On 18.05.2022 13:54, Rob Clark wrote:
> On Tue, May 17, 2022 at 10:42 AM Adrián Larumbe
> wrote:
> >
> > In the event of a job timeout, debug dump information will be written into
> > /sys/class/devcoredump.
> >
> > Inspired by etnaviv's similar featu
Hi Steven,
Thanks a lot for your feedback, it was quite useful.
Also I'm sorry about having taken so long to write a reply, but other things
held me
back from working on Panfrost for way too long already.
On 18.05.2022 12:03, Steven Price wrote:
>On 17/05/2022 18:42, Adrián Larumbe wrot
r for header fields to something more representative.
- Changed the dump header format to a union whose members are the different
header types.
Adrián Larumbe (1):
drm/panfrost: Add support for devcoredump
drivers/gpu/drm/panfrost/Kconfig | 1 +
drivers/gpu/drm/panfrost/Makefile
In the event of a job timeout, debug dump information will be written into
/sys/class/devcoredump.
Inspired by etnaviv's similar feature.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panfrost/Kconfig | 1 +
drivers/gpu/drm/panfrost/Makefile| 3 +-
drivers/gp
In the event of a job timeout, debug dump information will be written into
/sys/class/devcoredump.
Inspired by etnaviv's similar feature.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panfrost/Kconfig | 1 +
drivers/gpu/drm/panfrost/Makefile| 3 +-
drivers/gp
patch:
- Added JS and MMU offset macros for selecting the right register in the dump
rather
than using magic numbers.
- Added sanity check to make sure Panfrost's page size is the same as the CPU's
- Added major and minor dump header version numbers
Adrián Larumbe (1):
drm/panfrost: A
.
Adrián Larumbe (2):
drm/panfrost: Add specific register offset macros for JS and MMU AS
drm/panfrost: Add support for devcoredump
drivers/gpu/drm/panfrost/Kconfig | 1 +
drivers/gpu/drm/panfrost/Makefile| 3 +-
drivers/gpu/drm/panfrost/panfrost_dump.c | 249
Each Panfrost job has its own job slot and MMU address space set of
registers, which are selected with a job-specific index.
Turn the shift and stride used for selection of the right register set base
into a define rather than using magic numbers.
Signed-off-by: Adrián Larumbe
---
drivers/gpu
In the event of a job timeout, debug dump information will be written into
/sys/class/devcoredump.
Inspired by etnaviv's similar feature.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panfrost/Kconfig | 1 +
drivers/gpu/drm/panfrost/Makefile| 3 +-
drivers/gp
s videomode
supports YCBCR420 output format.
Fix the selection logic.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/meson/meson_dw_hdmi.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/meson/meson_dw_hdmi.c
b/drivers/gpu/drm/meson/meson_dw_hdmi.c
index 5cd2b2
.
Mesa MR under review can be found at:
https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14034
Adrián Larumbe (1):
drm/panfrost: Add support for devcoredump
drivers/gpu/drm/panfrost/Kconfig | 1 +
drivers/gpu/drm/panfrost/Makefile| 3 +-
drivers/gpu/drm/panfrost
In the event of a job timeout, debug dump information will be written into
/sys/class/devcoredump.
Inspired by etnaviv's similar feature.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panfrost/Kconfig | 1 +
drivers/gpu/drm/panfrost/Makefile| 3 +-
drivers/gp
the Mesa build.
Give it a name, and also rename pan_reg_hdr structure because it will
always be prefixed by the union name.
Bug: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7195
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panfrost/panfrost_dump.c | 20 ++--
i
eep nodes to freed memory.
All three patches have been tested on an Odroid N2+ plus SBC.
Adrián Larumbe (3):
drm/meson: reorder driver deinit sequence to fix use-after-free bug
drm/meson: explicitly remove aggregate driver at module unload time
drm/meson: remove drm bridges at aggregate driver
ge_detach is called for every single bridge linked to the encoder,
including the one whose memory had already been deallocated.
