On Thu, 2017-12-07 at 14:40 +0100, Maarten Lankhorst wrote:
> Instead of assuming each subtest cleans up after itself, assume it
> fails and doesn't. Now that igt_kms can clean up stale events, we
> can just force each subtest to only clean up its framebuffers,
> which isn't harmful if it failed.
>
From: Tvrtko Ursulin
A couple of small optimizations which altogether bring around 30%
improvement in my testing.
1. Do less string processing on tracepoints names and push more of the
check into the if-ladder.
2. Pull out common db key and ctx processing and cache common values in
local
On Thu, 2017-12-07 at 14:40 +0100, Maarten Lankhorst wrote:
> Using the fancy new DRM_CAP_CRTC_IN_VBLANK_EVENT cap I can finally
> make this test the work I originally intended to.
>
> For the !modeset case that means performing a pageflip on both
> crtc's,
> then requeueing as soon as the event i
On 2017.12.22 09:49:36 +0200, Jani Nikula wrote:
>
> Yes. We want a minimum of two people looking at each patch. It's pretty
> much irrelevant if the committer/maintainer is the author or not. 2*sob
> or sob+rb or sob+ack, or more for more complicated things.
>
> It's unfortunately common that th
Quoting Dave Airlie (2017-12-22 03:05:16)
> On 21 December 2017 at 19:28, Chris Wilson wrote:
> > The vk cts test:
> > dEQP-VK.api.external.semaphore.opaque_fd.export_multiple_times_temporary
> >
> > triggers a lot of
> > VFS: Close: file count is 0
> >
> > Dave pointed out that clearing the synco
On 12/21/2017 6:29 PM, Lionel Landwerlin wrote:
Some more findings I made while playing with this series & GPUTop.
Turns out the 2ms drift per second is due to timecounter. Adding the
delta this way :
https://github.com/djdeath/linux/commit/7b002cb360483e331053aec0f98433a5bd5c5c3f#diff-9b74b
== Series Details ==
Series: scripts/trace.pl: Optimize event parsing and processing (rev5)
URL : https://patchwork.freedesktop.org/series/35569/
State : success
== Summary ==
IGT patchset tested on top of latest successful build
beb26d89ff5c5621c1e6b6ac2a45439507af86b7 meson: Install .testlis
On 12/21/2017 6:37 AM, Sujaritha Sundaresan wrote:
Including GEM_BUG_ON for GuC reset function in
intel_uncore.
Can be reframed - "Instead of returning -EINVAL, GEM_BUG_ON when GuC
reset is invoked for
platforms not supporting as we don't expect to invoke it"
Subject can be - "GEM_BUG_ON on i
Hi guys,
Anyone has the bandwidth to review the below tiny patch, or suggest some
else as potential reviewer?
Truth be told I only tested it lightly, but it passed the perf_pmu unit
tests. Goal is to bring engine busyness PMU reporting accuracy to the
same level as execlists.
Regards,
Tv
Hi Dave, just flushing my queue of a couple of late arrivals before
xmas. Pull and send on at your leisure.
Happy holidays,
Jani.
The following changes since commit a4ffdc2b6726958c07d535318400124e3a3bc19b:
drm/i915: Protect DDI port to DPLL map from theoretical race. (2017-12-19
14:59:53 +0
On 22/12/17 09:30, Sagar Arun Kamble wrote:
On 12/21/2017 6:29 PM, Lionel Landwerlin wrote:
Some more findings I made while playing with this series & GPUTop.
Turns out the 2ms drift per second is due to timecounter. Adding the
delta this way :
https://github.com/djdeath/linux/commit/7b002
== Series Details ==
Series: scripts/trace.pl: Optimize event parsing and processing (rev5)
URL : https://patchwork.freedesktop.org/series/35569/
State : success
== Summary ==
Test gem_tiled_swapping:
Subgroup non-threaded:
incomplete -> PASS (shard-hsw) fdo#10421
On 11/29/2017 6:03 PM, Tvrtko Ursulin wrote:
From: Tvrtko Ursulin
Wire up the engine busy stats accounting to the GuC submission backend.
