Just uploaded a snapshot of stable tree with a number of "local" fixes to
sparc-2.0.35.
You can fetch by anonymous ftp at:
ftp://vger.rutgers.edu/pub/linux/Sparc/kernel/v2.0/980805/
sparclinux-2.0-980805.tar.gzentire source tree
sparclinux-980719-980805.diff.gzpatch against 980719 snapshot
sparclinux-2.0.35-980805.diff.gzpatch against plain 2.0.35 source
vmlinux-980805.gz
modules-980805.tar.gz
System.map-980805.gz
vmlinuxSMP-980805.gz
modulesSMP-980805.tar.gz
System.mapSMP-980805.gz
What's new:
- complete turbosparc support, though not well optimized it works
correctly. Thanks to Stephen Zander for testing about a dozen kernels
and providing me with useful feedback. (Someone is working on porting
these to 2.1; He can speak up if he wishes)
- stupid fix to dummy audio device source; the module distributed last
time was correct, but the source was out of date. oops.
- patch to esp driver to make egcs-built esp.o not make sun4c unhappy.
thanks to Aron Griffis for testing this, and to Eddie Dost (who
suggested the final version of the patch) and David Miller (whose
insight clued me in to what was really happening, even if I still don't
entirely understand the mechanics of it). Also in 2.1.
- another patch to the esp driver, to disable sync to a device when a
certain error occurs. Also in 2.1.
- patch to scsi.c to enable single_lun based on BLIST_SINGLELUN a few
lines earlier, as in 2.1, so concurrent accesses to devices which are
not safe for this sort of thing do not crash, hang, or I/O Error.
What I hope to have ready for the next snapshot:
- fixed bpp driver, or at least know why it doesn't, won't or can't work
- fixed OSS ioctls in the sparc audio support, as a few are currently
broken.
Future work:
- audio ringbuffers instead of non-resizable, fixed number of buffers.
- dbri audio driver (LX, 10, 20)
- and then it's on to 2.1 development for me, I think...
-D
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]