Acked-by: Aaron Liu
BR,
Aaron Liu
> -Original Message-
> From: Quan, Evan
> Sent: Tuesday, November 19, 2019 11:57 AM
> To: amd-gfx@lists.freedesktop.org
> Cc: Liu, Aaron ; Quan, Evan
> Subject: [PATCH] drm/amd/powerplay: add missing header file declaration
>
> This can fix the compil
Reviewed-by: Hawking Zhang
Regards,
Hawking
-Original Message-
From: amd-gfx On Behalf Of Evan Quan
Sent: 2019年11月19日 11:57
To: amd-gfx@lists.freedesktop.org
Cc: Quan, Evan ; Liu, Aaron
Subject: [PATCH] drm/amd/powerplay: add missing header file declaration
This can fix the compile err
This can fix the compile errors below:
drivers/gpu/drm/amd/amdgpu/../powerplay/smu_v11_0.c: In function
‘smu_v11_0_baco_set_state’:
drivers/gpu/drm/amd/amdgpu/../powerplay/smu_v11_0.c:1674:27: error: implicit
declaration of function ‘amdgpu_ras_get_context’
[-Werror=implicit-function-declaration
Series is reviewed-by: Evan Quan
-Original Message-
From: amd-gfx On Behalf Of Alex Deucher
Sent: Friday, November 15, 2019 12:42 AM
To: amd-gfx@lists.freedesktop.org
Cc: Deucher, Alexander
Subject: [PATCH 2/2] drm/amdgpu: disable gfxoff when using register read
interface
When gfxoff
Hi Chris,
The info ioctl code path seems already covered at libdrm initialization time:
amdgpu_device_initialize()
-> amdgpu_queury_gpu_info_init()
-> amdgpu_read_mm_registers()
-> drmCommandWrite(fd, DRM_AMDGPU_INFO, ...)
BR,
Xiaojie
From: C
Acked-by: Xiaojie Yuan
BR,
Xiaojie
From: Alex Deucher
Sent: Tuesday, November 19, 2019 1:24 AM
To: Yuan, Xiaojie
Cc: Deucher, Alexander; amd-gfx@lists.freedesktop.org
Subject: Re: [PATCH 2/2] drm/amdgpu: disable gfxoff when using register read
interface
On 2019-11-18 17:24, Alex Sierra wrote:
Only for the debugger use case.
[why]
Avoid endless translation retries, after an invalid address access has
been issued to the GPU. Instead, the trap handler is forced to enter by
generating a no-retry-fault.
A s_trap instruction is inserted in the debugg
Only for the debugger use case.
[why]
Avoid endless translation retries, after an invalid address access has
been issued to the GPU. Instead, the trap handler is forced to enter by
generating a no-retry-fault.
A s_trap instruction is inserted in the debugger case to let the wave to
enter trap hand
Flag added to indicate if the amdgpu vm context is used for compute or
graphics.
Change-Id: Ia813037fda2ec2947d73f5c7328388078fbeebe5
Signed-off-by: Alex Sierra
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 3 +++
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h | 2 ++
2 files changed, 5 insertions(+)
di
Ye, I noticed - I just relied on the documentation in this case.
Andrey
On 11/18/19 3:01 PM, Christian König wrote:
Well then we should probably update the documentation.
Take a look at the implementation, there is no compiler or SMP barrier
at all.
Christian.
Am 18.11.19 um 18:01 schrieb
Applied. Thanks!
Alex
On Mon, Nov 18, 2019 at 3:15 AM Quan, Evan wrote:
>
> Reviewed-by: Evan Quan
>
> > -Original Message-
> > From: Chen Wandun
> > Sent: Monday, November 18, 2019 4:04 PM
> > To: Quan, Evan ; Deucher, Alexander
> > ; amd-gfx@lists.freedesktop.org; linux-
> > ker...@
Am 18.11.19 um 18:52 schrieb Andrey Grodzovsky:
Problem:
Due to a race between drm_sched_cleanup_jobs in sched thread and
drm_sched_job_timedout in timeout work there is a possiblity that
bad job was already freed while still being accessed from the
timeout thread.
