linaro-dev December 2012

linaro-dev@lists.linaro.org

55 participants
57 discussions

by Dave Pigott

Hi all, I was just discussing IPv6 with Philip Colmer, our new IT Services Manager (cc'd on this mail), and it strikes me that we should at least be considering dual running at some point in the future, i.e. providing both v4 and v6. I'm not clear what the ramifications are, or as yet whether Zen will support it. Philip has experience with this, and seems to remember that Zen do support it, but I'll bang an e-mail out to them to check. The reason for this e-mail is to start a discussion as to whether we think it's worth raising a BP, or if we can ignore this issue. Thoughts, comments and brickbats welcome. Thanks Dave

12 years

Test Result Summary for Calendar Week 50 for Linux Linaro ubuntu Quantal.

by Botao Sun

Calendar Week 50: Here is test result summary for Linux Linaro ubuntu Quantal image on following boards: 1) ARM Versatile Express A9; 2) Samsung Origen; 3) TI Panda 4430; 4) TI Panda 4460; 5) ST Ericsson Snowball. Synopsis: All Linux Linaro ubuntu Quantal builds for above boards are available. No network connection on Samsung Origen; ST Ericsson Snowball boots failed. Device Tree is unavailable except on Samsung Origen. 1. ARM Versatile Express A9 + Linux Linaro Quantal (Column D): https://docs.google.com/a/linaro.org/spreadsheet/ccc?key=0AroPySpr4FnEdFNmV… Only "Halt" & "Device Tree" test failed, all other features work well. 2. Samsung Origen + Linux Linaro Quantal (Column C): https://docs.google.com/a/linaro.org/spreadsheet/ccc?key=0AroPySpr4FnEdEowN… No network connection in this image, neither Ethernet nor WiFi works. Also, "reboot" & "halt" doesn't work. Power Management test would hang system when run test "cpuhotplug_02". 3. TI Panda 4430 + Linux Linaro Quantal (Column D): https://docs.google.com/a/linaro.org/spreadsheet/ccc?key=0AroPySpr4FnEdEwwZ… Data streaming failed in DS-5 due to Gator version mismatch. Device Tree is unavailable, the directory "/proc/device-tree" is empty. All other features work well. 4. TI Panda 4460 + Linux Linaro Quantal (Column D): https://docs.google.com/a/linaro.org/spreadsheet/ccc?key=0AroPySpr4FnEdEwwZ… Exactly same status as TI Panda 4430 test result - "Device Tree" & "DS5 Data Streaming" test failed. 5. ST Ericsson Snowball + Linux Linaro Quantal (Column D): https://docs.google.com/a/linaro.org/spreadsheet/ccc?key=0AroPySpr4FnEdFJ4X… Same as last week, system boots failed. This blocked all other test cases. For the previous week (Calendar week 49) summary, please refer to attachment. Thank you. Best Regards Botao Sun QA Team WiKi Page: https://wiki.linaro.org/Platform/QA

12 years

Re: [PATCH] cpuidle: reinitialize power_usage values when adding/removing C-states

