On Wed, Nov 02, 2022 at 10:26:40AM +0100, Rasmus Villemoes wrote: > On 01/11/2022 23.33, Kees Cook wrote: > > Add __realloc_size() hint to kmemdup() so the compiler can reason about > > the length of the returned buffer. (These must not use __alloc_size, > > since those include __malloc which says the contents aren't defined[1]). > > > > [1] https://lore.kernel.org/linux-hardening/d199c2af-06af-8a50-a6a1-00eefa0b67b4@xxxxxxxxx/ > > > > > extern char *kstrndup(const char *s, size_t len, gfp_t gfp); > > -extern void *kmemdup(const void *src, size_t len, gfp_t gfp) __alloc_size(2); > > +extern void *kmemdup(const void *src, size_t len, gfp_t gfp) __realloc_size(2); > > What tree is this based on? I see that kmemdup() has grown that bogus > __alloc_size in next-20221101, but in next-20221102 this commit seems to > DTRT, namely > > -extern void *kmemdup(const void *src, size_t len, gfp_t gfp); > +extern void *kmemdup(const void *src, size_t len, gfp_t gfp) > __realloc_size(2); > > (i.e. there should never be an intermediate commit where kmemdup has > __alloc_size()). Right -- I fixed in in my -next tree to not use __alloc_size. -- Kees Cook
