Subject: Via UDMA5 3/4/5 is not functional!

Hi, guys!

I want to report a problem.

I have an Athlon 900MHz / 256 MB RAM, chipset: VIA VT82c686B, IBM
harddrive (IBM-DTLA-307030).

At first I tried kernel 2.2.16:
- hdparm -u1 -d1 -X69 /dev/hda => I get 36MB/s

Then I tried kernel 2.4.1. I issued exactly the same hdparm command.
i got in syslog the message: "ide0: Speed warnings UDMA 3/4/5 is not
functional"!

What is the problem?

Thanks in advance!

---
Catalin(ux) BOIE
[email protected]
A new Linux distribution: http://l13plus.deuroconsult.ro
http://www2.deuroconsult.ro/~catab


2001-02-22 08:43:22

by Vojtech Pavlik

[permalink] [raw]
Subject: Re: Via UDMA5 3/4/5 is not functional!

On Thu, Feb 22, 2001 at 10:09:29AM +0200, Catalin BOIE wrote:
> Hi, guys!
>
> I want to report a problem.
>
> I have an Athlon 900MHz / 256 MB RAM, chipset: VIA VT82c686B, IBM
> harddrive (IBM-DTLA-307030).
>
> At first I tried kernel 2.2.16:
> - hdparm -u1 -d1 -X69 /dev/hda => I get 36MB/s
>
> Then I tried kernel 2.4.1. I issued exactly the same hdparm command.
> i got in syslog the message: "ide0: Speed warnings UDMA 3/4/5 is not
> functional"!
>
> What is the problem?

The 80-wire cable was not detected. Try this driver ...

--
Vojtech Pavlik
SuSE Labs


Attachments:
(No filename) (569.00 B)
ide-timing.h (9.94 kB)
via82cxxx.c (17.23 kB)
Download all attachments

2001-02-22 15:38:32

by Ricardo Galli

[permalink] [raw]
Subject: Re: Via UDMA5 3/4/5 is not functional!

> Then I tried kernel 2.4.1. I issued exactly the same hdparm command.
> i got in syslog the message: "ide0: Speed warnings UDMA 3/4/5 is not
> functional"!


I had the same problem.

Add
append="ide0=ata66 ide1=ata66 ide0=autotune ide1=autotune hda=autotune
hdb=autotune hdc=autotune"
to lilo.conf.


BTW, I am having now CRC errors in IDE1 on the VIA, IDE0 it's ok at udma5
and 30MB/sec.

Regards,

--ricardo
http://m3d.uib.es/~gallir/

2001-02-22 15:55:46

by Gregory Maxwell

[permalink] [raw]
Subject: Re: Via UDMA5 3/4/5 is not functional!

On Thu, Feb 22, 2001 at 04:38:48PM +0100, Ricardo Galli wrote:
> > Then I tried kernel 2.4.1. I issued exactly the same hdparm command.
> > i got in syslog the message: "ide0: Speed warnings UDMA 3/4/5 is not
> > functional"!
> I had the same problem.
> Add
> append="ide0=ata66 ide1=ata66 ide0=autotune ide1=autotune hda=autotune
> hdb=autotune hdc=autotune"
> to lilo.conf.
>
> BTW, I am having now CRC errors in IDE1 on the VIA, IDE0 it's ok at udma5
> and 30MB/sec.

Are you not using a udma cable? :)

2001-02-22 22:57:58

by Mike Fedyk

[permalink] [raw]
Subject: Re: Via UDMA5 3/4/5 is not functional!

Catalin BOIE wrote:
> i got in syslog the message: "ide0: Speed warnings UDMA 3/4/5 is not
> functional"!
>
> What is the problem?
>
> Thanks in advance!
>
Wasn't dma disabled on all VIA ide chipsets because of sporadic corruption?

2001-02-23 18:17:15

by Matthias Andree

[permalink] [raw]
Subject: Re: Via UDMA5 3/4/5 is not functional!

On Thu, 22 Feb 2001, Mike Fedyk wrote:

> Wasn't dma disabled on all VIA ide chipsets because of sporadic corruption?

That's not the point, the quoted message shows up when you try -X67..
-X69 in hdparm; cable detection seems to not work out properly.

--
Matthias Andree

2001-02-23 19:54:12

by Jasmeet Sidhu

[permalink] [raw]
Subject: Re: Via UDMA5 3/4/5 is not functional!

CRC errors are due to faulty cables. Go buy some cables that are
good. I've had this problem and the only fix is to use better
cables. Infact, If you get really bad cables, on lets say hdc which is on
the same controller as hda, then when the kernel resets the harddrive
giving crc errors hdc, hda will also hang and you will get DMA
timeouts. This will leave your machine in a frozen state, much like the
BSOD of the windows world...