Since there is not
no
context out interrupt we need to place the accounting
callbacks per-request in order to correctly pair with user interrupts.
v2
Sleep for a known duration. In particular, CI once saw a measurement for
busyness greater than the intended batch_duration!
v2: Go back to starting pmu sampling outside of spinner; the GPU should
be idle.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104241
Signed-off-by: Chris Wilson
C
Create all the spinners before starting the sampler and then measure how
long we sleep.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104160
Signed-off-by: Chris Wilson
Cc: Tvrtko Ursulin
---
tests/perf_pmu.c | 45 +
1 file changed, 25 insert
Don't rely on the timer being precise when we can sleep for a known
duration.
Signed-off-by: Chris Wilson
Cc: Tvrtko Ursulin
---
tests/perf_pmu.c | 10 +-
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/tests/perf_pmu.c b/tests/perf_pmu.c
index 96091b853..b3b289656 100644
Use the normal batch_duration_ns and display the sampled frequency:
Frequency: min=100, max=750, boost=750 MHz
Min frequency: requested 100.0, actual 100.0
Max frequency: requested 755.6, actual 755.6
v2: Remove the early spin_batch_end and assert the measured frequencies
Quoting changbin...@intel.com (2017-12-22 06:31:49)
> From: Changbin Du
>
> Our QA reported a problem caused by movntdqa instructions. Currently,
> the KVM hypervisor doesn't support VEX-prefix instructions emulation.
> If users passthrough a GPU to guest with vfio option 'x-no-mmap=on',
> then a
Don't rely on the timer being precise when we can sleep for a known
duration.
Signed-off-by: Chris Wilson
Cc: Tvrtko Ursulin
---
tests/perf_pmu.c | 38 ++
1 file changed, 26 insertions(+), 12 deletions(-)
diff --git a/tests/perf_pmu.c b/tests/perf_pmu.c
inde
We want to expose the parameters for controlling how long it takes for
us to notice and park the GPU after a stream of requests in order to try
and tune the optimal power-efficiency vs latency of a mostly idle system.
Signed-off-by: Chris Wilson
Cc: Tvrtko Ursulin
---
drivers/gpu/drm/i915/Kcon
An interesting discussion regarding "hybrid interrupt polling" for NVMe
came to the conclusion that the ideal busyspin before sleeping was half
of the expected request latency (and better if it was already halfway
through that request). This suggested that we too should look again at
our tradeoff b
On 22/12/2017 10:50, Sagar Arun Kamble wrote:
On 11/29/2017 6:03 PM, Tvrtko Ursulin wrote:
From: Tvrtko Ursulin
Wire up the engine busy stats accounting to the GuC submission backend.
Since there is not
no
Yeah. :)
context out interrupt we need to place the accounting
callbacks per-re
On 22/12/2017 11:33, Chris Wilson wrote:
We want to expose the parameters for controlling how long it takes for
us to notice and park the GPU after a stream of requests in order to try
and tune the optimal power-efficiency vs latency of a mostly idle system.
Extra space.
Signed-off-by: Ch
Quoting Tvrtko Ursulin (2017-12-22 11:52:28)
>
> On 22/12/2017 11:33, Chris Wilson wrote:
> > +static inline long retire_delay(void)
> > +{
> > + const long delay = msecs_to_jiffies(CONFIG_DRM_I915_GEM_RETIRE_DELAY);
> > +
> > + if (CONFIG_DRM_I915_GEM_RETIRE_DELAY >= 1000)
> > +
== Series Details ==
Series: series starting with [1/2] drm/i915: Expose the busyspin durations for
i915_wait_request
URL : https://patchwork.freedesktop.org/series/35716/
State : failure
== Summary ==
Series 35716v1 series starting with [1/2] drm/i915: Expose the busyspin
durations for i915
The ultimate goal of this series is to promote our definition
of fixed16.16 type to include/linux. However, before doing so
we need to do some preparations/cleanups locally.
Note that some helper functions may require additional renames.
Any suggestions are welcomed.
Suggested-by: Chris Wilson
C
Rename uint_fixed_16_16_t to fixed16_16_t to match header name.