Fix:
Instead of just peeking a
Well then we should probably update the documentation.
Take a look at the implementation, there is no compiler or SMP barrier
at all.
Christian.
Am 18.11.19 um 18:01 schrieb Andrey Grodzovsky:
The documentation states it can be used safely with concurrent
list_del_init so I assume it's true
On Mon, Nov 18, 2019 at 6:17 AM Changfeng.Zhu wrote:
>
> From: changzhu
>
> To get the value of gpu_busy_percent, it needs to realize
> get_current_activity_percent and get_metrics_table.
> The framework of renoir smu is different from old ones like raven. It
> needs to realize get_current_activi
On Mon, Nov 18, 2019 at 5:07 AM Hawking Zhang wrote:
>
> Drop all the IP specific cmd_submit callback function
> and use the common helper instead
>
> Change-Id: I865b26fd25a696f213e888beb927717dec884546
> Signed-off-by: Hawking Zhang
Nice cleanup!
Series is:
Reviewed-by: Alex Deucher
> ---
>
Problem:
Due to a race between drm_sched_cleanup_jobs in sched thread and
drm_sched_job_timedout in timeout work there is a possiblity that
bad job was already freed while still being accessed from the
timeout thread.
Fix:
Instead of just peeking at the bad job in the mirror list
remove it from th
Applied. Thanks!
Alex
On Mon, Nov 18, 2019 at 1:56 AM Quan, Evan wrote:
>
> Reviewed-by: Evan Quan
>
> -Original Message-
> From: Colin King
> Sent: Friday, November 15, 2019 5:48 PM
> To: Rex Zhu ; Quan, Evan ; Deucher,
> Alexander ; Koenig, Christian
> ; Zhou, David(ChunMing) ;
>
On Mon, Nov 18, 2019 at 3:37 AM Frederick Lawler wrote:
>
> Commit 8c0d3a02c130 ("PCI: Add accessors for PCI Express Capability")
> added accessors for the PCI Express Capability so that drivers didn't
> need to be aware of differences between v1 and v2 of the PCI
> Express Capability.
>
> Replace
Applied. thanks!
Alex
On Mon, Nov 18, 2019 at 10:07 AM zhengbin wrote:
>
> Fixes coccicheck warning:
>
> drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c:64:13-31: WARNING: dma_alloc_coherent
> use in ih -> ring already zeroes out memory, so memset is not needed
>
> Reported-by: Hulk Robot
> Signed-of
Can I get an RB or AB on the patch from anyone?
Thanks!
Alex
On Mon, Nov 18, 2019 at 9:09 AM Yuan, Xiaojie wrote:
>
> Hi Alex,
>
> I tried on Navi14 with Gfxoff enabled (gfx in 'OFF' state when I run the test
> program) and used a test program to read GRBM_STATUS/CP_STAT/GB_ADDR_CONFIG
> via
Applied the series. Thanks!
Alex
On Fri, Nov 15, 2019 at 9:56 AM zhengbin wrote:
>
> zhengbin (6):
> drm/radeon: remove set but not used variable 'size','relocs_chunk'
> drm/radeon: remove set but not used variable 'backbias_response_time'
> drm/radeon: remove set but not used variable 'd
The documentation states it can be used safely with concurrent
list_del_init so I assume it's true - but I think my even bigger mistake
is that without locking i just do list_first_entry right after
list_empty_careful and by this can grab pointer to the same job as
concurrent drm_sched_job_time
list_empty_careful() should only be used for optimizing cases, but never
if you need to rely on the result.
The problem is that the function doesn't has any memory barriers
whatsoever, it just checks if the next and prev pointer are both empty
instead of just the next pointer.
Christian.