Fix by calling component_unbind_all after drm_dev_put.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/meson/meson_drv.c | 2 +-
1 file changed, 1 insertion(+), 1 de
007] ==
Fix by keeping track of which encoders were initialised in the meson_drm
structure and manually removing their bridges at aggregate driver's unbind
time.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/meson/mes
x27;remove' driver callback for meson-drm, and explicitly deleting
the
aggregate device.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/meson/meson_drv.c | 8
1 file changed, 8 insertions(+)
diff --git a/drivers/gpu/drm/meson/meson_drv.c
b/drivers/gpu/drm/meson/meson_d
Hi Steven,
On 13.09.2022 09:45, Steven Price wrote:
>On 12/09/2022 17:44, Adrián Larumbe wrote:
>> Building Mesa's Perfetto requires including the panfrost drm uAPI header in
>> C++ code, but the C++ compiler requires anonymous unions to have only
>> public non-static d
On 19.09.2022 15:03, Neil Armstrong wrote:
> On 19/09/2022 03:09, Adrián Larumbe wrote:
> > drm bridges added by meson_encoder_hdmi_init and meson_encoder_cvbs_init
> > were not manually removed at module unload time, which caused dangling
> > references to freed memory to
/mesa/mesa/-/issues/7252
Fixes: 730c2bf4ad39 ("drm/panfrost: Add support for devcoredump")
Signed-off-by: Adrián Larumbe
---
include/uapi/drm/panfrost_drm.h | 30 +++---
1 file changed, 15 insertions(+), 15 deletions(-)
diff --git a/include/uapi/drm/panfrost_drm.
From: Steven Price
The two structs internal to struct panfrost_dump_object_header were
named, but sadly that is incompatible with C++, causing an error: "an
anonymous union may only have public non-static data members".
However nothing refers to struct pan_reg_hdr and struct pan_bomap_hdr
and th
007] ==
Fix by keeping track of which encoders were initialised in the meson_drm
structure and manually removing their bridges at aggregate driver's unbind
time.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/meson/mes
Hi Steven, once again thanks for the feedback. I was off for some time and then
busy with other stuff, but I can finally work on a new revision for the patch
On 23.06.2022 13:23, Steven Price wrote:
> On 22/06/2022 15:36, Adrián Larumbe wrote:
> > In the event of a job timeout, d
patch:
- Added MMU and JS base and stride macros for range checks within
panfrost_core_dump_registers
- Fixed dereference-before-assignment BO pointer bug
- Chose the right-sized endianness conversion function for header's valid bit
Adrián Larumbe (2):
drm/panfrost: Add specific reg
In the event of a job timeout, debug dump information will be written into
/sys/class/devcoredump.
Inspired by etnaviv's similar feature.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panfrost/Kconfig | 1 +
drivers/gpu/drm/panfrost/Makefile| 3 +-
drivers/gp
Each Panfrost job has its own job slot and MMU address space set of
registers, which are selected with a job-specific index.
Turn the shift and stride used for selection of the right register set base
into a define rather than using magic numbers.
Signed-off-by: Adrián Larumbe
---
drivers/gpu
In the event of a job timeout, debug dump information will be written into
/sys/class/devcoredump.
Inspired by etnaviv's similar feature.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panfrost/Kconfig | 1 +
drivers/gpu/drm/panfrost/Makefile| 3 +-
drivers/gp
:
- Expressed MMU_AS_STRIDE as a bitshift to avoid yet another magic number and
removed misleading comment on the same line.
Adrián Larumbe (2):
drm/panfrost: Add specific register offset macros for JS and MMU AS
drm/panfrost: Add support for devcoredump
drivers/gpu/drm/panfrost/Kconfig
Each Panfrost job has its own job slot and MMU address space set of
registers, which are selected with a job-specific index.
Turn the shift and stride used for selection of the right register set base
into a define rather than using magic numbers.
Signed-off-by: Adrián Larumbe
---
drivers/gpu
ree commits to accurately
reflect the modified files' subsystem.
- Add v1's Acked-by tags from subsystem maintainer
v2:
- Add commit message to patch number 3 in the series
Adrián Larumbe (3):
drm/bridge: dw-hdmi: change YUV420 selection logic at clock setup
drm
The current output bus format selection logic is enforcing YUV420 even
when the drm mode allows for other bus formats as well.
