On Wed, May 23, 2012 at 07:19:17PM +0100, Al Viro wrote:
> Grr... That's a conflict between uml gaining TIF_NOTIFY_RESUME (needed
> prereq to task_work_add() series) and patch in said series doing away
> with explicit calls of key_replace_session_keyring(). Fixup is to remove
> those two lines in arch/um/process.c, same as done on other architectures
> by "keys: kill the dummy key_replace_session_keyring()" (commit c1cb001).
>
> Not an issue for mainline merge, since task_work_add patchset goes later,
> but I think I'll have to cherry-pick that series into signal.git. And
> probably reorder it a bit, moving the calls into tracehook_notify_resume()
> first, with "kill the dummy..." commit removing just that single call.
Hmm... Two solutions, take your pick:
1)
I think the minimal solution is this: I add the "move
key_replace_session_keyring() into tracehook_notify_resume()" into signal.git
for-next, which yields one conflict with next/akpm. With conflict resolution
being "take tracehook_notify_resume() from next/akpm". I've put that
into for-next-variant1
2)
Cherry-picked these guys into signal.git, along with the rest
of signal prereqs for them. Merge with next/akpm-base yields a couple
of trivial conflicts in kernel/fork.c (with
sched, mm: Rework sched_{fork,exec} node assignment
removing INIT_LIST_HEAD right next to the place where we add one; conflict
resolution being just keep the one Oleg adds and remove the one Peter removes)
and in kernel/irq/manage.c (with
genirq: Be more informative on irq type mismatch
changing a couple of printks in there; conflict resolution: just remove
exit_irq_thread() in merged variant). That's for-next-variant2. With that
variant we get 5 more duplicates with next/akpm, obviously.
Stephen, which way would you prefer it handled?
--
To unsubscribe from this list: send the line "unsubscribe linux-next" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
