RE: [PATCH] drm/amdgpu: Add judgement to avoid infinite loop
It have solution to solve this defect, I am debugging the modifications. -Original Message- From: Zhou1, Tao Sent: Saturday, January 29, 2022 3:54 PM To: Chai, Thomas ; amd-gfx@lists.freedesktop.org Cc: Zhang, Hawking ; Clements, John Subject: RE: [PATCH] drm/amdgpu: Add judgement to avoid infinite loop [AMD Official Use Only] For quick workaround, I agree with the solution. But regarding the root cause, the list is still messed up. Can we make ras_list to be a global variable across all cards, and add list empty check (or add a flag to indicate the register status of ras block) before list add to avoid redundant register? Regards, Tao > -Original Message- > From: Chai, Thomas > Sent: Saturday, January 29, 2022 11:53 AM > To: amd-gfx@lists.freedesktop.org > Cc: Chai, Thomas ; Zhang, Hawking > ; Zhou1, Tao ; Clements, > John ; Chai, Thomas > Subject: [PATCH] drm/amdgpu: Add judgement to avoid infinite loop > > 1. The infinite loop causing soft lock occurs on multiple amdgpu cards >supporting ras feature. > 2. This a workaround patch. It is valid for multiple amdgpu cards of the >same type. > 3. The root cause is that each GPU card device has a separate .ras_list >link header, but the instance and linked list node of each ras block >are unique. When each device is initialized, each ras instance will >repeatedly add link node to the device every time. In this way, only >the .ras_list of the last initialized device is completely correct. >the .ras_list->prev and .ras_list->next of the device initialzied >before can still point to the correct ras instance, but the prev >pointer and next pointer of the pointed ras instance both point to >the last initialized device's .ras_ list instead of the beginning >.ras_ list. When using list_for_each_entry_safe searches for >non-existent Ras nodes on devices other than the last device, the >last ras instance next pointer cannot always be equal to the >beginning .ras_list, so that the loop cannot be terminated, the >program enters a infinite loop. > BTW: Since the data and initialization process of each card are the same, > the link list between ras instances will not be destroyed every time > the device is initialized. > 4. The soft locked logs are as follows: > [ 262.165690] CPU: 93 PID: 758 Comm: kworker/93:1 Tainted: G OE > 5.13.0-27-generic #29~20.04.1-Ubuntu > [ 262.165695] Hardware name: Supermicro AS -4124GS-TNR/H12DSG-O-CPU, > BIOS T20200717143848 07/17/2020 [ 262.165698] Workqueue: events > amdgpu_ras_do_recovery [amdgpu] [ 262.165980] RIP: > 0010:amdgpu_ras_get_ras_block+0x86/0xd0 [amdgpu] [ 262.166239] Code: > 68 > d8 4c 8d 71 d8 48 39 c3 74 54 49 8b 45 38 48 85 c0 74 32 44 89 fa 44 > 89 e6 4c 89 ef e8 82 e4 9b dc 85 c0 74 3c 49 8b 46 28 <49> 8d 56 28 4d > 89 f5 48 83 e8 28 48 > 39 d3 74 25 49 89 c6 49 8b 45 [ 262.166243] RSP: > 0018:ac908fa87d80 > EFLAGS: 0202 [ 262.166247] RAX: c1394248 RBX: > 91e4ab8d6e20 > RCX: c1394248 [ 262.166249] RDX: 91e4aa356e20 RSI: > 000e RDI: 91e4ab8c [ 262.166252] RBP: > ac908fa87da8 R08: 0007 R09: 0001 [ > 262.166254] R10: 91e4930b64ec R11: R12: > 000e [ 262.166256] R13: 91e4aa356df8 R14: > c1394320 > R15: 0003 [ 262.166258] FS: () > GS:92238fb4() knlGS: [ 262.166261] CS: > 0010 > DS: ES: CR0: 80050033 [ 262.166264] CR2: > 0001004865d0 CR3: 00406d796000 CR4: 00350ee0 [ > 262.166267] Call Trace: > [ 262.166272] amdgpu_ras_do_recovery+0x130/0x290 [amdgpu] [ > 262.166529] ? psi_task_switch+0xd2/0x250 [ 262.166537] ? > __switch_to+0x11d/0x460 [ 