== Series Details ==
Series: series starting with [1/4] drm/i915/fbc: Nuke bogus single pipe fbc1
restriction
URL : https://patchwork.freedesktop.org/series/69366/
State : success
== Summary ==
CI Bug Log - changes from CI_DRM_7323_full -> Patchwork_15241_full
FIXUP to "drm/i915/dsi: add support for DSC".
Signed-off-by: Jani Nikula
---
drivers/gpu/drm/i915/display/icl_dsi.c | 7 ++-
1 file changed, 2 insertions(+), 5 deletions(-)
diff --git a/drivers/gpu/drm/i915/display/icl_dsi.c
b/drivers/gpu/drm/i915/display/icl_dsi.c
index 823dbd9d4d06..52bb
On 08/11/2019 14:22, Chris Wilson wrote:
Within this set of fence execution tests, we never once try to modeset;
being KMS master is not required.
Signed-off-by: Chris Wilson
---
tests/i915/gem_exec_fence.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tests/i915/gem_e
On 08/11/2019 14:22, Chris Wilson wrote:
The pair of *-hang-all will generate sufficient hangs for the kernel to
ban the client. This is unfortunate as it means all further tests are
skipped. Ask nicely not to be banned.
Signed-off-by: Chris Wilson
---
tests/i915/gem_exec_fence.c | 7 +++
On 08/11/2019 14:22, Chris Wilson wrote:
The kernel is now enforcing that clients are not allowed to block higher
priority contexts from accessing the GPU; one is no longer allowed to
sleep for a second hogging the GPU. Reduce the sleep down to 50ms, short
enough not to anger the preempt-off che
Quoting Tvrtko Ursulin (2019-11-13 10:54:58)
>
> On 08/11/2019 14:22, Chris Wilson wrote:
> > The kernel is now enforcing that clients are not allowed to block higher
> > priority contexts from accessing the GPU; one is no longer allowed to
> > sleep for a second hogging the GPU. Reduce the sleep
As we want to be able to run inside atomic context for retiring the
i915_active, and we are no longer allowed to abuse mutex_trylock, split
the tree management portion of i915_active.mutex into an irq-safe
spinlock.
References: a0855d24fc22d ("locking/mutex: Complain upon mutex API misuse in
IRQ
== Series Details ==
Series: series starting with [1/4] drm/i915/gt: Set unused mocs entry to follow
PTE on tgl as on all others
URL : https://patchwork.freedesktop.org/series/69382/
State : success
== Summary ==
CI Bug Log - changes from CI_DRM_7323_full -> Patchwork_15244_full
=
On 2019-10-31 at 17:14:22 -0700, José Roberto de Souza wrote:
> A recent change in BSpec allow us to change EXTLINE while transcoder
> is enabled so this allow us to change it even when doing the first
> fastset after taking over previous hardware state set by BIOS.
> BIOS don't enable PSR, so if s
== Series Details ==
Series: series starting with [v3,1/2] Revert "drm/i915/ehl: Update MOCS table
for EHL"
URL : https://patchwork.freedesktop.org/series/69383/
State : failure
== Summary ==
CI Bug Log - changes from CI_DRM_7323_full -> Patchwork_15245_full
==
On Wed, Nov 13, 2019 at 05:08:21PM +0530, Anshuamn Gupta wrote:
> On 2019-10-31 at 17:14:22 -0700, José Roberto de Souza wrote:
> > A recent change in BSpec allow us to change EXTLINE while transcoder
> > is enabled so this allow us to change it even when doing the first
> > fastset after taking ov
On Tue, Nov 12, 2019 at 04:21:43PM +, Patchwork wrote:
> == Series Details ==
>
> Series: drm/i915: Fix detection for a CMP-V PCH
> URL : https://patchwork.freedesktop.org/series/69345/
> State : success
>
> == Summary ==
Thanks for the review, pushed to -dinq.
>
> CI Bug Log - changes f
Looks good to me, there is a minor comment see below.
