linaro-dev May 2013

linaro-dev@lists.linaro.org

70 participants
66 discussions

Re: [PATCH 3/3] sched: cfs.nr_running does not contain the intended metric

by Viresh Kumar

Fixing linaro dev address On 16 May 2013 22:18, Chris Redpath <chris.redpath(a)arm.com> wrote: > rq->nr_running is the actual number of runnable tasks we wish to use > to determine if a task is alone on a CPU. > > Change-Id: Icaf3022e02924ecdc94e14d4146c6fadd9580e2b > Signed-off-by: Chris Redpath <chris.redpath(a)arm.com> > --- > kernel/sched/fair.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c > index 3a53343..861c98d 100644 > --- a/kernel/sched/fair.c > +++ b/kernel/sched/fair.c > @@ -3871,7 +3871,7 @@ static inline unsigned int hmp_offload_down(int cpu, struct sched_entity *se) > return NR_CPUS; > > /* Is the task alone on the cpu? */ > - if (cpu_rq(cpu)->cfs.nr_running < 2) > + if (cpu_rq(cpu)->nr_running < 2) > return NR_CPUS; > > /* Is the task actually starving? */ > -- > 1.7.9.5 > >

11 years, 7 months

Re: [PATCH 2/3] sched: Ignore offline CPUs in HMP migration & load stats

by Viresh Kumar

Fixing linaro dev address On 16 May 2013 22:18, Chris Redpath <chris.redpath(a)arm.com> wrote: > Previously, an offline CPU would always appear to have a zero load > and this would distort the offload functionality used for balancing > big and little domains. > > Maintain a mask of online CPUs in each domain and use this instead. > > Change-Id: I639b564b2f40cb659af8ceb8bd37f84b8a1fe323 > Signed-off-by: Chris Redpath <chris.redpath(a)arm.com> > --- > arch/arm/kernel/topology.c | 6 ++++-- > include/linux/sched.h | 1 + > kernel/sched/fair.c | 39 +++++++++++++++++++++++++++++++++++++-- > 3 files changed, 42 insertions(+), 4 deletions(-) > > diff --git a/arch/arm/kernel/topology.c b/arch/arm/kernel/topology.c > index fa45fb4..93ad2cd 100644 > --- a/arch/arm/kernel/topology.c > +++ b/arch/arm/kernel/topology.c > @@ -381,12 +381,14 @@ void __init arch_get_hmp_domains(struct list_head *hmp_domains_list) > if(!cpumask_empty(&hmp_slow_cpu_mask)) { > domain = (struct hmp_domain *) > kmalloc(sizeof(struct hmp_domain), GFP_KERNEL); > - cpumask_copy(&domain->cpus, &hmp_slow_cpu_mask); > + cpumask_copy(&domain->possible_cpus, &hmp_slow_cpu_mask); > + cpumask_and(&domain->cpus, cpu_online_mask, &domain->possible_cpus); > list_add(&domain->hmp_domains, hmp_domains_list); > } > domain = (struct hmp_domain *) > kmalloc(sizeof(struct hmp_domain), GFP_KERNEL); > - cpumask_copy(&domain->cpus, &hmp_fast_cpu_mask); > + cpumask_copy(&domain->possible_cpus, &hmp_fast_cpu_mask); > + cpumask_and(&domain->cpus, cpu_online_mask, &domain->possible_cpus); > list_add(&domain->hmp_domains, hmp_domains_list); > } > #endif /* CONFIG_SCHED_HMP */ > diff --git a/include/linux/sched.h b/include/linux/sched.h > index ea439ab..0553d76 100644 > --- a/include/linux/sched.h > +++ b/include/linux/sched.h > @@ -988,6 +988,7 @@ bool cpus_share_cache(int this_cpu, int that_cpu); > #ifdef CONFIG_SCHED_HMP > struct hmp_domain { > struct cpumask cpus; > + struct cpumask possible_cpus; > struct list_head hmp_domains; > }; > #endif /* CONFIG_SCHED_HMP */ > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c > index 0f9941e..3a53343 100644 > --- a/kernel/sched/fair.c > +++ b/kernel/sched/fair.c > @@ -3514,10 +3514,10 @@ static int __init hmp_cpu_mask_setup(void) > dc = 0; > list_for_each(pos, &hmp_domains) { > domain = list_entry(pos, struct hmp_domain, hmp_domains); > - cpulist_scnprintf(buf, 64, &domain->cpus); > + cpulist_scnprintf(buf, 64, &domain->possible_cpus); > pr_debug(" HMP domain %d: %s\n", dc, buf); > > - for_each_cpu_mask(cpu, domain->cpus) { > + for_each_cpu_mask(cpu, domain->possible_cpus) { > per_cpu(hmp_cpu_domain, cpu) = domain; > } > dc++; > @@ -3526,6 +3526,35 @@ static int __init hmp_cpu_mask_setup(void) > return 1; > } > > +static struct hmp_domain *hmp_get_hmp_domain_for_cpu(int cpu) > +{ > + struct hmp_domain *domain; > + struct list_head *pos; > + > + list_for_each(pos, &hmp_domains) { > + domain = list_entry(pos, struct hmp_domain, hmp_domains); > + if(cpumask_test_cpu(cpu, &domain->possible_cpus)) > + return domain; > + } > + return NULL; > +} > + > +static void hmp_online_cpu(int cpu) > +{ > + struct hmp_domain *domain = hmp_get_hmp_domain_for_cpu(cpu); > + > + if(domain) > + cpumask_set_cpu(cpu, &domain->cpus); > +} > + > +static void hmp_offline_cpu(int cpu) > +{ > + struct hmp_domain *domain = hmp_get_hmp_domain_for_cpu(cpu); > + > + if(domain) > + cpumask_clear_cpu(cpu, &domain->cpus); > +} > + > /* > * Migration thresholds should be in the range [0..1023] > * hmp_up_threshold: min. load required for migrating tasks to a faster cpu > @@ -6570,11 +6599,17 @@ void trigger_load_balance(struct rq *rq, int cpu) > > static void rq_online_fair(struct rq *rq) > { > +#ifdef CONFIG_SCHED_HMP > + hmp_online_cpu(rq->cpu); > +#endif > update_sysctl(); > } > > static void rq_offline_fair(struct rq *rq) > { > +#ifdef CONFIG_SCHED_HMP > + hmp_offline_cpu(rq->cpu); > +#endif > update_sysctl(); > > /* Ensure any throttled groups are reachable by pick_next_task */ > -- > 1.7.9.5 > >

