linaro-dev February 2011

linaro-dev@lists.linaro.org

120 participants
166 discussions

[PM] 02/02/2011 - Minutes for the Power Management WG weekly call

by Amit Kucheria

Hello Everyone, The minutes of the weekly call can be found at: https://wiki.linaro.org/WorkingGroups/PowerManagement/Meetings/2011-02-02 Regards, Amit

14 years, 10 months

Re: [PATCH] arm: Improve MMC performance on Versatile Express

by Russell King - ARM Linux

On Mon, Jan 24, 2011 at 12:27:16PM -0000, Pawel Moll wrote: > So - we'll try to enlarge FIFO. For the moment - playing with interrupts > affinity seem to be a viable workaround. I don't think enlarging the FIFO will help too much. The issue is whether the CPU can keep up with the data rate coming off the card. If it can't, then no matter how large the FIFO is, it will eventually overflow. The real answer is to avoid PIO mode, and use DMA support. However, I've had problems using DMA on the ARM development boards. You can find details my DMA issues internally within ARM by talking to Catalin. The alternative answer, I believe implemented by some of ARMs silicon partners, is to turn the card clock off when the FIFO becomes full/empty to stop it sending more data. I think this violates some of the MMC/SD requirements, but it seems to work for the silicon partners just fine.

14 years, 10 months

Fwd: [PATCH, v3 2/2] cgroups: introduce timer slack subsystem

by amit Kucheria

Seems like a start to the idea we discussed in Dallas about using cgroups to group together tasks with similar wakeup characteristics. Thanks Daniel. /Amit ----- Forwarded message from Daniel Lezcano <daniel.lezcano(a)free.fr> ----- Date: Wed, 02 Feb 2011 22:01:48 +0100 From: Daniel Lezcano <daniel.lezcano(a)free.fr> To: amit Kucheria <amit.kucheria(a)linaro.org> Subject: Fwd: [PATCH, v3 2/2] cgroups: introduce timer slack subsystem User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.2.13) Gecko/20101208 Thunderbird/3.1.7 Hi Amit, I supposed this patch may interest you ;) Cheers -- Daniel -------- Original Message -------- Subject: [PATCH, v3 2/2] cgroups: introduce timer slack subsystem Date: Wed, 2 Feb 2011 22:47:36 +0200 From: Kirill A. Shutsemov <kirill(a)shutemov.name> To: Paul Menage <menage(a)google.com>, Li Zefan <lizf(a)cn.fujitsu.com> CC: jacob.jun.pan(a)linux.intel.com, containers(a)lists.linux-foundation.org, linux-kernel(a)vger.kernel.org, Arjan van de Ven <arjan(a)linux.intel.com> From: Kirill A. Shutemov<kirill(a)shutemov.name> Provides a way of tasks grouping by timer slack value. Introduces per cgroup max and min timer slack value. When a task attaches to a cgroup, its timer slack value adjusts (if needed) to fit min-max range. It also provides a way to set timer slack value for all tasks in the cgroup at once. This functionality is useful in mobile devices where certain background apps are attached to a cgroup and minimum wakeups are desired. Signed-off-by: Kirill A. Shutemov<kirill(a)shutemov.name> Idea-by: Jacob Pan<jacob.jun.pan(a)linux.intel.com> --- include/linux/cgroup_subsys.h | 6 + include/linux/init_task.h | 4 +- init/Kconfig | 10 ++ kernel/Makefile | 1 + kernel/cgroup_timer_slack.c | 242 +++++++++++++++++++++++++++++++++++++++++ kernel/sys.c | 24 +++- 6 files changed, 280 insertions(+), 7 deletions(-) create mode 100644 