[AMD Official Use Only - Internal Distribution Only]
Series are Reviewed-by: Kenneth Feng
-Original Message-
From: Powell, Darren
Sent: Friday, April 23, 2021 11:23 AM
To: amd-gfx@lists.freedesktop.org
Cc: Powell, Darren
Subject: [PATCH v2 0/3] Prevent set of DCEFCLK on smu_v11 gpus
On Thu, 2021-04-22 at 18:33 -0400, Lyude Paul wrote:
> OK - talked with Ville a bit on this and did some of my own research, I
> actually think that moving i2c to drm_dp_aux_init() is the right decision
> for
> the time being. The reasoning behind this being that as shown by my previous
> work of
[AMD Official Use Only - Internal Distribution Only]
Hi Christian,
It will be used by a user space SMI-lib for GPU status query.
Hi David, please feel free to correct my statement since I'm not sure I have a
view of whole picture.
Thanks,
Jiawei
From: Christian König
Sent: Thursday, April
Writing to dcefclk causes the gpu to become unresponsive, and requires a reboot.
Patch ignores a .force_clk_levels(SMU_DCEFCLK) call and issues an
info message.
Signed-off-by: Darren Powell
---
drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c | 5 -
v2 : change condition to apply to all chips after NAVI10
Writing to dcefclk causes the gpu to become unresponsive, and requires a reboot.
Patch prevents user from successfully writing to file pp_dpm_dcefclk on parts
NAVI10 and newer, and gives better user feedback that this operation is not
=== Description ===
Set of simple patches to prevent attempts to set dcefclk on NAVI10
=== Test System ===
* DESKTOP(AMD FX-8350 + NAVI10(731F/ca), BIOS: F2)
+ ISO(Ubuntu 20.04.1 LTS)
+ Kernel(5.11.0-custom-amdinternal-dirty)
=== Patch Summary ===
linux: (g...@gitlab.freedesktop.org:agd5f)
Insert the value of the response to error message emitted when the
SMU msg pre-check failes
Signed-off-by: Darren Powell
---
drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c
From: Aric Cyr
This DC patchset brings along following fixes:
* [FW Promotion] Release 0.0.63
* fix HDCP reset sequence on reinitialize
* Fix BSOD with NULL check
* Added multi instance support for ABM
* Revert wait vblank on update dpp clock
* skip program clock when allow seamless boot
* Add
From: Anthony Koo
Signed-off-by: Anthony Koo
Reviewed-by: Anthony Koo
Acked-by: Wayne Lin
---
drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h | 9 ++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h
From: "Max.Tseng"
[Why] New platform. Need to add corresponding register control
Signed-off-by: Max.Tseng
Reviewed-by: Anthony Koo
Acked-by: Wayne Lin
---
drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dio_stream_encoder.h | 2 ++
1 file changed, 2 insertions(+)
diff --git
From: Yu-ting Shen
[why]
we don't support authentication with DEVICE_COUNT=0
[how]
check value DEVICE_COUNT before doing authentication
Signed-off-by: Yu-ting Shen
Reviewed-by: Wenjing Liu
Acked-by: Wayne Lin
---
drivers/gpu/drm/amd/display/modules/hdcp/hdcp1_execution.c | 5 +
From: Wenjing Liu
[why]
As hardware team suggested that we need to add a max dsc bw overhead
into existing stream bandwidth when DSC is used.
The formula as below:
max_dsc_bw_overhead =
v_addressable * slice_count * 256 bit * pixel clock / v_total / h_total
effective stream bandwidth = pixel
From: Mikita Lipski
[why]
Previous statement would always evaluate to true
making it meaningless
[how]
Just check if a connector is MST by checking if its port exists.
Reported-by: kernel test robot
Signed-off-by: Mikita Lipski
Reviewed-by: Nicholas Kazlauskas
Acked-by: Wayne Lin
---
From: Robin Chen
[Why]
It seems there's a typo to set MASTER_UPDATE_LOCK_DB_EN when disable
doublebuffer lock.
[How]
Clear MASTER_UPDATE_LOCK_DB_EN when disable doublebuffer lock
Signed-off-by: Robin Chen
Reviewed-by: Joshua Aberback
Acked-by: Wayne Lin
---
From: "Max.Tseng"
[Why]
Add register for programming in new platform
Signed-off-by: Max.Tseng
Reviewed-by: Anthony Koo
Acked-by: Wayne Lin
---
drivers/gpu/drm/amd/display/dc/dcn30/dcn30_dio_stream_encoder.h | 2 ++
1 file changed, 2 insertions(+)
diff --git
From: Chris Park
[Why]
CLK mgr is null for server settings.
