Re: Strange things happen when reading from CD-ROM
Steven Dickenson <firstname.lastname@example.org> writes:
> > The IDE-Devices connected to it are: hda: Traxdata CDRW2260+,
> > ATAPI CDROM drive hdb: TOSHIBA CD-ROM XM-6002B, ATAPI CDROM drive
> > hdc: IBM-DTLA-307030, ATA DISK drive hdd: SAMSUNG WU32163A
> > (2.16GB), ATA DISK drive
> You have a rather strange drive configuration. I'd normally put the
> hard drives on the primary IDE interface, and load my CD-ROM's on
> the second.
This has "historical" reasons.
> I guess it really doesn't matter, but it's just common practice.
That's what I thought, too. Do you think that connecting the hard
disks to ide0 could make any difference?
> Also, if I'm reading it correctly, you have a nice, fast IBM 30GB
> drive paired with an old, slow Samsung 2GB drive. I'd dump the
> Samsung like a bad habit. It's probably slowing down your entire
> IDE channel everytime you access it.
The Samsung is slave - can the slave slow down the master? And
besides that old Samsung drive is only used for crap like the windows
partition for my little brother. I don't want Windows to befoul my
nice new drive... :-)
It would be mechanically impossible to connect all the drives in any
other way using the chassis and cables I have.
> > Whenever the CD-drives are accessed in some way (reading, burning)
> > at least two strange effects happen:
> > 1) Sound played at 44kHz becomes noisy and literally slows down.
> > Sound played at 22kHz or less isn't affected.
> > 2) When connected to an ISP with the 56k external serial modem,
> > massive communication problems appear. E.g. ping doesn't work (it
> > simply ceases to produce any output). As soon as the CD-ROM
> > access stops, everything works again.
> Have you tried removing the CD Audio cable? I doubt it will make a
> difference, but it's one more thing out of the equation.
I will try it, but the problem doesn't seem to be interference, as the
sound output really gets _slower_ and not just noisy.
> Have you tried using your drives seperatly? Boot up the system
> with only one CD attached and see what happens, then try it with the
I'll try it, but I think I tried it already. What I surely tried was
using another CD I have and this didn't change anything.
> Also, it seems as though you may be experiencing some electrical
> noise problems, since both of these problems deal with very
> sensitive frequencies. Have you tried another power supply?
No, my chassis is very low-cost so this might be worth trying. I'll
just have to get another power supply from somewhere and that might be
> > Maybe changing the PIO-/DMA-Mode could help, but I don't know how
> > to set them for CD-ROM drives (hdparm is only for HDDs).
> You can usually set your drives for PIO mode in the BIOS, or at the
> very least turn off bus-mastering for that channel.
I've tried changing many things in BIOS. Even when I set my drives to
some PIO-mode the kernel still always says
ide0: BM-DMA at 0xe000-0xe007, BIOS settings: hda:DMA, hdb:DMA
ide1: BM-DMA at 0xe008-0xe00f, BIOS settings: hdc:DMA, hdd:DMA
> I know I'm just offering stabs in the dark here, but it's a very
> strange problem.
Yes, it's very strange indeed. The same drives didn't cause any
problems with my good old P200 board (RIP).
> I don't know what tools, extra parts, or knowledge you have at your
> disposal, but I'd do a part-by-part swap out until the I found the
I consider myself a rather experienced programmer and I have quite
often built computers and exchanged parts but I don't have any
> I'd also try it under another OS on a spare hard drive.
I have (or had) win95 on my second drive and I didn't experience any
of the problems there nor did my brother complain about noisy sound
while playing games.
This is why I'm quite sure it must be some software problem. Are
there any tools/ways like hdparm to tweak the CD-drives in Linux?
> Good luck!
Thanks! It's really annoying being unable to play MP3s from CD or
being unable to access the net and CDs at the same time... And I'm
really curious what the reason for such a strange behavior could be.