On 2019-10-31 at 17:14:20 -0700, José Roberto de Souza wrote:
> Both activate functions and the dc3co disable function were doing the
> same thing, so better move to a function and share.
> Also while at it adding a WARN_ON to catch invalid valu
I915_CONTEXT_PARAM_RINGSIZE specifies how large to create the command
ringbuffer for logical ring contects. This directly affects the number
of batches userspace can submit before blocking waiting for space.
Signed-off-by: Chris Wilson
---
tests/Makefile.sources| 3 +
tests/i915/gem_ct
Several tests depend upon the implicit engine->mmio_base but have no
means of determining the physical layout. Since the kernel has started
providing this information, start putting it to use.
Signed-off-by: Chris Wilson
---
lib/i915/gem_engine_topology.c | 84 ++
We [will] expose various per-engine scheduling controls. One of which,
'timeslice_duration_ms', defines the scheduling quantum. If a context
exhausts its timeslice, it will be preempted in favour of running one of
its compatriots.
Signed-off-by: Chris Wilson
---
tests/Makefile.sources
Check that if two contexts (one high priority, one low) share the same
buffer that has taken a page fault that we do not create an implicit
dependency between the two contexts for servicing that page fault and
binding the vma.
Signed-off-by: Chris Wilson
---
tests/i915/gem_exec_schedule.c | 166
We [will] expose various per-engine scheduling controls. One of which,
'preempt_timeout_ms', defines how we wait for a preemption request to be
honoured by the currently executing context. If it fails to relieve the
GPU within the required timeout, the engine is reset and the miscreant
forcibly evi
One of the hardest priority inversion tasks to both handle and to
simulate in testing is inversion due to resource contention. The
challenge is that a high priority context should never be blocked by a
low priority context, even if both are starving for resources --
ideally, at least for some RT OS
Some of the non-privileged registers are at the same offset on each
engine. We can improve our coverage for unknown HW layout by using the
reported engine->mmio_base for relative offsets.
Signed-off-by: Chris Wilson
---
tests/i915/gem_ctx_isolation.c | 164 -
1 fi
We [will] expose various per-engine scheduling controls. One of which,
'heartbeat_duration_ms', defines how often we send a heartbeat down the
engine to check upon the health of the engine. If a heartbeat does not
complete within the interval (or two), the engine is declared hung.
Signed-off-by: C
pi-ringfull uses 2 contexts that share a buffer. The intent was that the
contexts were independent, but it was the effect of the global lock held
by the low priority client that prevented the high priority client from
executing. I began to add a second variant where there was a shared
resource whic
On 2019-11-13 02:09, Lucas De Marchi wrote:
On Tue, Nov 12, 2019 at 02:47:57PM -0800, Matt Roper wrote:
The bspec was just updated with a minor correction to entry 61 (it
shouldn't have had the SCF bit set).
v2:
- Add a MOCS_ENTRY_UNUSED() and use it to declare the
explicitly-reserved MOCS e
== Series Details ==
Series: drm/i915: Fix a bug calling sleep function in atomic context
URL : https://patchwork.freedesktop.org/series/69385/
State : success
== Summary ==
CI Bug Log - changes from CI_DRM_7324_full -> Patchwork_15246_full
Check that even if userspace tries to sneak around the CPU caches of its
zeroed pages, it sees nothing but zeroes.
Suggested-by: Joonas Lahtinen
Signed-off-by: Chris Wilson
Cc: Joonas Lahtinen
Cc: Matthew Auld
---
tests/i915/gem_create.c | 45 +++--
1 file
Check that even if userspace tries to sneak around the CPU caches of its
zeroed pages, it sees nothing but zeroes.
Suggested-by: Joonas Lahtinen
Signed-off-by: Chris Wilson
Cc: Joonas Lahtinen
Cc: Matthew Auld
---
tests/i915/gem_create.c | 44 +++--
1 file
Check that even if userspace tries to sneak around the CPU caches of its
zeroed pages, it sees nothing but zeroes.
