Re: [PATCH v6 3/4] drm/xe/svm: Report terminal page-fault failures using SIGID

["Yadav, Arvind" <arvind.yadav@intel.com>]

On 02-09-2026 18:13, Michal Wajdeczko wrote:
>
> On 9/1/2026 8:30 AM, Arvind Yadav wrote:
>> Route terminal SVM page-fault failures through the structured SIGID logging
>> helper.
>>
>> These errors are reported only after the retry count is exceeded, so they
>> represent terminal page-fault handling failures rather than transient retry
>> conditions.
> that sounds like a LLM output
>
>> Use the PAGEFAULT component, which maps to XE_SIGID_MEM_FAULT, and pass the
>> real errno value to the helper. Pass the GT object so the report carries
>> tile/GT location. The helper renders the errno, so remove the manual
>> ERR_PTR() formatting from the message text.
> do we really need to explain here the details how xe_log() macro works?
> or that 'real errno' value will be used?


Noted, I will simplify the commit message.

>
>> Cc: Mallesh Koujalagi <mallesh.koujalagi@intel.com>
>> Cc: Badal Nilawar <badal.nilawar@intel.com>
>> Cc: Matthew Brost <matthew.brost@intel.com>
>> Cc: Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>
>> Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
>> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
>> Suggested-by: Mallesh Koujalagi <mallesh.koujalagi@intel.com>
>> Signed-off-by: Arvind Yadav <arvind.yadav@intel.com>
>> ---
>>   drivers/gpu/drm/xe/xe_svm.c | 13 +++++++------
>>   1 file changed, 7 insertions(+), 6 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
>> index 627a741293d5..349e307ca092 100644
>> --- a/drivers/gpu/drm/xe/xe_svm.c
>> +++ b/drivers/gpu/drm/xe/xe_svm.c
>> @@ -13,6 +13,7 @@
>>   #include "xe_bo.h"
>>   #include "xe_exec_queue_types.h"
>>   #include "xe_gt_stats.h"
>> +#include "xe_log.h"
>>   #include "xe_migrate.h"
>>   #include "xe_module.h"
>>   #include "xe_pagefault.h"
>> @@ -1361,9 +1362,9 @@ static int __xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
>>   				else
>>   					goto retry;
>>   			} else {
>> -				drm_err(&vm->xe->drm,
>> -					"VRAM allocation failed, retry count exceeded, asid=%u, errno=%pe\n",
>> -					vm->usm.asid, ERR_PTR(err));
>> +				xe_log_err(gt, PAGEFAULT, err,
>> +					   "VRAM allocation failed, retry count exceeded, asid=%u\n",
> s/asid/ASID


Noted,

>
>> +					   vm->usm.asid);
>>   				goto err_out;
>>   			}
>>   		}
>> @@ -1384,9 +1385,9 @@ static int __xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
>>   			range_debug(range, "PAGE FAULT - RETRY PAGES");
>>   			goto retry;
>>   		} else {
>> -			drm_err(&vm->xe->drm,
>> -				"Get pages failed, retry count exceeded, asid=%u, gpusvm=%p, errno=%pe\n",
>> -				vm->usm.asid, &vm->svm.gpusvm, ERR_PTR(err));
>> +			xe_log_err(gt, PAGEFAULT, err,
>> +				   "Get pages failed, retry count exceeded, asid=%u, gpusvm=%p\n",
>> +				   vm->usm.asid, &vm->svm.gpusvm);
> s/asid/ASID
>
> also, does it make sense to log the gpusv kernel pointer ??
> maybe log the gpusvm->name instead?


Noted,
I will log the GPUVM name instead of the kernel pointer.

Thanks,
Arvind

>
>>   		}
>>   	}
>>   	if (err) {