svn commit: r263318 - in head: sys/compat/freebsd32 sys/kern sys/sys usr.bin/truss

[Attilio Rao]

On Wed, Mar 19, 2014 at 7:10 PM, Jilles Tjoelker <jilles at stack.nl> wrote:
> On Tue, Mar 18, 2014 at 09:32:03PM +0000, Attilio Rao wrote:
>> Author: attilio
>> Date: Tue Mar 18 21:32:03 2014
>> New Revision: 263318
>> URL: http://svnweb.freebsd.org/changeset/base/263318
>
>> Log:
>>   Remove dead code from umtx support:
>>   - Retire long time unused (basically always unused) sys__umtx_lock()
>>     and sys__umtx_unlock() syscalls
>>   - struct umtx and their supporting definitions
>>   - UMUTEX_ERROR_CHECK flag
>>   - Retire UMTX_OP_LOCK/UMTX_OP_UNLOCK from _umtx_op() syscall
>
>>   __FreeBSD_version is not bumped yet because it is expected that further
>>   breakages to the umtx interface will follow up in the next days.
>>   However there will be a final bump when necessary.
>
>>   Sponsored by:       EMC / Isilon storage division
>>   Reviewed by:        jhb
>
>> [snip]
>> Modified: head/sys/sys/umtx.h
>> ==============================================================================
>> --- head/sys/sys/umtx.h       Tue Mar 18 20:14:13 2014        (r263317)
>> +++ head/sys/sys/umtx.h       Tue Mar 18 21:32:03 2014        (r263318)
>> [snip]
>> -static __inline int
>> -umtx_wait(u_long *p, long val, const struct timespec *timeout)
>> -{
>> -     if (_umtx_op(p, UMTX_OP_WAIT, val, 0,
>> -         __DECONST(void *, timeout)) == -1)
>> -             return (errno);
>> -     return (0);
>> -}
>> -
>> -/* Wake threads waiting on a user address. */
>> -static __inline int
>> -umtx_wake(u_long *p, int nr_wakeup)
>> -{
>> -     if (_umtx_op(p, UMTX_OP_WAKE, nr_wakeup, 0, 0) == -1)
>> -             return (errno);
>> -     return (0);
>> -}
>> -
>
> These two do not use struct umtx, are not obsolete and are used by some
> ports, such as lang/sbcl with certain options. They are similar to the
> Linux futex system call. The relatively recent libxshmfence uses the
> underlying umtx_op directly, mainly because there is no static inline
> function for UMTX_OP_WAIT_UINT.

They can just use _umtx_op() straight off.
I already mentioned in the commit message that I will bump
__FreeBSD_version appropriately when more API breakage will be
introduced and we will have a relatively "stable" API situation of
umtx. At that point ports maintainer can fix the ports.

Attilio


-- 
Peace can only be achieved by understanding - A. Einstein