[How]
Guard the function with NULL check.
Signed-off-by: Chris Park
Reviewed-by: Nicholas Kazlauskas
Acked-by: Wayne Lin
---
drivers/gpu/drm/amd/display/dc/core/dc.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff
From: Jake Wang
[WHY & HOW]
ABM assumes only 1 eDP is connected. Refactored existing
ABM interface to support multiple instances.
Signed-off-by: Jake Wang
Reviewed-by: Nicholas Kazlauskas
Acked-by: Wayne Lin
---
drivers/gpu/drm/amd/display/dc/dce/dmub_abm.c | 30 +++
From: Yu-ting Shen
[WHY]
to initial ddc structure before we use them to avoid error checking.
[HOW]
add 0 as structure initialization value in DCN resource construct
Signed-off-by: Yu-ting Shen
Reviewed-by: Meenakshikumar Somasundaram
Acked-by: Wayne Lin
---
From: Stylon Wang
[Why & How]
Add a per-connector debugfs entry to expose internal display flag,
which is indication that the display is "internally connected"
and not hotpluggable.
Signed-off-by: Stylon Wang
Reviewed-by: Wayne Lin
---
.../amd/display/amdgpu_dm/amdgpu_dm_debugfs.c | 20
From: Lewis Huang
[Why]
Driver program dpp clock calculate by pipe split config but hw config is single
pipe.
[How]
Skip programming clock when allow seamless boot.
After porgramming pipe config, seamless boot flag will be clear.
Signed-off-by: Lewis Huang
Reviewed-by: Eric Yang
Acked-by:
From: Lewis Huang
[Why]
This change only fix dpp clock switch to lower case.
New solution later can fix both case, which is "dc: skip
program clock when allow seamless boot"
[How]
This reverts commit "dc: wait vblank when stream enabled
and update dpp clock"
Signed-off-by: Lewis Huang
From: Michael Strauss
[WHY]
New minor revision needs to be handled
[HOW]
Add switch case and assert to catch missing switch cases in the future
Signed-off-by: Michael Strauss
Reviewed-by: Eric Yang
Acked-by: Wayne Lin
---
drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c | 2 ++
1 file
From: Brandon Syu
[why]
When setup is called after hdcp has already setup,
it would cause to disable HDCP flow won’t execute.
[how]
Don't clean up hdcp content to be 0.
Signed-off-by: Brandon Syu
Reviewed-by: Wenjing Liu
Acked-by: Wayne Lin
---
This DC patchset brings along following fixes:
* [FW Promotion] Release 0.0.63
* fix HDCP reset sequence on reinitialize
* Fix BSOD with NULL check
* Added multi instance support for ABM
* Revert wait vblank on update dpp clock
* skip program clock when allow seamless boot
* Add new DP_SEC
After draining the stale retry fault, or failed to validate the range
to recover, have to remove the fault address from fault filter ring, to
be able to handle subsequent retry interrupt on same address. Otherwise
the retry fault will not be processed to recover until timeout passed.
Add interface to remove address from fault filter ring by resetting
fault ring entry of the fault address timestamp to 0, then future vm
fault on the address will be processed to recover.
Use spinlock to protect fault hash ring access by interrupt handler and
interrupt scheduled deferred work for
On 2021-04-21 3:22 a.m., Christian
König wrote:
Am
20.04.21 um 22:21 schrieb Philip Yang:
Add interface to remove address from fault
filter ring by resetting
fault ring entry of the fault address timestamp to 0, then
On 2021-04-21 11:29 a.m., Felix
Kuehling wrote:
On
2021-04-21 3:55 a.m., Christian König wrote:
Am 21.04.21 um 03:20 schrieb Felix
Kuehling:
Am 2021-04-20 um 4:21 p.m. schrieb
Philip Yang:
Regards,
Oak
On 2021-04-21, 9:31 PM, "amd-gfx on behalf of Felix Kuehling"
wrote:
Do AQL queue double-mapping with a single attach call. That will make it
easier to create per-GPU BOs later, to be shared between the two BO VA
mappings on the same GPU.
Freeing the
On 2021-04-22 7:20 p.m., Harry Wentland wrote:
[Why]
This hasn't been well tested and leads to complete system hangs on DCN1
based systems, possibly others.
The system hang can be reproduced by gesturing the video on the YouTube
Android app on ChromeOS into full screen.
[How]
Reject atomic
[Why]
This hasn't been well tested and leads to complete system hangs on DCN1
based systems, possibly others.
