On Thu, Jan 20, 2011 at 05:01:13PM -0700, Eric Blake wrote:
Regression introduced in commit e6b68d7.
Prior to that point, handlesAlloc was always a multiple of
EVENT_ALLOC_EXTENT, and was an integer (so even if the
subtraction wrapped, a negative value was less than the
count and did not
On 01/21/2011 04:15 AM, Daniel P. Berrange wrote:
Prior to that point, handlesAlloc was always a multiple of
EVENT_ALLOC_EXTENT, and was an integer (so even if the
subtraction wrapped, a negative value was less than the
count and did not try to free the handles array). But after
that point,
Regression introduced in commit e6b68d7.
Prior to that point, handlesAlloc was always a multiple of
EVENT_ALLOC_EXTENT, and was an integer (so even if the
subtraction wrapped, a negative value was less than the
count and did not try to free the handles array). But after
that point, VIR_RESIZE_N
On Thu, Jan 20, 2011 at 05:01:13PM -0700, Eric Blake wrote:
Regression introduced in commit e6b68d7.
Prior to that point, handlesAlloc was always a multiple of
EVENT_ALLOC_EXTENT, and was an integer (so even if the
subtraction wrapped, a negative value was less than the
count and did not