Also switch into kernel integer types.
Suggested-by: Chris Wilson
Signed-off-by: Michal Wajdeczko
Cc: Chris Wilson
Cc: Rodrigo Vivi
Cc: Joonas Lahtinen
---
drivers/gpu/drm/i915/fixed16_16.h | 91 ++---
Zero and One are additional commonly used values that
can have its own definitions.
Signed-off-by: Michal Wajdeczko
Cc: Chris Wilson
Cc: Rodrigo Vivi
Cc: Joonas Lahtinen
---
drivers/gpu/drm/i915/fixed16_16.h | 15 ---
drivers/gpu/drm/i915/intel_pm.c | 22 +++---
If someone provides too large number for fixed16 type
we will WARN but we will not correctly clamp values
and that may lead to fully wrong calculations.
Signed-off-by: Michal Wajdeczko
Cc: Chris Wilson
Cc: Rodrigo Vivi
Cc: Joonas Lahtinen
---
drivers/gpu/drm/i915/fixed16_16.h | 13 +--
Tidy up fixed16_16 code before promoting it to the include/linux.
Signed-off-by: Michal Wajdeczko
Cc: Chris Wilson
Cc: Rodrigo Vivi
Cc: Joonas Lahtinen
---
drivers/gpu/drm/i915/fixed16_16.h | 96 ++-
1 file changed, 43 insertions(+), 53 deletions(-)
diff -
Our uint_fixed_16_16_t type definition and related helper functions
deserve dedicated header. While here cleanup indent.
v2: avoid too generic filename (Chris)
Signed-off-by: Michal Wajdeczko
Cc: Chris Wilson
Cc: Rodrigo Vivi
Cc: Joonas Lahtinen
Acked-by: Rodrigo Vivi
---
drivers/gpu/drm/i9
Quoting Patchwork (2017-12-21 21:36:17)
> == Series Details ==
>
> Series: Misc i915_drv.h cleanups (rev2)
> URL : https://patchwork.freedesktop.org/series/35637/
> State : failure
>
> == Summary ==
>
> Test gem_tiled_swapping:
> Subgroup non-threaded:
> incomplete -> P
== Series Details ==
Series: series starting with [1/5] igt/perf_pmu: Tighten busy measurement
URL : https://patchwork.freedesktop.org/series/35715/
State : success
== Summary ==
IGT patchset tested on top of latest successful build
865a47ca2b93b208ba016f3c4fc00831ec7bfec6 overlay: parse trace
Quoting Michal Wajdeczko (2017-12-22 12:25:54)
> Zero and One are additional commonly used values that
> can have its own definitions.
>
> Signed-off-by: Michal Wajdeczko
> Cc: Chris Wilson
> Cc: Rodrigo Vivi
> Cc: Joonas Lahtinen
> ---
> drivers/gpu/drm/i915/fixed16_16.h | 15 ---
Quoting Michal Wajdeczko (2017-12-22 12:25:55)
> If someone provides too large number for fixed16 type
> we will WARN but we will not correctly clamp values
> and that may lead to fully wrong calculations.
>
> Signed-off-by: Michal Wajdeczko
> Cc: Chris Wilson
> Cc: Rodrigo Vivi
> Cc: Joonas La
Quoting Michal Wajdeczko (2017-12-22 12:25:56)
> Tidy up fixed16_16 code before promoting it to the include/linux.
>
> Signed-off-by: Michal Wajdeczko
> Cc: Chris Wilson
> Cc: Rodrigo Vivi
> Cc: Joonas Lahtinen
> ---
> drivers/gpu/drm/i915/fixed16_16.h | 96
> ++--
== Series Details ==
Series: drm/i915: promote fixed16_16
URL : https://patchwork.freedesktop.org/series/35717/
State : success
== Summary ==
Series 35717v1 drm/i915: promote fixed16_16
https://patchwork.freedesktop.org/api/1.0/series/35717/revisions/1/mbox/
Test debugfs_test:
Subgrou
On Thu, 2017-12-07 at 14:40 +0100, Maarten Lankhorst wrote:
> Instead of assuming each subtest cleans up after itself, assume it
> fails and doesn't. Now that igt_kms can clean up stale events, we
> can just force each subtest to only clean up its framebuffers,
> which isn't harmful if it failed.
