On 09/21/2017 10:56 AM, Evan Quan wrote:
Change-Id: I28e9ca38b68234d0325a5b8a01d135649939c0af
Signed-off-by: Evan Quan
Reviewed-by: Junwei Zhang
---
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 23 +++
1 file changed, 23
SRIOV doesn't implement PMC capability of PCIe, so it can't update
power state by reading PMC register.
Currently, amdgpu driver doesn't disable pci device when removing
driver, the enable_cnt of pci device will not be decrease to 0.
When reloading driver, pci_enable_device will do nothing as
currently, not all asics implement this callback function
so not return error to avoid powerplay initialize failed
in those asices
Change-Id: I492b7ab54eebcc0d84c169edc0d5876c4529c619
Signed-off-by: Rex Zhu
---
drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c | 6 +++---
1
Change-Id: I28e9ca38b68234d0325a5b8a01d135649939c0af
Signed-off-by: Evan Quan
---
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 23 +++
1 file changed, 23 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
From: Yong Zhao
There are already CHIP_* definitions under amd_shared.h file on amdgpu
side, so KFD should reuse them rather than defining new ones.
Using enum for asic type requires default cases on switch statements
to prevent compiler warnings. WARN on unsupported ASICs.
Small self-contained KFD fixes that don't introduce any new
functionality or features. These may be suitable to include in
drm-fixes for 4.14.
v2:
- Rebased on gabbayo/amdkfd-next
- Addressed code review feedback
- Dropped "Set /dev/kfd permissions to 0666 by default"
Felix Kuehling (3):
From: Yong Zhao
The idea is to let kfd init and resume function share the same code path
as much as possible, rather than to have two copies of almost identical
code. That way improves the code readability and maintainability.
Signed-off-by: Yong Zhao
To avoid spamming the log.
Signed-off-by: Felix Kuehling
Reviewed-by: Oded Gabbay
---
drivers/gpu/drm/amd/amdkfd/kfd_events.c | 5 -
drivers/gpu/drm/amd/amdkfd/kfd_priv.h | 1 +
2 files changed, 5 insertions(+), 1 deletion(-)
diff --git
From: Yong Zhao
When we do suspend/resume through "sudo pm-suspend" while there is
HSA activity running, upon resume we will encounter HWS hanging, which
is caused by memory read/write failures. The root cause is that when
suspend, we neglected to unbind pasid from kfd device.
From: Yong Zhao
Several functions in DQM are shared between cpsch and nocpsch code.
Remove the misleading _nocpsch suffix from their names.
Signed-off-by: Yong Zhao
Signed-off-by: Felix Kuehling
Reviewed-by: Oded Gabbay
From: Yong Zhao
The hard-coded values related to VMID were removed in KFD, as those
values can be calculated in the KFD initialization function.
v2: remove unnecessary local variable
Signed-off-by: Yong Zhao
Signed-off-by: Felix Kuehling
From: Yong Zhao
Avoid intermediate negative numbers when doing calculations with a mix
of signed and unsigned variables where implicit conversions can lead
to unexpected results.
When kernel queue buffer wraps around to 0, we need to check that rptr
won't be overwritten by
From: Yong Zhao
The timeout in milliseconds should not be regarded as jiffies. This
commit fixed that.
v2:
- use msecs_to_jiffies
- change timeout_ms parameter to unsigned int to match msecs_to_jiffies
Signed-off-by: Yong Zhao
Signed-off-by: Felix
Adjust latencies and timeouts for dequeueing with HWS and consolidate
them in one place. Make them longer to allow long running waves to
complete without causing a timeout. The timeout is twice as long as the
latency plus some buffer to make sure we don't detect a timeout
prematurely.
Change
Am 20.09.2017 um 19:47 schrieb Li, Samuel:
No that isn't correct. Pinning just notes that the BO shouldn't be moved any
more. It doesn't wait for anything.
It does. The following is from amdgpu_gem_prime_pin(),
91 * Wait for all shared fences to complete before we switch to future
Am 20.09.2017 um 19:34 schrieb Li, Samuel:
If that is what this callback should do then this implementation would be
incorrect. Pinning doesn't wait for any GPU operation to finish.
During pining, it will all the fences to finish. That shall be OK.
No that isn't correct. Pinning just notes
Since newer asics have IP blocks with the same register we must specify both.
