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

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 02.05.17 at 16:12, <andrew.cooper3@xxxxxxxxxx> wrote:
> On 02/05/17 14:22, Jan Beulich wrote:
>> @@ -345,15 +344,20 @@ 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(.Lft2,  domain_crash_page_fault_48)
>> +        _ASM_EXTABLE(.Lft3,  domain_crash_page_fault_40)
>> +        _ASM_EXTABLE(.Lft4,  domain_crash_page_fault_24)
>> +        _ASM_EXTABLE(.Lft5,  domain_crash_page_fault_32)
>> +        _ASM_EXTABLE(.Lft6,  domain_crash_page_fault_16)
>> +        _ASM_EXTABLE(.Lft7,  domain_crash_page_fault_16)
>>          _ASM_EXTABLE(.Lft12, domain_crash_page_fault_8)
> 
> Given that you have altered the notation for a delta from (%rsi) (which
> is a good improvement), these labels should be similarly altered for
> consistency.

Actually, looking at this again I'm no longer sure factoring out the
8 here would be a good idea. I'm also not sure I understand you
saying "Given that you have altered the notation for a delta from
(%rsi)" - the notation didn't change at all, the numeric tag has
always been expressing the adjustment needed for %rsi to
represent the actual failing memory address.

So together with my earlier reply I'm not of the opinion that any
of the adjustments you ask for are actually warranted.

Jan


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