11 years, 7 months

Re: [PATCH 1/3] sched: Do not ignore grouped tasks during HMP forced migration.

by Viresh Kumar

Fixing Linaro dev addresss. On 16 May 2013 22:18, Chris Redpath <chris.redpath(a)arm.com> wrote: > If the entity is not a task, it is a cfs group rq. Iterate up to > find the task entity. > > Change-Id: I7cab7aba0798f6f14e38ad32e566d90e5937ffbc > Signed-off-by: Chris Redpath <chris.redpath(a)arm.com> > --- > kernel/sched/fair.c | 11 ++++++++++- > 1 file changed, 10 insertions(+), 1 deletion(-) > > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c > index 5ecfb1b..0f9941e 100644 > --- a/kernel/sched/fair.c > +++ b/kernel/sched/fair.c > @@ -6474,10 +6474,19 @@ static void hmp_force_up_migration(int this_cpu) > target = cpu_rq(cpu); > raw_spin_lock_irqsave(&target->lock, flags); > curr = target->cfs.curr; > - if (!curr || !entity_is_task(curr)) { > + if (!curr) { > raw_spin_unlock_irqrestore(&target->lock, flags); > continue; > } > + if (!entity_is_task(curr)) { > + struct cfs_rq *cfs_rq; > + > + cfs_rq = group_cfs_rq(curr); > + while (cfs_rq) { > + curr = cfs_rq->curr; > + cfs_rq = group_cfs_rq(curr); > + } > + } > p = task_of(curr); > if (hmp_up_migration(cpu, curr)) { > if (!target->active_balance) { > -- > 1.7.9.5 > >

11 years, 7 months

Release Critical Test Result Summary of Calendar Week 20, 2013 for Linux Linaro ubuntu Raring.

by Botao Sun

Calendar Week 20, 2013: Here is release critical test result summary for Linux Linaro ubuntu Raring image on following boards: 1) ARM Versatile Express A9; 2) Samsung Arndale; 3) TI Panda 4430; 4) TI Panda 4460; 5) ST Ericsson Snowball. Synopsis: All release critical test passed on all boards. 1. ARM Versatile Express A9 + Linux Linaro Raring (Column V): https://docs.google.com/a/linaro.org/spreadsheet/ccc?key=0AroPySpr4FnEdFNmV… 2. Samsung Arndale + Linux Linaro Raring (Column P): https://docs.google.com/a/linaro.org/spreadsheet/ccc?key=0AgB-fT5LL31CdGZJS… 3. TI Panda 4430 + Linux Linaro Raring (Column V): https://docs.google.com/a/linaro.org/spreadsheet/ccc?key=0AroPySpr4FnEdEwwZ… 4. TI Panda 4460 + Linux Linaro Raring (Column V): https://docs.google.com/a/linaro.org/spreadsheet/ccc?key=0AroPySpr4FnEdEwwZ… 5. ST Ericsson Snowball + Linux Linaro Raring (Column W): https://docs.google.com/a/linaro.org/spreadsheet/ccc?key=0AroPySpr4FnEdFJ4X… For the test summary of last week (Calendar Week 19), please refer to attachment. Thank you. Best Regards Botao Sun

11 years, 7 months

[PATCH] configs: linaro-base: Enable CRYPTO_[AES|SHA1]_ARM

by Ard Biesheuvel

These assembler implementations of SHA1 and AES have been in the upstream source tree since September 2012 but need to be selected explicitly in order to be enabled. Signed-off-by: Ard Biesheuvel <ard.biesheuvel(a)linaro.org> --- linaro/configs/linaro-base.conf | 2 ++ 1 file changed, 2 insertions(+) diff --git a/linaro/configs/linaro-base.conf b/linaro/configs/linaro-base.conf index 5c748a7..ba7b97b 100644 --- a/linaro/configs/linaro-base.conf +++ b/linaro/configs/linaro-base.conf @@ -86,3 +86,5 @@ CONFIG_HW_PERF_EVENTS=y CONFIG_FUNCTION_TRACER=y CONFIG_ENABLE_DEFAULT_TRACERS=y CONFIG_PROC_DEVICETREE=y +CONFIG_CRYPTO_AES_ARM=y +CONFIG_CRYPTO_SHA1_ARM=y -- 1.8.1.2

11 years, 7 months

Do you use (review.)android.git.linaro.org?

by Philip Colmer

This email is going to a rather wide audience so my apologies if it isn't relevant to you. If you don't use android.git.linaro.org or review.android.git.linaro.org, you can ignore the rest of this email. If you use either of these servers, please note that there will be planned maintenance between 2pm and 5pm, GMT+1, this Friday, 17th May. We are moving the server from Canonical to Amazon EC2. If you currently use review.android.git.linaro.org and sign in on it, please read the following information VERY CAREFULLY! At the moment, users are authenticated via OpenID with Launchpad. We are changing that to use OpenID with the Linaro Login service. If you don't know if you have a Linaro Login account, or you want to check that it is going to work properly: * Go to https://login.linaro.org:8443/openidserver. * If you are asked to log in, please enter your EMAIL ADDRESS and your Linaro Login password. If you don't know what your Linaro Login password is, or you cannot successfully log in, please contact me as soon as possible! * The web page should now display a URL of the format https://login.linaro.org:8443/openidserver/users/<your email address>. This is your OpenID URL and means you will be ready to follow the link identity instructions below. After the maintenance has been completed, you will need to take the following steps ONCE: 1. Go to https://login.linaro.org:8443/openidserver. If you are asked to log in, please enter your EMAIL ADDRESS and your Linaro Login password. 2. The web page should now display a URL of the format https://login.linaro.org:8443/openidserver/users/<your email address>. Make a note of that, please. 3. Go to http://review.android.git.linaro.org/. Sign in with your Launchpad ID. 4. When you have signed in, click on the Settings link in the top right corner of the web page. 5. Click on the Identities link on the left hand side of the web page. 6. Click on Link Another Identity. 7. Enter the URL you noted from step 2 and click on Link Identity. 8. If your web browser session from step 2 has expired, you will be asked to log in again - enter the same information you entered at step 1. 9. If your web browser session has not expired, or once you have logged in, you will see a web page with the title "OpenID Verification". On the right hand side, click on Allow Always. You should now be back at the Settings page. If you want to check that you can now sign in with your Linaro Login credentials, click on Sign Out in the corner of the Gerrit window, click Sign In and this time use the URL from step 2 rather than your Launchpad URL. You should be logged straight in if you have not closed any web browser windows. PLEASE NOTE: we will be supporting both Launchpad and Linaro Login OpenID authentication until the *end of May*. After that time, we will be switching over to just Linaro Login and you will no longer have to enter the OpenID URL - Gerrit will authenticate you directly with Linaro Login. Please do NOT attempt the above instructions until after the maintenance has been completed. I am providing the information now so that (a) you have it ready for use and (b) I can help anyone out if they do not know their Linaro Login details. If you have any questions, please email me. Thank you. Regards Philip Colmer IT Services Manager Linaro

