home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!sun-barr!olivea!charnel!rat!usc!zaphod.mps.ohio-state.edu!darwin.sura.net!ra!tantalus.nrl.navy.mil!eric
- From: eric@tantalus.nrl.navy.mil (Eric Youngdale)
- Newsgroups: comp.os.linux
- Subject: New scsi code.
- Message-ID: <Bxx7F9.MGD@ra.nrl.navy.mil>
- Date: 18 Nov 92 16:35:32 GMT
- Sender: usenet@ra.nrl.navy.mil
- Organization: Naval Research Laboratory
- Lines: 52
-
-
- The mail handler at niksula.hut.fi seems to be all foobar, so
- it appears as if I cannot post things to the SCSI channel and expect them to
- arrive. Thus I am going to do this the old fashion way...
-
- I have uploaded a new alpha version of the scsi code to tsx-11.mit.edu.
- The source code is on tsx-11.mit.edu in pub/linux/ALPHA/scsi in fscsi-06.tar.Z
- I do not have context diffs from fscsi-a05, but if people want them, I can
- generate them easily enough (the new future domain code would make the diffs
- rather long). There are a number of minor fixes here and there, which I
- described at the end of this message.
-
- My firm intention is that the new code be in place in the 0.99 kernel,
- and according to Linus, we have about 1-2 weeks (I want the code in 1.0, and it
- would be foolish to drop in the new code at the last minute). I STRONGLY
- encourage people to try and use the code, even if the performance enhancements
- are not in place for your specific scsi adapter. I do not intend to rush
- people working on the low-level code - if you do not have time to complete the
- job prior to 0.99, then we will just go with the slower code for the time
- being, and drop in the replacements once they are ready - after 1.0 if need be.
-
- -Eric
-
- Here is a short list of some of the changes that have been made since
- fscsi-a05:
-
- * The Makefile/Makefile.in has been cleaned up a little more - it
- should automatically regenerate the Makefile when you first drop the code into
- place. A "make clean" resets things so the next make forces a regeneration
- of the Makefile the next time around.
-
- * The future domain code has some scatter-gather enabled, and the
- performance has reportedly been improved quite significantly.
-
- * I added the framework for blacklisting certain devices, and skipping
- a lun scan for those devices.
-
- * Fixed the cdrom code so that it will in principle accept buffer sizes
- in any multiple of 512 bytes (in preparation for different size buffers in the
- buffer cache). The scsi disk code should already handle this correctly.
-
- * Commented out a couple of printks that had been there for debugging.
-
- * Added a few minor Ultrastor patches from Dave Gentzel (use
- request_dma to allocate dma channel).
-
- * Use the new linux/ioctl.h linux/mtio.h include files.
-
- * A bunch of other minor corrections here and there.
-
- --
- Eric Youngdale
-