Re: [PATCH] drm/ttm: fix busy memory to fail other user v5

2019-05-07 Thread zhoucm1
On 2019年05月07日 17:03, Christian König wrote: [CAUTION: External Email] Ping! in fact, already address your comments, just wait for Prike test result, anyway, send v6 first. -David Marek is going to need this for its GDS patches as well. Christina. Am 30.04.19 um 11:10 schrieb Christian

Re: [PATCH] drm/ttm: fix busy memory to fail other user v5

2019-05-07 Thread Christian König
Ping! Marek is going to need this for its GDS patches as well. Christina. Am 30.04.19 um 11:10 schrieb Christian König: Am 30.04.19 um 09:01 schrieb Chunming Zhou: heavy gpu job could occupy memory long time, which lead other user fail to get memory. basically pick up Christian idea: 1. Re

Re: [PATCH] drm/ttm: fix busy memory to fail other user v5

2019-04-30 Thread Christian König
Am 30.04.19 um 09:01 schrieb Chunming Zhou: heavy gpu job could occupy memory long time, which lead other user fail to get memory. basically pick up Christian idea: 1. Reserve the BO in DC using a ww_mutex ticket (trivial). 2. If we then run into this EBUSY condition in TTM check if the BO we

[PATCH] drm/ttm: fix busy memory to fail other user v5

2019-04-30 Thread Chunming Zhou
heavy gpu job could occupy memory long time, which lead other user fail to get memory. basically pick up Christian idea: 1. Reserve the BO in DC using a ww_mutex ticket (trivial). 2. If we then run into this EBUSY condition in TTM check if the BO we need memory for (or rather the ww_mutex of it

[PATCH] drm/ttm: fix busy memory to fail other user v5

2019-04-26 Thread Chunming Zhou
heavy gpu job could occupy memory long time, which lead other user fail to get memory. basically pick up Christian idea: 1. Reserve the BO in DC using a ww_mutex ticket (trivial). 2. If we then run into this EBUSY condition in TTM check if the BO we need memory for (or rather the ww_mutex of it

Re: [PATCH] drm/ttm: fix busy memory to fail other user v4

2019-04-26 Thread zhoucm1
please ignore v4. Will send v5 instead. On 2019年04月26日 17:53, Chunming Zhou wrote: heavy gpu job could occupy memory long time, which lead other user fail to get memory. basically pick up Christian idea: 1. Reserve the BO in DC using a ww_mutex ticket (trivial). 2. If we then run into this E

[PATCH] drm/ttm: fix busy memory to fail other user v4

2019-04-26 Thread Chunming Zhou
heavy gpu job could occupy memory long time, which lead other user fail to get memory. basically pick up Christian idea: 1. Reserve the BO in DC using a ww_mutex ticket (trivial). 2. If we then run into this EBUSY condition in TTM check if the BO we need memory for (or rather the ww_mutex of it

[PATCH] drm/ttm: fix busy memory to fail other user v4

2019-04-26 Thread Chunming Zhou
heavy gpu job could occupy memory long time, which lead other user fail to get memory. basically pick up Christian idea: 1. Reserve the BO in DC using a ww_mutex ticket (trivial). 2. If we then run into this EBUSY condition in TTM check if the BO we need memory for (or rather the ww_mutex of it

Re: [PATCH] drm/ttm: fix busy memory to fail other user v3

2019-04-26 Thread zhoucm1
On 2019年04月26日 17:11, Koenig, Christian wrote: Am 26.04.19 um 11:07 schrieb zhoucm1: [SNIP] + spin_lock(&glob->lru_lock); +    for (i = 0; i < TTM_MAX_BO_PRIORITY; ++i) { +    if (list_empty(&man->lru[i])) +    continue; +    bo = list_first_entry(&man->lru[i],

Re: [PATCH] drm/ttm: fix busy memory to fail other user v3

2019-04-26 Thread Koenig, Christian
Am 26.04.19 um 11:07 schrieb zhoucm1: > [SNIP] >>> + spin_lock(&glob->lru_lock); >>> +    for (i = 0; i < TTM_MAX_BO_PRIORITY; ++i) { >>> +    if (list_empty(&man->lru[i])) >>> +    continue; >>> +    bo = list_first_entry(&man->lru[i], >>> + 

Re: [PATCH] drm/ttm: fix busy memory to fail other user v3

2019-04-26 Thread zhoucm1
On 2019年04月26日 16:31, Christian König wrote: Am 25.04.19 um 09:39 schrieb Chunming Zhou: heavy gpu job could occupy memory long time, which lead other user fail to get memory. basically pick up Christian idea: 1. Reserve the BO in DC using a ww_mutex ticket (trivial). Any reason you don't

Re: [PATCH] drm/ttm: fix busy memory to fail other user v3

2019-04-26 Thread Christian König
Am 25.04.19 um 09:39 schrieb Chunming Zhou: heavy gpu job could occupy memory long time, which lead other user fail to get memory. basically pick up Christian idea: 1. Reserve the BO in DC using a ww_mutex ticket (trivial). Any reason you don't want to split this into a separate patch? 2.

[PATCH] drm/ttm: fix busy memory to fail other user v3

2019-04-25 Thread Chunming Zhou
heavy gpu job could occupy memory long time, which lead other user fail to get memory. basically pick up Christian idea: 1. Reserve the BO in DC using a ww_mutex ticket (trivial). 2. If we then run into this EBUSY condition in TTM check if the BO we need memory for (or rather the ww_mutex of it

Re: [PATCH] drm/ttm: fix busy memory to fail other user

2019-04-24 Thread Christian König
Am 24.04.19 um 13:40 schrieb Chunming Zhou: basically pick up Christian idea: 1. Reserve the BO in DC using a ww_mutex ticket (trivial). Please split that step into a separate patch. 2. If we then run into this EBUSY condition in TTM check if the BO we need memory for (or rather the ww_mute

[PATCH] drm/ttm: fix busy memory to fail other user

2019-04-24 Thread Chunming Zhou
basically pick up Christian idea: 1. Reserve the BO in DC using a ww_mutex ticket (trivial). 2. If we then run into this EBUSY condition in TTM check if the BO we need memory for (or rather the ww_mutex of its reservation object) has a ticket assigned. 3. If we have a ticket we grab a reference