From mboxrd@z Thu Jan 1 00:00:00 1970 From: sebastian.hesselbarth@gmail.com (Sebastian Hesselbarth) Date: Fri, 20 Nov 2015 21:19:46 +0100 Subject: [PATCH RESEND 1/8] arm: dts: berlin2q: add watchdog nodes In-Reply-To: <20151120113414.1aa1e48e@xhacker> References: <1447672194-483-1-git-send-email-jszhang@marvell.com> <1447672194-483-2-git-send-email-jszhang@marvell.com> <564E3549.2050700@gmail.com> <20151120113414.1aa1e48e@xhacker> Message-ID: <564F8062.8070609@gmail.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On 20.11.2015 04:34, Jisheng Zhang wrote: > On Thu, 19 Nov 2015 21:47:05 +0100 > Sebastian Hesselbarth wrote: >> On 16.11.2015 12:09, Jisheng Zhang wrote: >>> The Marvell Berlin BG2Q has 3 watchdogs which are compatible with the >>> snps,dw-wdt driver sit in the sysmgr domain. This patch adds the >>> corresponding device tree nodes. >>> >>> Signed-off-by: Jisheng Zhang >>> --- >>> arch/arm/boot/dts/berlin2q.dtsi | 24 ++++++++++++++++++++++++ >>> 1 file changed, 24 insertions(+) >>> >>> diff --git a/arch/arm/boot/dts/berlin2q.dtsi b/arch/arm/boot/dts/berlin2q.dtsi >>> index a3ecde5..fac4315 100644 >>> --- a/arch/arm/boot/dts/berlin2q.dtsi >>> +++ b/arch/arm/boot/dts/berlin2q.dtsi >>> @@ -483,6 +483,30 @@ >>> ranges = <0 0xfc0000 0x10000>; >>> interrupt-parent = <&sic>; >>> >>> + wdt0: watchdog at 1000 { >>> + compatible = "snps,dw-wdt"; >>> + reg = <0x1000 0x100>; >>> + clocks = <&refclk>; >>> + interrupts = <0>; >>> + status = "disabled"; >> >> as the watchdogs are internal and cannot be clock gated >> at all, how about we remove the status = "disabled" and >> make them always available? > > there are two issues here: > > 1. the dw-wdt can't support multiple variants now. I have rewrite the driver > with watchdog core supplied framework, but the patch isn't sent out and > may be need time to clean up and review. Ok. > 2. not all dw-wdt devices are available and functional. This depends on > board design and configuration. I understand that "board design and configuration" may hinder the wdt to issue a hard reset. But all others are able to issue a soft reset or just an interrupt, right? So, I still don't see why we should disable wdt nodes by default except for the driver issue above. > So IMHO status=disabled and patch5-8 is necessary, what do you think? No. I'd agree to enable wdt0 by default and leave wdt[1,2] disabled because of the driver issue. Patches 5-8 only enable wdt0 anyway. As soon as the driver issue is resolved, we enable all wdt nodes unconditionally. Sebastian >> I have applied patches 1-4 with the status property removed. >> This also renders patches 5-8 useless. >> >> So, for now tentatively >> >> Appled to berlin/dt and berlin64/dt respectivly >> >> with status property removed. >> >> Sebastian >> >