linaro-dev February 2012

linaro-dev@lists.linaro.org

150 participants
147 discussions

[ACTIVITY] Graphics WG weekly report - wk 05.2012 (20120130-20120203)

by Ilias Biris

Hello! Please find below the summary of the weekly report, which can be found in https://wiki.linaro.org/WorkingGroups/Middleware/Graphics/WeeklyReport with more details and links. The last meeting minutes: https://wiki.linaro.org/WorkingGroups/Middleware/Graphics/Notes/2012-02-01 Please check also the Connect Graphics and Multimedia sessions as lined up through the blueprints: https://blueprints.launchpad.net/sprints/lcq1-12?searchtext=linaro-gfxmm == Highlights == - 12.01 release completed, here are some statistics: + 3 out of 6 blueprints completed, + 2 Lava related blueprints moved to 12.02 (https://blueprints.launchpad.net/linaro-graphics-tests/+spec/lava-tests-run… and https://blueprints.launchpad.net/linaro-graphics-dashboard/+spec/linaro-gra… both those blueprints are still in need of infrastructure) + the dma-buf + v4l2 as a user blueprint was put on hold since there is discussion ongoing related to the evolution of dma-buf - Worked to resolve issues for the Nux/Compiz/Unity release, Unity with gles2.0 branch was approved and on its way to be merged - Fixed display issue when playing two desktop scenes in a row, exposed by the canvas-reset branch functionality == Issues == - Dashboard bugs: #881789 - timeout in image deployment and #904796 - provide staging/development lava server instances are a still there and pose a risk for our work with the LAVA dashboard for glmark2 tests - The following issues need to be addressed for 12.02 release: + compiz issues (eg organize and prioritize the "refactoring" - chunking it down into manageable patches - of the compiz changes so we can move forward sanely) + upstream (e.g. dmabuf) review cycles + data publishing policy and what it means for our dashboard work == Risks == - Even though there are actions now taken to deal with Unity/Nux/Compiz, it is possible that those packages are have issues which may require more time for the release - If the dashboard related bugs are not resolved during Connect it will be once again necessary to move those blueprints forward. This would be the 3rd milestone in a row where we move those blueprints, so we should work together with Validation team to make sure the bugs are resolved - Due to the current process of discussing the evolution of dma-buf with upstream (DRM and V4L2 communities) it is very probable to not have something ready for dma-buf for 12.02, we would then target 12.03 instead Best regards, -- Ilias Biris ilias.biris(a)linaro.org Project Manager, Linaro M: +358504839608, IRC: ibiris Skype: ilias_biris Linaro.org│ Open source software for ARM SoCs

12 years, 11 months

[PATCH] module: debugging check for runaway kthreads

by Dmitry Antipov

Debugging option CONFIG_MODULE_KTHREAD_CHECK provides a way to check whether all kernel threads created by the module and have used module code as a thread worker function are really exited when the module is unloaded. See http://marc.info/?l=linux-kernel&m=132811276131767&w=2 to check why it might be useful. Signed-off-by: Dmitry Antipov <dmitry.antipov(a)linaro.org> --- include/linux/kthread.h | 5 +++++ init/Kconfig | 9 +++++++++ kernel/kthread.c | 24 ++++++++++++++++++++++++ kernel/module.c | 45 +++++++++++++++++++++++++++++++++++++++++++++ 4 files changed, 83 insertions(+), 0 deletions(-) diff --git a/include/linux/kthread.h b/include/linux/kthread.h index 0714b24..33897c3 100644 --- a/include/linux/kthread.h +++ b/include/linux/kthread.h @@ -13,6 +13,11 @@ struct task_struct *kthread_create_on_node(int (*threadfn)(void *data), #define kthread_create(threadfn, data, namefmt, arg...) \ kthread_create_on_node(threadfn, data, -1, namefmt, ##arg) +#ifdef CONFIG_MODULE_KTHREAD_CHECK +unsigned long get_kthread_func(struct task_struct *tsk); +#else +#define get_kthread_func(tsk, addr, mod) (0) +#endif /** * kthread_run - create and wake a thread. diff --git a/init/Kconfig b/init/Kconfig index 3f42cd6..fa7c6e0 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -1397,6 +1397,15 @@ config MODULE_FORCE_UNLOAD rmmod). This is mainly for kernel developers and desperate users. If unsure, say N. +config MODULE_KTHREAD_CHECK + bool "Check for runaway kernel threads at module unload" + depends on MODULE_UNLOAD && EXPERIMENTAL && DEBUG_KERNEL + help + This option allows you to check whether all kernel threads created + by the module and have used module code as a thread worker function + are really exited when the module is unloaded. This is mainly for + module developers. If insure, say N. + config MODVERSIONS bool "Module versioning support" help diff --git a/kernel/kthread.c b/kernel/kthread.c index 3d3de63..5c53817 100644 --- a/kernel/kthread.c +++ b/kernel/kthread.c @@ -38,6 +38,9 @@ struct kthread_create_info struct kthread { int should_stop; +#ifdef CONFIG_MODULE_KTHREAD_CHECK + void *fn; +#endif void *data; struct completion exited; }; @@ -45,6 +48,24 @@ struct kthread { #define to_kthread(tsk) \ container_of((tsk)->vfork_done, struct kthread, exited) +#ifdef CONFIG_MODULE_KTHREAD_CHECK + +unsigned long get_kthread_func(struct task_struct *tsk) +{ + struct kthread *kt; + unsigned long addr; + + get_task_struct(tsk); + BUG_ON(!(tsk->flags & PF_KTHREAD)); + kt = to_kthread(tsk); + barrier(); + addr = tsk->vfork_done ? (unsigned long)kt->fn : 0UL; + put_task_struct(tsk); + return addr; +} + +#endif /* CONFIG_MODULE_KTHREAD_CHECK */ + /** * kthread_should_stop - should this kthread return now? * @@ -106,6 +127,9 @@ static int kthread(void *_create) int ret; self.should_stop = 0; +#ifdef CONFIG_MODULE_KTHREAD_CHECK + self.fn = threadfn; +#endif self.data = data; init_completion(&self.exited); current->vfork_done = &self.exited; diff --git a/kernel/module.c b/kernel/module.c index 2c93276..fe6637b 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -45,6 +45,7 @@ #include <linux/string.h> #include <linux/mutex.h> #include <linux/rculist.h> +#include <linux/kthread.h> #include <asm/uaccess.h> #include <asm/cacheflush.h> #include <asm/mmu_context.h> @@ -764,6 +765,49 @@ static void wait_for_zero_refcount(struct module *mod) mutex_lock(&module_mutex); } +#ifdef CONFIG_KALLSYMS +static const char *get_ksymbol(struct module *mod, unsigned long addr, + unsigned long *size, unsigned long *offset); +#else +#define get_ksymbol(mod, addr, size, offset) NULL +#endif + +#ifdef CONFIG_MODULE_KTHREAD_CHECK + +static void check_kthreads(struct module *mod) +{ + unsigned long flags; + struct task_struct *g, *p; + + read_lock_irqsave(&tasklist_lock, flags); + do_each_thread(g, p) { + const char *name; + unsigned long addr, offset, size; + + /* Note kthreadd is special. Other kthreads should + have their 'struct kthread' on the stack until + do_exit() calls schedule() for the last time. */ + if (p->mm || p == kthreadd_task) + continue; + + addr = get_kthread_func(p); + if (__module_text_address(addr) == mod) { + name = get_ksymbol(mod, addr, &size, &offset); + printk(KERN_WARNING "khread %p[%s:%d] running " + "0x%lx(%s) is still alive, fix module %s, " + "crash possible\n", p, p->comm, p->pid, addr, + (name ? name : "<unknown>"), mod->name); + } + } while_each_thread(g, p); + read_unlock_irqrestore(&tasklist_lock, flags); +} + +#else + +#define check_kthreads(mod) do { } while (0) + +#endif /* CONFIG_MODULE_KTHREAD_CHECK */ + SYSCALL_DEFINE2(delete_module, const char __user *, name_user, unsigned int, flags) { @@ -831,6 +875,7 @@ SYSCALL_DEFINE2(delete_module, const char __user *, name_user, blocking_notifier_call_chain(&module_notify_list, MODULE_STATE_GOING, mod); async_synchronize_full(); + check_kthreads(mod); /* Store the name of the last unloaded module for diagnostic purposes */ strlcpy(last_unloaded_module, mod->name, sizeof(last_unloaded_module)); -- 1.7.7.6

12 years, 11 months

Bugs report for Freescale i.MX53 Loco board run Android and Ubuntu with Linaro 11.12&12.01

by orientsusong

Dear Linaro Dev Team, Thanks a lot for your kindly and attentions, firstly! i test the Linaro Version 11.12&12.01 for MX53 Loco board this two months in Shenzhen, China. Bug 1: Boot Time too long, it needs 1 minutes past 53 seconds for Ubuntu on Linaro Version 11.12&12.01. and Android 4.0.1 and 4.0.3 boot time needs 2 minutes past 55 seconds. Bug 2: Android Launcher no response and unlock. Bug 3: HDMI sound interrupt. Bug 4: HDMI Display not fine and some noise signal on the monitor with Ubuntu 11.10 over Linaro 12.01. Thanks again! -------------------------------------------------------------------------------- If you have any question or requirement please feel free to contact me. Keep in touch. Thanks for your co-operations! 有任何疑问和问题，请及时跟宋超联系，谢谢支持！顺祝商祺 ! Best Regards! 宋超 Steven Song Tel:+86 755 86,638,386;86,638,396 +86 755 86,0298,880;86,639,890 Ext 816 Fax:+86 755 86,029,992 Mobile:+86 13,682,508,330 Email: orientsusong(a)126.com orientsusong(a)gmail.com MSN: orientsusong(a)hotmail.com QQ:472716555 Skype: orientsusong 深圳市海泰信科技发展有限公司 Shenzhen Hithing Technology Development Co., Ltd. Website: http://www.hithing.com/index.asp 中国深圳南山科技园南区长虹科技大厦 1301 室 518057 Rm1301, 13F,Changhong Technology Bldg.,South Block, Hi-Tech Park,Nanshan Dist.,Shenzhen 518057,P.R.China *************************************************************************************** The information in this email is confidential and may be legally privileged. Access to this email by anyone other than the intended addressee is unauthorized. If you are not the intended recipient of this message, any review, disclosure, copying, distribution, retention, or any action taken or omitted to be taken in reliance on it is prohibited and may be unlawful. If you are not the intended recipient, please reply to or forward a copy of this message to the sender and delete the message, any attachments, and any copies thereof from your system. ****************************************************************************************

12 years, 11 months

Linaro Android against vanilla linux kernel versions

by Guclu, Tugrul

Hi I have already asked thi question at http://ask.linaro.org/questions/607/looking-for-android-distribution-bui ld-on-top-of-linux-kernel-26359-for-imx53-board <http://ask.linaro.org/questions/607/looking-for-android-distribution-bu ild-on-top-of-linux-kernel-26359-for-imx53-board> but could not receive satisfactory answer. Exactly how can you determine against which vanilla linux kernel version linaro Android distribution is built ? Thanks for your support BR Tugrul

12 years, 11 months

ubuntu-desktop/sources.txt bug/typo?

by Scott Douglass

Hi, I've noticed what looks like a bug/typo in http://releases.linaro.org/images/12.01/oneiric/ubuntu-desktop/sources.txt and I wonder if this is the right place to report it? Thanks. The last entry has "linarotv-xbmc" where I would expect "ubuntu-desktop": ... panda-ubuntu-desktop: http://releases.linaro.org/12.01/ubuntu/oneiric-hwpacks/ hwpack_linaro-lt-panda-x11-base_20120123-1_armel_supported.tar.gz: md5sum(4dae5fc312b4e0a73d81511e4e4b1e55) http://releases.linaro.org/12.01/ubuntu/oneiric-images/ubuntu-desktop linaro-o-linarotv-xbmc-tar-20120125-0.tar.gz: md5sum(e27e3af3d00eefa5f64f72f990482053) -- IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.

12 years, 11 months

Connect Sessions from the Developer Platform Team

by Ricardo Salveti

Hey folks, As usual, follows a list of the sessions the Developer Platform Team will lead at this Connect: Ubuntu LEB and LAVA: Current status and future planning for proper image testing and validation https://blueprints.launchpad.net/linaro-ubuntu/+spec/linaro-platforms-q112-… Improvements and future discussions for LTs and the Ubuntu LEB https://blueprints.launchpad.net/linaro-ubuntu/+spec/linaro-platforms-q112-… Packaged Kernel CI: Current Status and Next Steps https://blueprints.launchpad.net/linaro-ubuntu/+spec/linaro-platforms-q112-… Sysroots: Automation, Maintenance and Future Work https://blueprints.launchpad.net/linaro-ubuntu/+spec/linaro-platforms-q112-… U-Boot-Linaro Future Planning https://blueprints.launchpad.net/linaro-ubuntu/+spec/linaro-platforms-q112-… Developer Platform: Future Planning https://blueprints.launchpad.net/linaro-ubuntu/+spec/linaro-platforms-q112-… Cross Build with Multi-Arch: Current state and future planning https://blueprints.launchpad.net/linaro-ubuntu/+spec/linaro-platforms-q112-… Please make sure you subscribe to the ones you'd like to attend, and also add yourself as 'essential' to the ones you really like to participate (so the scheduler can find a free spot that's also available for you). This is not yet the final list, as we'll probably have additional ones until the end of the week, but you can always look for all the sessions related with the Developer Platform Team at https://blueprints.launchpad.net/linaro-ubuntu?searchtext=linaro-platforms-… Thanks, -- Ricardo Salveti de Araujo

12 years, 11 months

Fwd: Template

by Zach Pfeffer

If anyone needs a template for presentations, this one looks nice. Thank Tony! ---------- Forwarded message ---------- From: Tony Mansson <tony.mansson(a)linaro.org> Date: 1 February 2012 11:44 Subject: Template To: Botao Sun <Botao.Sun(a)linaro.org>, Bernhard Rosenkranzer <Bernhard.Rosenkranzer(a)linaro.org>, Vishal Bhoj <vishal.bhoj(a)linaro.org>, Zach Pfeffer <zach.pfeffer(a)linaro.org>, Amit Pundir <amit.pundir(a)linaro.org> Here, use this as template. -- 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

12 years, 11 months

cpu_relax in busy loops question

by Daniel Lezcano

Hi Linus, While looking at the drivers/mfd/db8500-prcmu.c file I noticed: 595 /* Grab the HW semaphore. */ 596 while ((readl(PRCM_SEM) & PRCM_SEM_PRCM_SEM) != 0) 597 cpu_relax(); I was wondering why is cpu_relax needed here as readl does a memory barrier ? I thought the cpu_relax function was related to the x86 to consume less power with an optimization of the "rep nop" loops or/and do a memory barrier. Should a busy-loop be always with cpu_relax ? Thanks -- Daniel -- <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, 11 months

What our customers want from Android

by Zach Pfeffer

So I've been talking to our customers (10 at last count) and they really like what we're doing. There are three things that are getting in their way: linaro-android-media-create GCC4.6 frequent kernel upgrades They love our builds, but they'd really like it if they could get stock AOSP builds for their boards on stable kernels that they can work with using fastboot that have been CI tested and QA'd. I'm not advocating for the wholesale destruction of our current way of life, tip kernels, tip toolchains, linaro-android-media-create, but I would like to move in a direction that gives our users what they want. -- 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

12 years, 11 months

[PATCH v6 0/4] Add SMDK5250 board support

by Chander Kashyap

This patchset add support for Samsung's SMDK5250 board based on EXYNOS5250 based SoC. It also adds support for MMC SPL booting. The porting is done by Samsung engineers at HQ in System LSI Team. I am contributing in upstreaming the code for the board. Based upon discussions following patches are dropped in this version: Exynos: Add CONFIG_EXYNOS4 Macro to EXYNOS4 based boards Exynos: Clock.c: Replace exynos4 prefix with exynos SMDK5250: enable device tree support is squashed with EXYNOS: Add SMDK5250 board support Chander Kashyap (4): Exynos: Clock.c: Use CONFIG_SYS_CLK_FREQ macro ARM: EXYNOS: Add support for Exynos5 based SoCs EXYNOS: Add SMDK5250 board support EXYNOS: SMDK5250: Add MMC SPL support MAINTAINERS | 1 + arch/arm/cpu/armv7/exynos/clock.c | 214 ++++++++++- arch/arm/include/asm/arch-exynos/clock.h | 326 ++++++++++++++++ arch/arm/include/asm/arch-exynos/cpu.h | 35 ++- arch/arm/include/asm/arch-exynos/dmc.h | 146 +++++++ arch/arm/include/asm/arch-exynos/gpio.h | 99 +++++- arch/arm/include/asm/arch-exynos/tzpc.h | 52 +++ board/samsung/smdk5250/Makefile | 67 ++++ board/samsung/smdk5250/clock_init.c | 202 ++++++++++ board/samsung/smdk5250/dmc_init.c | 508 +++++++++++++++++++++++++ board/samsung/smdk5250/lowlevel_init.S | 96 +++++ board/samsung/smdk5250/mmc_boot.c | 58 +++ board/samsung/smdk5250/setup.h | 353 +++++++++++++++++ board/samsung/smdk5250/smdk5250.c | 152 ++++++++ board/samsung/smdk5250/tools/mkexynos_image.c | 117 ++++++ board/samsung/smdk5250/tzpc_init.c | 48 +++ boards.cfg | 1 + include/configs/s5pc210_universal.h | 1 + include/configs/smdk5250.h | 202 ++++++++++ include/configs/trats.h | 1 + 20 files changed, 2662 insertions(+), 17 deletions(-) create mode 100644 arch/arm/include/asm/arch-exynos/dmc.h create mode 100644 arch/arm/include/asm/arch-exynos/tzpc.h create mode 100644 board/samsung/smdk5250/Makefile create mode 100644 board/samsung/smdk5250/clock_init.c create mode 100644 board/samsung/smdk5250/dmc_init.c create mode 100644 board/samsung/smdk5250/lowlevel_init.S create mode 100644 board/samsung/smdk5250/mmc_boot.c create mode 100644 board/samsung/smdk5250/setup.h create mode 100644 board/samsung/smdk5250/smdk5250.c create mode 100644 board/samsung/smdk5250/tools/mkexynos_image.c create mode 100644 board/samsung/smdk5250/tzpc_init.c create mode 100644 include/configs/smdk5250.h -- 1.7.5.4

12 years, 11 months

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

linaro-dev February 2012