linaro-dev

linaro-dev@lists.linaro.org

4332 discussions

by Dave Turvene - Work

I worked on a pandaboard Linaro android release last year (13.07 I think) and now need to use a Beaglebone black (BBB). Google search indicates that BBB is supported but I can't seem to find the config files to build it. Some of the searchs show building kernel, u-boot, apps separately and then setting up sdcard partitions which I will do if need be. BUT I'd much rather repeat the pandaboard environment. I'm happy to build Ubuntu, Android, OE, Debian, anything with a linux kernel.... Dave Turvene

11 years

[ACTIVITY] Week 39

by Koen Kooi

== Progress == * Release (100%) == Plans == * Fix openjdk build -- Koen Kooi Builds and Baselines | Release Manager Linaro.org | Open source software for ARM SoCs

11 years

[ACTIVITY] Week 39

by Riku Voipio

== Progress == * Investigated kvm oops regression in 3.17-rcX (20%) * Fixed libvirt-native build (10%) * Fixed Qemu package version and kvm net test (20%) * Researched building packages and running services in docker (20%) * Created card for tracking Aarch64 optimizations (CARD-1643 10%) * gcov/arm support now in rmk/for-next (10%) == Plans == * Provide docker versions of crossbuild and nativebuild in ubuntu-packaging-tools * Juno/next builds for kvm (CARD-1611)

11 years

[ACTIVITY] week 40

by Wookey

== Progress == * Buildd-tending: Debian-ports arm64 is at 88% (and 100% SCC) built (DEVPLAT-226: 5%) Debian official rebuild is at 77% (3 SCC packages not built) (CARD-1586: 20%) http://buildd.debian-ports.org/stats/graph-quarter-big.png https://buildd.debian.org/stats/graph-quarter-big.png * Tracked Openjdk-7 build falure on Juno vs mustang: (DebianBug #761604) * Arranged buildd/porterbox hardware in Linaro Build farm: (CARD-1586: 2%) * Built updated cross-binutils packages - now in debian testing (DEVPLAT-220: 5%) * Uploaded cross-gcc packages at debconf (rejected): (DEVPLAT-220: 5%) * Updated cross-gcc packages to fix doc-links issue: (DEVPLAT-220: 15%) * Worked on cross-gcc-standalone packages: (DEVPLAT-220: 20%) * Updated sbuild with multiarch support, and linaro chroot-setup-commands patches (20%) * Tested cross-llvm usage on Debian (needs work - filed upstream bug) == Plans == * Get cross-toolchain packages into testing before freeze * Upload multiarch-ready sbuild * Enable new buildd hardware == Issues == Shortage of hours in day. Wookey -- Principal hats: Linaro, Emdebian, Wookware, Balloonboard, ARM http://wookware.org/

11 years

[ACTIVITY] Wek 39

by Bernhard Rosenkränzer

== Progress == * Upstreamed a number of AOSP+clang patches * Started investigating why AOSP fails to boot if init was built with clang, but no final results yet * Built a version of gcc trunk with patches needed for arm-linux-androideabi- and aarch64-linux-android- targets == Plans == * Upstream remaining AOSP+clang patches * Work on getting AOSP+clang to boot to UI * Start builds of AOSP with gcc 5.0 snapshots * Try to figure out where cbuild1 keeps its gpg keys == Issues == * none

