Wrong button make me removed others guys from the thread.
Sorry for this mistake.
On 13 September 2012 09:56, Mike Galbraith <efault(a)gmx.de> wrote:
> On Thu, 2012-09-13 at 09:44 +0200, Vincent Guittot wrote:
>> On 13 September 2012 09:29, Mike Galbraith <efault(a)gmx.de> wrote:
>> > On Thu, 2012-09-13 at 08:59 +0200, Vincent Guittot wrote:
>> >> On 13 September 2012 08:49, Mike Galbraith <efault(a)gmx.de> wrote:
>> >> > On Thu, 2012-09-13 at 06:11 +0200, Vincent Guittot wrote:
>> >> >> On tickless system, one CPU runs load balance for all idle CPUs.
>> >> >> The cpu_load of this CPU is updated before starting the load balance
>> >> >> of each other idle CPUs. We should instead update the cpu_load of the balance_cpu.
>> >> >>
>> >> >> Signed-off-by: Vincent Guittot <vincent.guittot(a)linaro.org>
>> >> >> ---
>> >> >> kernel/sched/fair.c | 11 ++++++-----
>> >> >> 1 file changed, 6 insertions(+), 5 deletions(-)
>> >> >>
>> >> >> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>> >> >> index 1ca4fe4..9ae3a5b 100644
>> >> >> --- a/kernel/sched/fair.c
>> >> >> +++ b/kernel/sched/fair.c
>> >> >> @@ -4794,14 +4794,15 @@ static void nohz_idle_balance(int this_cpu, enum cpu_idle_type idle)
>> >> >> if (need_resched())
>> >> >> break;
>> >> >>
>> >> >> - raw_spin_lock_irq(&this_rq->lock);
>> >> >> - update_rq_clock(this_rq);
>> >> >> - update_idle_cpu_load(this_rq);
>> >> >> - raw_spin_unlock_irq(&this_rq->lock);
>> >> >> + rq = cpu_rq(balance_cpu);
>> >> >> +
>> >> >> + raw_spin_lock_irq(&rq->lock);
>> >> >> + update_rq_clock(rq);
>> >> >> + update_idle_cpu_load(rq);
>> >> >> + raw_spin_unlock_irq(&rq->lock);
>> >> >>
>> >> >> rebalance_domains(balance_cpu, CPU_IDLE);
>> >> >>
>> >> >> - rq = cpu_rq(balance_cpu);
>> >> >> if (time_after(this_rq->next_balance, rq->next_balance))
>> >> >> this_rq->next_balance = rq->next_balance;
>> >> >> }
>> >> >
>> >> > Ew, banging locks and updating clocks to what good end?
>> >> The goal is to update the cpu_load table of the CPU before starting
>> >> the load balance. Other wise we will use outdated value in the load
>> >> balance sequence
>> > If there's load to distribute, seems it should all work out fine without
>> > doing that. What harm is being done that makes this worth while?
>> this_load and avg_load can be wrong and make an idle CPU set as
>> balanced compared to the busy one
> I think you need to present numbers showing benefit. Crawling all over
> a mostly idle (4096p?) box is decidedly bad thing to do.
Yep, let me prepare some figures
You should also notice that you are already crawling all over the idle
processor in rebalance_domains
A head's up for people submitting fastmodel jobs to LAVA. We've recently
decided on a different naming scheme for the device/device-types in the
lab. Previously it was:
device: fastmodel01, fastmodel02
We are changing this to be:
device: rtsm_ve-a15x4-a7x4_01, rtsm_ve-a15x4-a7x4_02
and I'll also be creating a new device/type:
device: rtsm_ve-a15x1-a7x1_01, rtsm_ve-a15x1-a7x1_02
How does this affect me? You'll need to change your job you submit to
LAVA to match this new device type.
Postmortem and lessons learned for Linaro's release 2012.08
Highlights and Key Successes
Linaro also held its first Virtual Connect event on 13 - 17 August,
2012. This online event used Google+ Hangouts on Air and was streamed
and recorded via YouTube. Members of the various Linaro Teams provided
informative, educational and exciting sessions that allowed for public
viewing and participation.
The Android team created stable Jelly Bean builds from AOSP and an
Android benchmarking platform to automate many of the tests in the
The Developer Platform Team announced USB Host enablement is now
available in LAVA and that all packaged linux-linaro kernels are now
cross-compiled and boot tested via LAVA before they are uploaded to
Linaro Overlay PPA. Additionally, U-Boot-Linaro is now based on the
latest upstream release--v2012.07.
Postmortem and Lessons Learned
This cycle was fairly low key: lots of people were out for vacations
and many others were attending external conferences such as Linux
One issue of concern was the deployment of an infrastructure project
on the last week before release.
A the new django application was deployed on snapshots.linaro.org that
disrupted several builds on the Friday before the release. The issues
took one week to resolve.
The issues are:
- The deployment was not communicated and broadcast properly
- Build maintainers were not given proper warning
- deploying new functional software on a Friday with no one
available on the weekend for emergency
- deploying during a release week
A big "thank you" is in order for Georgy Redkozubov (gesha) for his
help and availability in troubleshooting and getting it fixed.
An incident report has been filed here:
The number of high or essential priority blueprints that missed the cycle:
Android 3 out of 13
Developer Platform 0 out of 3
Infrastructure 1 out of 6
Lava 2 out of 4
Total 6 out of 26
23% of high or essential priority blueprints scheduled for this cycle
were not delivered.
Total blueprints: 14 out of 49 missed the cycle.
High priority missed blueprints recap:
12.05: 19 out of 48, 39%
12.06: 13 out of 31, 42%
12.07: 14 out of 31, 45%
12.08: 6 out of 26, 23%
* Not included is data from working groups and landing teams
Linaro Release Manager | Project Manager
Linaro.org | Open source software for ARM SoCs
I am trying to boot kernel on panda board. I got the source from:
Following is my branch HEAD:
Merge: 1d10459 8b29cd5
Author: Andrey Konovalov <andrey.konovalov(a)linaro.org>
Date: Mon Sep 3 17:10:24 2012 +0400
Merge branch 'tracking-ll-last-minute-fixes' into merge-linux-linaro
I am using following for compilation: omap2plus_defconfig, omap4-panda.dts
When i try to boot, kernel hangs after uncompressing linux...
If i just remove the 0x815f0000 from bootm command.. i.e. boot without dtb...
it boots fine.
Earlier i have flashed my card using linaro-media-create. Got, H/w package and
devel rootfs from 12.08 release. Replaced uImage and board.dtb in /dev/mmcblk0p1