The system hang can be reproduced by gesturing the video on the YouTube
Android app on ChromeOS into full screen.
[How]
Reject atomic commits with non-zero drm_plane_state.src_x or src_y
OK - talked with Ville a bit on this and did some of my own research, I
actually think that moving i2c to drm_dp_aux_init() is the right decision for
the time being. The reasoning behind this being that as shown by my previous
work of fixing drivers that call drm_dp_aux_register() too early - it
Am 2021-04-22 um 8:44 a.m. schrieb Colin King:
> From: Colin Ian King
>
> The variable r is being initialized with a value that is never read
> and it is being updated later with a new value. The initialization is
> redundant and can be removed.
>
> Addresses-Coverity: ("Unused value")
>
Am 2021-04-22 um 8:31 a.m. schrieb Colin King:
> From: Colin Ian King
>
> Currently the call to kfd_process_gpuidx_from_gpuid is returning an
> int value and this is being assigned to a uint32_t variable gpuidx
> and this is being checked for a negative error return which is always
> going to be
On Tue, 2021-04-20 at 02:16 +0300, Ville Syrjälä wrote:
>
> The init vs. register split is intentional. Registering the thing
> and allowing userspace access to it before the rest of the driver
> is ready isn't particularly great. For a while now we've tried to
> move towards an architecture
Am 2021-04-22 um 11:40 a.m. schrieb Nirmoy Das:
> Remove unused AMDGPU_VM_CONTEXT_GFX and AMDGPU_VM_CONTEXT_COMPUTE
> flags.
>
> Signed-off-by: Nirmoy Das
I saw this too late. Ignore my comment on the previous patch. Both
patches 2 and 3 are
Reviewed-by: Felix Kuehling
> ---
>
Am 2021-04-22 um 11:39 a.m. schrieb Nirmoy Das:
> Currently only way to create compute vm is through
> amdgpu_vm_make_compute(). So vm_context isn't required
> anymore for amdgpu_vm_init().
>
> Signed-off-by: Nirmoy Das
I believe you can also remove the AMDGPU_VM_CONTEXT_* #defines from
Remove unused AMDGPU_GEM_CREATE_SHADOW flag.
Signed-off-by: Nirmoy Das
---
drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 5 +
include/uapi/drm/amdgpu_drm.h | 2 --
2 files changed, 1 insertion(+), 6 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
Remove unused AMDGPU_VM_CONTEXT_GFX and AMDGPU_VM_CONTEXT_COMPUTE
flags.
Signed-off-by: Nirmoy Das
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h | 3 ---
1 file changed, 3 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
index
Shadow BOs are only needed for vm code so call amdgpu_bo_create_shadow()
directly instead of depending on amdgpu_bo_create().
Signed-off-by: Nirmoy Das
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 70 +-
1 file changed, 47 insertions(+), 23 deletions(-)
diff --git
Remove shadow bo related code as vm code is creating shadow bo
using proper API. Without shadow bo code, amdgpu_bo_create() is basically
a wrapper around amdgpu_bo_do_create(). So rename amdgpu_bo_do_create()
to amdgpu_bo_create().
Signed-off-by: Nirmoy Das
---
Currently only way to create compute vm is through
amdgpu_vm_make_compute(). So vm_context isn't required
anymore for amdgpu_vm_init().
Signed-off-by: Nirmoy Das
---
drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c | 3 ++-
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 16 +++-
Exposed amdgpu_bo_create_shadow() will be needed
for amdgpu_vm handling.
Signed-off-by: Nirmoy Das
---
drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 6 +++---
drivers/gpu/drm/amd/amdgpu/amdgpu_object.h | 3 +++
2 files changed, 6 insertions(+), 3 deletions(-)
diff --git
Am 2021-04-22 um 10:47 a.m. schrieb Alex Sierra:
> SVM ranges are created for unregistered memory, triggered
> by page faults. These ranges are migrated/mapped to
> GPU VRAM memory.
>
> Signed-off-by: Alex Sierra
> ---
> drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 107 ++-
>
Am 2021-04-21 um 3:51 p.m. schrieb Alex Sierra:
> Attribute access value for default ranges is set, based on
> process xnack on/off.
> XNACK ON has GPU access attribute for unregistered ranges through page
> fault. While XNACK OFF has no access attribute for unregistered ranges.