by Daniel Lezcano

On 10/17/2012 12:39 AM, Julius Werner wrote: > When cpuidle drivers do not supply explicit power_usage values, > cpuidle/driver.c inserts dummy values instead. When a running processor > dynamically gains new C-states (e.g. after ACPI events), the power_usage > values of those states will stay uninitialized, and cpuidle governors > will never choose to enter them. > > This patch moves the dummy value initialization from > cpuidle_register_driver to cpuidle_enable_device, which drivers such as > acpi/processor_idle.c will call again when they add or remove C-states. > Tested and verified on an Acer AC700 Chromebook, which supports the C3 > state only when it switches from AC to battery power. > > Signed-off-by: Julius Werner <jwerner(a)chromium.org> > --- This is specific to the acpi and should be handled in the processor_idle.c file instead of the cpuidle core code. Could be the function 'acpi_processor_cst_has_changed' the right place to set a dummy power value for the power in the new C-state ? > drivers/cpuidle/cpuidle.c | 24 ++++++++++++++++++++++++ > drivers/cpuidle/driver.c | 25 ------------------------- > 2 files changed, 24 insertions(+), 25 deletions(-) > > diff --git a/drivers/cpuidle/cpuidle.c b/drivers/cpuidle/cpuidle.c > index 7f15b85..bef3a31 100644 > --- a/drivers/cpuidle/cpuidle.c > +++ b/drivers/cpuidle/cpuidle.c > @@ -298,6 +298,28 @@ static void poll_idle_init(struct cpuidle_driver *drv) > static void poll_idle_init(struct cpuidle_driver *drv) {} > #endif /* CONFIG_ARCH_HAS_CPU_RELAX */ > > +static void set_power_states(struct cpuidle_driver *drv) > +{ > + int i; > + > + /* > + * cpuidle driver should set the drv->power_specified bit > + * before registering if the driver provides > + * power_usage numbers. > + * > + * If power_specified is not set, > + * we fill in power_usage with decreasing values as the > + * cpuidle code has an implicit assumption that state Cn > + * uses less power than C(n-1). > + * > + * With CONFIG_ARCH_HAS_CPU_RELAX, C0 is already assigned > + * an power value of -1. So we use -2, -3, etc, for other > + * c-states. > + */ > + for (i = CPUIDLE_DRIVER_STATE_START; i < drv->state_count; i++) > + drv->states[i].power_usage = -1 - i; > +} > + > /** > * cpuidle_enable_device - enables idle PM for a CPU > * @dev: the CPU > @@ -330,6 +352,8 @@ int cpuidle_enable_device(struct cpuidle_device *dev) > cpuidle_enter_tk : cpuidle_enter; > > poll_idle_init(drv); > + if (!drv->power_specified) > + set_power_states(drv); > > if ((ret = cpuidle_add_state_sysfs(dev))) > return ret; > diff --git a/drivers/cpuidle/driver.c b/drivers/cpuidle/driver.c > index 87db387..caaed27 100644 > --- a/drivers/cpuidle/driver.c > +++ b/drivers/cpuidle/driver.c > @@ -18,28 +18,6 @@ static struct cpuidle_driver *cpuidle_curr_driver; > DEFINE_SPINLOCK(cpuidle_driver_lock); > int cpuidle_driver_refcount; > > -static void set_power_states(struct cpuidle_driver *drv) > -{ > - int i; > - > - /* > - * cpuidle driver should set the drv->power_specified bit > - * before registering if the driver provides > - * power_usage numbers. > - * > - * If power_specified is not set, > - * we fill in power_usage with decreasing values as the > - * cpuidle code has an implicit assumption that state Cn > - * uses less power than C(n-1). > - * > - * With CONFIG_ARCH_HAS_CPU_RELAX, C0 is already assigned > - * an power value of -1. So we use -2, -3, etc, for other > - * c-states. > - */ > - for (i = CPUIDLE_DRIVER_STATE_START; i < drv->state_count; i++) > - drv->states[i].power_usage = -1 - i; > -} > - > /** > * cpuidle_register_driver - registers a driver > * @drv: the driver > @@ -58,9 +36,6 @@ int cpuidle_register_driver(struct cpuidle_driver *drv) > return -EBUSY; > } > > - if (!drv->power_specified) > - set_power_states(drv); > - > cpuidle_curr_driver = drv; > > spin_unlock(&cpuidle_driver_lock); -- <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs Follow Linaro: <http://www.facebook.com/pages/Linaro> Facebook | <http://twitter.com/#!/linaroorg> Twitter | <http://www.linaro.org/linaro-blog/> Blog

12 years

LAVA downtime

by Dave Pigott

Hi all, Just a warning that there may be a little disruption in LAVA today because I'm switching over to a wider IP address space in the lab. If all goes according to plan, there should be nothing noticeable, but just to be on the safe side I'm warning you. For the moment, all boards will be offline, but the scheduler will still be running and queuing jobs for when they come back. Thanks Dave

