[prev in list] [next in list] [prev in thread] [next in thread] 

List:       musl
Subject:    Re: [musl] [PATCH 3/9] use weak symbols for the POSIX functions that will be used by C threads
From:       Rich Felker <dalias () libc ! org>
Date:       2014-09-06 18:52:46
Message-ID: 20140906185246.GW23797 () brightrain ! aerifal ! cx
[Download RAW message or body]

On Mon, Sep 01, 2014 at 12:46:23AM +0200, Jens Gustedt wrote:
> The intent of this is to avoid name space pollution of the C threads
> implementation.

I found a few bugs in this which I'm fixing and about to commit if all
goes well:

> diff --git a/src/mman/mprotect.c b/src/mman/mprotect.c
> index f488486..535787b 100644
> --- a/src/mman/mprotect.c
> +++ b/src/mman/mprotect.c
> @@ -2,10 +2,12 @@
> #include "libc.h"
> #include "syscall.h"
> 
> -int mprotect(void *addr, size_t len, int prot)
> +int __mprotect(void *addr, size_t len, int prot)
> {
> 	size_t start, end;
> 	start = (size_t)addr & -PAGE_SIZE;
> 	end = (size_t)((char *)addr + len + PAGE_SIZE-1) & -PAGE_SIZE;
> 	return syscall(SYS_mprotect, start, end-start, prot);
> }
> +
> +weak_alias(__mprotect, mprotect);

I tested and inlining the syscall does not make pthread_create.o
smaller. It would eliminate a dependency on one extra file, so I might
make the change later, but for now I'm just doing it your way in the
interest of getting things committed without major edits to the
patches.

> diff --git a/src/thread/__timedwait.c b/src/thread/__timedwait.c
> index d6f1233..3d62272 100644
> --- a/src/thread/__timedwait.c
> +++ b/src/thread/__timedwait.c
> @@ -4,6 +4,9 @@
> #include "futex.h"
> #include "syscall.h"
> 
> +int __pthread_setcancelstate(int new, int *old);
> +int __clock_gettime(clockid_t clk, struct timespec *ts);

I've removed arg names from prototypes like this just for consistency
with how it's done elsewhere.

> @@ -139,7 +143,7 @@ static void init_file_lock(FILE *f)
> 
> void *__copy_tls(unsigned char *);
> 
> -int pthread_create(pthread_t *restrict res, const pthread_attr_t *restrict attrp, \
> void *(*entry)(void *), void *restrict arg) +static int __pthread_create(pthread_t \
> *restrict res, const pthread_attr_t *restrict attrp, void *(*entry)(void *), void \
> *restrict arg)

This static is wrong if the goal is to go ahead and get the files in
the form where they're "ready to use" for C11. I saw it was reverted
in the later patch, but I'm just taking it out here.

> diff --git a/src/thread/pthread_getspecific.c b/src/thread/pthread_getspecific.c
> index b2a282c..bfc4294 100644
> --- a/src/thread/pthread_getspecific.c
> +++ b/src/thread/pthread_getspecific.c
> @@ -1,7 +1,10 @@
> #include "pthread_impl.h"
> 
> -void *pthread_getspecific(pthread_key_t k)
> +static void *__pthread_getspecific(pthread_key_t k)
> {
> 	struct pthread *self = __pthread_self();
> 	return self->tsd[k];
> }
> +
> +weak_alias(__pthread_getspecific, pthread_getspecific);
> +weak_alias(__pthread_getspecific, tss_get);

This actually goes ahead and implements one C11 function, which is not
horrible, but mildly inconsistent. I think I'll move it to the later
patch that's actually supposed to implement these functions.

> diff --git a/src/thread/pthread_mutex_trylock.c \
> b/src/thread/pthread_mutex_trylock.c index e851517..9be7930 100644
> --- a/src/thread/pthread_mutex_trylock.c
> +++ b/src/thread/pthread_mutex_trylock.c
> @@ -50,9 +50,11 @@ int __pthread_mutex_trylock_owner(pthread_mutex_t *m)
> 	return 0;
> }
> 
> -int pthread_mutex_trylock(pthread_mutex_t *m)
> +static int __pthread_mutex_trylock(pthread_mutex_t *m)
> {
> 	if ((m->_m_type&15) == PTHREAD_MUTEX_NORMAL)
> 		return a_cas(&m->_m_lock, 0, EBUSY) & EBUSY;
> 	return __pthread_mutex_trylock_owner(m);
> }
> +
> +weak_alias(__pthread_mutex_trylock, pthread_mutex_trylock);

This static looks like an actual bug that would break the patch series
as a whole; I can't see where it's made visible for mtx_trylock to use
later. Removing static here.

Rich


[prev in list] [next in list] [prev in thread] [next in thread]