kernel/cgroup_timer_slack.c diff --git a/include/linux/cgroup_subsys.h b/include/linux/cgroup_subsys.h index ccefff0..e399228 100644 --- a/include/linux/cgroup_subsys.h +++ b/include/linux/cgroup_subsys.h @@ -66,3 +66,9 @@ SUBSYS(blkio) #endif /* */ + +#ifdef CONFIG_CGROUP_TIMER_SLACK +SUBSYS(timer_slack) +#endif + +/* */ diff --git a/include/linux/init_task.h b/include/linux/init_task.h index caa151f..48eca8f 100644 --- a/include/linux/init_task.h +++ b/include/linux/init_task.h @@ -124,6 +124,8 @@ extern struct cred init_cred; # define INIT_PERF_EVENTS(tsk) #endif +#define TIMER_SLACK_NS_DEFAULT 50000 + /* * INIT_TASK is used to set up the first task table, touch at * your own risk!. Base=0, limit=0x1fffff (=2MB) @@ -177,7 +179,7 @@ extern struct cred init_cred; .cpu_timers = INIT_CPU_TIMERS(tsk.cpu_timers), \ .fs_excl = ATOMIC_INIT(0), \ .pi_lock = __RAW_SPIN_LOCK_UNLOCKED(tsk.pi_lock), \ - .timer_slack_ns = 50000, /* 50 usec default slack */ \ + .timer_slack_ns = TIMER_SLACK_NS_DEFAULT, \ .pids = { \ [PIDTYPE_PID] = INIT_PID_LINK(PIDTYPE_PID), \ [PIDTYPE_PGID] = INIT_PID_LINK(PIDTYPE_PGID), \ diff --git a/init/Kconfig b/init/Kconfig index be788c0..f21b4ce 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -596,6 +596,16 @@ config CGROUP_FREEZER Provides a way to freeze and unfreeze all tasks in a cgroup. +config CGROUP_TIMER_SLACK + tristate "Timer slack cgroup subsystem" + help + Provides a way of tasks grouping by timer slack value. + Introduces per cgroup timer slack value which will override + the default timer slack value once a task is attached to a + cgroup. + It's useful in mobile devices where certain background apps + are attached to a cgroup and minimum wakeups are desired. + config CGROUP_DEVICE bool "Device controller for cgroups" help diff --git a/kernel/Makefile b/kernel/Makefile index 353d3fe..0b60239 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -61,6 +61,7 @@ obj-$(CONFIG_BACKTRACE_SELF_TEST) += backtracetest.o obj-$(CONFIG_COMPAT) += compat.o obj-$(CONFIG_CGROUPS) += cgroup.o obj-$(CONFIG_CGROUP_FREEZER) += cgroup_freezer.o +obj-$(CONFIG_CGROUP_TIMER_SLACK) += cgroup_timer_slack.o obj-$(CONFIG_CPUSETS) += cpuset.o obj-$(CONFIG_CGROUP_NS) += ns_cgroup.o obj-$(CONFIG_UTS_NS) += utsname.o diff --git a/kernel/cgroup_timer_slack.c b/kernel/cgroup_timer_slack.c new file mode 100644 index 0000000..a343a50 --- /dev/null +++ b/kernel/cgroup_timer_slack.c @@ -0,0 +1,242 @@ +/* + * cgroup_timer_slack.c - control group timer slack subsystem + * + * Copyright Nokia Corparation, 2011 + * Author: Kirill A. Shutemov + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + */ +#include<linux/cgroup.h> +#include<linux/init_task.h> +#include<linux/module.h> +#include<linux/slab.h> + +struct cgroup_subsys timer_slack_subsys; +struct timer_slack_cgroup { + struct cgroup_subsys_state css; + unsigned long min_slack_ns; + unsigned long max_slack_ns; +}; + +enum { + TIMER_SLACK_MIN, + TIMER_SLACK_MAX, +}; + +int dummy_timer_slack_check(struct task_struct *task, unsigned long slack_ns); +extern int (*timer_slack_check)(struct task_struct *task, + unsigned long slack_ns); + +static struct timer_slack_cgroup *cgroup_to_tslack_cgroup(struct cgroup *cgroup) +{ + struct cgroup_subsys_state *css; + + css = cgroup_subsys_state(cgroup, timer_slack_subsys.subsys_id); + return