Commit graph

193 commits

Author SHA1 Message Date
rsc
bea4b9f724 openbsd 2005-07-19 11:48:01 +00:00
rsc
9d654ebc8c openbsd 2005-07-19 11:43:12 +00:00
rsc
1041ea6f68 stolen from freebsd; maybe work on openbsd? 2005-07-19 11:43:01 +00:00
rsc
2b6040818e use full prototypes 2005-05-07 22:42:14 +00:00
rsc
bc7da02954 OS X Tiger changes 2005-05-07 22:41:59 +00:00
rsc
910ca53df0 try a different approach to .S files 2005-05-01 22:28:11 +00:00
rsc
cea10000c5 OpenBSD/power support. 2005-05-01 18:38:12 +00:00
rsc
8e4ae643c3 more netbsd support. 2005-03-28 19:37:33 +00:00
rsc
5ba33c04da NetBSD support. Thanks to Christoph Lohmann. 2005-03-28 15:58:14 +00:00
rsc
0ba4eb47d4 no need to setproc(0) 2005-03-18 19:05:50 +00:00
rsc
72f4d44f86 do not touch TSTP, TTIN, TTOU 2005-03-18 19:04:49 +00:00
rsc
8dd8a81f71 add threaddata 2005-03-18 18:56:17 +00:00
rsc
d7688cce37 add check for procscheduler returning 2005-02-16 17:15:56 +00:00
rsc
80b8842f3e hard-to-find locking bug 2005-02-15 18:08:28 +00:00
rsc
8ee6ad4d96 different fix for main proc bug 2005-02-14 18:58:56 +00:00
rsc
8f4871fd2e work around stupid linux bug 2005-02-13 18:35:19 +00:00
rsc
31f809b54d remove debugging 2005-02-08 19:13:34 +00:00
rsc
46cfcf550f stupid gcc 2005-02-03 15:41:39 +00:00
rsc
669a062b5a osprocid can be -1 2005-01-30 16:35:05 +00:00
rsc
1fdb4a5450 rename scheduler to have a more distinctive name, so stack traces can end there 2005-01-23 22:57:53 +00:00
rsc
a0e8d02d09 Apparently SuSE 9.1's X distribution uses even bigger
buffers than the standard ones.  64kB appears to be enough
for a stack in that case, but let's just go nuts and make
the stacks enormous, so that it takes a few more doublings
of X's stack needs before we run into this problem again.

The VM system should take care of not actually using most
of the memory anyway.
2005-01-23 16:02:04 +00:00
rsc
803292c631 proccreate race 2005-01-18 20:53:12 +00:00
rsc
515b6e5bff add threadgetname 2005-01-18 20:17:12 +00:00
rsc
668b32b76c explain nptl stuff 2005-01-18 18:14:14 +00:00
rsc
9a4f7bced2 explain nptl stuff 2005-01-18 18:13:24 +00:00
rsc
3d867865ef better exit handling 2005-01-18 18:13:16 +00:00
rsc
8650250482 64-bit nonsense 2005-01-18 05:57:35 +00:00
rsc
4753f0691a more exits 2005-01-17 21:30:15 +00:00
rsc
7a2c88509b allow config to override kernel version 2005-01-17 21:30:05 +00:00
rsc
0b561695d6 useful warning comment 2005-01-16 21:31:21 +00:00
rsc
a0d2173ce4 too big fd 2005-01-16 17:37:11 +00:00
rsc
c358f4aa5e license 2005-01-14 19:42:33 +00:00
rsc
2b3b1e8984 oops sunos 2005-01-14 18:11:21 +00:00
rsc
4a8edd8d41 add procwakeupandunlock 2005-01-14 17:56:21 +00:00
rsc
1ac1981659 where did this change go? 2005-01-14 17:31:01 +00:00
rsc
c8b6342d3c Many small edits. 2005-01-13 04:49:19 +00:00
rsc
37cc72d323 paranoia 2005-01-11 21:28:00 +00:00
rsc
e9dbe11dbf don't rfork(RFNOTEG) because then you lose
the ability to read from the console.  damn.

also, handle case where child exits before
fork returns in parent.  have to record that
sigchld was seen and then run the handler later.
2005-01-11 21:06:55 +00:00
rsc
b8f742db85 os x changes 2005-01-11 17:43:53 +00:00
rsc
515bd9ec76 do rfork(RFNOTEG) before exec 2005-01-10 00:28:14 +00:00
rsc
6af7b77611 oops 2005-01-07 22:59:51 +00:00
rsc
50f7ec339a initial darwin 2005-01-07 21:47:30 +00:00
rsc
1aa9c533e0 allow lock passing 2005-01-07 20:52:07 +00:00
rsc
c345061e8d sunos fixes 2005-01-07 07:47:37 +00:00
rsc
a0a331aad9 add _procwakeupandunlock
to help ease locking contention
on Linux 2.4.
2005-01-06 23:43:42 +00:00
rsc
815552b9ed go back to 1000 scheds 2005-01-06 23:13:56 +00:00
rsc
a620761c92 add some annoying lock loop prints -- let's see
if they ever fire.

