Re: Accessing CDs or cd9660 broken

On Mon, 2026-01-19 21:17:24 +0100, Thomas Klausner <[email protected]> wrote:
> On Mon, Jan 19, 2026 at 09:12:15PM +0100, Jan-Benedict Glaw wrote:
> > --> The problem seems to be clearly with NetBSD's cd9660 driver or
> > some underlaying mechanism (block-level access to the SCSI CD
> > interface, ...)  Would be nice if somebody could test with a recent
> > NetBSD current version whether or not CD / cd9660 (with RR
> > extensions?) still work for them? Maybe with my above two images?
> 
> On 11.99.4/x86_64:
> 
> # rump_cd9660 20260118213312_NetBSD-11.99.4-alpha.iso   /home/wiz/foo
> rump_cd9660: "20260118213312_NetBSD-11.99.4-alpha.iso" is a relative path.
> rump_cd9660: using "/home/wiz/20260118213312_NetBSD-11.99.4-alpha.iso" instead.
> [   1.0000000] entropy: ready
> # cd foo
> # ll
> total 11180
> drwxr-xr-x   3 root  wheel      2048 Jan  6 16:21 .
> drwx------  45 wiz   users      2560 Jan 19 21:15 ..
> drwxr-xr-x   4 root  wheel      2048 Jan  6 16:21 alpha
> -rw-r--r--   1 root  wheel     62920 Jan  6 16:21 boot
> -rw-r--r--   1 root  wheel  11377336 Jan  6 16:21 netbsd
> # mkdir /tmp/foo
> # cp -R * /tmp/foo/
> # du -sh /tmp/foo
> 306M    /tmp/foo
> # du -sh .
> 306M    .
> 
> Seems to work.

Thanks for testing!

So the cd9660 fs driver seems to work; I can reproduce that by copying
the CD image to a (PiSCSI simulated) HDD and mount it. That would
happily mount it and copy all contents (from within the installer's
shell) to /dev/null .  So this is either (now only testing on the
Alpha 3000/600, but I've seen the same on VAX) an Alpha (VAX) issue,
an issue with the SCSI chip (interface) or specifically when accessing
CD-ROM drives, as I have this issue with a real Plextor drive as well.
(The drive is jumpered to 512 byte blocks of course.)

[   1.0000000] DEC 3000 - M600, 175MHz
[   1.0000000] 8192 byte page size, 1 processor.
[   1.0000000] total memory = 192 MB
[   1.0000000] (2048 KB reserved for PROM, 190 MB used by NetBSD)
[   1.0000000] avail memory = 176 MB
[   1.0000000] asc0 at tcds0 chip 0: NCR53C94, 40MHz, SCSI ID 5
[   1.0000000] scsibus0 at asc0: 8 targets, 8 luns per target
[   3.1476652] cd0 at scsibus0 target 0 lun 0: <DEC, RRD42    (C) DEC, 4.5d> cdrom removable      <----- PiSCSI-simulated
[   3.2462929] cd0: async, 8-bit transfers
[   3.7746203] sd0 at scsibus0 target 2 lun 0: <SEAGATE, ST118273W, 5764> disk fixed              <----- Spinning rust
[   3.8673853] sd0: 17366 MB, 7501 cyl, 20 head, 237 sec, 512 bytes/sect x 35566480 sectors
[   3.9679705] sd0: sync (100.00ns offset 15), 8-bit (10.000MB/s) transfers, tagged queueing
[   4.0685627] sd1 at scsibus0 target 3 lun 0: <PiSCSI, SCSI HD 2048 MiB, 2512> disk fixed        <----- 2GB PiSCSI disk (ment to become root disk)
[   4.1652422] sd1: 2048 MB, 20971 cyl, 8 head, 25 sec, 512 bytes/sect x 4194304 sectors
[   4.2589879] sd1: async, 8-bit transfers
[   4.3127052] cd1 at scsibus0 target 4 lun 0: <DEC, RRD42   (C) DEC, 4.5d> cdrom removable       <----- Real DEC RRD42
[   4.4090893] cd1: async, 8-bit transfers
[   4.4660236] sd2 at scsibus0 target 6 lun 0: <PiSCSI, SCSI HD 301 MiB, 2512> disk fixed         <----- PiSCSI-simulated disk containing the CD image also available in the real and PiSCSI-simulated RRD42
[   4.5617214] sd2: 301 MB, 3086 cyl, 8 head, 25 sec, 512 bytes/sect x 617360 sectors
[   4.6525428] sd2: async, 8-bit transfers

As HDD access seems to work just fine, this would point to the CD /
MMC driver.

MfG, JBG

--

-----BEGIN PGP SIGNATURE-----

iF0EABECAB0WIQQlDTvPcScNjKREqWEdvV51g5nhuwUCaW6huQAKCRAdvV51g5nh
uy9iAKCKEZKNqXizjS8qNlMQvIYaO4gFtgCeNw4qm4hVNvhyAtkcm98SQMR36BY=
=yi8Z
-----END PGP SIGNATURE-----