>
> Signed-off-by:
On 2021-04-22 9:20 a.m., Felix Kuehling
wrote:
Am 2021-04-22 um 9:08 a.m. schrieb philip yang:
On 2021-04-20 9:25 p.m., Felix Kuehling wrote:
@@ -2251,14 +2330,34 @@ svm_range_restore_pages(struct amdgpu_device
*adev, unsigned int pasid,
SVM ranges are created for unregistered memory, triggered
by page faults. These ranges are migrated/mapped to
GPU VRAM memory.
Signed-off-by: Alex Sierra
---
drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 107 ++-
1 file changed, 104 insertions(+), 3 deletions(-)
diff --git
On 4/22/21 3:56 PM, Felix Kuehling wrote:
Am 2021-04-22 um 8:35 a.m. schrieb Nirmoy Das:
Fix vm->is_compute_context initialization in amdgpu_vm_init().
Signed-off-by: Nirmoy Das
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff
[AMD Official Use Only - Internal Distribution Only]
Looks good to me.
Reviewed-by: Dennis Li
Best Regards
Dennis Li
-Original Message-
From: Hawking Zhang
Sent: Thursday, April 22, 2021 10:06 PM
To: amd-gfx@lists.freedesktop.org; Li, Dennis
Cc: Zhang, Hawking
Subject: [PATCH]
aldebaran gfx ras is still under development
Signed-off-by: Hawking Zhang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
index bb0d027..f62873f
Am 2021-04-22 um 8:35 a.m. schrieb Nirmoy Das:
> Fix vm->is_compute_context initialization in amdgpu_vm_init().
>
> Signed-off-by: Nirmoy Das
> ---
> drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 6 +++---
> 1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git
On 4/22/21 2:48 PM, Christian König wrote:
Am 22.04.21 um 14:35 schrieb Nirmoy Das:
Shadow BOs are only needed for vm code so call amdgpu_bo_create_shadow()
directly instead of depending on amdgpu_bo_create().
Signed-off-by: Nirmoy Das
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 75
Is that useful to Vulkan/OpenGL/other clients in any way?
Christian.
Am 22.04.21 um 15:18 schrieb Gu, JiaWei (Will):
CC Tom.
On Apr 22, 2021, at 21:09, Gu, JiaWei (Will) wrote:
<[v2][umr]add-vbios-info-query.patch>
UMR patch which calls this new IOCTL attached.
Best regards,
Jiawei
On
Am 2021-04-22 um 9:08 a.m. schrieb philip yang:
>
>
> On 2021-04-20 9:25 p.m., Felix Kuehling wrote:
> @@ -2251,14 +2330,34 @@ svm_range_restore_pages(struct amdgpu_device
> *adev, unsigned int pasid,
}
mmap_read_lock(mm);
+retry_write_locked:
mutex_lock(>lock);
CC Tom.
> On Apr 22, 2021, at 21:09, Gu, JiaWei (Will) wrote:
>
> <[v2][umr]add-vbios-info-query.patch>
> UMR patch which calls this new IOCTL attached.
>
> Best regards,
> Jiawei
>
>> On Apr 22, 2021, at 10:34, Jiawei Gu wrote:
>>
>> Add AMDGPU_INFO_VBIOS_INFO subquery id for detailed
Am 2021-04-21 um 10:24 p.m. schrieb Jonathan Kim:
> In order to support multi-process debugging, HWS PM4 packet
> MAP_PROCESS requires an extension of 5 DWORDS to support targeting of
> per-vmid SPI debug control registers as well as watch points per process.
>
> v2: align upstream with internal
UMR patch which calls this new IOCTL attached.
Best regards,
Jiawei
> On Apr 22, 2021, at 10:34, Jiawei Gu wrote:
>
> Add AMDGPU_INFO_VBIOS_INFO subquery id for detailed vbios info.
>
> Signed-off-by: Jiawei Gu
> ---
> drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c| 19 +++
>
On 2021-04-20 9:25 p.m., Felix Kuehling
wrote:
Am 2021-04-20 um 8:45 p.m. schrieb Felix Kuehling:
Am 2021-04-19 um 9:52 p.m. schrieb Alex Sierra:
SVM ranges are created for unregistered memory, triggered
by page faults.
Am 22.04.21 um 14:35 schrieb Nirmoy Das:
Remove shadow bo related code as vm code is creating shadow bo
using proper API.
Signed-off-by: Nirmoy Das
---
drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 30 ++
1 file changed, 2 insertions(+), 28 deletions(-)
diff --git
From: Colin Ian King
The variable r is being initialized with a value that is never read
and it is being updated later with a new value. The initialization is
redundant and can be removed.
