linaro-dev

linaro-dev@lists.linaro.org

4332 discussions

by Bernhard Rosenkränzer

== Progress == * Toolchain binary release 2014.06 * NDK buildscripts * Started with Bionic optimizations * Misc. bughunting and code reviews == Plans == * NDK release * Finish Bionic optimizations == Issues == * None

11 years, 2 months

Idlestat output question

by Salman Ahmed

Hello, Running top command with idlestat tool doesn't show any frequency information sometimes. I have run the program with the following command - #sudo idlestat --trace -t 5 -f /tmp/myoutput /usr/bin/top and it ended up with the following output where there is no frequency state information. The file "myoutput" doesn't have any cpu_frequency entry either. Total trace buffer: 153896 kB Log is 4.999488 secs long with 675 events clusterA@state hits total(us) avg(us) min(us) max(us) C6-SNB 163 4955450.00 30401.53 0.00 141946.00 cpu0@state hits total(us) avg(us) min(us) max(us) C1-SNB 1 912.00 912.00 912.00 912.00 C6-SNB 115 4996270.00 43445.83 67.00 211819.00 cpu0 wakeups name count irq040 eth0 50 cpu1@state hits total(us) avg(us) min(us) max(us) C1-SNB 3 10809.00 3603.00 27.00 10742.00 C6-SNB 62 4985746.00 80415.26 59.00 223952.00 cpu1 wakeups name count irq018 ata_piix 6 cpu2@state hits total(us) avg(us) min(us) max(us) C1-SNB 3 10800.00 3600.00 26.00 10745.00 C6-SNB 52 4986684.00 95897.77 5724.00 223957.00 cpu2 wakeups name count cpu3@state hits total(us) avg(us) min(us) max(us) C1-SNB 20 15123.00 756.15 121.00 10895.00 C6-SNB 47 4962591.00 105587.04 1.00 223950.00 cpu3 wakeups name count Does that mean processors spent all time in idle state and none in active state? -- Best Regards Salman Ahmed

11 years, 2 months

[ACTIVITY] Week 25

by Vishal Bhoj

== Progress == * Enabled SELinux on Android builds for LSK (CARD-1452, 60 %) * ARMv8 Android RC build setup and testing (CARD-1104, 40 %) == Plans == * 14.06 Android RC. * Closeout CARD-1049 <https://cards.linaro.org/browse/CARD-1049> after setting the build. == Issues == * None. Regards, Vishal

11 years, 2 months

[ACTIVITY] week 25

by Harigopal Gollamudi

== Progress == CARD-1489: Migrating the arndale android builds to 4.9 toolchain. Built the code. need to verify on the HW. == Plans == CARD-1489: Test verify the binaries which got built with linaro toolchain v4.9. == Issues == None -- Thanks and Regards, HariGopal G Android Engineer

11 years, 2 months

[ACTIVITY] Week 25

by Riku Voipio

== Progress == * ARMv8 kvm booted and tested on custom kernel/fs (CARD-346 30%) * Fixing OE/Alip build failure 10% * Researched Lava vmgroups (BUG-1330469 10%) * Set up Xen loop on midway since arndale is broken (20%) * KVM loop for TC2 for mainline CI.. (20%) * Ported iaxclient to Aarch64, (BUG-1330436 10%) == Plans == * Preliminary KVM armv8 CI loop * ALIP image build fix (possibly involves upgrading firefox) * Escape to VAC on 29th == Issues == * Nobody seems to know how to boot TC2/ubuntu/lava/mainline