container_of(css, struct timer_slack_cgroup, css); +} + +static int cgroup_timer_slack_check(struct task_struct *task, + unsigned long slack_ns) +{ + struct cgroup_subsys_state *css; + struct timer_slack_cgroup *tslack_cgroup; + + css = task_subsys_state(task, timer_slack_subsys.subsys_id); + tslack_cgroup = container_of(css, struct timer_slack_cgroup, css); + + if (slack_ns< tslack_cgroup->min_slack_ns) + return -EPERM; + if (slack_ns> tslack_cgroup->max_slack_ns) + return -EPERM; + return 0; +} + +static struct cgroup_subsys_state * +tslack_cgroup_create(struct cgroup_subsys *subsys, struct cgroup *cgroup) +{ + struct timer_slack_cgroup *tslack_cgroup; + + tslack_cgroup = kmalloc(sizeof(*tslack_cgroup), GFP_KERNEL); + if (!tslack_cgroup) + return ERR_PTR(-ENOMEM); + + if (cgroup->parent) { + struct timer_slack_cgroup *parent; + parent = cgroup_to_tslack_cgroup(cgroup->parent); + tslack_cgroup->min_slack_ns = parent->min_slack_ns; + tslack_cgroup->max_slack_ns = parent->max_slack_ns; + } else { + tslack_cgroup->min_slack_ns = 0UL; + tslack_cgroup->max_slack_ns = ULONG_MAX; + } + + return&tslack_cgroup->css; +} + +static void tslack_cgroup_destroy(struct cgroup_subsys *subsys, + struct cgroup *cgroup) +{ + kfree(cgroup_to_tslack_cgroup(cgroup)); +} + +/* + * Adjust ->timer_slack_ns and ->default_max_slack_ns of the task to fit + * limits of the cgroup. + */ +static void tslack_adjust_task(struct timer_slack_cgroup *tslack_cgroup, + struct task_struct *tsk) +{ + if (tslack_cgroup->min_slack_ns> tsk->timer_slack_ns) + tsk->timer_slack_ns = tslack_cgroup->min_slack_ns; + else if (tslack_cgroup->max_slack_ns< tsk->timer_slack_ns) + tsk->timer_slack_ns = tslack_cgroup->max_slack_ns; + + if (tslack_cgroup->min_slack_ns> tsk->default_timer_slack_ns) + tsk->default_timer_slack_ns = tslack_cgroup->min_slack_ns; + else if (tslack_cgroup->max_slack_ns< tsk->default_timer_slack_ns) + tsk->default_timer_slack_ns = tslack_cgroup->max_slack_ns; +} + +static void tslack_cgroup_attach(struct cgroup_subsys *subsys, + struct cgroup *cgroup, struct cgroup *prev, + struct task_struct *tsk, bool threadgroup) +{ + tslack_adjust_task(cgroup_to_tslack_cgroup(cgroup), tsk); +} + +static int tslack_write_set_slack_ns(struct cgroup *cgroup, struct cftype *cft, + u64 val) +{ + struct timer_slack_cgroup *tslack_cgroup; + struct cgroup_iter it; + struct task_struct *task; + + tslack_cgroup = cgroup_to_tslack_cgroup(cgroup); + if (!val || val< tslack_cgroup->min_slack_ns || + val> tslack_cgroup->max_slack_ns ) + return -EINVAL; + + /* Change timer slack value for all tasks in the cgroup */ + cgroup_iter_start(cgroup,&it); + while ((task = cgroup_iter_next(cgroup,&it))) + task->timer_slack_ns = val; + cgroup_iter_end(cgroup,&it); + + return 0; +} + +static u64 tslack_read_range(struct cgroup *cgroup, struct cftype *cft) +{ + struct timer_slack_cgroup *tslack_cgroup; + + tslack_cgroup = cgroup_to_tslack_cgroup(cgroup); + switch (cft->private) { + case TIMER_SLACK_MIN: + return tslack_cgroup->min_slack_ns; + case TIMER_SLACK_MAX: + return tslack_cgroup->max_slack_ns; + default: + BUG(); + }; +} + +static int tslack_write_range(struct cgroup *cgroup, struct cftype *cft, + u64 val) +{ + struct timer_slack_cgroup *tslack_cgroup; + struct cgroup_iter it; + struct task_struct *task; + + if (!val) + return -EINVAL; + + tslack_cgroup = cgroup_to_tslack_cgroup(cgroup); + switch (cft->private) { + case TIMER_SLACK_MIN: + if (val> tslack_cgroup->max_slack_ns) + return -EINVAL; + tslack_cgroup->min_slack_ns = val; + break; + case TIMER_SLACK_MAX: + if (val< tslack_cgroup->min_slack_ns) + return -EINVAL; + tslack_cgroup->max_slack_ns = val; + break; + default: + BUG(); + } + + /* + * Adjust timer slack value for all tasks in the cgroup to fit + * min-max range. + */ + cgroup_iter_start(cgroup,&it); + while ((task = cgroup_iter_next(cgroup,&it))) + tslack_adjust_task(tslack_cgroup, task); + cgroup_iter_end(cgroup,&it); + + return 0; +} + +static struct cftype files[] = { + { + .name = "set_slack_ns", + .write_u64 = tslack_write_set_slack_ns, + }, + { + .name = "min_slack_ns", + .private = TIMER_SLACK_MIN, + .read_u64 = tslack_read_range, + .write_u64 = tslack_write_range, + }, + { + .name = "max_slack_ns", + .private = TIMER_SLACK_MAX, + .read_u64 = tslack_read_range, + .write_u64 = tslack_write_range, + }, +}; + +static int tslack_cgroup_populate(struct cgroup_subsys *subsys, + struct cgroup *cgroup) +{ + return cgroup_add_files(cgroup, subsys, files, ARRAY_SIZE(files)); +} + +struct cgroup_subsys timer_slack_subsys = { + .name = "timer_slack", + .module = THIS_MODULE, +#ifdef CONFIG_CGROUP_TIMER_SLACK + .subsys_id = timer_slack_subsys_id, +#endif + .create = tslack_cgroup_create, + .destroy = tslack_cgroup_destroy, + .attach = tslack_cgroup_attach, + .populate = tslack_cgroup_populate, +}; + +static int __init init_cgroup_timer_slack(void) +{ + BUG_ON(timer_slack_check != dummy_timer_slack_check); + timer_slack_check = cgroup_timer_slack_check; + return cgroup_load_subsys(&timer_slack_subsys); +} + +static void __exit exit_cgroup_timer_slack(void) +{ + BUG_ON(timer_slack_check != cgroup_timer_slack_check); + timer_slack_check = dummy_timer_slack_check; + cgroup_unload_subsys(&timer_slack_subsys); +} + +module_init(init_cgroup_timer_slack); +module_exit(exit_cgroup_timer_slack); +MODULE_LICENSE("GPL"); diff --git a/kernel/sys.c b/kernel/sys.c index 18da702..57322a7 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -118,6 +118,16 @@ EXPORT_SYMBOL(cad_pid); void (*pm_power_off_prepare)(void); +int dummy_timer_slack_check(struct task_struct *task, unsigned long slack_ns) +{ + return 0; +} +EXPORT_SYMBOL_GPL(dummy_timer_slack_check); + +int (*timer_slack_check)(struct task_struct *task, unsigned long slack_ns) = + dummy_timer_slack_check; +EXPORT_SYMBOL_GPL(timer_slack_check); + /* * set the priority of a task * - the caller must hold the RCU read lock @@ -1694,12 +1704,14 @@ SYSCALL_DEFINE5(prctl, int, option, unsigned long, arg2, unsigned long, arg3, error = current->timer_slack_ns; break; case PR_SET_TIMERSLACK: - if (arg2<= 0) - current->timer_slack_ns = - current->default_timer_slack_ns; - else - current->timer_slack_ns = arg2; - error = 0; + if (arg2<= 0) { + me->timer_slack_ns = me->default_timer_slack_ns; + break; + } + + error = timer_slack_check(me, arg2); + if (!error) + me->timer_slack_ns = arg2; break; case PR_MCE_KILL: if (arg4 | arg5) -- 1.7.3.5 _______________________________________________ Containers mailing list Containers(a)lists.linux-foundation.org https://lists.linux-foundation.org/mailman/listinfo/containers ----- End forwarded message -----

