
=== Release notes for ftape-0.9.9d ===

This release will probably fix most of the problems people were having
with the 0.9.9 release. If any problem remains, let me know !

o  The ftape module is locked in memory now so it can't be accidentally
   removed when the drives is in use. You'll need the latest version of
   the modules utilities: modutils-0.99.14.tar.gz, the older ones from
   the ftape distribution won't work anymore !
o  The fdc not ready problems at the end of track are solved.
o  The start-tape code that failed one some (slow starting ?) drives is fixed.
o  Fdc timing is relaxed a bit. There shouldn't be any "fdc_write timeout,
   retry" messages anymore.
o  The kernel_version string in kernel-interface.c is moved into a
   separate include file (kernel-version.h) that's created in the Makefile.
   This way, after a `make clean', a correct version will be built.


=== Release notes for ftape-0.9.9 ===


This version has changed a lot and I haven't had the chance to test
it for a long period, so I'll leave that up to you <grin> !

o  First of all, there are no modules patches anymore, since as of kernel
   pl14d these are included in the standard kernel :-)
o  Another important change has to do with the fdc timing: If you experience
   problems now and the older version did work, let me know and include
   the timing information you get when loading the driver.
o  The ecc code seems to work very reliable now (although there are
   limitations on what it can do). The chance of getting wrong data
   is minimized with a slight tradeoff to reduce the number of the retries.
o  Multiple volume tar dumps work fine (as long as you don't use a 512
   bytes blocksize, but since that doesn't make much sense anyhow I'll
   look into it later).
o  Trace levels have been changed. Now `mt -f fsf 0' can be used to
   disable all output except bugs. Default is level 4 now.

        set it to:     to get:
         0              bugs
         1              + errors
         2              + warnings
         3              + information
         4              + more information
         5              + program flow
         6              + fdc/dma info
         7              + data flow
         8              + everything else

o  Fdc overruns on systems doing heavy dma bursts or other bus mastering
   activities is implemented. If I don't hear anything I'll assume it works.
o  The drives changed name from driver.o into ftape.o.
o  The trace messages now include a modulo 256 sequence number to show
   if any messages are missing (syslog overrun).
o  Ioctl MTRESET is implemented, needs a special version of `mt' though.
o  Write protected tapes will be detected.
o  All userft ioctls are now removed from the code, so even people who
   saved the old userft sources are out of luck now ;-)
o  In general: error recovery has improved much. I tried also to speed up
   the retry code, but this seems to be hardware dependent. If you get
   a lot of rewinds during a retry, let me know (please include the log).
o  Note that the driver should compile without any warnings or errors !
   so if you get any chances are there is something wrong, don't ignore it.
o  There are some compile time options added to the Makefile, most
   importantly a fix for the badly formatted Conner tapes. (Hope it works).
o  Kai Harrekilde-Petersen wrote a ftape FAQ <khp@mic.dth.dk>, thanks Kai !


Bas.


BTW: This morning I got some sample code from Colorado Memory Systems to
     look at and see how the tape formatting should be done.
     Although CMS stated they wouldn't support Linux in the near future,
     they do value a working driver for their drives, thanks Ron !

