On Thu, 10 Aug 2006, Klaus Muth wrote:
> Dear mailing list, dear Alan,
>
> please forgive me for sending tons of debugging log. Today I learned about
> dmesg -c, which helps a lot...
>
> Am Donnerstag, 10. August 2006 08:11 schrieb Klaus Muth:
> > > Anyway, here's something you can try that might help. Before plugging in
> > > the device, do this:
> > >
> > > modprobe scsi-mod
> > > echo 'SRE:OPTAC Tool:0x80000' >/proc/scsi/device_info
> > >
> > > If it still doesn't work, please provide another verbose usb-storage
> > > debugging log.
> >
> > NoGo. I'll provide full log with CONFIG_USB_DEBUG=y later
> Ok. dmesg output after
> # modprobe scsi-mod
> # echo 'SRE:OPTAC Tool:0x80000' >/proc/scsi/device_info
> and plugin of the device:
...
> Hmmm. As far I can see, this ALLOW_MEDIUM_REMOVAL is issued despite of the
> blacklist entry. So:
> # cat /proc/scsi/device_info | grep OPTAC
> 'SRE ' 'OPTAC Tool ' 0x80000
> Maybe OPTAC Tool does not match because of the spaces?
The spaces shouldn't cause a problem; they are supposed to be there.
This should have worked. I don't understand why the command was issued.
If you feel up to a little kernel debugging, you can try to figure out
the reason. Here are the things you need to know. Most of the files
mentioned below are in the drivers/scsi/ directory of the kernel source.
When you do the "echo" command, it is handled by scsi_devinfo.c:
proc_scsi_devinfo_read(). This much worked correctly, as verified by your
"cat" command.
When the device is detected, the system calls scsi_devinfo.c:
scsi_get_device_flags(). This routine should have found the entry you
created. The interesting part of the code is at the end, the case where
devinfo->compatible is false. If you put a printk statement there, you
should be able to see what happens when the device's information is
compared with your entry.
That routine is called from scsi_scan.c:scsi_probe_lun(), the line that
says
*bflags = scsi_get_device_flags(sdev, &inq_result[8],
&inq_result[16]);
This should set *bflags to the 0x80000 value that you wrote. Again, you
can test that. The value is used later in scsi_add_lun(), where the code
says:
if (*bflags & BLIST_NOT_LOCKABLE)
sdev->lockable = 0;
(Note that include/scsi/scsi_devinfo.h defines BLIST_NOT_LOCKABLE as
0x80000.) So sdev->lockable should be set to 0.
Finally, sdev->lockable is used in scsi_ioctl.c:scsi_set_medium_removal().
It's tested right at the start of the routine, and when the value is 0 the
routine will return immediately without sending the PREVENT-ALLOW command.
Again, you can test that.
Alan Stern
-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
linux-usb-devel@lists.sourceforge.net
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
