From mboxrd@z Thu Jan 1 00:00:00 1970 From: swarren@wwwdotorg.org (Stephen Warren) Date: Fri, 24 Apr 2015 22:39:06 -0600 Subject: [PATCH 2/2] ARM: BCM2835: Move the restart/power_off handling to the WDT driver. In-Reply-To: <1429902534-2348-2-git-send-email-eric@anholt.net> References: <1429902534-2348-1-git-send-email-eric@anholt.net> <1429902534-2348-2-git-send-email-eric@anholt.net> Message-ID: <553B1A6A.6010205@wwwdotorg.org> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On 04/24/2015 01:08 PM, Eric Anholt wrote: > Since the WDT is what's used to drive restart and power off, it makes > more sense to keep it there, where the regs are already mapped and > definitions for them provided. Note that this means you may need to > add CONFIG_BCM2835_WDT to retain functionality of your kernel. The series, Acked-by: Stephen Warren Tested-by: Stephen Warren > Note that power off has never worked for me, and just reboots as well. > So I can't say that I've *really* tested the power off code. The RPi can't actually power itself off, but it used to be the case that if you rebooted it after setting up a certain register configuration, the firmware would put the device into a low-power state. This did work when it was first upstreamed. However, it no longer works. I believe this was due to a change in the firmware, which is why I don't always trust the firmware. I should really check what the downstream kernel does for power off now; I assume it must have changed since the code was upstreamed.