On 07/04/2007 01:35 PM, Jiri Kosina wrote:On Wed, 4 Jul 2007, Jakub Jelinek wrote:
The above highlighted changes are the cause of random segfaults of PIE binaries. See https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=246623Thanks a lot for pointing this out. Andrew, could this be folded into pie-randomization.patch please?
From: Jiri Kosina <jkosina@xxxxxxx>
pie randomization: fix BAD_ADDR macro
pie-randomization.patch makes the load_addr in load_elf_interp() the load bias of ld.so (difference between the actual load base address and first PT_LOAD segment's p_vaddr). If the difference equals (on x86) to 0xfffff000 (which is valid [1]), SIGSEGV is incorrectly sent.
This patch changes the BAD_ADDR so that it catches the mappings to the error-area properly.
But what about this patch that made the opposite change:
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=ce51059be56f63762089412b3ece348067afda85