From mboxrd@z Thu Jan 1 00:00:00 1970 From: Sebastian Andrzej Siewior Subject: Re: [ANNOUNCE] 3.12.9-rt13 Date: Fri, 7 Feb 2014 11:41:07 +0100 Message-ID: <20140207104107.GC23668@linutronix.de> References: <20140203194841.GA17108@linutronix.de> <20140205092657.GA21590@opentech.at> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Cc: linux-rt-users , LKML , Thomas Gleixner , rostedt@goodmis.org, John Kacur To: Nicholas Mc Guire Return-path: Content-Disposition: inline In-Reply-To: <20140205092657.GA21590@opentech.at> Sender: linux-kernel-owner@vger.kernel.org List-Id: linux-rt-users.vger.kernel.org * Nicholas Mc Guire | 2014-02-05 10:26:57 [+0100]: >Sorry - this one causes a build failure with PREEMPT_RT_BASE=y and >PREEMPT_RT_FULL not set. At some point we are going to drop PREEMPT_RT_BASE because it was only meant for debugging in the beginning. However I don't know when this will be. >The patch below fixes this build failure for 3.12.9-rt13. > >Not sure what the clean way of resolving this is - this patch proposes to >move the spin_*_local into linux/locallock.h and map to spin_*lock for >the "CONFIG_PREEMPT_RT_FULL not set" case. > >This was build tested with Preempt none,voluntary,low-lat,base,full and >otherweise got only limited testing. > >I'm also not sure if putting the rt specific locks into locallock.h in >this way is the proper way to deal with this #include dependency. I think I keep it that way. However I split into two patches and add the spin_lock_local => rt_spin_lock mapping in the rt-add-rt-locks.patch where the locks are introduced. Sebastian