11 years, 7 months

[PATCH: cpuidle: arm_big_little] pass target residency to mcpm

by Sebastian Capella

Hi, This patch will forward target residency information from the arm_big_little driver to mcpm. If multiple powerdown states are used, the vendor specific code will need a way to distinguish the intended c-state information. I do not have TC2 hardware to verify this. Would someone be able to help verify this change on TC2? Thanks! Sebastian Sebastian Capella (1): cpuidle: arm_big_little: route target residency to mcpm drivers/cpuidle/arm_big_little.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) -- 1.7.9.5

11 years, 7 months

Re: [OE-core] Changes in Linaro layers for OpenEmbedded

by Paul Sherwood

On 10/05/2013 13:00, linaro-dev-request(a)lists.linaro.org wrote: > At Linaro we have many teams and some of them are using OpenEmbedded or > will do it soon. One of reasons is lack of ARM big-endian distributions > on a market. I don't normally comment here, but I'm confused by the above statement... why does lack of big-endian distributions lead Linaro to use OpenEmbedded? fwiw we believe that Baserock is currently the only commercial ARM big-endian solution, but arguably baserock is not exactly a "distribution". br Paul -- Paul Sherwood Codethink Ltd. Tel: +44 788 798 4900 302 Ducie House, Ducie Street, http://www.codethink.co.uk/ Manchester, M1 2JW, United Kingdom. Codethink provides advanced software design, development, integration & test services: from embedded systems to high performance apps to cloud.

11 years, 7 months

big.LITTLE MP status May 7, 2013

by Serge Broslavsky

https://wiki.linaro.org/projects/big.LITTLE.MP Work Items for functional completion: https://wiki.linaro.org/projects/big.LITTLE.MP/WorkItems Roadmap Cards: Epic big.LITTLE MP card - http://cards.linaro.org/browse/CARD-190 Main Project page: https://launchpad.net/linaro-big-little-system Development ============ 1. big.LITTLE IKS Current IKS snapshot: git://git.linaro.org/landing-teams/working/arm/kernel.git More information: http://lists.linaro.org/pipermail/linaro-dev/2013-May/015882.html 2. big.LITTLE MP There are two branches, which are functionally equivalent - they tweak the scheduler to select the right processes to move to the A15 and try to decrease power consumption while maintaining performance. * A Reference (stable) big.LITTLE MP patchset (merged with linux-linaro): https://git.linaro.org/gitweb?p=arm/big.LITTLE/mp.git;a=shortlog;h=refs/hea… Recommended to be used for testing. * And an upstream (work in progress) big.LITTLE MP patchset: https://git.linaro.org/gitweb?p=arm/big.LITTLE/mp.git;a=shortlog;h=refs/hea… Being worked on in order to achieve the performance/power numbers of the reference one. * Documentation: Code flow is already covered by several LWN articles. See https://wiki.linaro.org/WorkingGroups/PowerManagement/Doc/ in the Articles section. We will provide a tuning guide once we have reached agreement upstream. * Automated benchmarking code is ready: - Workload results automated upload is being worked on. Blueprint Updates ------------------ * https://blueprints.launchpad.net/linaro-big-little-system/+spec/big-little-… Should be much simpler now. * https://blueprints.launchpad.net/linaro-big-little-system/+spec/cpuidle-add… Done. * https://blueprints.launchpad.net/linaro-big-little-system/+spec/small-task-… Discussion is ongoing in ML. Testing on a 64 core x86 system is ongoing, results to be available in the next report. * https://blueprints.launchpad.net/linaro-big-little-system/+spec/cpu-hot-plu… Unlocked last week. Discussions around data extraction are ongoing. * https://blueprints.launchpad.net/linaro-big-little-system/+spec/sched-coope… Long-lasting BP (6+ month to get something), no updates. QA == big.LITTLE MP LAVA Dashboard http://validation.linaro.org/lava-server/dashboard/image-reports/linaro-and… Testing (week 19): sched_tests.git #Cases #Run Pass Fail Absolute pass rate Regression 20 19 19 0 100 % mpbasicsuite 14 10 10 0 100 % mpcoresuite 10 6 6 0 100 % mpextendedsuite 20 10 7 3 70 % mploadbalance 2 2 2 0 100 % http://lists.linaro.org/pipermail/linaro-dev/2013-May/015923.html Bugs: https://bugs.launchpad.net/linaro-big-little-system/+bugs?field.tag=bl-mp Highlight: Bug 1166246 - Lockup on post 3.9 IKS kernel - isolated the problem (both A15 CPUs suspended by cpuidle, upon returning the status is not hitting the RAM). - cpuidle is a prerequisite for this issue to happen. - lockup is happening 800+ GHz, lower frequencies are being tested. - Dave Martin has a script that triggers the issue easily. - analysis is ongoing. -- Best Regards, Serge Broslavsky <serge.broslavsky(a)linaro.org> Project Manager, Linaro M: +37129426328 IRC: ototo Skype: serge.broslavsky http://linaro.org | Open source software for ARM SoCs

