Bug 112956

Summary: DVD RW +- Firmware overwritten
Product: [openSUSE] SUSE LINUX 10.0 Reporter: Patrick Claven <pat>
Component: InstallationAssignee: Andreas Jaeger <aj>
Status: RESOLVED INVALID QA Contact: Klaus Kämpf <kkaempf>
Severity: Major    
Priority: P5 - None CC: aj, behlert, kernel01, yast2-maintainers
Version: Beta 2   
Target Milestone: ---   
Hardware: i686   
OS: SUSE Other   
Whiteboard:
Found By: Other Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---

Description Patrick Claven 2005-08-25 12:39:39 UTC
After installing the Suse 10.0 Beta 2, my DVD RW ceased functioning. The BIOS
was unable to detect it, even when attempting to boot a windows cd. It was
completely undetected by any method I tried (live cd's, boot floppy's etc),
totally fried. I believe it was due to firmware being overwritten, as I had the
laptop replaced (it was a compaq V4004AP), with an HP Pavilion DV4000 and the
same scenario occurred. Both laptops are centrino 740, 1.73Ghz with integrate
intel 915 graphics etc, and the HP proprietary DVD burner, and both suffered the
same fate. I have now had to pay for a replacement DVD drive on the second
laptop due to my own stupidity in installing this beta product.
Comment 3 Andreas Jaeger 2005-08-25 14:57:15 UTC
Hi Patrick, I have never heard of this.  This surprises me totally.

What kind of DVD is this?  SATA or IDE drive?

Stefan, do you have a similar laptop that you tested a beta on?  Is this known.
Comment 4 Stefan Behlert 2005-08-25 15:16:09 UTC
I don't have any of these models mentioned. I'll see if we have the DVD RW  
drive. We didn't see anything as reported in the bug. Patrick, can you give me 
some more information about the model of the DVD-RW drives? 
Comment 6 Jens Axboe 2005-08-25 17:36:19 UTC
We had an issue like this before, with an older Samsung drive in the 7.x series.
Back then the problem was that the cdrom driver issued a FLUSH_CACHE command to
a non-writer, which had used that opcode for the 'update firmware' command. This
was a silly firmware issue, since the drive really should not have reused that
opcode and because the update firmware logic should be a lot smarter and not
brick the drive if it doesn't receive any data.

This being a DVD-RW, the same problem is unlikely (since it will definitely
understand FLUSH_CACHE as just that), so it's hard to say what is going on
here... With a new such drive at hand, I could software trace it and see where
it goes wrong. Right now I can only guess.
Comment 7 Patrick Claven 2005-08-25 23:50:57 UTC
Hi, sorry, I should have provided more information. The model as near as i can
tell is the GCC-4243N - there is a firmware update patch for it on the HP
website, and i think it's possible i may have been able to recover the drive if
i had left windows on the machine and run their crappy winflash, as they dont
seem to provide a dos utility, and whats worse, is that even my external floppy
drive which worked with my dell inspiron 5150 couldnt be used as a boot option
from the custom boot menu. I believe the DVD drive in both laptops is in fact
Samsung based also. I will continue to dig around for more info if you can tell
me any other specific information you'd like to know. Basically, the installer
ran beautifully up until the point where it rebooted and initialised the kernel
etc. 
Comment 8 Patrick Claven 2005-08-25 23:51:40 UTC
dammit, its also an ide drive
Comment 11 Stefan Behlert 2005-08-26 11:28:49 UTC
Ok, I don't have this device available. I'll raise severity, aj, please adjust  
back if you feel it's not appropriate. 
Comment 14 Stefan Behlert 2005-08-29 16:17:06 UTC
AJ, I assign this to you for further tracking. Without Hardware I see no 
chance to fix this. 
Comment 16 Andreas Jaeger 2005-09-03 14:12:18 UTC
We're trying to get Hardware and analyze the problem.

The descriped behaviour indicates a bug in this hardware as delivered by the
vendor were they do not follow the specification.  This means we would have to
blacklist this drive somehow once we figured out the hardware bug.
Comment 19 Andreas Jaeger 2005-09-13 12:23:01 UTC
We tested this now with the following models:
- GCC-4243N(S05D) FW: 1.07 HW: B MECHA: C TYPE: MCGO
- GCC-4243N(M36C) FW: 1.03 HW: B MECHA: C TYPE: MCBO

and cannot reproduce this on a HP Crystal 1.0.  We could not get exactly the
same models that you have but these are close enough.
Comment 20 Andreas Jaeger 2005-12-21 10:58:41 UTC
NOthing we can do further here.