Re: 2.6.10-rc1-bk page allocation failure. order:2, mode:0x20

[Brad Campbell]

Nick Piggin wrote:
> Brad Campbell wrote:
>
> > G'day all,
> >
> > I'm still getting quite a lot of these come up in the logs when the 
> > system is under mild load.
> > I suspect it might have something to do with running an MTU of 9000 on 
> > the main ethernet port which is directly feeding a workstation with an 
> > NFS root (and thus gets quite a high load at times)
> >
> > It's not so much an issue but it does cause the workstation to stall 
> > for up to a second while it waits for data every time it occurs.
> >
> > The loaded ethernet port is this one on an PCI card
> >
> > 0000:00:0d.0 Ethernet controller: Marvell Technology Group Ltd. Yukon 
> > Gigabit Ethernet 10/100/1000Base-T Adapter (rev 12)
> >
> > This started rearing its ugly head when I moved from 2.6.5 to 
> > 2.6.9-preX and persists with BK as of about 2 days ago.
>
> There are patches in the newest -mm kernels that should help the
> problem. If you're willing to test them, the feedback would be
> welcome.

Always willing to test specific patches. Can I just grab the broken out patches, or pull some 
specific csets from a bk tree? I'm not particularly keen on running an -mm kernel on this box if I 
can avoid it (It's a server in 24hr use with 2.5TB of data where the backup media is 7,000km away).

Regards,
Brad
-
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/