Samsung SC-152A CD-ROM not working with FreeBSD 4.8/5.0/5.1

Per von Zweigbergk pvz at e.kth.se
Sat Jun 14 11:15:14 PDT 2003


First of all, i offer my sincerest apologies if this is sent to an 
inappropriate list. I got no helpful responses to this question / bug 
report neither on the freebsd-current mailing list nor on IRC.

Also, since the web based bug reporting software was offline, I had no 
way to send the bug report, not having a working FreeBSD machine 
anywhere.

I've tried submitting it to freebsd-bugs at freebsd.org, but I received no 
confirmation of the message having been received and recorded in the 
bug tracking database, nor did it turn up on a GNATS search. So, I 
figured it was just "lost in the noise".

So, on the recommendation by a person on IRC, I decided to post it 
here, even though it may not be the best forum to post this in.


Anyway... I have recently built two computers:

1) shiba:

Microstar 865PE-Neo2 LS motherboard
Intel Pentium 4 2.4 GHz processor with 800 MHz bus
Three Intel Pro/100 (fxp) adapters
One cheap Geforce4MX AGP card
Western Digital 120 GB hard drive
Some floppy disk drive
Samsung SC-152A CD-ROM drive

2) danneman:

Asus P4P800 motherbord (also with the same Intel 865PE chipset)
Intel Pentium 4 2.4 GHz processor with 800 MHz bus
One Intel Pro/100 (fxp) adapters
One cheap Geforce4MX AGP card
3ware Escalade 7500-4 IDE RAID card
Three Western Digital 120 GB hard drives in a RAID-5 configuration
Some pretty generic Tekram card for SCSI (can't remember exact model 
number. It seems to work though.)
One HP DAT72 drive connected via SCSI
Some floppy disk drive
Samsung SC-152A CD-ROM drive


The Samsung drives are on a seperate IDE buses, connected with UDMA33 
40-conductor cables to the IDE bus on the motherboard. The CD-ROM 
drives are all properly jumpered for IDE master operation.

The problem I'm experiencing with both these computers is with the 
CD-ROM drive.  Linux has no problem with the drive (as tested using 
Debian GNU/Linux rescue disks as well as GNU parted boot disks).

However, upon attempting to run it with FreeBSD 5.1-RELEASE 
(5.0-RELEASE gave the same results), I got the following error messages 
on the console:

ata1-master: timeout waiting for interrupt
ata1-master: ATAPI identify failed

(Booting verbosely gave no more information.)

And the CD-ROM was nowhere to be found in the system. Even though the 
system had just booted of that very disk.

This in itself however, is not surprising. I have very little knowledge 
about bootable CD-ROM:s, but from what I've understood, (this could be 
wrong, but the general idea is probably valid), you basically put some 
bootstrap code on a floppy image in a special place on the CD, and the 
BIOS "pretends" that the part of the CD is actually a bootable floppy.

This allows FreeBSD to boot off the pseudo-floppy-on-CD without knowing 
how to access the rest of the CD without the help of the BIOS.


All right, so FreeBSD basically can't access the CD-ROM drive. You'd 
generally suspect FreeBSD not liking the IDE controller or the chipset. 
Not so. When the drive is swapped for a CD-ROM drive of another 
(non-)brand, it works flawlessly. The same behaviour is displayed on 
both systems (not surprising, since the motherboards use the same 
chipsets.)


What about 4.8-RELEASE I hear you think? Well, 4.8 didn't really like 
the drive either, but it disliked it in a different way. :-) I don't 
have the exact error messages available, but I believe it just got 
stuck in a long loop complaining about something similar to 
"ATAPI_SEEK_BIG". If you find it relevant, I can make bootdisks for 
FreeBSD 4.8 and check this out and post the exact error messages 
sequence.


Either way, if there is any other pertinent information I can give you 
for resolving this bug, I'll be more than happy to provide it.

And again, I apologise if this e-mail has been sent to an inappropriate 
e-mail address.

-- 
Per von Zweigbergk <pvz at e.kth.se>



More information about the freebsd-hackers mailing list