>
From: Tvrtko Ursulin
Rather than calibrate and emit nop batches, use a manually signalled chain
of spinners to generate the desired interrupts.
v2: Two flavours of interrupt generation. (Chris Wilson)
Signed-off-by: Tvrtko Ursulin
Cc: Chris Wilson
---
tests/perf_pmu.c | 141 +
From: Tvrtko Ursulin
Support creating spin batches which return an output fence using new
__igt_spin_batch_new_fence / igt_spin_batch_new_fence API.
This will be used fromthe perf_pmu@interrupts test to ensure user
interrupt generation from a batch with controlled duration.
v2: Support out fenc
On 22/12/2017 11:03, Chris Wilson wrote:
Sleep for a known duration. In particular, CI once saw a measurement for
busyness greater than the intended batch_duration!
v2: Go back to starting pmu sampling outside of spinner; the GPU should
be idle.
Bugzilla: https://bugs.freedesktop.org/show_bug.
On 22/12/2017 11:03, Chris Wilson wrote:
Create all the spinners before starting the sampler and then measure how
long we sleep.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104160
Signed-off-by: Chris Wilson
Cc: Tvrtko Ursulin
---
tests/perf_pmu.c | 45 +--
On 22/12/2017 11:03, Chris Wilson wrote:
Don't rely on the timer being precise when we can sleep for a known
duration.
Signed-off-by: Chris Wilson
Cc: Tvrtko Ursulin
---
tests/perf_pmu.c | 10 +-
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/tests/perf_pmu.c b/tests
On 22/12/2017 11:03, Chris Wilson wrote:
Don't rely on the timer being precise when we can sleep for a known
duration.
Signed-off-by: Chris Wilson
Cc: Tvrtko Ursulin
---
tests/perf_pmu.c | 38 ++
1 file changed, 26 insertions(+), 12 deletions(-)
diff --
On 22/12/2017 11:03, Chris Wilson wrote:
Use the normal batch_duration_ns and display the sampled frequency:
Frequency: min=100, max=750, boost=750 MHz
Min frequency: requested 100.0, actual 100.0
Max frequency: requested 755.6, actual 755.6
v2: Remove the early spin_ba
We already emit a GEM_TRACE for when we start preemption, but we lack
one to show when the preemption is completed and we return to the regular
queue. This is to continue the investigation into the mysertious
<0>[ 197.854177] -0 1..s1 197837017us :
execlists_submission_tasklet: rcs0 cs-i
== Series Details ==
Series: series starting with [1/2] lib/dummyload: Support returning output fence
URL : https://patchwork.freedesktop.org/series/35718/
State : failure
== Summary ==
IGT patchset tested on top of latest successful build
865a47ca2b93b208ba016f3c4fc00831ec7bfec6 overlay: pars
Quoting Tvrtko Ursulin (2017-12-22 13:13:47)
> From: Tvrtko Ursulin
>
> Support creating spin batches which return an output fence using new
> __igt_spin_batch_new_fence / igt_spin_batch_new_fence API.
>
> This will be used fromthe perf_pmu@interrupts test to ensure user
> interrupt generation f
On 22/12/2017 13:27, Chris Wilson wrote:
We already emit a GEM_TRACE for when we start preemption, but we lack
one to show when the preemption is completed and we return to the regular
queue. This is to continue the investigation into the mysertious
Un-permute mysterious :)
<0>[ 197.854177]
== Series Details ==
Series: series starting with [1/5] igt/perf_pmu: Tighten busy measurement
URL : https://patchwork.freedesktop.org/series/35715/
State : success
== Summary ==
Test kms_frontbuffer_tracking:
Subgroup fbc-1p-offscren-pri-shrfb-draw-blt:
fail -> P
Quoting Tvrtko Ursulin (2017-12-22 13:13:48)
> From: Tvrtko Ursulin
>
> Rather than calibrate and emit nop batches, use a manually signalled chain
> of spinners to generate the desired interrupts.
