Re: [Xen-devel] [PATCH] libx86: Work around GCC bug with ebx output constrants

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 19.11.18 at 15:33, <andrew.cooper3@xxxxxxxxxx> wrote:
> On 19/11/2018 13:51, Jan Beulich wrote:
>>>  static inline void cpuid_leaf(uint32_t leaf, struct cpuid_leaf *l)
>>>  {
>>> -    asm ( "cpuid"
>>> -          : "=a" (l->a), "=b" (l->b), "=c" (l->c), "=d" (l->d)
>>> +    asm ( XCHG_BX
>>> +          "cpuid;"
>>> +          XCHG_BX
>>> +          : "=a" (l->a), BX_CON (l->b), "=c" (l->c), "=d" (l->d)
>> Strictly speaking all other outputs also need to use =& in the
>> 32-bit case. But I wouldn't insist on such an adjustment. With
>> the others done
>> Reviewed-by: Jan Beulich <jbeulich@xxxxxxxx>
> 
> Using =& for the constraints which also have inputs results "operand has
> impossible constraints"

Yes, I had (silently) assumed this would be the case.

> It should only matter for the output-only operands, to prevent GCC
> allocating %[er]dx for the bx constraint, but that shouldn't matter
> anyway because =&r can't be allocated to conflict with =d

Hence me having said "strictly speaking".

Jan



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