lkml.org 
[lkml]   [2006]   [Apr]   [2]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRe: [RESEND][2.6.15] New ATA error messages on upgrade to 2.6.15
Date
Due to Mark Hahn's repeated demands that I leave his replies offlist,  
I have not included the contents of his message, however as I am
_quite_ sure that my problem is not due to failing or misconfigured
hardware, I am sending my own post to the LKML and a few others in
hopes of receiving additional information that can help me debug the
kernel's confusion.

Like I have tirelessly repeated to you before, Mark:

(1) This same behavior occurs on both controllers regardless of
which drive and cable combination is attached to the controller. I
spent several hours when I first set the system up trying every drive
and cable permutation possible, given the 3 drives I had at the time
and the 5 cables from 4 different manufacturers. Each and every time
it was the PDC controller that gave errors for its two drives until
it reset itself, and from that point in boot on everything was fine.
There is no possible way for it to be a cable or drive specific
problem, I have since replaced 2 of the drives one-by-one with
different drives as they failed due to old age, both times using a
new drive and the shipped cable, in every case it gave the same errors.

(2) It's extremely unlikely that the card itself is faulty; it
exhibits identical symptoms on both drives and has ever since I
originally purchased the card and installed 2.4.X on the system.

(3) It's not possible that it's the power supply. As I said before
I checked the power supply with an oscilliscope _during_boot_. The
levels were correct (12V and 5V), and virtually noise-free. The
hardware was originally designed to reliably support the power load
placed on it, it's a 350W power supply with the 3 drives, a low-power
400MHz CPU, a Rage128 GPU, and 800MB or so of PC-100 RAM.

(4) I'm not an idiot, I've administered a whole lab full of
workstations with a wide variety of IDE hardware from the very old to
the very new; I know what I'm doing and how to properly connect the
cables. It's not like it takes a genius anyways given that the host-
side socket is blue and the drive-side one is black. I know how a
single drive always goes at the end of a cable. Etc. I've even had
to deal with old systems with old-style Y-shaped ATA cables that had
a drive at each end; and I know how to tell the difference between them.

(5) The cables themselves run flat against the grounded metal of the
case, the machine itself has little enough EMI as it is; no wireless
chips and a low-speed CPU and bus, the wires are well away from the
power supply. Even if that were the case, I've rearranged the wire
routing many times and _never_ seen a change in behavior, even
removing the CD-ROM drive and placing a drive there had no effect.

If (as Alan Cox suggested) the problem lies with the kernel
incorrectly detecting IDE settings from the controller or not
noticing a buggy firmware, then I'm hoping that with some help I can
determine where the kernel driver is getting confused and correct the
programmed bus timings (or whatever the hell isn't working). As my
practical knowledge of the ATA and IDE protocols is quite limited I'm
well out of my depth here and unable to debug without further
assistance.

Cheers,
Kyle Moffett

-
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/

\
 
 \ /
  Last update: 2006-04-03 14:38    [W:0.042 / U:0.452 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site