11 years, 2 months

idlestat: irq count reported is wrong

by Mohammad Merajul Islam Molla

Hello Amit/Daniel, I think there is a bug in idlestat due to which it reports wrong irq counts in output. Log is 9.998478 secs long with 1518 events clusterA@state hits total(us) avg(us) min(us) max(us) C6-SNB 455 9921571.00 21805.65 0.00 176303.00 cpu0@state hits total(us) avg(us) min(us) max(us) C1-SNB 10 2032.00 203.20 27.00 1219.00 C3-SNB 2 5909.00 2954.50 161.00 5748.00 C6-SNB 320 9978546.00 31182.96 48.00 237467.00 cpu0 wakeups name count irq040 eth0 150 cpu1@state hits total(us) avg(us) min(us) max(us) C6-SNB 140 9981689.00 71297.78 80.00 311045.00 cpu1 wakeups name count irq018 ata_piix 19 cpu2@state hits total(us) avg(us) min(us) max(us) C6-SNB 87 9995015.00 114885.23 11935.00 323889.00 cpu2 wakeups name count cpu3@state hits total(us) avg(us) min(us) max(us) C1-SNB 3 383.00 127.67 122.00 139.00 C6-SNB 97 9954601.00 102624.75 51.00 323890.00 cpu3 wakeups name count from my trace output file - grep 'irq=40' /tmp/trace | wc -l returns 308. Half should be the irq count = 154. grep 'irq=18' /tmp/trace | wc -l returns 80. Half should be irq count = 40. I have made a patch to fix the issue below. Below check causes early unwanted return from store_irq() function. In fact, the whole 'wakeirq' member seems unnecessary, hence the related code. if (cstates->wakeirq != NULL) return 0; Would you please check? diff --git a/idlestat.c b/idlestat.c index 5e8f8d5..d4ea859 100644 --- a/idlestat.c +++ b/idlestat.c @@ -654,7 +654,6 @@ static int store_data(double time, int state, int cpu, cstates->cstate[state].data = data; cstates->cstate_max = MAX(cstates->cstate_max, state); cstates->last_cstate = state; - cstates->wakeirq = NULL; /* update P-state stats if supported */ if (pstate) @@ -684,9 +683,6 @@ static int store_irq(int cpu, int irqid, char *irqname, struct wakeup_irq *irqinfo; struct wakeup_info *wakeinfo = &cstates->wakeinfo; - if (cstates->wakeirq != NULL) - return 0; - irqinfo = find_irqinfo(wakeinfo, irqid); if (NULL == irqinfo) { irqinfo = realloc(wakeinfo->irqinfo, @@ -705,8 +701,6 @@ static int store_irq(int cpu, int irqid, char *irqname, irqinfo->count++; - cstates->wakeirq = irqinfo; - return 0; } diff --git a/idlestat.h b/idlestat.h index 1977ab4..bb27e3a 100644 --- a/idlestat.h +++ b/idlestat.h @@ -76,7 +76,6 @@ struct cpuidle_cstates { struct wakeup_info wakeinfo; int last_cstate; int cstate_max; - struct wakeup_irq *wakeirq; }; struct cpufreq_pstate { -- Thanks, -Meraj

11 years, 2 months

Test Result Summary of Linux Linaro ubuntu in Calendar Week 25.

by Botao Sun

Calendar Week 25: Here is test result summary for Linux Linaro ubuntu Trusty image on following boards: 1) Samsung Arndale; 2) TI Panda 4430; 3) TI Panda 4460. 1. Samsung Arndale + Linux Linaro Trusty (Column BA): https://docs.google.com/a/linaro.org/spreadsheet/ccc?key=0AgB-fT5LL31CdGZJS… 2. TI Panda 4430 + Linux Linaro Trusty (Column BH): https://docs.google.com/a/linaro.org/spreadsheet/ccc?key=0AroPySpr4FnEdEwwZ… 3. TI Panda 4460 + Linux Linaro Trusty (Column BH): https://docs.google.com/a/linaro.org/spreadsheet/ccc?key=0AroPySpr4FnEdEwwZ… Thank you. Best Regards Botao Sun

11 years, 2 months

idlestat recent fix related query

by Mohammad Merajul Islam Molla

Recently there was a patch (commit f5b9bdcb530a8871e6c65a617d41001b39bf4912), to fix resource leak on dir_topology. Now, the below code segment is either unnecessary or it should call opendir on 'newpath' instead of 'path'? dir_topology = opendir(path); if (!dir_topology) continue; closedir(dir_topology); Please comment. -- Thanks, -Meraj

11 years, 2 months

[PATCH POWERDEBUG] Fix resource leaks in export_free_gpios()

by Mohammad Merajul Islam Molla

Fix for the below issues - 1. valgrind reports memory leaks in export_free_gpios. 'line' pointer was never freed. ==5732== 120 bytes in 1 blocks are definitely lost in loss record 130 of 271 ==5732== at 0x482CBB8: malloc (vg_replace_malloc.c:270) ==5732== by 0x48CC633: getdelim (iogetdelim.c:67) ==5732== by 0xB303: export_free_gpios (stdio.h:117) ==5732== by 0xB3DD: gpio_init (gpio.c:376) ==5732== by 0x97F5: main (powerdebug.c:264) 2. Open files need to be closed. diff --git a/gpio.c b/gpio.c index 39480b0..076b260 100644 --- a/gpio.c +++ b/gpio.c @@ -353,7 +353,15 @@ void export_free_gpios(void) if (system(command) < 0) printf("error: failed to export gpio-%d\n", i); } + + free(line); out: + if(fgpio) + fclose(fgpio); + + if(fgpio_export) + fclose(fgpio_export); + return; } -- Thanks, -Meraj

11 years, 2 months

Releases and Snapshots maintenance

by Benjamin Copeland

Scheduled: June 23, 2014 9:00AM - 11:00AM UTC Components Affected: releases.linaro.org, snapshots.linaro.org Description: Scheduled maintenance for releases.linaro.org and snapshots.linaro.org. We will be updating the Linaro license production code. https://bugs.launchpad.net/linaro-license-protection/+bug/1328998 For updates and notifications please subscribe to: http://dash.linaro.org/ Thanks Ben

11 years, 2 months

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

linaro-dev