11 years

[PATCH v2] idlestat: Allocate struct cpufreq_pstate dynamically

by pi-cheng.chen

When using intel_pstate driver, "scaling_available_freqs" attr is not exported to sysfs. It causes assertion of idlestat due to memory of struct cpufreq_pstate was not allocated. Allocate struct cpufreq_pstate dynamically when getting frequency information from trace file instead of parsing available frequencies from sysfs Changes v1 to v2: Sort the cpufreq_pstate list when parsing events Signed-off-by: Pi-Cheng Chen <pi-cheng.chen(a)linaro.org> --- idlestat.c | 118 +++++++++++++++++++++++++++++-------------------------------- 1 file changed, 56 insertions(+), 62 deletions(-) diff --git a/idlestat.c b/idlestat.c index da615cb..d03f12e 100644 --- a/idlestat.c +++ b/idlestat.c @@ -561,8 +561,8 @@ static void release_pstate_info(struct cpufreq_pstates *pstates, int nrcpus) } /** - * build_pstate_info - parse cpufreq sysfs entries and build per-CPU - * structs to maintain statistics of P-state transitions + * build_pstate_info - allocate and initialize per-CPU structs to + * maintain statistics of P-state transitions * @nrcpus: number of CPUs * * Return: per-CPU array of structs (success) or NULL (error) @@ -578,55 +578,8 @@ static struct cpufreq_pstates *build_pstate_info(int nrcpus) memset(pstates, 0, sizeof(*pstates) * nrcpus); for (cpu = 0; cpu < nrcpus; cpu++) { - struct cpufreq_pstate *pstate; - int nrfreq; - char *fpath, *freq, line[256]; - FILE *sc_av_freq; - - if (asprintf(&fpath, CPUFREQ_AVFREQ_PATH_FORMAT, cpu) < 0) - goto clean_exit; - - /* read scaling_available_frequencies for the CPU */ - sc_av_freq = fopen(fpath, "r"); - free(fpath); - if (!sc_av_freq) { - fprintf(stderr, "warning: P-states not supported for " - "CPU%d\n", cpu); - continue; - } - freq = fgets(line, sizeof(line)/sizeof(line[0]), sc_av_freq); - fclose(sc_av_freq); - if (!freq) { - /* unlikely to be here, but just in case... */ - fprintf(stderr, "warning: P-state info not found for " - "CPU%d\n", cpu); - continue; - } - - /* tokenize line and populate each frequency */ - nrfreq = 0; - pstate = NULL; - while ((freq = strtok(freq, "\n ")) != NULL) { - struct cpufreq_pstate *tmp = realloc(pstate, sizeof(*pstate) * (nrfreq+1)); - if (!tmp) - goto clean_exit; - pstate = tmp; - - /* initialize pstate record */ - pstate[nrfreq].id = nrfreq; - pstate[nrfreq].freq = atol(freq); - pstate[nrfreq].count = 0; - pstate[nrfreq].min_time = DBL_MAX; - pstate[nrfreq].max_time = 0.; - pstate[nrfreq].avg_time = 0.; - pstate[nrfreq].duration = 0.; - nrfreq++; - freq = NULL; - } - - /* now populate cpufreq_pstates for this CPU */ - pstates[cpu].pstate = pstate; - pstates[cpu].max = nrfreq; + pstates[cpu].pstate = NULL; + pstates[cpu].max = 0; pstates[cpu].current = -1; /* unknown */ pstates[cpu].idle = -1; /* unknown */ pstates[cpu].time_enter = 0.; @@ -634,10 +587,52 @@ static struct cpufreq_pstates *build_pstate_info(int nrcpus) } return pstates; +} -clean_exit: - release_pstate_info(pstates, nrcpus); - return NULL; +/** + * alloc_pstate - allocate, sort, and initialize pstate struct + * to maintain statistics of P-state transitions + * @pstates: per-CPU P-state statistics struct + * @freq: frequency for which the newly pstate is allocated + * + * Return: the index of the newly allocated pstate struct + */ +static int alloc_pstate(struct cpufreq_pstates *pstates, unsigned int freq) +{ + struct cpufreq_pstate *pstate, *tmp; + int nrfreq, i, next = 0; + + pstate = pstates->pstate; + nrfreq = pstates->max; + + tmp = realloc(pstate, sizeof(*pstate) * (nrfreq + 1)); + if (!tmp) { + perror("realloc pstate"); + return -1; + } + pstate = tmp; + pstates->pstate = tmp; + pstates->max = nrfreq + 1; + + for (i = 0; i < nrfreq && freq <= pstate[i].freq; i++) + ; + + next = i; + for (i = nrfreq; i > next && i > 0; i--) { + pstate[i] = pstate[i - 1]; + pstate[i].id = i; + pstates->current = (pstates->current == i - 1)? i : pstates->current; + } + + pstate[next].id = next; + pstate[next].freq = freq; + pstate[next].count = 0; + pstate[next].min_time = DBL_MAX; + pstate[next].max_time = 0; + pstate[next].avg_time = 0; + pstate[next].duration = 0; + + return next; } static int get_current_pstate(struct cpuidle_datas *datas, int cpu, @@ -712,6 +707,9 @@ static void cpu_change_pstate(struct cpuidle_datas *datas, int cpu, cur = get_current_pstate(datas, cpu, &ps, &p); next = freq_to_pstate_index(ps, freq); + if (next < 0) + next = alloc_pstate(ps, freq); + assert(next >= 0); switch (cur) { case 1: @@ -763,7 +761,6 @@ static int store_data(double time, int state, int cpu, struct cpuidle_datas *datas, int count) { struct cpuidle_cstates *cstates = &datas->cstates[cpu]; - struct cpufreq_pstate *pstate = datas->pstates[cpu].pstate; struct cpuidle_cstate *cstate; struct cpuidle_data *data, *tmp; int nrdata, last_cstate = cstates->last_cstate; @@ -826,9 +823,8 @@ static int store_data(double time, int state, int cpu, /* need indication if CPU is idle or not */ cstates->last_cstate = -1; - /* update P-state stats if supported */ - if (pstate) - cpu_pstate_running(datas, cpu, time); + /* update P-state stats */ + cpu_pstate_running(datas, cpu, time); return 0; } @@ -846,9 +842,8 @@ static int store_data(double time, int state, int cpu, cstates->cstate_max = MAX(cstates->cstate_max, state); cstates->last_cstate = state; cstates->wakeirq = NULL; - /* update P-state stats if supported */ - if (pstate) - cpu_pstate_idle(datas, cpu, time); + /* update P-state stats*/ + cpu_pstate_idle(datas, cpu, time); return 0; } @@ -1018,7 +1013,6 @@ static struct cpuidle_datas *idlestat_load(struct program_options *options) } else if (strstr(buffer, "cpu_frequency")) { assert(sscanf(buffer, TRACE_FORMAT, &time, &freq, &cpu) == 3); - assert(datas->pstates[cpu].pstate != NULL); cpu_change_pstate(datas, cpu, freq, time); count++; continue; -- 1.9.1

