The adev can be found from bo by amdgpu_ttm_adev(bo->tbo.bdev),
and adev is also not used in the function
amdgpu_amdkfd_map_gtt_bo_to_gart().
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h | 2 +-
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 3 +--
driv
TLB flush after unmap accidentially was removed on
gfx9.4.2. It is to add it back.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdkfd/kfd_priv.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
b/drivers/gpu/drm/amd/amdkfd
amdgpu_job_ring may return NULL, which causes kernel NULL
pointer error, using another way to print ring name instead
of ring->name.
Suggested-by: Lijo Lazar
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c | 6 ++
1 file changed, 2 insertions(+), 4 deleti
reset cause is requested by customer as additional
info for gpu reset smi event.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c| 34 +
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h| 17 +++--
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c| 9 ++-
drivers
-
From: amd-gfx On Behalf Of Eric Huang
Sent: Friday, May 31, 2024 8:38 PM
To: amd-gfx@lists.freedesktop.org
Cc: Kasiviswanathan, Harish ; Huang, JinHuiEric
Subject: [PATCH] drm/amdkfd: add reset cause in gpu pre-reset smi event
reset cause is requested by customer as additional info for gpu
reset source or reset cause is very useful info
for reset context, it will be used by events API.
Suggested-by: Lijo Lazar
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_reset.c | 34 +++
drivers/gpu/drm/amd/amdgpu/amdgpu_reset.h | 13 +
2 files
reset cause is requested by customer as additional
info for gpu reset smi event.
v2: integerate reset sources suggested by Lijo Lazar
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c | 3 +++
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h | 10 +++---
drivers/gpu/drm
Thanks for your review Lijo, I will send a patch with reset source in
another places.
Regards,
Eric
On 2024-06-04 03:26, Lazar, Lijo wrote:
On 6/3/2024 11:42 PM, Eric Huang wrote:
reset cause is requested by customer as additional
info for gpu reset smi event.
v2: integerate reset sources
To fullfill the reset event description.
Suggested-by: Lijo Lazar
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c | 1 +
drivers/gpu/drm/amd/amdgpu/amdgpu_job.c | 1 +
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 1 +
3 files changed, 3 insertions(+)
diff --git a
On 2023-11-17 00:20, David Yat Sin wrote:
Fixes issue where user events of type KFD_EVENT_TYPE_HW_EXCEPTION do not
have valid data
Signed-off-by: David Yat Sin
---
drivers/gpu/drm/amd/amdkfd/kfd_events.c | 4
1 file changed, 4 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_e
MES on process termination.
Note that the flush call and the MES debugger calls use the same MES
interface but are separated as KFD calls to avoid conflicting with each
other.
Signed-off-by: Jonathan Kim
Tested-by: Alice Wong
Reviewed-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
The field adev->mes.funcs is NULL in function amdgpu_mes_flush_shader_debugger
on non-mes asics, add mes enabling check for call this func to
resolve the error.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c | 3 ++-
1 file changed, 2 insertions(+)
Signed-off-by: Jonathan Kim
Reviewed-by: Eric Huang
---
drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manag
gpus can't be evicted.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdkfd/kfd_device.c | 2 +-
drivers/gpu/drm/amd/amdkfd/kfd_process.c | 24 +---
2 files changed, 2 insertions(+), 24 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c
b/drivers/g
On 2021-07-30 5:26 p.m., Felix Kuehling wrote:
Am 2021-07-28 um 1:31 p.m. schrieb Eric Huang:
It is to fix a bug of gpu_recovery on multiple GPUs,
When one gpu is reset, the application running on other
gpu hangs, because kfd post reset doesn't restore the
running process.
This will r
It is to differenciate case scenario for proper behavior when
calling evict queues, such as GPU reset doesn't need to roll
back restoring partial evicted queues.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c | 2 +-
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
Due to changes of HW memory model, we need to change Aldebaran
MTYPEs to meet HW changes.