Signed-off-by: Tom St Denis
---
doc/umr.1| 2 +-
src/app/main.c | 3 ++-
src/app/umr_lookup.c | 33 -
3 files changed, 27 insertions(+), 11
> If that is what this callback should do then this implementation would be
> incorrect. Pinning doesn't wait for any GPU operation to finish.
During pining, it will all the fences to finish. That shall be OK.
Sam
> -Original Message-
> From: Koenig, Christian
> Sent: Wednesday,
> What happens when another thread is using amdgpu_dmabuf_ops to call
> begin_cpu_access/end_cpu_access when you are fixing it up again?
Right, that is an issue.
> I would just completely drop the two callbacks, pinning is not necessary for
> CPU access and thinking more about it it actually has
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/lib/chash.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/amd/lib/chash.c b/drivers/gpu/drm/amd/lib/chash.c
index e07e6f3..b8e45f3 100644
--- a/drivers/gpu/drm/amd/lib/chash.c
repeated defining in hwmgr.h
Change-Id: Id4fdbdb7df727516d9bdb2ca6c621b97b5ce509f
Signed-off-by: Rex Zhu
---
drivers/gpu/drm/amd/powerplay/inc/smumgr.h | 5 -
drivers/gpu/drm/amd/powerplay/smumgr/ci_smc.c | 2 +-
Change-Id: I722ecb827144adf42b149e41595580ca8d4385ae
Signed-off-by: Rex Zhu
---
drivers/gpu/drm/amd/powerplay/inc/smumgr.h| 31 --
drivers/gpu/drm/amd/powerplay/smumgr/smumgr.c | 87 ---
2 files changed, 118 deletions(-)
diff --git
repeated defining in hwmgr.h
use PHM_WAIT_INDIRECT_FIELD instand.
Change-Id: I4c06638c4fb2af2bae18a0c0f4f825b555f2bea3
Signed-off-by: Rex Zhu
---
drivers/gpu/drm/amd/powerplay/inc/smumgr.h| 11 ---
drivers/gpu/drm/amd/powerplay/smumgr/ci_smc.c | 2
repeated defining in hwmgr.h
Signed-off-by: Rex Zhu
Change-Id: I62316f35417784c4cf30a75e148bd1895a184a5b
---
drivers/gpu/drm/amd/powerplay/inc/smumgr.h | 5 -
drivers/gpu/drm/amd/powerplay/smumgr/cz_smumgr.c | 12 ++--
2 files changed, 6 insertions(+), 11
repeated defining in hwmgr.h
Change-Id: I03b1adfb164d43e9f33e6442f666c013f2e98a05
Signed-off-by: Rex Zhu
---
drivers/gpu/drm/amd/powerplay/inc/smumgr.h | 6 --
drivers/gpu/drm/amd/powerplay/smumgr/fiji_smumgr.c | 14 +++---
the macro is as same as PHM_WRITE_INDIRECT_FIELD
Change-Id: I62389ad2d699530e3a7d6a81f7427c21fd88b7be
Signed-off-by: Rex Zhu
---
drivers/gpu/drm/amd/powerplay/inc/smumgr.h| 15 ---
drivers/gpu/drm/amd/powerplay/smumgr/ci_smc.c | 4 ++--
Am 20.09.2017 um 13:42 schrieb Tom St Denis:
We discovered that on some devices even with iommu enabled
you can access all of system memory through the iommu translation.
Therefore, we revert the read method to the translation only service
and drop the write method completely.
