Very Computer

As you might be able to guess from the title, I blew it; my new
machine that I'm bringing Linux up on has one; when I bought it
I noted "Sony CD-ROM, good, those are supported" and of course
only certain CD-ROMs are supported and the 55E isn't one of them
(at least by Slackware 2.1).

However, after doing a bit of research (CD-ROM HOWTO, hints in
the documentation) it appears that all is not lost: the Sony
55E is an "ATAPI" drive and so is the NEC CDR-260, and there's
an alpha driver for the latter.  This is supposed to be a new
standard CD-ROM interface.  So maybe I'm in luck.

Before I start patching this thing into my kernel, though, I thought
I'd ask here first to see if anyone else has any experience with this.
So, some questions:

1. Did anyone get a Sony CDU-55E working under Linux, with this or
   another driver?

2. Is anyone using the above-mentioned NEC CDR-260 driver with an NEC
   CD-ROM and does it work reasonably well for that one?

Post answers if you like, but I'd like to get mail responses -- I'll
summarize whatever I find out to the net.  If no one answers, I'll
just go ahead and try the sucker.

Thanks in advance.

--

Phone: +1 415 694 1729

Uh, I don't think so....

The NEC-260 is NOT, to the best of my knowledge, ATAPI.  The
NEC-260R is ATAPI and is NOT supported in Linux.  Dell shipped
me one when I ordered a Sony 33A... and it will NOT work with
the NEC260 disk from Slackware.  The NEC260 appparently is another
type of EIDE which Linux supports, but of course is long out of
production.

Wrong.  The nec-260 is not 100% compliant  with the latest ATAPI specs,
but the linux ATAPI driver works just fine with it.  I dunno about the Sonys,
but the new MITSUMI ATAPI drives, as well as the VERSA(?), also work under
the Linux ATAPI driver (part of the (E)ide driver:  ide-3.4.patch.75+.gz).

Please give me a hand here.  There's the NEC-260 and there's the NEC-260R.
The latter claims to be ATAPI, but does not work with the NEC260 boot disk.
I presume it _should_ work with the "ATAPI" driver.  If not, will a Sony 55E?

I'm trying to get a 260R working.  Dell tells me it is IRQ15 and i/o port 1F0.
I have a working source tree of 1.1.74 with the 72+ patch applied.  What device
shouldthe ATAPI 260R be seen as, and what is the LILO command to make it
visible?  I suspect it's hdc but maybe hdc... Dell has "PCI IDE" for the HD and
"ISA HD" for the CDROM, both motherboard jacks with 2-device cables.  Thanks.
___

Bolt Beranek & Newman Inc., Cambridge MA  USA   +1 617 873 3850
Opinions are mine alone; sharing requires permission.

Quote:>Please give me a hand here.  There's the NEC-260 and there's the NEC-260R.
>The latter claims to be ATAPI, but does not work with the NEC260 boot disk.
>I presume it _should_ work with the "ATAPI" driver.  If not, will a Sony 55E?

The 260 and 260R are both `ATAPI', but they implement different versions
of the specification.  The `nec260' driver, which i released in September
(and which i think has been used for boot disks for some distributions)
supports only the old version of the spec which was implemented by the 260.
The newer version of the driver which is included with the EIDE patches
should support both versions.

I've heard mixed stories about the Sony drive.  One person said that he was
able to get it to work, but only by first loading mess-dos and then
warm-booting linux.

Quote:>I'm trying to get a 260R working.  Dell tells me it is IRQ15 and i/o port 1F0.

The usual assignment is primary interface = 1f0 = irq14
                      secondary interface = 170 = irq15

Quote:>I have a working source tree of 1.1.74 with the 72+ patch applied.  What device
>shouldthe ATAPI 260R be seen as, and what is the LILO command to make it
>visible?  I suspect it's hdc but maybe hdc... Dell has "PCI IDE" for the HD and

hda, hdb = master, slave on first controller
hdc, hdd = master, slave on second controller

Thus if the drive is the master on the second controller, it should be
called hdc.  If it's not found by the autoprobe when booting you should
be able to force it by saying `hdc=cdrom' to lilo.

sss

55E?

Quote:>The 260 and 260R are both `ATAPI', but they implement different versions
>of the specification.  The `nec260' driver, which i released in September
>(and which i think has been used for boot disks for some distributions)
>supports only the old version of the spec which was implemented by the 260.
>The newer version of the driver which is included with the EIDE patches
>should support both versions.
>>I'm trying to get a 260R working.  Dell tells me it is IRQ15 and i/o port

1F0.

Quote:>The usual assignment is primary interface = 1f0 = irq14
>                      secondary interface = 170 = irq15
>>I have a working source tree of 1.1.74 with the 72+ patch applied.  What device
>>shouldthe ATAPI 260R be seen as, and what is the LILO command to make it
>>visible?  I suspect it's hdc but maybe hdc... Dell has "PCI IDE" for the HD and
>hda, hdb = master, slave on first controller
>hdc, hdd = master, slave on second controller
>Thus if the drive is the master on the second controller, it should be
>called hdc.  If it's not found by the autoprobe when booting you should
>be able to force it by saying `hdc=cdrom' to lilo.

Well, it's closer, but not quite there yet.

I built a Slackware boot kernel with the 3.2 patches and 1.1.74 kernel.
Autoprobe does nothing, but "ramdisk hdc=cdrom" does indeed show "ide1: cdrom"
during bootup.  During the phase where the second (ramdisk) floppy reads in,
the CDROM actually spins a bit, so something's going on.

