diff mbox

drm/i915: Disable THP until we have a read BW W/A

Message ID 20171124112750.29957-1-joonas.lahtinen@linux.intel.com (mailing list archive)
State New, archived
Headers show

Commit Message

Joonas Lahtinen Nov. 24, 2017, 11:27 a.m. UTC
We seem to be missing some W/A for 2M pages and are getting
a hit on raw read bandwidths (even 30%) even though the 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.

Reported-by: Valtteri Rantala <valtteri.rantala@intel.com>
Fixes: b901bb89324a ("drm/i915/gemfs: enable THP")
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Matthew Auld <matthew.auld@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Valtteri Rantala <valtteri.rantala@intel.com>
Cc: Eero Tamminen <eero.t.tamminen@intel.com>
---
 drivers/gpu/drm/i915/i915_gemfs.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

Comments

Chris Wilson Nov. 24, 2017, 12:12 p.m. UTC | #1
Quoting Joonas Lahtinen (2017-11-24 11:27:50)
> We seem to be missing some W/A for 2M pages and are getting
> a hit on raw read bandwidths (even 30%) even though the 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.
> 
> Reported-by: Valtteri Rantala <valtteri.rantala@intel.com>
> Fixes: b901bb89324a ("drm/i915/gemfs: enable THP")
> Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
> Cc: Matthew Auld <matthew.auld@intel.com>
> Cc: Chris Wilson <chris@chris-wilson.co.uk>
> Cc: Jani Nikula <jani.nikula@linux.intel.com>
> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
> Cc: Valtteri Rantala <valtteri.rantala@intel.com>
> Cc: Eero Tamminen <eero.t.tamminen@intel.com>

I don't have much to say, a lot of unknowns and you've already concisely
summed up everything that is known.
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
-Chris
Chris Wilson Nov. 24, 2017, 3:03 p.m. UTC | #2
Quoting Chris Wilson (2017-11-24 12:12:33)
> Quoting Joonas Lahtinen (2017-11-24 11:27:50)
> > We seem to be missing some W/A for 2M pages and are getting
> > a hit on raw read bandwidths (even 30%) even though the 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.
> > 
> > Reported-by: Valtteri Rantala <valtteri.rantala@intel.com>
> > Fixes: b901bb89324a ("drm/i915/gemfs: enable THP")
> > Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
> > Cc: Matthew Auld <matthew.auld@intel.com>
> > Cc: Chris Wilson <chris@chris-wilson.co.uk>
> > Cc: Jani Nikula <jani.nikula@linux.intel.com>
> > Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
> > Cc: Valtteri Rantala <valtteri.rantala@intel.com>
> > Cc: Eero Tamminen <eero.t.tamminen@intel.com>
> 
> I don't have much to say, a lot of unknowns and you've already concisely
> summed up everything that is known.
> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>

Bah, what works on the cmdline isn't allowed as an option!
Apparently, s/deny/never/
-Chris
Matthew Auld Nov. 24, 2017, 4:56 p.m. UTC | #3
On 24 November 2017 at 11:27, Joonas Lahtinen
<joonas.lahtinen@linux.intel.com> wrote:
> We seem to be missing some W/A for 2M pages and are getting
> a hit on raw read bandwidths (even 30%) even though the 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.
>
> Reported-by: Valtteri Rantala <valtteri.rantala@intel.com>
> Fixes: b901bb89324a ("drm/i915/gemfs: enable THP")
> Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
> Cc: Matthew Auld <matthew.auld@intel.com>
> Cc: Chris Wilson <chris@chris-wilson.co.uk>
> Cc: Jani Nikula <jani.nikula@linux.intel.com>
> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
> Cc: Valtteri Rantala <valtteri.rantala@intel.com>
> Cc: Eero Tamminen <eero.t.tamminen@intel.com>

With s/deny/never/:
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
diff mbox

Patch

diff --git a/drivers/gpu/drm/i915/i915_gemfs.c b/drivers/gpu/drm/i915/i915_gemfs.c
index e2993857df37..851d827d83a8 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=deny";
 		int flags = 0;
 		int err;