[Intel-gfx] [PATCH v2 4/6] drm/i915/shrinker: Restrict vmap purge to objects with vmaps

[Dave Gordon]

On 11/04/16 17:40, Chris Wilson wrote:
> On Mon, Apr 11, 2016 at 03:57:21PM +0100, Tvrtko Ursulin wrote:
>>
>> On 11/04/16 15:44, Chris Wilson wrote:
>>> On Mon, Apr 11, 2016 at 03:25:41PM +0100, Tvrtko Ursulin wrote:
>>>>
>>>> On 08/04/16 12:11, Chris Wilson wrote:
>>>>> When called because we have run out of vmap address space, we only need
>>>>> to recover objects that have vmappings and not all.
>>>>>
>>>>> v2: Start using is_vmalloc_addr()
>>>>>
>>>>> Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
>>>>> Cc: Tvrtko Ursulin <tvrtko.ursulin at linux.intel.com>
>>>>> ---
>>>>>   drivers/gpu/drm/i915/i915_drv.h          |  1 +
>>>>>   drivers/gpu/drm/i915/i915_gem_shrinker.c | 10 +++++++++-
>>>>>   2 files changed, 10 insertions(+), 1 deletion(-)
>>>>
>>>> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
>>>
>>> Having started to use the obj->mapping cache in anger, do you think there
>>> is any left before pushing this part of the puzzle? I'm hitting an
>>> interesting scaling problem with vmalloc's lazy free list (that's a
>>> patch for later!).
>>
>> Looks ready and useful to me so I vote for merging it.
>
> Done. Hopefully I can make some headway on bugfixes and then start
> chasing some of the more interesting improvements.
> -Chris

Yay! Now we'll try using it for the various GuC objects that could do 
with a long-term mapping rather than repeated kmap_atomic() :)

.Dave.