Il 18/06/2013 16:38, Richard Henderson ha scritto:
+#ifndef atomic_read
+#define atomic_read(ptr) (*(__typeof__(*ptr) *volatile) (ptr))
#endif
+#ifndef atomic_set
+#define atomic_set(ptr, i) ((*(__typeof__(*ptr) *volatile) (ptr)) =
(i))
+#endif
On 06/18/2013 01:36 AM, Paolo Bonzini wrote:
>> Why all the ifdefs? If __atomic support is present, then __ATOMIC defines
>> will
>> exist.
>
> Then I can just use "#ifdef __ATOMIC_RELAXED" instead of
> "#if QEMU_GNUC_PREREQ(4, 8)"?
I'd say so.
> I have no idea, but I can say which semantics I
Il 18/06/2013 10:36, Paolo Bonzini ha scritto:
> Important Note: Note that it is important for both threads to access
> the same volatile variable in order to properly set up the happens-before
> relationship. It is not the case that everything visible to thread A
> when it writes v
Il 17/06/2013 20:57, Richard Henderson ha scritto:
> On 06/16/2013 04:21 AM, Liu Ping Fan wrote:
>> +#if QEMU_GNUC_PREREQ(4, 8)
>> +#ifndef __ATOMIC_RELAXED
>> +#define __ATOMIC_RELAXED 0
>> +#endif
>
> Why all the ifdefs? If __atomic support is present, then __ATOMIC defines
> will
> exist.
Th
On 06/16/2013 04:21 AM, Liu Ping Fan wrote:
> +#if QEMU_GNUC_PREREQ(4, 8)
> +#ifndef __ATOMIC_RELAXED
> +#define __ATOMIC_RELAXED 0
> +#endif
Why all the ifdefs? If __atomic support is present, then __ATOMIC defines will
exist.
> +/* Compiler barrier */
> +#define barrier() ({ asm volatile(""
From: Paolo Bonzini
We're already using them in several places, but __sync builtins are just
too ugly to type, and do not provide seqcst load/store operations.
Signed-off-by: Paolo Bonzini
---
docs/atomics.txt | 322 +++
hw/display/qxl.c
