Re: stdio thread safety

[Corinna Vinschen <vinschen at redhat dot com>]

On Jun  3 11:04, Carl Norum wrote:
> 
> Hi newlib list,
> 
> I'm working on an embedded ARM project, and I'm trying to figure out
> my C library options situation.  I've used newlib before, but this is
> the first time I'll have tried to use it in a multithreaded
> environment.  sys/lock.h appears to just make all of the locking
> primitives no-ops, though.  Is the thread-safety part of newlib still
> a work in progress?  Or am I missing something with how to hook in?
> It looks like the fprintf implementation has the calls to lock/unlock
> files in the right places, but those calls just don't do anything.
> The __malloc_lock and __malloc_unlock, in contrast, do appear to work
> and do the right thing - I implement those functions and they get
> called.  However, if I implement "__lock_init_recursive", for example,
> newlib doesn't end up making any calls to it.  I'm guessing that's a
> result of the library being built with that lock.h header mentioned
> above.

Just override the definitions in sys/lock.h by providing your own
sys/lock.h, along the lines of:

  typedef __your_lock_type _LOCK_T;
  typedef __your_lock_recursive_type _LOCK_T;

  #define __lock_init(__lock)  __your_lock_init(__lock)
  #define __lock_init_recursive(__lock) __your_lock_init_recursive(&__lock)
  #define __lock_close(__lock) __your_lock_fini(&__lock)
  [...]

> Right now, I've worked around it by using the linker to make wrap
> functions that lock access to stdio, but that's not very pretty.  I
> may have to do something like this anyway, though - the new
> nano-malloc looks good, but it's explicitly not thread safe (the
> MALLOC_LOCK and MALLOC_UNLOCK functions are defined to be nothing in
> the source).

Yes, that has been discussed and the original developer is hopefully
looking into adding thread safety soon.


Corinna

-- 
Corinna Vinschen
Cygwin Maintainer
Red Hat