mirrors/linux
1
0
Fork 0
mirror of https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git synced 2026-02-27 12:55:59 -05:00
Code Issues Packages Projects Releases Wiki Activity

drm/amdgpu: optimize insert_nop using multi dwords

Browse Source 
Optimize the ring_insert_nop fn for n dwords in one
step rather then call to amdgpu_ring_write for each
nop packet. This avoid function call for each nop
packet and also wptr is updated once only.

Signed-off-by: Sunil Khatri <sunil.khatri@amd.com>
Suggested-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This commit is contained in:
Sunil Khatri
2024-10-08 18:32:16 +05:30
committed by Alex Deucher
parent ed3dac4bf9
commit ea4e4754c9
1 changed files with 19 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c
View File

@@ -108,10 +108,26 @@ int amdgpu_ring_alloc(struct amdgpu_ring *ring, unsigned int ndw)
*/
void amdgpu_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
{
int i;
uint32_t occupied, chunk1, chunk2;
uint32_t *dst;
for (i = 0; i < count; i++)
amdgpu_ring_write(ring, ring->funcs->nop);
occupied = ring->wptr & ring->buf_mask;
dst = (void *)&ring->ring[occupied];
chunk1 = ring->buf_mask + 1 - occupied;
chunk1 = (chunk1 >= count) ? count : chunk1;
chunk2 = count - chunk1;
if (chunk1)
memset32(dst, ring->funcs->nop, chunk1);
if (chunk2) {
dst = (void *)ring->ring;
memset32(dst, ring->funcs->nop, chunk2);
}
ring->wptr += count;
ring->wptr &= ring->ptr_mask;
ring->count_dw -= count;
}
/**