Re: [PATCH 7/6] Lost bits - fix PG_writeback vs PG_private race in NFS

[Peter Zijlstra <a.p.zijlstra@chello.nl>]

On Fri, 2006-08-25 at 16:11 -0400, Trond Myklebust wrote:
> On Fri, 2006-08-25 at 18:36 +0200, Peter Zijlstra wrote:
> > Make sure we clear PG_writeback after we clear PG_private, otherwise
> > weird and wonderfull stuff will happen.
> > 
> NACK.
> 
> Look carefully at the case of unstable writes: your patch does nothing
> to guarantee that PG_writeback is cleared after PG_private for that
> case.

Ah, right. Thanks for pointing this out.

> Anyhow, you don't explain exactly what is wrong with clearing
> PG_writeback before PG_private.

Yes, this was a rather hasty patch, I was mortified to find that I
missed a few changes and my patch-set would crash instantly someone
would try it.

The VM doesn't really like PG_private set on PG_swapcache pages, I guess
I'll have to rectify that and leave the NFS behaviour as is.

Will correct this in the next round.

Thanks for the feedback,

Peter

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