On 24 Feb 2001, Thomas Graichen wrote:
> Marcelo Tosatti <marcelo@xxxxxxxxxxxxxxxx> wrote:
> > On Fri, 23 Feb 2001, Steve Lord wrote:
> > <snip>
> >> Other changes coming down the pipeline - probably next week will help too,
> >> we have better dbench and bonnie numbers on some internal code right now.
>
> > It looks we're allocating the page to hold the page pointers for the
> > cluster unconditionally in __pagebuf_write_full_page (my last message
> > talks a bit more about that kind of stuff). This page, as far as I can
> > see, is not used at all without kiobuf io. (Thomas is not using kiobuf IO
> > IIRC).
>
> correct - all my testing was without kio
>
> btw. does the kiobuf stuff now work again without any risk in 2.4.2?
> (i remember there was a problem which was i think somehow related to
> the multiblock think in pio mode which steve wanted to fix for 2.4.2
> ... somewhat in that direction i think it was)
Yes; there was an issue with doing kiobuf I/O via multwrite PIO to IDE
disks. That has been fixed, AFAICT in the 2.4.2 XFS tree. Testers are
always welcome! :^)
Cheers,
-Chait.
|