>
> v2: Two flavours of interrupt generation. (Chris Wilson)
>
> Signed-off-by: Tvrtko Ursulin
>
== Series Details ==
Series: drm/i915/execlists: Show preemption progress in GEM_TRACE
URL : https://patchwork.freedesktop.org/series/35719/
State : success
== Summary ==
Series 35719v1 drm/i915/execlists: Show preemption progress in GEM_TRACE
https://patchwork.freedesktop.org/api/1.0/series/3
We should only attempt to remove requests from the execution queue that
are on the execution queue. These are the requests that have been
assigned a global_seqno, so we can assert that we only attempt to remove
requests with a nonzero global_seqno. Afterwards we assert that we
remove them in order,
== Series Details ==
Series: drm/i915: promote fixed16_16
URL : https://patchwork.freedesktop.org/series/35717/
State : failure
== Summary ==
Test perf:
Subgroup blocking:
fail -> PASS (shard-hsw) fdo#102252
Test gem_tiled_swapping:
Subgroup non-thre
Looking at a CI failure with an ominous line of
[ 362.550715] hangcheck current seqno ff6b, last ff8c, hangcheck
ff6b [6016 ms], inflight 118
with no apparent cause for the seqno to be negative, left me wondering
if someone had scribbled over the HWSP. So include the HWSP in the
engin
From: Johnson Lin
Some panels support limited range output (16-235) compared
to full range RGB values (0-255). Also userspace can control
the RGB range using "Broadcast RGB" property. Currently the
code to handle full range to limited range is broken. This
patch fixes the same by properly scaling
== Series Details ==
Series: drm/i915: Assert that the request is on the execution queue before
being removed
URL : https://patchwork.freedesktop.org/series/35723/
State : success
== Summary ==
Series 35723v1 drm/i915: Assert that the request is on the execution queue
before being removed
ht
Quoting Tvrtko Ursulin (2017-12-21 17:13:16)
> From: Tvrtko Ursulin
>
> Switch over to dynamically creating device attributes, which are in turn
> used by the perf core to expose available counters in sysfs.
>
> This way we do not expose counters which are not avaiable on the current
> platform,
== Series Details ==
Series: drm/i915: Show HWSP in intel_engine_dump()
URL : https://patchwork.freedesktop.org/series/35724/
State : success
== Summary ==
Series 35724v1 drm/i915: Show HWSP in intel_engine_dump()
https://patchwork.freedesktop.org/api/1.0/series/35724/revisions/1/mbox/
Test d
== Series Details ==
Series: drm/i915: Fix Limited Range Color Handling
URL : https://patchwork.freedesktop.org/series/35725/
State : success
== Summary ==
Series 35725v1 drm/i915: Fix Limited Range Color Handling
https://patchwork.freedesktop.org/api/1.0/series/35725/revisions/1/mbox/
Test g
== Series Details ==
Series: drm/i915/execlists: Show preemption progress in GEM_TRACE
URL : https://patchwork.freedesktop.org/series/35719/
State : failure
== Summary ==
Test gem_mmap_wc:
Subgroup set-cache-level:
skip -> PASS (shard-snb)
Test kms_flip_even
On Fri, 22 Dec 2017 13:34:58 +0100, Chris Wilson
wrote:
Quoting Michal Wajdeczko (2017-12-22 12:25:54)
Zero and One are additional commonly used values that
can have its own definitions.
Signed-off-by: Michal Wajdeczko
Cc: Chris Wilson
Cc: Rodrigo Vivi
Cc: Joonas Lahtinen
---
drivers/g
On Fri, Dec 22, 2017 at 08:34:47PM +0530, Uma Shankar wrote:
> From: Johnson Lin
>
> Some panels support limited range output (16-235) compared
> to full range RGB values (0-255). Also userspace can control
> the RGB range using "Broadcast RGB" property. Currently the
> code to handle full range
On Fri, 22 Dec 2017 13:39:11 +0100, Chris Wilson
wrote:
Quoting Michal Wajdeczko (2017-12-22 12:25:55)
If someone provides too large number for fixed16 type
we will WARN but we will not correctly clamp values
and that may lead to fully wrong calculations.
