lkml.org 
[lkml]   [2010]   [Jul]   [14]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH] x86: fix ordering constraints on crX read/writes
On 07/14/2010 05:28 PM, Zachary Amsden wrote:
>
>> static inline void native_write_cr2(unsigned long val)
>> {
>> - asm volatile("mov %0,%%cr2": : "r" (val), "m" (__force_order));
>> + asm volatile("mov %1,%%cr2": "+m" (__force_order) : "r" (val) :
>> "memory");
>> }
>>
>
>
> You don't need the memory clobber there. Technically, this should
> never be used, however.

Yes. I just did it for consistency. Likewise, I didn't pore over the
manuals to work out whether writes to any crX could really have memory
side-effects.

>>
>> static inline unsigned long native_read_cr3(void)
>> {
>> unsigned long val;
>> - asm volatile("mov %%cr3,%0\n\t" : "=r" (val), "=m"
>> (__force_order));
>> + asm volatile("mov %%cr3,%0\n\t" : "=r" (val) : "m"
>> (__force_order));
>> return val;
>> }
>>
>> static inline void native_write_cr3(unsigned long val)
>> {
>> - asm volatile("mov %0,%%cr3": : "r" (val), "m" (__force_order));
>> + asm volatile("mov %1,%%cr3": "+m" (__force_order) : "r" (val) :
>> "memory");
>> }
>>
>> static inline unsigned long native_read_cr4(void)
>> {
>> unsigned long val;
>> - asm volatile("mov %%cr4,%0\n\t" : "=r" (val), "=m"
>> (__force_order));
>> + asm volatile("mov %%cr4,%0\n\t" : "=r" (val) : "m"
>> (__force_order));
>> return val;
>> }
>>
>> @@ -271,7 +286,7 @@ static inline unsigned long
>> native_read_cr4_safe(void)
>> asm volatile("1: mov %%cr4, %0\n"
>> "2:\n"
>> _ASM_EXTABLE(1b, 2b)
>> - : "=r" (val), "=m" (__force_order) : "0" (0));
>> + : "=r" (val) : "m" (__force_order), "0" (0));
>> #else
>> val = native_read_cr4();
>> #endif
>> @@ -280,7 +295,7 @@ static inline unsigned long
>> native_read_cr4_safe(void)
>>
>> static inline void native_write_cr4(unsigned long val)
>> {
>> - asm volatile("mov %0,%%cr4": : "r" (val), "m" (__force_order));
>> + asm volatile("mov %1,%%cr4": "+m" (__force_order) : "r" (val) :
>> "memory");
>> }
>>
>> #ifdef CONFIG_X86_64
>>
>>
>>
>
> Looks good. I really hope __force_order gets pruned however. Does it
> actually?

There's a couple of instances in my vmlinux. I didn't try to track them
back to specific .o files. gcc tends to generate references by putting
its address into a register and passing that into the asms.

J



\
 
 \ /
  Last update: 2010-07-15 02:57    [from the cache]
©2003-2014 Jasper Spaans. Advertise on this site