From: Jesse Zhang <jesse.zhang@xxxxxxx>
Fix the similar warning:
[ 155.585721] kernfs: can not remove 'enforce_isolation', no directory
[ 155.592201] WARNING: CPU: 3 PID: 6960 at fs/kernfs/dir.c:1683 kernfs_remove_by_name_ns+0xb9/0xc0
[ 155.601145] Modules linked in: xt_MASQUERADE xt_comment nft_compat veth bridge stp llc overlay nft_fib_inet nft_fib_ipv4 nft_fib_ipv6 nft_fib nft_reject_inet nf_reject_ipv4 nf_reject_ipv6 nft_reject nft_ct nft_chain_nat nf_nat nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 ip_set nf_tables nfnetlink qrtr intel_rapl_msr amd_atl intel_rapl_common amd64_edac edac_mce_amd amdgpu kvm_amd kvm ipmi_ssif amdxcp rapl drm_exec gpu_sched drm_buddy i2c_algo_bit drm_suballoc_helper drm_ttm_helper ttm pcspkr drm_display_helper acpi_cpufreq drm_kms_helper video wmi k10temp i2c_piix4 acpi_ipmi ipmi_si drm zram ip_tables loop squashfs dm_multipath crct10dif_pclmul crc32_pclmul crc32c_intel ghash_clmulni_intel sha512_ssse3 sha256_ssse3 sha1_ssse3 sp5100_tco ixgbe rfkill ccp dca sunrpc be2iscsi bnx2i cnic uio cxgb4i cxgb4 tls cxgb3i cxgb3 mdio libcxgbi libcxgb qla4xxx iscsi_boot_sysfs iscsi_tcp libiscsi_tcp libiscsi scsi_transport_iscsi ipmi_devintf ipmi_msghandler fuse
[ 155.685224] systemd-journald[1354]: Compressed data object 957 -> 524 using ZSTD
[ 155.685687] CPU: 3 PID: 6960 Comm: amd_pci_unplug Not tainted 6.10.0-1148853.1.zuul.164395107d6642bdb451071313e9378d #1
[ 155.704149] Hardware name: TYAN B8021G88V2HR-2T/S8021GM2NR-2T, BIOS V1.03.B10 04/01/2019
[ 155.712383] RIP: 0010:kernfs_remove_by_name_ns+0xb9/0xc0
[ 155.717805] Code: a0 00 48 89 ef e8 37 96 c7 ff 5b b8 fe ff ff ff 5d 41 5c 41 5d e9 f7 96 a0 00 0f 0b eb ab 48 c7 c7 48 ba 7e 8f e8 f7 66 bf ff <0f> 0b eb dc 0f 1f 00 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90
[ 155.736766] RSP: 0018:ffffb1685d7a3e20 EFLAGS: 00010296
[ 155.742108] RAX: 0000000000000038 RBX: ffff929e94c80000 RCX: 0000000000000000
[ 155.749363] RDX: ffff928e1efaf200 RSI: ffff928e1efa18c0 RDI: ffff928e1efa18c0
[ 155.756612] RBP: 0000000000000008 R08: 0000000000000000 R09: 0000000000000003
[ 155.763855] R10: ffffb1685d7a3cd8 R11: ffffffff8fb3e1c8 R12: ffffffffc1ef5341
[ 155.771104] R13: ffff929e94cc5530 R14: 0000000000000000 R15: 0000000000000000
[ 155.778357] FS: 00007fd9dd8d9c40(0000) GS:ffff928e1ef80000(0000) knlGS:0000000000000000
[ 155.786594] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 155.792450] CR2: 0000561245ceee38 CR3: 0000000113018000 CR4: 00000000003506f0
[ 155.799702] Call Trace:
[ 155.802254] <TASK>
[ 155.804460] ? __warn+0x80/0x120
[ 155.807798] ? kernfs_remove_by_name_ns+0xb9/0xc0
[ 155.812617] ? report_bug+0x164/0x190
[ 155.816393] ? handle_bug+0x3c/0x80
[ 155.819994] ? exc_invalid_op+0x17/0x70
[ 155.823939] ? asm_exc_invalid_op+0x1a/0x20
[ 155.828235] ? kernfs_remove_by_name_ns+0xb9/0xc0
[ 155.833058] amdgpu_gfx_sysfs_fini+0x59/0xd0 [amdgpu]
[ 155.838637] gfx_v9_0_sw_fini+0x123/0x1c0 [amdgpu]
[ 155.843887] amdgpu_device_fini_sw+0xbc/0x3e0 [amdgpu]
[ 155.849432] amdgpu_driver_release_kms+0x16/0x30 [amdgpu]
[ 155.855235] drm_dev_put.part.0+0x3c/0x60 [drm]
[ 155.859914] drm_release+0x8b/0xc0 [drm]
[ 155.863978] __fput+0xf1/0x2c0
[ 155.867141] __x64_sys_close+0x3c/0x80
[ 155.870998] do_syscall_64+0x64/0x170
Check if the device is unplugged before deleting sysfs files.
