mirrors/linux
2
0
Fork 0
mirror of https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git synced 2026-08-31 05:49:47 -04:00
Code Issues Packages Projects Releases Wiki Activity

drm/amdkfd: fix SDMA queue counter read on non-gfx9.4.3 ASICs

Browse Source 
The SDMA queue counter read was dispatched by GC version: anything newer
than gfx 9.4.2 was routed to the kfd2kgd->hqd_sdma_get_counter hook.
However that hook is only implemented for gfx 9.4.3, so gfx 10.3, gfx 11
and gfx 12 fell into the else branch with a NULL hook and got
-EOPNOTSUPP. This spammed "Failed to read SDMA queue counter" on every
SDMA queue teardown and left sdma_val at 0, so the per-process SDMA
activity accounting stopped working on those ASICs.

Dispatch based on whether the hook is implemented instead of the GC
version, so ASICs without the hook keep using read_sdma_queue_counter()
as before.

Fixes: 8f09c0ec21 ("drm/amdkfd: add sdma queue counter for gfxv9.4.3")
Reviewed-by: Eric Huang <jinhuieric.huang@amd.com>
Signed-off-by: Jesse Zhang <Jesse.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This commit is contained in:
Jesse Zhang
2026-06-15 18:58:09 +08:00
committed by Alex Deucher
parent 6cfa412680
commit 296ebc46de
2 changed files with 17 additions and 23 deletions
Download Patch File Download Diff File Expand all files Collapse all files

28
drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
View File

@@ -1027,17 +1027,15 @@ static int destroy_queue_nocpsch(struct device_queue_manager *dqm,
/* Get the SDMA queue stats */
if ((q->properties.type == KFD_QUEUE_TYPE_SDMA) ||
(q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI)) {
if (KFD_GC_VERSION(dqm->dev) <= IP_VERSION(9, 4, 2))
if (dqm->dev->kfd2kgd->hqd_sdma_get_counter)
retval = dqm->dev->kfd2kgd->hqd_sdma_get_counter(
dqm->dev->adev, q->mqd,
dqm->dev->kfd->device_info.num_sdma_queues_per_engine,
&sdma_val);
else
retval = read_sdma_queue_counter(
(uint64_t __user *)q->properties.read_ptr,
&sdma_val);
else
retval = dqm->dev->kfd2kgd->hqd_sdma_get_counter ?
dqm->dev->kfd2kgd->hqd_sdma_get_counter(
dqm->dev->adev, q->mqd,
dqm->dev->kfd->device_info.num_sdma_queues_per_engine,
&sdma_val) :
-EOPNOTSUPP;
if (retval)
dev_err(dev, "Failed to read SDMA queue counter for queue: %d\n",
q->properties.queue_id);
@@ -2675,17 +2673,15 @@ static int destroy_queue_cpsch(struct device_queue_manager *dqm,
/* Get the SDMA queue stats */
if ((q->properties.type == KFD_QUEUE_TYPE_SDMA) ||
(q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI)) {
if (KFD_GC_VERSION(dqm->dev) <= IP_VERSION(9, 4, 2))
if (dqm->dev->kfd2kgd->hqd_sdma_get_counter)
retval = dqm->dev->kfd2kgd->hqd_sdma_get_counter(
dqm->dev->adev, q->mqd,
dqm->dev->kfd->device_info.num_sdma_queues_per_engine,
&sdma_val);
else
retval = read_sdma_queue_counter(
(uint64_t __user *)q->properties.read_ptr,
&sdma_val);
else
retval = dqm->dev->kfd2kgd->hqd_sdma_get_counter ?
dqm->dev->kfd2kgd->hqd_sdma_get_counter(
dqm->dev->adev, q->mqd,
dqm->dev->kfd->device_info.num_sdma_queues_per_engine,
&sdma_val) :
-EOPNOTSUPP;
if (retval)
dev_err(dev, "Failed to read SDMA queue counter for queue: %d\n",

12
drivers/gpu/drm/amd/amdkfd/kfd_process.c
View File

@@ -192,15 +192,13 @@ static void kfd_sdma_activity_worker(struct work_struct *work)
list_for_each_entry(sdma_q, &sdma_q_list.list, list) {
val = 0;
if (KFD_GC_VERSION(dqm->dev) <= IP_VERSION(9, 4, 2))
ret = read_sdma_queue_counter(sdma_q->rptr, &val);
else
ret = dqm->dev->kfd2kgd->hqd_sdma_get_counter ?
dqm->dev->kfd2kgd->hqd_sdma_get_counter(
if (dqm->dev->kfd2kgd->hqd_sdma_get_counter)
ret = dqm->dev->kfd2kgd->hqd_sdma_get_counter(
dqm->dev->adev, sdma_q->mqd,
dqm->dev->kfd->device_info.num_sdma_queues_per_engine,
&val) :
-EOPNOTSUPP;
&val);
else
ret = read_sdma_queue_counter(sdma_q->rptr, &val);
if (ret) {
pr_debug("Failed to read SDMA queue active counter for queue id: %d",