Am
Can you explain why ? As I see it - list_empty_careful is specifically
designed for the case where the only other concurrent operation in
progress is list_del_init
(https://www.kernel.org/doc/htmldocs/kernel-api/API-list-empty-careful.html)
- which is exactly what happens in this patch, no othe
Hi Xiaojie,
could you add that test to the unit tests we have in libdrm?
Would be rather nice to have,
Christian.
Am 18.11.19 um 15:09 schrieb Yuan, Xiaojie:
Hi Alex,
I tried on Navi14 with Gfxoff enabled (gfx in 'OFF' state when I run the test
program) and used a test program to read GRBM_S
Hi Andrey,
the only thing which doesn't looks so good is the switch to
list_empty_careful in drm_sched_cleanup_jobs.
We either take the lock here or we don't, but please not that extra
checking.
Christian.
Am 18.11.19 um 15:07 schrieb Andrey Grodzovsky:
Thanks Emily.
Christan - ping for
Fixes coccicheck warning:
drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c:64:13-31: WARNING: dma_alloc_coherent
use in ih -> ring already zeroes out memory, so memset is not needed
Reported-by: Hulk Robot
Signed-off-by: zhengbin
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c | 1 -
1 file changed, 1 dele
Am 18.11.19 um 15:05 schrieb Andrey Grodzovsky:
Christian - ping.
Andrey
On 11/14/19 11:41 AM, Andrey Grodzovsky wrote:
On 11/13/19 10:09 PM, Zhou1, Tao wrote:
Two questions:
1. "we lose all reservation during ASIC reset"
Are you sure of it? I remember the content of vram may be lost after
Thank you Christian. An intentional use of invalid PTE setting sounds
reasonable for debugger case. I wasn't able to understand because I though this
is preparation work for page migration.
Hi Alex,
Please add some comments in the commit message to say that this is
intentionally invalid PTE se
Hi Alex,
I tried on Navi14 with Gfxoff enabled (gfx in 'OFF' state when I run the test
program) and used a test program to read GRBM_STATUS/CP_STAT/GB_ADDR_CONFIG via
DRM_IOCTL_AMDGPU_INFO.
All read out values are valid as below (in this read cycle, I saw gfx block is
first awakened and then p
Thanks Emily.
Christan - ping for review.
Andrey
On 11/14/19 11:39 PM, Deng, Emily wrote:
Hi Andrey,
Currently, I am busying with another issue, maybe will try next week.
Best wishes
Emily Deng
-Original Message-
From: Grodzovsky, Andrey
Sent: Friday, November 15, 2019 6:14
Christian - ping.
Andrey
On 11/14/19 11:41 AM, Andrey Grodzovsky wrote:
On 11/13/19 10:09 PM, Zhou1, Tao wrote:
Two questions:
1. "we lose all reservation during ASIC reset"
Are you sure of it? I remember the content of vram may be lost after
reset but the allocated status could be reserved
Hi Oak,
well your questions are really valid. This is a follow up from an
internal discussion on how to handle the debug trap handler on Vega and
newer.
The key missing information in the commit message is that we are
intentionally using an invalid combination of flags (F and P set at the
s
Am 18.11.19 um 13:27 schrieb Yintian Tao:
There is one regression from 042f3d7b745cd76aa
To put flush_delayed_work after adev->shutdown = true
which will make amdgpu_ih_process not response the irq
At last, all ib ring tests will be failed just like below
[drm] amdgpu: finishing device.
[drm] Fe
There is one regression from 042f3d7b745cd76aa
To put flush_delayed_work after adev->shutdown = true
which will make amdgpu_ih_process not response the irq
At last, all ib ring tests will be failed just like below
[drm] amdgpu: finishing device.
[drm] Fence fallback timer expired on ring gfx
[drm]
Hi Christian
Thanks, I got it, I will use flush_delayed_work to replace canceling.