v2: squash in mem_info_preempt_used fix
Signed-off-by: Jesse Zhang <jesse.zhang@xxxxxxx>
Suggested-by: Lijo Lazar <lijo.lazar@xxxxxxx>
Reviewed-by: Tim Huang <tim.huang@xxxxxxx>
Signed-off-by: Alex Deucher <alexander.deucher@xxxxxxx>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c | 12 +++++++++---
drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c | 10 ++++++++--
drivers/gpu/drm/amd/amdgpu/amdgpu_preempt_mgr.c | 8 ++++++--
drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c | 10 ++++++++--
drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 9 +++++++--
drivers/gpu/drm/amd/amdgpu/amdgpu_vpe.c | 9 +++++++--
drivers/gpu/drm/amd/amdgpu/df_v3_6.c | 8 ++++++--
7 files changed, 51 insertions(+), 15 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
index 3c89c74d67e0..c65feb97167d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
@@ -25,6 +25,7 @@
#include <linux/firmware.h>
#include <linux/pm_runtime.h>
+#include <drm/drm_drv.h>
#include "amdgpu.h"
#include "amdgpu_gfx.h"
@@ -1778,9 +1779,14 @@ int amdgpu_gfx_sysfs_init(struct amdgpu_device *adev)
void amdgpu_gfx_sysfs_fini(struct amdgpu_device *adev)
{
- amdgpu_gfx_sysfs_xcp_fini(adev);
- amdgpu_gfx_sysfs_isolation_shader_fini(adev);
- amdgpu_gfx_sysfs_reset_mask_fini(adev);
+ int idx;
+
+ if (drm_dev_enter(adev_to_drm(adev), &idx)) {
+ amdgpu_gfx_sysfs_xcp_fini(adev);
+ amdgpu_gfx_sysfs_isolation_shader_fini(adev);
+ amdgpu_gfx_sysfs_reset_mask_fini(adev);
+ drm_dev_exit(idx);
+ }
}
int amdgpu_gfx_cleaner_shader_sw_init(struct amdgpu_device *adev,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c
index 04eb51674596..29449d9f8e55 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c
@@ -24,6 +24,7 @@
*
*/
+#include <drm/drm_drv.h>
#include "amdgpu.h"
#include "amdgpu_jpeg.h"
#include "amdgpu_pm.h"
@@ -447,6 +448,11 @@ int amdgpu_jpeg_sysfs_reset_mask_init(struct amdgpu_device *adev)
void amdgpu_jpeg_sysfs_reset_mask_fini(struct amdgpu_device *adev)
{
- if (adev->jpeg.num_jpeg_inst)
- device_remove_file(adev->dev, &dev_attr_jpeg_reset_mask);
+ int idx;
+
+ if (drm_dev_enter(adev_to_drm(adev), &idx)) {
+ if (adev->jpeg.num_jpeg_inst)
+ device_remove_file(adev->dev, &dev_attr_jpeg_reset_mask);
+ drm_dev_exit(idx);
+ }
}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_preempt_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_preempt_mgr.c
index e8adfd0a570a..9a0346ed6ea4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_preempt_mgr.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_preempt_mgr.c
@@ -23,6 +23,7 @@
* Authors: Christian König, Felix Kuehling
*/
+#include <drm/drm_drv.h>
#include "amdgpu.h"
/**
@@ -129,7 +130,7 @@ int amdgpu_preempt_mgr_init(struct amdgpu_device *adev)
void amdgpu_preempt_mgr_fini(struct amdgpu_device *adev)
{
struct ttm_resource_manager *man = &adev->mman.preempt_mgr;
- int ret;
+ int idx, ret;
ttm_resource_manager_set_used(man, false);
@@ -137,7 +138,10 @@ void amdgpu_preempt_mgr_fini(struct amdgpu_device *adev)
if (ret)
return;
- device_remove_file(adev->dev, &dev_attr_mem_info_preempt_used);
