diff mbox

arm64: enable generic idle loop

Message ID 1440128422-4170-1-git-send-email-leo.yan@linaro.org (mailing list archive)
State New, archived
Headers show

Commit Message

Leo Yan Aug. 21, 2015, 3:40 a.m. UTC
Enable generic idle loop for ARM64, so can support for hlt/nohlt
command line options to override default idle loop behavior.

Signed-off-by: Leo Yan <leo.yan@linaro.org>
---
 arch/arm64/Kconfig | 1 +
 1 file changed, 1 insertion(+)

Comments

Catalin Marinas Aug. 21, 2015, 9 a.m. UTC | #1
On Fri, Aug 21, 2015 at 11:40:22AM +0800, Leo Yan wrote:
> Enable generic idle loop for ARM64, so can support for hlt/nohlt
> command line options to override default idle loop behavior.

Just curious, what do you need this for?
Leo Yan Aug. 21, 2015, 9:14 a.m. UTC | #2
On Fri, Aug 21, 2015 at 10:00:43AM +0100, Catalin Marinas wrote:
> On Fri, Aug 21, 2015 at 11:40:22AM +0800, Leo Yan wrote:
> > Enable generic idle loop for ARM64, so can support for hlt/nohlt
> > command line options to override default idle loop behavior.
> 
> Just curious, what do you need this for?

With this config, usually we can add "nohlt" in command line to force
the cpu will never be powered off; so can quickly narrow down if some
issues are related with cpuidle or not.

And if CPU is powered off, then its debug logic also will be powered off.
So we can simply use this method to let debugger always hot connect with
the target CPU.

Thanks,
Leo Yan
Catalin Marinas Aug. 21, 2015, 3:54 p.m. UTC | #3
On Fri, Aug 21, 2015 at 05:14:46PM +0800, Leo Yan wrote:
> On Fri, Aug 21, 2015 at 10:00:43AM +0100, Catalin Marinas wrote:
> > On Fri, Aug 21, 2015 at 11:40:22AM +0800, Leo Yan wrote:
> > > Enable generic idle loop for ARM64, so can support for hlt/nohlt
> > > command line options to override default idle loop behavior.
> > 
> > Just curious, what do you need this for?
> 
> With this config, usually we can add "nohlt" in command line to force
> the cpu will never be powered off; so can quickly narrow down if some
> issues are related with cpuidle or not.
> 
> And if CPU is powered off, then its debug logic also will be powered off.
> So we can simply use this method to let debugger always hot connect with
> the target CPU.

OK.

Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Leo Yan Sept. 15, 2015, 11:20 a.m. UTC | #4
Hi Will,

On Fri, Aug 21, 2015 at 04:54:25PM +0100, Catalin Marinas wrote:
> On Fri, Aug 21, 2015 at 05:14:46PM +0800, Leo Yan wrote:
> > On Fri, Aug 21, 2015 at 10:00:43AM +0100, Catalin Marinas wrote:
> > > On Fri, Aug 21, 2015 at 11:40:22AM +0800, Leo Yan wrote:
> > > > Enable generic idle loop for ARM64, so can support for hlt/nohlt
> > > > command line options to override default idle loop behavior.
> > > 
> > > Just curious, what do you need this for?
> > 
> > With this config, usually we can add "nohlt" in command line to force
> > the cpu will never be powered off; so can quickly narrow down if some
> > issues are related with cpuidle or not.
> > 
> > And if CPU is powered off, then its debug logic also will be powered off.
> > So we can simply use this method to let debugger always hot connect with
> > the target CPU.
> 
> OK.
> 
> Acked-by: Catalin Marinas <catalin.marinas@arm.com>

Suppose you have missed this patch at previous merge window,
so just for reminding.

Thanks,
Leo Yan
Catalin Marinas Sept. 15, 2015, 2:31 p.m. UTC | #5
On Tue, Sep 15, 2015 at 07:20:08PM +0800, Leo Yan wrote:
> On Fri, Aug 21, 2015 at 04:54:25PM +0100, Catalin Marinas wrote:
> > On Fri, Aug 21, 2015 at 05:14:46PM +0800, Leo Yan wrote:
> > > On Fri, Aug 21, 2015 at 10:00:43AM +0100, Catalin Marinas wrote:
> > > > On Fri, Aug 21, 2015 at 11:40:22AM +0800, Leo Yan wrote:
> > > > > Enable generic idle loop for ARM64, so can support for hlt/nohlt
> > > > > command line options to override default idle loop behavior.
> > > > 
> > > > Just curious, what do you need this for?
> > > 
> > > With this config, usually we can add "nohlt" in command line to force
> > > the cpu will never be powered off; so can quickly narrow down if some
> > > issues are related with cpuidle or not.
> > > 
> > > And if CPU is powered off, then its debug logic also will be powered off.
> > > So we can simply use this method to let debugger always hot connect with
> > > the target CPU.
> > 
> > OK.
> > 
> > Acked-by: Catalin Marinas <catalin.marinas@arm.com>
> 
> Suppose you have missed this patch at previous merge window,
> so just for reminding.

