Am 11.01.22 um 08:09 schrieb Somalapuram Amaranath:
trace_amdgpu_vm_update_ptes trace unable to log when nptes too large
Signed-off-by: Somalapuram Amaranath <amaranath.somalapu...@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index 0e7dc23f78e7..0630bd43238c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -1632,6 +1632,7 @@ static int amdgpu_vm_update_ptes(struct
amdgpu_vm_update_params *params,
* silent to stop fault floods.
*/
nptes = max(nptes, 1u);
+ nptes = nptes > 32 ? 32 : nptes;
Please do something like min(nptes, 32) in the parameter list of
trace_amdgpu_vm_update_ptes instead. I'm not sure if the nptes are not
needed any more after this.
Then please also remove the previous workaround in amdgpu_trace.h.
Apart from that sounds like the right idea to me.
Christian.
trace_amdgpu_vm_update_ptes(params, frag_start, upd_end,
nptes, dst, incr, upd_flags,