Here is what happens when one of the drives has a bad cable...

In this setup, hdo and hdm are both master IBM DLTA drives connected to a
Promise ATA100 controller. Both drives are masters as there are no slave
drives configured in this setup. The promise controller has 2 ide channels,
ide0 would be /dev/hdm and /dev/hdn and ide1 would be then /dev/hdo and
/dev/hdp. As said above, only masters /dev/hdm and /dev/hdo are
used. When /dev/hdo has DMA problems, mainly CRC, the kernel will try to
fix the controller by sending a reset, which also takes /dev/hdm down with it.

Feb 20 22:37:29 bertha kernel: hdo: dma_intr: status=0x51 { DriveReady
SeekComplete Error }
Feb 20 22:37:29 bertha kernel: hdo: dma_intr: error=0x84 { DriveStatusError
BadCRC }
Feb 20 22:37:29 bertha kernel: hdo: dma_intr: status=0x51 { DriveReady
SeekComplete Error }
Feb 20 22:37:29 bertha kernel: hdo: dma_intr: error=0x84 { DriveStatusError
BadCRC }
Feb 20 22:37:29 bertha kernel: hdo: dma_intr: status=0x51 { DriveReady
SeekComplete Error }
Feb 20 22:37:29 bertha kernel: hdo: dma_intr: error=0x84 { DriveStatusError
BadCRC }
Feb 20 22:37:29 bertha kernel: hdo: dma_intr: status=0x51 { DriveReady
SeekComplete Error }
Feb 20 22:37:29 bertha kernel: hdo: dma_intr: error=0x84 { DriveStatusError
BadCRC }
Feb 20 22:37:29 bertha kernel: hdo: dma_intr: status=0x51 { DriveReady
SeekComplete Error }
Feb 20 22:37:29 bertha kernel: hdo: dma_intr: error=0x84 { DriveStatusError
BadCRC }
Feb 20 22:37:29 bertha kernel: hdo: dma_intr: status=0x51 { DriveReady
SeekComplete Error }
Feb 20 22:37:29 bertha kernel: hdo: dma_intr: error=0x84 { DriveStatusError
BadCRC }
Feb 20 22:37:29 bertha kernel: ide7: reset: success
Feb 20 22:37:49 bertha kernel: hdm: timeout waiting for DMA
Feb 20 22:37:49 bertha kernel: ide_dmaproc: chipset supported
ide_dma_timeout func only: 14
Feb 20 22:37:49 bertha kernel: hdm: irq timeout: status=0x52 { DriveReady
SeekComplete Index }
Feb 20 22:37:59 bertha kernel: hdm: timeout waiting for DMA
Feb 20 22:37:59 bertha kernel: ide_dmaproc: chipset supported
ide_dma_timeout func only: 14
Feb 20 22:37:59 bertha kernel: hdm: irq timeout: status=0x50 { DriveReady
SeekComplete }
Feb 20 22:38:19 bertha kernel: hdm: timeout waiting for DMA
Feb 20 22:38:19 bertha kernel: ide_dmaproc: chipset supported
ide_dma_timeout func only: 14
Feb 20 22:38:19 bertha kernel: hdm: irq timeout: status=0x50 { DriveReady
SeekComplete }
Feb 20 22:38:39 bertha kernel: hdm: timeout waiting for DMA
Feb 20 22:38:39 bertha kernel: ide_dmaproc: chipset supported
ide_dma_timeout func only: 14
Feb 20 22:38:39 bertha kernel: hdm: irq timeout: status=0x50 { DriveReady
SeekComplete }
Feb 20 22:38:39 bertha kernel: hdm: DMA disabled
Feb 20 22:38:39 bertha kernel: ide6: reset: success


At 04:38 PM 2/22/2001 +0100, Ricardo Galli wrote:
> > Then I tried kernel 2.4.1. I issued exactly the same hdparm command.
> > i got in syslog the message: "ide0: Speed warnings UDMA 3/4/5 is not
> > functional"!
>
>
>I had the same problem.
>
>Add
>append="ide0=ata66 ide1=ata66 ide0=autotune ide1=autotune hda=autotune
>hdb=autotune hdc=autotune"
>to lilo.conf.
>
>
>BTW, I am having now CRC errors in IDE1 on the VIA, IDE0 it's ok at udma5
>and 30MB/sec.
>
>Regards,
>
>--ricardo
>http://m3d.uib.es/~gallir/
>
>-
>To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>the body of a message to [email protected]
>More majordomo info at http://vger.kernel.org/majordomo-info.html
>Please read the FAQ at http://www.tux.org/lkml/


- - -
Jasmeet Sidhu
Unix Systems Administrator
ArrayComm, Inc.
[email protected]
http://www.arraycomm.com