Suggested-by: Joonas Lahtinen
Signed-off-by: Chris Wilson
Cc: Joonas Lahtinen
Cc: Matthew Auld
---
tests/i915/gem_create.c | 48 +++--
1 file
Hi Dave, Daniel,
Here's a PR for this week's drm-misc-fixes.
Maxime
drm-misc-fixes-2019-11-13:
- One fix to the dotclock dividers range for sun4i
The following changes since commit 105401b659b7eb9cb42d6b5b75d5c049ad4b3dca:
drm/shmem: Add docbook comments for drm_gem_shmem_object madvise field
Still the saga of the hsw live_blt incoherency continues. While it did
seem that the invalidate before the breadcrumb had improved the mtbf,
nevertheless live_blt still failed. Mika's next idea was to pull the
invalidate-stall into the breadcrumb write itself.
References: 860afa086841 ("drm/i915/g
== Series Details ==
Series: drm/i915: Split i915_active.mutex into an irq-safe spinlock for the
rbtree
URL : https://patchwork.freedesktop.org/series/69399/
State : warning
== Summary ==
$ dim checkpatch origin/drm-tip
68bd26fc5933 drm/i915: Split i915_active.mutex into an irq-safe spinlock
I'm observing incoherence metric values, changing from run to run.
It appears the patches introducing noa wait & reconfiguration from
command stream switched places in the series multiple times during the
review. This lead to the dependency of one onto the order to go
missing...
Signed-off-by: Li
From: Imre Deak
Add a test to exercise the kernel's mechanism to detection of RC6
context corruptions, take the necessary action in response (disable
RC6 and runtime PM) and recover when possible (after system
suspend/resume).
v2:
- Skip test on non-existing engines.
- Fix for old kernels where
From: "Kuoppala, Mika"
If we can't do secure execbuf, there is no point in trying.
Signed-off-by: Kuoppala, Mika
---
tests/i915/gem_exec_params.c | 16
tests/i915/gem_mocs_settings.c | 14 ++
tests/perf_pmu.c | 11 +++
3 files changed, 41 in
From: "Kuoppala, Mika"
If cmd parser is mandatory, batch can't be modified post execbuf.
Some tests rely on modifying batch post execbuf. Give those
tests a method to query if those modifications ever reach
the actual engine command stream.
v2: pull in the test changes, doh
v3: class based query
From: Jon Bloomfield
gen9+ introduces a cmdparser for the BLT engine which copies the
incoming BB to a kmd owned buffer for submission (to prevent changes
being made after the bb has been safely scanned). This breaks the
spin functionality because it relies on changing the submitted spin
buffers
From: Mika Kuoppala
For testing blitter engine command parser on gen9.
v2: bad jump offset
v3: rebase
v4: improve bb start and subcase it
v5: fix presumed offsets (Jon)
Signed-off-by: Mika Kuoppala
---
tests/Makefile.sources| 3 +
tests/i915/gem_blt_parse.c| 997
To simplify emitting the recursive batch, make batch
always the first object on the execbuf list.
v2: set handles early, poll_ptr indecency (Chris)
v3: allow dep with poll
v4: fix gem_exec_schedule
v5: rebase
v6: rebase
v6: gem_ctx_shared
v7: conditional close of poll handle
Cc: Chris Wilson
Sig
From: Imre Deak
Add a helper to query the supported and currently selected
suspend-to-mem modes.
v2:
- Fix for old kernels where the mem_sleep sysfs file didn't yet exist.
Signed-off-by: Imre Deak
---
lib/igt_aux.c | 81 +++
lib/igt_aux.h | 24 +
Quoting Mika Kuoppala (2019-11-13 15:49:07)
> To simplify emitting the recursive batch, make batch
> always the first object on the execbuf list.
Requires v4.13. Useful to leave as a note. Fwiw, looks like 3.16 and
4.14 are still rotting away slowly.
-Chris
Chris Wilson writes:
> Still the saga of the hsw live_blt incoherency continues. While it did
> seem that the invalidate before the breadcrumb had improved the mtbf,
> nevertheless live_blt still failed. Mika's next idea was to pull the
> invalidate-stall into the breadcrumb write itself.
