Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 30 Oct 2000 12:01:59 -0500 (EST)
From:      Daniel Eischen <eischen@vigrid.com>
To:        Max Khon <fjoe@iclub.nsu.ru>
Cc:        Roman Shterenzon <roman@harmonic.co.il>, freebsd-stable@FreeBSD.ORG, "Daniel M. Eischen" <eischen@vigrid.com>, jkh@FreeBSD.ORG
Subject:   Re: Is it late already? (-pthread)
Message-ID:  <Pine.SUN.3.91.1001030111421.2759A-100000@pcnet1.pcnet.com>
In-Reply-To: <Pine.BSF.4.21.0010301942160.90486-100000@iclub.nsu.ru>

next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, 30 Oct 2000, Max Khon wrote:
> hi, there!
> 
> On Mon, 30 Oct 2000, Roman Shterenzon wrote:
> 
> > The pthreads in RELENG_4 are quite broken (I mean, very-very broken).
> > Daniel M. Eischen <eischen@vigrid.com> has produced some patches which are
> > in -current for some time now, and were tested by me in -stable.
> > They were found to work great for all things I've tested it on.
> > It works *better* than the original implementation _anyway_, which is,
> > like I said is broken.
> > So, perhaps it wasn't in -current for a long time, but leaving the current
> > implementation is like shipping broken pthreads apriori
> 
> what problems are you talking about?
> maybe this will explain problems we have with ACE+TAO on RELENG_4

Most of the problems have to do with signal handling.  The
threads library in -current (hopefully) fixes these problems.
I am hesitent to MFC to -stable without further testing
in -stable.  I have had two reports of problems in -current
that I cannot reproduce.

Another problem that I found was with languages (GNAT) that use
builtin setjmp/longjmp in conjunction with pthread_kill() to implement
asynchronous transfers of control.  I'm currently trying to fix
this but I won't be done before Nov 1.

Latest diffs to -stable are at:

  http://people.freebsd.org/~deischen/libc_r.diffs-stable

Feedback from those with big mysql databases is appreciated.

-- 
Dan Eischen



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.SUN.3.91.1001030111421.2759A-100000>