262.166542] ? __switch_to_asm+0x36/0x70 [ > 262.166549] process_one_work+0x220/0x3c0 [ 262.166556] > worker_thread+0x4d/0x3f0 [ 262.166560] ? > process_one_work+0x3c0/0x3c0 [ 262.166563] kthread+0x12b/0x150 [ > 262.166568] ? > set_kthread_struct+0x40/0x40 [ 262.166571] ret_from_fork+0x22/0x30 > > Signed-off-by: yipechai > --- > drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 4 > 1 file changed, 4 insertions(+) > > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c > b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c > index d4e07d0acb66..3d533ef0783d 100644 > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c > @@ -884,6 +884,7 @@ static int amdgpu_ras_block_match_default(struct > amdgpu_ras_block_object *block_ static struct amdgpu_ras_block_object > *amdgpu_ras_get_ras_block(struct amdgpu_device *adev, >
RE: [PATCH] drm/amdgpu: Add judgement to avoid infinite loop
[AMD Official Use Only] For quick workaround, I agree with the solution. But regarding the root cause, the list is still messed up. Can we make ras_list to be a global variable across all cards, and add list empty check (or add a flag to indicate the register status of ras block) before list add to avoid redundant register? Regards, Tao > -Original Message- > From: Chai, Thomas > Sent: Saturday, January 29, 2022 11:53 AM > To: amd-gfx@lists.freedesktop.org > Cc: Chai, Thomas ; Zhang, Hawking > ; Zhou1, Tao ; Clements, > John ; Chai, Thomas > Subject: [PATCH] drm/amdgpu: Add judgement to avoid infinite loop > > 1. The infinite loop causing soft lock occurs on multiple amdgpu cards >supporting ras feature. > 2. This a workaround patch. It is valid for multiple amdgpu cards of the >same type. > 3. The root cause is that each GPU card device has a separate .ras_list >link header, but the instance and linked list node of each ras block >are unique. When each device is initialized, each ras instance will >repeatedly add link node to the device every time. In this way, only >the .ras_list of the last initialized device is completely correct. >the .ras_list->prev and .ras_list->next of the device initialzied >before can still point to the correct ras instance, but the prev >pointer and next pointer of the pointed ras instance both point to >the last initialized device's .ras_ list instead of the beginning >.ras_ list. When using list_for_each_entry_safe searches for >non-existent Ras nodes on devices other than the last device, the >last ras instance next pointer cannot always be equal to the >beginning .ras_list, so that the loop cannot be terminated, the >program enters a infinite loop. > BTW: Since the data and initialization process of each card are the same, > the link list between ras instances will not be destroyed every time > the device is initialized. > 4. The soft locked logs are as follows: > [ 262.165690] CPU: 93 PID: 758 Comm: kworker/93:1 Tainted: G OE > 5.13.0-27-generic #29~20.04.1-Ubuntu > [ 262.165695] Hardware name: Supermicro AS -4124GS-TNR/H12DSG-O-CPU, > BIOS T20200717143848 07/17/2020 [ 262.165698] Workqueue: events > amdgpu_ras_do_recovery [amdgpu] [ 262.165980] RIP: > 0010:amdgpu_ras_get_ras_block+0x86/0xd0 [amdgpu] [ 262.166239] Code: 68 > d8 4c 8d 71 d8 48 39 c3 74 54 49 8b 45 38 48 85 c0 74 32 44 89 fa 44 89 e6 4c > 89 > ef e8 82 e4 9b dc 85 c0 74 3c 49 8b 46 28 <49> 8d 56 28 4d 89 f5 48 83 e8 28 > 48 > 39 d3 74 25 49 89 c6 49 8b 45 [ 262.166243] RSP: 0018:ac908fa87d80 > EFLAGS: 0202 [ 262.166247] RAX: c1394248 RBX: 91e4ab8d6e20 > RCX: c1394248 [ 262.166249] RDX: 91e4aa356e20 RSI: > 000e RDI: 91e4ab8c [ 262.166252] RBP: > ac908fa87da8 R08: 0007 R09: 0001 > [ 262.166254] R10: 91e4930b64ec R11: R12: > 000e [ 262.166256] R13: 91e4aa356df8 R14: c1394320 > R15: 0003 [ 262.166258] FS: () > GS:92238fb4() knlGS: [ 262.166261] CS: 0010 > DS: ES: CR0: 80050033 [ 262.166264] CR2: > 0001004865d0 CR3: 00406d796000 CR4: 00350ee0 > [ 262.166267] Call Trace: > [ 262.166272] amdgpu_ras_do_recovery+0x130/0x290 [amdgpu] > [ 262.166529] ? psi_task_switch+0xd2/0x250 [ 262.166537] ? > __switch_to+0x11d/0x460 [ 262.166542] ? __switch_to_asm+0x36/0x70 > [ 262.166549] process_one_work+0x220/0x3c0 [ 262.166556] > worker_thread+0x4d/0x3f0 [ 262.166560] ? process_one_work+0x3c0/0x3c0 > [ 262.166563] kthread+0x12b/0x150 [ 262.166568] ? > set_kthread_struct+0x40/0x40 [ 262.166571] ret_from_fork+0x22/0x30 > > Signed-off-by: yipechai > --- > drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 4 > 1 file changed, 4 insertions(+) > > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c > b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c > index d4e07d0acb66..3d533ef0783d 100644 > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c > @@ -884,6 +884,7 @@ static int amdgpu_ras_block_match_default(struct > amdgpu_ras_block_object *block_ static struct amdgpu_ras_block_object > *amdgpu_ras_get_ras_block(struct amdgpu_device *adev, > enum amdgpu_ras_block block, > uint32_t sub_block_index) { > + int loop_cnt = 0; > struct amdgpu_ras_block_object *obj, *tmp; > > if (block >= AMDGPU_RAS_BLOCK__LAST) > @@ -900,6 +901,9 @@ static struct amdgpu_ras_block_object > *amdgpu_ras_get_ras_block(struct amdgpu_de > if (amdgpu_ras_block_match_default(obj, block) == 0) > return obj; > } > + > + if (++loop_cnt >= AMDGPU_RAS_BLOCK__LAST) > + break; > } > > return NULL; > -- > 2.25.1
RE: [PATCH] drm/amdgpu: Add judgement to avoid infinite loop
OK -Original Message- From: Chen, Guchun Sent: Saturday, January 29, 2022 12:02 PM To: Chai, Thomas ; amd-gfx@lists.freedesktop.org Cc: Zhou1, Tao ; Zhang, Hawking ; Clements, John ; Chai, Thomas ; Chai, Thomas Subject: RE: [PATCH] drm/amdgpu: Add judgement to avoid infinite loop [Public] Please add a Fixes tag, as it should fix a regression from former patch. Regards, Guchun -Original Message- From: amd-gfx On Behalf Of yipechai Sent: Saturday, January 29, 2022 11:53 AM To: amd-gfx@lists.freedesktop.org Cc: Zhou1, Tao ; Zhang, Hawking ; Clements, John ; Chai, Thomas ; Chai, Thomas Subject: [PATCH] drm/amdgpu: Add judgement to avoid infinite loop 1. The infinite loop causing soft lock occurs on multiple amdgpu cards supporting ras feature. 2. This a workaround patch. It is valid for multiple amdgpu cards of the same type. 3. The root cause is that each GPU card device has a separate .ras_list link header, but the instance and linked list node of each ras block are unique. When each device is initialized, each ras instance will repeatedly add link node to the device every time. In this way, only the .ras_list of the last initialized device is completely correct. the .ras_list->prev and .ras_list->next of the device initialzied before can still point to the correct ras instance, but the prev pointer and next pointer of the pointed ras instance both point to the last initialized device's .ras_ list instead of the beginning .ras_ list. When using list_for_each_entry_safe searches for non-existent Ras nodes on devices other than the last device, the last ras instance next pointer cannot always be equal to the beginning .ras_list, so that the loop cannot be terminated, the program enters a infinite loop. BTW: Since the data and initialization process of each card are the same, the link list between ras instances will not be destroyed every time the device is initialized. 4. The soft locked logs are as follows: [ 262.165690] CPU: 93 PID: 758 Comm: kworker/93:1 Tainted: G OE 5.13.0-27-generic #29~20.04.1-Ubuntu [ 262.165695] Hardware name: Supermicro AS -4124GS-TNR/H12DSG-O-CPU, BIOS T20200717143848 07/17/2020 [ 262.165698] Workqueue: events amdgpu_ras_do_recovery [amdgpu] [ 262.165980] RIP: 0010:amdgpu_ras_get_ras_block+0x86/0xd0 [amdgpu] [ 262.166239] Code: 68 d8 4c 8d 71 d8 48 39 c3 74 54 49 8b 45 38 48 85 c0 74 32 44 89 fa 44 89 e6 4c 89 ef e8 82 e4 9b dc 85 c0 74 3c 49 8b 46 28 <49> 8d 56 28 4d 89 f5 48 83 e8 28 48 39 d3 74 25 49 89 c6 49 8b 45 [ 262.166243] RSP: 0018:ac908fa87d80 EFLAGS: 0202 [ 262.166247] RAX: c1394248 RBX: 91e4ab8d6e20 RCX: c1394248 [ 262.166249] RDX: 91e4aa356e20 RSI: 000e RDI: 91e4ab8c [ 262.166252] RBP: ac908fa87da8 R08: 0007 R09: 0001 [ 262.166254] R10: 91e4930b64ec R11: R12: 000e [ 262.166256] R13: 91e4aa356df8 R14: c1394320 R15: 0003 [ 262.166258] FS: () GS:92238fb4() knlGS: [ 262.166261] CS: 0010 DS: ES: CR0: 80050033 [ 262.166264] CR2: 0001004865d0 CR3: 00406d796000 CR4: 00350ee0 [ 262.166267] Call Trace: [ 262.166272] amdgpu_ras_do_recovery+0x130/0x290 [amdgpu] [ 262.166529] ? psi_task_switch+0xd2/0x250 [ 262.166537] ? __switch_to+0x11d/0x460 [ 262.166542] ? __switch_to_asm+0x36/0x70 [ 262.166549] process_one_work+0x220/0x3c0 [ 262.166556] worker_thread+0x4d/0x3f0 [ 262.166560] ? process_one_work+0x3c0/0x3c0 [ 262.166563] kthread+0x12b/0x150 [ 262.166568] ? set_kthread_struct+0x40/0x40 [ 262.166571] ret_from_fork+0x22/0x30 Signed-off-by: yipechai --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 4 1 file changed, 4 insertions(+) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index d4e07d0acb66..3d533ef0783d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -884,6 +884,7 @@ static int amdgpu_ras_block_match_default(struct amdgpu_ras_block_object *block_ static struct amdgpu_ras_block_object *amdgpu_ras_get_ras_block(struct amdgpu_device *adev, enum amdgpu_ras_block block, uint32_t sub_block_index) { + int loop_cnt = 0; struct amdgpu_ras_block_object *obj, *tmp; if (block >= AMDGPU_RAS_BLOCK__LAST) @@ -900,6 +901,9 @@ static struct amdgpu_ras_block_object *amdgpu_ras_get_ras_block(struct amdgpu_de if (amdgpu_ras_block_match_default(obj, block) == 0) return obj; } + + if (++loop_cnt >= AMDGPU_RAS_BLOCK__LAST) + break; } return NULL; -- 2.25.1
RE: [PATCH] drm/amdgpu: Add judgement to avoid infinite loop
[AMD Official Use Only] Reviewed-by: John Clements -Original Message- From: Chai, Thomas Sent: Saturday, January 29, 2022 11:53 AM To: amd-gfx@lists.freedesktop.org Cc: Chai, Thomas ; Zhang, Hawking ; Zhou1, Tao ; Clements, John ; Chai, Thomas Subject: [PATCH] drm/amdgpu: Add judgement to avoid infinite loop 1. The infinite loop causing soft lock occurs on multiple amdgpu cards supporting ras feature. 2. This a workaround patch. It is valid for multiple amdgpu cards of the same type. 3. The root cause is that each GPU card device has a separate .ras_list link header, but the instance and linked list node of each ras block are unique. When each device is initialized, each ras instance will repeatedly add link node to the device every time. In this way, only the .ras_list of the last initialized device is completely correct. the .ras_list->prev and .ras_list->next of the device initialzied before can still point to the correct ras