[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: ide-scsi -> write_g1: scsi sendcmd: fatal error

Joerg Schilling wrote:

"Thomas Schmitt" <scdbackup@gmx.net> wrote:


Hi,

Fog_Watch:

cdrecord using ide-scsi returns an error that I don't understand.
...
cdrecord: No such device or address. Cannot send SCSI cmd via ioctl.

Joerg Schilling:

This is a Linux kernel problem, please ask the related people.

2.6.24-gentoo-r8

Well, ide-scsi is said to be deprecated with
kernel 2.6.


This is not a helpful reply.


One should at least find out a bit more
before exposing oneself at LKML.

What do you get from this command ?
  cdrecord -scanbus


If the OP did create an own rotten kernel this will not help.
Custom does not imply "rotten" many people build kernels without support things for hardware they lack and features they avoid. 


ide-scsi seems to be buggy and unmaintained in newer Linux releases.
I would say "lightly tested" for sure, based on one machine still using that feature, it seems to work using a ZIP100 drive as a scsi disk.
cdrecord -scanbus on a _vanilla_ Linux kernel will work as well as
There *is* no vanilla kernel, virtually every distribution and developer picks different configuration options to build the kernel, even though the code base is the same.
just calling "cdrecord ..." _without_ dev= parameter Most users have only one drive and in this case, cdrecord will automagically search for the right device.
I doubt your assumption is right about "most" users, but in the case where only a single drive is present the application does identify it and use it. 


This is how the typical output from a cdrecord -scanbus call looks on a
vanilla 2.6.13 Linux kernel:

cdrecord -scanbus
Cdrecord-ProDVD-ProBD-Clone 2.01.01a60 (i686-pc-linux-gnu) Copyright (C) 1995-2009 Jörg Schilling
Linux sg driver version: 3.5.27
Using libscg version 'schily-0.9'.
scsibus0:
        0,0,0     0) 'QUANTUM ' 'ATLAS10K2-TY184L' 'DA40' Disk
        0,1,0     1) *
        0,2,0     2) *
        0,3,0     3) *
        0,4,0     4) *
        0,5,0     5) *
        0,6,0     6) *
        0,7,0     7) *
scsibus1001:
        1001,0,0 100100) 'MITSUMI ' 'CD-ROM FX4830T!B' 'R02J' Removable CD-ROM
        1001,1,0 100101) *
        1001,2,0 100102) *
        1001,3,0 100103) *
        1001,4,0 100104) *
        1001,5,0 100105) *
        1001,6,0 100106) *
        1001,7,0 100107) *

As you see, there is no need to hand craft e linux kernel - cdrecord works
around the oddities in the Linux device addressing ;-)
The usual reason for trimming the configuration is to get a compile done before the next revision of the kernel comes out.
Related question: does cdrecord do the right thing if the only burner is on rscsi? 

--
E. Robert Bogusta
 It seemed like a good idea at the time
Reply to: