Fix another sysfs duplication error when reinitializing the device:
sysfs: cannot create duplicate filename '.../enforce_isolation'
Fix this by:
- improve amdgpu_gfx_sysfs_init error unwinding so partially-created
gfx sysfs files are removed on registration failure.
- added adev->gfx.sysfs_registered tracking so gfx sysfs cleanup is
idempotent when harware teardown runs before the existing software
teardown cleanup path.
- calling amdgpu_gfx_sysfs_fini from amdgpu_device_fini_hw before the
general device sysfs cleanup.
Signed-off-by: Geoffrey McRae <[email protected]>
Cc: Alex Deucher <[email protected]>
Cc: Christian König <[email protected]>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 1 +
drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c | 32 +++++++++++++++++-----
drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h | 2 ++
3 files changed, 28 insertions(+), 7 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 5442a1fc1c37..7278d68f8021 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -4242,6 +4242,7 @@ void amdgpu_device_fini_hw(struct amdgpu_device *adev)
}
amdgpu_fence_driver_hw_fini(adev);
+ amdgpu_gfx_sysfs_fini(adev);
amdgpu_device_sys_interface_fini(adev);
/* disable ras feature must before hw fini */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
index a5b835d0c166..a4616d54adcc 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
@@ -2195,6 +2195,8 @@ int amdgpu_gfx_sysfs_init(struct amdgpu_device *adev)
{
int r;
+ adev->gfx.sysfs_registered = false;
+
r = amdgpu_gfx_sysfs_xcp_init(adev);
if (r) {
dev_err(adev->dev, "failed to create xcp sysfs files");
@@ -2202,23 +2204,39 @@ int amdgpu_gfx_sysfs_init(struct amdgpu_device *adev)
}
r = amdgpu_gfx_sysfs_isolation_shader_init(adev);
- if (r)
+ if (r) {
dev_err(adev->dev, "failed to create isolation sysfs files");
+ goto err_isolation_shader;
+ }
r = amdgpu_gfx_sysfs_reset_mask_init(adev);
- if (r)
+ if (r) {
dev_err(adev->dev, "failed to create reset mask sysfs files");
+ goto err_reset_mask;
+ }
+ adev->gfx.sysfs_registered = true;
+
+ return 0;
+
+err_reset_mask:
+ amdgpu_gfx_sysfs_reset_mask_fini(adev);
+ amdgpu_gfx_sysfs_isolation_shader_fini(adev);
+err_isolation_shader:
+ amdgpu_gfx_sysfs_xcp_fini(adev);
return r;
}
void amdgpu_gfx_sysfs_fini(struct amdgpu_device *adev)
{
- if (adev->dev->kobj.sd) {
- amdgpu_gfx_sysfs_xcp_fini(adev);
- amdgpu_gfx_sysfs_isolation_shader_fini(adev);
- amdgpu_gfx_sysfs_reset_mask_fini(adev);
- }
+ if (!adev->gfx.sysfs_registered)
+ return;
+
+ amdgpu_gfx_sysfs_xcp_fini(adev);
+ amdgpu_gfx_sysfs_isolation_shader_fini(adev);
+ amdgpu_gfx_sysfs_reset_mask_fini(adev);
+
+ adev->gfx.sysfs_registered = false;
}
static void amdgpu_gfx_reset_start_compute_scheds(struct amdgpu_device *adev,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
index 9432107c96a1..64cb1c1e51fb 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
@@ -477,6 +477,8 @@ struct amdgpu_gfx {
struct amdgpu_irq_src rlc_poison_irq;
struct sq_work sq_work;
+ bool sysfs_registered;
+
/* gfx status */
uint32_t gfx_current_status;
/* ce ram size*/
--
2.43.0
