It's valid to add the same fence multiple times to a dma-resv object and we shouldn't need one extra slot for each.
Signed-off-by: Christian König christian.koenig@amd.com Fixes: a3f7c10a269d5 ("dma-buf/dma-resv: check if the new fence is really later") Cc: stable@vger.kernel.org # v5.19+ --- drivers/dma-buf/dma-resv.c | 2 +- include/linux/dma-fence.h | 15 +++++++++++++++ 2 files changed, 16 insertions(+), 1 deletion(-)
diff --git a/drivers/dma-buf/dma-resv.c b/drivers/dma-buf/dma-resv.c index 38b4110378de..eb8b733065b2 100644 --- a/drivers/dma-buf/dma-resv.c +++ b/drivers/dma-buf/dma-resv.c @@ -301,7 +301,7 @@ void dma_resv_add_fence(struct dma_resv *obj, struct dma_fence *fence,
dma_resv_list_entry(fobj, i, obj, &old, &old_usage); if ((old->context == fence->context && old_usage >= usage && - dma_fence_is_later(fence, old)) || + dma_fence_is_later_or_same(fence, old)) || dma_fence_is_signaled(old)) { dma_resv_list_set(fobj, i, fence, usage); dma_fence_put(old); diff --git a/include/linux/dma-fence.h b/include/linux/dma-fence.h index ebe78bd3d121..b3772edca2e6 100644 --- a/include/linux/dma-fence.h +++ b/include/linux/dma-fence.h @@ -498,6 +498,21 @@ static inline bool dma_fence_is_later(struct dma_fence *f1, return __dma_fence_is_later(f1->seqno, f2->seqno, f1->ops); }
+/** + * dma_fence_is_later_or_same - return true if f1 is later or same as f2 + * @f1: the first fence from the same context + * @f2: the second fence from the same context + * + * Returns true if f1 is chronologically later than f2 or the same fence. Both + * fences must be from the same context, since a seqno is not re-used across + * contexts. + */ +static inline bool dma_fence_is_later_or_same(struct dma_fence *f1, + struct dma_fence *f2) +{ + return f1 == f2 || dma_fence_is_later(f1, f2); +} + /** * dma_fence_later - return the chronologically later fence * @f1: the first fence from the same context
Ping? Can I get an rb or acked-by for that?
Thanks, Christian.
Am 15.11.23 um 10:30 schrieb Christian König:
It's valid to add the same fence multiple times to a dma-resv object and we shouldn't need one extra slot for each.
Signed-off-by: Christian König christian.koenig@amd.com Fixes: a3f7c10a269d5 ("dma-buf/dma-resv: check if the new fence is really later") Cc: stable@vger.kernel.org # v5.19+
drivers/dma-buf/dma-resv.c | 2 +- include/linux/dma-fence.h | 15 +++++++++++++++ 2 files changed, 16 insertions(+), 1 deletion(-)
diff --git a/drivers/dma-buf/dma-resv.c b/drivers/dma-buf/dma-resv.c index 38b4110378de..eb8b733065b2 100644 --- a/drivers/dma-buf/dma-resv.c +++ b/drivers/dma-buf/dma-resv.c @@ -301,7 +301,7 @@ void dma_resv_add_fence(struct dma_resv *obj, struct dma_fence *fence, dma_resv_list_entry(fobj, i, obj, &old, &old_usage); if ((old->context == fence->context && old_usage >= usage &&
dma_fence_is_later(fence, old)) ||
dma_fence_is_signaled(old)) { dma_resv_list_set(fobj, i, fence, usage); dma_fence_put(old);dma_fence_is_later_or_same(fence, old)) ||
diff --git a/include/linux/dma-fence.h b/include/linux/dma-fence.h index ebe78bd3d121..b3772edca2e6 100644 --- a/include/linux/dma-fence.h +++ b/include/linux/dma-fence.h @@ -498,6 +498,21 @@ static inline bool dma_fence_is_later(struct dma_fence *f1, return __dma_fence_is_later(f1->seqno, f2->seqno, f1->ops); } +/**
- dma_fence_is_later_or_same - return true if f1 is later or same as f2
- @f1: the first fence from the same context
- @f2: the second fence from the same context
- Returns true if f1 is chronologically later than f2 or the same fence. Both
- fences must be from the same context, since a seqno is not re-used across
- contexts.
- */
+static inline bool dma_fence_is_later_or_same(struct dma_fence *f1,
struct dma_fence *f2)
+{
- return f1 == f2 || dma_fence_is_later(f1, f2);
+}
- /**
- dma_fence_later - return the chronologically later fence
- @f1: the first fence from the same context
On 11/27/23 14:24, Christian König wrote:
Ping? Can I get an rb or acked-by for that?
Thanks, Christian.
Am 15.11.23 um 10:30 schrieb Christian König:
It's valid to add the same fence multiple times to a dma-resv object and we shouldn't need one extra slot for each.
Signed-off-by: Christian König christian.koenig@amd.com Fixes: a3f7c10a269d5 ("dma-buf/dma-resv: check if the new fence is really later") Cc: stable@vger.kernel.org # v5.19+
Reviewed-by: Thomas Hellström thomas.hellstrom@linux.intel.com
drivers/dma-buf/dma-resv.c | 2 +- include/linux/dma-fence.h | 15 +++++++++++++++ 2 files changed, 16 insertions(+), 1 deletion(-)
diff --git a/drivers/dma-buf/dma-resv.c b/drivers/dma-buf/dma-resv.c index 38b4110378de..eb8b733065b2 100644 --- a/drivers/dma-buf/dma-resv.c +++ b/drivers/dma-buf/dma-resv.c @@ -301,7 +301,7 @@ void dma_resv_add_fence(struct dma_resv *obj, struct dma_fence *fence, dma_resv_list_entry(fobj, i, obj, &old, &old_usage); if ((old->context == fence->context && old_usage >= usage && - dma_fence_is_later(fence, old)) || + dma_fence_is_later_or_same(fence, old)) || dma_fence_is_signaled(old)) { dma_resv_list_set(fobj, i, fence, usage); dma_fence_put(old); diff --git a/include/linux/dma-fence.h b/include/linux/dma-fence.h index ebe78bd3d121..b3772edca2e6 100644 --- a/include/linux/dma-fence.h +++ b/include/linux/dma-fence.h @@ -498,6 +498,21 @@ static inline bool dma_fence_is_later(struct dma_fence *f1, return __dma_fence_is_later(f1->seqno, f2->seqno, f1->ops); } +/**
- dma_fence_is_later_or_same - return true if f1 is later or same
as f2
- @f1: the first fence from the same context
- @f2: the second fence from the same context
- Returns true if f1 is chronologically later than f2 or the same
fence. Both
- fences must be from the same context, since a seqno is not
re-used across
- contexts.
- */
+static inline bool dma_fence_is_later_or_same(struct dma_fence *f1, + struct dma_fence *f2) +{ + return f1 == f2 || dma_fence_is_later(f1, f2); +}
/** * dma_fence_later - return the chronologically later fence * @f1: the first fence from the same context
Linaro-mm-sig mailing list -- linaro-mm-sig@lists.linaro.org To unsubscribe send an email to linaro-mm-sig-leave@lists.linaro.org
linaro-mm-sig@lists.linaro.org