[PATCH 1/2] drm/xe/vm: Use xe_vm_lock()/xe_vm_unlock() helpers

Nilawar, Badal badal.nilawar at intel.com
Wed Apr 24 04:09:30 UTC 2024



On 24-04-2024 09:35, Nilawar, Badal wrote:
> 
> 
> On 24-04-2024 09:27, Ghimiray, Himal Prasad wrote:
>>
>> On 24-04-2024 09:24, Matthew Brost wrote:
>>> On Wed, Apr 24, 2024 at 09:27:55AM +0530, Himal Prasad Ghimiray wrote:
>>>> There is no change in functionality. Using the helper function
>>>> defined within the driver.
>>>>
>>>> -v2
>>>> Use xe_vm_unlock() (Ashutosh/Matt)
>>>>
>>>> Reviewed-by: Badal Nilawar <badal.nilawar at intel.com>
>>>> Cc: Matthew Brost <matthew.brost at intel.com>
>>>> Cc: Ashutosh Dixit <ashutosh.dixit at intel.com>
>>>> Signed-off-by: Himal Prasad Ghimiray <himal.prasad.ghimiray at intel.com>
>>>> ---
>>>>   drivers/gpu/drm/xe/xe_vm.c | 4 ++--
>>>>   1 file changed, 2 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c
>>>> index 85d6f359142d..c350c561e7d9 100644
>>>> --- a/drivers/gpu/drm/xe/xe_vm.c
>>>> +++ b/drivers/gpu/drm/xe/xe_vm.c
>>>> @@ -1279,7 +1279,7 @@ struct xe_vm *xe_vm_create(struct xe_device 
>>>> *xe, u32 flags)
>>>>       drm_gem_object_put(vm_resv_obj);
>>>> -    err = dma_resv_lock_interruptible(xe_vm_resv(vm), NULL);
>>>> +    err = xe_vm_lock(vm, true);
>>>>       if (err)
>>>>           goto err_close;
>>>> @@ -1323,7 +1323,7 @@ struct xe_vm *xe_vm_create(struct xe_device 
>>>> *xe, u32 flags)
>>>>           xe_pt_populate_empty(tile, vm, vm->pt_root[id]);
>>>>       }
>>>> -    dma_resv_unlock(xe_vm_resv(vm));
>>>> +    xe_vm_unlock(vm);
>>> There is another dma_resv_unlock below the label 'err_unlock_close' too.
>>>
>>> Matt
>>
>>
>> Thanks Matt, Addressing it in next version.
> I think dma_resv_lock_interruptible also be replaced with wrapper.
> 
> xe_vm.c:1282:   err = dma_resv_lock_interruptible(xe_vm_resv(vm), NULL);
> xe_vm.c:3117:   return dma_resv_lock_interruptible(xe_vm_resv(vm), NULL);
> 
> xe_bo.c:2050:  return dma_resv_lock_interruptible(bo->ttm.base.resv, NULL);
Please ignore this, it is already addressed.

> 
> Regards,
> Badal
>>
>>
>>>
>>>>       /* Kernel migration VM shouldn't have a circular loop.. */
>>>>       if (!(flags & XE_VM_FLAG_MIGRATION)) {
>>>> -- 
>>>> 2.25.1
>>>>


More information about the Intel-xe mailing list