CVS log for dietlibc/libpthread/thread_internal.h

Up to [DVB] / dietlibc / libpthread

Request diff between arbitrary revisions

Keyword substitution: kv
Default branch: MAIN

  first stab at getting TLS to work in actual threads
  use fs: to find current thread on x86_64

fix pthread_cond_timedwait

  update syscalls
  lots of fixes from Markus FX Oberhumer:
    * better signal handling error detection
    * setjmp.S on i386
    * getenv on i386
    * sys/stat.h -malign-double fix
    * update fcntl.h for MIPS to 2.6.13 version
    * jmp_buf is 40*8 instead of 58*8 on ppc64.
    * This patch fixes the section of the TOC entries on ppc64
    * libstdc++ updates for gcc 4
    * remove cstddef and cwchar (no longer needed)
    * int -> ssize_t for read and write
    * sigjmp.c: __mask_was_saved was not set in all cases!
    * another round of ILP64 cleanups

  Lots of fixes and patches from Markus Oberhumer.
    libstdc++ compatibility
    ilp64 compatibility
    make diet libc compile with -Werror on 13 platforms
  ppc64 now compiles again

- some more stdio unlocked functions and cleanup with the cancel-points
- "new"-internal pthread_mutex_{un,}lock functions

fixing the pthread_join function
minor of the thread_specific data handling

add the semaphore support.

removed problematic longjmp/setjmp for timedwait

- checkin of the new libpthread
- add confstr

be more alpha and gcc 2.96 for SuSE 7.1 AXP friendly, add fdatasync.

found an other bug in the condition (signal) handling

major change:
 - fixed a misunderstanding bug / errors are returned not set to errno
 - new pthread_join struct (now it passes the glib check)
 - changed (and probably broke) the pthread_key* struct (but now it passes the glib test)

a ton of warnings fixed ;)

lots of errno cleanups,
lots of errno related thread-safeness cleanups in librpc (don't ask).

fixed a bug...
if the thread-life was shorter then one time-slice it had slipped the
creator...

typo, WANT_THREAD_SAVE -> WANT_THREAD_SAFE

add msg* library and pthread check for WANT_THREAD_SAFE in
dietfeatures.h

and now the thread specific key handling...
this is,was and will be a pain...

here the pthread conditional variables...

cncellation in a more less pain :)
fixed some interna...

malloc/free is now cancellaton save...
now the cleanup-stack is implemented to do the cleanup (ups)

hamm hope this'll do it.
removed the pthread_mutices from logging (syslog implementation)

changed some...
1. removed some pthread dependencies for dietlibc
2. added a pthread_sys_alloc to compensate

here are the first cancel-point-syscall wrapper.... (N.C.)

ok.
1. fixed a copy-paste bug
2. added atfork handling (pain...)
3. added cleanup stack (even more pain...)

some little changes I'd forgotten...
hopefully a deadlock preventer (theoretical deadlock, not yet encountered)

arg... after many heisen-bugs it should work now.
there is now a manager thread.

so now the pthread_attr access functions are complete. (and hopefully
functional as man-page said)

So, here now the modified create and some additional mutex stuff...
it should now be a base to start tests. (next step)

testandset -> __testandset
__testandset is in the dietlibc

ups... some non statics

ok here are the FIRST release and not yet complete...
(should not work until the architecture dependend parts are ready)