From mboxrd@z Thu Jan 1 00:00:00 1970 From: khilman@ti.com (Kevin Hilman) Date: Mon, 27 Jun 2011 12:04:17 -0700 Subject: [PATCHv2] omap3: beagle: Use GPTIMER1 for clockevents In-Reply-To: (Sanjeev Premi's message of "Mon, 27 Jun 2011 18:58:54 +0530") References: <1309171009-8075-1-git-send-email-premi@ti.com> <20110627111556.GM23145@atomide.com> <20110627123140.GP23145@atomide.com> Message-ID: <87vcvrayfi.fsf@ti.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org "Premi, Sanjeev" writes: >> -----Original Message----- >> From: Tony Lindgren [mailto:tony at atomide.com] >> Sent: Monday, June 27, 2011 6:02 PM >> To: Premi, Sanjeev >> Cc: linux-omap at vger.kernel.org; >> linux-arm-kernel at lists.infradead.org; Gregoire Gentil; >> Bhandiwad, Hrishikesh; Jason Lam; Thomas Weber >> Subject: Re: [PATCHv2] omap3: beagle: Use GPTIMER1 for clockevents >> >> * Premi, Sanjeev [110627 05:08]: >> > > From: Tony Lindgren [mailto:tony at atomide.com] >> > > >> > > I don't think omap3_beagle_init_rev is even called when >> > > the timer is set? >> > >> > [sp] I verified the patch based on the print indicating that >> > GPTIMER1 being used as clockevent source. >> > http://marc.info/?l=linux-omap&m=130893319726456&w=2 >> >> I suspect the test always fails though, so it probably never >> gets set to gptimer12 on any board :) >> > > [sp] While I take my time understanding things on devel-timer; > I had a quick question - at risk of being flamed. > > Adding a new machine ID would trickle to u-boot and same > uImage (default) may not work across board revisions. > > How does this scheme look like: > - GPTIMER1 is used as default - as it works for most boards. > - GPTIMER12 is used based on a static config option OR a > board specific bootarg > > I know both these options aren't general practice. Still > wanted to know your views in the current context. > Another option may be to always use GPTIMER1 on early boot for all boards. Later in the boot, when the old rev boards are detected, a new GPT12 clockevent could be registered (with a higher rating) such that the clockevent core would switch to the new source. Kevin