[Xenomai-core] Xenomai v2.4.4

[Xenomai-core] Xenomai v2.4.4

[Philippe Gerum]

Here is the fourth maintenance release for the v2.4.x branch.  Short
log follows:

	[nucleus]

	* Prevent drifts between large calculated versus measured dates.
	* Fix potential deadlock on SMP when ptracing shadow threads.
	* Thaw timers when a ptraced shadow thread exits.
	* Fix race introduced by xnsynch ownership preemption.
	* Fix RPI update.
	* Fix priority propagation issue with mixed interfaces.
	* Fix compilation issues with uClibc.

	[native]

	* Output outstanding number of heap mappings from the registry.
	* Fix tiny memory leak for rt_task_self().
	* Ensure rt_task_suspend() returns -EINTR upon signal receipt.
	* Allow rt_task_sleep_until() to sleep indefinitely
          (TM_INFINITE).
	* Fix rt_task_send() in the non-blocking case (TM_NONBLOCK).
	* Documentation updates.

	[posix]

	* Fix pthread_cond_init(), pthread_mutex_init() for NULL
          attributes.
	* Restart silently upon EINTR error.
	* Fix signal handling when directed at shadow threads.
	* Fix recursion count of mutex after suspension on a condvar.

	[psos]

	* Fix calendar date to ticks conversion.
	* Return ERR_NOBUF when	required.
	* Fix system call number of tm_evevery service.

	[vxworks]

	* Make sem_give() a conforming syscall.
	* Make semaphore support callable from non-VxWorks tasks.
	* Rework mutex support.
	* Provide taskInfoGet() service.
	* Fix byte count in msgQReceive() wrapper.

	[powerpc]

	* Upgrade to kernel 2.6.25.
	* Add legacy ppc/ branch support to 2.6.20.

	[arm]

	* Upgrade to kernel 2.6.25.

	[blackfin]

	* Resync on 2008R1 (kernel 2.6.22). Add support for bf54x and
          bf52x.

	[x86]

	* Upgrade to kernel 2.6.25.

See the ChangeLog for details.

http://download.gna.org/xenomai/stable/xenomai-2.4.4.tar.bz2

-- 
Philippe.

Re: [Xenomai-core] Xenomai v2.4.4

[Gilles Chanteperdrix]

Philippe Gerum wrote:
 > 
 > Here is the fourth maintenance release for the v2.4.x branch.  Short
 > log follows:
 > 
 > 	[nucleus]
 > 
 > 	* Prevent drifts between large calculated versus measured dates.
 > 	* Fix potential deadlock on SMP when ptracing shadow threads.
 > 	* Thaw timers when a ptraced shadow thread exits.
 > 	* Fix race introduced by xnsynch ownership preemption.
 > 	* Fix RPI update.
 > 	* Fix priority propagation issue with mixed interfaces.
 > 	* Fix compilation issues with uClibc.
 > 
 > 	[native]
 > 
 > 	* Output outstanding number of heap mappings from the registry.
 > 	* Fix tiny memory leak for rt_task_self().
 > 	* Ensure rt_task_suspend() returns -EINTR upon signal receipt.
 > 	* Allow rt_task_sleep_until() to sleep indefinitely
 >           (TM_INFINITE).
 > 	* Fix rt_task_send() in the non-blocking case (TM_NONBLOCK).
 > 	* Documentation updates.
 > 
 > 	[posix]
 > 
 > 	* Fix pthread_cond_init(), pthread_mutex_init() for NULL
 >           attributes.
 > 	* Restart silently upon EINTR error.
 > 	* Fix signal handling when directed at shadow threads.
 > 	* Fix recursion count of mutex after suspension on a condvar.
 > 
 > 	[psos]
 > 
 > 	* Fix calendar date to ticks conversion.
 > 	* Return ERR_NOBUF when	required.
 > 	* Fix system call number of tm_evevery service.
 > 
 > 	[vxworks]
 > 
 > 	* Make sem_give() a conforming syscall.
 > 	* Make semaphore support callable from non-VxWorks tasks.
 > 	* Rework mutex support.
 > 	* Provide taskInfoGet() service.
 > 	* Fix byte count in msgQReceive() wrapper.
 > 
 > 	[powerpc]
 > 
 > 	* Upgrade to kernel 2.6.25.
 > 	* Add legacy ppc/ branch support to 2.6.20.
 > 
 > 	[arm]
 > 
 > 	* Upgrade to kernel 2.6.25.

A special note for Roland: the configure script now allows configuring
Xenomai on ARM for a "generic" machine, which should be suitable for
generating Debian packages for ARM.

-- 


					    Gilles.

Re: [Xenomai-core] Xenomai v2.4.4

[Roland Stigge]

High,

Gilles Chanteperdrix wrote:
> A special note for Roland: the configure script now allows configuring
> Xenomai on ARM for a "generic" machine, which should be suitable for
> generating Debian packages for ARM.

Thanks! :-)

By the way: lintian(1) detects several empty *.map files which weren't
empty in 2.4.3.

bye,
  Roland

Re: [Xenomai-core] Xenomai v2.4.4

[Roland Stigge]

Roland Stigge wrote:
> By the way: lintian(1) detects several empty *.map files which weren't
> empty in 2.4.3.

i.e.

arm_2hal_8c__incl.map
blackfin_2hal_8c__incl.map
blackfin_2nmi_8c__incl.map
generic_2hal_8c__incl.map
generic_2nmi_8c__incl.map
hal__32_8c__incl.map
hal__64_8c__incl.map
hal-common_8c__incl.map
ia64_2hal_8c__incl.map
nmi__32_8c__incl.map
powerpc_2hal_8c__incl.map
rtdm_8h__incl.map
smi_8c__incl.map

in doc/generated/html/api/

Re: [Xenomai-core] Xenomai v2.4.4

[Gilles Chanteperdrix]

Roland Stigge wrote:
 > Roland Stigge wrote:
 > > By the way: lintian(1) detects several empty *.map files which weren't
 > > empty in 2.4.3.
 > 
 > i.e.
 > 
 > arm_2hal_8c__incl.map
 > blackfin_2hal_8c__incl.map
 > blackfin_2nmi_8c__incl.map
 > generic_2hal_8c__incl.map
 > generic_2nmi_8c__incl.map
 > hal__32_8c__incl.map
 > hal__64_8c__incl.map
 > hal-common_8c__incl.map
 > ia64_2hal_8c__incl.map
 > nmi__32_8c__incl.map
 > powerpc_2hal_8c__incl.map
 > rtdm_8h__incl.map
 > smi_8c__incl.map
 > 
 > in doc/generated/html/api/

It looks like these map files contain no href in the 2.4.3 release, so,
there is nothing "clickable" and I think we do not loose anything. I
wonder why Doxygen needs these empty files though.


-- 


					    Gilles.