From fc08075dc61787269743a1abb8ff8221868c3210 Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Sun, 7 May 2000 22:04:36 +0000 Subject: Update. (ucs4_internal_loop): Likewise. (ucs4_internal_loop_unaligned): Likewise. (internal_ucs4le_loop): Likewise. (internal_ucs4le_loop_unaligned): Likewise. (ucs4le_internal_loop): Likewise. (ucs4le_internal_loop_unaligned): Likewise. --- linuxthreads/ChangeLog | 21 +++++++++++++++++++++ 1 file changed, 21 insertions(+) (limited to 'linuxthreads/ChangeLog') diff --git a/linuxthreads/ChangeLog b/linuxthreads/ChangeLog index 5c1c1e1298..8a0310dadb 100644 --- a/linuxthreads/ChangeLog +++ b/linuxthreads/ChangeLog @@ -1,3 +1,24 @@ +2000-05-06 Kaz Kylheku + + * internals.h: New thread manager request type, REQ_KICK. + * join.c (pthread_exit): main thread now calls exit() instead + of _exit() in order to proper process cleanup. + * manager.c (__pthread_manager): Do not terminate manager + after unblocking main thread; wait for main thread's + REQ_PROCESS_EXIT request instead. + Also, added REQ_KICK case to handle new request; this just does + nothing. + * manager.c (pthread_exited): Do not terminate manager after + unblocking main thread. + * manager.c (__pthread_manager_sighandler): If the main thread + is waiting for all other threads to die, send a REQ_KICK into + the thread manager request pipe to get it to clean out the threads + and unblock the main thread as soon as possible. This fixes + the 2000 millisecond hang on shutdown bug. + * Examples/ex7.c: New file, tests shutdown behavior when all threads + including the main one call pthread_exit(), or implicitly do so. + * Makefile (tests): Add ex7. + 2000-05-05 Andreas Jaeger * sysdeps/unix/sysv/linux/i386/getcpuclockid.c -- cgit v1.2.3-70-g09d2