14 years, 10 months

Panda Board power supplies

by Joey Stanford

Hi All, For your records.... As you may have heard, the official power supply for Panda Boards has been discontinued. Digi-Key contacted me today with a suitable replacement in case we need them. This is the part: http://search.digikey.com/scripts/DkSearch/dksus.dll?Pname?Name=T1075-P5P-ND Joey

14 years, 10 months

query regarding perf

by Sudip Jain

Hi, I am struggling to run perf tool on u8500 as hwpacks for 2.6.35 kernel are still not out. Can someone help me know if anyone is using perf on arm-based board such as panda, beagle etc. Ideally, I would like to start a firefox application, and run command 'perf stat -p <pid_of_firefox> -e cycles sleep 1' Regards, Sudip

14 years, 10 months

Notes & Actions: Linaro Developer Platform Weekly Status Meeting 2/2

by Tom Gall

Enclosed you'll find a link to the agenda, notes and actions from the Linaro Developer Platforms Weekly Status meeting held on February 2nd in #linaro-meeting on irc.freenode.net at 16:00 UTC. https://wiki.linaro.org/Platform/Foundations/2011-02-02 Actions from the meeting where as follows: * JamieBennett to come up with test assignments for the new developer image * jcrigby to write up landing team u-boot, kernel workflows and send for review after first draft is done * slangasek to review https://wiki.linaro.org/Platform/DevPlatform/Specs/MemoryFootprintMethodolo… for dmart * ppearse to figure out what version number we want for 'helios' to be in natty * dmart to send a beagle kernel config to aviksil1 * aviksil1 to send pointer to lttng-enabled tree to slangasek Regards, Tom (tgall_foo) Developer Platforms Team "We want great men who, when fortune frowns will not be discouraged." - Colonel Henry Knox w) tom.gall att linaro.org w) tom_gall att vnet.ibm.com h) tom_gall att mac.com

14 years, 10 months

Re: linaro-n-validation-job-dispatcher

by Paul Larson

Moving this thread to the linaro-dev mailing list because I'd like it to be public discussion, and specifically to involve Rob since we talked about his interest in getting involved with LAVA development and especially the job dispatcher. 2. One queue TOTAL. One queue may seem like a bottleneck, but I don't think >> it has to be in practice. One process can monitor that queue, then launch a >> process or thread to handle each new job that comes in. >> > I think RabbitMQ message can have the ability to include the board > information in it to use one queue, and also we can use different queues for > different types of boards. > Right, but my question was, if you are already encoding information about which host to run on in the job stream, what is the advantage to having a queue for each board type, rather than a single queue? > >> >> >> Job description: >> I'd like to see some more detail here. Can you provide an example of a >> job file that would illustrate how you see it working? We also need to >> specify the underlying mechanisms that will handle parsing what's in the job >> file, and calling [something?] to do those tasks. What we have here feels >> like it might be a bit cumbersome. >> > I added an detailed one on the spec, like: > > 1. > > [Beagle-1, > http://snapshots.linaro.org/11.05-daily/linaro-hwpacks/imx51/20110131/0/ima…, > > http://snapshots.linaro.org/11.05-daily/linaro-headless/20110131/0/images/t…, > 900, ["abrek", "LTP", 180], ["none", "echo 100 > abc", 1], ... ] > 2. > > [IMX51-2, 20100131, 900, ["abrek", "LTP", 180], ["none", "echo 100 > > abc", 1], ... ] > > This looks almost like JSON, which is what Zygmunt was originally pushing for IIRC. If we are already going down that path, seems it would be sensible to take it a step further and have defined sections. For example: Tests:[ { name:"LTP", testsuite:"abrek", subtest:"ltp", <---- not thrilled about that name "subtest", but can't think of something better to call it at the moment timeout:180, reboot_after:True }, { name:"echo test", testsuite:"shell", timeout:20 reboot_after:False } ] What do you think? Obviously, there would be other sections for things like defining the host characteristics, the image to deploy, etc. > >> >> > Other questions... >> What if we have a dependency, how does that get installed on the target >> image? For example, if I want to do a test that requires bootchart be >> installed before the system is booted, we should be able to specify that, >> and have it installed before booting. >> >> What about installing necessary test suites? How do we tell it, in >> advance, what we need to have installed, and how does it get on the image >> before we boot, or before we start testing? >> > I think validation tools, test suites and necessary files are likely to > install after test image deployment. > Yes, clearly they have to be installed after deployment, but we may also need to consider them installing BEFORE we actually boot the test image. Thanks, Paul Larson

