------- Additional Comments From ak at muc dot de 2004-11-01 03:07 ------- I must correct myself slightly - the kernel oops I see is not the second loop, but in the check before:
ret = *pathp[0].slot; <--- references either NULL or a bogus pointer if (ret == NULL) goto out; I tried various options, but only -O0 resolves it. -O1, -O2 -fno-ivopts or -O1 -fno-tree-loop-ivcanon doesn't (depending on the options the crash is a NULL pointer or a bogus non NULL pointer) It's hard unfortunately to construct a run time test because it would need to create a simulated radixtree. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18241