[paranoia] SEMI-SOLVED many scsi timeout errors

[Ian Pilcher]

Jack Pistachio wrote:
> Well, after reading some of the archives, I decided to
> apply the same solution Ian did for the 686B Southbride. 
> That is, I put the CDRW on the Promise controller.  It
> worked... thanks to Ian for his posts.  I thought that
> solution might apply only to older Via chipsets like the
> KT133, but Via's woes with CDROMs and Linux seems to
> continue.
> I'm not to happy about this, seeing as how a pretty penny
> was originally spent to get the best reviewed motherboard
> and chipset available at the time.  It was a couple of
> months before I could solve this problem.  Also, the
> Promise controller is not where I would like my CDRW to be.
>  There it can only be used in PIO modes since the CDRW
> isn't a UDMA 100/133 drive, which are the DMA modes the
> controller supports.  It also take away the space for my
> RAID array.  There are other reasons too, but they are more
> minor.
> Now I suppose the question is: for anyone that has had
> success with the via chipset, what is your setup?  Can you
> get a CDROM to rip on it?  And when will there be more
> support for Via, especially in the kernel?  
> 
> Why, oh why do companies refuse to support the future of
> Linux when it is in their own good?
> 

Just FYI, I continue to fiddle around with this on occasion, usually
when Red Hat releases a new kernel.  My current setup is as follows:

VIA VT82C686B:

     ide0:   hda:  WDC WD300AB-00BPA1  (Western Digital 30GB HDD)
             hdb:  (none)

     ide1:   hdc:  CW038D ATAPI CD-R/RW  (CyberDrv CD-R/W)
             hdd:  (none)

Maxtor Ultra133TX2 (Promise PDC20269):

     ide2:   hde:  CREATIVE CD5233E-N  (Creative CD-ROM)
             hdf:  (none)

     ide3:   hdg:  WDC WD1200BB-00CAA1  (Western Digital 120GB HDD)
             hdh:  (none)

Maxtor Ultra133TX2 (Promise PDC20269):

     ide4:   hdi:  Maxtor 4G120J6  (Maxtor 120GB HDD)
             hdj:  (none)

     ide5:   hdk:  Maxtor 4G120J6  (Maxtor 120GB HDD)
             hdl:  (none)

Although it's somewhat aggravating that I can't use DMA with hde, it's
not really a problem, since Linux doesn't use DMA for DAE anyway (at
least on ATAPI devices), and that's all I use that drive for.

The really interesting thing is that ripping with cdparanoia works just
fine from either hdc or hde in this configuration.  When I discovered
this (running Red Hat's 2.4.18-19.8.0smp kernel), I thought that the
problem had been fixed, and I swapped hda and hde.  As soon as I tried
ripping (from hdc?) in the swapped configuration, however, the lost
interrupts were back in all their glory.

So ripping from hdc seems to be dependant on what type of device is on
hda.  This just confirms my general impression that this VIA chipset is
a POS, and like you I'm pretty irked about it.

I'm really not sure what I'm going to do when the time comes to upgrade.
When Intel killed SMP capability in the Pentium line, I figured I that I
would eventually end up with some kind of a dual-Athlon system -- and
the x86-64 should be available in 12 to 18 months.  The stability I've
always gotten from Intel chipsets keeps calling me back, however.  If a
dual Xeon is remotely reasonable at the time, I may well feel compelled
to go in that direction.

In the end, of course, it may not matter.  cdparanoia is so dated that I
find myself having to boot Windows NT and run EAC to rip quite a few
CDs.


-- 
========================================================================
Ian Pilcher                                           pilchman at attbi.com
========================================================================

<p>--- >8 ----
List archives:  http://www.xiph.org/archives/
Paranoia homepage: http://www.xiph.org/paranoia/
To unsubscribe from this list, send a message to 'paranoia-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.