Signed-off-by: Michal Wajdeczko
Cc
Quoting Michal Wajdeczko (2017-12-22 15:51:49)
> On Fri, 22 Dec 2017 13:34:58 +0100, Chris Wilson
> wrote:
>
> > Quoting Michal Wajdeczko (2017-12-22 12:25:54)
> >> Zero and One are additional commonly used values that
> >> can have its own definitions.
> >>
> >> Signed-off-by: Michal Wajdeczko
== Series Details ==
Series: drm/i915: Assert that the request is on the execution queue before
being removed
URL : https://patchwork.freedesktop.org/series/35723/
State : failure
== Summary ==
Test kms_flip:
Subgroup vblank-vs-modeset-suspend-interruptible:
pass
On Fri, 22 Dec 2017 13:25:53 +0100, Michal Wajdeczko
wrote:
Rename uint_fixed_16_16_t to fixed16_16_t to match header name.
Also switch into kernel integer types.
Suggested-by: Chris Wilson
Signed-off-by: Michal Wajdeczko
Cc: Chris Wilson
Cc: Rodrigo Vivi
Cc: Joonas Lahtinen
---
driver
On Fri, Dec 22, 2017 at 05:54:31PM +0200, Ville Syrjälä wrote:
> On Fri, Dec 22, 2017 at 08:34:47PM +0530, Uma Shankar wrote:
> > From: Johnson Lin
> >
> > Some panels support limited range output (16-235) compared
> > to full range RGB values (0-255). Also userspace can control
> > the RGB range
== Series Details ==
Series: drm/i915: Show HWSP in intel_engine_dump()
URL : https://patchwork.freedesktop.org/series/35724/
State : success
== Summary ==
Test pm_rpm:
Subgroup system-suspend-execbuf:
skip -> PASS (shard-hsw) fdo#103375
Test kms_flip:
== Series Details ==
Series: drm/i915: Fix Limited Range Color Handling
URL : https://patchwork.freedesktop.org/series/35725/
State : warning
== Summary ==
Test kms_flip:
Subgroup vblank-vs-modeset-suspend-interruptible:
pass -> SKIP (shard-snb)
On 21/12/2017 15:48, Chris Wilson wrote:
Ensure that the terminating write into WC-memory is flushed before we
might trigger a wait for ring space.
Signed-off-by: Chris Wilson
---
tests/gem_exec_await.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/tests/gem_exec_await.c b/tests/gem_e
Quoting Michal Wajdeczko (2017-12-22 16:08:51)
> On Fri, 22 Dec 2017 13:25:53 +0100, Michal Wajdeczko
> wrote:
>
> > Rename uint_fixed_16_16_t to fixed16_16_t to match header name.
> > Also switch into kernel integer types.
> >
> > Suggested-by: Chris Wilson
> > Signed-off-by: Michal Wajdeczko
Quoting Chris Wilson (2017-12-22 17:11:00)
> Quoting Michal Wajdeczko (2017-12-22 16:08:51)
> > On Fri, 22 Dec 2017 13:25:53 +0100, Michal Wajdeczko
> > wrote:
> >
> > > Rename uint_fixed_16_16_t to fixed16_16_t to match header name.
> > > Also switch into kernel integer types.
> > >
> > > Sugg
On 22/12/2017 14:52, Chris Wilson wrote:
Quoting Tvrtko Ursulin (2017-12-21 17:13:16)
From: Tvrtko Ursulin
Switch over to dynamically creating device attributes, which are in turn
used by the perf core to expose available counters in sysfs.
