=== Releases notes for ftape-1.13b === - Fixed fix that broke error recovery in fdc-isr.c. - Increased head seek timeout after problem report. === Releases notes for ftape-1.13a === - Fixed FC-10 support (wrong DMA channel). - Fixed FLOPPY_HACK requiring extra symbols in ksyms.c. === Releases notes for ftape-1.13 === - Added tape drive offline mode: Once offline the driver will refuse to perform any tape operation until a new tape is inserted. Use mt -f /dev/ftape rewoffl to rewind the tape and soft-eject the cartridge. - Added code implementing status ioctl. `mt -f /dev/ftape status' now returns some usefull (?) information instead of an I/O error. - Changed eof handling code so that eof mark will be skipped when closing device just before eof mark is read. This makes Gnu tar respond `more natural' to file marks. - Added FC-10 activation code. The FC-10 support might or might not work, there seems to be a problem with the data rate selection. I can't test it myself, so FC-10 users, let me know ! - Added support for 82077 (1 Mbps) floppy disk controllers. The driver will now use the highest data rate available ! - Activating the fifo on 82077 fdc's. This should cure the overrun problems caused by bus mastering controllers like the Adaptec 1542-CF. Special thanks to Jon Tombs for finding out how to activate the FC-10 card (If this doesn't work it's probably because I've messed with the code :-) and help with the status ioctl and eof handling ! IMPORTANT: I moved to gcc version 2.5.8 so I won't be able to test if the code from the 2.4.5 compiler is working. Bas, 29/06/94. === Releases notes for ftape-1.12c === - Fixed ioctl reset command: The tape unit gets deselected by a soft reset command (although not specified in qic-117 doc). - Fixed data-rate problem with new Colorado drives: Added set data-rate commands for fdc and tape drive, defaulting to 500 Kbps. - Fixed ftape_ready_wait timing: Timeout was too long because of overhead. - Added new vendor id's: 0x00021 AWAI CT-803 0x0014e Conner C250MQ - Fixed failing multiple file access on Iomega drives: Iomega drives have a firmware bug that makes them return a new-cartridge inserted status after every soft reset. The ftape_open code is changed so that the soft reset is no longer needed. === Release notes for ftape-1.12 === Okay, I finally decided to declare this code beta :-) After a long period of being not able to work on the driver I bought myself a QIC-80 drive last saturday and fixed some remaining problems in the driver. To prevent confusion with ftape-1.02 which has nothing to do with ftape-0.9.x, I'll start the numbering of the beta versions with 1.12 ! There are no new features with this release, only bug fixes. All features released in the past, including multi-volume backups and end-of-file marks, should be working as advertised. Note: Some tape drives (Iomega) falsely set the new-cartridge- loaded status every time the driver is opened. These drives won't work together with file-marks as every file ends up at the beginning of the tape. Summary of bugs fixed: - 'unforseen write error' caused by a data underrun causes retry now (I hope, this is to be tested in the field). - aborting the driver should work better now. - fast seek problems solved for most drives. - some 'I/O errors' due to missing recovery fixed. Future plans: - Fix more bugs when new ones get reported. - Add code for 1 Mbps controllers as soon as I find a really cheap one. Have fun, Bas. === Release notes for ftape-0.9.12 === - Improved signal handling, fdc_interrupt_wait now restores original mask. - Improved fast seek, not relying on tape speed anymore because these differ from drive to drive. It will not work with some older drives that will need a timed function (not implemented yet). === Release notes for ftape-0.9.11 === - Changed some trace levels from 4 to 5. - Added write underrun error handler (untested !!!). - Fixed failing read_id because of missing queue pulses. - Added recovery if wait_segment fails in read_segment. - Better end of tape handling in read_id, smart_stop & abort_operation. - Changed some messages in fdc_isr. === Release notes for ftape-0.9.10a === - Fixed Makefile to prevent some versions of make to loop. - Fixed failing write retry bug (typo in write_segment). - Fixed Wangtek 3080F wakeup method (colorado instead of none). - Implemented trial version of fast seek. - Fixed bad timing of ftape_sleep(), expecting msec's instead of whole seconds now ! Fixed all calls. - Integrated faster ecc code. === Release notes for ftape-0.9.10 === So here's another release of the QIC-117 floppy tape driver. Two reasons prevented me of calling this code BETA yet: - Implementing new features, improving the error recovery and removing bugs I've changed a lot of code again (I hope this was the last time, can't promise anything though ;-). - I haven't got the slightest idea of how many people are using the driver successfully (I mostly get problem reports) :-( I reckon this version will trigger a new flood of problem reports. Beside knowing about these and working on them, I would like to know how many of you are successfully using the driver (so I can see if it's worth the effort ;-). So if you're happy with the driver, send me a postcard saying so and mention the driver version and tape drive you're using. Bas. ---- email: Bas Laarhoven smail: Bas Laarhoven, Mr. v. Boemellaan 39, 5237 KA 's-Hertogenbosch, Netherlands -------------------- New features: o Full support for file marks while remaining QIC-40/80 compatible. All (?) mtio options are implemented so it's possible to store multiple (e.g. tar) files on one tape and read them back at random. o The tape is stopped when the dataflow stagnates. This allows creation of compressed tapes with much less overhead (and tape wear). Other changes: o The tracing level is no longer set with `mt fsf' but with `mt fsr' now! o If using the gcc-2.4.5 compiler, due to a bug in the compiler you'll need to set a flag in the Makefile! o Distributed the code over a new set of files. o Rewritten Makefile using dependency file (need `make dep')! o Improved trace readability although less efficient now. o Tape unreferenced status doesn't abort ftape_open anymore. This allows formatting and other non-read/write commands on a unformatted tape. o Prepared code for single sector read/write. o Rewritten interrupt service routine from scratch. o Extended report_vendor_id code to recognize Iomega drives. o Fixed bug in fdc-io.c incorrectly handling overrun errors. o Increased retry count for writes. Still to do: o Integrate faster ecc code. o Integrate high-speed-controller code. o Support for 1 Mbps floppydisk controllers. o Handle my mail backlog ;-) --------------------