Re:Re: [PATCH V3] amdgpu:optimization-- reduce no need mutex_lock area
Sure, this seems to be a lot more professional than my previous modification.
My original intention is to make the code easier to read, and I learned a lot
from
submitting these patches. Thank you very much for all your guidance!
Regards,
Bernard
发件人：Felix Kuehling
发送日期：2020-04-22 10:27:16
收件人：1587181464-114215-1-git-send-email-bern...@vivo.com,Alex Deucher
,"Christian König" ,"David
(ChunMing) Zhou" ,David Airlie ,Daniel
Vetter
,amd-gfx@lists.freedesktop.org,dri-de...@lists.freedesktop.org,linux-ker...@vger.kernel.org
抄送人：opensource.ker...@vivo.com,Bernard Zhao
主题：Re: [PATCH V3] amdgpu:optimization-- reduce no need mutex_lock area
Thanks again for the patch. I'm going
to apply this with some minor fixes. The headline should start
with "drm/amdgpu:". I'll also change the wording of the headline
and commit message:
drm/amdgpu: shrink critical section
in amdgpu_amdkfd_gpuvm_free_memory_of_gpu
Reduce the mem->lock`s protected
code area, no need to protect pr_debug.
This also simplifies error handling.
There is one more cosmetic change I'm
going to make, see inline. I'll apply your patch with those
updates if you're OK with that.
On 2020-04-21 2:48, Bernard Zhao wrote:
Maybe we could reduce the mutex_lock(>lock)`s protected code area,
and no need to protect pr_debug.
Signed-off-by: Bernard Zhao
Changes since V1:
*commit message improve
Changes since V2:
*move comment along with the mutex_unlock
Link for V1:
*https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Flore.kernel.org%2Fpatchwork%2Fpatch%2F1226588%2Fdata=02%7C01%7CFelix.Kuehling%40amd.com%7Cc715f5d02aae40846b8f08d7e5c001fd%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637230485152789682sdata=4UdZiWMAbW8eR1BS2%2F6qMvs5K6cHWy5c8I32ReQ4uz0%3Dreserved=0
Link for V2:
*https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Flore.kernel.org%2Fpatchwork%2Fpatch%2F1227907%2Fdata=02%7C01%7CFelix.Kuehling%40amd.com%7Cc715f5d02aae40846b8f08d7e5c001fd%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637230485152799673sdata=Wt%2Bk7k4MtSX4zIDgmLEOB6ZKzfuqAd6GJZ3Creqf1aY%3Dreserved=0
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 14 +++---
1 file changed, 7 insertions(+), 7 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index 327317c54f7c..f03d9843d723 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -1285,21 +1285,21 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
struct bo_vm_reservation_context ctx;
struct ttm_validate_buffer *bo_list_entry;
int ret;
+ unsigned int mapped_to_gpu_memory;
I'll move this before the "int ret;" line. It makes the code more
readable if long declarations come before short ones.
Regards,
Felix
mutex_lock(>lock);
+ mapped_to_gpu_memory = mem->mapped_to_gpu_memory;
+ mutex_unlock(>lock);
+ /* lock is not needed after this, since mem is unused and will
+* be freed anyway
+*/
- if (mem->mapped_to_gpu_memory > 0) {
+ if (mapped_to_gpu_memory > 0) {
pr_debug("BO VA 0x%llx size 0x%lx is still mapped.\n",
mem->va, bo_size);
- mutex_unlock(>lock);
return -EBUSY;
}
- mutex_unlock(>lock);
- /* lock is not needed after this, since mem is unused and will
-* be freed anyway
-*/
-
/* No more MMU notifiers */
amdgpu_mn_unregister(mem->bo);
___
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
Re: [PATCH V3] amdgpu:optimization-- reduce no need mutex_lock area
Thanks again for the patch. I'm going to apply this with some minor
fixes. The headline should start with "drm/amdgpu:". I'll also change
the wording of the headline and commit message:
drm/amdgpu: shrink critical section in
amdgpu_amdkfd_gpuvm_free_memory_of_gpu
Reduce the mem->lock`s protected code area, no need to protect pr_debug.
