If we fail to reset the GPU, we declare the machine wedged. However, the
GPU may well still be running in the background with an in-flight
request. So despite our efforts in cleaning up the request queue and
faking the breadcrumb in the HWSP, the GPU may eventually write the
in-flght seqno there breaking all of our assumptions and throwing the
driver into a deep turmoil, wedging beyond wedged.
To avoid this we ideally want to reset the GPU. Since that has already
failed, make sure the rings have the stop bit set instead. This is part
of the normal GPU reset sequence, but that is actually disabled by
igt/gem_eio to force the wedged state. If we assume the worst, we must
poke at the bit again before we give up.
v2: Move the intel_gpu_reset() from set-wedged in the reset error path
into i915_gem_set_wedged() itself. Even if the reset fails (e.g. if it is
disabled by gem_eio), it still tries to make sure the engines are
stopped. For i915_gem_set_wedged() callers from outside of i915_reset(),
this should make sure the GPU is disabled while the driver is marked as
being wedged.
Testcase: igt/gem_eio
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Michał Winiarski <michal.winiarski@intel.com>
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Cc: Michel Thierry <michel.thierry@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180315151015.22741-1-chris@chris-wilson.co.uk
GuC WOPCM registers are write-once registers. Current driver code accesses
these registers without checking the accessibility to these registers which
will lead to unpredictable driver behaviors if these registers were touch
by other components (such as faulty BIOS code).
This patch moves the GuC WOPCM registers updating code into intel_wopcm.c
and adds check before and after the update to GuC WOPCM registers so that
we can make sure the driver is in a known state after writing to these
write-once registers.
v6:
- Made sure module reloading won't bug the kernel while doing
locking status checking
v7:
- Fixed patch format issues
v8:
- Fixed coding style issue on register lock bit macro definition (Sagar)
v9:
- Avoided to use redundant !! to cast uint to bool (Chris)
- Return error code instead of GEM_BUG_ON for locked with invalid register
values case (Sagar)
- Updated guc_wopcm_hw_init to use guc_wopcm as first parameter (Michal)
- Added code to set and validate the HuC_LOADING_AGENT_GUC bit in GuC
WOPCM offset register based on the presence of HuC firmware (Michal)
- Use bit fields instead of macros for GuC WOPCM flags (Michal)
v10:
- Refined variable names, removed redundant comments (Joonas)
- Introduced lockable_reg to handle the write once register write and
propagate the write error to caller (Joonas)
- Used lockable_reg abstraction to avoid locking bit check on generic
i915_reg_t (Michal)
- Added log message for error paths (Michal)
- Removed hw_updated flag and only relies on real hardware status
v11:
- Replaced lockable_reg with simplified function (Michal)
- Used new macros for locking bits of WOPCM size/offset registers instead
of using BIT(0) directly (Michal)
- use intel_wopcm_init_hw() called from intel_gem_init_hw() to do GuC
WOPCM register setup instead of calling from intel_uc_init_hw() (Michal)
v12:
- Updated function kernel-doc to align with code changes (Michal)
- Updated code to use wopcm pointer directly (Michal)
v13:
- Updated the ordering of s-o-b/cc/r-b tags (Sagar)
BSpec: 10875, 10833
Signed-off-by: Jackie Li <yaodong.li@intel.com>
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Michal Wajdeczko <michal.wajdeczko@intel.com> (v11)
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> (v12)
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1520987574-19351-5-git-send-email-yaodong.li@intel.com
Hardware may have specific restrictions on GuC WOPCM offset and size. On
Gen9, the value of the GuC WOPCM size register needs to be larger than the
value of GuC WOPCM offset register + a Gen9 specific offset (144KB) for
reserved GuC WOPCM. Fail to enforce such a restriction on GuC WOPCM size
will lead to GuC firmware execution failures. On the other hand, with
current static GuC WOPCM offset and size values (512KB for both offset and
size), the GuC WOPCM size verification will fail on Gen9 even if it can be
fixed by lowering the GuC WOPCM offset by calculating its value based on
HuC firmware size (which is likely less than 200KB on Gen9), so that we can
have a GuC WOPCM size value which is large enough to pass the GuC WOPCM
size check.
