[PATCH v3 1/4] drm/ttm: Create pinned list
Christian König
ckoenig.leichtzumerken at gmail.com
Mon Aug 30 17:05:06 UTC 2021
Am 30.08.21 um 19:02 schrieb Andrey Grodzovsky:
>
> On 2021-08-30 12:51 p.m., Christian König wrote:
>> Am 30.08.21 um 16:16 schrieb Andrey Grodzovsky:
>>>
>>> On 2021-08-30 4:58 a.m., Christian König wrote:
>>>> Am 27.08.21 um 22:39 schrieb Andrey Grodzovsky:
>>>>> This list will be used to capture all non VRAM BOs not
>>>>> on LRU so when device is hot unplugged we can iterate
>>>>> the list and unmap DMA mappings before device is removed.
>>>>>
>>>>> v2: Reanme function to ttm_bo_move_to_pinned
>>>>> v3: Move the pinned list to ttm device
>>>>
>>>> As far as I can see there is not list_del() remaining. So this
>>>> won't work correctly.
>>>
>>>
>>> It's in ttm_bo_release, there was no code change there hence it's
>>> not captured in the patch.
>>
>> Ah! So you keep the logic as is there. Sorry totally missed that.
>>
>> In this case the patch is Reviewed-by: Christian König
>> <christian.koenig at amd.com>
>>
>> Can you push this to drm-misc-next?
>>
>> Thanks,
>> Christian.
>
>
> I think It's supposed to go on top of your changes you mention here
> which are not pushed yet.
> I will need to apply all the patches on top of yours and retest (I was
> doing everything in amd-staging-drm-next)
> until now.
Works for me as well. Alternatively you can just push this patch here to
drm-misc-next so that I can rebase on top and merge the rest through
amd-staging-drm-next.
The patch pushed to drm-misc-next should automatically fall out when
Alex rebases his stuff on upstream the next time.
Christian.
>
> Andrey
>
>
>>
>>>
>>> Andrey
>>>
>>>
>>>>
>>>> I suggest to rather rebase on top of the stuff I'm working on for a
>>>> while to move the LRU into the resource instead.
>>>>
>>>> Just send out the latest patch set of this with you in CC.
>>>>
>>>> Christian.
>>>>
>>>>>
>>>>> Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky at amd.com>
>>>>> Suggested-by: Christian König <christian.koenig at amd.com>
>>>>> ---
>>>>> drivers/gpu/drm/ttm/ttm_bo.c | 18 ++++++++++++++----
>>>>> drivers/gpu/drm/ttm/ttm_device.c | 1 +
>>>>> include/drm/ttm/ttm_device.h | 1 +
>>>>> 3 files changed, 16 insertions(+), 4 deletions(-)
>>>>>
>>>>> diff --git a/drivers/gpu/drm/ttm/ttm_bo.c
>>>>> b/drivers/gpu/drm/ttm/ttm_bo.c
>>>>> index 1b950b45cf4b..1fedd0eb67ba 100644
>>>>> --- a/drivers/gpu/drm/ttm/ttm_bo.c
>>>>> +++ b/drivers/gpu/drm/ttm/ttm_bo.c
>>>>> @@ -69,7 +69,17 @@ static void ttm_bo_mem_space_debug(struct
>>>>> ttm_buffer_object *bo,
>>>>> }
>>>>> }
>>>>> -static void ttm_bo_del_from_lru(struct ttm_buffer_object *bo)
>>>>> +static inline void ttm_bo_move_to_pinned(struct ttm_buffer_object
>>>>> *bo)
>>>>> +{
>>>>> + struct ttm_device *bdev = bo->bdev;
>>>>> +
>>>>> + list_move_tail(&bo->lru, &bdev->pinned);
>>>>> +
>>>>> + if (bdev->funcs->del_from_lru_notify)
>>>>> + bdev->funcs->del_from_lru_notify(bo);
>>>>> +}
>>>>> +
>>>>> +static inline void ttm_bo_del_from_lru(struct ttm_buffer_object *bo)
>>>>> {
>>>>> struct ttm_device *bdev = bo->bdev;
>>>>> @@ -98,7 +108,7 @@ void ttm_bo_move_to_lru_tail(struct
>>>>> ttm_buffer_object *bo,
>>>>> dma_resv_assert_held(bo->base.resv);
>>>>> if (bo->pin_count) {
>>>>> - ttm_bo_del_from_lru(bo);
>>>>> + ttm_bo_move_to_pinned(bo);
>>>>> return;
>>>>> }
>>>>> @@ -339,7 +349,7 @@ static int ttm_bo_cleanup_refs(struct
>>>>> ttm_buffer_object *bo,
>>>>> return ret;
>>>>> }
>>>>> - ttm_bo_del_from_lru(bo);
>>>>> + ttm_bo_move_to_pinned(bo);
>>>>> list_del_init(&bo->ddestroy);
>>>>> spin_unlock(&bo->bdev->lru_lock);
>>>>> ttm_bo_cleanup_memtype_use(bo);
>>>>> @@ -1154,7 +1164,7 @@ int ttm_bo_swapout(struct ttm_buffer_object
>>>>> *bo, struct ttm_operation_ctx *ctx,
>>>>> return 0;
>>>>> }
>>>>> - ttm_bo_del_from_lru(bo);
>>>>> + ttm_bo_move_to_pinned(bo);
>>>>> /* TODO: Cleanup the locking */
>>>>> spin_unlock(&bo->bdev->lru_lock);
>>>>> diff --git a/drivers/gpu/drm/ttm/ttm_device.c
>>>>> b/drivers/gpu/drm/ttm/ttm_device.c
>>>>> index 5f31acec3ad7..530a9c36be37 100644
>>>>> --- a/drivers/gpu/drm/ttm/ttm_device.c
>>>>> +++ b/drivers/gpu/drm/ttm/ttm_device.c
>>>>> @@ -208,6 +208,7 @@ int ttm_device_init(struct ttm_device *bdev,
>>>>> struct ttm_device_funcs *funcs,
>>>>> INIT_DELAYED_WORK(&bdev->wq, ttm_device_delayed_workqueue);
>>>>> spin_lock_init(&bdev->lru_lock);
>>>>> INIT_LIST_HEAD(&bdev->ddestroy);
>>>>> + INIT_LIST_HEAD(&bdev->pinned);
>>>>> bdev->dev_mapping = mapping;
>>>>> mutex_lock(&ttm_global_mutex);
>>>>> list_add_tail(&bdev->device_list, &glob->device_list);
>>>>> diff --git a/include/drm/ttm/ttm_device.h
>>>>> b/include/drm/ttm/ttm_device.h
>>>>> index cd592f8e941b..03fb44d061e0 100644
>>>>> --- a/include/drm/ttm/ttm_device.h
>>>>> +++ b/include/drm/ttm/ttm_device.h
>>>>> @@ -265,6 +265,7 @@ struct ttm_device {
>>>>> */
>>>>> spinlock_t lru_lock;
>>>>> struct list_head ddestroy;
>>>>> + struct list_head pinned;
>>>>> /*
>>>>> * Protected by load / firstopen / lastclose /unload sync.
>>>>
>>
More information about the amd-gfx
mailing list