260ddbefb7
Rearrange ed_init() slightly, to better correspond to the National manual.
...
Optimize ed_rint() slightly in the case of multiple packets. (Avoid an outb
and an inb for each packet).
1994-10-23 23:24:56 +00:00
a6426e2b27
oops; went overboard, and did the user_ldt calls the wrong way.
1994-10-23 21:57:12 +00:00
603fe7b5a6
Slight optimization.
1994-10-23 21:38:00 +00:00
0114159848
Fix the packet length correction for buggy chips so that it actually works,
...
and always use it (otherwise short packets could still be corrupted). Also
produce a diagnostic message if DIAGNOSTIC, since it *is* a hardware bug.
Add an ED_PAGE_SHIFT constant, and use it, to avoid GCC doing horrendous
things with divides and multiplies.
1994-10-23 21:22:13 +00:00
a9ef20740a
Make this compile with USER_LDT again.
1994-10-20 23:48:21 +00:00
be18e790ff
Get the label sector right if there is no NetBSD partition in the MBR.
1994-10-20 22:44:37 +00:00
3e30a415ae
Be careful not to stack multiple timeouts.
1994-10-20 19:22:01 +00:00
4d072c5435
Remove the local timeout mechanism, and just use explicit timeout()/untimeout()
...
calls.
1994-10-20 18:37:45 +00:00
c11a06cbc2
Always set d_secsize to 512 for now.
1994-10-20 17:03:09 +00:00
0862d07daf
Don't switch to single-sector mode just because we got a corrected error.
1994-10-20 16:36:21 +00:00
d4153fff4d
Rearrange wdcstart() in a more logical fashion, and eliminate many redundant
...
calculations. Also, speed up bad block searching a little.
1994-10-20 16:19:08 +00:00
cd66683989
openpart -> openmask
1994-10-20 14:08:07 +00:00
cae71ba91d
Various changes:
...
* Add a dkdevice; move sc_*openpart, sc_label, and sc_cpulabel into it.
* Turn sc_wlabel into a flag, and don't set it automatically in DIOCWLABEL.
* If there is no label, disallow all I/O except to the raw partition.
* Don't allow I/O to `unused' partitions.
* Beginnings of support for block sizes other than 512 bytes.
Other minor changes.
1994-10-20 13:44:46 +00:00
cb4e5ecaa5
Print a single `soft error' message when an operation was retried and
...
eventually completed successfully. Also fix fsbn numbers in diskerr() calls.
1994-10-20 04:46:17 +00:00
77e4fe47ef
update for new syscall args description mechanism
1994-10-20 04:43:21 +00:00
b2d8a64ebd
Make another printf() #ifdef FD_DEBUG.
1994-10-20 03:38:13 +00:00
872999d392
Make the verbose diagnostics #ifdef FD_DEBUG.
1994-10-20 00:52:23 +00:00
70e27da16b
Fix up some printfs and make them consistent.
1994-10-18 18:24:35 +00:00
8ae8b4ab29
Bug fixes from Rafal Boni:
...
* Set promiscuous mode flag before configuring card, so it actually works.
* Handle failed TDR commands correctly.
* Remove some apparently defunct and bogus code.
(This driver needs more work.)
1994-10-18 13:40:30 +00:00
1725b07a35
there are no flags between us (and compiling) any more.
1994-10-18 06:11:15 +00:00
805605ce56
kill le0. *sigh*
1994-10-17 21:49:51 +00:00
320359f8e2
change, requested by Jarle Greipsland
1994-10-17 00:08:29 +00:00
d00c05633a
Wait on the correct bit in the lower half of kbc_get8042cmd(); and some other
...
mostly stylistic changes. Based on suggestions from John Hood.
1994-10-16 15:46:13 +00:00
06b14c27fd
Substantially smaller and faster version, from me.
1994-10-16 15:37:41 +00:00
f501e4e128
make <stdarg.h> a symlink, and clean up ports' stdarg.h and varargs.h files.
1994-10-15 04:57:29 +00:00
49c34d3337
move arch dependence to archs. clean up a bit. deal with weird MAXPARTITONS.
1994-10-14 18:26:22 +00:00
6f5794ecba
Fix bugs in dealing with odd-length mbufs in ed_pio_write_mbufs(). From
...
John Hood.
1994-10-14 13:05:13 +00:00
7c01f7ce51
Changes from DG to deal with some old chips that corrupt the packet length.
1994-10-12 13:42:00 +00:00
4252ca6644
Correct multicast filter calculation.
1994-10-09 21:19:04 +00:00
ee5cde2719
Oops; slight optimization to the previous.
1994-10-09 18:10:55 +00:00
8c5f95b10b
Pretend that trap() and syscall() have no args.
1994-10-09 16:09:27 +00:00
88501e66d2
Fix post-decrement loop to not overshoot, so the last change works right.
1994-10-09 15:49:10 +00:00
867f51ab31
Only print the return address of an interrupt, trap, or syscall separately if
...
it was the last frame in the trace; otherwise the info is duplicated.
1994-10-09 15:40:51 +00:00
236f105ccd
Remove unneeded #includes.
1994-10-09 15:06:23 +00:00
c3fa5122d6
#include cpufunc.h from pmap.h.
1994-10-09 15:00:38 +00:00
2013b5f349
Combine INTERRUPT case with TRAP and SYSCALL.
1994-10-09 14:51:17 +00:00
f72ff16e5b
Oops. Just fake %ebp; nothing else.
1994-10-09 14:45:40 +00:00
c9cb256e1d
If DDB is configured, make a fake stack frame.
1994-10-09 14:43:00 +00:00
5d0226231f
Do something sensible with interrupt frames.
1994-10-09 14:37:54 +00:00
aef750bbec
Fix slight bogosity in trap frame handling, and combine with syscall case.
1994-10-09 13:48:40 +00:00
e19b0472c8
Make multiple-inclusion protection consistent.
1994-10-09 13:33:53 +00:00
c87fbde918
Fix typo.
1994-10-09 13:31:19 +00:00
701c0f0058
Eliminate i386_saved_state.
1994-10-09 13:27:07 +00:00
8295e4b8f2
Nuke dirty().
1994-10-09 13:23:21 +00:00
ebbc8e0476
Make the conversion macros more consistent.
1994-10-09 13:20:37 +00:00
18830ce0ea
Define pmap_update() using tlbflush().
1994-10-09 13:11:18 +00:00
e165978b36
Use pmap_update(), not tlbflush().
1994-10-09 13:09:11 +00:00
47db98b4d9
Eliminate register restrictions where possible.
1994-10-09 13:02:56 +00:00
87041c0044
Make reg, frame, and sigcontext more similar.
1994-10-09 12:57:11 +00:00
112dac786c
Make reg, frame, and sigcontext more similar.
1994-10-09 12:39:05 +00:00
mycroft