Suse 8 and maxtor 120GB

vassilis · 08-11-2002, 04:36 AM

I searched the forum and I couldn't find a solution to this one so here it goes.

I have an old motherboard which does not support ATA 133 so I used the maxtor controlled for ATA 133 and installed a Maxtor 120GB HD. on that. During boot time the bios reports no IDE drives and the Maxtor ATA controller detects and reports the drive properly (as expected). When the SuSe 8.0 installer tries to partition the drive however it says that it has a 7.8 Terabytes size!!! And obviously all attempts to partition (even with values that just try to partition the first 120GB) fail miserably. Suse says that teh drive has a non conventional patirtion table and stops.

I know that there is an entry in the Suse database saying if you are trying to install 7.3 use the floppies from 7.2 and the install a customised kernel. But surely there must be a new proper driver somewhere for the Maxtor drives. There have been around for quite some time now. Can somebody help? Windows XP has no problem with the drive (although it needed the latest driver from Maxtor) but I have a problem with Windows XP.

Help please

finegan · 08-11-2002, 03:09 PM

What's the exact model of this controller, it might take a custom kernel as a lot of ata133 controllers weren't in the 2.4.18 kernel that SuSe 8.0 shipped with. 2.4.18 came out in February.

Cheers,

Finegan

vassilis · 08-12-2002, 12:18 AM

from what I can see from the controller board the only meaningful numbering (if you exclude the various serial numbers) is as follows:
Maxtor ATA133
with a PROMISE PDC20269 chip which has a stick on label saying ATA133 V2.20 B10

Vassilis

finegan · 08-12-2002, 09:12 AM

From the make menuconfig of 2.4.18:

CONFIG_BLK_DEV_PDC202XX: x
x x
x Promise Ultra33 or PDC20246 x
x Promise Ultra66 or PDC20262 x
x Promise Ultra100 or PDC20265/PDC20267/PDC20268 x
x x
x This driver adds up to 4 more EIDE devices sharing a single x
x interrupt. This add-on card is a bootable PCI UDMA controller. Since x
x multiple cards can be installed and there are BIOS ROM problems that x
x happen if the BIOS revisions of all installed cards (three-max) do x
x not match, the driver attempts to do dynamic tuning of the chipset x
x at boot-time for max-speed. Ultra33 BIOS 1.25 or newer is required x
x for more than one card. This card may require that you say Y to x
x "Special UDMA Feature". x
x x
x If you say Y here, you need to say Y to "Use DMA by default when x
x available" as well. x
x x
x Please read the comments at the top of x
x drivers/ide/pdc202xx.c. x
x x
x If unsure, say N.

It seems the PDC20268 was the last card supported by the 2.4.18 kernel.

Here's the easiest solution I can think of, I did this same thing for my Highpoint, same problem:

Install with the giganto drive on the normal old IDE controller, which will be /dev/hda.

Download and compile 2.4.19, which has support for the controller, in make menuconfig under ATA/IDE/MFM/RLL support, make certain to compile the support for the card as part of the kernel '*' not just as a module, 'm'. Make certain 2.4.19 boots fine, and look through 'dmesg' to see if it noticed the card.

Switch the drive from the normal old vanilla IDE controller to the Promise card, and then go into BIOS and DISABLE the primary and secondary IDE controllers. They would normally be /dev/hda, hdb, hdc, and hdd, with the promise card starting at hde. With them off, and also set the computer to be able to boot from the promise card, which should be /dev/hda now!

Cheers,

Finegan

vassilis · 08-12-2002, 04:45 PM

Many thanks Finegan,

I will try that. The trouble is that the on board IDE hangs when it tries to identify the drive. But My solution is to use an old drive smaller drive I have as the boot drive install the teh Suse kernel I have then try to compile the newer kernel as you suggested and try to connect the MAXTOR drive to its controller afterwards. Long winded but it may work. I will report back.

Best regards

Vassilis

dirgon · 10-25-2002, 02:17 PM

Quote:

from what I can see from the controller board the only meaningful numbering (if you exclude the various serial numbers) is as follows:
Maxtor ATA133 with a PROMISE PDC20269 chip which has a stick on label saying ATA133 V2.20 B10

Well, I wonder if this is why I am having problems installing RH8 to a fairly new Maxtor 40 gig ATA133 drive on an Asus A7V333 motherboard. The drive is currently connected to the onboard controller and the board sees it as an ATA133 drive (UDMA-6). I wonder if RH8 just does not have support for the ATA133 on that board?

I kept getting errors at the point of the install where it should format and install the packages. I was getting a dialog saying that it could not re-read the partition table: error 16: device or resource busy.

So, I stuck the drive in my server machine, which has an older Asus P3V4X motherboard. I believe it is an ATA66 controller. I was able to fdisk the Maxtor with no problems.

I did an hdparm -i on the drive and udma4 has an asterisk in the results. I am assuming that means it is running in udma4 mode on that machine.

finegan · 10-25-2002, 05:07 PM

What is the onboard controller on that new board, the 2.4.18 kernel that RedHat 8.0 shipped with is about... oh, 7 months old. The drivers for it and an install method for RedHat 8.0 are probably available from the vendors site. Most likely its a Promise or a Highpoint controller.

