[Intel-gfx] [PATCH v2 2/5] drm/i915: Remove __GFP_NORETRY from our buffer allocator

[Chris Wilson]

I tried __GFP_NORETRY in the belief that __GFP_RECLAIM was effective. It
struggles with handling reclaim of our dirty buffers and relies on
reclaim via kswapd. As a result, a single pass of direct reclaim is
unreliable when i915 occupies the majority of available memory, and the
only means of effectively waiting on kswapd to amke progress is by not
setting the __GFP_NORETRY flag and lopping. That leaves us with the
dilemma of invoking the oomkiller instead of propagating the allocation
failure back to userspace where it can be handled more gracefully (one
hopes).  In the future we may have __GFP_MAYFAIL to allow repeats up until
we genuinely run out of memory and the oomkiller would have been invoked.
Until then, let the oomkiller wreck havoc.

v2: Stop playing with side-effects of gfp flags and await __GFP_MAYFAIL
v3: Update comments that direct reclaim only appears to be ignoring our
dirty buffers!

Fixes: 24f8e00a8a2e ("drm/i915: Prefer to report ENOMEM rather than incur the oom for gfx allocations")
Testcase: igt/gem_tiled_swapping
Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala at linux.intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen at linux.intel.com>
Cc: Daniel Vetter <daniel.vetter at ffwll.ch>
Cc: Michal Hocko <mhocko at suse.com>
---
 drivers/gpu/drm/i915/i915_gem.c | 15 ++++++++++++++-
 1 file changed, 14 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index ca61a0be1458..b22145a876c5 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -2404,7 +2404,20 @@ i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj)
 			if (!*s) {
 				/* reclaim and warn, but no oom */
 				gfp = mapping_gfp_mask(mapping);
-				gfp |= __GFP_NORETRY;
+
+				/* Our bo are always dirty and so we require
+				 * kswapd to reclaim our pages (direct reclaim
+				 * does not effectively begin pageout of our
+				 * buffers on its own). However, direct reclaim
+				 * only waits for kswapd when under allocation
+				 * congestion. So as a result __GFP_RECLAIM is
+				 * unreliable and fails to actually reclaim our
+				 * dirty pages -- unless you try over and over
+				 * again with !__GFP_NORETRY. However, we still
+				 * want to fail this allocation rather than
+				 * trigger the out-of-memory killer and for
+				 * this we want the future __GFP_MAYFAIL.
+				 */
 			}
 		} while (1);
 
-- 
2.11.0