>
> Ref
== Series Details ==
Series: drm/i915: Split i915_active.mutex into an irq-safe spinlock for the
rbtree
URL : https://patchwork.freedesktop.org/series/69399/
State : failure
== Summary ==
CI Bug Log - changes from CI_DRM_7331 -> Patchwork_15248
Quoting Mika Kuoppala (2019-11-13 15:49:08)
> From: Jon Bloomfield
>
> gen9+ introduces a cmdparser for the BLT engine which copies the
> incoming BB to a kmd owned buffer for submission (to prevent changes
> being made after the bb has been safely scanned). This breaks the
> spin functionality b
Chris Wilson writes:
> Be consistent in our mocs setup on Tigerlake and set the unused control
> value to follow the PTE entry as we previously have done. The unused
> values are beyond the defines of the ABI, the consistency simplifies our
> checking.
>
> Signed-off-by: Chris Wilson
Reviewed-b
== Series Details ==
Series: drm/i915/gt: Invalidate as we write the gen7 breadcrumb
URL : https://patchwork.freedesktop.org/series/69408/
State : warning
== Summary ==
$ dim checkpatch origin/drm-tip
f30aa89cb9aa drm/i915/gt: Invalidate as we write the gen7 breadcrumb
-:11: ERROR:GIT_COMMIT_I
Quoting Mika Kuoppala (2019-11-13 15:49:11)
> From: Mika Kuoppala
>
> For testing blitter engine command parser on gen9.
>
> v2: bad jump offset
> v3: rebase
> v4: improve bb start and subcase it
> v5: fix presumed offsets (Jon)
>
> Signed-off-by: Mika Kuoppala
> ---
> tests/Makefile.sources
Quoting Mika Kuoppala (2019-11-13 15:49:13)
> From: Imre Deak
>
> Add a test to exercise the kernel's mechanism to detection of RC6
> context corruptions, take the necessary action in response (disable
> RC6 and runtime PM) and recover when possible (after system
> suspend/resume).
>
> v2:
> - S
Quoting Mika Kuoppala (2019-11-13 15:59:53)
> Chris Wilson writes:
>
> > Still the saga of the hsw live_blt incoherency continues. While it did
> > seem that the invalidate before the breadcrumb had improved the mtbf,
> > nevertheless live_blt still failed. Mika's next idea was to pull the
> > in
On Tue, Nov 12, 2019 at 06:19:35PM -0800, Lucas De Marchi wrote:
> Tiger Lake supports HDMI on port A. For other platforms we ignore what the VBT
> says regarding HDMI to workaround broken VBTs, see commit
> 2ba7d7e04371 ("drm/i915/bios: ignore HDMI on port A"). Make this apply
> gen12+ so they inh
== Series Details ==
Series: drm/i915/gt: Invalidate as we write the gen7 breadcrumb
URL : https://patchwork.freedesktop.org/series/69408/
State : success
== Summary ==
CI Bug Log - changes from CI_DRM_7331 -> Patchwork_15249
Summary
--
Quoting Lionel Landwerlin (2019-11-13 15:46:39)
> I'm observing incoherence metric values, changing from run to run.
>
> It appears the patches introducing noa wait & reconfiguration from
> command stream switched places in the series multiple times during the
> review. This lead to the dependency
Chris Wilson writes:
> Quoting Mika Kuoppala (2019-11-13 15:49:11)
>> From: Mika Kuoppala
>>
>> For testing blitter engine command parser on gen9.
>>
>> v2: bad jump offset
>> v3: rebase
>> v4: improve bb start and subcase it
>> v5: fix presumed offsets (Jon)
>>
>> Signed-off-by: Mika Kuoppal
On Wed, Nov 13, 2019 at 11:49:53AM +, Patchwork wrote:
> == Series Details ==
>
> Series: series starting with [v3,1/2] Revert "drm/i915/ehl: Update MOCS table
> for EHL"
> URL : https://patchwork.freedesktop.org/series/69383/
> State : failure
>
> == Summary ==
>
> CI Bug Log - changes f
The TGL MOCS table was corrected in the bspec and the kernel. Since
this test hardcodes its own copy of the MOCS table, we need to make
corresponding fixes here.
