Matthew Dharm wrote:
>
> Did someone change something between pre4 and pre6? I'm seeing some odd
> behavior when dealing with sd_mod.
>
> Specifically, I have SCSI compiled in, and sd_mod as a module. I manually
> load my virtual HBA (usb-storage) and it detects a device. But, when I go
> to access the device (or if the SCSI layer decides to probe it
> automatically for capacity, etc)., things just stop.
>
> The system is running at this point, but I can't access the device on the
> virtual HBA. The sd_mod module is listed as unitialized by lsmod, and it
> just sits there. The modprobe which is trying to load sd_mod is still
> running. It looks like it's waiting for the module to finish
> initializing.
>
> The strange thing is, if I manually load sd_mod first, it works just fine.
> No problems whatsoever.
WCHAN in "ps axl" should show where it is hanging. If the truncated
WCHAN name is unclear it can be fetched in hex with "ps axln" and you
can search for a nearby label in the System.map or /proc/ksyms .
Not much change in the mid-level in pre4 and pre5. In pre6-2 there is
a new "old_undeflow" flag in the Scsi_Cmnd structure which I assume
Eric Y. added.
Recently I got burnt using devfs and the MODLOAD option in devfsd
that failed in a similar way to what you have described. The solution
was to upgrade to modutils-2.3.10 .
While on the subject of change, has the module dependency between
cdrom and sr_mod been switched around? On my system loading cdrom
used to have the side effect of loading sr_mod (I think), now it
is the other way around (which is probably more sensible).
Doug Gilbert
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to [EMAIL PROTECTED]
