Running the Cyberpunk 2077 benchmark we can observe that the lookup helper
is relatively hot, but the 97% of the calls are for a single object. (~3%
for two points, and never more than three points. While a more trivial
workload like vkmark under Plasma is even more skewed to single point
lookups.)
Therefore lets add a fast path to bypass the kmalloc_array/kfree and use a
pre-allocated stack array for those cases.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursu...@igalia.com>
---
drivers/gpu/drm/drm_syncobj.c | 53 +++++++++++++++++++++++++++--------
1 file changed, 41 insertions(+), 12 deletions(-)
diff --git a/drivers/gpu/drm/drm_syncobj.c b/drivers/gpu/drm/drm_syncobj.c
index 94932b89298f..233bdef53c87 100644
--- a/drivers/gpu/drm/drm_syncobj.c
+++ b/drivers/gpu/drm/drm_syncobj.c
@@ -1223,6 +1223,8 @@ EXPORT_SYMBOL(drm_timeout_abs_to_jiffies);
static int drm_syncobj_array_find(struct drm_file *file_private,
u32 __user *handles,
uint32_t count,
+ struct drm_syncobj **stack_syncobjs,
+ u32 stack_count,
struct drm_syncobj ***syncobjs_out)
{
struct drm_syncobj **syncobjs;
@@ -1232,9 +1234,13 @@ static int drm_syncobj_array_find(struct drm_file
*file_private,
if (!access_ok(handles, count * sizeof(*handles)))
return -EFAULT;
- syncobjs = kmalloc_array(count, sizeof(*syncobjs), GFP_KERNEL);
- if (!syncobjs)
- return -ENOMEM;
+ if (count > stack_count) {
+ syncobjs = kmalloc_array(count, sizeof(*syncobjs), GFP_KERNEL);
+ if (!syncobjs)
+ return -ENOMEM;
+ } else {
+ syncobjs = stack_syncobjs;
+ }
for (i = 0; i < count; i++) {
u64 handle;
@@ -1260,25 +1266,31 @@ static int drm_syncobj_array_find(struct drm_file
*file_private,
drm_syncobj_put(syncobjs[i]);
i--;
}
- kfree(syncobjs);
+
+ if (syncobjs != stack_syncobjs)
+ kfree(syncobjs);
return ret;
}
static void drm_syncobj_array_free(struct drm_syncobj **syncobjs,
- uint32_t count)
+ uint32_t count,
+ struct drm_syncobj **stack_syncobjs)
{
uint32_t i;
for (i = 0; i < count; i++)
drm_syncobj_put(syncobjs[i]);
- kfree(syncobjs);
+
+ if (syncobjs != stack_syncobjs)
+ kfree(syncobjs);
}
int
drm_syncobj_wait_ioctl(struct drm_device *dev, void *data,
struct drm_file *file_private)
{
+ struct drm_syncobj *stack_syncobjs[4];
struct drm_syncobj_wait *args = data;
ktime_t deadline, *pdeadline = NULL;
u32 count = args->count_handles;
@@ -1304,6 +1316,8 @@ drm_syncobj_wait_ioctl(struct drm_device *dev, void *data,
ret = drm_syncobj_array_find(file_private,
u64_to_user_ptr(args->handles),
count,
+ stack_syncobjs,
+ ARRAY_SIZE(stack_syncobjs),
&syncobjs);
if (ret < 0)
return ret;
@@ -1321,7 +1335,7 @@ drm_syncobj_wait_ioctl(struct drm_device *dev, void *data,
&first,
pdeadline);
- drm_syncobj_array_free(syncobjs, count);
+ drm_syncobj_array_free(syncobjs, count, stack_syncobjs);
if (timeout < 0)
return timeout;
@@ -1336,6 +1350,7 @@ drm_syncobj_timeline_wait_ioctl(struct drm_device *dev,
void *data,
struct drm_file *file_private)
{
struct drm_syncobj_timeline_wait *args = data;
+ struct drm_syncobj *stack_syncobjs[4];
ktime_t deadline, *pdeadline = NULL;
u32 count = args->count_handles;
struct drm_syncobj **syncobjs;
@@ -1361,6 +1376,8 @@ drm_syncobj_timeline_wait_ioctl(struct drm_device *dev,
void *data,
ret = drm_syncobj_array_find(file_private,
u64_to_user_ptr(args->handles),
count,
+ stack_syncobjs,
+ ARRAY_SIZE(stack_syncobjs),
&syncobjs);
if (ret < 0)
return ret;
@@ -1378,7 +1395,7 @@ drm_syncobj_timeline_wait_ioctl(struct drm_device *dev,