Signed-off-by: Eric Huang
Reviewed-by: Harish Kasiviswanathan
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 13 +
1 file changed, 5 insertions(+), 8 deletions(-)
diff --git a/drivers
From: Philip Yang
Due to changes of HW memory model, we need to change Aldebaran MTYPEs to
meet HW changes.
Change-Id: I0b566ccd1d406abe93ee84da91216794fc650ed4
Signed-off-by: Philip Yang
Reviewed-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 11 ---
1 file changed, 4
In NPS4 BIOS we need to find the closest numa node when creating
topology io link between cpu and gpu, if PCI driver doesn't set
it.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdkfd/kfd_crat.c | 94 ++-
1 file changed, 91 insertions(+), 3 deletions(-)
diff
ping...
On 2021-04-28 11:11 a.m., Eric Huang wrote:
In NPS4 BIOS we need to find the closest numa node when creating
topology io link between cpu and gpu, if PCI driver doesn't set
it.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdkfd/kfd_crat.c | 94 ++
Thanks Felix for your review. I will send another patch.
Eric
On 2021-04-30 7:42 p.m., Felix Kuehling wrote:
Am 2021-04-28 um 11:11 a.m. schrieb Eric Huang:
In NPS4 BIOS we need to find the closest numa node when creating
topology io link between cpu and gpu, if PCI driver doesn't s
In NPS4 BIOS we need to find the closest numa node when creating
topology io link between cpu and gpu, if PCI driver doesn't set
it.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdkfd/kfd_crat.c | 95 ++-
1 file changed, 93 insertions(+), 2 deletions(-)
diff
acpi_table_parse_srat) and the acpi
subsystem should expose APIs for rest drivers to query such numa information.
Regards,
Oak
On 2021-04-28, 11:12 AM, "amd-gfx on behalf of Eric Huang"
wrote:
In NPS4 BIOS we need to find the closest numa node when creating
topology io link b
On 2021-05-03 3:13 p.m., Felix Kuehling wrote:
Am 2021-05-03 um 10:47 a.m. schrieb Eric Huang:
In NPS4 BIOS we need to find the closest numa node when creating
topology io link between cpu and gpu, if PCI driver doesn't set
it.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/a
also will avoid parsing ACPI_SRAT_TYPE_CPU_AFFINITY structs.
Thanks,
Lijo
-Original Message-
From: amd-gfx On Behalf Of Eric Huang
Sent: Wednesday, April 28, 2021 8:42 PM
To: amd-gfx@lists.freedesktop.org
Cc: Huang, JinHuiEric
Subject: [PATCH] drm/amdkfd: add ACPI SRAT parsing for to
so will avoid parsing ACPI_SRAT_TYPE_CPU_AFFINITY structs.
>
> Thanks,
> Lijo
>
>
> -Original Message-
> From: amd-gfx On Behalf Of
Eric Huang
> Sent: Wednesday, April 28, 2021 8:42 PM
> To: amd-gfx@lists.freedesktop.org
> Cc: Huang, JinHuiEric
> Subject:
On 2021-05-03 3:34 p.m., Felix Kuehling wrote:
Am 2021-05-03 um 3:27 p.m. schrieb Eric Huang:
On 2021-05-03 3:13 p.m., Felix Kuehling wrote:
Am 2021-05-03 um 10:47 a.m. schrieb Eric Huang:
In NPS4 BIOS we need to find the closest numa node when creating
topology io link between cpu and gpu
In NPS4 BIOS we need to find the closest numa node when creating
topology io link between cpu and gpu, if PCI driver doesn't set
it.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdkfd/kfd_crat.c | 91 +++
1 file changed, 91 insertions(+)
diff --git a/driver
SDMA FW fixes the hang issue for adding heavy-weight TLB
flush on Arcturus, so we can enable it.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 8 +---
drivers/gpu/drm/amd/amdkfd/kfd_chardev.c | 3 ++-
2 files changed, 7 insertions(+), 4 deletions
SDMA FW fixes the hang issue for adding heavy-weight TLB
flush on Arcturus, so we can enable it.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 8 +---
drivers/gpu/drm/amd/amdkfd/kfd_chardev.c | 3 ++-
2 files changed, 7 insertions(+), 4 deletions
The SDMA fix is generic and not in a specific version of FW, so we don't
have to check.
Thanks,
Eric
On 2022-01-18 11:35, Alex Deucher wrote:
On Tue, Jan 18, 2022 at 11:16 AM Eric Huang wrote:
SDMA FW fixes the hang issue for adding heavy-weight TLB
flush on Arcturus, so we can enab
e Only]
I think what he means is that if we are using SDMA v17, this will cause issues,
won't it? Should we check that SDMA version is >=18 before enabling it? Or am I
misunderstanding the fix?
Yes, that was my concern.
Alex
Kent
-Original Message-
From: amd-gfx On Behalf
SDMA FW fixes the hang issue for adding heavy-weight TLB
flush on Arcturus, so we can enable it.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 9 ++---
drivers/gpu/drm/amd/amdkfd/kfd_chardev.c | 4 +++-
2 files changed, 9 insertions(+), 4 deletions
SDMA FW fixes the hang issue for adding heavy-weight TLB
flush on Arcturus, so we can enable it.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 6 --
drivers/gpu/drm/amd/amdkfd/kfd_chardev.c | 10 --
2 files changed, 8 insertions(+), 8
*From:* amd-gfx on behalf of
Felix Kuehling
*Sent:* Tuesday, January 18, 2022 6:36 PM
*To:* Huang, JinHuiEric ;
amd-gfx@lists.freedesktop.org
*Subject:* Re: [PATCH] drm/amdkfd: enable heavy-weight TLB flush on
Arcturus
Am 2022-01-18 um 5:45 p.m. schrieb Eric Huang:
SDMA FW fixes the
It is to meet the requirement for memory allocation
optimization on MI50.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdkfd/kfd_chardev.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
b/drivers/gpu/drm/amd/amdkfd
] drm/amdkfd: enable heavy-weight TLB flush on Vega20
On Fri, Jan 21, 2022 at 11:17 AM Eric Huang wrote:
It is to meet the requirement for memory allocation optimization on
MI50.
Signed-off-by: Eric Huang
Assuming there is no firmware version requirement, the patch is:
Acked-by: Alex De
It it to optimize memory allocation latency.
Signed-off-by: Eric Huang
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
index 960913a35ee4..ab73741edb97 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
+++ b/drivers/gpu/drm/amd/amdkfd
With XGMI connection flushing HDP on PCIe is unnecessary,
it is also to optimize memory allocation latency.
Signed-off-by: Eric Huang
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_cpu.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_cpu.c
index ac45d9c7a4e9..aefb3d2927d5 100644
--- a/drivers/gpu
tee it done before
memory is remapped? if remapping depends on it, then more latency will
be introduced in map.
Regards,
Eric
Regards,
Felix
Am 2021-05-25 um 1:53 p.m. schrieb Eric Huang:
It it to optimize memory allocation latency.
Signed-off-by: Eric Huang
diff --git a/drivers/g
On 2021-05-28 11:23 a.m., Christian König wrote:
Am 27.05.21 um 16:05 schrieb philip yang:
On 2021-05-26 5:25 p.m., Felix Kuehling wrote:
Am 2021-05-26 um 3:21 p.m. schrieb Eric Huang:
On 2021-05-25 3:16 p.m., Felix Kuehling wrote:
Similar to a recent fix by Philip Yang 76e08b37d0aa
1. conditionally flush TLBs after map.