Signed-off-by:
the macro is as same as PHM_WRITE_FIELD
Signed-off-by: Rex Zhu
Change-Id: I678e6277b6d582ab6a651a5fba07989634bf1aee
---
drivers/gpu/drm/amd/powerplay/inc/smumgr.h| 3 ---
drivers/gpu/drm/amd/powerplay/smumgr/ci_smc.c | 6 +++---
the SMUM_WAIT_VFPF_INDIRECT_FIELD_UNEQUAL
is irrelated to SMU, so move to hwmgr.h
and rename to
PHM_WAIT_VFPF_INDIRECT_FIELD_UNEQUAL
Signed-off-by: Rex Zhu
Change-Id: I2036309e8dd4cc3b211ae06a3d0ddfd9f16a92e7
---
drivers/gpu/drm/amd/powerplay/inc/hwmgr.h | 13
the SMUM_WAIT_INDIRECT_FIELD_UNEQUAL
is irrelated to SMU, so move to hwmgr.h
and rename to PHM_WAIT_INDIRECT_FIELD_UNEQUAL
Signed-off-by: Rex Zhu
Change-Id: Ieeb89b022edc9f3897c39641dd00e7409aa0faf9
---
drivers/gpu/drm/amd/powerplay/inc/hwmgr.h | 13 +
Change-Id: I81163535116224db66b20d0afe33909aa1d4bf11
Signed-off-by: Rex Zhu
---
drivers/gpu/drm/amd/powerplay/inc/smumgr.h| 4 ++--
drivers/gpu/drm/amd/powerplay/smumgr/ci_smc.c | 2 +-
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git
Change-Id: I37f87d31c8467b1c8b9a2bfa5e40083634f19fdb
Signed-off-by: Rex Zhu
---
drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c| 42 --
drivers/gpu/drm/amd/powerplay/inc/hwmgr.h | 10 --
drivers/gpu/drm/amd/powerplay/smumgr/rv_smumgr.c | 2
Change-Id: I82f9d6e6d59609d47124e868c53e1cc7717c45d1
Signed-off-by: Rex Zhu
---
drivers/gpu/drm/amd/powerplay/inc/hwmgr.h | 11 ---
1 file changed, 11 deletions(-)
diff --git a/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h
b/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h
index
Am 20.09.2017 um 11:31 schrieb Monk Liu:
fix missing finish uvd enc_ring.
v2:
since the adev pointer check in already in ring_fini
so drop the check outsider
Change-Id: Ib74237ca5adcb3b128c9b751fced0b7db7b09e86
Signed-off-by: Monk Liu
Reviewed-by: Christian König
the smumgr layer is redundant in powerplay.
so delete struct smumgr, move smu callback functions and
backend data to hwmgr.
the macros SMUM_* in smumgr.h is functionally repeated
with macros PHM_* in hwmgr.h, and the macros is irrelated
to smu. so delete the macros in smumgr.h
Rex Zhu (18):
This reverts commit 0e1c378fce66db833e08770d888dda1c5ec7936a.
---
src/lib/CMakeLists.txt | 1 +
src/lib/close_asic.c | 2 +-
src/lib/discover.c | 3 -
src/lib/free_maps.c| 44 ++
src/lib/read_vram.c| 218 -
src/umr.h
We discovered that on some devices even with iommu enabled
you can access all of system memory through the iommu translation.
Therefore, we revert the read method to the translation only service
and drop the write method completely.
Signed-off-by: Tom St Denis
---
On 20/09/17 03:07 AM, Christian König wrote:
Am 19.09.2017 um 23:38 schrieb Tom St Denis:
On 19/09/17 02:33 PM, Christian König wrote:
[root@carrizo ~]# xxd -l 1024 -s 0xC
/sys/kernel/debug/dri/0/amdgpu_iova
Actually 0xC is a special address, e.g. video BIOS if I'm not
completely
fix missing finish uvd enc_ring.
v2:
since the adev pointer check in already in ring_fini
so drop the check outsider
Change-Id: Ib74237ca5adcb3b128c9b751fced0b7db7b09e86
Signed-off-by: Monk Liu
---
drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c | 4
1 file changed, 4
> -Original Message-
> From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
> Of Eric Huang
> Sent: Tuesday, September 19, 2017 2:07 PM
> To: amd-gfx@lists.freedesktop.org
> Cc: Huang, JinHuiEric
> Subject: [PATCH 0/4] some changes for thermal interrupts
>
> Eric Huang
Am 19.09.2017 um 23:38 schrieb Tom St Denis:
On 19/09/17 02:33 PM, Christian König wrote:
[root@carrizo ~]# xxd -l 1024 -s 0xC
/sys/kernel/debug/dri/0/amdgpu_iova
Actually 0xC is a special address, e.g. video BIOS if I'm not
completely mistaken.
Not sure why that would be mapped
On 09/20/2017 02:22 PM, Evan Quan wrote:
Change-Id: Ia41bf64501557723fa811ad98a7b5630f12d9ed8
Signed-off-by: Evan Quan
Reviewed-by: Junwei Zhang
---
drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff
No ,we cannot do this in SW init, because PSP need all other component finish
their HW init and get the fw_size, before it can call ucode_init, so if we put
this in SW init, the fw_size is still 0
BR Monk
-Original Message-
From: Christian König
Change-Id: I96cd1d463a5743f918a03cad5160ea0bbd908ad0
Signed-off-by: Evan Quan
---
drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
index
44 matches
Mail list logo