Re: ez.c in 2.1.74

Grant R. Guenther (
Sat, 20 Dec 1997 10:23:23 -0500 (EST)

> Shouldn't the following function in ez.c pass inode->i_sb to the
> invalidate_inodes() function?
> Does the code even work the way it currently is (I don't have access to a
> Syquest drive to test)?

No it doesn't. It ceased working when someone (no trace of who) threw
__initfunc all through the driver without regard for what the functions
do. You are correct that ez_release() is also wrong.

The ez driver is actually quite obsolete. It was current at the time 1.3
was frozen for 2.0, but owing to some communication problems, it missed
inclusion and was dropped into 2.1 instead. Since 90% of the users of the
parallel port devices are using 2.0, all development has gone on outside
the official kernel distribution. About six months ago, 'ep' was released -
it's a much more general driver that handles the EZ-drives as well as
a number of other products based on Shuttle's EPAT adapter. But even that
is obsolete.

I announced the new "paride" subsystem here a week ago. This subsystem
supports eleven different parallel port IDE protocols and it contains
high-level drivers for IDE disks, ATAPI cdroms and disks. Parallel port
ATAPI tape drivers are in preparation and I expect to have a parallel port
ATAPI generic driver (which has an interface identical to the SCSI generic
driver) to support the new crop of parallel port CD-RW devices.

Providing the beta testing currently in progress is successful, I expect
to submit most of the paride module for inclusion within the next ten days.

Grant R. Guenther