Re: [CVS] commit

Date view Thread view Subject view Author view

From: Alexandre Oliva (oliva@dcc.unicamp.br)
Date: Thu Feb 04 1999 - 08:24:27 EST


On Feb 4, 1999, Godmar Back <gback@cs.utah.edu> wrote:

> Hmmm, I was finally able to reproduce the library "not found" failure on a
> slower Linux machine with RedHat 5.1. Apparently, it's dying somewhere
> in libltdl.

Probably the context switch during malloc problem :-(

One simple (yet quite ugly) way to avoid this problem is to ensure
that we can only context switch by alarm when the interrupt handler is
(implicitly) called from within Java code. I.e., there could be a
global variable that would enable/disable preemption, and it would be
disabled whenever we entered native code, and re-enabled whenever we
left it (possibly switching context at that point). Of course,
blocking calls would have to allow context switching explicitly, but
this would be much easier to handle than handling each special case
manually.

> What would be nice to have is a verbose mode in libltdl, maybe
> activated by setting an env variable. Or at least an error string a la
> strerror() if lt_whatever fails.

There is lt_dlerror() already, and Kaffe uses it.

> Unfortunately, it never happens under gdb.

Signals are handled differently within gdb :-(

-- 
Alexandre Oliva  http://www.dcc.unicamp.br/~oliva  aoliva@{acm.org}
oliva@{dcc.unicamp.br,gnu.org,egcs.cygnus.com,samba.org}
Universidade Estadual de Campinas, SP, Brasil


Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Sat Sep 23 2000 - 19:57:57 EDT