On Sun, Jun 22, 2008 at 7:11 PM, Roland Dreier <[EMAIL PROTECTED]> wrote:
>
> realistically I think we leave this workaround forever.  it's not like
> it costs very much.

I was thinking about applications which create and destroy HCA
resources at a high rate which might be affected.
>
> Does mlx4 have any similar problem?
>
No, it does not.

>
>  > This fixes the bug reported by Arthur from SGI here:
>  > http://lists.openfabrics.org/pipermail/general/2008-May/050026.html
>
> This should be in the changelog itself... there's no reason to throw
> away this useful information when merging the patch.
OK.
>
>
> Anyway, thanks for debugging this... one question, why not just use the
> following (completely untested) change instead?  Does this not work?
>
> At least using clear_highpage() instead of vmap() by hand seems much
> simpler too if __GFP_ZERO isn't usable for some reason.

Looks to me like using __GFP_ZERO is the cleanest approach. I like
less clear_highpage()  since it uses kmap_atomic() which could fail.
_______________________________________________
general mailing list
[email protected]
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general

To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general

Reply via email to