also fix race in delaystack.  it was actually happening!
that just goes to show that while i can reason very well
about what can and cannot happen, i have no intuition
at all for how likely they are.
2005-01-06 23:08:55 +00:00
rsc
af89fc108d more debugging 2005-01-06 23:07:19 +00:00
rsc
c7acb53e03 add _threaddaemonize 2005-01-04 22:23:24 +00:00
rsc
15cd8255b5 more debugging 2005-01-04 22:22:18 +00:00
rsc
230ead2fbe better handling 2005-01-04 22:22:09 +00:00
rsc
f6127ed209 more tests 2005-01-04 22:15:53 +00:00
wkj
0debe1c681 Use FreeBSD-approved method for determining FreeBSD version. 2004-12-29 01:29:42 +00:00
rsc
1d2533d010 restore old plan 9 property that when the
last thread exits the main proc, the remaining
program ends up in the background and the
program appears to have exited.
2004-12-28 22:36:24 +00:00
rsc
d54ead7fb9 more freebsd work 2004-12-28 04:20:39 +00:00
rsc
31cf0ab11f attempt at FreeBSD pre-5 support 2004-12-28 03:50:11 +00:00
rsc
93eb807ac2 remove debugging 2004-12-28 03:42:11 +00:00
rsc
ddb664da1b add threaddaemonize 2004-12-28 01:45:39 +00:00
rsc
96d6b19dcd new test program 2004-12-28 01:36:53 +00:00
rsc
2c87dda8f8 add threaddaemonize(), more bug fixes 2004-12-28 01:35:38 +00:00
rsc
8cbd854a8a better exec handling 2004-12-27 19:11:33 +00:00
rsc
955a2ca78d starting to work on Linux 2004-12-27 18:21:58 +00:00
rsc
e127e40bb1 shuffle to allow use of execchan in non-pthreads impls 2004-12-27 17:19:44 +00:00
rsc
8cd0011323 figure out which files to put in the library 2004-12-27 16:53:01 +00:00
rsc
f1e8706de8 linux pre-2.6 implementation 2004-12-27 16:52:46 +00:00
rsc
4dbefdd41c start linux pre-2.6 port 2004-12-27 16:52:26 +00:00
rsc
73722a8bbf confine pthreads to pthread.c 2004-12-27 03:49:03 +00:00
rsc
dcfa5aba0d remove debugging 2004-12-27 01:22:33 +00:00
rsc
e1dc7e4511 signal handling, add prototype for pthreadinit 2004-12-27 00:13:48 +00:00
rsc
6a5c5d4870 remove empty qlock.c 2004-12-26 02:08:53 +00:00
rsc
619085f0b4 more new libthread 2004-12-25 21:57:50 +00:00
rsc
1544f90960 New thread library 2004-12-25 21:56:33 +00:00
rsc
24f4e66b12 still more files 2004-11-08 16:04:01 +00:00
rsc
cf4f3eafc6 extra files 2004-11-08 16:03:42 +00:00
rsc
1956455367 more thread work 2004-11-08 16:03:20 +00:00
rsc
f6c9a57869 figure out which thread implementation to use 2004-10-22 18:46:18 +00:00
rsc
5093c3fa40 try to implement the daemonize hack. 2004-10-22 18:45:08 +00:00
rsc
793d93626b new test program 2004-10-22 17:15:47 +00:00
rsc
ba15d71b0c debugging, more pthreads crap 2004-10-22 17:15:30 +00:00
rsc
ad6638adc7 add this. 2004-09-23 21:40:21 +00:00
rsc
e7821682ea add pthread 2004-09-23 19:03:46 +00:00
rsc
7966faa931 Continue fighting pthreads.
Clean up thread library a bit too.
2004-09-23 03:01:36 +00:00
rsc
c6687d4591 Continue the pthreads torture. 2004-09-21 01:11:28 +00:00
rsc
e5c26b01f7 various forgotten files 2004-09-20 15:08:57 +00:00
rsc
bcf527a98e Continue switching library over to pthreads when possible.
Tprimes works on Linux 2.6.  You can only have 128 procs
though.
2004-09-17 03:34:32 +00:00
rsc
06bb4ed20d Rewrite to remove dependence on rendezvous and its bizarre
data structures.  Makes it easier to use pthreads too.
Still need to add code for non-pthreads systems.

Just a checkpoint to switch work to another machine.
2004-09-17 00:38:29 +00:00
rsc
984e353160 PowerPC Linux support from ericvh.
Mainly adding va_copy/va_end.
Also fix bug in sprint wrapping
around top of memory.
2004-08-22 15:39:56 +00:00
rsc
73bef9b76b More consistent use of _exits and exits
in place of _exit and exit.
2004-08-11 01:56:35 +00:00
wkj
a87638642c Compare function pointers against 0 rather than nil. 2004-07-09 01:54:06 +00:00
rsc
99dfa1dad2 va_arg fix 2004-06-17 03:33:33 +00:00
rsc
65de82a30f clean up children more often 2004-06-09 14:23:34 +00:00
rsc
9a4289cfdc add network dialing 2004-06-09 14:11:28 +00:00
rsc
1be1539ea8 make sure only one guy can call exit. 2004-05-23 01:00:12 +00:00
rsc
f1bfc54e12 More little bug fixes 2004-05-14 17:45:39 +00:00
rsc
1a8f27c350 various bug fixes 2004-05-14 15:14:21 +00:00
rsc
30387d7ba6 Noblock is nonsense. 2004-05-12 18:22:54 +00:00
rsc
c4097c2951 Fix small bugs. 2004-05-11 17:51:27 +00:00
rsc
2e965b3324 various bug fixes 2004-05-05 04:22:16 +00:00
rsc
be36ff6885 add -W to specify window size.
various other little fixes.
2004-04-29 17:13:24 +00:00
rsc
10495dd324 add stubs for other systems.
add threadspawn.
2004-04-25 21:31:32 +00:00