[uClibc]vfork test fail in ppc target

[shaowei dai]

Hi, Erik and all:


The following problem seems caused by vfork/exec. I noticed that there
is a similar discussion about this issue in
http://www.uclibc.org/lists/uclibc/2001-September/004187.html 
Does anybody know the status of this issue? 

Following Erik's recommendation, I did the same vfork test included in
my distribution. The output is as follows:

#./vfork
Hi.  I'm the child process...
#bin  etc  lib  proc  usr  vfork  
dev   home  linuxrc sbin   var

Then wait here. Only after i press the Enter button, it will return to
prompt #. 

The uClibc version i'm using is cvs version 0.9.12  20 June 2002.
Busybox version is 0.61.pre.


Could anybody using the same version do a test? Or is there any version
without this problem? 

Appreciate your help!

shaowei

> Message: 2
> From: shaowei dai <dshaowei at rfnetech.com>
> To: uclibc at uclibc.org
> Date: 26 Jun 2002 18:33:16 +0800
> Subject: [uClibc]popen fail to return in ppc environment
> 
> Hi, all:
> 
> I'm using the latest cvs version uclibc for powerpc environment. When i
> compile pcmcia_cs package against it, it can run up to modprobe without
> problem. The symptom is after loading the corresponding module, i find
> that /var/run/stab still tells me "Socket 0: empty". After tracing down,
> i found out that it is due to a function call in cardmgr,
> popen(cmd,"r"), never return. This call is responsible for loading
> driver modules. Because cardmgr blocks here, there is no write to the
> stab file. 
> 
> When i switch to hardhat platform to do the same test, it does return. 
> 
> What interesting is that although 'popen' doesn't return, the loading of
> modules success. As i can assure this by issuing 'lsmod' to check.
> 
> Doesn't anybody know what maybe the cause of this strange behavior of
> 'popen'? 
> 
> Any input will be appreciated!
> David
> 
> 
> 
> 
> --__--__--
> 
> _______________________________________________
> uClibc mailing list
> uClibc at uclibc.org
> http://uclibc.org/mailman/listinfo/uclibc
> 
> 
> End of uClibc Digest