Best Regards
Yintian Tao
-Original Message-
From: Koenig, Christian
Sent: 2019年11月18日 19:32
To: Tao, Yintian
Cc: amd-gfx@lists.freedesktop.org
Subject: Re: [PATCH] drm/amdgpu: put cancel dealyed work a
Good catch, but I would still prefer to use flush_delayed_work() instead
of canceling it.
Regards,
Christian.
Am 18.11.19 um 09:21 schrieb Yintian Tao:
There is one regression from 042f3d7b745cd76aa
and one improvement here.
-regression:
put flush_delayed_work after adev->shutdown = true
which
Am 18.11.19 um 10:19 schrieb Liu, Aaron:
BR,
Aaron Liu
-Original Message-
From: Koenig, Christian
Sent: Friday, November 15, 2019 8:50 PM
To: Huang, Ray ; Liu, Aaron
Cc: amd-gfx@lists.freedesktop.org; Deucher, Alexander
; Olsak, Marek ;
Liu, Leo ; Tuikov, Luben
Subject: Re: [PATCH 0
From: changzhu
To get the value of gpu_busy_percent, it needs to realize
get_current_activity_percent and get_metrics_table.
The framework of renoir smu is different from old ones like raven. It
needs to realize get_current_activity_percent and get_metrics_table in
renoir_ppt.c like navi10.
Chan
Reviewed-by: John Clements
-Original Message-
From: Hawking Zhang
Sent: Monday, November 18, 2019 6:20 PM
To: amd-gfx@lists.freedesktop.org; Clements, John ;
Deucher, Alexander
Cc: Zhang, Hawking
Subject: [PATCH] drm/amdgpu: pull ras controller int status only when ras
enabled
ras_
ras_controller_irq and athub_err_event_irq are only registered
when PCIE_BIF ras is marked as supported. as the result, the driver
also just need pull the int status in such case.
Change-Id: Ibd1f29be253e0e60f9be7ff2208f5cf2b78a56a4
Signed-off-by: Hawking Zhang
---
drivers/gpu/drm/amd/amdgpu/amd
[Why]
HDMI 2.0 adds aspect ratio attribute to distinguish different
4k modes. According to Appendix E of HDMI 2.0 spec, source should
use VSIF to indicate video mode only when the mode is one defined
in HDMI 1.4b 4K modes. Otherwise, use AVI infoframes to convey VIC.
Current code doesn't take aspe
[Why]
In hdmi_mode_alternate_clock(), it adds an exception for VIC 4
mode (4096x2160@24) due to there is no alternate clock defined for
that mode in HDMI1.4b. But HDMI2.0 adds 23.98Hz for that mode.
[How]
Remove the exception
v2: Adjust the comment description of hdmi_mode_alternate_clock()
due t
Series Reviewed-by: John Clements
-Original Message-
From: Hawking Zhang
Sent: Monday, November 18, 2019 6:07 PM
To: amd-gfx@lists.freedesktop.org; Clements, John ;
Deucher, Alexander
Cc: Zhang, Hawking
Subject: [PATCH 1/3] drm/amdgpu: add psp funcs for ring write pointer read/write
The ring write pointer regsiter update is the only part that
is IP specific ones in psp_cmd_submit function.
Add two callbacks for wptr read/write so that we unify the
psp_cmd_submit function for all the ASICs.
Change-Id: Idd0b70bd8682f8bd2f9b92098d6ce55fa4cee310
Signed-off-by: Hawking Zhang
---
Except for ring wptr update, the psp ring cmd submission
function shouldn't be IP specific one. Create a common
helper function to be shared for all the ASICs.
Change-Id: Ic59f22f762fc7580fd9438d865d37d5996ce8b84
Signed-off-by: Hawking Zhang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c | 50 +
Drop all the IP specific cmd_submit callback function
and use the common helper instead
Change-Id: I865b26fd25a696f213e888beb927717dec884546
Signed-off-by: Hawking Zhang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c | 2 +-
drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h | 5 ---
drivers/gpu/drm/amd/amd
Series is
Reviewed-by: Hawking Zhang
Regards,
Hawking
-Original Message-
From: amd-gfx On Behalf Of Evan Quan
Sent: 2019年11月18日 17:20
To: amd-gfx@lists.freedesktop.org
Cc: Quan, Evan
Subject: [PATCH 2/2] drm/amd/powerplay: add Arcturus baco reset support
Enable baco reset support on Ar
Thanks for reviews.