12 years

[PATCH] merge_config.sh: Add option to specify output dir

by Zhangfei Gao

Provide a -O option to specify dir to put generated .config Signed-off-by: Zhangfei Gao <zhangfei.gao(a)linaro.org> Tested-by: Jon Medhurst (Tixy) <tixy(a)linaro.org> --- scripts/kconfig/merge_config.sh | 18 +++++++++++++++--- 1 file changed, 15 insertions(+), 3 deletions(-) diff --git a/scripts/kconfig/merge_config.sh b/scripts/kconfig/merge_config.sh index 974d5cb..9ad5744 100755 --- a/scripts/kconfig/merge_config.sh +++ b/scripts/kconfig/merge_config.sh @@ -32,11 +32,13 @@ usage() { echo " -m only merge the fragments, do not execute the make command" echo " -n use allnoconfig instead of alldefconfig" echo " -r list redundant entries when merging fragments" + echo " -O dir to put generated output files" } MAKE=true ALLTARGET=alldefconfig WARNREDUN=false +OUTPUT=. while true; do case $1 in @@ -59,6 +61,16 @@ while true; do shift continue ;; + "-O") + if [ -d $2 ];then + OUTPUT=$2 + else + echo "output directory $2 does not exist" 1>&2 + exit 1 + fi + shift 2 + continue + ;; *) break ;; @@ -100,7 +112,7 @@ for MERGE_FILE in $MERGE_LIST ; do done if [ "$MAKE" = "false" ]; then - cp $TMP_FILE .config + cp $TMP_FILE $OUTPUT/.config echo "#" echo "# merged configuration written to .config (needs make)" echo "#" @@ -111,14 +123,14 @@ fi # Use the merged file as the starting point for: # alldefconfig: Fills in any missing symbols with Kconfig default # allnoconfig: Fills in any missing symbols with # CONFIG_* is not set -make KCONFIG_ALLCONFIG=$TMP_FILE $ALLTARGET +make KCONFIG_ALLCONFIG=$TMP_FILE O=$OUTPUT $ALLTARGET # Check all specified config values took (might have missed-dependency issues) for CFG in $(sed -n "$SED_CONFIG_EXP" $TMP_FILE); do REQUESTED_VAL=$(grep -w -e "$CFG" $TMP_FILE) - ACTUAL_VAL=$(grep -w -e "$CFG" .config) + ACTUAL_VAL=$(grep -w -e "$CFG" $OUTPUT/.config) if [ "x$REQUESTED_VAL" != "x$ACTUAL_VAL" ] ; then echo "Value requested for $CFG not in final .config" echo "Requested value: $REQUESTED_VAL" -- 1.7.9.5

