Re: kernel BUG at drivers/xen/grant-table.c:923!

[Ian Campbell <ian.campbell@citrix.com>]

On Fri, 2015-04-17 at 11:31 +0100, David Vrabel wrote:
> On 17/04/15 11:28, Ian Campbell wrote:
> > (Was Re: [osstest test] 50423: regressions - FAIL)
> > 
> > This cropped up in an osstest flight (the results only go to Ian and I).
> > 
> > On Fri, 2015-04-17 at 11:10 +0100, Ian Campbell wrote:
> >> On Thu, 2015-04-16 at 21:36 +0000, osstest service user wrote:
> >>> flight 50423 osstest real [real]
> >>> http://logs.test-lab.xenproject.org/osstest/logs/50423/
> >>>
> >>> Regressions :-(
> >>>
> >>> Tests which did not succeed and are blocking,
> >>> including tests which could not be run:
> >>>  test-armhf-armhf-xl          15 guest-start/debian.repeat fail REGR. vs. 50331
> >>
> >>
> >> Looks to have been a real issue:
> >>
> >> http://logs.test-lab.xenproject.org/osstest/logs/50423/test-armhf-armhf-xl/serial-cubietruck-gleizes.log
> >> Apr 14 19:01:21.834851 [ 1229.470934] ------------[ cut here ]------------
> >> Apr 14 19:01:22.445308 [ 1229.471001] kernel BUG at drivers/xen/grant-table.c:923!
> > 
> > This seems familiar, anyone remember it? (CCing ARM, netback and generic
> > Linux maintainers).
> 
> This doesn't look familiar to me.  What BUG_ON() is this?

The one in gnttab_batch_copy:

        void gnttab_batch_copy(struct gnttab_copy *batch, unsigned count)
        {
                struct gnttab_copy *op;
        
                if (HYPERVISOR_grant_table_op(GNTTABOP_copy, batch, count))
                        BUG();
                for (op = batch; op < batch + count; op++)
                        if (op->status == GNTST_eagain)
                                gnttab_retry_eagain_gop(GNTTABOP_copy, op,
                                                        &op->status, __func__);
        }
        EXPORT_SYMBOL_GPL(gnttab_batch_copy);
        
The actual hypercall has failed, which is a pretty unusual failure case.
r0 == fffffff2 is likely the return code, which is -14 == EFAULT.

So I suppose something is up with the pointer at batch.

Ian.