]> git.baikalelectronics.ru Git - kernel.git/commit
[PATCH] shmdt: check address alignment
authorHugh Dickins <hugh@veritas.com>
Fri, 24 Mar 2006 11:18:06 +0000 (03:18 -0800)
committerLinus Torvalds <torvalds@g5.osdl.org>
Fri, 24 Mar 2006 15:33:25 +0000 (07:33 -0800)
commite66d9f70a4ebfe54e331ad7044b2db4e5f33fcf9
treeaae81d11a102030772ab5a1a6bfa55820fbaf0da
parent2e4093fb831eed167b67b1cd6c004c4aa6614492
[PATCH] shmdt: check address alignment

SUSv3 says the shmdt() function shall fail with EINVAL if the value of
shmaddr is not the data segment start address of a shared memory segment:
our sys_shmdt needs to reject a shmaddr which is not page-aligned.

Does it have the potential to break existing apps?

Hugh says

  "sys_shmdt() just does the wrong (unexpected) thing with a misaligned
  address: it'll fail on what you might expect it to succeed on, and only
  succeed on what it should definitely fail on.

  "That is, I think it behaves as if shmaddr gets rounded up, when the only
  understandable behaviour would be if it rounded it down.

  "Which does mean you'd have to be devious to see anything but EINVAL from
  a misaligned shmaddr there, so it's not terribly important."

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
ipc/shm.c