References: 046091758b50 ("Revert "drm/i915/ehl: Update MOCS table for EHL"")
References: bfb0e8e63d86 ("drm/i915/tgl: MOCS table updat
== Series Details ==
Series: drm/i915/perf: don't forget noa wait after oa config
URL : https://patchwork.freedesktop.org/series/69409/
State : success
== Summary ==
CI Bug Log - changes from CI_DRM_7331 -> Patchwork_15250
Summary
---
Chris Wilson writes:
> Quoting Mika Kuoppala (2019-11-13 15:49:08)
>> From: Jon Bloomfield
>>
>> gen9+ introduces a cmdparser for the BLT engine which copies the
>> incoming BB to a kmd owned buffer for submission (to prevent changes
>> being made after the bb has been safely scanned). This bre
Do not leak our internal kernel address for random userspace to abuse.
Daniel added the support to fbdev to filter out the physical addresses
being exposed by fbdev, put that to use to protect ourselves.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=112256
Fixes: 5f889b9a61dd ("drm/i915:
== Series Details ==
Series: drm/i915/fbdev: Hide smem_start from userspace
URL : https://patchwork.freedesktop.org/series/69415/
State : warning
== Summary ==
$ dim checkpatch origin/drm-tip
0392ce3a5721 drm/i915/fbdev: Hide smem_start from userspace
-:12: ERROR:GIT_COMMIT_ID: Please use git
fbdev uses the physical address of our framebuffer for its fb_mmap()
routine. While we need to adapt this address for the new io BAR, we have
to fix v5.4 first! The simplest fix is to restore the smem back to v5.4
and we will then probably have to implement our fbops->fb_mmap() call
back to handle
On 13/11/2019 18:35, Chris Wilson wrote:
Quoting Lionel Landwerlin (2019-11-13 15:46:39)
I'm observing incoherence metric values, changing from run to run.
It appears the patches introducing noa wait & reconfiguration from
command stream switched places in the series multiple times during the
r
Quoting Lionel Landwerlin (2019-11-13 18:07:59)
> On 13/11/2019 18:35, Chris Wilson wrote:
> > Quoting Lionel Landwerlin (2019-11-13 15:46:39)
> >> I'm observing incoherence metric values, changing from run to run.
> >>
> >> It appears the patches introducing noa wait & reconfiguration from
> >> co
Quoting Chris Wilson (2019-11-13 18:10:22)
> Quoting Lionel Landwerlin (2019-11-13 18:07:59)
> > On 13/11/2019 18:35, Chris Wilson wrote:
> > > Quoting Lionel Landwerlin (2019-11-13 15:46:39)
> > >> I'm observing incoherence metric values, changing from run to run.
> > >>
> > >> It appears the patc
== Series Details ==
Series: drm/i915/fbdev: Hide smem_start from userspace
URL : https://patchwork.freedesktop.org/series/69415/
State : success
== Summary ==
CI Bug Log - changes from CI_DRM_7332 -> Patchwork_15251
Summary
---
**SU
From: Mika Kuoppala
To simplify emitting the recursive batch, make batch
always the first object on the execbuf list.
This will require kernel v4.13 or greater.
v2: set handles early, poll_ptr indecency (Chris)
v3: allow dep with poll
v4: fix gem_exec_schedule
v5: rebase
v6: rebase
v6: gem_ctx_
From: Jon Bloomfield
gen9+ introduces a cmdparser for the BLT engine which copies the
incoming BB to a kmd owned buffer for submission (to prevent changes
being made after the bb has been safely scanned). This breaks the
spin functionality because it relies on changing the submitted spin
buffers
From: "Kuoppala, Mika"
If we can't do secure execbuf, there is no point in trying.
