[PATCH v2] drm/amdgpu: Fix lockdep warning in RAS SYSFS v2

Andrey Grodzovsky andrey.grodzovsky at amd.com
Fri Mar 8 22:29:36 UTC 2019


Problem:
When loading driver with debug lockdep enabled the WARN_ON as bellow
was observed. Gooling about this warning i found the follwing explanation -
https://git.sphere.ly/tucstwo/cam-test/commit/671ee198b38694cf1dfbaa0b9ea823929517c367

Fix:
Switch all debugfs attributes in RAS to static

v2: Add correct WARN_ON message to description.

Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.670622] DEBUG_LOCKS_WARN_ON(1)
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.670630] WARNING: CPU: 5 PID: 1100 at kernel/locking/lockdep.c:3129 lockdep_init_map+0x288/0x290
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.670761] Modules linked in: amdgpu(O+) chash gpu_sched(O) ttm(O) drm_kms_helper(O) drm(O) i2c_algo_bit fb_sys_fops syscopyarea sysfillrect sysimgblt intel_rapl x86_pkg_temp_thermal intel_powerclamp coretemp kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul snd_hda_codec_realtek ghash_clmulni_intel snd_hda_codec_generic ledtrig_audio snd_hda_intel snd_hda_codec snd_hda_core snd_hwdep snd_pcm snd_seq_midi snd_seq_midi_event aesni_intel snd_rawmidi aes_x86_64 crypto_simd cryptd glue_helper eeepc_wmi snd_seq asus_wmi sparse_keymap wmi_bmof snd_seq_device snd_timer serio_raw joydev snd soundcore mei_me mei acpi_pad mac_hid binfmt_misc nfsd auth_rpcgss nfs_acl parport_pc lockd ppdev grace lp parport sunrpc autofs4 hid_generic psmouse e1000e r8169 ahci libahci usbhid hid mxm_wmi wmi video
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.670922] CPU: 5 PID: 1100 Comm: modprobe Tainted: G           O      5.0.0-rc1-dev+ #37
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.670991] Hardware name: System manufacturer System Product Name/Z170-PRO, BIOS 1902 06/27/2016
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.671061] RIP: 0010:lockdep_init_map+0x288/0x290
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.671127] Code: c7 34 3d 1b 83 e8 08 fc 21 00 83 3d 35 4e 05 02 00 0f 85 df fe ff ff 48 c7 c6 00 99 48 82 48 c7 c7 20 94 48 82 e8 b8 87 f6 ff <0f> 0b e9 c5 fe ff ff 90 49 89 f0 31 c9 31 d2 31 f6 e9 32 8f 07 00
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.671219] RSP: 0018:ffff8883e0faf0e8 EFLAGS: 00010286
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.671291] RAX: 0000000000000000 RBX: ffff8883b07df348 RCX: ffffffff81165de4
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.671364] RDX: 0000000000000003 RSI: dffffc0000000000 RDI: 0000000000000246
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.671437] RBP: ffffffff8253e4e0 R08: fffffbfff05a271d R09: fffffbfff05a271d
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.671510] R10: 0000000000000001 R11: fffffbfff05a271c R12: ffff8883b07d2140
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.671583] R13: 0000000000001001 R14: 0000000000000000 R15: 0000000000000000
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.671656] FS:  00007fcc21833700(0000) GS:ffff8883f4080000(0000) knlGS:0000000000000000
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.671737] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.671809] CR2: 000055b90a13b000 CR3: 00000003d96de005 CR4: 00000000003606e0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.671882] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.671955] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.672027] Call Trace:
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.672099]  __kernfs_create_file+0x9d/0x150
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.672171]  sysfs_add_file_mode_ns+0x11d/0x270
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.672246]  internal_create_group+0x218/0x600
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.672324]  ? remove_files.isra.1+0xa0/0xa0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.672396]  ? rcu_read_lock_sched_held+0x75/0x80
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.672555]  ? amdgpu_ras_create_obj+0x10c/0x130 [amdgpu]
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.672703]  ? __amdgpu_ras_feature_enable+0x109/0x200 [amdgpu]
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.672862]  amdgpu_ras_init+0x41e/0x560 [amdgpu]
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.673012]  ? amdgpu_ras_reserve_bad_pages+0x5e0/0x5e0 [amdgpu]
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.673086]  ? __mutex_unlock_slowpath+0xda/0x420
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.673162]  ? wait_for_completion+0x200/0x200
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.673240]  ? idr_alloc_u32+0x1b0/0x1b0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.673349]  ? drm_property_create+0x18a/0x200 [drm]
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.673500]  amdgpu_device_init+0x15fc/0x2950 [amdgpu]
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.673646]  ? amdgpu_device_has_dc_support+0x30/0x30 [amdgpu]
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.673720]  ? __alloc_pages_nodemask+0x232/0x460
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.673792]  ? __alloc_pages_slowpath+0x1370/0x1370
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.673864]  ? __mutex_unlock_slowpath+0xda/0x420
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.673936]  ? policy_nodemask+0x19/0xa0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.674008]  ? kasan_unpoison_shadow+0x36/0x50
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.674079]  ? kasan_kmalloc_large+0x9a/0xe0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.674222]  amdgpu_driver_load_kms+0x101/0x540 [amdgpu]
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.674371]  ? amdgpu_driver_unload_kms+0x220/0x220 [amdgpu]
