From mboxrd@z Thu Jan 1 00:00:00 1970 From: linux@arm.linux.org.uk (Russell King - ARM Linux) Date: Thu, 28 Apr 2011 19:15:02 +0100 Subject: [PATCH 09/14] at91: switch pit timer to early platform devices In-Reply-To: <1304009792.3081.8.camel@redbox> References: <20110425180847.GA12904@game.jcrosoft.org> <1303756284-26529-9-git-send-email-plagnioj@jcrosoft.com> <20110428113427.GG17290@n2100.arm.linux.org.uk> <20110428131538.GA10594@game.jcrosoft.org> <1304009792.3081.8.camel@redbox> Message-ID: <20110428181502.GE17290@n2100.arm.linux.org.uk> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Thu, Apr 28, 2011 at 06:56:32PM +0200, Andrew Victor wrote: > hi, > > > > And that's another reason to say no to this. Please stop inventing new > > > ways to do things unless you're prepared to provide it as a replacement > > > for all the (ARM) platforms we have in the kernel. > > Personally I do not change the current arm way > > > > I init the timer during system_timer->init > > > > I just the the early device their to pass the resources > > but if switch all arm timer this way is fine to you I'm ready to do the whole > > update > > Well, attached is a simpler way of passing the base-address to the timer > drivers. It will solve the timer issues with a single kernel supporting > multiple AT91 processors. > > I think it would be cleaner still if the "struct sys_timer" could > include a register-base-address field. Russell? That's not really what all platforms want, and with the advent of clocksources and clock events, I think the sys_timer thing is mostly dead. I've been thinking about killing it off for a while for those platforms using the GENERIC_CLOCKEVENTS stuff. From what I can see, the only real users of it are: acorn (riscpc), at91x40, clps711x, ebsa110, ep93xx, h720x, ixp2000, ixp23xx, ks8695, nuc93x, pnx4008, and samsung. It would be nice to get the modern stuff out of that converted to the clockevents/clocksource stuff - any chance of that happening for AT91x40?