12 years

[GIT PULL]; big LITTLE MP master v13

by Viresh Kumar

[Already Pulled by Andrey] Hi Guys, This is an formal announcement of v13 release. Updates: - Created more branches with same stuff to keep topics separated. - Applied additional patch from Morten to task-placement-v2 branch --------------x---------------------x-------------------- The following changes since commit 77b67063bb6bce6d475e910d3b886a606d0d91f7: Linux 3.7-rc5 (2012-11-11 13:44:33 +0100) are available in the git repository at: git://git.linaro.org/arm/big.LITTLE/mp.git big-LITTLE-MP-master-v13 for you to fetch changes up to fba4f1ff31d4303efa7270936930c4e01931172a: Merge branches 'arm-multi_pmu_v2', 'hw-bkp-v7.1-debug-v1', 'task-placement-v2-sysfs', 'misc-patches', 'config-fragments' and 'sched-pack-small-tasks-v1-arm' into big-LITTLE-MP-master-v13 (2012-12-07 20:09:17 +0530) ---------------------------------------------------------------- Ben Segall (1): sched: Maintain per-rq runnable averages Chris Redpath (2): ARM: Experimental Frequency-Invariant Load Scaling Patch ARM: Fix build breakage when big.LITTLE.conf is not used. Dietmar Eggemann (1): ARM: hw_breakpoint: v7.1 self-hosted debug powerdown support Jon Medhurst (1): ARM: sched: Avoid empty 'slow' HMP domain Liviu Dudau (2): linaro/configs: big-LITTLE-MP: Enable the new tunable sysfs interface by default. Revert "sched: secure access to other CPU statistics" Lorenzo Pieralisi (1): ARM: kernel: provide cluster to logical cpu mask mapping API Marc Zyngier (1): ARM: perf: add guest vs host discrimination Mark Rutland (1): ARM: perf: register cpu_notifier at driver init Morten Rasmussen (15): sched: entity load-tracking load_avg_ratio sched: Task placement for heterogeneous systems based on task load-tracking sched: Forced task migration on heterogeneous systems sched: Introduce priority-based task migration filter ARM: Add HMP scheduling support for ARM architecture ARM: sched: Use device-tree to provide fast/slow CPU list for HMP ARM: sched: Setup SCHED_HMP domains sched: Add ftrace events for entity load-tracking sched: Add HMP task migration ftrace event sched: SCHED_HMP multi-domain task migration control sched: Enable HMP priority filter by default sched: Only down migrate low priority tasks if allowed by affinity mask linaro/configs: Enable HMP priority filter by default ARM: TC2: Re-enable SD_SHARE_POWERLINE sched: Basic global balancing support for HMP Olivier Cozette (1): ARM: Change load tracking scale using sysfs Paul Turner (15): sched: Track the runnable average on a per-task entity basis sched: Aggregate load contributed by task entities on parenting cfs_rq sched: Maintain the load contribution of blocked entities sched: Add an rq migration call-back to sched_class sched: Account for blocked load waking back up sched: Aggregate total task_group load sched: Compute load contribution by a group entity sched: Normalize tg load contributions against runnable time sched: Maintain runnable averages across throttled periods sched: Replace update_shares weight distribution with per-entity computation sched: Refactor update_shares_cpu() -> update_blocked_avgs() sched: Update_cfs_shares at period edge sched: Make __update_entity_runnable_avg() fast sched: Introduce temporary FAIR_GROUP_SCHED dependency for load-tracking sched: implement usage tracking Peter Zijlstra (1): sched: Describe CFS load-balancer Sudeep KarkadaNagesha (9): ARM: perf: allocate CPU PMU dynamically at probe time ARM: perf: consistently use struct perf_event in arm_pmu functions ARM: perf: check ARMv7 counter validity on a per-pmu basis ARM: perf: replace global CPU PMU pointer with per-cpu pointers ARM: perf: register CPU PMUs with idr types ARM: perf: set cpu affinity to support multiple PMUs ARM: perf: set cpu affinity for the irqs correctly ARM: perf: remove spaces in CPU PMU names ARM: perf: save/restore pmu registers in pm notifier Thomas Gleixner (1): genirq: Add default affinity mask command line option Vincent Guittot (7): sched: add a new SD SHARE_POWERLINE flag for sched_domain sched: pack small tasks sched: secure access to other CPU statistics sched: pack the idle load balance ARM: sched: clear SD_SHARE_POWERLINE sched: pack small tasks: fix printk formating sched: pack small tasks: fix update packing domain Viresh Kumar (5): Revert "sched: Introduce temporary FAIR_GROUP_SCHED dependency for load-tracking" configs: Add config fragments for big LITTLE MP linaro/configs: Update big LITTLE MP fragment for task placement work config-frag/big-LITTLE: Use device-tree to provide fast/slow CPU list for HMP Merge branches 'arm-multi_pmu_v2', 'hw-bkp-v7.1-debug-v1', 'task-placement-v2-sysfs', 'misc-patches', 'config-fragments' and 'sched-pack-small-tasks-v1-arm' into big-LITTLE-MP-master-v13 Will Deacon (2): ARM: perf: return NOTIFY_DONE from cpu notifier when no available PMU ARM: perf: consistently use arm_pmu->name for PMU name Documentation/devicetree/bindings/arm/pmu.txt | 3 + Documentation/kernel-parameters.txt | 9 + arch/arm/Kconfig | 85 ++ arch/arm/include/asm/perf_event.h | 5 + arch/arm/include/asm/pmu.h | 40 +- arch/arm/include/asm/topology.h | 34 + arch/arm/kernel/hw_breakpoint.c | 57 + arch/arm/kernel/perf_event.c | 103 +- arch/arm/kernel/perf_event_cpu.c | 169 +- arch/arm/kernel/perf_event_v6.c | 130 +- arch/arm/kernel/perf_event_v7.c | 295 ++-- arch/arm/kernel/perf_event_xscale.c | 161 +- arch/arm/kernel/topology.c | 125 ++ arch/ia64/include/asm/topology.h | 1 + arch/tile/include/asm/topology.h | 1 + include/linux/sched.h | 29 + include/linux/topology.h | 3 + include/trace/events/sched.h | 153 ++ kernel/irq/irqdesc.c | 21 +- kernel/sched/core.c | 16 + kernel/sched/debug.c | 39 +- kernel/sched/fair.c | 2036 +++++++++++++++++++++++-- kernel/sched/sched.h | 65 +- linaro/configs/big-LITTLE-MP.conf | 13 + 24 files changed, 3037 insertions(+), 556 deletions(-) create mode 100644 linaro/configs/big-LITTLE-MP.conf