Fix it by adding check for 420-only drm modes.
Signed-off-by: Adrián Larumbe
Acked-by: Neil Armstrong
---
drivers/gpu/drm/bridge/synopsys/dw-hdmi.c | 7 ---
1 file
The hdmi_datamap enum is no longer in use. Also reindent enable_audio's
call params.
Signed-off-by: Adrián Larumbe
Acked-by: Neil Armstrong
---
drivers/gpu/drm/bridge/synopsys/dw-hdmi.c | 22 --
1 file changed, 4 insertions(+), 18 deletions(-)
diff --git a/drivers/gp
will configure a colour space based on the bus
format that doesn't match the pixel value we had calculated as described
above.
Fix it by bringing back dw-hdmi bus format check when picking the right
pixel clock.
Signed-off-by: Adrián Larumbe
Acked-by: Neil Armstrong
---
drivers/gpu/drm/b
I fixed it by bringing back dw-hdmi bus format check in dw-meson.
The second patch makes sure YUV420 bus format is the only one being returned by
dw-hdmi's output format bridge function when that's the only drm mode allowed.
Adrián Larumbe (3):
drm/meson: dw-hdmi: change YUV420 selec
will configure a colour space based on the bus
format that doesn't match the pixel value we had calculated as described
above.
Fix it by bringing back dw-hdmi bus format check when picking the right
pixel clock.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/bridge/synopsys/dw-hdmi.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/bridge/synopsys/dw-hdmi.c | 22 --
1 file changed, 4 insertions(+), 18 deletions(-)
diff --git a/drivers/gpu/drm/bridge/synopsys/dw-hdmi.c
b/drivers/gpu/drm/bridge/synopsys/dw-hdmi.c
index 1afb8f2603a0..0accfb51509c 100644
The current output bus format selection logic is enforcing YUV420 even
when the drm mode allows for other bus formats as well.
Fix it by adding check for 420-only drm modes.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/bridge/synopsys/dw-hdmi.c | 7 ---
1 file changed, 4 insertions
ing the bridge enablement stage.
Signed-off-by: Adrián Larumbe
Fixes: 5d844091f237 ("drm/scdc-helper: Pimp SCDC debugs")
---
drivers/gpu/drm/bridge/synopsys/dw-hdmi.c | 9 +
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/drivers/gpu/drm/bridge/synopsys/dw-hdmi.c
b
The hdmi_datamap enum is no longer in use. Also reindent enable_audio's
call params.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/bridge/synopsys/dw-hdmi.c | 22 --
1 file changed, 4 insertions(+), 18 deletions(-)
diff --git a/drivers/gpu/drm/bridge/synopsys/dw-h
The current output bus format selection logic is enforcing YUV420 even
when the drm mode allows for other bus formats as well.
Fix it by adding check for 420-only drm modes.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/bridge/synopsys/dw-hdmi.c | 7 ---
1 file changed, 4 insertions
This is v2 of:
https://lore.kernel.org/dri-devel/20230528140001.1057084-1-adrian.laru...@collabora.com/
The only difference is having added an actual commit message to patch
number 3 in the series.
Adrián Larumbe (3):
drm/meson: dw-hdmi: change YUV420 selection logic at clock setup
dw-hdmi
will configure a colour space based on the bus
format that doesn't match the pixel value we had calculated as described
above.
Fix it by bringing back dw-hdmi bus format check when picking the right
pixel clock.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/bridge/synopsys/dw-hdmi.
privileged user or accessed with IGT's gputop
utility.
Adrián Larumbe (2):
drm/panfrost: Add fdinfo support to Panfrost
drm/panfrost: Add drm memory stats display through fdinfo
drivers/gpu/drm/panfrost/panfrost_device.c | 12 +++
drivers/gpu/drm/panfrost/panfrost_device.h
For drm_show_memory_stats to produce a more accurate report, provide a new
Panfrost DRM object callback that decides whether an object is resident in
memory or eligible for purging.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panfrost/panfrost_drv.c | 8 ++--
drivers/gpu/drm/panfrost
We calculate the amount of time the GPU spends on a job with ktime samples,
and then add it to the cumulative total for the open DRM file, which is
what will be eventually exposed through the 'fdinfo' DRM file descriptor.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/dr
'fdinfo' DRM file descriptor.