11 years, 7 months

Linaro Virtualization status 8 May 2013 (Week 19)

by Serge Broslavsky

Linaro Virtualization Team's Weekly Report - Week 19 2013 ========================================================= During the last week (2 - 8 May 2013) the team has been working on the following areas. More detailed information can be found on the meeting reports Wiki page: [https://wiki.linaro.org/LEG/Engineering/Virtualization/TeamMeetingReports]. Ongoing work ------------ * BOOTLOADER HYP MODE SUPPORT FOR ARMv7 [http://cards.linaro.org/browse/CARD-361] by Andre Przywara * rebase existing Versatile Express patches against upstream u-boot [http://cards.linaro.org/browse/VIRT-70] Done: - final version of the patch set - sent to the ML Working on: - testing an polishing * ARMv7 KVM GUEST MIGRATION [http://cards.linaro.org/browse/CARD-363] by Peter Maydell and Andre Przywara * QEMU: implement full cp15 sync with kernel [http://cards.linaro.org/browse/VIRT-49] by Peter Maydell Done: - can run KVM with this code now Working on: - GDB is behaving oddly connected to QEMU's stub, which is suspected to be a GDB bug (LP:1175525) but may also be a QEMU side issue - test migration proper (requires kernel to support arch timer state save/load) - clean up some TODOs and other cruft - debug & patch cleanup. * kernel: agree on a data structure to describe timer state [http://cards.linaro.org/browse/VIRT-42] * kernel: implement save and restore of timer state in kernel [http://cards.linaro.org/browse/VIRT-43] by Andre Przywara Done: - first version implemented and reviewed on the ML. - created few unit tests (to be added to LAVA). Working on: - simplifying the code according to the feedback received. * kernel: agree on a data structure to hold GIC state [http://cards.linaro.org/browse/VIRT-44] by Andre Przywara Working on: - VGIC investigation. * ARMv8 KVM QEMU prototype [http://cards.linaro.org/browse/CARD-221] by john Rigby * Add arm64 mach-virt system model skeleton [http://cards.linaro.org/browse/VIRT-53] * Add kvm guest creation and launching of armv8/mach-virt [http://cards.linaro.org/browse/VIRT-56] Done: - submittied a new rev of armv7 mach-virt and aarch64 patch sets to ML Working on: - finishing the implementation work * KVM ARMV8 SMP HARDWARE BRING-UP [http://cards.linaro.org/browse/CARD-555] by Anup Patel and Pranavkumar Sawargaonkar Done: - Found cleaner fixes instead of hacky cache flushes for KVM ARMv8: Issue was with HCR_GUEST_FLAGS: in arm64 there are two new bits ID[33] and CD[32]; ID is for Stage 2 Instruction cache disable. Unfortunately these bits were getting set by small macro defination mistake thus making VM's stage 2 cache disable. - Done with arch_timer changes required for KVM ARMv8 - Sent out a new patch for emergency (early) print support for virtio console. Working on: - Submit patch to fix Stage2 MemAttr attribute (HCR) programmed by KVM ARMv8 - Try ARM32 Guest (UP) on v8 Foundation Model - Try ARM32 Guest (UP) on APM Board (UP) * XEN - MULTI-PLATFORM ARMV7 BINARIES [http://cards.linaro.org/browse/CARD-560] by Jullien Grall Done: - Reworked multi ARM platform patch series - Sent minor fixes for Xen on ARM Working on: - Sending a new version of the patch series - Still need to update wiki page for the arndale (both Xen and U-boot) - VFP save/restore in Xen * CI BRING UP: KVM ARMv7 [http://cards.linaro.org/browse/CARD-233] by Riku Voipio Done: - Real tests to be added in the scope of a new card (CARD-599). Working on: - Card close out. -- Best Regards, Serge Broslavsky <serge.broslavsky(a)linaro.org> Project Manager, Linaro M: +37129426328 IRC: ototo Skype: serge.broslavsky http://linaro.org | Open source software for ARM SoCs

11 years, 7 months

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

linaro-dev May 2013