14 years, 10 months

Linaro 11.05 Alpha-2 (part 2) candidate testing

by Jamie Bennett

Hi, Due to an xserver issue in the Ubuntu archive (xserver packages built against old xserver-xorg-core) the Alpha-2 images have been re-rolled and are ready for your testing. The links are: Hardware Packs: http://releases.linaro.org/platform/linaro-n/hwpacks/alpha-2-rc2/ Headless: http://releases.linaro.org/platform/linaro-n/headless/alpha-2-rc2/ ALIP: http://releases.linaro.org/platform/linaro-n/alip/alpha-2-rc2/ Netbook: http://releases.linaro.org/platform/linaro-n/netbook/alpha-2-rc2/ Developer Tools image: http://releases.linaro.org/platform/linaro-n/developer/alpha-2-rc2/ As always, for an explanation of how to test the images on your hardware head on over to: http://wiki.linaro.org/Releases/DailyBuilds and submit your results to: http://qatracker.linaro.org For an explanation of how to use the qatracker please see: https://wiki.linaro.org/QA/QATracker Regards, Jamie. -- Linaro Release Manager

14 years, 10 months

RFC: Try moving kernel wg call to IRC?

by Dave Martin

Hi there, Just a thought--- is it worth experimenting with running the kernel wg meeting on IRC instead of the conf call? Personally, I find IRC meetings _much_ more digestible than conf calls (better "audio quality"; ability to scroll back when you don't hear / don't understand something etc.), but I don't know whether other people feel the same way -- that may just be my personal preference. Comparing the dev platform group's weekly IRC meeting with the kernel conf call, I've found that the IRC format works quite a lot better for me, but again, that may just be my experience. Cheers ---Dave

14 years, 10 months

Fwd: query regarding perf

by Sudip Jain

---------- Forwarded message ---------- From: Dave Martin <dave.martin(a)linaro.org> Date: Wed, Feb 2, 2011 at 4:55 PM Subject: Re: query regarding perf To: Sudip Jain <sudip.jain(a)linaro.org> On Wed, Feb 2, 2011 at 11:09 AM, Sudip Jain <sudip.jain(a)linaro.org> wrote: > I am getting the error when I execute perf on U8500 > No permission to collect stats. > Consider tweaking /proc/sys/kernel/perf_event_paranoid. > The problem I think is perf version and kernel version are not same.On > ubuntu host x86 machine, if I do sudo apt-get install linux-tools-common, > then perf runs succesfully. Even I downloaded the source code using sudo > apt-get source linux-tools-common and re-built perf. It works. Yes, you'll need the perf tools to be (reasonably) in sync with the kernel. The linux-linaro-tools package provides the tools for the released kernels, but if you're working with a development kernel it may be better to build the tools yourself, as you describe. > Recently my ubuntu host crashed, I re-installed with 10.04, the kernel > version is 2.6.32-21 and perf version is 2.6.35.8. So again I am getting the > same error (now on ubuntu host) which I mentioned above. Though now I have > the opportunity to debug the perf code, but that is not my target. > I am hoping that panda/beagle is fully supported for ubuntu mainstream AFAIK panda and beagle both ought to work, but please shout if you run into problems! Cheers ---Dave

14 years, 10 months

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

linaro-dev February 2011