This patch updates the reserved GuC WOPCM size for RC6 context on Gen9 to
24KB to strictly align with the Gen9 GuC WOPCM layout. It also adds support
to verify the GuC WOPCM size aganist the Gen9 hardware restrictions. To
meet all above requirements, let's provide dynamic partitioning of the
WOPCM that will be based on platform specific HuC/GuC firmware sizes.
v2:
- Removed intel_wopcm_init (Ville/Sagar/Joonas)
- Renamed and Moved the intel_wopcm_partition into intel_guc (Sagar)
- Removed unnecessary function calls (Joonas)
- Init GuC WOPCM partition as soon as firmware fetching is completed
v3:
- Fixed indentation issues (Chris)
- Removed layering violation code (Chris/Michal)
- Created separat files for GuC wopcm code (Michal)
- Used inline function to avoid code duplication (Michal)
v4:
- Preset the GuC WOPCM top during early GuC init (Chris)
- Fail intel_uc_init_hw() as soon as GuC WOPCM partitioning failed
v5:
- Moved GuC DMA WOPCM register updating code into intel_wopcm.c
- Took care of the locking status before writing to GuC DMA
Write-Once registers. (Joonas)
v6:
- Made sure the GuC WOPCM size to be multiple of 4K (4K aligned)
v8:
- Updated comments and fixed naming issues (Sagar/Joonas)
- Updated commit message to include more description about the hardware
restriction on GuC WOPCM size (Sagar)
v9:
- Minor changes variable names and code comments (Sagar)
- Added detailed GuC WOPCM layout drawing (Sagar/Michal)
- Refined macro definitions to be reader friendly (Michal)
- Removed redundent check to valid flag (Michal)
- Unified first parameter for exported GuC WOPCM functions (Michal)
- Refined the name and parameter list of hardware restriction checking
functions (Michal)
v10:
- Used shorter function name for internal functions (Joonas)
- Moved init-ealry function into c file (Joonas)
- Consolidated and removed redundant size checks (Joonas/Michal)
- Removed unnecessary unlikely() from code which is only called once
during boot (Joonas)
- More fixes to kernel-doc format and content (Michal)
- Avoided the use of PAGE_MASK for 4K pages (Michal)
- Added error log messages to error paths (Michal)
v11:
- Replaced intel_guc_wopcm with more generic intel_wopcm and attached
intel_wopcm to drm_i915_private instead intel_guc (Michal)
- dynamic calculation of GuC non-wopcm memory start (a.k.a WOPCM Top
offset from GuC WOPCM base) (Michal)
- Moved WOPCM marco definitions into .c source file (Michal)
- Exported WOPCM layout diagram as kernel-doc (Michal)
v12:
- Updated naming, function kernel-doc to align with new changes (Michal)
v13:
- Updated the ordering of s-o-b/cc/r-b tags (Sagar)
- Corrected one tense error in comment (Sagar)
- Corrected typos and removed spurious comments (Joonas)
Bspec: 12690
Signed-off-by: Jackie Li <yaodong.li@intel.com>
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Cc: Sagar Arun Kamble <sagar.a.kamble@intel.com>
Cc: Sujaritha Sundaresan <sujaritha.sundaresan@intel.com>
Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: John Spotswood <john.a.spotswood@intel.com>
Cc: Oscar Mateo <oscar.mateo@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> (v8)
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> (v9)
Reviewed-by: Michal Wajdeczko <michal.wajdeczko@intel.com> (v11)
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> (v12)
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1520987574-19351-2-git-send-email-yaodong.li@intel.com
tasklet_kill() will spin waiting for the current tasklet to be executed.
However, if tasklet_disable() has been called, then the tasklet is never
executed but permanently put back onto the runlist until
tasklet_enable() is called. Ergo, we cannot use tasklet_kill() inside a
disable/enable pair. This is the case when we call set-wedge from inside
i915_reset(), and another request was submitted to us concurrent to the
reset.
Fixes: 963ddd63c3 ("drm/i915: Suspend submission tasklets around wedging")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180307134226.25492-6-chris@chris-wilson.co.uk
Currently, we only allow ourselves to prune the fences so long as
all the waits completed (i.e. all the fences we checked were signaled),
and that the reservation snapshot did not change across the wait.
However, if we only waited for a subset of the reservation object, i.e.
just waiting for the last writer to complete as opposed to all readers
as well, then we would erroneously conclude we could prune the fences as
indeed although all of our waits were successful, they did not represent
the totality of the reservation object.
v2: We only need to check the shared fences due to construction (i.e.
all of the shared fences will be later than the exclusive fence, if
any).
