From mboxrd@z Thu Jan 1 00:00:00 1970 From: Philippe Gerum In-Reply-To: <4580381E.4060501@domain.hid> References: <4576842B.1090203@domain.hid> <1165398116.7218.31.camel@domain.hid> <458032AC.4060707@domain.hid> <1166030688.5360.171.camel@domain.hid> <4580381E.4060501@domain.hid> Content-Type: text/plain Date: Wed, 13 Dec 2006 18:37:32 +0100 Message-Id: <1166031452.5360.180.camel@domain.hid> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Subject: [Xenomai-core] Re: [Adeos-main] [RFC] type-based spinlock op selection Reply-To: rpm@xenomai.org List-Id: "Xenomai life and development \(bug reports, patches, discussions\)" List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Jan Kiszka Cc: adeos-main@gna.org, xenomai-core On Wed, 2006-12-13 at 18:27 +0100, Jan Kiszka wrote: > Philippe Gerum wrote: > > On Wed, 2006-12-13 at 18:04 +0100, Jan Kiszka wrote: > >> Philippe Gerum wrote: > >>> On Wed, 2006-12-06 at 09:49 +0100, Jan Kiszka wrote: > >>>> Hi, > >>>> > >>>> reading through the I-pipe patch I noticed that there are quite a few > >>>> spin_lock->spin_lock_hw and similar conversions. Having the conversion > >>>> mechanism of the -rt patch in mind, I wrote the attached I-pipe cleanup > >>>> that applies a similar pattern on this issue. Gives more than 10k patch > >>>> size reduction for x86 and a bit more safety against minor variations of > >>>> the patched code. Still, review of the lock usages keeps on being > >>>> required as the lock user could assume running IRQs off while I-pipe may > >>>> not guarantee this. > >>>> > >>>> The required patch for Xenomai is attached as well. > >>>> > >>> I definitely like this. Queued for testing, thanks. > >>> > >> Small update of the ipipe part: typecasts were missing for the > >> SMP||DEBUG_SPINLOCK case. > > > > I saw this recently, thanks. I'm currently working on 1.6-02, which > > includes the spinlock op selection, and yet-another-rework of the > > interrupt flow. Did anything changed regarding the tracer in the > > attached patch? > > > > Nope, only the linux/spinlock.h should be different (unless I messed > something up). Do you get rejects? No, that's ok. I was wondering whether I should pick changes aside of spinlock.h, which I have reworked. > > I'm currently preparing to post a few cleanup patches for I-pipe (and > Xenomai). Not a must for -02, but maybe worth a glance. Just a minute... > -- Philippe.