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.674458]  ? drm_dev_register+0x1a4/0x320 [drm]
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.674530]  ? __kasan_slab_free+0x138/0x170
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.674617]  drm_dev_register+0x1fd/0x320 [drm]
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.674761]  amdgpu_pci_probe+0xef/0x1a0 [amdgpu]
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.674904]  ? amdgpu_pci_remove+0x60/0x60 [amdgpu]
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.674977]  local_pci_probe+0x76/0xe0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.675048]  pci_device_probe+0x205/0x300
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.675119]  ? kernfs_create_link+0xae/0x100
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.675191]  ? pci_device_remove+0x1c0/0x1c0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.675264]  really_probe+0x382/0x5e0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.675336]  driver_probe_device+0x171/0x1b0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.675408]  __driver_attach+0x193/0x1a0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.675479]  ? driver_probe_device+0x1b0/0x1b0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.675550]  bus_for_each_dev+0xe4/0x160
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.675621]  ? lock_downgrade+0x2f0/0x2f0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.675691]  ? subsys_dev_iter_exit+0x10/0x10
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.675764]  bus_add_driver+0x322/0x3a0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.675836]  driver_register+0xc6/0x1a0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.675907]  ? 0xffffffffa1090000
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.675978]  do_one_initcall+0xb8/0x29f
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.676049]  ? trace_event_raw_event_initcall_finish+0x150/0x150
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.676122]  ? kasan_unpoison_shadow+0x36/0x50
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.676193]  ? kasan_kmalloc+0xae/0xf0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.676264]  ? kmem_cache_alloc_trace+0x14d/0x2b0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.676336]  ? do_init_module+0x35/0x335
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.676406]  ? kasan_unpoison_shadow+0x36/0x50
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.676479]  do_init_module+0xec/0x335
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.676550]  load_module+0x3d5d/0x4780
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.676627]  ? module_frob_arch_sections+0x20/0x20
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.676700]  ? ima_read_file+0x10/0x10
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.676770]  ? vfs_read+0x127/0x190
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.676842]  ? kernel_read+0x74/0xa0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.676913]  ? kernel_read_file+0x16c/0x350
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.676986]  ? apparmor_task_free+0xc0/0xc0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.677057]  ? do_mmap+0x55e/0x790
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.677130]  ? __do_sys_finit_module+0x175/0x1b0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.677201]  __do_sys_finit_module+0x175/0x1b0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.677273]  ? __ia32_sys_init_module+0x40/0x40
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.677344]  ? check_chain_key+0x131/0x1e0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.677416]  ? syscall_trace_enter+0x1fc/0x530
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.677491]  ? vtime_user_exit+0xc8/0xe0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.677563]  do_syscall_64+0x7d/0x1f0
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.677634]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.677706] RIP: 0033:0x7fcc213654d9
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.677776] Code: 00 f3 c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 40 00 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 8f 29 2c 00 f7 d8 64 89 01 48
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.677873] RSP: 002b:00007ffc8d3c2888 EFLAGS: 00000246 ORIG_RAX: 0000000000000139
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.677954] RAX: ffffffffffffffda RBX: 000055b90a1363b0 RCX: 00007fcc213654d9
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.678027] RDX: 0000000000000000 RSI: 000055b9091b926b RDI: 000000000000000d
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.678100] RBP: 000055b9091b926b R08: 0000000000000000 R09: 0000000000000000
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.678173] R10: 000000000000000d R11: 0000000000000246 R12: 0000000000000000
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.678246] R13: 000055b90a13aa30 R14: 0000000000040000 R15: 0000000000040000
Mar  5 12:27:01 ubuntu-1604-test kernel: [   21.678322] ---[ end trace d006c1f8e03b5e65 ]---

Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky at amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 54 ++++++++++++++++++---------------
 drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h |  2 --
 2 files changed, 29 insertions(+), 27 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
index bf462c5..b0575b6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
@@ -77,8 +77,7 @@ struct ras_manager {
 	struct amdgpu_device *adev;
 	/* debugfs */
 	struct dentry *ent;
-	/* sysfs */
-	struct device_attribute sysfs_attr;
+
 	int attr_inuse;
 
