Re: [Xen-devel] [bug]xen 4.10 + dom0 4.15 couldn't boot up

[Juergen Gross <jgross@xxxxxxxx>]

On 01/02/18 10:14, Jan Beulich wrote:
>>>> On 01.02.18 at 08:29, <jgross@xxxxxxxx> wrote:
>> On 01/02/18 07:20, Zhang, Xiong Y wrote:
>>> (XEN) Domain 0 (vcpu#0) crashed on cpu#0:
>>> (XEN) ----[ Xen-4.11-unstable  x86_64  debug=y   Not tainted ]----
>>> (XEN) CPU:    0
>>> (XEN) RIP:    e033:[<ffffffff8103ca29>]
>>> (XEN) RFLAGS: 0000000000000292   EM: 1   CONTEXT: pv guest (d0v0)
>>> (XEN) rax: 0000000000000000   rbx: ffffffff81e05fe0   rcx: 0000000000000000
>>> (XEN) rdx: 0000000000000030   rsi: ffffffff82203f04   rdi: ffffffff82451f60
>>> (XEN) rbp: ffffffff82203f08   rsp: ffffffff82203e20   r8:  ffffffff82203f08
>>> (XEN) r9:  00000000ffffffff   r10: ffffffff82203f0c   r11: 0000000000000000
>>> (XEN) r12: ffffffff82203f0c   r13: ffffffff82203e88   r14: ffffffff82203f00
>>> (XEN) r15: ffffffff82203e98   cr0: 000000008005003b   cr4: 00000000003526e0
>>> (XEN) cr3: 00000003facc5000   cr2: 0000000000000028
>>> (XEN) fsb: 0000000000000000   gsb: 0000000000000000   gss: 0000000000000000
>>> (XEN) ds: 0000   es: 0000   fs: 0000   gs: 0000   ss: e02b   cs: e033
>>> (XEN) Guest stack trace from rsp=ffffffff82203e20:
>>> (XEN)    0000000000000000 0000000000000000 0000000000000000 ffffffff8103ca29
>>> (XEN)    000000010000e030 0000000000010092 ffffffff82203e68 000000000000e02b
>>> (XEN)    0000000000000000 0000000000000000 0000000000000000 0000000000000000
>>> (XEN)    ffffffff82451f60 0000000000000000 0000000000000000 0000000000000000
>>> (XEN)    0000000000000000 0000000000000000 ffffffff82451f60 ffffffff82203f08
>>> (XEN)    ffffffff82203f0c ffffffff82203f04 ffffffff82203f00 ffffffff82203f1c
>>> (XEN)    ffffffff8103d611 ffffffff82203f10 ffffffff82203f14 ffffffff82203f18
>>> (XEN)    0000000000003027 0000000000000000 0000000080000008 0000000000000000
>>> (XEN)    0000000000000000 0000000000000000 0000000000000000 0000000000000000
>>> (XEN)    0000000000000000 0000000000000000 ffffffff8281ba11 0000000000000000
>>> (XEN)    0000000000000000 0000000000000000 0000000000000000 0000000000000000
>>> (XEN)    0000000000000000 0000000000000000 0000000000000000 0000000000000000
>>> (XEN)    0000000000000000 0000000000000000 0000000000000000 0000000000000000
>>> (XEN)    0000000000000000 0000000000000000 0000000000000000 0000000000000000
>>> (XEN)    0000000000000000 0000000000000000 0000000000000000 0000000000000000
>>> (XEN)    0f00000060c0c748 ccccccccccccc305 cccccccccccccccc cccccccccccccccc
>>> (XEN)    cccccccccccccccc cccccccccccccccc cccccccccccccccc cccccccccccccccc
>>> (XEN)    cccccccccccccccc cccccccccccccccc cccccccccccccccc cccccccccccccccc
>>> (XEN)    cccccccccccccccc cccccccccccccccc cccccccccccccccc cccccccccccccccc
>>> (XEN)    cccccccccccccccc cccccccccccccccc cccccccccccccccc cccccccccccccccc
>>> (XEN) Hardware Dom0 crashed: rebooting machine in 5 seconds.
>>> (XEN) APIC error on CPU0: 40(00)
>>> (XEN) Resetting with ACPI MEMORY or I/O RESET_REG.
>>>
>>> addr2line -f -e vmlinux ffffffff8103ca29
>>> init_scattered_cpuid_features
>>> arch/x86/kernel/cpu/scattered.c:38
>>
>> And this is the problem: it seems as if the compiler now generates an
>> access via %gs in this function, but this segment hasn't been setup at
>> this stage of the boot process.
>>
>> I'm trying to move up initialization of the %gs segment.
> 
> This can be done _really_ early, as our old XenoLinux kernel shows
> (4.4 based):
> 
> startup_64:
>       movq $(init_thread_union+THREAD_SIZE-8),%rsp
> 
>       /* rsi is pointer to startup info structure.
>          pass it to C */
>       movq %rsi,%rdi
> 
>       /* Set up %gs.
>        *
>        * The base of %gs always points to the bottom of the irqstack
>        * union.  If the stack protector canary is enabled, it is
>        * located at %gs:40.  Note that, on SMP, the boot cpu uses
>        * init data section till per cpu areas are set up.
>        */
>       movl    $MSR_GS_BASE,%ecx
>       movq    $INIT_PER_CPU_VAR(irq_stack_union),%rax
>       cdq
>       wrmsr
> 
> I won't make any guarantees on the applicability of everything the
> comment says, though.

Thanks for the confirmation that I'm on the right track. :-)


Juergen

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