instance, but the prev pointer and next pointer of the pointed ras instance both point to the last initialized device's .ras_ list instead of the beginning .ras_ list. When using list_for_each_entry_safe searches for non-existent Ras nodes on devices other than the last device, the last ras instance next pointer cannot always be equal to the beginning .ras_list, so that the loop cannot be terminated, the program enters a infinite loop. BTW: Since the data and initialization process of each card are the same, the link list between ras instances will not be destroyed every time the device is initialized. 4. The soft locked logs are as follows: [ 262.165690] CPU: 93 PID: 758 Comm: kworker/93:1 Tainted: G OE 5.13.0-27-generic #29~20.04.1-Ubuntu [ 262.165695] Hardware name: Supermicro AS -4124GS-TNR/H12DSG-O-CPU, BIOS T20200717143848 07/17/2020 [ 262.165698] Workqueue: events amdgpu_ras_do_recovery [amdgpu] [ 262.165980] RIP: 0010:amdgpu_ras_get_ras_block+0x86/0xd0 [amdgpu] [ 262.166239] Code: 68 d8 4c 8d 71 d8 48 39 c3 74 54 49 8b 45 38 48 85 c0 74 32 44 89 fa 44 89 e6 4c 89 ef e8 82 e4 9b dc 85 c0 74 3c 49 8b 46 28 <49> 8d 56 28 4d 89 f5 48 83 e8 28 48 39 d3 74 25 49 89 c6 49 8b 45 [ 262.166243] RSP: 0018:ac908fa87d80 EFLAGS: 0202 [ 262.166247] RAX: c1394248 RBX: 91e4ab8d6e20 RCX: c1394248 [ 262.166249] RDX: 91e4aa356e20 RSI: 000e RDI: 91e4ab8c [ 262.166252] RBP: ac908fa87da8 R08: 0007 R09: 0001 [ 262.166254] R10: 91e4930b64ec R11: R12: 000e [ 262.166256] R13: 91e4aa356df8 R14: c1394320 R15: 0003 [ 262.166258] FS: () GS:92238fb4() knlGS: [ 262.166261] CS: 0010 DS: ES: CR0: 80050033 [ 262.166264] CR2: 0001004865d0 CR3: 00406d796000 CR4: 00350ee0 [ 262.166267] Call Trace: [ 262.166272] amdgpu_ras_do_recovery+0x130/0x290 [amdgpu] [ 262.166529] ? psi_task_switch+0xd2/0x250 [ 262.166537] ? __switch_to+0x11d/0x460 [ 262.166542] ? __switch_to_asm+0x36/0x70 [ 262.166549] process_one_work+0x220/0x3c0 [ 262.166556] worker_thread+0x4d/0x3f0 [ 262.166560] ? process_one_work+0x3c0/0x3c0 [ 262.166563] kthread+0x12b/0x150 [ 262.166568] ? set_kthread_struct+0x40/0x40 [ 262.166571] ret_from_fork+0x22/0x30 Signed-off-by: yipechai --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 4 1 file changed, 4 insertions(+) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index d4e07d0acb66..3d533ef0783d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -884,6 +884,7 @@ static int amdgpu_ras_block_match_default(struct amdgpu_ras_block_object *block_ static struct amdgpu_ras_block_object *amdgpu_ras_get_ras_block(struct amdgpu_device *adev, enum amdgpu_ras_block block, uint32_t sub_block_index) { + int loop_cnt = 0; struct amdgpu_ras_block_object *obj, *tmp; if (block >= AMDGPU_RAS_BLOCK__LAST) @@ -900,6 +901,9 @@ static struct amdgpu_ras_block_object *amdgpu_ras_get_ras_block(struct amdgpu_de if (amdgpu_ras_block_match_default(obj, block) == 0) return obj; } + + if (++loop_cnt >= AMDGPU_RAS_BLOCK__LAST) + break; } return NULL; -- 2.25.1
RE: [PATCH] drm/amdgpu: Add judgement to avoid infinite loop