12 years

Fwd: IPv6 support [5405065:4109486]

by Dave Pigott

Adding Philip this time. :) Begin forwarded message: > From: "Zen - Technical Support" <support(a)zen.co.uk> > Subject: RE: IPv6 support [5405065:4109486] > Date: 10 December 2012 12:17:28 GMT > To: "Dave Pigott" <dave.pigott(a)linaro.org> > > Hello, > > The Zen core network has not yet had IPv6 turned on which would give you a native IPv6 connection as yet. Although our network could handle IPv6 Our admin systems are not yet fully compliant. > > You can however set up a 6in4 tunnel using a provider such as sixxs or hurricane electric which would give two-way IPv6 connectivity. > > > David Flint > Service Support Consultant > Zen Internet > T: 0845 058 9009 > F: 0845 058 9005 > W: http://www.zen.co.uk > > >> -----Original Message----- >> From: "Dave Pigott" <dave.pigott(a)linaro.org> >> Received: 10/12/2012 10:12 >> To: support(a)zen.co.uk >> Subject: RE: IPv6 support >> >> Hi, >> >> Just wondering if our leased line supports IPv6 or not? We would like to support both v4 and v6 if at all possible. >> >> Kind regards >> >> Dave Pigott > > ------------------ > Take a look inside and explore our new state of the art, award-winning Data Centre -http://www.zen.co.uk/business/default.aspx?page=11592 > > Please consider your environmental responsibility before printing this email. > > This message is private and confidential. If you have received this message in error please notify us and remove it from your system. > > Zen Internet Limited may monitor email traffic data to manage billing, to handle customer enquiries and for the prevention and detection of fraud. We may also monitor the content of emails sent to and/or from Zen Internet Limited for the purposes of security, staff training and to monitor quality of service. > > Zen Internet Limited is registered in England and Wales > Sandbrook Park, Sandbrook Way, Rochdale OL11 1RY > Company No. 03101568 > VAT Reg No. 686 0495 01

12 years

Fwd: IPv6 support [5405065:4109486]