This way we do not expose counters which are not av
On 22/12/2017 14:19, Chris Wilson wrote:
We should only attempt to remove requests from the execution queue that
are on the execution queue. These are the requests that have been
assigned a global_seqno, so we can assert that we only attempt to remove
requests with a nonzero global_seqno. Afterw
Looking at a CI failure with an ominous line of
[ 362.550715] hangcheck current seqno ff6b, last ff8c, hangcheck
ff6b [6016 ms], inflight 118
with no apparent cause for the seqno to be negative, left me wondering
if someone had scribbled over the HWSP. So include the HWSP in the
engin
On 22/12/2017 17:28, Chris Wilson wrote:
Looking at a CI failure with an ominous line of
[ 362.550715] hangcheck current seqno ff6b, last ff8c, hangcheck
ff6b [6016 ms], inflight 118
with no apparent cause for the seqno to be negative, left me wondering
if someone had scribbled ove
== Series Details ==
Series: drm/i915: Show HWSP in intel_engine_dump() (rev2)
URL : https://patchwork.freedesktop.org/series/35724/
State : success
== Summary ==
Series 35724v2 drm/i915: Show HWSP in intel_engine_dump()
https://patchwork.freedesktop.org/api/1.0/series/35724/revisions/2/mbox/
On Fri, Dec 22, 2017 at 08:51:41AM +, Zhenyu Wang wrote:
> On 2017.12.22 09:49:36 +0200, Jani Nikula wrote:
> >
> > Yes. We want a minimum of two people looking at each patch. It's pretty
> > much irrelevant if the committer/maintainer is the author or not. 2*sob
> > or sob+rb or sob+ack, or m
Looking at a CI failure with an ominous line of
[ 362.550715] hangcheck current seqno ff6b, last ff8c, hangcheck
ff6b [6016 ms], inflight 118
with no apparent cause for the seqno to be negative, left me wondering
if someone had scribbled over the HWSP. So include the HWSP in the
engin
Looking at a CI failure with an ominous line of
[ 362.550715] hangcheck current seqno ff6b, last ff8c, hangcheck
ff6b [6016 ms], inflight 118
with no apparent cause for the seqno to be negative, left me wondering
if someone had scribbled over the HWSP. So include the HWSP in the
engin
From: Ville Syrjälä
The LG 4k TV I have doesn't deassert HPD when I turn the TV off, but
when I turn it back on it will pulse the HPD line. By that time it has
forgotten everything we told it about scrambling and the clock ratio.
Hence if we want to get a picture out if it again we have to tell i
From: Ville Syrjälä
Doing link reatrining from the short pulse handler is problematic since
that might introduce deadlocks with MST sideband processing. Currently
we don't retrain MST links from this code, but we want to change that.
So better to move the entire thing to the hotplug work. We can
From: Ville Syrjälä
During hpd processing we may want to do things both before and
after the display detection. To that end split the encoder->hot_plug()
hooks.
Signed-off-by: Ville Syrjälä
---
drivers/gpu/drm/i915/intel_drv.h | 3 ++-
drivers/gpu/drm/i915/intel_hotplug.c | 6 --
drive
== Series Details ==
Series: drm/i915: Show HWSP in intel_engine_dump() (rev3)
URL : https://patchwork.freedesktop.org/series/35724/
State : success
== Summary ==
Series 35724v3 drm/i915: Show HWSP in intel_engine_dump()
https://patchwork.freedesktop.org/api/1.0/series/35724/revisions/3/mbox/
On 22/12/2017 18:25, Chris Wilson wrote:
Looking at a CI failure with an ominous line of
[ 362.550715] hangcheck current seqno ff6b, last ff8c, hangcheck
ff6b [6016 ms], inflight 118
with no apparent cause for the seqno to be negative, left me wondering
if someone had scribbled ove
Quoting Tvrtko Ursulin (2017-12-22 18:52:29)
>
> On 22/12/2017 18:25, Chris Wilson wrote:
> > Looking at a CI failure with an ominous line of
> > [ 362.550715] hangcheck current seqno ff6b, last ff8c, hangcheck
> > ff6b [6016 ms], inflight 118
> > with no apparent cause for the seqno
Quoting Chris Wilson (2017-12-22 19:00:24)
> Quoting Tvrtko Ursulin (2017-12-22 18:52:29)
> >
> > On 22/12/2017 18:25, Chris Wilson wrote:
> > > Looking at a CI failure with an ominous line of
> > > [ 362.550715] hangcheck current seqno ff6b, last ff8c, hangcheck
> > > ff6b [6016 ms]
== Series Details ==
Series: drm/i915: Show HWSP in intel_engine_dump() (rev2)
URL : https://patchwork.freedesktop.org/series/35724/
State : warning
== Summary ==
Test kms_flip:
Subgroup vblank-vs-modeset-suspend-interruptible:
skip -> INCOMPLETE (shard-hsw) fdo#1
From: Ville Syrjälä
Here's a rebased series of the remaining CCS patches. Apart from the rebase
and some commit message changes nothing has changed.