+ if (drm_dev_enter(adev_to_drm(adev), &idx)) {
+ device_remove_file(adev->dev, &dev_attr_mem_info_preempt_used);
+ drm_dev_exit(idx);
+ }
ttm_resource_manager_cleanup(man);
ttm_set_driver_manager(&adev->mman.bdev, AMDGPU_PL_PREEMPT, NULL);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c
index 8c89b69edc20..24e9daacaabb 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.c
@@ -21,6 +21,7 @@
*
*/
+#include <drm/drm_drv.h>
#include <linux/firmware.h>
#include "amdgpu.h"
#include "amdgpu_sdma.h"
@@ -448,9 +449,14 @@ int amdgpu_sdma_sysfs_reset_mask_init(struct amdgpu_device *adev)
void amdgpu_sdma_sysfs_reset_mask_fini(struct amdgpu_device *adev)
{
+ int idx;
+
if (!amdgpu_gpu_recovery)
return;
- if (adev->sdma.num_instances)
- device_remove_file(adev->dev, &dev_attr_sdma_reset_mask);
+ if (drm_dev_enter(adev_to_drm(adev), &idx)) {
+ if (adev->sdma.num_instances)
+ device_remove_file(adev->dev, &dev_attr_sdma_reset_mask);
+ drm_dev_exit(idx);
+ }
}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
index c813dc94a7bc..b4a550795470 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
@@ -1315,6 +1315,11 @@ int amdgpu_vcn_sysfs_reset_mask_init(struct amdgpu_device *adev)
void amdgpu_vcn_sysfs_reset_mask_fini(struct amdgpu_device *adev)
{
- if (adev->vcn.num_vcn_inst)
- device_remove_file(adev->dev, &dev_attr_vcn_reset_mask);
+ int idx;
+
+ if (drm_dev_enter(adev_to_drm(adev), &idx)) {
+ if (adev->vcn.num_vcn_inst)
+ device_remove_file(adev->dev, &dev_attr_vcn_reset_mask);
+ drm_dev_exit(idx);
+ }
}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vpe.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vpe.c
index 74e671c74142..fa0fd784dd13 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vpe.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vpe.c
@@ -904,8 +904,13 @@ int amdgpu_vpe_sysfs_reset_mask_init(struct amdgpu_device *adev)
void amdgpu_vpe_sysfs_reset_mask_fini(struct amdgpu_device *adev)
{
- if (adev->vpe.num_instances)
- device_remove_file(adev->dev, &dev_attr_vpe_reset_mask);
+ int idx;
+
+ if (drm_dev_enter(adev_to_drm(adev), &idx)) {
+ if (adev->vpe.num_instances)
+ device_remove_file(adev->dev, &dev_attr_vpe_reset_mask);
+ drm_dev_exit(idx);
+ }
}
static const struct amdgpu_ring_funcs vpe_ring_funcs = {
diff --git a/drivers/gpu/drm/amd/amdgpu/df_v3_6.c b/drivers/gpu/drm/amd/amdgpu/df_v3_6.c
index 483a441b46aa..54c05af2eed2 100644
--- a/drivers/gpu/drm/amd/amdgpu/df_v3_6.c
+++ b/drivers/gpu/drm/amd/amdgpu/df_v3_6.c
@@ -20,6 +20,7 @@
* OTHER DEALINGS IN THE SOFTWARE.
*
*/
+#include <drm/drm_drv.h>
#include "amdgpu.h"
#include "df_v3_6.h"
@@ -254,9 +255,12 @@ static void df_v3_6_sw_init(struct amdgpu_device *adev)
static void df_v3_6_sw_fini(struct amdgpu_device *adev)
{
+ int idx;
- device_remove_file(adev->dev, &dev_attr_df_cntr_avail);
-
+ if (drm_dev_enter(adev_to_drm(adev), &idx)) {
+ device_remove_file(adev->dev, &dev_attr_df_cntr_avail);
+ drm_dev_exit(idx);
+ }
}
static void df_v3_6_enable_broadcast_mode(struct amdgpu_device *adev,
--
2.47.0
