We seem to be missing some W/A for 2M pages and are getting
a hit on raw GPU read bandwidths (even 30%) even though the
GPU write bandwidths improve (even 10%).
For now, disable THP, which is our only practical source of
2M pages until we have a W/A for the issue.
v2:
- Be explicit that we talk about GPU bandwidths (Eero)
- s/deny/never/ because that's why (Chris)
Reported-by: Valtteri Rantala <valtteri.rantala@xxxxxxxxx>
Fixes: b901bb89324a ("drm/i915/gemfs: enable THP")
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@xxxxxxxxxxxxxxx>
Cc: Matthew Auld <matthew.auld@xxxxxxxxx>
Cc: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
Cc: Jani Nikula <jani.nikula@xxxxxxxxxxxxxxx>
Cc: Rodrigo Vivi <rodrigo.vivi@xxxxxxxxx>
Cc: Valtteri Rantala <valtteri.rantala@xxxxxxxxx>
Cc: Eero Tamminen <eero.t.tamminen@xxxxxxxxx>
Reviewed-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
Reviewed-by: Matthew Auld <matthew.auld@xxxxxxxxx>
---
drivers/gpu/drm/i915/i915_gemfs.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/i915/i915_gemfs.c b/drivers/gpu/drm/i915/i915_gemfs.c
index e2993857df37..888b7d3f04c3 100644
--- a/drivers/gpu/drm/i915/i915_gemfs.c
+++ b/drivers/gpu/drm/i915/i915_gemfs.c
@@ -52,7 +52,8 @@ int i915_gemfs_init(struct drm_i915_private *i915)
if (has_transparent_hugepage()) {
struct super_block *sb = gemfs->mnt_sb;
- char options[] = "huge=within_size";
+ /* FIXME: Disabled until we get W/A for read BW issue. */
+ char options[] = "huge=never";
int flags = 0;
int err;
--
2.14.3
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
