[PATCH 5/7] drm/xe: Cleanup unwind of gt initialization

[Lucas De Marchi]

On Sat, Feb 01, 2025 at 08:24:50AM +0200, Raag Jadav wrote:
>On Fri, Jan 31, 2025 at 02:31:38PM -0800, Lucas De Marchi wrote:
>> Move the xe_gt_remove() to be handled by xe_gt.c itself so the caller,
>> xe_device_probe() doesn't have to unwind the gt loop. It's also more in
>> line with the xe_device_probe() style.
>>
>> Signed-off-by: Lucas De Marchi <lucas.demarchi at intel.com>
>> ---
>>  drivers/gpu/drm/xe/xe_device.c | 21 ++-----------------
>>  drivers/gpu/drm/xe/xe_gt.c     | 37 ++++++++++++++++------------------
>>  drivers/gpu/drm/xe/xe_gt.h     |  1 -
>>  3 files changed, 19 insertions(+), 40 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c
>> index e519f435b1606..bea626f6b4cbf 100644
>> --- a/drivers/gpu/drm/xe/xe_device.c
>> +++ b/drivers/gpu/drm/xe/xe_device.c
>> @@ -743,7 +743,6 @@ int xe_device_probe(struct xe_device *xe)
>>  	struct xe_tile *tile;
>>  	struct xe_gt *gt;
>>  	int err;
>> -	u8 last_gt;
>>  	u8 id;
>>
>>  	xe_pat_init_early(xe);
>> @@ -851,18 +850,16 @@ int xe_device_probe(struct xe_device *xe)
>>  		return err;
>>
>>  	for_each_gt(gt, xe, id) {
>> -		last_gt = id;
>> -
>>  		err = xe_gt_init(gt);
>>  		if (err)
>> -			goto err_fini_gt;
>> +			return err;
>
>So it's either all or nothing? Can't we operate with atleast what we have?

it was already like that, this is just moving the check to be in a
better place.

But yes, we've decided long ago that it's better to fail early and fix
it rather than having semi-working driver in production.

Lucas De Marchi

>
>Raag