* [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini
@ 2022-08-12 9:13 YiPeng Chai
2022-08-12 9:13 ` [PATCH 2/2] drm/amdgpu: fix hive reference leak when adding xgmi device YiPeng Chai
2022-08-15 7:02 ` [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini Chai, Thomas
0 siblings, 2 replies; 7+ messages in thread
From: YiPeng Chai @ 2022-08-12 9:13 UTC (permalink / raw)
To: amd-gfx; +Cc: YiPeng Chai, Hawking.Zhang, KevinYang.Wang, yipechai
The amdgpu_xgmi_remove_device function will send unload command
to psp through psp ring to terminate xgmi, but psp ring has been
destroyed in psp_hw_fini.
Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index c84fdef0ac45..2445255bbf01 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -2787,6 +2787,9 @@ static int amdgpu_device_ip_fini_early(struct amdgpu_device *adev)
amdgpu_amdkfd_suspend(adev, false);
+ if (adev->gmc.xgmi.num_physical_nodes > 1)
+ amdgpu_xgmi_remove_device(adev);
+
/* Workaroud for ASICs need to disable SMC first */
amdgpu_device_smu_fini_early(adev);
@@ -2830,9 +2833,6 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev)
if (amdgpu_sriov_vf(adev) && adev->virt.ras_init_done)
amdgpu_virt_release_ras_err_handler_data(adev);
- if (adev->gmc.xgmi.num_physical_nodes > 1)
- amdgpu_xgmi_remove_device(adev);
-
amdgpu_amdkfd_device_fini_sw(adev);
for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
--
2.25.1
^ permalink raw reply related [flat|nested] 7+ messages in thread
* [PATCH 2/2] drm/amdgpu: fix hive reference leak when adding xgmi device
2022-08-12 9:13 [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini YiPeng Chai
@ 2022-08-12 9:13 ` YiPeng Chai
2022-08-15 7:02 ` [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini Chai, Thomas
1 sibling, 0 replies; 7+ messages in thread
From: YiPeng Chai @ 2022-08-12 9:13 UTC (permalink / raw)
To: amd-gfx; +Cc: YiPeng Chai, Hawking.Zhang, KevinYang.Wang, yipechai
Only amdgpu_get_xgmi_hive but no amdgpu_put_xgmi_hive
which will leak the hive reference.
Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 2445255bbf01..4cdc50873621 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -2456,12 +2456,14 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev)
if (!hive->reset_domain ||
!amdgpu_reset_get_reset_domain(hive->reset_domain)) {
r = -ENOENT;
+ amdgpu_put_xgmi_hive(hive);
goto init_failed;
}
/* Drop the early temporary reset domain we created for device */
amdgpu_reset_put_reset_domain(adev->reset_domain);
adev->reset_domain = hive->reset_domain;
+ amdgpu_put_xgmi_hive(hive);
}
}
--
2.25.1
^ permalink raw reply related [flat|nested] 7+ messages in thread
* RE: [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini
2022-08-12 9:13 [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini YiPeng Chai
2022-08-12 9:13 ` [PATCH 2/2] drm/amdgpu: fix hive reference leak when adding xgmi device YiPeng Chai
@ 2022-08-15 7:02 ` Chai, Thomas
2022-08-16 3:49 ` Zhang, Hawking
2022-08-25 1:03 ` Zhang, Hawking
1 sibling, 2 replies; 7+ messages in thread
From: Chai, Thomas @ 2022-08-15 7:02 UTC (permalink / raw)
To: amd-gfx; +Cc: Wang, Yang(Kevin), Zhang, Hawking
[AMD Official Use Only - General]
Ping on this series.
-----Original Message-----
From: Chai, Thomas <YiPeng.Chai@amd.com>
Sent: Friday, August 12, 2022 5:13 PM
To: amd-gfx@lists.freedesktop.org
Cc: Chai, Thomas <YiPeng.Chai@amd.com>; Zhang, Hawking <Hawking.Zhang@amd.com>; Wang, Yang(Kevin) <KevinYang.Wang@amd.com>; Chai, Thomas <YiPeng.Chai@amd.com>
Subject: [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini
The amdgpu_xgmi_remove_device function will send unload command to psp through psp ring to terminate xgmi, but psp ring has been destroyed in psp_hw_fini.
Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index c84fdef0ac45..2445255bbf01 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -2787,6 +2787,9 @@ static int amdgpu_device_ip_fini_early(struct amdgpu_device *adev)
amdgpu_amdkfd_suspend(adev, false);
+ if (adev->gmc.xgmi.num_physical_nodes > 1)
+ amdgpu_xgmi_remove_device(adev);
+
/* Workaroud for ASICs need to disable SMC first */
amdgpu_device_smu_fini_early(adev);
@@ -2830,9 +2833,6 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev)
if (amdgpu_sriov_vf(adev) && adev->virt.ras_init_done)
amdgpu_virt_release_ras_err_handler_data(adev);
- if (adev->gmc.xgmi.num_physical_nodes > 1)
- amdgpu_xgmi_remove_device(adev);
-
amdgpu_amdkfd_device_fini_sw(adev);
for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
--
2.25.1
^ permalink raw reply related [flat|nested] 7+ messages in thread
* RE: [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini
2022-08-15 7:02 ` [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini Chai, Thomas
@ 2022-08-16 3:49 ` Zhang, Hawking
2022-08-16 3:50 ` Zhang, Hawking
2022-08-25 1:03 ` Zhang, Hawking
1 sibling, 1 reply; 7+ messages in thread
From: Zhang, Hawking @ 2022-08-16 3:49 UTC (permalink / raw)
To: Chai, Thomas, amd-gfx; +Cc: Wang, Yang(Kevin)
[AMD Official Use Only - General]
Alternatively, we might be better split xgmi ta terminate from xgmi_remove_device. In psp_hw_fini, check ta->fw and num_of_physical_mode to terminate xgmi ta. and make amdgpu_xgmi_remove_device only deal with software fini, like add_device.
Regards,
Hawking
-----Original Message-----
From: Chai, Thomas <YiPeng.Chai@amd.com>
Sent: Monday, August 15, 2022 15:03
To: amd-gfx@lists.freedesktop.org
Cc: Zhang, Hawking <Hawking.Zhang@amd.com>; Wang, Yang(Kevin) <KevinYang.Wang@amd.com>
Subject: RE: [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini
[AMD Official Use Only - General]
Ping on this series.
-----Original Message-----
From: Chai, Thomas <YiPeng.Chai@amd.com>
Sent: Friday, August 12, 2022 5:13 PM
To: amd-gfx@lists.freedesktop.org
Cc: Chai, Thomas <YiPeng.Chai@amd.com>; Zhang, Hawking <Hawking.Zhang@amd.com>; Wang, Yang(Kevin) <KevinYang.Wang@amd.com>; Chai, Thomas <YiPeng.Chai@amd.com>
Subject: [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini
The amdgpu_xgmi_remove_device function will send unload command to psp through psp ring to terminate xgmi, but psp ring has been destroyed in psp_hw_fini.
Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index c84fdef0ac45..2445255bbf01 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -2787,6 +2787,9 @@ static int amdgpu_device_ip_fini_early(struct amdgpu_device *adev)
amdgpu_amdkfd_suspend(adev, false);
+ if (adev->gmc.xgmi.num_physical_nodes > 1)
+ amdgpu_xgmi_remove_device(adev);
+
/* Workaroud for ASICs need to disable SMC first */
amdgpu_device_smu_fini_early(adev);
@@ -2830,9 +2833,6 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev)
if (amdgpu_sriov_vf(adev) && adev->virt.ras_init_done)
amdgpu_virt_release_ras_err_handler_data(adev);
- if (adev->gmc.xgmi.num_physical_nodes > 1)
- amdgpu_xgmi_remove_device(adev);
-
amdgpu_amdkfd_device_fini_sw(adev);
for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
--
2.25.1
^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini
2022-08-16 3:49 ` Zhang, Hawking
@ 2022-08-16 3:50 ` Zhang, Hawking
2022-08-16 5:24 ` Chai, Thomas
0 siblings, 1 reply; 7+ messages in thread
From: Zhang, Hawking @ 2022-08-16 3:50 UTC (permalink / raw)
To: Chai, Thomas, amd-gfx; +Cc: Wang, Yang(Kevin)
[-- Attachment #1: Type: text/plain, Size: 2869 bytes --]
Fixed typo
Regards,
Hawking
From: Zhang, Hawking <Hawking.Zhang@amd.com>
Date: Tuesday, August 16, 2022 at 11:49
To: Chai, Thomas <YiPeng.Chai@amd.com>, amd-gfx@lists.freedesktop.org <amd-gfx@lists.freedesktop.org>
Cc: Wang, Yang(Kevin) <KevinYang.Wang@amd.com>
Subject: RE: [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini
[AMD Official Use Only - General]
Alternatively, it might be better split xgmi ta terminate from xgmi_remove_device. In psp_hw_fini, check ta->fw and num_of_physical_mode to terminate xgmi ta. and make amdgpu_xgmi_remove_device only deal with software fini, like add_device.
Regards,
Hawking
-----Original Message-----
From: Chai, Thomas <YiPeng.Chai@amd.com>
Sent: Monday, August 15, 2022 15:03
To: amd-gfx@lists.freedesktop.org
Cc: Zhang, Hawking <Hawking.Zhang@amd.com>; Wang, Yang(Kevin) <KevinYang.Wang@amd.com>
Subject: RE: [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini
[AMD Official Use Only - General]
Ping on this series.