2. add heavy weight TLBs flushing after unmap.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h| 1 +
.../gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 2 ++
drivers/gpu/drm/amd/amdgpu/amdgpu_object.h| 1 +
drivers/gpu/drm
it is for large bar/xgmi which has cpu update function,
for small bar that has sdma update function it will be done
later.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h | 1 +
drivers/gpu/drm/amd/amdgpu/amdgpu_vm_cpu.c | 21 -
drivers/gpu/drm
table_freed will be always true when mapping a memory with size
bigger than 2MB. Using a check of turnning valid PDE into PTE
will resolve the issue.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 16 +++-
1 file changed, 15 insertions(+), 1 deletion(-)
diff
Please ignore those patches, I find a better resolution and will send it
out later.
Regards,
Eric
On 2021-05-28 4:49 p.m., Eric Huang wrote:
1. conditionally flush TLBs after map.
2. add heavy weight TLBs flushing after unmap.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu
1. conditionally flush TLBs after map.
2. add heavy weight TLBs flushing after unmap.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h| 1 +
.../gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 2 ++
drivers/gpu/drm/amd/amdgpu/amdgpu_object.h| 1 +
drivers/gpu/drm
table_freed will be always true when mapping a memory with size
bigger than 2MB. The problem is page table's entries are always
existed, but existing mapping depends on page talbe's bo, so
using a check of page table's bo existed will resolve the issue.
Signed-off-by: Eric Huang
-
On 2021-05-30 12:54 p.m., Christian König wrote:
Am 30.05.21 um 00:51 schrieb Eric Huang:
table_freed will be always true when mapping a memory with size
bigger than 2MB. The problem is page table's entries are always
existed, but existing mapping depends on page talbe's bo, so
usi
On 2021-05-31 10:08 a.m., Christian König wrote:
Am 30.05.21 um 20:29 schrieb Eric Huang:
On 2021-05-30 12:54 p.m., Christian König wrote:
Am 30.05.21 um 00:51 schrieb Eric Huang:
table_freed will be always true when mapping a memory with size
bigger than 2MB. The problem is page table&#
With XGMI connection flushing HDP on PCIe is unnecessary,
it is also to optimize memory allocation latency.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.h | 1 +
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 1 +
drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c | 3 +++
3 files
table_freed will be always true when mapping a memory with size
bigger than 2MB. The problem is page table's entries are always
existed, but existing mapping depends on page talbe's bo, so
using a check of page table's bo existed will resolve the issue.
Signed-off-by: Eric Huang
-
With XGMI connection flushing HDP on PCIe is unnecessary,
it is also to optimize memory allocation latency.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.h | 1 +
drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c | 3 ++-
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 1 +
drivers/gpu
On 2021-06-01 3:05 a.m., Christian König wrote:
Am 01.06.21 um 02:06 schrieb Eric Huang:
With XGMI connection flushing HDP on PCIe is unnecessary,
it is also to optimize memory allocation latency.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_hdp.h | 1 +
drivers/gpu/drm
1. conditionally flush TLBs after map.
2. add heavy weight TLBs flushing after unmap.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h| 2 +-
.../gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 21 ++-
drivers/gpu/drm/amd/amdgpu/amdgpu_object.h| 1
1. conditionally flush TLBs after map.
2. add heavy weight TLBs flush after unmap.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h| 2 +-
.../gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 19 +++--
drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c| 6
It is a part of memory mapping optimization.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdkfd/kfd_chardev.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
index 4da8aff3df27..98f1d2b586c5 100644
It is to provide more tlb flush types opotion for different
case scenario.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdkfd/kfd_chardev.c | 2 +-
drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c | 6 +++---
drivers/gpu/drm/amd/amdkfd/kfd_priv.h | 2
It is to pass the flag to KFD, and optimize table_freed in
amdgpu_vm_bo_update_mapping.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 6 +++---
drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c | 2 +-
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 10 +-
drivers/gpu/drm
It is to optimize memory mapping latency, and to aviod
a page fault in a corner case of changing valid PDE into
PTE.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h| 2 +-
.../gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 19 --
drivers/gpu/drm/amd/amdkfd
Integrate two macros into two generic functions and add
no_flush flag to determine if HDP flush is needed for
all Asics.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h| 9 ---
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 29 ++
drivers/gpu/drm
On 2021-06-02 3:00 a.m., Christian König wrote:
Am 02.06.21 um 04:04 schrieb Eric Huang:
Integrate two macros into two generic functions and add
no_flush flag to determine if HDP flush is needed for
all Asics.
