Re: [PATCH 2/6] mm: mlocking in try_to_unmap_one

[KOSAKI Motohiro]

> On Fri, Nov 13, 2009 at 05:26:14PM +0900, KOSAKI Motohiro wrote:
> > Probably we can remove VM_NONLINEAR perfectly. I've never seen real user of it.
> 
> Do you mean as a whole or in the mlock logic? databases are using
> remap_file_pages on 32bit archs to avoid generating zillon of vmas on
> tmpfs scattered mappings. On 64bits it could only be useful to some
> emulators but with real shadow paging and nonlinear rmap already
> created on shadow pagetables, it looks pretty useless on 64bit archs
> to me.

Hehe, you point out kosaki is stupid and knoledgeless.
thanks correct me.

Probaby we have to maintain VM_NONLINEAR for one or two year. two years
later, nobody use database on 32bit machine.
(low-end DB might be use on 32bit, but that's ok. it doesn't use VM_NONLINEAR)




--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/