[PATCH v3] sched: fix wrong rq's runnable_avg update with rt task

2013-02-13 Thread Vincent Guittot
When a RT task is scheduled on an idle CPU, the update of the rq's load is not done because CFS's functions are not called. Then, the idle_balance, which is called just before entering the idle function, updates the rq's load and makes the assumption that the elapsed time since the last update, was

Re: [PATCH] sched: fix env->src_cpu for active migration

2013-02-13 Thread Vincent Guittot
Damien, Regarding your sched_domain config and especially the flags field, you should not be impacted by my patch because - need_active_balance is the only new place that use env->src_cpu in the load_balance function - and your machine will never test the condition: "env->src_cpu > env->dst_cpu" i

Re: [PATCH] sched: fix env->src_cpu for active migration

2013-02-13 Thread Vincent Guittot
On 13 February 2013 15:08, Damien Wyart wrote: > * Vincent Guittot [2013-02-13 13:08]: >> Damien, >> Regarding your sched_domain config and especially the flags field, you >> should not be impacted by my patch because >> - need_active_balance is the only new place that use env->src_cpu in >> the

[PATCH] Samsung/Arndale: fix RELEASE build with ARMLINUXGCC

2013-02-13 Thread Ryan Harkin
The attached patch from Reece has already been applied to uefi-next as part of 13.02-rc5. But it's still open for review... 0001-Samsung-Arndale-fix-RELEASE-build-with-ARMLINUXGCC.patch Description: Binary data ___ linaro-dev mailing list linaro-dev@li

Re: [PATCH v3] sched: fix wrong rq's runnable_avg update with rt task

2013-02-13 Thread Paul Turner
On Wed, Feb 13, 2013 at 12:54 AM, Vincent Guittot wrote: > When a RT task is scheduled on an idle CPU, the update of the rq's load is > not done because CFS's functions are not called. Then, the idle_balance, > which is called just before entering the idle function, updates the > rq's load and mak

RE: [PATCH] Samsung/Arndale: fix RELEASE build with ARMLINUXGCC

2013-02-13 Thread Olivier Martin
Instead of using: #if !defined(MDEPKG_NDEBUG) (...) #endif It would be better to use: DEBUG_CODE_BEGIN(); (...) DEBUG_CODE_END(); The reason is you can be in debug build and disable debug code (and/or disable other things such as ASSERT). See definition of PcdDebugPropertyMask for the full li

Re: [PATCH v3] sched: fix wrong rq's runnable_avg update with rt task

2013-02-13 Thread Vincent Guittot
On 13 February 2013 16:28, Paul Turner wrote: > On Wed, Feb 13, 2013 at 12:54 AM, Vincent Guittot > wrote: >> When a RT task is scheduled on an idle CPU, the update of the rq's load is >> not done because CFS's functions are not called. Then, the idle_balance, >> which is called just before enter

Re: [PATCH] sched: fix env->src_cpu for active migration

2013-02-13 Thread Vincent Guittot
On 13 February 2013 15:28, Vincent Guittot wrote: > On 13 February 2013 15:08, Damien Wyart wrote: >> * Vincent Guittot [2013-02-13 13:08]: >>> Damien, >>> Regarding your sched_domain config and especially the flags field, you >>> should not be impacted by my patch because >>> - need_active_bala

Re: [PATCH] sched: fix env->src_cpu for active migration

2013-02-13 Thread Paul Turner
On Tue, Feb 12, 2013 at 5:19 AM, Vincent Guittot wrote: > need_active_balance uses env->src_cpu which is set only if there is more > than 1 task on the run queue. > We must set the src_cpu field unconditionnally > otherwise the test "env->src_cpu > env->dst_cpu" will always fail if there is > only

Re: [PATCH] Samsung/Arndale: fix RELEASE build with ARMLINUXGCC

2013-02-13 Thread Reece R. Pollack
On 02/13/2013 10:44 AM, Olivier Martin wrote: > Instead of using: > #if !defined(MDEPKG_NDEBUG) > (...) > #endif > > It would be better to use: > DEBUG_CODE_BEGIN(); > (...) > DEBUG_CODE_END(); > > The reason is you can be in debug build and disable debug code (and/or > disable other things suc