Signed-off-by: Kuoppala, Mika
---
tests/i915/gem_exec_params.c | 22 --
tests/i915/gem_mocs_settings.c | 14 ++
tests/perf_pmu.c | 14 ++
3 files chang
Quoting Chris Wilson (2019-11-13 18:28:06)
> From: "Kuoppala, Mika"
>
> If we can't do secure execbuf, there is no point in trying.
>
> Signed-off-by: Kuoppala, Mika
Reviewed-by: Chris Wilson
There are a few more I915_EXEC_SECURE users, as MI_STORE_DWORD requires
it on gen4/5; but they are of
On 13/11/2019 20:11, Chris Wilson wrote:
Quoting Chris Wilson (2019-11-13 18:10:22)
Quoting Lionel Landwerlin (2019-11-13 18:07:59)
On 13/11/2019 18:35, Chris Wilson wrote:
Quoting Lionel Landwerlin (2019-11-13 15:46:39)
I'm observing incoherence metric values, changing from run to run.
It a
Quoting Chris Wilson (2019-11-13 18:28:07)
> From: Mika Kuoppala
>
> To simplify emitting the recursive batch, make batch
> always the first object on the execbuf list.
>
> This will require kernel v4.13 or greater.
>
> v2: set handles early, poll_ptr indecency (Chris)
> v3: allow dep with poll
Quoting Chris Wilson (2019-11-13 18:28:08)
> From: Jon Bloomfield
>
> gen9+ introduces a cmdparser for the BLT engine which copies the
> incoming BB to a kmd owned buffer for submission (to prevent changes
> being made after the bb has been safely scanned). This breaks the
> spin functionality be
Quoting Chris Wilson (2019-11-13 18:06:33)
> fbdev uses the physical address of our framebuffer for its fb_mmap()
> routine. While we need to adapt this address for the new io BAR, we have
> to fix v5.4 first! The simplest fix is to restore the smem back to v5.4
> and we will then probably have to
== Series Details ==
Series: drm/i915/fbdev: Restore physical addresses for fb_mmap()
URL : https://patchwork.freedesktop.org/series/69416/
State : failure
== Summary ==
CI Bug Log - changes from CI_DRM_7332 -> Patchwork_15252
Summary
-
As we want to be able to run inside atomic context for retiring the
i915_active, and we are no longer allowed to abuse mutex_trylock, split
the tree management portion of i915_active.mutex into an irq-safe
spinlock.
References: a0855d24fc22d ("locking/mutex: Complain upon mutex API misuse in
IRQ
On Thu, Nov 07, 2019 at 01:45:58PM -0800, Jose Souza wrote:
Adding pipe D support to DSI transcoder.
Not adding it for EDP transcoder code paths as only TGL has 4 pipes
and it do not have a EDP transcoder.
Cc: Lucas De Marchi
Signed-off-by: José Roberto de Souza
Reviewed-by: Lucas De Marchi
On Thu, Nov 07, 2019 at 11:18:37PM +, Jose Souza wrote:
On Thu, 2019-11-07 at 15:10 -0800, Lucas De Marchi wrote:
On Thu, Nov 07, 2019 at 10:56:09PM +, Jose Souza wrote:
> On Thu, 2019-11-07 at 14:44 -0800, Lucas De Marchi wrote:
> > On Thu, Nov 07, 2019 at 01:45:59PM -0800, Jose Souza w
On Tue, Nov 05, 2019 at 05:45:00PM -0800, Jose Souza wrote:
PSR2 HW only support a limited number of bits per pixel, if mode has
more than supported PSR2 should not be enabled.
BSpec: 50422
BSpec: 7713
matches both specs
Reviewed-by: Lucas De Marchi
Lucas De Marchi
Cc: Gwan-gyeong Mun
S
On Tue, Nov 12, 2019 at 09:16:57AM -0800, Matt Roper wrote:
On Tue, Nov 05, 2019 at 05:45:00PM -0800, José Roberto de Souza wrote:
PSR2 HW only support a limited number of bits per pixel, if mode has
more than supported PSR2 should not be enabled.