As Christian's suggestion, I added drm_version checking, for the older driver,
security test suites are disabled.
I create a remote branch for libdrm: remotes/origin/aaliu/for-tmz-support
BR,
Aaron Liu
From: Olsak, Marek
Sent: Saturday, November 16, 2019 3:53 AM
To: Koenig,
Enable baco reset support on Arcturus.
Change-Id: I7b69016ee0d238e0fcb323aa10215e29924a6ca6
Signed-off-by: Evan Quan
---
drivers/gpu/drm/amd/amdgpu/soc15.c | 1 +
drivers/gpu/drm/amd/powerplay/arcturus_ppt.c | 7 +++
drivers/gpu/drm/amd/powerplay/smu_v11_0.c| 14 ++---
Added bif doorbell interrupt setting and applied different
settings for BACO reset for RAS recovery.
Change-Id: I823b2d478699d469ecc7746e2a8fb1110a4a146f
Signed-off-by: Evan Quan
---
drivers/gpu/drm/amd/powerplay/smu_v11_0.c | 30 ---
1 file changed, 27 insertions(+), 3 delet
BR,
Aaron Liu
> -Original Message-
> From: Koenig, Christian
> Sent: Friday, November 15, 2019 8:50 PM
> To: Huang, Ray ; Liu, Aaron
> Cc: amd-gfx@lists.freedesktop.org; Deucher, Alexander
> ; Olsak, Marek ;
> Liu, Leo ; Tuikov, Luben
> Subject: Re: [PATCH 07/12] amdgpu: using IB flag
return errno code to caller when error occur, and meanwhile
remove gcc '-Wunused-but-set-variable' warning.
drivers/gpu/drm/amd/amdgpu/../powerplay/smumgr/vegam_smumgr.c: In function
vegam_populate_smc_boot_level:
drivers/gpu/drm/amd/amdgpu/../powerplay/smumgr/vegam_smumgr.c:1364:6: warning:
var
OK, it indeed should return the 'result' to caller,
I will make a modification and repost the patch.
Thanks
Chen Wandun
On 2019/11/18 14:50, Quan, Evan wrote:
Thanks. But it's better to return the 'result' out on 'result != 0'.
Regards,
Evan
-Original Message-
From: Chen Wandun
Sent:
Commit 8c0d3a02c130 ("PCI: Add accessors for PCI Express Capability")
added accessors for the PCI Express Capability so that drivers didn't
need to be aware of differences between v1 and v2 of the PCI
Express Capability.
Replace pci_read_config_word() and pci_write_config_word() calls with
pcie_ca
program_hpd_filter() currently fails to check for the errors
returned in construct(). This patch returns error in
case of failure.
Signed-off-by: Aditya Pakki
---
drivers/gpu/drm/amd/display/dc/core/dc_link.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/a
There is one regression from 042f3d7b745cd76aa
and one improvement here.
-regression:
put flush_delayed_work after adev->shutdown = true
which will make amdgpu_ih_process not response the irq
At last, all ib ring tests will be failed just like below
[drm] amdgpu: finishing device.
[drm] Fence fall
Reviewed-by: Evan Quan
> -Original Message-
> From: Chen Wandun
> Sent: Monday, November 18, 2019 4:04 PM
> To: Quan, Evan ; Deucher, Alexander
> ; amd-gfx@lists.freedesktop.org; linux-
> ker...@vger.kernel.org; dri-de...@lists.freedesktop.org
> Cc: chenwan...@huawei.com
> Subject: [PATC
58 matches
Mail list logo
