| Message ID | 02b510a8a0b43b2aa55b0438859fb9fbd0f89f25.1570547447.git.robin.murphy@arm.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | drm/panfrost: Quiet shrinker messages | expand |
On Tue, Oct 8, 2019 at 10:13 AM Robin Murphy <robin.murphy@arm.com> wrote: > > As brought up on IRC, logging a vague and unattributed message for a > normal and expected operation looks a bit spammy. Use a dev_* variant > to clarify it as a driver message, and downgrade the level to debug to > avoid cluttering up end users' logs. If it was good enough for msm, it's not good enough for us? > Signed-off-by: Robin Murphy <robin.murphy@arm.com> > --- > drivers/gpu/drm/panfrost/panfrost_gem_shrinker.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-)
On Tue, Oct 8, 2019 at 2:46 PM Rob Herring <robh@kernel.org> wrote: > > On Tue, Oct 8, 2019 at 10:13 AM Robin Murphy <robin.murphy@arm.com> wrote: > > > > As brought up on IRC, logging a vague and unattributed message for a > > normal and expected operation looks a bit spammy. Use a dev_* variant > > to clarify it as a driver message, and downgrade the level to debug to > > avoid cluttering up end users' logs. > > If it was good enough for msm, it's not good enough for us? > I suppose there are differing opinions about shrinker spam.. I kinda like to know when it is happening, OTOH it is a much more normal situation on devices with less RAM or android.. is there a ratelimited version of dev_dbg(), though? You might want that.. BR, -R
diff --git a/drivers/gpu/drm/panfrost/panfrost_gem_shrinker.c b/drivers/gpu/drm/panfrost/panfrost_gem_shrinker.c index 458f0fa68111..d1c0bd81c9f7 100644 --- a/drivers/gpu/drm/panfrost/panfrost_gem_shrinker.c +++ b/drivers/gpu/drm/panfrost/panfrost_gem_shrinker.c @@ -74,7 +74,7 @@ panfrost_gem_shrinker_scan(struct shrinker *shrinker, struct shrink_control *sc) mutex_unlock(&pfdev->shrinker_lock); if (freed > 0) - pr_info_ratelimited("Purging %lu bytes\n", freed << PAGE_SHIFT); + dev_dbg(pfdev->dev, "Purging %lu bytes\n", freed << PAGE_SHIFT); return freed; }
As brought up on IRC, logging a vague and unattributed message for a normal and expected operation looks a bit spammy. Use a dev_* variant to clarify it as a driver message, and downgrade the level to debug to avoid cluttering up end users' logs. Signed-off-by: Robin Murphy <robin.murphy@arm.com> --- drivers/gpu/drm/panfrost/panfrost_gem_shrinker.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)