BSpec: 50422
BSpec: 7713
Cc: Gwan-gyeong Mun
S
Disable VS Unit Clockgating.
BSpec: 52857
Cc: Mika Kuoppala
Signed-off-by: Radhakrishna Sripada
---
drivers/gpu/drm/i915/i915_reg.h | 1 +
drivers/gpu/drm/i915/intel_pm.c | 4
2 files changed, 5 insertions(+)
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
i
Extend disabling SAMPLER_STATE prefetch workaround to gen12.
BSpec: 52890
Cc: Mika Kuoppala
Signed-off-by: Radhakrishna Sripada
---
drivers/gpu/drm/i915/gt/intel_workarounds.c | 12 +++-
1 file changed, 7 insertions(+), 5 deletions(-)
diff --git a/drivers/gpu/drm/i915/gt/intel_workarou
== Series Details ==
Series: drm/i915: Split i915_active.mutex into an irq-safe spinlock for the
rbtree (rev2)
URL : https://patchwork.freedesktop.org/series/69399/
State : warning
== Summary ==
$ dim checkpatch origin/drm-tip
70ca71d22f04 drm/i915: Split i915_active.mutex into an irq-safe sp
I broke fb_mmap() proving that we need a test in CI!
References: https://bugs.freedesktop.org/show_bug.cgi?id=112256
Signed-off-by: Chris Wilson
---
tests/Makefile.sources| 1 +
tests/fbdev.c | 69 +++
tests/intel-ci/fast-feedback.
Quoting Chris Wilson (2019-11-13 18:06:33)
> fbdev uses the physical address of our framebuffer for its fb_mmap()
> routine. While we need to adapt this address for the new io BAR, we have
> to fix v5.4 first! The simplest fix is to restore the smem back to v5.4
> and we will then probably have to
On 11/12/2019 4:28 PM, Bruce Chang wrote:
> There are quite a few reports regarding "BUG: sleeping function called from
> invalid context at mm/page_alloc.c"
>
> Basically after the io_mapping_map_atomic_wc/kmap_atomic, it enters atomic
> context, but compress_page cannot be called in atomic con
Quoting Chris Wilson (2019-11-13 18:42:06)
> Quoting Chris Wilson (2019-11-13 18:28:08)
> > From: Jon Bloomfield
> >
> > gen9+ introduces a cmdparser for the BLT engine which copies the
> > incoming BB to a kmd owned buffer for submission (to prevent changes
> > being made after the bb has been s
From: Jon Bloomfield
gen9+ introduces a cmdparser for the BLT engine which copies the
incoming BB to a kmd owned buffer for submission (to prevent changes
being made after the bb has been safely scanned). This breaks the
spin functionality because it relies on changing the submitted spin
buffers
On Wed, 13 Nov 2019 at 14:05, Chris Wilson wrote:
>
> Check that even if userspace tries to sneak around the CPU caches of its
> zeroed pages, it sees nothing but zeroes.
