linux-mips-fnet
[Top] [All Lists]

Re: problem of pthread

Subject: Re: problem of pthread
From: Masami Komiya <mkomiya@crossnet.co.jp>
Date: Mon, 19 Jun 2000 18:10:15 +0900
Cc: linux-mips@fnet.fr
References: <394DCF1B.DB3ED8A3@crossnet.co.jp> <394DDA87.E57F6DDC@niisi.msk.ru>
The processor is NEC VR5000 and endian type is little. 
The result using pre-compiled glibc for DEC stations is the same. 

Masami Komiiya

"Gleb O. Raiko" wrote:
> 
> Masami Komiya wrote:
> >
> > I am try to port some applications to our embedded system linux-mips
> > version 2.2.10 is running.
> > We use glibc-2.0.6-5 for applications. Glibc is downloaded from
> > SGI's site and compiled by myself.
> > Many applications are running with no problems. But the application
> > that uses pthread cannot make a success the execution of
> > pthread_create().
> > Though function returns the value EAGAIN, the system resources is
> > enough.
> > Does somebody suggestion ?
> 
> Please, report the processor. I use latest Ralf's 2.0.6-51m w/o problems
> in ptherad_create on R3k. However, the IDT chip, for example, has a bug
> in logic and raises cpu unesable instead of illegal instruction. This
> prevents emulation of ll/sc and, thus, handling of atomic stuff for r3k
> processors. Certainly, there is a problem w/ IDT R3k chips and it should
> be handled by the kernel.
> 
> Also, it's usually bad idea to recompile glibc yourself until you know
> exactly all caveats.
> 
> Regards,
> Gleb.
> 
> BTW, Does anybody know where I may get an errata list for IDT chips ?
> 
> Gleb.

<Prev in Thread] Current Thread [Next in Thread>