From: Rich Felker Date: Thu, 10 Mar 2011 01:07:24 +0000 (-0500) Subject: fix race condition in raise - just mask signals X-Git-Tag: v0.7.0~26 X-Git-Url: https://git.librecmc.org/?a=commitdiff_plain;h=0bed7e0acfd34e3fb63ca0e4d99b7592571355a9;p=oweals%2Fmusl.git fix race condition in raise - just mask signals a signal handler could fork after the pid/tid were read, causing the wrong process to be signalled. i'm not sure if this is supposed to have UB or not, but raise is async-signal-safe, so it probably is allowed. the current solution is slightly expensive so this implementation is likely to be changed in the future. --- diff --git a/src/signal/raise.c b/src/signal/raise.c index f437d23f..cc2b19b7 100644 --- a/src/signal/raise.c +++ b/src/signal/raise.c @@ -2,17 +2,17 @@ #include #include "syscall.h" +int __sigprocmask(int, const sigset_t *, sigset_t *); + int raise(int sig) { int pid, tid, ret; - /* Getting the pid/tid pair is not atomic, and could give wrong - * result if a fork occurs in a signal handler between the two - * syscalls. Use the tgkill syscall's ESRCH semantics to detect - * this condition and retry. */ - do { - tid = syscall0(__NR_gettid); - pid = syscall0(__NR_getpid); - ret = syscall3(__NR_tgkill, pid, tid, sig); - } while (ret<0 && errno == ESRCH); + sigset_t set; + sigfillset(&set); + __sigprocmask(SIG_BLOCK, &set, &set); + tid = syscall0(__NR_gettid); + pid = syscall0(__NR_getpid); + ret = syscall3(__NR_tgkill, pid, tid, sig); + __sigprocmask(SIG_SETMASK, &set, 0); return ret; }