>
> Suggested-by: Joonas Lahtinen
> Signed-off-by: Chris Wilson
> Cc: Joonas Lahtinen
> Cc: Matthew Auld
Reviewed-by: Matth
There are quite a few reports regarding "BUG: sleeping function called from
invalid context at mm/page_alloc.c"
Basically after the io_mapping_map_atomic_wc/kmap_atomic, it enters atomic
context, but compress_page cannot be called in atomic context as it will
call pool_alloc with GFP_KERNEL flag w
== Series Details ==
Series: drm/i915: Split i915_active.mutex into an irq-safe spinlock for the
rbtree (rev2)
URL : https://patchwork.freedesktop.org/series/69399/
State : failure
== Summary ==
CI Bug Log - changes from CI_DRM_7332 -> Patchwork_15253
=
Quoting Bruce Chang (2019-11-13 19:52:44)
> There are quite a few reports regarding "BUG: sleeping function called from
> invalid context at mm/page_alloc.c"
>
> Basically after the io_mapping_map_atomic_wc/kmap_atomic, it enters atomic
> context, but compress_page cannot be called in atomic conte
> -Original Message-
> From: Chris Wilson
> Sent: Wednesday, November 13, 2019 11:47 AM
> To: intel-gfx@lists.freedesktop.org
> Cc: Bloomfield, Jon ; Lahtinen, Joonas
> ; Vivi, Rodrigo ;
> Kuoppala, Mika ; Mika Kuoppala
>
> Subject: [PATCH i-g-t] igt: Use COND_BBEND for busy spinning on g
From: Jon Bloomfield
gen9+ introduces a cmdparser for the BLT engine which copies the
incoming BB to a kmd owned buffer for submission (to prevent changes
being made after the bb has been safely scanned). This breaks the
spin functionality because it relies on changing the submitted spin
buffers
From: Jon Bloomfield
gen9+ introduces a cmdparser for the BLT engine which copies the
incoming BB to a kmd owned buffer for submission (to prevent changes
being made after the bb has been safely scanned). This breaks the
spin functionality because it relies on changing the submitted spin
buffers
On Sat, 2019-11-09 at 15:10 +, Chris Wilson wrote:
> igt_aux.h already provides the optimal igt_fls(), so use that in
> preference to open coding the brute force version.
>
> Reported-by: Stuart Summers
> Signed-off-by: Chris Wilson
> Cc: Stuart Summers
Thanks for the look here Chris :)
-
Hi Dave & Daniel,
Just one msm patch this week. Looks like -misc is going to be perfect when merge
window rolls around :-)
drm-misc-next-fixes-2019-11-13:
- Fix memory leak in gpu debugfs node's release (Johan)
Cc: Johan Hovold
Cheers, Sean
The following changes since commit 3ca3a9eab7085b3
On Fri, Nov 08, 2019 at 03:45:00PM +0200, Stanislav Lisovskiy wrote:
> Also implemented algorithm for choosing DBuf slice configuration
> based on active pipes, pipe ratio as stated in BSpec 12716.
>
> Now pipe allocation still stays proportional to pipe width as before,
> however within allowed D
On Sat, 2019-11-09 at 10:47 +, Patchwork wrote:
> == Series Details ==
>
> Series: series starting with [1/3] drm/i915/display: Fix
> TRANS_DDI_MST_TRANSPORT_SELECT definition
> URL : https://patchwork.freedesktop.org/series/69160/
> State : success
>
> == Summary ==
>
> CI Bug Log - chang
On Wed, Nov 13, 2019 at 08:51:03AM -0800, Matt Roper wrote:
The TGL MOCS table was corrected in the bspec and the kernel. Since
this test hardcodes its own copy of the MOCS table, we need to make
corresponding fixes here.
References: 046091758b50 ("Revert "drm/i915/ehl: Update MOCS table for EH
Quoting Chris Wilson (2019-11-13 20:20:31)
> From: Jon Bloomfield
>
> gen9+ introduces a cmdparser for the BLT engine which copies the
> incoming BB to a kmd owned buffer for submission (to prevent changes
> being made after the bb has been safely scanned). This breaks the
> spin functionality be
From: Jon Bloomfield
gen9+ introduces a cmdparser for the BLT engine which copies the
incoming BB to a kmd owned buffer for submission (to prevent changes
being made after the bb has been safely scanned). This breaks the
spin functionality because it relies on changing the submitted spin
buffers
Quoting Radhakrishna Sripada (2019-11-13 19:18:39)
> Extend disabling SAMPLER_STATE prefetch workaround to gen12.
>
> BSpec: 52890
> Cc: Mika Kuoppala
> Signed-off-by: Radhakrishna Sripada
> ---
> drivers/gpu/drm/i915/gt/intel_workarounds.c | 12 +++-
> 1 file changed, 7 insertions(+),
In case of tiled displays, if we hotplug just one connector,
fbcon currently just selects the preferred mode and if it is
tiled mode then that becomes a problem if rest of the tiles are
not present.
So in the fbdev driver on hotplug when we probe the client modeset,
if we dont find all the connecto
1 - 100 of 131 matches
Mail list logo