On Mon, Oct 03, 2022 at 11:09:12AM +0200, Geert Uytterhoeven wrote:
Hi Greg,
On Mon, Oct 3, 2022 at 9:28 AM Greg Kroah-Hartman gregkh@linuxfoundation.org wrote:
From: Al Viro viro@zeniv.linux.org.uk
[ Upstream commit 06bbaa6dc53cb72040db952053432541acb9adc7 ]
passing kmap_local_page() result to __kernel_write() is unsafe - random ->write_iter() might (and 9p one does) get unhappy when passed ITER_KVEC with pointer that came from kmap_local_page().
Fix by providing a variant of __kernel_write() that takes an iov_iter from caller (__kernel_write() becomes a trivial wrapper) and adding dump_emit_page() that parallels dump_emit(), except that instead of __kernel_write() it uses __kernel_write_iter() with ITER_BVEC source.
Fixes: 3159ed57792b "fs/coredump: use kmap_local_page()" Signed-off-by: Al Viro viro@zeniv.linux.org.uk Signed-off-by: Sasha Levin sashal@kernel.org
This will need a follow-up patch, which I have just posted[1], to not break the build if CONFIG_ELF_CORE is not set.
[1] https://lore.kernel.org/20221003090657.2053236-1-geert@linux-m68k.org
Thanks, now dropped from 5.19 and 5.15 queues. When this gets merged, can you ping stable@kernel.org to add them both back?
thanks,
greg k-h