From mboxrd@z Thu Jan 1 00:00:00 1970 From: Sebastian Andrzej Siewior Subject: Re: [PATCH RT 3/6] rt: Make cpu_chill() use hrtimer instead of msleep() Date: Fri, 07 Mar 2014 16:23:10 +0100 Message-ID: <5319E45E.20900@linutronix.de> References: <20140305003333.402759240@goodmis.org> <20140305003346.602499217@goodmis.org> <531996BB.2050401@linutronix.de> <20140307095223.2f412a1a@gandalf.local.home> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: linux-kernel@vger.kernel.org, linux-rt-users , Thomas Gleixner , Carsten Emde , John Kacur , Paul Gortmaker , stable-rt@vger.kernel.org To: Steven Rostedt Return-path: In-Reply-To: <20140307095223.2f412a1a@gandalf.local.home> Sender: linux-kernel-owner@vger.kernel.org List-Id: linux-rt-users.vger.kernel.org On 03/07/2014 03:52 PM, Steven Rostedt wrote: >> Now that you posted "cpu_chill: Add a UNINTERRUPTIBLE >> hrtimer_nanosleep" wouldn't it make sense to delay this patches from >> the stable series until we get them all in one go? > > Sure, say this on the day I'm about to release ;-) Haven't noticed that earlier. > Nah, I'll release these today anyway. Otherwise I can't update to the > mainline stables. When the rc candidates are out, it holds up any new > updates to the versions. > > Also, the stable 3.10 has this bug already. Might as well update all > the stables with the same fix. It makes it easier on my side, as the > way I do the updates is to use the same quilt queue for all releases. I > start with 3.10, and get them working, and then apply the same queue to > 3.8. Any conflicts in the patch I do a quilt fork, with a -v3.8 > appended, and then after getting that working I go to 3.4, and so on. > This also lets me drop patches that are not applicable for earlier > releases. > > If I delay this for a new update, then 3.10 will be out of sync. I'm > just going to release these, and then we can start a new one right away. No worries. Do as it is easier for you. > > Thanks! > > -- Steve > Sebastian