You guessed dead right: udma4 is ATA66, the asterik is the current running speed, udma6 is ATA133, but only brand spankin new versions of hdparm can read it right... who knows if redhat shipped with that. They're kinda conservative on running anything that might be considered beta.

Cheers,

Finegan

fomyers · 10-27-2002, 02:18 AM

The maxtor drives vary with model, but on many models, the strapping plugs/switches can be used to effect the size reported to the bios during boot. This can get past most hangs due to reported size incompatabilities on older BIOS. I believe that the linux drivers use the partition table to build their picture of the drive, and do not usually use the bios size reported, but it will cause most boot loaders to have a fit. In that case defining a small (/boot) partition at the front of the drive and placing the kernel there will usually bypass that problem.

Good luck.

Frank

pd1acf · 12-19-2002, 07:40 PM

Hi Guys,

I recently installed the same sort of setup.
Older Mainboard (P3 550Mhz) with a new Promise ATA133TX2 card, and a Maxtor 80Gb disk.

Both should support the UDMA6/133 Mhz speed, but when my systems boots up, it is detecting the card + disk nicely, but it only detects it as 33Mhz...(UDMA2)

I tried several things to overrule this :
-in lilo.conf 'append="ide2=ata133 ide3=ata133"
-similar command when booting before kernel is loaded.

But this still does not improve.
Also i should get the warning while booting that i am overiding some things but none of this all.
hdparm -tT /dev/hde gives max speed of +/- 28Mb/s for the disk.

Any help is welcome.

Grtz,

ron

finegan · 12-20-2002, 02:19 AM

What does hdparm give you as far as information on the drive's currently flagged setting, for instance:

Code:

root@tyler:/home/bob# hdparm -i /dev/hde

/dev/hde:

 Model=WDC WD400BB-00AUA1, FwRev=18.20D18, SerialNo=WD-WMA6R2754114
 Config={ HardSect NotMFM HdSw>15uSec SpinMotCtl Fixed DTR>5Mbs FmtGapReq }
 RawCHS=16383/16/63, TrkSize=57600, SectSize=600, ECCbytes=40
 BuffType=DualPortCache, BuffSize=2048kB, MaxMultSect=16, MultSect=16
 CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=78165360
 IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
 PIO modes:  pio0 pio1 pio2 pio3 pio4 
 DMA modes:  mdma0 mdma1 mdma2 
 UDMA modes: udma0 udma1 udma2 udma3 udma4 *udma5 
 AdvancedPM=no WriteCache=enabled
 Drive conforms to: device does not report version:  1 2 3 4 5

is an ata100 drive set at ATA100, or udma5 as that's what's flagged. You have to remember test data pull may not be accurate if a lot of stuff is running... also, you never actually see 133Megs/sec, that's just burst speed, but you should be able to get into the 60s or 70s with that setup...

Cheers,

Finegan

pd1acf · 12-20-2002, 08:21 AM

Hi Finegan,

Here's my output of your request.
I totally understand that i will not be able to reach a full 133Mb/s troughput, but at least the drive should be in the UDAM5 mode.....
I am wondering if it would change something if i would load the driver as a module. sometimes this seems to make a difference...

root@pd1acf:~# hdparm -i /dev/hde

/dev/hde:

Model=Maxtor 6Y060L0, FwRev=YAR41VW0, SerialNo=Y2SCB1YE
Config={ Fixed }
RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=57
BuffType=DualPortCache, BuffSize=2048kB, MaxMultSect=16, MultSect=16
CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=120103200
IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
PIO modes: pio0 pio1 pio2 pio3 pio4
DMA modes: mdma0 mdma1 mdma2
UDMA modes: udma0 udma1 *udma2 udma3 udma4 udma5 udma6
AdvancedPM=yes: disabled (255) WriteCache=enabled
Drive conforms to: : 1 2 3 4 5 6 7

root@pd1acf:~#

Thanx for any help,

ron

pd1acf · 12-20-2002, 08:36 AM

Hi Finigan,

Another interesting thing.
I'am using kernel 2.4.20, and patched it with the 2.4.19-patch from promise, but i'm not able to modularize the driver.
It can only be selected with the '*' , not 'M'....
I'am wondering if i did the right thing with regards to the pathing and the kernel.
Would i really need to use 2.4.19 ?

Grtz,

ron

finegan · 12-20-2002, 08:51 AM

From what I can tell no, and since laziness is a universal constant, it makes sense that Promise didn't include the extra code necessary to modularize the driver. They by default expect you to boot to it, so it makes sense that they built it so you could only make it part of the kernel proper.

Cheers,

Finegan

pd1acf · 12-20-2002, 01:05 PM

Finigan,

What is your reaction to my included '/dev/hdparm -i /dev/hde' ?
Maybe you overlooked it beceause i sent two replys.....

Thnx,

ron