Messages in this thread | | | Date | Wed, 10 Jun 2015 11:54:11 -0400 | From | Steven Rostedt <> | Subject | Re: [RFC][PATCH] printk: Fixup the nmi printk mess |
| |
On Wed, 10 Jun 2015 17:32:53 +0200 Peter Zijlstra <peterz@infradead.org> wrote:
> > Note that the backtraces used to be serialized via > > static arch_spinlock_t lock = __ARCH_SPIN_LOCK_UNLOCKED > > in the past. See the commit > > a9edc8809328 ("x86/nmi: Perform a safe NMI stack trace on all CPUs") > > We could easily add a static raw_spinlock_t to __printk_nmi_flush() and > serialize its invocations if people think that is important.
Yes please. One thing about the original change is that it made the NMI backtraces readable. I hated the interleaving of the backtraces, as it made it almost impossible to to debug.
-- Steve
| |