Yes that starts looks like it should work, just a few comments below.
Signed
Integrate two generic functions to determine if HDP
flush is needed for all Asics.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h| 5
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c | 2 +-
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 33
On 2021-06-04 7:31 a.m., Christian König wrote:
Am 02.06.21 um 21:18 schrieb Eric Huang:
Integrate two generic functions to determine if HDP
flush is needed for all Asics.
Signed-off-by: Eric Huang
Nice work, just one more idea below.
But patch is Reviewed-by: Christian König
either
Add the parameter table_freed description on function description.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index b4f189ab672e
It is to fix a failure for SDMA updating PTEs.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdkfd/kfd_chardev.c | 26 +---
1 file changed, 23 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
b/drivers/gpu/drm/amd/amdkfd
Fix non-coherent bit of iolink properties flag
which always is 0.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdkfd/kfd_topology.c | 23 +++
1 file changed, 23 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
b/drivers/gpu/drm/amd/amdkfd
Fix non-coherent bit of p2plink properties flag
which always is 0.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdkfd/kfd_topology.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
index
It is based on reverting two patches back.
drm/amdkfd: Make TLB flush conditional on mapping
drm/amdgpu: Add table_freed parameter to amdgpu_vm_bo_update
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdkfd/kfd_chardev.c | 6 ++
1 file changed, 6 insertions(+)
diff --git a/drivers
It is based on reverting two patches back.
drm/amdkfd: Make TLB flush conditional on mapping
drm/amdgpu: Add table_freed parameter to amdgpu_vm_bo_update
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 6 ++
1 file changed, 6 insertions(+)
diff --git a
It is to workaround HW bug on other Asics and based on
reverting two commits:
drm/amdkfd: Add heavy-weight TLB flush after unmapping
drm/amdkfd: Add memory sync before TLB flush on unmap
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdkfd/kfd_chardev.c | 37 +---
1
I am converting codes into amd-staging-drm-next. Theoretically it will
improve a lot on the latency, the size of the array allocated is 24
(PAGE_SIZE/struct drm_mm_node) with this patch, and it was 8192 before.
So the latency should be reduced by 98 us.
Regards,
Eric
On 2021-07-13 12:11 p.m.,
ff-by: Eric Huang
Reviewed-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c | 71 +++-
1 file changed, 53 insertions(+), 18 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
index be4261c
Hi Christian/Felix,
If you don't have objection, it will be pushed into
amd-staging-dkms-5.11 and amd-staging-drm-next.
Thanks,
Eric
On 2021-07-13 3:17 p.m., Eric Huang wrote:
For allocations larger than 48MiB we need more than a page for the
housekeeping in the worst case resulting i
Aldebaran as well, so this workaround
is still invalid.
Regards,
Guchun
-Original Message-
From: amd-gfx On Behalf Of Eric Huang
Sent: Friday, July 9, 2021 3:54 AM
To: amd-gfx@lists.freedesktop.org
Cc: Huang, JinHuiEric ; Kuehling, Felix
Subject: [PATCH] drm/amdkfd: Only apply heavy-weight
Correction inline.
On 2021-07-14 11:22 a.m., Eric Huang wrote:
Hi Felix,
I was not able to reproduce the VM fault issue of SWDEV-292611(not
SWDEV-249241), which is the only regression reported on MI200. So the
patch is valid to review. Please take a look.
Thanks,
Eric
On 2021-07-09 1:45
gards,
Christian.