Ville Syrjälä (8):
drm/i915: Add a comment exlaining CCS hsub/vsub
drm/i915: Nuke a pointless unreachable()
drm/i915: Add the missing Y/Yf mo
From: Ville Syrjälä
The unreachable() is very much unreachable and the compiler knows
that, so there's no point in having it.
Cc: Ben Widawsky
Cc: Jason Ekstrand
Cc: Daniel Stone
Signed-off-by: Ville Syrjälä
---
drivers/gpu/drm/i915/intel_display.c | 2 --
1 file changed, 2 deletions(-)
di
From: Ville Syrjälä
Split the g4x and snb cases into separate functions to match how we deal
with all other platforms. Also sort the switch cases to match the format
lists we've declared earlier, to ease comparisons.
Cc: Ben Widawsky
Cc: Jason Ekstrand
Cc: Daniel Stone
Signed-off-by: Ville Sy
From: Ville Syrjälä
Let's document why we claim hsub==8,vsub==16 for CCS even though the
memory layout would suggest that we use 16x8 instead.
Cc: Daniel Vetter
Cc: Ben Widawsky
Cc: Jason Ekstrand
Cc: Daniel Stone
Signed-off-by: Ville Syrjälä
---
drivers/gpu/drm/i915/intel_display.c | 7 ++
From: Ville Syrjälä
Y/Yf were dropped out from the SKL+ sprite modifier list on account
of some watermark issues Daniel Stone was having. My subsequent testing
seemed to indicate that things work better now, so add the modifiers
back in.
v2: Update the commit message with a better explanation
C
From: Ville Syrjälä
The core now checks that the plane supports the fb's format+modifier
combination, so we can drop the related checks from
skl_check_ccs_aux_surface(). These checks were specific to
SKL/KBL/BXT anyway.
Cc: Ben Widawsky
Cc: Jason Ekstrand
Cc: Daniel Stone
Signed-off-by: Ville
From: Ville Syrjälä
SKL+ "sprites" no longer have 16KB max stride limit that earlier
platforms had. Bump up the limit to 32KB.
Cc: Ben Widawsky
Cc: Jason Ekstrand
Cc: Daniel Stone
Signed-off-by: Ville Syrjälä
---
drivers/gpu/drm/i915/intel_sprite.c | 3 ++-
1 file changed, 2 insertions(+),
From: Ville Syrjälä
Allow sprites to scan out compressed framebuffers.
Since different platforms have a different set of planes that
support CCS let's add a small helper to determine whether a
specific plane supports CCS or not. Currently that information
is spread around in many places, and not
From: Ville Syrjälä
Currently we only check that the plane supports the pixel format of the
fb we're about to feed to it. Extend it to check also the modifier, and
more specifically that the combination of the format and modifier is
supported.
Cc: dri-de...@lists.freedesktop.org
Cc: Ben Widawsky
== Series Details ==
Series: drm/i915: Show HWSP in intel_engine_dump() (rev4)
URL : https://patchwork.freedesktop.org/series/35724/
State : failure
== Summary ==
Applying: drm/i915: Show HWSP in intel_engine_dump()
error: Failed to merge in the changes.
Using index info to reconstruct a base
== Series Details ==
Series: series starting with [1/3] drm/i915: Split encoder->hot_plug() into pre
and post variants
URL : https://patchwork.freedesktop.org/series/35732/
State : failure
== Summary ==
Series 35732v1 series starting with [1/3] drm/i915: Split encoder->hot_plug()
into pre an
1 - 100 of 157 matches
Mail list logo
