On 02/24/2014 01:07 PM, Gilles Chanteperdrix wrote:
On 02/24/2014 09:56 AM, Jan Kiszka wrote:
On 2014-02-24 01:29, Gilles Chanteperdrix wrote:
Hi,
The following changes since commit ce5bfa8a3282c8f97d5ec4b36802333b29e8c241:
ipipe: fix wrong assumption in startup code of rt-timers (SMP) (2014-02-13
15:22:15 +0100)
are available in the git repository at:
git://git.xenomai.org/ipipe-gch.git for-ipipe-3.10
for you to fetch changes up to da6384562ba94d28ffa7931e016a26b7db0b8ef5:
ipipe/omap4: use the L2 cache write-allocate option (2014-02-24 01:26:47
+0100)
----------------------------------------------------------------
Charles Steinkuehler (1):
arm/ipipe: fix compilation error with THUMB2
Gilles Chanteperdrix (6):
arm/ipipe: fetch processor id from threadinfo in non-legacy mode
ipipe: introduce the ipipe_smp_p predicate
ipipe/spinlocks: use ipipe_smp_p to optimize-out spinlocks code
ipipe/arm: optimize the SMP_ON_UP case
ipipe/arm: add an option to disable L2 cache write-allocate
ipipe/omap4: use the L2 cache write-allocate option
arch/arm/include/asm/barrier.h | 44
+++++++++++++++++++++++++------
arch/arm/include/asm/hardware/cache-l2x0.h | 2 ++
arch/arm/include/asm/ipipe.h | 6 +++++
arch/arm/include/asm/ipipe_base.h | 32 +++++++++++++++++++++-
arch/arm/include/asm/percpu.h | 4 +--
arch/arm/kernel/entry-armv.S | 2 ++
arch/arm/kernel/ipipe.c | 22 +++++++++++++++-
arch/arm/mach-omap2/omap4-common.c | 3 ++-
arch/arm/mm/cache-l2x0.c | 13 +++++++++
include/linux/ipipe.h | 6 +++++
kernel/ipipe/core.c | 20 ++++++++------
11 files changed, 133 insertions(+), 21 deletions(-)
Could we do a stable update at this chance? There is a trivial conflict
in generic code and one in ARM in order to use recent 3.10. Maybe you
can have a look at the latter, Gilles:
http://git.xenomai.org/ipipe-jki.git/commit/?h=for-upstream/3.10&id=63866844101456d3f2f2405620eee7c4ad0bd88d
Since we are going to have to retest everything anyway when I am done
with the fd stuff, I would tend to prefer postponing the I-pipe stable
update at the same time. Even further than that: when we are ready to
release 3.0, so that if stable had a few more commits in the mean-time,
we will avoid useless tests.
Although time gets closer to the first 3.0-rc release, this is still a
few months ahead, and we are already lagging significantly behind
mainline wrt kernel updates.
It looks like Jan's team is ok to validate 3.10-stable on x86, and I can
do the same for other archs in the coming week. Let's roll out the
patches for these archs, postponing arm until you get a chance to fixup
the change impacting fcse there (does not seem bad at first glance).
Next on my plate will be to move the generic pipeline core to 3.14,
unless Jan has this brewing already.
Sorting out the issues which might remain regarding the RTDM API updates
will go a long way toward getting 3.x out.
--
Philippe.
_______________________________________________
Xenomai mailing list
[email protected]
http://www.xenomai.org/mailman/listinfo/xenomai
