On Thu, Jan 2, 2014 at 4:02 PM, Bingfeng Mei <b...@broadcom.com> wrote: > Hi, > This patch fixes pr59651. The original regression for pr52943 only appears on > AArch64 target. I constructed a similar test that also exposes bug on x86-64. > The problem is that calculation of address range in alias versioning for > loops with negative step is wrong during vectorization. For example, for a > loop that access int a[3] -> a[1], the old calculated address range is [a, > a+12). It should be [a+4, a+16) instead. > > Bootstrapped and passed testsuite on x86-64. Get confirmed by original > reporter that regression is fixed on AArch64. OK to commit?
Ok. Thanks, Richard. > Thanks, > Bingfeng
