On Thu, Jun 08 2017, Mikulas Patocka wrote:
> On Thu, 8 Jun 2017, NeilBrown wrote:
>
>> On Wed, Jun 07 2017, Mikulas Patocka wrote:
>>
>> > The function flush_signals clears all pending signals for the process. It
>> > may be used by kernel threads when we need to prepare a kernel thread for
>> > responding to signals. However using this function for an userspaces
>> > processes is incorrect - clearing signals without the program expecting it
>> > can cause misbehavior.
>> >
>> > The raid1 and raid5 code uses flush_signals in its request routine because
>> > it wants to prepare for an interruptible wait. This patch drops
>> > flush_signals and uses sigprocmask instead to block all signals (including
>> > SIGKILL) around the schedule() call. The signals are not lost, but the
>> > schedule() call won't respond to them.
>> >
>> > Signed-off-by: Mikulas Patocka <mpatocka@xxxxxxxxxx>
>> > Cc: stable@xxxxxxxxxxxxxxx
>>
>> Thanks for catching that!
>>
>> Acked-by: NeilBrown <neilb@xxxxxxxx>
>>
>> NeilBrown
>
> BTW. why does md_thread do "allow_signal(SIGKILL)" and then
> "if (signal_pending(current)) flush_signals(current)"?
>
> Does userspace really send SIGKILL to MD kernel threads? The SIGKILL will
> be lost when flush_signals is called, so it looks quite dubious.
>
This is for md_check_recovery() which does do something on a signal.
Chances are good that it will get to handle the signal before
md_thread() flushed them, but not guaranteed. I could be improved I
guess.
Or maybe it could be discarded - the md_check_recovery() thing.
The idea was that if you alt-sysrq-K to kill all processes, md arrays
would go into immediate-safe-mode where the metadata is marked clean
immediately after writes finish, rather than waiting a few seconds. The
chance of having a clean array after shutdown is hopefully improved.
I've never actually used this though, and I doubt many people know about
it. And bitmaps make it fairly pointless.
So I wouldn't object much if
allow_signal(SIGKILL);
and
if (signal_pending(current)) {
if (mddev->pers->sync_request && !mddev->external) {
pr_debug("md: %s in immediate safe mode\n",
mdname(mddev));
mddev->safemode = 2;
}
flush_signals(current);
}
were removed.
Thanks,
NeilBrown
Attachment:
signature.asc
Description: PGP signature
