Re: [RFC][PATCHv2 1/4] panic: avoid deadlocks in re-entrant console drivers

From: Sergey Senozhatsky
Date: Tue Oct 23 2018 - 08:04:49 EST


On (10/23/18 20:54), Sergey Senozhatsky wrote:
> So I did look at what lib/bust_spinlocks.c does; and I agree that waking
> up klogd makes little sense, on the other hand it just sets per-cpu
> pending bit, so not a big deal. console_unlock() should do there the
> same thing as console_flush_on_panic(). Yes. However, a bit of a bigger
> argument:
> __attribute__((weak)) suggests that bust_spinlocks() is arch-dependent
> and it's up to arch to do some extra stuff there [if needed]. So that's
> why I decided to keep bust_spinlocks(0) in panic() and, thus, call into
> arch-specific code (or common bust_spinlocks); then bump oops_in_progress
> so serial consoles become re-entrant and finally call
> console_flush_on_panic().

Seems that s390 is the only arch which defines its own bust_spinlocks().
Not sure why... Just to play games with console_loglevel?

---

void bust_spinlocks(int yes)
{
if (yes) {
oops_in_progress = 1;
} else {
int loglevel_save = console_loglevel;
console_unblank();
oops_in_progress = 0;
/*
* OK, the message is on the console. Now we call printk()
* without oops_in_progress set so that printk will give klogd
* a poke. Hold onto your hats...
*/
console_loglevel = 15;
printk(" ");
console_loglevel = loglevel_save;
}
}

---

The "printk(" "); without oops_in_progress" part is a bit worrisome.
This thing technically can deadlock. Unless s390 has no NMI panic().

-ss