[Xen-devel] Regression introduced by a3db233 x86emul: use DstEax also for {, I}{MUL, DIV

[Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx>]

Hey Jan,

The commit a3db233 x86emul: use DstEax also for {,I}{MUL,DIV}
introduces an regression when doing SR-IOV passthrough.

Specifically I get:
..
(d1) Invoking ROMBIOS ...
(XEN) stdvga.c:174:d1v0 entering stdvga mode
(d1) VGABios $Id: vgabios.c,v 1.67 2008/01/27 09:44:12 vruppert Exp $
(XEN) realmode.c:119:d1v0 Failed to emulate insn.
(XEN) Real-mode emulation failed: d1v0 Real @ c000:00000154 -> 0f 75 05 e8 b0 
5f eb 7c 80 fc 1a 75 05 e8 d3 6a
(XEN) domain_crash called from realmode.c:165
(XEN) Domain 1 (vcpu#0) crashed on cpu#10:
(XEN) ----[ Xen-4.8-unstable  x86_64  debug=y  Not tainted ]----
(XEN) CPU:    10
(XEN) RIP:    c000:[<0000000000000154>]
(XEN) RFLAGS: 0000000000010082   CONTEXT: hvm guest (d1v0)
(XEN) rax: 0000000000089e00   rbx: 000000000000e0fe   rcx: 000000000008ff04
(XEN) rdx: 00000000000003c6   rsi: 0000000000000009   rdi: 000000000000b5ba
(XEN) rbp: 0000000000002040   rsp: 000000000000dadf   r8:  0000000000000000
(XEN) r9:  0000000000000000   r10: 0000000000000000   r11: 0000000000000000
(XEN) r12: 0000000000000000   r13: 0000000000000000   r14: 0000000000000000
(XEN) r15: 0000000000000000   cr0: 0000000000000010   cr4: 0000000000000000
(XEN) cr3: 0000000000800000   cr2: 0000000000000000
(XEN) ds: c000   es: f000   fs: 0000   gs: 0000   ss: 9e00   cs: c000

And only if using qemu-traditional. PV or qemu-xen (SeaBIOS) run just fine.

I spotted this on 'staging' yesterday and git bisected it to this
particular commit. To double-check I built also an hypervisor with
commit d9c91cc domctl: relax getdomaininfo permissions - which worked
fine (see attached logs).

The setup is rather simple: the machine is an SuperMicro X8DTN with E5520 and
two 1GB 82576 NICs. Each NIC gets its SR-IOV functionality enabled meaning
it creates 7 VFS for a total of 14 VFs. Then I create 14 guests with each
guest having one VF. They all start doing netperf [not that we get to this
point].

Attached are the two 'xl dmesg' from good and bad run. More than happy
to test any patches (the testcase is fully automated).

Attachment: a3db233.log
Description: Text document

Attachment: d9c91cc.log
Description: Text document

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel