[PATCH] drm/ttm: Make sure BOs being swapped out are cacheable

[Christian König]

Am 25.01.2017 um 10:25 schrieb Thomas Hellstrom:
> On 01/25/2017 09:21 AM, Michel Dänzer wrote:
>> From: Michel Dänzer <michel.daenzer at amd.com>
>>
>> The current caching state may not be tt_cached, even though the
>> placement contains TTM_PL_FLAG_CACHED, because placement can contain
>> multiple caching flags. Trying to swap out such a BO would trip up the
>>
>> 	BUG_ON(ttm->caching_state != tt_cached);
>>
>> in ttm_tt_swapout.
>>
>> Cc: stable at vger.kernel.org
>> Signed-off-by: Michel Dänzer <michel.daenzer at amd.com>
> Reviewed-by: Thomas Hellstrom <thellstrom at vmware.com>

Reviewed-by: Christian König <christian.koenig at amd.com>.

>
>> ---
>>   drivers/gpu/drm/ttm/ttm_bo.c | 4 ++--
>>   1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
>> index d5063618efa7..86e3b233b722 100644
>> --- a/drivers/gpu/drm/ttm/ttm_bo.c
>> +++ b/drivers/gpu/drm/ttm/ttm_bo.c
>> @@ -1670,7 +1670,6 @@ static int ttm_bo_swapout(struct ttm_mem_shrink *shrink)
>>   	struct ttm_buffer_object *bo;
>>   	int ret = -EBUSY;
>>   	int put_count;
>> -	uint32_t swap_placement = (TTM_PL_FLAG_CACHED | TTM_PL_FLAG_SYSTEM);
>>   
>>   	spin_lock(&glob->lru_lock);
>>   	list_for_each_entry(bo, &glob->swap_lru, swap) {
>> @@ -1701,7 +1700,8 @@ static int ttm_bo_swapout(struct ttm_mem_shrink *shrink)
>>   	 * Move to system cached
>>   	 */
>>   
>> -	if ((bo->mem.placement & swap_placement) != swap_placement) {
>> +	if (bo->mem.mem_type != TTM_PL_SYSTEM ||
>> +	    bo->ttm->caching_state != tt_cached) {
>>   		struct ttm_mem_reg evict_mem;
>>   
>>   		evict_mem = bo->mem;
>