It looks like it missed the merging window. Will, can this still be
picked for 4.3? If not, it will be 4.4.
Will Deacon Sept. 15, 2015, 2:44 p.m. UTC | #6
On Tue, Sep 15, 2015 at 03:31:35PM +0100, Catalin Marinas wrote:
> On Tue, Sep 15, 2015 at 07:20:08PM +0800, Leo Yan wrote:
> > On Fri, Aug 21, 2015 at 04:54:25PM +0100, Catalin Marinas wrote:
> > > On Fri, Aug 21, 2015 at 05:14:46PM +0800, Leo Yan wrote:
> > > > On Fri, Aug 21, 2015 at 10:00:43AM +0100, Catalin Marinas wrote:
> > > > > On Fri, Aug 21, 2015 at 11:40:22AM +0800, Leo Yan wrote:
> > > > > > Enable generic idle loop for ARM64, so can support for hlt/nohlt
> > > > > > command line options to override default idle loop behavior.
> > > > > 
> > > > > Just curious, what do you need this for?
> > > > 
> > > > With this config, usually we can add "nohlt" in command line to force
> > > > the cpu will never be powered off; so can quickly narrow down if some
> > > > issues are related with cpuidle or not.
> > > > 
> > > > And if CPU is powered off, then its debug logic also will be powered off.
> > > > So we can simply use this method to let debugger always hot connect with
> > > > the target CPU.
> > > 
> > > OK.
> > > 
> > > Acked-by: Catalin Marinas <catalin.marinas@arm.com>
> > 
> > Suppose you have missed this patch at previous merge window,
> > so just for reminding.
> 
> It looks like it missed the merging window. Will, can this still be
> picked for 4.3? If not, it will be 4.4.

Yeah, it's harmless enough to pick up with the current crop of fixes.
Sorry I missed it.

Will
Leo Yan Sept. 15, 2015, 2:51 p.m. UTC | #7
On Tue, Sep 15, 2015 at 03:44:09PM +0100, Will Deacon wrote:
> On Tue, Sep 15, 2015 at 03:31:35PM +0100, Catalin Marinas wrote:
> > On Tue, Sep 15, 2015 at 07:20:08PM +0800, Leo Yan wrote:
> > > On Fri, Aug 21, 2015 at 04:54:25PM +0100, Catalin Marinas wrote:
> > > > On Fri, Aug 21, 2015 at 05:14:46PM +0800, Leo Yan wrote:
> > > > > On Fri, Aug 21, 2015 at 10:00:43AM +0100, Catalin Marinas wrote:
> > > > > > On Fri, Aug 21, 2015 at 11:40:22AM +0800, Leo Yan wrote:
> > > > > > > Enable generic idle loop for ARM64, so can support for hlt/nohlt
> > > > > > > command line options to override default idle loop behavior.
> > > > > > 
> > > > > > Just curious, what do you need this for?
> > > > > 
> > > > > With this config, usually we can add "nohlt" in command line to force
> > > > > the cpu will never be powered off; so can quickly narrow down if some
> > > > > issues are related with cpuidle or not.
> > > > > 
> > > > > And if CPU is powered off, then its debug logic also will be powered off.
> > > > > So we can simply use this method to let debugger always hot connect with
> > > > > the target CPU.
> > > > 
> > > > OK.
> > > > 
> > > > Acked-by: Catalin Marinas <catalin.marinas@arm.com>
> > > 
> > > Suppose you have missed this patch at previous merge window,
> > > so just for reminding.
> > 
> > It looks like it missed the merging window. Will, can this still be
> > picked for 4.3? If not, it will be 4.4.
> 
> Yeah, it's harmless enough to pick up with the current crop of fixes.
> Sorry I missed it.

It's fine and thanks :)
diff mbox

Patch

diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
index 0f6edb1..c404e74 100644
--- a/arch/arm64/Kconfig
+++ b/arch/arm64/Kconfig
@@ -31,6 +31,7 @@  config ARM64
 	select GENERIC_CLOCKEVENTS_BROADCAST if SMP
 	select GENERIC_CPU_AUTOPROBE
 	select GENERIC_EARLY_IOREMAP
+	select GENERIC_IDLE_POLL_SETUP
 	select GENERIC_IRQ_PROBE
 	select GENERIC_IRQ_SHOW
 	select GENERIC_IRQ_SHOW_LEVEL