Re: [Pan-users] pan 0.14.2 - Freezing

[Budman]

On Tue, 2003-09-09 at 16:47, Jeff Vian wrote:
> For future reference
> 
> On ANY IDE controller, the entire bus runs at the speed of the slowest 
> device. The IDE controller is not capable of running 2 different speed 
> on the same cable.
> 
> Thus, if you have a CDROM on the cable, all devices on that cable will 
> run at 33.
> 

Yep.  I understand that.  SCSI works the same way.  1 channel, 1 speed.

But the IDE drivers for the VIA chipset are very kludgy. Even with the
ATA100 drive, its not running to full potential.  I've always had the
CDROM on a non-embedded controller just so it would not hinder the other
drives, but its the VIA chipset that blows.  VIA sound is even worse -
non-configurable IRQs... that really is not cool.  

Since I changed the drives over to the non-motherboard IDE controller, I
was able to get the speed back.  I tried using the VIA chipset, since it
was ATA100, but it just did not work properly.  The ATA66 card has
faster speed than the ATA100 motherboard controller. 

I read many of the Linux sites, and there are issues with the VIA
chipset not working and not properly switching to the proper speed per
channel.  Prior, I had the both 33 devices on Pri, and then the 66 and
100 on the secondary, now it should downgrade to 66 right... well.. it
would not go higher than 33.  Even when I removed the 66 drive, the 100
would still run at 33. Definite bug in the kernel driver.  Not sure when
it will be addressed.

Now, on the ATA66 controller, Pri is running at 33, secondary is at 66,
and now the embedded controller drive Pri is running 100, and the Sec
cdrom is running 33.  But, the ATA100 is still running slower than the
ATA66 drive on the other controller, and its a 7200RPM!  I think it may
be worth it to get an additional ATA100 controller, and disable the MB
IDE all together. :)


-- 
Budman <address@hidden>

signature.asc
Description: This is a digitally signed message part