-----Original Message-----
From: Chai, Thomas <YiPeng.Chai@amd.com>
Sent: Friday, August 12, 2022 5:13 PM
To: amd-gfx@lists.freedesktop.org
Cc: Chai, Thomas <YiPeng.Chai@amd.com>; Zhang, Hawking <Hawking.Zhang@amd.com>; Wang, Yang(Kevin) <KevinYang.Wang@amd.com>; Chai, Thomas <YiPeng.Chai@amd.com>
Subject: [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini
The amdgpu_xgmi_remove_device function will send unload command to psp through psp ring to terminate xgmi, but psp ring has been destroyed in psp_hw_fini.
Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index c84fdef0ac45..2445255bbf01 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -2787,6 +2787,9 @@ static int amdgpu_device_ip_fini_early(struct amdgpu_device *adev)
amdgpu_amdkfd_suspend(adev, false);
+ if (adev->gmc.xgmi.num_physical_nodes > 1)
+ amdgpu_xgmi_remove_device(adev);
+
/* Workaroud for ASICs need to disable SMC first */
amdgpu_device_smu_fini_early(adev);
@@ -2830,9 +2833,6 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev)
if (amdgpu_sriov_vf(adev) && adev->virt.ras_init_done)
amdgpu_virt_release_ras_err_handler_data(adev);
- if (adev->gmc.xgmi.num_physical_nodes > 1)
- amdgpu_xgmi_remove_device(adev);
-
amdgpu_amdkfd_device_fini_sw(adev);
for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
--
2.25.1
[-- Attachment #2: Type: text/html, Size: 5754 bytes --]
^ permalink raw reply related [flat|nested] 7+ messages in thread
* RE: [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini
2022-08-16 3:50 ` Zhang, Hawking
@ 2022-08-16 5:24 ` Chai, Thomas
0 siblings, 0 replies; 7+ messages in thread
From: Chai, Thomas @ 2022-08-16 5:24 UTC (permalink / raw)
To: Zhang, Hawking, amd-gfx; +Cc: Wang, Yang(Kevin)
[-- Attachment #1: Type: text/plain, Size: 3762 bytes --]
[AMD Official Use Only - General]
OK, I will update the patch.
From: Zhang, Hawking <Hawking.Zhang@amd.com>
Sent: Tuesday, August 16, 2022 11:51 AM
To: Chai, Thomas <YiPeng.Chai@amd.com>; amd-gfx@lists.freedesktop.org
Cc: Wang, Yang(Kevin) <KevinYang.Wang@amd.com>
Subject: Re: [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini
Fixed typo
Regards,
Hawking
From: Zhang, Hawking <Hawking.Zhang@amd.com<mailto:Hawking.Zhang@amd.com>>
Date: Tuesday, August 16, 2022 at 11:49
To: Chai, Thomas <YiPeng.Chai@amd.com<mailto:YiPeng.Chai@amd.com>>, amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org> <amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>>
Cc: Wang, Yang(Kevin) <KevinYang.Wang@amd.com<mailto:KevinYang.Wang@amd.com>>
Subject: RE: [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini
[AMD Official Use Only - General]
Alternatively, it might be better split xgmi ta terminate from xgmi_remove_device. In psp_hw_fini, check ta->fw and num_of_physical_mode to terminate xgmi ta. and make amdgpu_xgmi_remove_device only deal with software fini, like add_device.
Regards,
Hawking
-----Original Message-----
From: Chai, Thomas <YiPeng.Chai@amd.com<mailto:YiPeng.Chai@amd.com>>
Sent: Monday, August 15, 2022 15:03
To: amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Cc: Zhang, Hawking <Hawking.Zhang@amd.com<mailto:Hawking.Zhang@amd.com>>; Wang, Yang(Kevin) <KevinYang.Wang@amd.com<mailto:KevinYang.Wang@amd.com>>
Subject: RE: [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini
[AMD Official Use Only - General]
Ping on this series.
-----Original Message-----
From: Chai, Thomas <YiPeng.Chai@amd.com<mailto:YiPeng.Chai@amd.com>>
Sent: Friday, August 12, 2022 5:13 PM
To: amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Cc: Chai, Thomas <YiPeng.Chai@amd.com<mailto:YiPeng.Chai@amd.com>>; Zhang, Hawking <Hawking.Zhang@amd.com<mailto:Hawking.Zhang@amd.com>>; Wang, Yang(Kevin) <KevinYang.Wang@amd.com<mailto:KevinYang.Wang@amd.com>>; Chai, Thomas <YiPeng.Chai@amd.com<mailto:YiPeng.Chai@amd.com>>
Subject: [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini
The amdgpu_xgmi_remove_device function will send unload command to psp through psp ring to terminate xgmi, but psp ring has been destroyed in psp_hw_fini.
Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com<mailto:YiPeng.Chai@amd.com>>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index c84fdef0ac45..2445255bbf01 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -2787,6 +2787,9 @@ static int amdgpu_device_ip_fini_early(struct amdgpu_device *adev)
amdgpu_amdkfd_suspend(adev, false);
+ if (adev->gmc.xgmi.num_physical_nodes > 1)
+ amdgpu_xgmi_remove_device(adev);
+
/* Workaroud for ASICs need to disable SMC first */
amdgpu_device_smu_fini_early(adev);
@@ -2830,9 +2833,6 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev)
if (amdgpu_sriov_vf(adev) && adev->virt.ras_init_done)
amdgpu_virt_release_ras_err_handler_data(adev);
- if (adev->gmc.xgmi.num_physical_nodes > 1)
- amdgpu_xgmi_remove_device(adev);
-
amdgpu_amdkfd_device_fini_sw(adev);
for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
--
2.25.1
[-- Attachment #2: Type: text/html, Size: 8198 bytes --]
^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini
2022-08-15 7:02 ` [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini Chai, Thomas
2022-08-16 3:49 ` Zhang, Hawking
@ 2022-08-25 1:03 ` Zhang, Hawking
1 sibling, 0 replies; 7+ messages in thread
From: Zhang, Hawking @ 2022-08-25 1:03 UTC (permalink / raw)
To: Chai, Thomas, amd-gfx; +Cc: Wang, Yang(Kevin)
[-- Attachment #1: Type: text/plain, Size: 2447 bytes --]
I thought I reviewed this one together with another one from you that fixed hive refcount leak. You sent them in series.
Anyway, go ahead to submit with my RB.
Thanks.
Regards,
Hawking
From: amd-gfx <amd-gfx-bounces@lists.freedesktop.org> on behalf of Chai, Thomas <YiPeng.Chai@amd.com>
Date: Thursday, August 25, 2022 at 00:37
To: amd-gfx@lists.freedesktop.org <amd-gfx@lists.freedesktop.org>
Cc: Wang, Yang(Kevin) <KevinYang.Wang@amd.com>, Zhang, Hawking <Hawking.Zhang@amd.com>
Subject: RE: [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini
[AMD Official Use Only - General]
Ping on this series.
-----Original Message-----
From: Chai, Thomas <YiPeng.Chai@amd.com>
Sent: Friday, August 12, 2022 5:13 PM
To: amd-gfx@lists.freedesktop.org
Cc: Chai, Thomas <YiPeng.Chai@amd.com>; Zhang, Hawking <Hawking.Zhang@amd.com>; Wang, Yang(Kevin) <KevinYang.Wang@amd.com>; Chai, Thomas <YiPeng.Chai@amd.com>
Subject: [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini
The amdgpu_xgmi_remove_device function will send unload command to psp through psp ring to terminate xgmi, but psp ring has been destroyed in psp_hw_fini.
Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index c84fdef0ac45..2445255bbf01 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -2787,6 +2787,9 @@ static int amdgpu_device_ip_fini_early(struct amdgpu_device *adev)
amdgpu_amdkfd_suspend(adev, false);
+ if (adev->gmc.xgmi.num_physical_nodes > 1)
+ amdgpu_xgmi_remove_device(adev);
+
/* Workaroud for ASICs need to disable SMC first */
amdgpu_device_smu_fini_early(adev);
@@ -2830,9 +2833,6 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev)
if (amdgpu_sriov_vf(adev) && adev->virt.ras_init_done)
amdgpu_virt_release_ras_err_handler_data(adev);
- if (adev->gmc.xgmi.num_physical_nodes > 1)
- amdgpu_xgmi_remove_device(adev);
-
amdgpu_amdkfd_device_fini_sw(adev);
for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
--
2.25.1
[-- Attachment #2: Type: text/html, Size: 5570 bytes --]
^ permalink raw reply related [flat|nested] 7+ messages in thread
end of thread, other threads:[~2022-08-25 7:34 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-08-12 9:13 [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini YiPeng Chai
2022-08-12 9:13 ` [PATCH 2/2] drm/amdgpu: fix hive reference leak when adding xgmi device YiPeng Chai
2022-08-15 7:02 ` [PATCH 1/2] drm/amdgpu: The call to amdgpu_xgmi_remove_device needs to be earlier than psp_hw_fini Chai, Thomas
2022-08-16 3:49 ` Zhang, Hawking
2022-08-16 3:50 ` Zhang, Hawking
2022-08-16 5:24 ` Chai, Thomas
2022-08-25 1:03 ` Zhang, Hawking
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.