Addresses-Coverity: ("Unused value")
Signed-off-by: Colin Ian King
---
Am 22.04.21 um 14:35 schrieb Nirmoy Das:
Shadow BOs are only needed for vm code so call amdgpu_bo_create_shadow()
directly instead of depending on amdgpu_bo_create().
Signed-off-by: Nirmoy Das
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 75 +-
1 file changed, 49
From: Colin Ian King
Currently the call to kfd_process_gpuidx_from_gpuid is returning an
int value and this is being assigned to a uint32_t variable gpuidx
and this is being checked for a negative error return which is always
going to be false. Fix this by making gpuidx an int32_t. This makes
Remove unused AMDGPU_GEM_CREATE_SHADOW flag.
Signed-off-by: Nirmoy Das
---
drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 5 +
include/uapi/drm/amdgpu_drm.h | 2 --
2 files changed, 1 insertion(+), 6 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
Remove shadow bo related code as vm code is creating shadow bo
using proper API.
Signed-off-by: Nirmoy Das
---
drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 30 ++
1 file changed, 2 insertions(+), 28 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
Shadow BOs are only needed for vm code so call amdgpu_bo_create_shadow()
directly instead of depending on amdgpu_bo_create().
Signed-off-by: Nirmoy Das
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 75 +-
1 file changed, 49 insertions(+), 26 deletions(-)
diff --git
Fix vm->is_compute_context initialization in amdgpu_vm_init().
Signed-off-by: Nirmoy Das
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index
Exposed amdgpu_bo_create_shadow() will be needed
for amdgpu_vm handling.
Signed-off-by: Nirmoy Das
---
drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 6 +++---
drivers/gpu/drm/amd/amdgpu/amdgpu_object.h | 3 +++
2 files changed, 6 insertions(+), 3 deletions(-)
diff --git
Starting with Vega the hardware supports concurrent flushes
of VMID which can be used to implement per process VMID
allocation.
But concurrent flushes are mutual exclusive with back to
back VMID allocations, fix this to avoid a VMID used in
two ways at the same time.
Signed-off-by: Christian
Am 22.04.21 um 03:30 schrieb Felix Kuehling:
The dmabuf attachment should be updated by moving the SG BO to DOMAIN_CPU
and back to DOMAIN_GTT. This does not necessarily invoke the
populate/unpopulate callbacks. Do this in backend_bind/unbind instead.
Signed-off-by: Felix Kuehling
Am 22.04.21 um 12:49 schrieb Victor Zhao:
Add sdma single packet invalidation
Well NAK on many aspects of this.
Signed-off-by: Victor Zhao
---
drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c | 8
drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h | 4
Add sdma single packet invalidation
Signed-off-by: Victor Zhao
---
drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c | 8
drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h | 4
drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c | 14 +-
drivers/gpu/drm/amd/amdgpu/sdma_v5_0.c | 12
PSP clear_vf_fw feature is outdated and has been removed.
Remove the related functions.
Signed-off-by: Victor Zhao
---
drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c | 32 -
drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h | 1 -
2 files changed, 33 deletions(-)
diff --git
[AMD Official Use Only - Internal Distribution Only]
Reviewed-by: Tao Zhou mailto:tao.zh...@amd.com>>
From: amd-gfx on behalf of Kenneth Feng
Sent: Thursday, April 22, 2021 5:33 PM
To: amd-gfx@lists.freedesktop.org
Cc: Feng, Kenneth
Subject: [PATCH]
enable cgls to improve the runtime power efficiency.
Signed-off-by: Kenneth Feng
---
drivers/gpu/drm/amd/amdgpu/nv.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c
index 0142f6760ad2..9c4f232e81c0 100644
---
On 4/22/21 9:08 AM, Christian König wrote:
Am 21.04.21 um 18:35 schrieb Nirmoy Das:
Shadow BOs are only needed for vm code so call amdgpu_bo_create_shadow()
directly instead of depending on amdgpu_bo_create().
As far as I can see that won't work correctly.
Now you only create the shadow for
Am 21.04.21 um 18:35 schrieb Nirmoy Das:
Shadow BOs are only needed for vm code so call amdgpu_bo_create_shadow()
directly instead of depending on amdgpu_bo_create().
As far as I can see that won't work correctly.
Now you only create the shadow for the root BO, but not for anything else.
Am 21.04.21 um 05:46 schrieb Roy Sun:
Update the timestamp of scheduled fence on HW
completion of the previous fences
This allow more accurate tracking of the fence
execution in HW
Signed-off-by: David M Nieto
Signed-off-by: Roy Sun
Reviewed-by: Christian König for the series.
---
77 matches
Mail list logo
