-off-by:
-
Ville Syrjälä authored
Let's try not to abuse plane->plane for sprites on VLV/CHV and instead use plane->id. Since out watermark structures aren't entirely plane type agnostic (for now) and start indexing sprites from 0 we'll add a small helper to convert between the two bases. Signed-off-by:
Paulo Zanoni <paulo.r.zanoni@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1479830524-7882-7-git-send-email-ville.syrjala@linux.intel.com
-
