Date: Tue, 24 Sep 2002 13:06:09 -0400 (EDT) From: Daniel Eischen <eischen@pcnet1.pcnet.com> To: Terry Lambert <tlambert2@mindspring.com> Cc: Andriy Gapon <avg@icyb.net.ua>, julian@elischer.org, hackers@FreeBSD.org Subject: Re: libc_r in stable Message-ID: <Pine.GSO.4.10.10209241304490.29920-100000@pcnet1.pcnet.com> In-Reply-To: <3D90999D.14E02BF0@mindspring.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 24 Sep 2002, Terry Lambert wrote: > Andriy Gapon wrote: > > I have noticed that it is not uncommon for a multithreaded program that > > invokes a lot of system calls to get EINTR from a syscall even if no > > non-default signal handling is installed and no 'real' signals are > > received. Looks like some syscalls get interrupted by the scheduling > > signal. In my understanding, this should be something hidden from a user. > > Is there any way to restart syscalls interrupted by the scheduling signal > > without returning a libc_r call ? > > Set SA_RESTART in the sa_flags of the sigaction for the scheduling > signal. Use "man sigaction" for details. Yes, for application installed signal handlers. Try this for library installed signal handlers (patch to -current). http://people.freebsd.org/~deischen/libc_r.diffs -- Dan Eischen To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.GSO.4.10.10209241304490.29920-100000>