Am 13.07.21 um 21:19 schrieb Eric Huang:
Hi Christian/Felix,
If you don't have objection, it will be pushed into
amd-staging-dkms-5.11 and amd-staging-drm-next.
Thanks,
Eric
On 2021-07-13 3:17 p.m., Eric Huang wrote:
For allocations larger than 48MiB we need more than
Driver doesn't keep the default bootup power limit and expose it
to user. As requested, we add it in driver.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/pm/amdgpu_pm.c| 43 ++-
drivers/gpu/drm/amd/pm/inc/amdgpu_smu.h | 2 +
drivers/gpu/drm/amd/p
There is no difference betweeen max and default to get
power limit in powerplay, adding a new parameter in the
funciton fixes it.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/include/kgd_pp_interface.h | 3 ++-
drivers/gpu/drm/amd/pm/amdgpu_pm.c | 12
drivers
Please ignore those two patches. I found a mistake and will send new
ones for review.
Regards,
Eric
On 2021-03-12 4:03 p.m., Eric Huang wrote:
There is no difference betweeen max and default to get
power limit in powerplay, adding a new parameter in the
funciton fixes it.
Signed-off-by: Eric
Driver doesn't keep the default bootup power limit and expose it
to user. As requested we add it in driver.
Signed-off-by: Eric Huang
---
.../gpu/drm/amd/include/kgd_pp_interface.h| 3 +-
drivers/gpu/drm/amd/pm/amdgpu_pm.c| 53 +--
drivers/gpu/drm/amd/p
that do not support the
current exception handling and running KFD tests.
The only time ADD_QUEUE.skip_process_ctx_clear is required is for
debugger use cases where a debugged process is always runtime enabled
when adding a queue.
Tested-by: Shikai Guo
Signed-off-by: Jonathan Kim
Reviewed-by: Eric
On 2023-09-26 23:00, Xiaogang.Chen wrote:
From: Xiaogang Chen
prange->svm_bo unref can happen in both mmu callback and a callback after
migrate to system ram. Both are async call in different tasks. Sync svm_bo
unref operation to avoid random "use-after-free".
Signed-off-by: Xiaogang.Chen
-
On 2022-04-14 04:19, Lang Yu wrote:
The idea is from commit a50fe7078035 ("drm/amdkfd: Only apply heavy-weight
TLB flush on Aldebaran") and commit f61c40c0757a ("drm/amdkfd: enable
heavy-weight TLB flush on Arcturus"). Otherwise, we will run into problems
on some ASICs when running SVM applica
Two changes:
1. reducing unnecessary evict/unmap when range is not mapped to gpu.
2. adding always evict when flags is set to always_mapped.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 10 --
1 file changed, 8 insertions(+), 2 deletions(-)
diff --git a/drivers
It is to add new options for always keeping gpu mapping
and custom of coarse grain allocation intead of fine
grain as default.
Signed-off-by: Eric Huang
---
include/uapi/linux/kfd_ioctl.h | 4
1 file changed, 4 insertions(+)
diff --git a/include/uapi/linux/kfd_ioctl.h b/include/uapi/linux
It is to give a chance for user to change default
flags setting, such as fine grain to coarse grain.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
b/drivers/gpu/drm
http://gerrit-git.amd.com/c/compute/ec/libhsakmt/+/697296
Regards,
Eric
On 2022-06-27 11:33, Alex Deucher wrote:
On Fri, Jun 24, 2022 at 12:03 PM Eric Huang wrote:
It is to add new options for always keeping gpu mapping
and custom of coarse grain allocation intead of fine
grain as default
No. There is only internal link for now, because it is under review.
Once it is submitted, external link should be in gerritgit for libhsakmt.
Regards,
Eric
On 2022-06-27 11:58, Alex Deucher wrote:
On Mon, Jun 27, 2022 at 11:36 AM Eric Huang wrote:
https://nam11
Thank you, Felix.
I will send all libhsakmt changes and amdkfd changes to amd-gfx.
