Le lundi 03 mai 2010 à 11:49 +0300, Avi Kivity a écrit :
> A kmalloc()ed page is virtually contiguous, satisfying your requirement.
>
Still, a different function pair is needed, at least for documentation
reasons.
Changli, _many_ patches like yours were suggested and refused in the
past for the reason it should be generic. Maybe its time...
So first introduce a pair of funtions, then we can use them.
At free time we can use is_vmalloc_addr() so that no extra bit is needed
to tell if a zone was vmalloced() or kmalloced()
One remark :
+ data = kmalloc(size, GFP_KERNEL);
If allocation fails (because of fragmentation), we would have a kernel
log.
Still, vmalloc() might be OK.
So you should add a __GFP_NOWARN flag.
/*
* Warning: if size is not a power of two, kmalloc() might
* waste memory because of its requirements.
*/
void *kvmalloc(size_t size)
{
void *ptr = kmalloc(size, GFP_KERNEL | __GFP_NOWARN);
if (ptr)
return ptr;
return vmalloc(size);
}
void kvfree(void *ptr)
{
BUG_ON(in_interrupt());
if (is_vmalloc_addr(ptr))
vfree(ptr);
else
kfree(ptr);
}
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