This also simplifies error handling.
There is one more cosmetic change I'm going to make, see inline. I'll
apply your patch with those updates if you're OK with that.
On 2020-04-21 2:48, Bernard Zhao wrote:
Maybe we could reduce the mutex_lock(>lock)`s protected code area,
and no need to protect pr_debug.
Signed-off-by: Bernard Zhao
Changes since V1:
*commit message improve
Changes since V2:
*move comment along with the mutex_unlock
Link for V1:
*https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Flore.kernel.org%2Fpatchwork%2Fpatch%2F1226588%2Fdata=02%7C01%7CFelix.Kuehling%40amd.com%7Cc715f5d02aae40846b8f08d7e5c001fd%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637230485152789682sdata=4UdZiWMAbW8eR1BS2%2F6qMvs5K6cHWy5c8I32ReQ4uz0%3Dreserved=0
Link for V2:
*https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Flore.kernel.org%2Fpatchwork%2Fpatch%2F1227907%2Fdata=02%7C01%7CFelix.Kuehling%40amd.com%7Cc715f5d02aae40846b8f08d7e5c001fd%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637230485152799673sdata=Wt%2Bk7k4MtSX4zIDgmLEOB6ZKzfuqAd6GJZ3Creqf1aY%3Dreserved=0
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 14 +++---
1 file changed, 7 insertions(+), 7 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index 327317c54f7c..f03d9843d723 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -1285,21 +1285,21 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
struct bo_vm_reservation_context ctx;
struct ttm_validate_buffer *bo_list_entry;
int ret;
+ unsigned int mapped_to_gpu_memory;
I'll move this before the "int ret;" line. It makes the code more
readable if long declarations come before short ones.
Regards,
Felix
mutex_lock(>lock);
+ mapped_to_gpu_memory = mem->mapped_to_gpu_memory;
+ mutex_unlock(>lock);
+ /* lock is not needed after this, since mem is unused and will
+* be freed anyway
+*/
- if (mem->mapped_to_gpu_memory > 0) {
+ if (mapped_to_gpu_memory > 0) {
pr_debug("BO VA 0x%llx size 0x%lx is still mapped.\n",
mem->va, bo_size);
- mutex_unlock(>lock);
return -EBUSY;
}
- mutex_unlock(>lock);
- /* lock is not needed after this, since mem is unused and will
-* be freed anyway
-*/
-
/* No more MMU notifiers */
amdgpu_mn_unregister(mem->bo);
___
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
[PATCH V3] amdgpu:optimization-- reduce no need mutex_lock area
Maybe we could reduce the mutex_lock(>lock)`s protected code area,
and no need to protect pr_debug.
Signed-off-by: Bernard Zhao
Changes since V1:
*commit message improve
Changes since V2:
*move comment along with the mutex_unlock
Link for V1:
*https://lore.kernel.org/patchwork/patch/1226588/
Link for V2:
*https://lore.kernel.org/patchwork/patch/1227907/
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 14 +++---
1 file changed, 7 insertions(+), 7 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index 327317c54f7c..f03d9843d723 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -1285,21 +1285,21 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
struct bo_vm_reservation_context ctx;
struct ttm_validate_buffer *bo_list_entry;
int ret;
+ unsigned int mapped_to_gpu_memory;
mutex_lock(>lock);
+ mapped_to_gpu_memory = mem->mapped_to_gpu_memory;
+ mutex_unlock(>lock);
+ /* lock is not needed after this, since mem is unused and will
+* be freed anyway
+*/
- if (mem->mapped_to_gpu_memory > 0) {
+ if (mapped_to_gpu_memory > 0) {
pr_debug("BO VA 0x%llx size 0x%lx is still mapped.\n",
mem->va, bo_size);
- mutex_unlock(>lock);
return -EBUSY;
}
- mutex_unlock(>lock);
- /* lock is not needed after this, since mem is unused and will
-* be freed anyway
-*/
-
/* No more MMU notifiers */
amdgpu_mn_unregister(mem->bo);
--
2.26.2
___
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
