the wonders of SCSI

Gregory G. Woodbury ggw at wolves.uucp
Mon May 28 04:51:14 AEST 1990

cpcahil at virtech.UUCP (Conor P. Cahill) writes:
>ggw at wolves.UUCP (Gregory G. Woodbury) writes:
>>These drives fsck and mount quite well.
>>  The maxtor and adaptec 1542A (intr and dma and such are
>>all correct.) are in place and work (sort of!).  I can mkfs/fsck/fsstat
>>the scsi drive (as /dev/dsk/c1d0s0 or c1d0p0) BUT! no variant of
>>/etc/mount will let me mount the device any where on the filesystem!
>>	mount: /dev/dsk/c1d0s0 no such device
>"no such device" is the ENXIO error number which usually means the 
>device driver for /dev/dsk/c1d0s0 is not configured into the kernel.

	Having just spent most of the night persuing the problem....
A newly configured kernel is in place with the aha1540 driver, includes
all the available filesystem types, etc. etc...  (Kernels are easy to

>I would verify the major/minor device number for the c1d0s0 entry.
>I find it hard to believe that the fsck actually worked on the 
>disk drive through that /dev entry (c1d0s0).

	The strangest part of this whole mess is that the drive can be
mkfs's and fsck's perfectly, mkpart seems to talk to the drive some of
the time and fsstat report the filesystems are fine - they all talk to
the device (evidenced by activity of the appropriate duration.)

	The only program that is complaining about the drive is
/etc/mount!  The first time through, it seems to take a bit longer than
subsequent attempts and they consistently report the "no such device"
(which is nominally the ENXIO - errno #6 "io on special file refers to
subdevice which does not exist or is beyond the limits of the device.")

	I have tried various combinations of c1d0s0 and c1d0p0 and other
things.  At this point my only options may be to back up everything and
start over from scratch!

>Another possibility may be that you set up the file system to be a 2K
>file system and the 2K stuff isn't compiled into the kernel.
	NO.  The 2k stuff is in place in kernal and the fsys was made
with 1K blocks.

>>stringing the mount command is non-revealing.  It doesnt even go look at
>>the device before telling that.
>What mount does do is pass you arguments as appropriate to the mount 
>system call.  It does not directly access your file system.

	That I understand, I may try writing a simplistic test prog or
two to probe the device. 

	Thank you, Connor, for a rapid response,  I hope things are a
bit clearer in the light of day (rather than at the end of a long,
frustrating, session.)

Gregory G. Woodbury @ The Wolves Den UNIX, Durham NC
UUCP: ...dukcds!wolves!ggw   ...mcnc!wolves!ggw           [use the maps!]
Domain: ggw at     ggw%wolves at
[The line eater is a boojum snark! ]           <standard disclaimers apply>

More information about the Comp.unix.i386 mailing list