On Thursday 08 September 2011 11:27 PM, Kevin Hilman wrote:
Hi Santosh,
Santosh Shilimkar<santosh.shilim...@ti.com> writes:
These notifiers can be used to isolate non-cpu specific PM
code from CPU idle code.
To start with, we can already leverage VFP handlers using notifier
chain for low power idle code. So use them.
Signed-off-by: Santosh Shilimkar<santosh.shilim...@ti.com>
Cc: Kevin Hilman<khil...@ti.com>
A comment below related to our other discussion in the CPU PM notifiers
thread...
After the discussion, this patch needs to be updated.
---
arch/arm/mach-omap2/cpuidle34xx.c | 7 +++++++
1 files changed, 7 insertions(+), 0 deletions(-)
diff --git a/arch/arm/mach-omap2/cpuidle34xx.c
b/arch/arm/mach-omap2/cpuidle34xx.c
index 4bf6e6e..1112519 100644
--- a/arch/arm/mach-omap2/cpuidle34xx.c
+++ b/arch/arm/mach-omap2/cpuidle34xx.c
@@ -24,6 +24,7 @@
#include<linux/sched.h>
#include<linux/cpuidle.h>
+#include<linux/cpu_pm.h>
#include<plat/prcm.h>
#include<plat/irqs.h>
@@ -118,9 +119,15 @@ static int omap3_enter_idle(struct cpuidle_device *dev,
pwrdm_for_each_clkdm(core_pd, _cpuidle_deny_idle);
}
+ if (mpu_state == PWRDM_POWER_OFF)
+ cpu_pm_enter();
Making all CPU PM notifiers conditional on the MPU power state is not
right since upon an MPU transition, other domains might also transition.
What if there are notifiers for other drivers in other power domains
whose states might be transitioning? Notifiers for those devices should
be called no matter what the MPU's target state is.
Also triggering the notifiers here has the same problem that I mentioned
with the syscore PM ops. Namely, that next power states are not yet
determined, so any device notifier decisions based on that will be wrong.
The way I envision this working is (at least on OMAP) is CPU PM
notifiers are triggered only *after* the next power states are
determined and programmed. This allows the device-specific notifiers to
check their target power state and take appropriate action.
I've been testing the CPU PM notifiers with the patch below (after also
dropping [PATCH 2/5] from your the CPU PM notifiers series.) This patch works
for both suspend and idle by using the code common to them, and
triggering the event only after the target states are programmed.
I'm testing this patch now along with the 2nd patch below which removes
all the UART calls from the idle/suspend path and just adds a notifier
in the UART driver.
Will post these officially after a bit more testing, but so far it's
working on OMAP2420/n810 and OMAP3430/n900.
Agree. I will drop my patch in favour of below patch.
Regards
Santosh
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html