bd890814cf
Removed cpu_set_init_frame() use.
...
Increased default limits of memory use, ridiculously small before.
1995-05-05 10:53:53 +00:00
f58eb36f82
Removed kernel stack relocation and double-mapping.
...
Speeded up fork() significantly.
1995-05-05 10:47:35 +00:00
214234c937
Fix MAXine interrupt mask routine
1995-05-05 06:48:14 +00:00
331e2bd82b
Add some code/defines to differentiate Asante and Interlan/Gator cards
...
correctly with the new nubus code.
1995-05-05 05:40:13 +00:00
8703076975
define BROKEN_SWAP and/or cpu_swapout as appropriate.
1995-05-05 03:41:51 +00:00
3ff601e7a1
Don't conditionalize utility routines based on DEBUG flag
1995-05-05 02:45:31 +00:00
03c9dc7a48
Fix long-standing bug in NetBSD/pmax interrupt initalization on
...
Turbochannel machines with an IOASIC.
After an interrupt is taken, the IOASIC interrupt enable mask is
and'ed with the kernel's interrupt-mask variable. This masks
off any interrupts that were enabled after the hardware interrupt-enable
mask was set. Due to this bug all iynterrupts must be enabled before
the first interrupt is taken. (Interrupts enabled later aren't on in the
ioasic intr-mask register, so they aren't on in (sirm & kernelmask),
which is what used to get written back to the ioasic interrupt-mask
register. Fixed, and tested on on 3MAX+ but not xine/3min.
If this patch perchance breaks on Xines or 3mins, the old code can be
re-enabled by changing the #if 0 in machdep.c to #if 1.
1995-05-04 19:48:41 +00:00
7d244351cf
Add `evcnt' field.
1995-05-04 19:43:27 +00:00
089ab8e266
define audio `evcnt' field.
1995-05-04 19:42:36 +00:00
9e61204b61
Tally interrupts.
1995-05-04 19:41:35 +00:00
2e0e8fc49d
use the machine-independent definitions for the mc146818.
...
wrap the various uses of the chip (including the uses of the NVRAM on it)
around the new defintions, structures, and functions specified in the header.
1995-05-04 19:39:08 +00:00
7153343684
two cases to quiet the compiler. perhaps should be fixed differently.
1995-05-04 19:35:20 +00:00
40b72cc7a2
Wrap NGDT and NLDT in parens.
1995-05-04 05:37:39 +00:00
0175317fc3
If we decide to use IRQ 13, explicitly turn off CR0_NE to be sure.
1995-05-04 00:30:56 +00:00
94f4346537
Update a comment to match that last change.
1995-05-04 00:20:40 +00:00
f0a42753e4
npxinit() should not be called during autoconfig now. Watch for this
...
condition.
1995-05-04 00:16:40 +00:00
e03a827a0a
Move some comments from before npxsave1() to before npxsave().
1995-05-04 00:00:25 +00:00
bfef4c14c5
Slight rearrangement, to make the TS handling easier to follow.
1995-05-03 23:09:37 +00:00
99224602b4
Only dump the FPU state if the process actually used the FPU.
1995-05-03 23:05:46 +00:00
d23aebf3bb
Fix pccnpollc().
1995-05-03 20:46:32 +00:00
a412d4803c
mtpr support for 11/780 added. DELAY added. mcount() corrected.
...
trapframe fixed to support all registers.
1995-05-03 19:53:40 +00:00
35d394d7e4
Changed conf.c to new style. Took away double saving of regs in traps.
...
Changed boot() to use vfs_shutdown(). Updated setregs.
1995-05-03 19:20:09 +00:00
6db4e1b58a
By default, don't frob the ICU masks here. The software masking should
...
be sufficient.
1995-05-03 14:31:35 +00:00
c6a24163a3
Use Alpha cdefs.h
1995-05-03 06:04:54 +00:00
f57d609a52
Use PCB_USEDFPU, and a small performance tweak.
1995-05-03 00:17:16 +00:00
66927585fc
make an __indr_references macro, that expands to nothing...
1995-05-03 00:13:55 +00:00
e772f7c5ef
Set PCB_USEDFPU.
1995-05-03 00:09:13 +00:00
029176e96b
Add PCB_USEDFPU.
1995-05-03 00:08:19 +00:00
64b99b333d
The pmax 3100 interrupt-init code is broken: cpu_startup() enables
...
interrupts before it calls configure(). On 3100s, this can result in
the ethernet interface interrupting before leprobe() is ever called.
Be a bit more defensive in leintr(), where it was dereferencing a null
pointer. This reputedly fixes the oft-reported problems of 3100s not
ooting unless they're on an idle net or had a hard reset done before boot.
(Reworking the config code completely would be nice too.)
1995-05-02 20:00:09 +00:00
be2c7f3d13
If we panic inside trap(), Do a stack traceback before printing the trap log.
...
Also change the stack-traceback code to avoid having multiple returns
(and thus multiple stack pops) because with gcc -O2 that breaks the
heuristic that a "jr ra" preceding the PC precedes code to push the
current stack frame. Which breaks stacktrace() before it even
traces past itself :-(. Use a goto instead.
1995-05-02 19:51:52 +00:00
0c8be3a288
Make sure kbdopen() ships the keyboard reset cmd.
1995-05-02 07:56:42 +00:00
176d4dd6e5
Added -T option.
1995-05-02 05:54:28 +00:00
73fead01fc
Fix typos.
1995-05-01 19:45:42 +00:00
edab8dc1d9
Oops, took out a bit too much
1995-05-01 17:35:45 +00:00
7bada980a5
Oops. Undo some console code checkins that didn't belong.
1995-05-01 17:22:20 +00:00
88f28c3996
Fix typo.
1995-05-01 14:19:14 +00:00
8792c8e284
Save and restore %fs and %gs across signals.
1995-05-01 14:15:07 +00:00
4eea138a48
Add sc_gs and sc_fs. Move sc_onstack and sc_mask to end.
1995-05-01 14:14:11 +00:00
c78dd5a369
New location & definitions for audio driver.
1995-05-01 14:13:45 +00:00
bbfcbb8483
Add VM86 segment registers to trapframe.
1995-05-01 14:13:26 +00:00
cef0c43dee
Stylistic changes, and we don't need KERNEL any more.
1995-05-01 13:16:16 +00:00
c55ef2fd64
Always reload the LDT iff we reallocated its storage.
1995-05-01 13:13:20 +00:00
953e339aa0
Stylistic changes tp cpu_exit().
1995-05-01 13:09:43 +00:00
f745c10020
Eliminate _gsel_tss and _default_ldt.
1995-05-01 13:02:24 +00:00
e5d7ebde66
Update to match include files.
1995-05-01 11:07:40 +00:00
0ea347bf51
Update for changes in segments.h.
1995-05-01 10:43:01 +00:00
bce8e94dd2
Correct comments in gdt_segs[] and ldt_segs[] initializers.
1995-05-01 09:08:27 +00:00
5d17674312
Remove IdlePTD.
1995-05-01 08:53:13 +00:00
87b25c7f5c
Update to match include files. Eliminate _ucodesel, _udatasel, and IdlePTD.
1995-05-01 08:25:30 +00:00
ae11bd216c
Remove historical garbage.
1995-05-01 08:14:25 +00:00
ragge