Re: [Xen-devel] [PATCH v2] x86: correct create_bounce_frame

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 05.05.17 at 10:41, <andrew.cooper3@xxxxxxxxxx> wrote:
> On 04/05/17 14:35, Jan Beulich wrote:
>> @@ -345,24 +344,30 @@ UNLIKELY_START(z, create_bounce_frame_ba
>>  __UNLIKELY_END(create_bounce_frame_bad_bounce_ip)
>>          movq  %rax,UREGS_rip+8(%rsp)
>>          ret
>> -        _ASM_EXTABLE(.Lft2,  domain_crash_page_fault_32)
>> -        _ASM_EXTABLE(.Lft3,  domain_crash_page_fault_24)
>> -        _ASM_EXTABLE(.Lft4,  domain_crash_page_fault_8)
>> -        _ASM_EXTABLE(.Lft5,  domain_crash_page_fault_16)
>> -        _ASM_EXTABLE(.Lft6,  domain_crash_page_fault)
>> -        _ASM_EXTABLE(.Lft7,  domain_crash_page_fault)
>> -        _ASM_EXTABLE(.Lft12, domain_crash_page_fault_8)
>> -        _ASM_EXTABLE(.Lft13, domain_crash_page_fault)
>> +        _ASM_EXTABLE(.Lft2,  domain_crash_page_fault_6x8)
>> +        _ASM_EXTABLE(.Lft3,  domain_crash_page_fault_5x8)
>> +        _ASM_EXTABLE(.Lft4,  domain_crash_page_fault_4x8)
>> +        _ASM_EXTABLE(.Lft5,  domain_crash_page_fault_3x8)
> 
> Do you perhaps mean to swap the labels for 4 and 5?

Oh, yes, of course. I remember that something was odd when
doing the conversion, but I wasn't able to spot it because things
looked well ordered.

Having made a mistake like this I wonder whether it wouldn't be
better to move these next to their instructions anyway? The
way we build them now both insn and extable generation could
actually be bundled into a macro, eliminating any risk for the
two parts to go out of sync ...

Jan


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