Fixes: e54ca97747 ("drm/i915: Remove completed fences after a wait")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180307171303.29466-1-chris@chris-wilson.co.uk
During igt, we frequently call into the driver to reset both HW and
driver state (idling the device, waiting for it to become idle and
freeing off old objects) to ensure that we start each test/subtest/pass
from known state. This process incurs an RCU barrier or two to ensure
that any such pending frees are indeed flushed before we return.
However, unconditionally waiting on the RCU barrier adds needless delay
to many callers, which adds up to several seconds when repeated thousands
of times. We can skip the rcu_barrier() if by tracking how many outstanding
frees we have, we know there are none.
The same path is used along suspend, where we may be able to save the
unconditional RCU barrier.
To put it into perspective with a completely meaningless
microbenchmark, igt/gem_sync/idle is improved from 50ms to 30us on bdw.
v2: Remove the extra synchronize_rcu() inside i915_drop_caches_set()
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180219220631.25001-1-chris@chris-wilson.co.uk
i915 is the only driver using those fields in the drm_gem_object
structure, so they only waste memory for all other drivers.
Move the fields into drm_i915_gem_object instead and patch the i915 code
with the following sed commands:
sed -i "s/obj->base.read_domains/obj->read_domains/g" drivers/gpu/drm/i915/*.c drivers/gpu/drm/i915/*/*.c
sed -i "s/obj->base.write_domain/obj->write_domain/g" drivers/gpu/drm/i915/*.c drivers/gpu/drm/i915/*/*.c
Change is only compile tested.
v2: move fields around as suggested by Chris.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20180216124338.9087-1-christian.koenig@amd.com
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Since commit 5896a5c8c9 (drm/i915: Always stop the rings before a
missing GPU reset) we attempt to stop the engines during gem_sanitize
even if reset=0 and nothing bad happened on the gpu.
The specs says that the STOP_RINGS bit needs to be cleared to resume
normal operation, but for some reason the value of the bit seems to be
changing without us writing to it (maybe rc6 entry/exit?), so normal
operation resumes correctly. However, it still feels incorrect to stop
the engines if there hasn't been any issue so skip the whole reset
call in gem_sanitize if i915.reset=0
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180207212440.13438-1-daniele.ceraolospurio@intel.com
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
On blb and pnv, we are seeing sporadic
i915 0000:00:02.0: Resetting chip after gpu hang
[drm:intel_gpu_reset [i915]] rcs0: timed out on STOP_RING
[drm:i915_reset [i915]] *ERROR* Failed hw init on reset -5
which notably lack the actual root cause of the error. Ostensibly it
should be the init_ring_common() that failed, but it's error paths are
covered by DRM_ERROR.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180207111545.17078-1-chris@chris-wilson.co.uk
Since commit 7b6da818d8 ("drm/i915: Restore the kernel context after a
GPU reset on an idle engine") we submit a request following the engine
reset. The intent is that we don't submit a request if the engine is
busy (as it will restart active by itself) but we only checked to see if
there were remaining requests in flight on the hardware and skipped
checking to see if there were any ready requests that would be
immediately submitted on restart (the same time as our new request would
be). Having convinced the engine to appear idle in the previous patch,
we can use intel_engine_is_idle() as a better test to only submit a new
request if there are no pending requests.
As it happens, this is tripping up igt/drv_selftest/live_hangcheck in CI
as we overfill the kernel_context ringbuffer trigger an infinite
recursion from within the reset.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104786
References: 7b6da818d8 ("drm/i915: Restore the kernel context after a GPU reset on an idle engine")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Michel Thierry <michel.thierry@intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180205152431.12163-4-chris@chris-wilson.co.uk
Previously, we relied on only running the hangcheck while somebody was
waiting on the GPU, in order to minimise the amount of time hangcheck
had to run. (If nobody was watching the GPU, nobody would notice if the
GPU wasn't responding -- eventually somebody would care and so kick
hangcheck into action.) However, this falls apart from around commit
4680816be3 ("drm/i915: Wait first for submission, before waiting for
request completion"), as not all waiters declare themselves to hangcheck
and so we could switch off hangcheck and miss GPU hangs even when
waiting under the struct_mutex.
If we enable hangcheck from the first request submission, and let it run
until the GPU is idle again, we forgo all the complexity involved with
only enabling around waiters. We just have to remember to be careful that
we do not declare a GPU hang when idly waiting for the next request to
be come ready, as we will run hangcheck continuously even when the
engines are stalled waiting for external events. This should be true
already as we should only be tracking requests submitted to hardware for
execution as an indicator that the engine is busy.
Fixes: 4680816be3 ("drm/i915: Wait first for submission, before waiting for request completion"
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104840
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180129144104.3921-1-chris@chris-wilson.co.uk
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Watching a light workload on Baytrail (running glxgears and a 1080p
decode), instead of the system remaining at low frequency, the glxgears
would regularly trigger waitboosting after which it would have to spend
a few seconds throttling back down. In this case, the waitboosting is
counter productive as the minimal wait for glxgears doesn't prevent it
from functioning correctly and delivering frames on time. In this case,
glxgears happens to almost always be waiting on the current request,
which we already expect to complete quickly (see i915_spin_request) and
so avoiding the waitboost on the active request and spinning instead
provides the best latency without overcommitting to upclocking.
However, if the system falls behind we still force the waitboost.
Similarly, we will also trigger upclocking if we detect the system is
not delivering frames on time - again using a mechanism that tries to
detect a miss and not preemptively upclock.
v2: Also skip boosting for after missed vblank if the desired request is
already active.
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>
Cc: Radoslaw Szwichtenberg <radoslaw.szwichtenberg@intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180118131609.16574-1-chris@chris-wilson.co.uk
Since Michal introduced new user controllable errors other than -EIO
during i915_gem_init(), we need to actually unwind on the error path as
we have to abort the module load (and we expect to do so cleanly!).
As we now teardown key state and then mark the driver as wedged (on
EIO), we have to be careful to not allow ourselves to resume and
unwedge, thus attempting to use the uninitialised driver.
v2: Try not to free driver state for the suppressed EIO
v3: Use load-fault-injection to test both error/recovery paths.
References: 8620eb1dbb ("drm/i915/uc: Don't use -EIO to report missing firmware")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Sagar Arun Kamble <sagar.a.kamble@intel.com>
Reviewed-by: Michał Winiarski <michal.winiarski@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171213134347.4608-1-chris@chris-wilson.co.uk
It seems that the DMC likes to transition between the DC states a lot when
there are no connected displays (no active power domains) during command
submission.
This activity on DC states has a negative impact on the performance of the
chip with huge latencies observed in the interrupt handlers and elsewhere.
Simple tests like igt/gem_latency -n 0 are slowed down by a factor of
eight.
Work around it by introducing a new power domain named,
POWER_DOMAIN_GT_IRQ, associtated with the "DC off" power well, which is
held for the duration of command submission activity.
CNL has the same problem which will be addressed as a follow-up. Doing
that requires a fix for a DC6 context corruption problem in the CNL DMC
firmware which is yet to be released.
v2:
* Add commit text as comment in i915_gem_mark_busy. (Chris Wilson)
* Protect macro body with braces. (Jani Nikula)
v3:
* Add dedicated power domain for clarity. (Chris, Imre)
* Commit message and comment text updates.
* Apply to all big-core GEN9 parts apart for Skylake which is pending DMC
firmware release.
v4:
* Power domain should be inner to device runtime pm. (Chris)
* Simplify NEEDS_CSR_GT_PERF_WA macro. (Chris)
* Handle async DMC loading by moving the GT_IRQ power domain logic into
intel_runtime_pm. (Daniel, Chris)
* Include small core GEN9 as well. (Imre)
v5
* Special handling for async DMC load is not needed since on failure the
power domain reference is kept permanently taken. (Imre)
v6:
* Drop the NEEDS_CSR_GT_PERF_WA macro since all firmwares have now been
deployed. (Imre, Chris)
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100572
Testcase: igt/gem_exec_nop/headless
Cc: Imre Deak <imre.deak@intel.com>
Acked-by: Chris Wilson <chris@chris-wilson.co.uk> (v2)
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> (v5)
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
[Imre: Add note about applying the WA on CNL as a follow-up]
Signed-off-by: Imre Deak <imre.deak@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171205132854.26380-1-tvrtko.ursulin@linux.intel.com
In quite a few places, we have a list iteration over the vma on an
object that only want to inspect GGTT vma. By construction, these are
placed at the start of the list, so we have copied that knowledge into
many callsites. Pull that knowledge back to i915_vma.h and provide a
for_each_ggtt_vma() to tidy up the code.
v2: Add a backreference from vma_create() to remind ourselves why we put
ggtt vma at the head of the obj->vma_list (and ppgtt vma at the tail).
v3: Fixup s/vma/V/
Suggested-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20171207211407.31549-1-chris@chris-wilson.co.uk