[Public] Please add a Fixes tag, as it should fix a regression from former patch. Regards, Guchun -Original Message- From: amd-gfx On Behalf Of yipechai Sent: Saturday, January 29, 2022 11:53 AM To: amd-gfx@lists.freedesktop.org Cc: Zhou1, Tao ; Zhang, Hawking ; Clements, John ; Chai, Thomas ; Chai, Thomas Subject: [PATCH] drm/amdgpu: Add judgement to avoid infinite loop 1. The infinite loop causing soft lock occurs on multiple amdgpu cards supporting ras feature. 2. This a workaround patch. It is valid for multiple amdgpu cards of the same type. 3. The root cause is that each GPU card device has a separate .ras_list link header, but the instance and linked list node of each ras block are unique. When each device is initialized, each ras instance will repeatedly add link node to the device every time. In this way, only the .ras_list of the last initialized device is completely correct. the .ras_list->prev and .ras_list->next of the device initialzied before can still point to the correct ras instance, but the prev pointer and next pointer of the pointed ras instance both point to the last initialized device's .ras_ list instead of the beginning .ras_ list. When using list_for_each_entry_safe searches for non-existent Ras nodes on devices other than the last device, the last ras instance next pointer cannot always be equal to the beginning .ras_list, so that the loop cannot be terminated, the program enters a infinite loop. BTW: Since the data and initialization process of each card are the same, the link list between ras instances will not be destroyed every time the device is initialized. 4. The soft locked logs are as follows: [ 262.165690] CPU: 93 PID: 758 Comm: kworker/93:1 Tainted: G OE 5.13.0-27-generic #29~20.04.1-Ubuntu [ 262.165695] Hardware name: Supermicro AS -4124GS-TNR/H12DSG-O-CPU, BIOS T20200717143848 07/17/2020 [ 262.165698] Workqueue: events amdgpu_ras_do_recovery [amdgpu] [ 262.165980] RIP: 0010:amdgpu_ras_get_ras_block+0x86/0xd0 [amdgpu] [ 262.166239] Code: 68 d8 4c 8d 71 d8 48 39 c3 74 54 49 8b 45 38 48 85 c0 74 32 44 89 fa 44 89 e6 4c 89 ef e8 82 e4 9b dc 85 c0 74 3c 49 8b 46 28 <49> 8d 56 28 4d 89 f5 48 83 e8 28 48 39 d3 74 25 49 89 c6 49 8b 45 [ 262.166243] RSP: 0018:ac908fa87d80 EFLAGS: 0202 [ 262.166247] RAX: c1394248 RBX: 91e4ab8d6e20 RCX: c1394248 [ 262.166249] RDX: 91e4aa356e20 RSI: 000e RDI: 91e4ab8c [ 262.166252] RBP: ac908fa87da8 R08: 0007 R09: 0001 [ 262.166254] R10: 91e4930b64ec R11: R12: 000e [ 262.166256] R13: 91e4aa356df8 R14: c1394320 R15: 0003 [ 262.166258] FS: () GS:92238fb4() knlGS: [ 262.166261] CS: 0010 DS: ES: CR0: 80050033 [ 262.166264] CR2: 0001004865d0 CR3: 00406d796000 CR4: 00350ee0 [ 262.166267] Call Trace: [ 262.166272] amdgpu_ras_do_recovery+0x130/0x290 [amdgpu] [ 262.166529] ? psi_task_switch+0xd2/0x250 [ 262.166537] ? __switch_to+0x11d/0x460 [ 262.166542] ? __switch_to_asm+0x36/0x70 [ 262.166549] process_one_work+0x220/0x3c0 [ 262.166556] worker_thread+0x4d/0x3f0 [ 262.166560] ? process_one_work+0x3c0/0x3c0 [ 262.166563] kthread+0x12b/0x150 [ 262.166568] ? set_kthread_struct+0x40/0x40 [ 262.166571] ret_from_fork+0x22/0x30 Signed-off-by: yipechai --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 4 1 file changed, 4 insertions(+) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index d4e07d0acb66..3d533ef0783d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -884,6 +884,7 @@ static int amdgpu_ras_block_match_default(struct amdgpu_ras_block_object *block_ static struct amdgpu_ras_block_object *amdgpu_ras_get_ras_block(struct amdgpu_device *adev, enum amdgpu_ras_block block, uint32_t sub_block_index) { + int loop_cnt = 0; struct amdgpu_ras_block_object *obj, *tmp; if (block >= AMDGPU_RAS_BLOCK__LAST) @@ -900,6 +901,9 @@ static struct amdgpu_ras_block_object *amdgpu_ras_get_ras_block(struct amdgpu_de if (amdgpu_ras_block_match_default(obj, block) == 0) return obj; } + + if (++loop_cnt >= AMDGPU_RAS_BLOCK__LAST) + break; } return NULL; -- 2.25.1
[PATCH] drm/amdgpu: Add judgement to avoid infinite loop