void *data,
&first,
pdeadline);
- drm_syncobj_array_free(syncobjs, count);
+ drm_syncobj_array_free(syncobjs, count, stack_syncobjs);
if (timeout < 0)
return timeout;
@@ -1496,6 +1513,7 @@ drm_syncobj_reset_ioctl(struct drm_device *dev, void
*data,
struct drm_file *file_private)
{
struct drm_syncobj_array *args = data;
+ struct drm_syncobj *stack_syncobjs[4];
struct drm_syncobj **syncobjs;
uint32_t i;
int ret;
@@ -1512,6 +1530,8 @@ drm_syncobj_reset_ioctl(struct drm_device *dev, void
*data,
ret = drm_syncobj_array_find(file_private,
u64_to_user_ptr(args->handles),
args->count_handles,
+ stack_syncobjs,
+ ARRAY_SIZE(stack_syncobjs),
&syncobjs);
if (ret < 0)
return ret;
@@ -1519,7 +1539,7 @@ drm_syncobj_reset_ioctl(struct drm_device *dev, void
*data,
for (i = 0; i < args->count_handles; i++)
drm_syncobj_replace_fence(syncobjs[i], NULL);
- drm_syncobj_array_free(syncobjs, args->count_handles);
+ drm_syncobj_array_free(syncobjs, args->count_handles, stack_syncobjs);
return 0;
}
@@ -1529,6 +1549,7 @@ drm_syncobj_signal_ioctl(struct drm_device *dev, void
*data,
struct drm_file *file_private)
{
struct drm_syncobj_array *args = data;
+ struct drm_syncobj *stack_syncobjs[4];
struct drm_syncobj **syncobjs;
uint32_t i;
int ret;
@@ -1545,6 +1566,8 @@ drm_syncobj_signal_ioctl(struct drm_device *dev, void
*data,
ret = drm_syncobj_array_find(file_private,
u64_to_user_ptr(args->handles),
args->count_handles,
+ stack_syncobjs,
+ ARRAY_SIZE(stack_syncobjs),
&syncobjs);
if (ret < 0)
return ret;
@@ -1555,7 +1578,7 @@ drm_syncobj_signal_ioctl(struct drm_device *dev, void
*data,
break;
}
- drm_syncobj_array_free(syncobjs, args->count_handles);
+ drm_syncobj_array_free(syncobjs, args->count_handles, stack_syncobjs);
return ret;
}
@@ -1567,6 +1590,7 @@ drm_syncobj_timeline_signal_ioctl(struct drm_device *dev,
void *data,
struct drm_syncobj_timeline_array *args = data;
uint64_t __user *points = u64_to_user_ptr(args->points);
uint32_t i, j, count = args->count_handles;
+ struct drm_syncobj *stack_syncobjs[4];
struct drm_syncobj **syncobjs;
struct dma_fence_chain **chains;
int ret;
@@ -1586,6 +1610,8 @@ drm_syncobj_timeline_signal_ioctl(struct drm_device *dev,
void *data,
ret = drm_syncobj_array_find(file_private,
u64_to_user_ptr(args->handles),
count,
+ stack_syncobjs,
+ ARRAY_SIZE(stack_syncobjs),
&syncobjs);
if (ret < 0)
return ret;
@@ -1622,7 +1648,7 @@ drm_syncobj_timeline_signal_ioctl(struct drm_device *dev,
void *data,
err_chains:
kfree(chains);
out:
- drm_syncobj_array_free(syncobjs, count);
+ drm_syncobj_array_free(syncobjs, count, stack_syncobjs);
return ret;
}
@@ -1631,6 +1657,7 @@ int drm_syncobj_query_ioctl(struct drm_device *dev, void
*data,
struct drm_file *file_private)
{
struct drm_syncobj_timeline_array *args = data;
+ struct drm_syncobj *stack_syncobjs[4];
struct drm_syncobj **syncobjs;
uint64_t __user *points = u64_to_user_ptr(args->points);
uint32_t i;
@@ -1651,6 +1678,8 @@ int drm_syncobj_query_ioctl(struct drm_device *dev, void
*data,
ret = drm_syncobj_array_find(file_private,
u64_to_user_ptr(args->handles),
args->count_handles,
+ stack_syncobjs,
+ ARRAY_SIZE(stack_syncobjs),
&syncobjs);
if (ret < 0)
return ret;
@@ -1694,7 +1723,7 @@ int drm_syncobj_query_ioctl(struct drm_device *dev, void
*data,
break;
}
}
- drm_syncobj_array_free(syncobjs, args->count_handles);
+ drm_syncobj_array_free(syncobjs, args->count_handles, stack_syncobjs);
return ret;
}
--
2.48.0
