drm/amd/pm: Enable smu v13_0_6 eccinfo in firmware query mode
Subsystems (3)
PATCH 1/1 drm/amd/pm: Enable smu v13_0_6 eccinfo in firmware query mode
smu v13_0_6 eccinfo is supported in firmware query mode only.
--- drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c +5 -2
@@ -2896,8 +2896,11 @@ static int smu_v13_0_6_select_xgmi_plpd_policy(struct smu_context *smu,static ssize_t smu_v13_0_6_get_ecc_info (struct smu_context *smu,void *table) { -/* Support ecc info by default */ -return 0 ; +struct amdgpu_device *adev = smu->adev; +unsigned int error_query_mode; + +return (amdgpu_ras_get_error_query_mode (adev, &error_query_mode) && + error_query_mode ==AMDGPU_RAS_FIRMWARE_ERROR_QUERY ) ?0 : -EOPNOTSUPP ; }static const struct pptable_funcs smu_v13_0_6_ppt_funcs = {
Signature · 1 line
2.25.1
[AMD Official Use Only - General]
Let' s see if we can retire the get_ecc_info callback since from smu_v13. (i.e., do not provide the callback implementation). What we just need to ensure the driver can fallback to direct poll mode when aca debug mode is enabled, this applies to both error query and bad page retirement.
Thoughts?
Regards,
Hawking
-----Original Message-----
From: amd-gfx <[email protected]> On Behalf Of Candice Li
Sent: Wednesday, January 10, 2024 10:26
To: [email protected]
Cc: Li, Candice <[email protected]>
Subject: [PATCH] drm/amd/pm: Enable smu v13_0_6 eccinfo in firmware query mode
smu v13_0_6 eccinfo is supported in firmware query mode only.
--- drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c +6 -3
@@ -2896,8 +2896,11 @@ static int smu_v13_0_6_select_xgmi_plpd_policy(struct smu_context *smu, static ssize_t smu_v13_0_6_get_ecc_info(struct smu_context *smu, void *table) { -/* Support ecc info by default */ -return 0 ; +struct amdgpu_device *adev = smu->adev; +unsigned int error_query_mode; + +return (amdgpu_ras_get_error_query_mode (adev, &error_query_mode) && + error_query_mode ==AMDGPU_RAS_FIRMWARE_ERROR_QUERY ) ?0 : +-EOPNOTSUPP ; }static const struct pptable_funcs smu_v13_0_6_ppt_funcs = { --
2.25.1
[AMD Official Use Only - General]
I agree.
Thanks,
Candice
-----Original Message-----
From: Zhang, Hawking <[email protected]>
Sent: Wednesday, January 10, 2024 10:48 AM
To: Li, Candice <[email protected]>; [email protected]
Cc: Li, Candice <[email protected]>
Subject: RE: [PATCH] drm/amd/pm: Enable smu v13_0_6 eccinfo in firmware query mode
[AMD Official Use Only - General]
Let' s see if we can retire the get_ecc_info callback since from smu_v13. (i.e., do not provide the callback implementation). What we just need to ensure the driver can fallback to direct poll mode when aca debug mode is enabled, this applies to both error query and bad page retirement.
Thoughts?
Regards,
Hawking
-----Original Message-----
From: amd-gfx <[email protected]> On Behalf Of Candice Li
Sent: Wednesday, January 10, 2024 10:26
To: [email protected]
Cc: Li, Candice <[email protected]>
Subject: [PATCH] drm/amd/pm: Enable smu v13_0_6 eccinfo in firmware query mode
smu v13_0_6 eccinfo is supported in firmware query mode only.
--- drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c +6 -3
@@ -2896,8 +2896,11 @@ static int smu_v13_0_6_select_xgmi_plpd_policy(struct smu_context *smu, static ssize_t smu_v13_0_6_get_ecc_info(struct smu_context *smu, void *table) { -/* Support ecc info by default */ -return 0 ; +struct amdgpu_device *adev = smu->adev; +unsigned int error_query_mode; + +return (amdgpu_ras_get_error_query_mode (adev, &error_query_mode) && + error_query_mode ==AMDGPU_RAS_FIRMWARE_ERROR_QUERY ) ?0 : +-EOPNOTSUPP ; }static const struct pptable_funcs smu_v13_0_6_ppt_funcs = { --
2.25.1