Re: [Xen-devel] [xen-4.9-testing test] 123343: regressions - FAIL

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 30.05.18 at 07:54, <osstest-admin@xxxxxxxxxxxxxx> wrote:
> flight 123343 xen-4.9-testing real [real]
> http://logs.test-lab.xenproject.org/osstest/logs/123343/ 
> 
> Regressions :-(
> 
> Tests which did not succeed and are blocking,
> including tests which could not be run:
>  test-amd64-amd64-rumprun-amd64  7 xen-boot               fail REGR. vs. 
> 123122
>  test-amd64-amd64-xl-qemut-ws16-amd64  7 xen-boot         fail REGR. vs. 
> 123122
>  test-amd64-i386-xl-qemuu-ovmf-amd64  7 xen-boot          fail REGR. vs. 
> 123122
>  test-amd64-i386-xl            7 xen-boot                 fail REGR. vs. 
> 123122
>  test-amd64-i386-xl-qemut-stubdom-debianhvm-amd64-xsm 7 xen-boot fail REGR. 
> vs. 123122
>  test-amd64-amd64-migrupgrade 11 xen-boot/dst_host        fail REGR. vs. 
> 123122
>  test-amd64-amd64-libvirt-vhd  7 xen-boot                 fail REGR. vs. 
> 123122
>  test-amd64-amd64-qemuu-nested-intel  7 xen-boot          fail REGR. vs. 
> 123122
>  test-amd64-amd64-xl-multivcpu  7 xen-boot                fail REGR. vs. 
> 123122
>  test-amd64-amd64-xl-qemuu-ws16-amd64  7 xen-boot         fail REGR. vs. 
> 123122
>  test-arm64-arm64-libvirt-xsm  6 xen-install              fail REGR. vs. 
> 123122
>  test-amd64-i386-xl-qemut-ws16-amd64  7 xen-boot          fail REGR. vs. 
> 123122
>  test-arm64-arm64-xl-xsm       6 xen-install              fail REGR. vs. 
> 123122
>  test-amd64-amd64-xl           7 xen-boot                 fail REGR. vs. 
> 123122

Here and presumably everywhere else:

(XEN) Assertion 'va >= XEN_VIRT_START' failed at 
...3.build-amd64/xen/xen/include/asm/x86_64/page.h:67

with

(XEN) Xen call trace:
(XEN)    [<ffff82d080283c50>] write_ptbase+0x7d/0x1b9
(XEN)    [<ffff82d0802d57eb>] cpu_init+0x7d/0xa0
(XEN)    [<ffff82d080406b1c>] trap_init+0xf0/0x117
(XEN)    [<ffff82d080403bfd>] __start_xen+0x1ffd/0x250d
(XEN)    [<ffff82d0802000f3>] __high_start+0x53/0x55

I admit I did only run release build tests so far.

Considering

void write_ptbase(struct vcpu *v)
{
    struct cpu_info *cpu_info = get_cpu_info();
    unsigned long new_cr4;

    new_cr4 = (is_pv_vcpu(v) && !is_idle_vcpu(v))
              ? pv_guest_cr4_to_real_cr4(v)
              : ((read_cr4() & ~(X86_CR4_PCIDE | X86_CR4_TSD)) | X86_CR4_PGE);

    if ( is_pv_vcpu(v) && v->domain->arch.pv_domain.xpti )
    ...

I assume we're wrongly setting xpti for the idle domain. I didn't figure yet
why this would be the case in 4.9, but not in 4.10.

Jan





_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel