Re: 2.2.17 Lockup and ATA-66/100 forced bit set (WARNING)

Michael B. Allen (mballen@erols.com)
Wed, 21 Feb 2001 18:33:13 -0500


On Wed, Feb 21, 2001 at 03:48:06PM -0800, Wayne Whitney wrote:
> > append="idebus=66 ide0=ata66"
>
> The idebus=66 part is incorrect. This option refers to the clock of
> the PCI bus the IDE controller is on and should rarely be changed from
> the default of 33MHz (i.e., only if you are overclocking the PCI bus).

Ah, well I just added that based on the end of this boot message:

ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx

> >kernel: VP_IDE: ATA-66/100 forced bit set (WARNING)!!
>
> I'm sure this is just telling you that you passed the ide0=ata66
> parameter. Usually it is best not to do this--the driver should run
> your chipset/drive as fast as possible without 'forcing' the
> configuration. Of course, testing with hdparm -t is considered the
> definitive way to check how fast the interface is running.

If I don't add it hdparm -t /dev/hda reports:

Timing buffered disk reads: 64 MB in 16.46 seconds = 3.89 MB/sec

Then if I do hdparm -d1 -X66 /dev/hda I get:

/dev/hda:
setting using_dma to 1 (on)
setting xfermode to 66 (UltraDMA mode2)
using_dma = 1 (on)

[root@nano /root]# hdparm -t /dev/hda

/dev/hda:
Timing buffered disk reads: 64 MB in 4.04 seconds = 15.84 MB/sec

Now if I add it via append, reboot and do hdparm -t I get:

Timing buffered disk reads: 64 MB in 3.70 seconds = 17.30 MB/sec

This last difference is consistently better. Weird.

All your other bets were right too.

Thanks Wayne,
Mike

-- 
signature pending
-
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/