Commit 9e4a91cd authored by Alex Sierra's avatar Alex Sierra Committed by Alex Deucher

drm/amdkfd: inc counter on child ranges with xnack off

During GPU page table invalidation with xnack off, new ranges
split may occur concurrently in the same prange. Creating a new
child per split. Each child should also increment its
invalid counter, to assure GPU page table updates in these
ranges.
Signed-off-by: default avatarAlex Sierra <alex.sierra@amd.com>
Reviewed-by: default avatarFelix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: default avatarAlex Deucher <alexander.deucher@amd.com>
parent 1d40ef90
...@@ -1590,6 +1590,7 @@ svm_range_evict(struct svm_range *prange, struct mm_struct *mm, ...@@ -1590,6 +1590,7 @@ svm_range_evict(struct svm_range *prange, struct mm_struct *mm,
unsigned long start, unsigned long last) unsigned long start, unsigned long last)
{ {
struct svm_range_list *svms = prange->svms; struct svm_range_list *svms = prange->svms;
struct svm_range *pchild;
struct kfd_process *p; struct kfd_process *p;
int r = 0; int r = 0;
...@@ -1601,7 +1602,19 @@ svm_range_evict(struct svm_range *prange, struct mm_struct *mm, ...@@ -1601,7 +1602,19 @@ svm_range_evict(struct svm_range *prange, struct mm_struct *mm,
if (!p->xnack_enabled) { if (!p->xnack_enabled) {
int evicted_ranges; int evicted_ranges;
atomic_inc(&prange->invalid); list_for_each_entry(pchild, &prange->child_list, child_list) {
mutex_lock_nested(&pchild->lock, 1);
if (pchild->start <= last && pchild->last >= start) {
pr_debug("increment pchild invalid [0x%lx 0x%lx]\n",
pchild->start, pchild->last);
atomic_inc(&pchild->invalid);
}
mutex_unlock(&pchild->lock);
}
if (prange->start <= last && prange->last >= start)
atomic_inc(&prange->invalid);
evicted_ranges = atomic_inc_return(&svms->evicted_ranges); evicted_ranges = atomic_inc_return(&svms->evicted_ranges);
if (evicted_ranges != 1) if (evicted_ranges != 1)
return r; return r;
...@@ -1618,7 +1631,6 @@ svm_range_evict(struct svm_range *prange, struct mm_struct *mm, ...@@ -1618,7 +1631,6 @@ svm_range_evict(struct svm_range *prange, struct mm_struct *mm,
schedule_delayed_work(&svms->restore_work, schedule_delayed_work(&svms->restore_work,
msecs_to_jiffies(AMDGPU_SVM_RANGE_RESTORE_DELAY_MS)); msecs_to_jiffies(AMDGPU_SVM_RANGE_RESTORE_DELAY_MS));
} else { } else {
struct svm_range *pchild;
unsigned long s, l; unsigned long s, l;
pr_debug("invalidate unmap svms 0x%p [0x%lx 0x%lx] from GPUs\n", pr_debug("invalidate unmap svms 0x%p [0x%lx 0x%lx] from GPUs\n",
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment