linaro-dev

linaro-dev@lists.linaro.org

4332 discussions

Participating Remotely at Linaro Connect Q1.12

by Stephen Doel

Hi All, If you're one of the unlucky ones who won't be attending Linaro Connect Q1.12 next week, you'll be pleased to hear we're enabling remote participation via Google Hangouts for the 5 morning Session rooms. As Hangouts allow use of webcams, this will hopefully make it a more interactive experience for you - for example with webcams located in the rooms you can see who's in the room and who's talking at any particular time (and you'll better know who to tell to speak up if they're too quiet J). We've provided guidelines for remote participation at the bottom of the Schedule page here: http://connect.linaro.org/schedule or directly here: http://connect.linaro.org/remote I'm conscious this is the first time we've used this facility, so do let me know if the guidelines aren't clear. And then I'll also appreciate your feedback during the week to help us improve next time. Thx Stephen Doel Chief Operating Officer T: +44 1223 45 00 23 │ M: +44 77 66 014 247 <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs Follow Linaro: <http://www.facebook.com/pages/Linaro/155974581091106> Facebook | <http://twitter.com/#!/linaroorg> Twitter | <http://www.linaro.org/linaro-blog/> Blog

13 years, 11 months

External tarballs can now be unpacked into a build

by James Tunnicliffe

Hi, https://wiki.linaro.org/Platform/Android/LinaroAndroidBuildService has been updated with these instructions about a new build option: EXTERNAL_TARBALL Use to request that the build system fetch an archive from the location that you set EXTERNAL_TARBALL to and unpack it into build/external_tarballs (build is the build root directory. This exact path can be written as $BUILD_SCRIPT_ROOT/../../build/external_tarballs). Multiple archives can be requested, separated by ";", e.g. EXTERNAL_TARBALL="http://foo.com/ball1.tar.gz;http://foo.com/ball2.tar.bz2" There has been some discussion about if the path that is being unpacked to is easy to use. We could modify the build scripts so we either set an environment variable to point to where we unpack to ($EXT_TARBALL_DIR) or have a fixed path in /tmp that we soft link from (/tmp/external_tarballs). If you have a strong opinion, please let me know. Thanks, -- James Tunnicliffe

13 years, 11 months

[PATCH 1/3] percpu: use ZERO_SIZE_PTR / ZERO_OR_NULL_PTR

by Dmitry Antipov

Fix pcpu_alloc() to return ZERO_SIZE_PTR if requested size is 0; fix free_percpu() to check passed pointer with ZERO_OR_NULL_PTR. Signed-off-by: Dmitry Antipov <dmitry.antipov(a)linaro.org> --- mm/percpu.c | 16 +++++++++++----- 1 files changed, 11 insertions(+), 5 deletions(-) diff --git a/mm/percpu.c b/mm/percpu.c index f47af91..e903a19 100644 --- a/mm/percpu.c +++ b/mm/percpu.c @@ -702,7 +702,8 @@ static struct pcpu_chunk *pcpu_chunk_addr_search(void *addr) * Does GFP_KERNEL allocation. * * RETURNS: - * Percpu pointer to the allocated area on success, NULL on failure. + * ZERO_SIZE_PTR if @size is zero, percpu pointer to the + * allocated area on success or NULL on failure. */ static void __percpu *pcpu_alloc(size_t size, size_t align, bool reserved) { @@ -713,7 +714,10 @@ static void __percpu *pcpu_alloc(size_t size, size_t align, bool reserved) unsigned long flags; void __percpu *ptr; - if (unlikely(!size || size > PCPU_MIN_UNIT_SIZE || align > PAGE_SIZE)) { + if (unlikely(!size)) + return ZERO_SIZE_PTR; + + if (unlikely(size > PCPU_MIN_UNIT_SIZE || align > PAGE_SIZE)) { WARN(true, "illegal size (%zu) or align (%zu) for " "percpu allocation\n", size, align); return NULL; @@ -834,7 +838,8 @@ fail_unlock_mutex: * Does GFP_KERNEL allocation. * * RETURNS: - * Percpu pointer to the allocated area on success, NULL on failure. + * ZERO_SIZE_PTR if @size is zero, percpu pointer to the + * allocated area on success, NULL on failure. */ void __percpu *__alloc_percpu(size_t size, size_t align) { @@ -856,7 +861,8 @@ EXPORT_SYMBOL_GPL(__alloc_percpu); * Does GFP_KERNEL allocation. * * RETURNS: - * Percpu pointer to the allocated area on success, NULL on failure. + * ZERO_SIZE_PTR if @size is zero, percpu pointer to the + * allocated area on success or NULL on failure. */ void __percpu *__alloc_reserved_percpu(size_t size, size_t align) { @@ -917,7 +923,7 @@ void free_percpu(void __percpu *ptr) unsigned long flags; int off; - if (!ptr) + if (unlikely(ZERO_OR_NULL_PTR(ptr))) return; kmemleak_free_percpu(ptr); -- 1.7.7.6

13 years, 11 months

Initial list of Blueprints for 12.02, from the Developer Platform Team and Android Team

by Zach Pfeffer

Actually including linaro-android and modifying the subject a little... On 30 January 2012 22:24, Zach Pfeffer <zach.pfeffer(a)linaro.org> wrote: > Linaro Android 12.02 is also ready to roll: > > https://launchpad.net/linaro-android/+milestone/12.02 > > We're generally working through the issues from the last cycle listed > in the 12.01 spreadsheet entry with a distinct Snowball bent this > cycle - since we're trying and get Google Hangouts working on it. For > the future, we'll generally put a heavy emphasis on working to make > all features green. > > The test sheets are here: > > Versatile Express A9 > https://docs.google.com/a/linaro.org/spreadsheet/ccc?key=0AnpUtxWjZbP9dENmW… > > iMX53 > https://docs.google.com/a/linaro.org/spreadsheet/ccc?key=0AkxwyUNxNaAadHlNN… > > Panda > https://docs.google.com/a/linaro.org/spreadsheet/ccc?key=0AkxwyUNxNaAadHFPb… > > Snowball > https://docs.google.com/a/linaro.org/spreadsheet/ccc?key=0AkxwyUNxNaAadHJoV… > > Origen > https://docs.google.com/a/linaro.org/spreadsheet/ccc?key=0AkxwyUNxNaAadHJXV… > > booya > > On 30 January 2012 21:21, Ricardo Salveti <ricardo.salveti(a)linaro.org> wrote: >> Hi guys, >> >> As we're almost done with 12.02 planning, I'd like to share the >> current blueprints we intend to work during the 12.02 cycle: >> https://launchpad.net/linaro-dev-platform/+milestone/12.02 >> >> I know this is a quite short cycle, as we'll have the Linaro Connect, >> Android Builders and ELC, but I really hope to get most of the >> blueprints done during Connect's hacking sessions. >> >> Some interesting topics we'll cover this month: >> * Adding the USB-Host enablement test cases at LAVA >> * Start testing and validating the daily Linaro Native and Cross GCC packages >> * Enable ARMHF Precise based builds >> * Update XBMC to the latest 11.0 Beta release (currently Beta 2) >> * Sysroot build automation >> * Backporting of Unity 5, bringing the latest development from the GWG >> * Finalizing the Cross Buildd with Multi-Arch support >> >> And a few other bug fixes :-) >> >> Hope you enjoy ;-) >> >> Cheers, >> -- >> Ricardo Salveti de Araujo >> >> _______________________________________________ >> linaro-dev mailing list >> linaro-dev(a)lists.linaro.org >> http://lists.linaro.org/mailman/listinfo/linaro-dev > > > > -- > Zach Pfeffer > Android Platform Team Lead, Linaro Platform Teams > Linaro.org | Open source software for ARM SoCs > Follow Linaro: http://www.facebook.com/pages/Linaro > http://twitter.com/#!/linaroorg - http://www.linaro.org/linaro-blog -- Zach Pfeffer Android Platform Team Lead, Linaro Platform Teams Linaro.org | Open source software for ARM SoCs Follow Linaro: http://www.facebook.com/pages/Linaro http://twitter.com/#!/linaroorg - http://www.linaro.org/linaro-blog

13 years, 11 months

Initial list of Blueprints for 12.02, from the Developer Platform Team

by Ricardo Salveti

Hi guys, As we're almost done with 12.02 planning, I'd like to share the current blueprints we intend to work during the 12.02 cycle: https://launchpad.net/linaro-dev-platform/+milestone/12.02 I know this is a quite short cycle, as we'll have the Linaro Connect, Android Builders and ELC, but I really hope to get most of the blueprints done during Connect's hacking sessions. Some interesting topics we'll cover this month: * Adding the USB-Host enablement test cases at LAVA * Start testing and validating the daily Linaro Native and Cross GCC packages * Enable ARMHF Precise based builds * Update XBMC to the latest 11.0 Beta release (currently Beta 2) * Sysroot build automation * Backporting of Unity 5, bringing the latest development from the GWG * Finalizing the Cross Buildd with Multi-Arch support And a few other bug fixes :-) Hope you enjoy ;-) Cheers, -- Ricardo Salveti de Araujo

13 years, 11 months

[PATCH 3/3] module: use ZERO_OR_NULL_PTR allocation pointer checking

by Dmitry Antipov

Use ZERO_OR_NULL_PTR allocation pointer checking where allocation function may return ZERO_SIZE_PTR. Signed-off-by: Dmitry Antipov <dmitry.antipov(a)linaro.org> --- kernel/module.c | 12 ++++++------ 1 files changed, 6 insertions(+), 6 deletions(-) diff --git a/kernel/module.c b/kernel/module.c index 2c93276..5183f91 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -438,7 +438,7 @@ static int percpu_modalloc(struct module *mod, } mod->percpu = __alloc_reserved_percpu(size, align); - if (!mod->percpu) { + if (unlikely(ZERO_OR_NULL_PTR(mod->percpu))) { printk(KERN_WARNING "%s: Could not allocate %lu bytes percpu data\n", mod->name, size); @@ -572,7 +572,7 @@ EXPORT_TRACEPOINT_SYMBOL(module_get); static int module_unload_init(struct module *mod) { mod->refptr = alloc_percpu(struct module_ref); - if (!mod->refptr) + if (unlikely(ZERO_OR_NULL_PTR(mod->refptr))) return -ENOMEM; INIT_LIST_HEAD(&mod->source_list); @@ -2322,14 +2322,14 @@ static void dynamic_debug_remove(struct _ddebug *debug) void * __weak module_alloc(unsigned long size) { - return size == 0 ? NULL : vmalloc_exec(size); + return vmalloc_exec(size); } static void *module_alloc_update_bounds(unsigned long size) { void *ret = module_alloc(size); - if (ret) { + if (likely(!ZERO_OR_NULL_PTR(ret))) { mutex_lock(&module_mutex); /* Update module bounds. */ if ((unsigned long)ret < module_addr_min) @@ -2638,7 +2638,7 @@ static int move_module(struct module *mod, struct load_info *info) * leak. */ kmemleak_not_leak(ptr); - if (!ptr) + if (unlikely(ZERO_OR_NULL_PTR(ptr))) return -ENOMEM; memset(ptr, 0, mod->core_size); @@ -2652,7 +2652,7 @@ static int move_module(struct module *mod, struct load_info *info) * after the module is initialized. */ kmemleak_ignore(ptr); - if (!ptr && mod->init_size) { + if (unlikely(ZERO_OR_NULL_PTR(ptr)) && mod->init_size) { module_free(mod, mod->module_core); return -ENOMEM; } -- 1.7.7.6

13 years, 11 months

Linaro Audio development ideas for 12.02 and beyond

by Kurt Taylor

Hi all, Here is a brain dump for some things I see on the horizon for audio work. Is this complete? Absolutely not. This is meant to be a place to capture and refine ideas before creating cards and/or blueprints for them. In other words, this should compliment the existing work and backlog already in LP. https://wiki.linaro.org/KurtTaylor/AudioBacklog Please edit or send any comments to me. Thanks! -- Kurt Taylor (irc krtaylor) Linaro Multimedia Linaro.org <http://www.linaro.org/>* **│ *Open source software for ARM SoCs Follow *Linaro: *Facebook <http://www.facebook.com/pages/Linaro> | Twitter<http://twitter.com/#%21/linaroorg>| Blog <http://www.linaro.org/linaro-blog/>

13 years, 11 months

Linaro Release 12.01 Postmortem Summary

by David Zinman

Greeting, Here is the postmortem and lessons learned for Linaro's first release for 2012. https://wiki.linaro.org/Cycles/1201/Release/Review Highlights and Key Successes ============================ The Ubuntu development platform team has delivered Ubuntu TV fully accelerated on a PandaBoard with Ubuntu LEB and the Ubuntu TV UI and XBMC also for PandaBoard. http://www.youtube.com/watch?feature=player_embedded&v=uQMOETMZA5Y The members demonstration was very successful with the graphics WG, landing teams, verification team, platform teams and the Igloo community all contributing some very fine achievements to the effort. Internal communications have been getting better and there is now a weekly sync between the platform groups and the working groups that is similar to the platform - landing team syncs. Each team brings a high priority issue or but to the table in order to find a plan for resolution. Postmortem and Lessons Learned ============================== This cycle seemed to present a challenge to most of the teams to stay focused on the delivery, There was no release from the Power Management group, and output was down for the Multimedia, Graphics and Platform teams. An issue here is that some release procedures are not being followed by some of the teams. An important issue that came up this cycle is that platform teams are requesting the source code for binary blobs. At integration, when an issue has been found, the path to the person that can fix and regenerate the binary blob is too long to actually fit the integration into one cycle. Add to this the requirement to use kernels and toolchains that are not aligned with the binary blobs and a conflict is created between development and schedule. Some of the lessons learned that came out of the cycle are: * There needs to be offspring knowledge within the infrastructure. The resource that had this domain knowledge is no longer with Linaro. To be effective, this knowledge should be brought in house again. * Listen to community users and incorporate useful information and go to people directly to get their reviews. * Need to become better at declaring scope which has good probability to be delivered at the beginning of the cycle, rather than declaring all pieces of work that we want to work upon and shed items which are impeded on the way. Impediments should be highlighted early on to get resolved. -- David Zinman Linaro Release Manager | Project Manager Linaro.org | Open source software for ARM SoCs

13 years, 11 months

Transcript from our Kick-off meeting today.

by Tony Mansson

<pfefferz> all right so welcome to 12.02! the goal is to mkae snowball work great and origen iMX53 is semi blocked <bero> There's still stuff to try there <pfefferz> and panda gets a lot of attention <bero> But it would be really useful to get some more information <pfefferz> bero, lets hold off on iMX53 until we have more info <jackh> the three bugs assigned to me, i tried this afternoon, two of them still there <pfefferz> can you send the current issue set in? jackh, cool <jackh> the gallery one does not there <pfefferz> did you update the thrid? that's good feel free to list the build where you don't see it in the bug everyone we have a new team member jackh, aka Chenglie has joined us from IBM <jackh> it's Chengjie He <pfefferz> welcome Chengjie <bero> Welcome jackh <pabhishek> welcome jackh <pfefferz> right ;) <jackh> haha, just call me jacky for short... <bhoj2> welcom jackh :) <botao_sun> jackh: I will give you a brief later about compiling the code. <pfefferz> its great to have you on the Android team Chengjie He we'll do another welcome round on wednesday ;) <jackh> pls do it face to face, zach... haha, joking <pfefferz> :) so we're gonna focus on getting our tests all green its a big job as usuual :) we've had some good pickup by the member companies of our concerns so we shouldn't be as blocked as we were we also have connect abs and elc as a reminder all presentations for elc and abs have to be submitted a week before the event also for everyone doing a session please have a rough agenda listed something like here's where we are here's where we need to be I'm going to take a more passive role during other people's sessions this time :) any questions or comments? <mansson__> Prepare a "lightning talk" version of your session (5 min.) good to have. <pfefferz> good point mansson__ <pfefferz> the highest prio thing we have right now is the demo at connect most everyone is involoved in getting it to happen I suspect we'll be spending a bunch of time in the hacking sessions making it work I have to get a poster together so one more thing people arte really starting to use our builds 1. because they're easy to use 2. they work 3. they can talk to us on irc this means they're looking at our BPs to see how our work is progressing so that's all I had thanks for setting the meeting up Tony anyone have anything else? <mansson__> I think everything is pretty clear. We continue to make a lasting impression on human evolution. Making ARM Linux history. <pfefferz> exactly! <mansson__> 15 out of 50 talks at ELC are by Linaro people. <pfefferz> hehe not bad for a non-sponser seriously guys we're changing the world every little thing we get working has a giant impact okay <pfefferz> mansson__, would you send this transcript out to everyone? <mansson__> Yes. <pfefferz> thanks!

13 years, 11 months

[PATCH 2/3] vmalloc: use ZERO_SIZE_PTR / ZERO_OR_NULL_PTR

by Dmitry Antipov

- Fix vmap() to return ZERO_SIZE_PTR if 0 pages are requested; - fix __vmalloc_node_range() to return ZERO_SIZE_PTR if 0 bytes are requested; - fix __vunmap() to check passed pointer with ZERO_OR_NULL_PTR. Signed-off-by: Dmitry Antipov <dmitry.antipov(a)linaro.org> --- mm/vmalloc.c | 10 +++++++--- 1 files changed, 7 insertions(+), 3 deletions(-) diff --git a/mm/vmalloc.c b/mm/vmalloc.c index 86ce9a5..040a9cd 100644 --- a/mm/vmalloc.c +++ b/mm/vmalloc.c @@ -1456,7 +1456,7 @@ static void __vunmap(const void *addr, int deallocate_pages) { struct vm_struct *area; - if (!addr) + if (unlikely(ZERO_OR_NULL_PTR(addr))) return; if ((PAGE_SIZE-1) & (unsigned long)addr) { @@ -1548,7 +1548,9 @@ void *vmap(struct page **pages, unsigned int count, might_sleep(); - if (count > totalram_pages) + if (unlikely(!count)) + return ZERO_SIZE_PTR; + if (unlikely(count > totalram_pages)) return NULL; area = get_vm_area_caller((count << PAGE_SHIFT), flags, @@ -1648,8 +1650,10 @@ void *__vmalloc_node_range(unsigned long size, unsigned long align, void *addr; unsigned long real_size = size; + if (unlikely(!size)) + return ZERO_SIZE_PTR; size = PAGE_ALIGN(size); - if (!size || (size >> PAGE_SHIFT) > totalram_pages) + if (unlikely((size >> PAGE_SHIFT) > totalram_pages)) goto fail; area = __get_vm_area_node(size, align, VM_ALLOC | VM_UNLIST, -- 1.7.7.6

13 years, 11 months

Jump to page:

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

linaro-dev