11 years, 1 month

[ACTIVITY] Week 38

by Bernhard Rosenkränzer

== Progress == * Attended connect, did talks * Finished toolchain 2014.09 binary release * Built Android toolchain 2014.09 * Built NDK 2014.09 * Got Android to build and boot to text mode+splash screen with clang * Found and fixed a couple of AOSP bugs * Worked on bug 667 == Plans == * Recover from Connect ;) * Upstream AOSP+clang patches * Work on getting AOSP+clang to boot to UI * Start builds of AOSP with gcc 5.0 snapshots * Start replacing things cbuild1 did for us that cbuild2 currently can't do == Issues == * outside of still waiting for 48h days, none.

11 years, 1 month

[Activity] Week 37

by Dileep Katta

== Progress == * Enabling UEFI Fastboot on Panda (CARD-198: 0 %) - On Hold * CARD 1103: Flash support on u-boot fastboot (CARD-1103: 100%) - submitted v3 - worked on wiki == Plans == * connect, LCU14 (80%) * CARD 1103: fastboot flash command (CARD-1103: 20%) - work on comments from Wolfgang == Issues == *None

11 years, 1 month

Ubuntu Utopic Unicorn test rebuild (main, all architectures)

by Matthias Klose

A test rebuild of Utopic Unicorn was started this week for all architectures, and is finished for amd64, i386, armhf and ppc64el. The rebuild for arm64 and powerpc will finish this week. Results can be found at http://people.ubuntuwire.org/~wgrant/rebuild-ftbfs-test/test-rebuild-201409… The archive for the test rebuild is https://launchpad.net/ubuntu/+archive/test-rebuild-20140914 Additional build failures for packages in utopic-proposed (not yet in utopic) can be found at http://qa.ubuntuwire.com/ftbfs/ Please help fixing the build failures for the final release. Matthias

11 years, 1 month

[ACTIVITY] Week 37

by Andrey Konovalov

== Progress == * vacations (20%) * llct and linux-linaro trees have been updated to v3.17-rc4. The both are now using the Android topic by Amit P. 2014.09 linux-linaro kernel release notes are done (CARD-788: 80%) == Plans == * Linaro Connect

11 years, 1 month

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

linaro-dev