by Dave Pigott

FYI Begin forwarded message: > From: "Zen - Technical Support" <support(a)zen.co.uk> > Subject: RE: IPv6 support [5405065:4109486] > Date: 10 December 2012 12:17:28 GMT > To: "Dave Pigott" <dave.pigott(a)linaro.org> > > Hello, > > The Zen core network has not yet had IPv6 turned on which would give you a native IPv6 connection as yet. Although our network could handle IPv6 Our admin systems are not yet fully compliant. > > You can however set up a 6in4 tunnel using a provider such as sixxs or hurricane electric which would give two-way IPv6 connectivity. > > > David Flint > Service Support Consultant > Zen Internet > T: 0845 058 9009 > F: 0845 058 9005 > W: http://www.zen.co.uk > > >> -----Original Message----- >> From: "Dave Pigott" <dave.pigott(a)linaro.org> >> Received: 10/12/2012 10:12 >> To: support(a)zen.co.uk >> Subject: RE: IPv6 support >> >> Hi, >> >> Just wondering if our leased line supports IPv6 or not? We would like to support both v4 and v6 if at all possible. >> >> Kind regards >> >> Dave Pigott > > ------------------ > Take a look inside and explore our new state of the art, award-winning Data Centre -http://www.zen.co.uk/business/default.aspx?page=11592 > > Please consider your environmental responsibility before printing this email. > > This message is private and confidential. If you have received this message in error please notify us and remove it from your system. > > Zen Internet Limited may monitor email traffic data to manage billing, to handle customer enquiries and for the prevention and detection of fraud. We may also monitor the content of emails sent to and/or from Zen Internet Limited for the purposes of security, staff training and to monitor quality of service. > > Zen Internet Limited is registered in England and Wales > Sandbrook Park, Sandbrook Way, Rochdale OL11 1RY > Company No. 03101568 > VAT Reg No. 686 0495 01

