Re: [Xen-devel] [PATCH] linux/x86: avoid casting hypercallargume

>>> Keir Fraser <Keir.Fraser@xxxxxxxxxxxx> 21.01.08 17:27 >>>
>On 21/1/08 15:18, "Jan Beulich" <jbeulich@xxxxxxxxxx> wrote:
>
>> -  : "1" ((long)(a1))    \
>> +  : "1" ((a1)?:0)     \
>
>Most of the patch seems to be placing this construct all over the place.
>What's it for? How is x?:0 different from x?

In its type: This way, the type gets promoted to int for anything smaller
than int, but left alone if long (or wider) or pointer. Various gcc versions
appear to have various problems when not promoting to at least int...

>> + register typeof((a4)?:0) __arg4 asm("r10") = (a4); \
>> + register typeof((a5)?:0) __arg5 asm("r8") = (a5); \
>> +    "=d" (__ign3), "+r" (__arg4), "+r" (__arg5) \
>
>Is it guaranteed that gcc will keep the register variables in their assigned
>registers across the inline asm?

That's how they document it, so I suppose this is the canonical way
and can be expected to not break.

Jan

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel

<Prev in Thread]

Current Thread

[Next in Thread>

Previous by Date:

Re: [Xen-devel] [PATCH] linux/x86: avoid casting hypercall arguments to long, Keir Fraser

Next by Date:

[Xen-devel] Re: [PATCH 0 of 2] Make tools cross-friendly, try 2, Christoph Egger

Previous by Thread:

Re: [Xen-devel] [PATCH] linux/x86: avoid casting hypercall arguments to long, Keir Fraser

Next by Thread:

Re: [Xen-devel] [PATCH] linux/x86: avoid casting hypercallarguments to long, Keir Fraser

Indexes:

[Date] [Thread] [Top] [All Lists]