| Message ID | 20160217142032.4895.52234.stgit@localhost.localdomain (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Wed, Feb 17, 2016 at 02:20:35PM +0000, Alan wrote: > The timeout is 64bit but the maths against it is done 32bit wrapped. Force > 64bit. Where does this wrap? Iirc the lowest HZ we have is 100, and that should still comfortably fit into 32 bits. I assume this is due to some overenthusiastic static checker. If so ok if I adjust the commit message to make it clear it never wrapped, and we only shut up a real tool here? Thanks, Daniel > > Signed-off-by: Alan Cox <alan@linux.intel.com> > --- > drivers/gpu/drm/i915/i915_gem.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c > index de57e7f..dc81045 100644 > --- a/drivers/gpu/drm/i915/i915_gem.c > +++ b/drivers/gpu/drm/i915/i915_gem.c > @@ -1358,9 +1358,9 @@ out: > * bit of mismatch in the jiffies<->nsecs<->ktime loop. So patch > * things up to make the test happy. We allow up to 1 jiffy. > * > - * This is a regrssion from the timespec->ktime conversion. > + * This is a regression from the timespec->ktime conversion. > */ > - if (ret == -ETIME && *timeout < jiffies_to_usecs(1)*1000) > + if (ret == -ETIME && *timeout < jiffies_to_usecs(1)*1000ULL) > *timeout = 0; > } > > > _______________________________________________ > dri-devel mailing list > dri-devel@lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/dri-devel
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index de57e7f..dc81045 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -1358,9 +1358,9 @@ out: * bit of mismatch in the jiffies<->nsecs<->ktime loop. So patch * things up to make the test happy. We allow up to 1 jiffy. * - * This is a regrssion from the timespec->ktime conversion. + * This is a regression from the timespec->ktime conversion. */ - if (ret == -ETIME && *timeout < jiffies_to_usecs(1)*1000) + if (ret == -ETIME && *timeout < jiffies_to_usecs(1)*1000ULL) *timeout = 0; }
The timeout is 64bit but the maths against it is done 32bit wrapped. Force 64bit. Signed-off-by: Alan Cox <alan@linux.intel.com> --- drivers/gpu/drm/i915/i915_gem.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)