Regards,
Eric
On 2022-06-28 16:44, Felix Kuehling wrote:
Am 2022-06-27 um 12:01 schrieb Eric Huang:
No. There is only internal link for now, because it is under review.
Once it is submitted, external link
It is to add new option for always keeping gpu mapping.
Signed-off-by: Eric Huang
---
include/uapi/linux/kfd_ioctl.h | 2 ++
1 file changed, 2 insertions(+)
diff --git a/include/uapi/linux/kfd_ioctl.h b/include/uapi/linux/kfd_ioctl.h
index fd49dde4d5f4..eba04ebfd9a8 100644
--- a/include/uapi
amdkfd changes:
Eric Huang (2):
drm/amdkfd: add new flag for svm
drm/amdkfd: change svm range evict
drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 10 --
include/uapi/linux/kfd_ioctl.h | 2 ++
2 files changed, 10 insertions(+), 2 deletions(-)
libhsakmt(thunk) changes:
which are
It is to add new option for always keeping gpu mapping.
Signed-off-by: Eric Huang
Change-Id: Iebee35e6de4d52fa29f82dd19f6bbf5640249492
---
include/linux/kfd_ioctl.h | 2 ++
1 file changed, 2 insertions(+)
diff --git a/include/linux/kfd_ioctl.h b/include/linux/kfd_ioctl.h
index 8a0ed49..5c45f58
To improve performance on queue preemption, allocate ctx s/r
area in VRAM instead of system memory, and migrate it back
to system memory when VRAM is full.
Signed-off-by: Eric Huang
Change-Id: If775782027188dbe84b6868260e429373675434c
---
include/hsakmttypes.h | 1 +
src/queues.c
Two changes:
1. reducing unnecessary evict/unmap when range is not mapped to gpu.
2. adding always evict when flags is set to always_mapped.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 10 --
1 file changed, 8 insertions(+), 2 deletions(-)
diff --git a/drivers
To improve performance on queue preemption, allocate ctx s/r
area in VRAM instead of system memory, and migrate it back
to system memory when VRAM is full.
Signed-off-by: Eric Huang
Change-Id: If775782027188dbe84b6868260e429373675434c
---
include/hsakmttypes.h | 1 +
src/queues.c
It is to add new option for always keeping gpu mapping.
Signed-off-by: Eric Huang
Change-Id: Iebee35e6de4d52fa29f82dd19f6bbf5640249492
---
include/linux/kfd_ioctl.h | 2 ++
1 file changed, 2 insertions(+)
diff --git a/include/linux/kfd_ioctl.h b/include/linux/kfd_ioctl.h
index 8a0ed49..5c45f58
On 2022-06-29 18:20, Felix Kuehling wrote:
On 2022-06-28 17:43, Eric Huang wrote:
Two changes:
1. reducing unnecessary evict/unmap when range is not mapped to gpu.
2. adding always evict when flags is set to always_mapped.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdkfd
On 2022-06-29 19:29, Felix Kuehling wrote:
On 2022-06-29 18:53, Eric Huang wrote:
On 2022-06-29 18:20, Felix Kuehling wrote:
On 2022-06-28 17:43, Eric Huang wrote:
Two changes:
1. reducing unnecessary evict/unmap when range is not mapped to gpu.
2. adding always evict when flags is set to
amdkfd changes:
Eric Huang (3):
drm/amdkfd: add new flag for svm
drm/amdkfd: change svm range evict
drm/amdkfd: optimize svm range evict
drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 13 +++--
include/uapi/linux/kfd_ioctl.h | 2 ++
2 files changed, 13 insertions(+), 2 deletions
Adding always evict queues when flag is set to
KFD_IOCTL_SVM_FLAG_GPU_ALWAYS_MAPPED as if XNACK off.
Signed-off-by: Eric Huang
---
drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
b/drivers/gpu
1 - 100 of 230 matches
Mail list logo
