[PATCH] drm/amdkfd: move flushing TLBs from map to unmap

[Eric Huang <jinhuieric.huang@xxxxxxx>]

It it to optimize memory allocation latency.

Signed-off-by: Eric Huang <jinhuieric.huang@xxxxxxx>

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c 
b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
index 960913a35ee4..ab73741edb97 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
@@ -1657,20 +1657,6 @@ static int kfd_ioctl_map_memory_to_gpu(struct 
file *filep,
                goto sync_memory_failed;
        }

-       /* Flush TLBs after waiting for the page table updates to 
complete */
-       for (i = 0; i < args->n_devices; i++) {
-               peer = kfd_device_by_id(devices_arr[i]);
-               if (WARN_ON_ONCE(!peer))
-                       continue;
-               peer_pdd = kfd_get_process_device_data(peer, p);
-               if (WARN_ON_ONCE(!peer_pdd))
-                       continue;
-               if (!amdgpu_read_lock(peer->ddev, true)) {
-                       kfd_flush_tlb(peer_pdd);
-                       amdgpu_read_unlock(peer->ddev);
-               }
-       }
-
        kfree(devices_arr);

        trace_kfd_map_memory_to_gpu_end(p,
@@ -1766,6 +1752,7 @@ static int kfd_ioctl_unmap_memory_from_gpu(struct 
file *filep,
                        amdgpu_read_unlock(peer->ddev);
                        goto unmap_memory_from_gpu_failed;
                }
+               kfd_flush_tlb(peer_pdd);
                amdgpu_read_unlock(peer->ddev);
                args->n_success = i+1;
        }
_______________________________________________
amd-gfx mailing list
amd-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/amd-gfx