However, slackware's install script can't mount it.  Manually mounting it
causes one of two errors.  The lesser one is a multi-line dump concerning
"segmentation fault".  ("Oops".)  This generally returns a # prompt.  The
greater error is to leave the system red-button hung.  (That's kind of usual.)
I tried a few variants of the basic manual mount
mount /dev/hdc1 -t iso9660 -o ro /system_cd
and never got it to read in.

So close, and yet so far away...  Any clues?  Thanks!
___

Bolt Beranek & Newman Inc., Cambridge MA  USA   +1 617 873 3850
Opinions are mine alone; sharing requires permission.

..
<I built a Slackware boot kernel with the 3.2 patches and 1.1.74 kernel.
<Autoprobe does nothing, but "ramdisk hdc=cdrom" does indeed show "ide1: cdrom"
<during bootup.  During the phase where the second (ramdisk) floppy reads in,
<the CDROM actually spins a bit, so something's going on.
<
<However, slackware's install script can't mount it.  Manually mounting it
<causes one of two errors.  The lesser one is a multi-line dump concerning
<"segmentation fault".  ("Oops".)  This generally returns a # prompt.  The
<greater error is to leave the system red-button hung.  (That's kind of usual.)
<I tried a few variants of the basic manual mount
<mount /dev/hdc1 -t iso9660 -o ro /system_cd
<and never got it to read in.

Okay, now read the included README.ide (from the ide-3.2 patch)
and do what *it* tells you to do to mount the cdrom (block=2048).
--

<
<Anybody get the ATAPI drivers to work on the secondary controller of
<a SCSI-less system?

Lots of folks, including the author of ide-cd.c on his own NEC-260.

Email him for help.  His address is in README.ide and ide-cd.c.
--

<
<Please give me a hand here.  There's the NEC-260 and there's the NEC-260R.
<The latter claims to be ATAPI, but does not work with the NEC260 boot disk.
<I presume it _should_ work with the "ATAPI" driver.  If not, will a Sony 55E?

All three apparently work.  Ask Scott for more info if you need it.

<I'm trying to get a 260R working.  Dell tells me it is IRQ15 and i/o port 1F0.

That sounds funny.. 0x1f0 is the primary IDE port, usually on IRQ14,
and is the same port that your IDE disk drive uses, if you have one.
Assuming your cdrom and disk are on the same ribbon cable, on 0x1f0 and IRQ14,
then jumper the cdrom for "slave" and the harddisk for "master".
The cdrom will appear as /dev/hdb

<visible?  I suspect it's hdc but maybe hdc... Dell has "PCI IDE" for the HD and
<"ISA HD" for the CDROM, both motherboard jacks with 2-device cables.  Thanks.

In this case, it would sound as if the cdrom is at 0x170 and IRQ15,
so you should jumper it as "single" or "master", and it should appear
as /dev/hdc

If it doesn't, then help it along with  hdc=cdrom  using LILO.
--

Not that it's easy, since the source tree is elsewhere... but in fact,
I did try that block size option:
mount /dev/cdrom -t iso9660 /cd -o ro,block=2048
(and various other spellings of that line, just in case)

That immediately results in red button hang.  This with /dev/cdrom
linked to /dev/hdc (the default is /dev/scd0) and a valid mount point.
(I tried /system_cd as predefined as well as a *al "/cd".)
If I mention block size, it never even gives the segmentation fault
error.  It just crashes!  Note that both the HD and CDROM drive
do a little spinning during the boot process.  For grins & chuckles
I also tried /hdc1 but it did the same.

Anybody get the ATAPI drivers to work on the secondary controller of
a SCSI-less system?
___

Bolt Beranek & Newman Inc., Cambridge MA  USA   +1 617 873 3850
Opinions are mine alone; sharing requires permission.

1F0.

Quote:>That sounds funny.. 0x1f0 is the primary IDE port, usually on IRQ14,
>and is the same port that your IDE disk drive uses, if you have one.
>Assuming your cdrom and disk are on the same ribbon cable, on 0x1f0 and IRQ14,
>then jumper the cdrom for "slave" and the harddisk for "master".
>The cdrom will appear as /dev/hdb

I checked again.  The DOS driver installs with this line:
DEVICEHIGH /L:1,24320 =C:\NEC_IDE.SYS /D:NECCD0 /N:1 /M:S /P:1F0S /I:15

Note the port 1F0, "slave", and IRQ15 settings.  It is on a different ribbon
cable from the HD.  I'm pretty sure the drive is jumpered "slave" even though
it's alone on the "ISA IDE" ribbon.  No, this doesn't sound logical, but it's
Dell, and that's the way it works in DOS.  I can't break it in DOS just to
get Linux to run.

Quote:>If it doesn't, then help it along with  hdc=cdrom  using LILO.

hdc=cdrom gets it to spin during boot but later to crash hard during mount.
LILO otherwise doesn't locate it.  Given the settings, though, is this really
hdc? hdb?  hdd?  What determines that, the port address?  Perhaps EIDE isn't
all that well standardized yet, or maybe (likely?) Mike Dell "don't need no
s*kin' standards".

Note that the IDE hard drive needs no driver to boot DOS, though it's on
the separate "PCI IDE" ribbon.
___

Bolt Beranek & Newman Inc., Cambridge MA  USA   +1 617 873 3850
Opinions are mine alone; sharing requires permission.