Re: Status on CPU hotplug issues

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Fri, 6 Oct 2006 16:10:12 -0700
Bryce Harrington <[email protected]> wrote:

> On Fri, Mar 17, 2006 at 01:04:12AM -0800, Andrew Morton wrote:
> > Srivatsa Vaddagiri <[email protected]> wrote:
> > > Well ..other arch-es need to have a similar check if they get around to
> > > implement physical hot-add (even if they allow offlining of all CPUs). This is 
> > > required since a user can (by mistake maybe) try to bring up an already online 
> > > CPU by writing a '1' to it's sysfs 'online' file. 'store_online' 
> > > (drivers/base/cpu.c) unconditionally calls 'smp_prepare_cpu' w/o checking for 
> > > this error condition. The check added in the patch catches such error 
> > > conditions as well.
> > 
> > OK..  I guess we should fix those architectures while we're thinking about it.
> >
> > > +	/* Check if CPU is already online. This can happen if user tries to 
> > > +	 * bringup an already online CPU or a previous offline attempt
> > > +	 * on this CPU has failed.
> > > +	 */
> > > +	if (cpu_online(cpu)) {
> > > +		ret = -EINVAL;
> > > +		goto exit;
> > > +	}
> > 
> > How well tested is this?  From my reading, this will cause
> > enable_nonboot_cpus() to panic.  Is that intended?
> 
> Andrew,
> 
> I wanted to give you an update on results of cpu testing I've done on
> recent kernels and several architectures.  Since -rc1 is out, I wanted
> to give added visibility to the few issues that remain.
> 
> The full results are available here:
> 
>     http://crucible.osdl.org/runs/hotplug_report.html
> 
> This is actually a report for cpu hotplug tests generated hourly,
> however we run it against all of the kernel -git snapshots posted to
> kernel.org.  Whereever you see a blank square, it indicates the kernel
> either failed to build or boot.
> 

Can you describe the nature of the cpu-hotplug tests you're running?  I'd
be fairly staggered if the kernel was able to survive a full-on cpu-hotplug
stress test for more than one second, frankly.  There's a lot of code in
there which is non-hotplug-aware.  Running a non-preemptible kernel would
make things appear more stable, perhaps.

iirc Pavel did some testing a month or two ago and was seeing userspace
misbehaviour?

> 
> Issues were found in four areas: General kernel, cpu hotplug, sysstat,
> and the test harness itself.
>

It's surprising that AMD and Intel CPUs behave differently.  Also a good
start on diagnosing things.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[Index of Archives]     [Kernel Newbies]     [Netfilter]     [Bugtraq]     [Photo]     [Stuff]     [Gimp]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Video 4 Linux]     [Linux for the blind]     [Linux Resources]
  Powered by Linux