Messages in this thread | | | Date | Tue, 14 Oct 2008 00:16:57 +0100 | From | Alan Cox <> | Subject | Re: [Resend][2.6.27-git-a447c093244] BUG: spinlock bad magic on CPU#0, setserial/3527 |
| |
On Mon, 13 Oct 2008 16:10:14 -0700 (PDT) Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > If this started today, then the most likely culprit is Alan's series of 80 > serial patches. > > Alan? I can only assume it's one of the > > spin_lock_irq(&up->port.lock); > > in serial8250_startup() (or quite likely in one of the functions it calls > that get inlined by the compiler, like enable_rsa()). > > And it sounds like it didn't get initialized.
Yeah I know whats going on I think. When I dropped out the NR_IRQ changes I didn't drop back in the spin lock init. Will sort it tomorrow (actually 'later today' as its midnight here)
Alan
| |