I still think there should be support for mixed ABI calls.
See also thread http://lists.debian.org/debian-hppa/2008/09/msg00018.html
It sounds like a good feature, but my time is restricted so I
implemented the simplest interface possible that I could prove worked.
If you have ample free time I suggest you post a version of the patch
with your enhancements. I'd be more than willing to test your patches.
I have to admin I will not write the code, I only propose the idea.
1) put "__lt_compat" field in a place where originally is LT initializer.
--- sysdeps/unix/sysv/linux/hppa/nptl/bits/pthreadtypes.h
+++ sysdeps/unix/sysv/linux/hppa/nptl/bits/pthreadtypes.h.compat
@@ -70,6 +70,7 @@ pthread_mutex_t
/* KIND must stay at this position in the structure to maintain
binary compatibility. */
int __kind;
+ int __lt_compat;
unsigned int __nusers;
__extension__ union
{
@@ -95,10 +96,11 @@ pthread_cond_t
struct
{
int __lock;
- unsigned int __futex;
+ int __lt_compat;
__extension__ unsigned long long int __total_seq;
__extension__ unsigned long long int __wakeup_seq;
__extension__ unsigned long long int __woken_seq;
+ unsigned int __futex;
void *__mutex;
unsigned int __nwaiters;
unsigned int __broadcast_seq;
@@ -130,6 +132,9 @@ pthread_rwlock_t
struct
{
int __lock;
+ int __lt_compat;
+ int __hppa_pad1;
+ int __hppa_pad2;
unsigned int __nr_readers;
unsigned int __readers_wakeup;
unsigned int __writer_wakeup;
2) create hppa specific pthread.h with just altered initializers (more zeroes)
3) At the begin of each affected function
if (v->__lt_compat)
{
lock_global_pt_lock();
if (v->__lt_compat)
{
zero_the_same_fields_as_NPTL_but_lt_compat_field();
barrier();
v->__lt_compat = 0;
};
unlock_global_pt_lock();
};
This just atomically converts linuxthreads initialized variable into NPTL
initialized variable.
For the new NPTL layout the zeroing code will not be taken,
therefore almost no overhead.
There might be a congestion on global_pt_lock, but I do not believe,
there are many static initilized mutexes in any real program.
There will be no need for allocation any new structures.
There are tons of statically initialized mutexs, gnome and gdk are
full of them as an example.
Hopefully not many will try to lock global_pt_lock in the same time.
Also rebuild against NPTL enabled libc-dev lowers the number
of possible linuxthreads initialized variables.
Petr