>>
>> Signed-off-by: Adrián Larumbe
>> ---
>> drivers/gpu/drm/panfrost/panfrost_device.c | 12
>> drivers/gpu/drm/panfrost/panfrost_device.h | 10 +++
>> drivers/gpu/drm/panfrost/panfrost_drv.c| 32 ++
These GPU registers will be used when programming the cycle counter, which
we need for providing accurate fdinfo drm-cycles values to user space.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panfrost/panfrost_regs.h | 5 +
1 file changed, 5 insertions(+)
diff --git a/drivers/gpu/drm
calculations.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panfrost/panfrost_devfreq.c | 8
drivers/gpu/drm/panfrost/panfrost_devfreq.h | 3 ++
drivers/gpu/drm/panfrost/panfrost_device.h | 13 ++
drivers/gpu/drm/panfrost/panfrost_drv.c | 45 -
drivers/gpu
ly be backed by system memory
after successive GPU page faults.
Provide a new DRM object generic function that would allow drivers to
return a more accurate RSS size for their BOs.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/drm_file.c | 5 -
include/drm/drm_gem.h | 9 +
2 fi
f RSS size for BO's.
- Implemente selection of drm-memory region size units
- Removed locking of shrinker's mutex in GEM obj status function
Adrián Larumbe (6):
drm/panfrost: Add cycle count GPU register definitions
drm/panfrost: Add fdinfo support GPU load metrics
drm/panfr
multiplier of a 10-power of 2, so 1 would give us size in Kib and 2
in Mib. If we want drm-file functions to pick the highest unit, then 0
should be passed.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/drm_file.c | 22 +-
drivers/gpu/drm/msm/msm_drv.c | 2
e's object handle database spinlock, so there's potential
for a race condition here.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panfrost/panfrost_drv.c | 9 +++--
drivers/gpu/drm/panfrost/panfrost_gem.c | 12
drivers/gpu/drm/panfrost/panfrost_gem.h | 1 +
ng
storage for it might have already been allocated by the exporting driver.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panfrost/panfrost_gem.c | 22 ++
drivers/gpu/drm/panfrost/panfrost_gem.h | 5 +
drivers/gpu/drm/panfrost/panfrost_mmu.c | 16 +++-
3 fil
On 29.08.2024 15:57, Liviu Dudau wrote:
> Hi Adrián,
>
> On Sat, Aug 17, 2024 at 11:10:17AM +0200, Boris Brezillon wrote:
> > On Fri, 16 Aug 2024 19:52:49 +0100
> > Adrián Larumbe wrote:
> >
> > > In the off-chance that waiting for the firmware to signal its
("drm/panthor: Add the MMU/VM logical block")
Cc: sta...@vger.kernel.org
Signed-off-by: Adrián Larumbe
Acked-by: Liviu Dudau
---
drivers/gpu/drm/panthor/panthor_fw.c | 8 +++-
drivers/gpu/drm/panthor/panthor_mmu.c | 21 ++---
drivers/gpu/drm/panthor/panthor_mmu.h | 1
t definition.
- Precompute size of group's syncobj BO to avoid code duplication.
- Some minor nits.
Adrián Larumbe (4):
drm/panthor: introduce job cycle and timestamp accounting
drm/panthor: add DRM fdinfo support
drm/panthor: enable fdinfo for memory stats
drm/panthor: add sysfs knob fo
or a
given set of enablement mask, and these are passed as the number of credits
when initialising a DRM scheduler job.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panthor/panthor_device.h | 22 ++
drivers/gpu/drm/panthor/panthor_sched.c | 327 ---
2 files ch
commit, and supporting manual toggle of it will be
the matter of a later commit.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panthor/panthor_devfreq.c | 18 -
drivers/gpu/drm/panthor/panthor_device.h | 14 +++
drivers/gpu/drm/panthor/panthor_drv.c | 35
Implement drm object's status callback.
