Heads-up: Possible regression between 10.0-RELEASE and 10.1-BETA1 with '/ on ZFS' setup

Steven Hartland smh at freebsd.org
Sat Oct 4 15:03:47 UTC 2014


----- Original Message ----- 
From: "Steven Hartland" <smh at freebsd.org>
> ----- Original Message ----- 
> From: "Glen Barber" <gjb at FreeBSD.org>
>> On Sat, Oct 04, 2014 at 03:51:39AM +0100, Steven Hartland wrote:
>> > > [...]
>> > >A possible solution to this is to set kern.kstack_pages=4 via loader(8),
>> > >however in my tests (solely in VirtualBox), this has been ineffective.
>> > >
>> > 
>> > This has been a known issue on i386 since the switch to Clang see UPDATING:
>> > 20121223:
>> >        After switching to Clang as the default compiler some users of ZFS
>> >        on i386 systems started to experience stack overflow kernel panics.
>> >        Please consider using 'options KSTACK_PAGES=4' in such configurations.
>> > 
>> > In my experience your millage may vary but essentially without 4 stack pages
>> > all bets are off in terms of stability.
>> > 
>> 
>> Did you read my email?
> 
> Cleary 4am was too late to be reading emails and my brain didn't ingest the
> "has been ineffective" so sorry for that.
> 
> That said setting options KSTACK_PAGES=4 has more reaching effects than
> kern.kstack_pages=4 so I wonder if you recompiled the kernel with this
> option if you would have any more luck?

Oh and just looking at the code kern.kstack_pages is read only so wont have
any effect, hence you will definitely need to set the kernel option as per
the UPDATING entry.

    Regards
    Steve


More information about the freebsd-stable mailing list