[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

RE: Is it a bug of PTHREAD_MUTEX_ERRORCHECK_NP ?



I am really confused by __pthread_initialize_minimal() now.

# ld -M tst-mutex2 > a 
# grep "__pthread_initialize_minimal" a     // nothing ???
# ld -M tst-join1 > b
# grep "__pthread_initialize_minimal" b   
0x0000a920                __pthread_initialize_minimal
0x0000a920                __pthread_initialize_minimal_internal

I renamed __pthread_initialize_minimal() -->
__pthread_initialize_minimal2() in libc-start.c 
and do the same to /nptl/init.c. At the same time, I will
print out sth at the begining of __pthread_initialize_minimal2().
However, the tst-mutex2 will always segment fault without 
printing anything. [for the arm toolchain issue, I couldn't 
debug it with gdb now, so it is hard to get more details :( ] 
any hints? thanks. 

by the way, here both tst-mutex2 & tst-join1 are statically linked. 

boris

> -----Original Message-----
> From: Roland McGrath [mailto:roland redhat com]
> Sent: 2003?3?7? 14:43
> To: phil-list redhat com
> Subject: Re: Is it a bug of PTHREAD_MUTEX_ERRORCHECK_NP ?
> 
> 
> __pthread_initialize_minimal is always called.  If you are 
> using mutexes,
> you must link with -lpthread and so should be getting the libpthread
> definition rather than the libc definition.  This might be broken with
> static linking because the reference is from inside libc.  
> Get a link map
> (-M linker option, or -Wl,-M through gcc) to see where
> __pthread_initialize_minimal is coming from.
> 
> 
> 
> -- 
> Phil-list mailing list
> Phil-list redhat com
> https://listman.redhat.com/mailman/listinfo/phil-list
> 





[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]