Also, we consider a PRIME imported BO to be resident if its matching
dma_buf has an open attachment, which means its backing storage had already
been allocated.
Signed-off-by: Adrián Larumbe
Reviewed-by: Liviu Dudau
---
drivers/gpu/drm/pa
mate input
values and output format in a documentation file.
Signed-off-by: Adrián Larumbe
---
Documentation/gpu/panthor.rst | 46 +++
drivers/gpu/drm/panthor/panthor_drv.c | 39 +++
2 files changed, 85 insertions(+)
create mode 100644 Documen
ts is to be interpreted as the
>> integer multiplier of a 10-power of 2, so 1 would give us size in Kib and 2
>> in Mib. If we want drm-file functions to pick the highest unit, then 0
>> should be passed.
>>
>> Signed-off-by: Adrián Larumbe
>> ---
>> drivers
On 31.08.2023 16:54, Steven Price wrote:
>On 24/08/2023 02:34, Adrián Larumbe wrote:
>> The drm-stats fdinfo tags made available to user space are drm-engine,
>> drm-cycles, drm-max-freq and drm-curfreq, one per job slot.
>>
>> This deviates from standard practice in
On 30.08.2023 12:31, Boris Brezillon wrote:
>On Thu, 24 Aug 2023 02:34:46 +0100
>Adrián Larumbe wrote:
>
>> A new DRM GEM object function is added so that drm_show_memory_stats can
>> provider more accurate memory usage numbers.
>
> s/provider/provide/
>
>>
On 30.08.2023 12:17, Boris Brezillon wrote:
>On Thu, 24 Aug 2023 02:34:45 +0100
>Adrián Larumbe wrote:
>
>> The drm-stats fdinfo tags made available to user space are drm-engine,
>> drm-cycles, drm-max-freq and drm-curfreq, one per job slot.
>
>Pretty sure this has
On 30.08.2023 12:52, Boris Brezillon wrote:
>On Thu, 24 Aug 2023 02:34:48 +0100
>Adrián Larumbe wrote:
>
>> BO's RSS is updated every time new pages are allocated and mapped for the
>> object, either in its entirety at creation time for non-heap buffers, or
>> el
These GPU registers will be used when programming the cycle counter, which
we need for providing accurate fdinfo drm-cycles values to user space.
Signed-off-by: Adrián Larumbe
Reviewed-by: Boris Brezillon
Reviewed-by: Steven Price
---
drivers/gpu/drm/panfrost/panfrost_regs.h | 5 +
1 file
previous multiplier for more accurate BO size numbers
- Removed special handling of Prime imported BO RSS
- Use rss_size only for heap objects
- Use bo->base.madv instead of specific purgeable flag
- Fixed kernel test robot warnings
Adrián Larumbe (8):
drm/panfrost: Add cycle count GPU
job IRQ and the actual time of the sample.
- Time spent in the engine queue waiting for the GPU to pick up the next
job.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panfrost/panfrost_devfreq.c | 8 +++
drivers/gpu/drm/panfrost/panfrost_devfreq.h | 3 ++
drivers/gpu/drm/panfrost
.
Toggling of GPU cycle counting has to be done through a module parameter.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panfrost/panfrost_device.c | 5 +++
drivers/gpu/drm/panfrost/panfrost_device.h | 6 +++
drivers/gpu/drm/panfrost/panfrost_gpu.c| 43 ++
drivers
e's object handle database spinlock, so there's potential
for a race condition here.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panfrost/panfrost_drv.c | 1 +
drivers/gpu/drm/panfrost/panfrost_gem.c | 14 ++
2 files changed, 15 insertions(+)
diff --git a/driv
Allow user space to decide whether the cycle counting register should be
enabled. The main goal is letting tools like nvtop or IGT's gputop access
this information in debug builds to obtain engine utilisation numbers.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panfrost/Mak
their entirety at
BO creation time.