 	/* fs node name */
@@ -374,10 +373,17 @@ static const struct file_operations amdgpu_ras_debugfs_ctrl_ops = {
 	.llseek = default_llseek
 };
 
+static struct ras_sysfs_attr {
+	struct device_attribute sysfs_attrs;
+	struct ras_manager *obj;
+} ras_sysfs_attrs[AMDGPU_RAS_BLOCK__LAST];
+
 static ssize_t amdgpu_ras_sysfs_read(struct device *dev,
 		struct device_attribute *attr, char *buf)
 {
-	struct ras_manager *obj = container_of(attr, struct ras_manager, sysfs_attr);
+	struct ras_sysfs_attr *ras_sysfs_attr = container_of(attr, struct ras_sysfs_attr, sysfs_attrs);
+	struct ras_manager *obj = ras_sysfs_attr->obj;
+
 	struct ras_query_if info = {
 		.head = obj->head,
 	};
@@ -694,10 +700,9 @@ int amdgpu_ras_query_error_count(struct amdgpu_device *adev,
 static ssize_t amdgpu_ras_sysfs_features_read(struct device *dev,
 		struct device_attribute *attr, char *buf)
 {
-	struct amdgpu_ras *con =
-		container_of(attr, struct amdgpu_ras, features_attr);
 	struct drm_device *ddev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = ddev->dev_private;
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	struct ras_common_if head;
 	int ras_block_count = AMDGPU_RAS_BLOCK_COUNT;
 	int i;
@@ -724,11 +729,12 @@ static ssize_t amdgpu_ras_sysfs_features_read(struct device *dev,
 	return s;
 }
 
+static DEVICE_ATTR(features, S_IRUGO, amdgpu_ras_sysfs_features_read, NULL);
+
 static int amdgpu_ras_sysfs_create_feature_node(struct amdgpu_device *adev)
 {
-	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	struct attribute *attrs[] = {
-		&con->features_attr.attr,
+		&dev_attr_features.attr,
 		NULL
 	};
 	struct attribute_group group = {
@@ -736,22 +742,13 @@ static int amdgpu_ras_sysfs_create_feature_node(struct amdgpu_device *adev)
 		.attrs = attrs,
 	};
 
-	con->features_attr = (struct device_attribute) {
-		.attr = {
-			.name = "features",
-			.mode = S_IRUGO,
-		},
-			.show = amdgpu_ras_sysfs_features_read,
-	};
-
 	return sysfs_create_group(&adev->dev->kobj, &group);
 }
 
 static int amdgpu_ras_sysfs_remove_feature_node(struct amdgpu_device *adev)
 {
-	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	struct attribute *attrs[] = {
-		&con->features_attr.attr,
+		&dev_attr_features.attr,
 		NULL
 	};
 	struct attribute_group group = {
@@ -778,17 +775,22 @@ int amdgpu_ras_sysfs_create(struct amdgpu_device *adev,
 			head->sysfs_name,
 			sizeof(obj->fs_data.sysfs_name));
 
-	obj->sysfs_attr = (struct device_attribute){
-		.attr = {
-			.name = obj->fs_data.sysfs_name,
-			.mode = S_IRUGO,
+
+
+	ras_sysfs_attrs[head->head.block] = (struct ras_sysfs_attr){
+		.sysfs_attrs = {
+			.attr = {
+				.name = obj->fs_data.sysfs_name,
+				.mode = S_IRUGO,
+			},
+				.show = amdgpu_ras_sysfs_read,
 		},
-			.show = amdgpu_ras_sysfs_read,
+		.obj = obj
 	};
 
 	if (sysfs_add_file_to_group(&adev->dev->kobj,
-				&obj->sysfs_attr.attr,
-				"ras")) {
+				    &ras_sysfs_attrs[head->head.block].sysfs_attrs.attr,
+				    "ras")) {
 		put_obj(obj);
 		return -EINVAL;
 	}
@@ -807,8 +809,10 @@ int amdgpu_ras_sysfs_remove(struct amdgpu_device *adev,
 		return -EINVAL;
 
 	sysfs_remove_file_from_group(&adev->dev->kobj,
-				&obj->sysfs_attr.attr,
+				&ras_sysfs_attrs[head->block].sysfs_attrs.attr,
 				"ras");
+
+	ras_sysfs_attrs[head->block].obj = NULL;
 	obj->attr_inuse = 0;
 	put_obj(obj);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
index 02cb9a1..b572bae 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
@@ -88,8 +88,6 @@ struct amdgpu_ras {
 	struct dentry *dir;
 	/* debugfs ctrl */
 	struct dentry *ent;
-	/* sysfs */
-	struct device_attribute features_attr;
 	/* block array */
 	struct ras_manager *objs;
 
-- 
2.7.4



More information about the amd-gfx mailing list