fb12ba70e2
Only #include other stuff #ifdef KERNEL.
1994-04-08 04:24:16 +00:00
3f5e4f423f
Implement dynamic IRQ configuration and IRQ sharing. Inline spl*() calls.
...
Reorganize and clean up the relevant code.
1994-04-07 06:48:19 +00:00
beb5680d81
This is #included multiple times.
1994-04-03 22:36:41 +00:00
0234c30a00
Clean up #includes.
1994-04-03 22:28:23 +00:00
3a8dd5d6b1
Clean up #includes.
1994-04-03 18:58:55 +00:00
3ce26e4f42
Make sure we get psl.h.
1994-04-03 18:54:21 +00:00
c075fee1b2
Clean this up a little.
1994-04-03 18:53:53 +00:00
91906bc019
Inline assembly [hn]to[nh][sl] functions used xchgb %h1,%b1 instead of the
...
faster (on the i486 & i586) rorw $8, %w1. The inline assembly for GCC 1.X
already used rorw. Using rorw is one byte longer, but we wouldn't be
inlining at all if we weren't optimizing for speed.
1994-03-16 17:20:24 +00:00
6c218e563f
add basic integral types (a la sparc port) that new nvi wants.
1994-03-14 02:10:40 +00:00
427a14083c
Added RCSID macro, for RCS ID's (what else?).
1994-03-12 01:23:21 +00:00
3a57542a30
Back out previous change for now.
1994-03-09 07:58:39 +00:00
a692d2185a
Inline spl*() calls.
1994-03-08 13:24:46 +00:00
917559fb12
Add `com' softintr.
1994-03-08 06:57:28 +00:00
6d35bd3f7f
Move some machine dependant disklabel stuff in here.
1994-02-22 17:16:41 +00:00
efa03b9a6c
fix from Thomas Eberhardt <thomas@mathematik.uni-Bremen.de>:
...
byte swap macros used a variable "X" internally, which would get
screwed up if user programs #defined "X". change uses of X to __x.
1994-02-12 07:23:49 +00:00
5235180083
Fix spelling error in Copyright notice
1994-01-28 23:43:26 +00:00
5444802ec5
getting and setting FP regs aren't appropriate yet
1994-01-28 07:06:32 +00:00
24e4cb4bad
move varargs stuff to arch/<machine>/include.
1994-01-27 00:18:18 +00:00
2a51c00a70
FPU control word of 0x137f doesn't work well in "real" world applications
...
so I've changed it to 0x127f. 64 bit precision causes errors in paranoia
and our math library, changing to 53 bit precision solves these problems.
I've been told that long double support in Gcc 2.5 requires 64 bit mode,
so a comprehensive FP strategy (compiler, library, FPU, emulator) will
have to be addressed in the future.
1994-01-25 19:58:30 +00:00
c4b7945dca
My previous checkin didn't really change the FPU Exception Mask -- it was
...
trapped inside an #ifdef -- this commit changes it as I had originally
intended.
1994-01-17 18:10:01 +00:00
900084ce24
Clean up deleted files.
1994-01-16 02:00:01 +00:00
ea52295d11
SHMMAXPGS now pegged at 1024, where it belongs
1994-01-16 01:36:07 +00:00
c8b8051b89
Gratuitous formatting changes.
1994-01-11 13:30:40 +00:00
cb875046b8
LBL "/dev/audio" include file.
...
probably will need to be hacked if "the right things" are done to bsd_audio.c
1994-01-09 19:30:17 +00:00
8cdc8d997d
define PT_STEP, and clean up some
1994-01-09 16:53:03 +00:00
4eb1bbfb0b
Use PSL_MBO and PSL_MBZ in the more traditional way.
1994-01-09 14:58:14 +00:00
23c144e8b3
code reorg
1994-01-08 11:13:52 +00:00
ceb7f9888d
Change default NetBSD NPX control word to 0x137f.
...
This masks all FP exceptions (like every other UNIX system), and makes
it possible for gcc 2.5's long double support to work.
1994-01-08 01:19:06 +00:00
6f81dd8136
Add .type directive to ENRTY().
...
Move PIC_ macros over here from SYS.h
1994-01-07 00:46:20 +00:00
d5bb16fdb2
Move kernel to 0xf8000000.
1994-01-05 16:02:36 +00:00
e254def3cf
Fix VM_{MIN,MAX}_KERNEL_ADDRESS.
1994-01-04 00:44:03 +00:00
c0e772492e
Inline all the [lr]cr[023]() `functions' in cpufunc.h.
1994-01-03 16:21:48 +00:00
7131e8517e
Garbage collection.
1994-01-03 16:20:47 +00:00
e75be43c5d
Make sure we get the correct `rep' prefix per the Intel manual (though really
...
GAS should figure this out itself).
1993-12-31 21:11:32 +00:00
89ea303b7c
fix __asm() clobber list for some instructions. fix from
...
Lennart Augustsson <augustss@cs.chalmers.se>
1993-12-30 05:16:43 +00:00
fd9de878ef
add NPTEPG back in. gdb wants it. Thanks to Rich Murphey for pointing it out.
1993-12-22 17:39:45 +00:00
d2ad7a00c2
should be 'struct reg' not 'struct regs' -- it's in reg.h, after all!
...
also, delete r_fs and r_gs; looking at the code and comments
describing/using this struct, they don't belong here
(though if support for them is ever put in here, they should be,
but for i486-en only... how do you do that 'reasonably,' though?
interface issues...)
1993-12-22 12:57:24 +00:00
7fbfc4bac0
add (possibly iffy) definitions of PSL_USERSET and PSL_USERCLR,
...
"standard" PSL masks.
1993-12-22 12:56:08 +00:00
dced52af6c
bump MAXSSIZ back down to 8M (from 256M, where charles put it).
...
execve() currently allocated MAXSSIZ of address space (AND PAGE TABLES)
for the stack. until that's changed, and the stack-handling code is
updated to match, MAXSSIZ==256M will make execve() significantly slower
than MAXSSIZ==8M.
1993-12-22 12:30:59 +00:00
6c32b87568
Remove two unused fields.
1993-12-21 09:47:45 +00:00
0cd3f0fdd2
Remove some things which should not be here.
1993-12-20 09:08:11 +00:00
de04e7e727
New, more complete version.
1993-12-20 09:07:22 +00:00
2250fff236
Cleanup and garbage collection; nothing significant. From magnum branch.
1993-12-20 05:25:02 +00:00
546d92cec9
Forgot to decrement these.
1993-12-19 07:54:52 +00:00
c857af00dc
Clean up trap handling and macroize interrupt entry and exit code. From
...
magnum branch.
1993-12-19 06:58:12 +00:00
8b7ff01b9e
Clean up deleted files.
1993-12-19 03:42:13 +00:00
90b25c90a1
Cleanup from magnum branch.
1993-12-19 03:41:30 +00:00
ea13c12adc
From magnum branch:
...
Remove Jolitz's netisr kluge. Make sure cpl == 0 really means base priority.
Other minor cleanup.
1993-12-17 00:10:06 +00:00
1cc1b73beb
Cleanup of startup and signal handling code from magnum branch.
1993-12-15 09:31:29 +00:00
fccc819b20
Fold in most of the magnum vm changes.
1993-12-14 05:31:17 +00:00
mycroft