This calculation is unnecessary for imported PRIME objects, since heap
buffers cannot be exported by our driver, and the actual BO RSS size is the
one reported in its attached dmabuf structure.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panfrost/panf
ly be backed by system memory
after successive GPU page faults.
Provide a new DRM object generic function that would allow drivers to
return a more accurate RSS size for their BOs.
Signed-off-by: Adrián Larumbe
Reviewed-by: Boris Brezillon
---
drivers/gpu/drm/drm_file.c | 5 -
include/drm/drm
The current implementation will try to pick the highest available size
display unit as soon as the BO size exceeds that of the previous
multiplier.
By selecting a higher threshold, we could show more accurate size numbers.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/drm_file.c | 2 +-
1
On 06.09.2023 09:57, Boris Brezillon wrote:
>On Tue, 5 Sep 2023 19:45:18 +0100
>Adrián Larumbe wrote:
>
>> diff --git a/drivers/gpu/drm/panfrost/panfrost_job.c
>> b/drivers/gpu/drm/panfrost/panfrost_job.c
>> index 033f5e684707..8b1bf6ac48f8 100644
>&g
On 06.09.2023 09:21, Boris Brezillon wrote:
>On Tue, 5 Sep 2023 19:45:18 +0100
>Adrián Larumbe wrote:
>
>> In a future development, we will want to keep track of the number of GPU
>> cycles spent on a given job. That means we should enable it only when the
>> GPU ha
On 06.09.2023 10:01, Boris Brezillon wrote:
>On Tue, 5 Sep 2023 19:45:23 +0100
>Adrián Larumbe wrote:
>
>> BO's RSS is updated every time new pages are allocated on demand and mapped
>> for the object at GPU page fault's IRQ handler, but only for heap buffers.
&g
On 06.09.2023 10:11, Boris Brezillon wrote:
>On Tue, 5 Sep 2023 19:45:24 +0100
>Adrián Larumbe wrote:
>
>> The current implementation will try to pick the highest available size
>> display unit as soon as the BO size exceeds that of the previous
>> multiplier.
&
These GPU registers will be used when programming the cycle counter, which
we need for providing accurate fdinfo drm-cycles values to user space.
Signed-off-by: Adrián Larumbe
Reviewed-by: Boris Brezillon
Reviewed-by: Steven Price
---
drivers/gpu/drm/panfrost/panfrost_regs.h | 5 +
1 file
their entirety at
BO creation time.
This calculation is unnecessary for imported PRIME objects, since heap
buffers cannot be exported by our driver, and the actual BO RSS size is the
one reported in its attached dmabuf structure.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panfrost/panf
sful mapping.
- drm-file picks an fdinfo memory object size unit that doesn't lose precision.
Adrián Larumbe (6):
drm/panfrost: Add cycle count GPU register definitions
drm/panfrost: Add fdinfo support GPU load metrics
drm/panfrost: Add fdinfo support for memory stats
drm/drm_file: A
power waste in case no
engine usage measuring is necessary.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panfrost/Makefile | 2 +
drivers/gpu/drm/panfrost/panfrost_debugfs.c | 20
drivers/gpu/drm/panfrost/panfrost_debugfs.h | 13 +
drivers/gpu/drm/pan
ly be backed by system memory
after successive GPU page faults.
Provide a new DRM object generic function that would allow drivers to
return a more accurate RSS size for their BOs.