1. The infinite loop causing soft lock occurs on multiple amdgpu cards supporting ras feature. 2. This a workaround patch. It is valid for multiple amdgpu cards of the same type. 3. The root cause is that each GPU card device has a separate .ras_list link header, but the instance and linked list node of each ras block are unique. When each device is initialized, each ras instance will repeatedly add link node to the device every time. In this way, only the .ras_list of the last initialized device is completely correct. the .ras_list->prev and .ras_list->next of the device initialzied before can still point to the correct ras instance, but the prev pointer and next pointer of the pointed ras instance both point to the last initialized device's .ras_ list instead of the beginning .ras_ list. When using list_for_each_entry_safe searches for non-existent Ras nodes on devices other than the last device, the last ras instance next pointer cannot always be equal to the beginning .ras_list, so that the loop cannot be terminated, the program enters a infinite loop. BTW: Since the data and initialization process of each card are the same, the link list between ras instances will not be destroyed every time the device is initialized. 4. The soft locked logs are as follows: [ 262.165690] CPU: 93 PID: 758 Comm: kworker/93:1 Tainted: G OE 5.13.0-27-generic #29~20.04.1-Ubuntu [ 262.165695] Hardware name: Supermicro AS -4124GS-TNR/H12DSG-O-CPU, BIOS T20200717143848 07/17/2020 [ 262.165698] Workqueue: events amdgpu_ras_do_recovery [amdgpu] [ 262.165980] RIP: 0010:amdgpu_ras_get_ras_block+0x86/0xd0 [amdgpu] [ 262.166239] Code: 68 d8 4c 8d 71 d8 48 39 c3 74 54 49 8b 45 38 48 85 c0 74 32 44 89 fa 44 89 e6 4c 89 ef e8 82 e4 9b dc 85 c0 74 3c 49 8b 46 28 <49> 8d 56 28 4d 89 f5 48 83 e8 28 48 39 d3 74 25 49 89 c6 49 8b 45 [ 262.166243] RSP: 0018:ac908fa87d80 EFLAGS: 0202 [ 262.166247] RAX: c1394248 RBX: 91e4ab8d6e20 RCX: c1394248 [ 262.166249] RDX: 91e4aa356e20 RSI: 000e RDI: 91e4ab8c [ 262.166252] RBP: ac908fa87da8 R08: 0007 R09: 0001 [ 262.166254] R10: 91e4930b64ec R11: R12: 000e [ 262.166256] R13: 91e4aa356df8 R14: c1394320 R15: 0003 [ 262.166258] FS: () GS:92238fb4() knlGS: [ 262.166261] CS: 0010 DS: ES: CR0: 80050033 [ 262.166264] CR2: 0001004865d0 CR3: 00406d796000 CR4: 00350ee0 [ 262.166267] Call Trace: [ 262.166272] amdgpu_ras_do_recovery+0x130/0x290 [amdgpu] [ 262.166529] ? psi_task_switch+0xd2/0x250 [ 262.166537] ? __switch_to+0x11d/0x460 [ 262.166542] ? __switch_to_asm+0x36/0x70 [ 262.166549] process_one_work+0x220/0x3c0 [ 262.166556] worker_thread+0x4d/0x3f0 [ 262.166560] ? process_one_work+0x3c0/0x3c0 [ 262.166563] kthread+0x12b/0x150 [ 262.166568] ? set_kthread_struct+0x40/0x40 [ 262.166571] ret_from_fork+0x22/0x30 Signed-off-by: yipechai --- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 4 1 file changed, 4 insertions(+) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index d4e07d0acb66..3d533ef0783d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -884,6 +884,7 @@ static int amdgpu_ras_block_match_default(struct amdgpu_ras_block_object *block_ static struct amdgpu_ras_block_object *amdgpu_ras_get_ras_block(struct amdgpu_device *adev, enum amdgpu_ras_block block, uint32_t sub_block_index) { + int loop_cnt = 0; struct amdgpu_ras_block_object *obj, *tmp; if (block >= AMDGPU_RAS_BLOCK__LAST) @@ -900,6 +901,9 @@ static struct amdgpu_ras_block_object *amdgpu_ras_get_ras_block(struct amdgpu_de if (amdgpu_ras_block_match_default(obj, block) == 0) return obj; } + + if (++loop_cnt >= AMDGPU_RAS_BLOCK__LAST) + break; } return NULL; -- 2.25.1