On Wed, Aug 15, 2012 at 11:17 AM, Oleg Nesterov <o...@redhat.com> wrote: > On 08/15, Kees Cook wrote: >> >> It sounds like get_task_comm shouldn't have locking at all then? It >> should just do a length-limited copy > > Without task_lock() get_task_comm() can copy incomplete new name. > > Honestly, I do not know any user which "strictly" needs the correct > name. may be proc.
Right, which is my point -- if the race to read against set_task_comm() isn't useful to anything, why lock in get_task_comm at all? > >> and make sure there is a trailing >> 0-byte? > > get_task_comm()->strncpy() should always see (and copy) 0-byte. > comm[TASK_COMM_LEN - 1] == '\0' and this byte is never changed. > > set_task_comm()->strlcpy() can write to this byte, but it can > only write 0 again. Right, and set_task_comm even does a memset() of 0 over the whole area before the strlcpy too. Regardless, it sounds like just using ->comm directly is fine; I'll send a patch. -Kees -- Kees Cook Chrome OS Security -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