Signed-off-by: Adrián Larumbe
Reviewed-by: Boris Brezillon
---
drivers/gpu/drm/drm_file.c | 5 -
include/drm/drm
igger BO's, their size will naturally be aligned on something
bigger than a 4 KiB page, so in practice it is very unlikely their display
unit would default to KiB.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/drm_file.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/d
e's object handle database spinlock, so there's potential
for a race condition here.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panfrost/panfrost_drv.c | 2 ++
drivers/gpu/drm/panfrost/panfrost_gem.c | 14 ++
2 files changed, 16 insertions(+)
diff --git a/driv
set of counters than the
current one implies a counter reset, which also messes up with the ring
buffer's extraction and insertion pointers. This is an unfortunate
hardware limitation.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panthor/Makefile | 3 +-
drivers/gpu/drm/p
When vm-binding an already-created BO, the entirety of its virtual size is
then backed by system memory, so its RSS is always the same as its virtual
size.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panthor/panthor_drv.c | 1 +
drivers/gpu/drm/panthor/panthor_gem.c | 12
2
queue slot number and the
queue's index within the group.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panthor/panthor_devfreq.c | 10 +
drivers/gpu/drm/panthor/panthor_device.h | 11 ++
drivers/gpu/drm/panthor/panthor_drv.c | 31
drivers/gpu/drm/panthor/panthor_
This patch series enables userspace utilities like gputop and nvtop to
query a render context's fdinfo file and figure out rates of engine
and memory utilisation.
Adrián Larumbe (2):
drm/panthor: Enable fdinfo for cycle and time measurements
drm/panthor: Enable fdinfo for memory
s knob and replace it with a sysfs file that accomplishes
the same functionality, and document its ABI in a separate file.
Signed-off-by: Adrián Larumbe
---
.../testing/sysfs-driver-panfrost-profiling | 10 +
Documentation/gpu/panfrost.rst| 9
drivers/gpu/drm/panfros
.laru...@collabora.com/
Adrián Larumbe (1):
drm/panfrost: Replace fdinfo's profiling debugfs knob with sysfs
.../testing/sysfs-driver-panfrost-profiling | 10 +
Documentation/gpu/panfrost.rst| 9
drivers/gpu/drm/panfrost/Makefile | 2 -
drive
On 11.03.2024 11:02, Boris Brezillon wrote:
> On Wed, 6 Mar 2024 08:33:47 +
> Tvrtko Ursulin wrote:
>
> > On 06/03/2024 01:56, Adrián Larumbe wrote:
> > > Debugfs isn't always available in production builds that try to squeeze
> > > every single byte
If job accounting is disabled, then both fdinfo's drm-engine and drm-cycle
key values will remain immutable. In that case, it makes more sense not to
display them at all to avoid confusing user space profiling tools.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panfrost/panfrost_drv.c
On 20.10.2023 11:44, Steven Price wrote:
> sg_page_iter_page() doesn't return an error code, so the IS_ERR() check
> is wrong and the error path will never be executed. This also allows
> simplifying the code to remove the local variable 'page'.
>
> CC: Adri
Hi Steven, thanks for the review.
On 13.06.2024 16:28, Steven Price wrote:
> On 06/06/2024 01:49, Adrián Larumbe wrote:
> > This patch series enables userspace utilities like gputop and nvtop to
> > query a render context's fdinfo file and figure out rates of engine
> &
cycle and timestamp
calculations and job accounting memory management, and a second one
that enables fdinfo.
- Moved NUM_INSTRS_PER_SLOT to the file prelude
- Removed nelem variable from the group's struct definition.
- Precompute size of group's syncobj BO to avoid code duplication.
- Some
when initialising the DRM scheduler for each queue, with a
profiled job contributing twice as many credits.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panthor/panthor_device.h | 2 +
drivers/gpu/drm/panthor/panthor_sched.c | 244 ---
2 files changed, 216 insertions(+), 30 d
This commit introduces a DRM device sysfs attribute that lets UM control
the job accounting status in the device. The knob variable had been brought
in as part of a previous commit, but now we're able to fix it manually.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panthor/panthor_
commit, and supporting manual toggle of it will be
the matter of a later commit.
Signed-off-by: Adrián Larumbe
---
drivers/gpu/drm/panthor/panthor_devfreq.c | 18 -
drivers/gpu/drm/panthor/panthor_device.h | 10 +
drivers/gpu/drm/panthor/panthor_drv.c | 33
drivers
Implement drm object's status callback.
Also, we consider a PRIME imported BO to be resident if its matching
dma_buf has an open attachment, which means its backing storage had already
been allocated.
Signed-off-by: Adrián Larumbe
Reviewed-by: Liviu Dudau
---
drivers/gpu/drm/pa
1 - 100 of 416 matches
Mail list logo