12 years

[RFC] cpufreq: Fix issues with hot-[un]plugging of cpus

by Viresh Kumar

Hi Rafael and other cpufreq masters, I am really not sure if this patch is a HACK or a valid solution :) Please have a look. Setup: ----- - ARM big LITTLE Platform with 5 cpus. cpus 2, 3 & 4 share clock and hence are part of policy->cpus. cpus 0 & 1 share clock but they aren't discussed here. - Interactive governor: I know its not mainlined and will never be. Problem should be same with all the governors we have. So, problem is not really governor dependent. I have added few prints at start/stop cases of governor, to see what policy->cpus is set to. Problem Statement: ----------------- There are many issues :( 1) I couldn't understand what we really want the value of policy->cpus be? Only the online cpus sharing clock? or all the possible cpus sharing clock? When we offline a cpu, we remove it from cpus field: cpumask_clear_cpu(cpu, data->cpus); Which shows that it is mask of online cpus only. But there are many places checking if a cpu is online or not (cpu in policy->cpus) in cpufreq.c. Which shows something else. 2) Kernel crashes when hot-unplugging cpus When we do: for i in 2 3 4; do echo 0 > /sys/devices/system/cpu/cpu$i/online; done We get: cpufreq_governor_interactive: CPUFREQ_GOV_STOP cpu 2 policy->cpus 2-4 cpufreq_vexpress: CPUFreq for CPU 3 initialized cpufreq_governor_interactive: CPUFREQ_GOV_START cpu 3 policy->cpus 2-4 CPU2: shutdown cpufreq_governor_interactive: CPUFREQ_GOV_STOP cpu 3 policy->cpus 3-4 cpufreq_vexpress: CPUFreq for CPU 4 initialized cpufreq_governor_interactive: CPUFREQ_GOV_START cpu 4 policy->cpus 2-4 And then crash from interactive governor. Note the following lines: ------------------------ cpufreq_governor_interactive: CPUFREQ_GOV_STOP cpu 2 policy->cpus 2-4 cpufreq_governor_interactive: CPUFREQ_GOV_START cpu 3 policy->cpus 2-4 cpufreq_governor_interactive: CPUFREQ_GOV_STOP cpu 3 policy->cpus 3-4 cpufreq_governor_interactive: CPUFREQ_GOV_START cpu 4 policy->cpus 2-4 So the governor is always started for CPUs 2 through 4 despite the fact that CPU 2 is removed. This makes the governor unstable due to wrong info passed. FIX: keep only online cpus in policy->cpus --- 3) With above issue solved, i tried to run the same stuff again and look what i got: cpufreq_governor_interactive: CPUFREQ_GOV_STOP cpu 2 policy->cpus 2-4 cpufreq_vexpress: CPUFreq for CPU 3 initialized cpufreq_governor_interactive: CPUFREQ_GOV_START cpu 3 policy->cpus 2-4 CPU2: shutdown a) Because cpu 2 is still not completely hot unplugged and is still present in cpu_online_mask. b) If we try to remove cpu 3, governor wouldn't be notified of it, because cpu 3 wasn't policy->cpu. FIX: Stop/Start governor on every non policy->cpu removal --- 4) Hot unplug is fine now, but not hotplug :( for i in 2 3 4; do echo 1 > /sys/devices/system/cpu/cpu$i/online; done CPU2: Booted secondary processor cpufreq_vexpress: CPUFreq for CPU 2 initialized cpufreq_governor_interactive: CPUFREQ_GOV_START cpu 2 policy->cpus 2 CPU3: Booted secondary processor cpufreq_vexpress: CPUFreq for CPU 3 initialized CPU4: Booted secondary processor cpufreq_vexpress: CPUFreq for CPU 4 initialized Hmmm, Governor is not informed about 3 and 4. :( FIX: Stop/Start governor on every managed_policy cpu, i.e. cpu for which --- policy struct exists Please share your comments, i know there would be many :) Signed-off-by: Viresh Kumar <viresh.kumar(a)linaro.org> --- drivers/cpufreq/cpufreq.c | 17 +++++++++++++++++ 1 file changed, 17 insertions(+) diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c index 0c3e29a..69bd739 100644 --- a/drivers/cpufreq/cpufreq.c +++ b/drivers/cpufreq/cpufreq.c @@ -749,11 +749,16 @@ static int cpufreq_add_dev_policy(unsigned int cpu, return -EBUSY; } + __cpufreq_governor(managed_policy, CPUFREQ_GOV_STOP); + spin_lock_irqsave(&cpufreq_driver_lock, flags); cpumask_copy(managed_policy->cpus, policy->cpus); per_cpu(cpufreq_cpu_data, cpu) = managed_policy; spin_unlock_irqrestore(&cpufreq_driver_lock, flags); + __cpufreq_governor(managed_policy, CPUFREQ_GOV_START); + __cpufreq_governor(managed_policy, CPUFREQ_GOV_LIMITS); + pr_debug("CPU already managed, adding link\n"); ret = sysfs_create_link(&dev->kobj, &managed_policy->kobj, @@ -968,6 +973,13 @@ static int cpufreq_add_dev(struct device *dev, struct subsys_interface *sif) pr_debug("initialization failed\n"); goto err_unlock_policy; } + + /* + * affected cpus must always be the one, which are online. We aren't + * managing offline cpus here. + */ + cpumask_and(policy->cpus, policy->cpus, cpu_online_mask); + policy->user_policy.min = policy->min; policy->user_policy.max = policy->max; @@ -1057,8 +1069,13 @@ static int __cpufreq_remove_dev(struct device *dev, struct subsys_interface *sif */ if (unlikely(cpu != data->cpu)) { pr_debug("removing link\n"); + __cpufreq_governor(data, CPUFREQ_GOV_STOP); cpumask_clear_cpu(cpu, data->cpus); spin_unlock_irqrestore(&cpufreq_driver_lock, flags); + + __cpufreq_governor(data, CPUFREQ_GOV_START); + __cpufreq_governor(data, CPUFREQ_GOV_LIMITS); + kobj = &dev->kobj; cpufreq_cpu_put(data); unlock_policy_rwsem_write(cpu); -- 1.7.12.rc2.18.g61b472e

12 years

[HMP][PATCH 0/1] Global balance

by Morten Rasmussen

Hi Viresh, Here is a patch that introduces global load balancing on top of the existing HMP patch set. It depends on the HMP patches already present in your task-placement-v2 branch. It can be applied on top of the HMP sysfs patches if needed. The fix should be trivial. Could you include in the MP branch for the 12.12 release? Testing with sysbench and coremark show significant performance improvements for parallel workloads as all cpus can now be used for cpu intensive tasks. Thanks, Morten Morten Rasmussen (1): sched: Basic global balancing support for HMP kernel/sched/fair.c | 101 +++++++++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 97 insertions(+), 4 deletions(-) -- 1.7.9.5

12 years

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

linaro-dev December 2012