linaro-dev April 2014

linaro-dev@lists.linaro.org

26 participants
18 discussions

[PATCH] Produce a situation where L1 cache flush missing vs cpu power down hangs the system

by Daniel Lezcano

As pointed by Lorenzo, when a cpu powers down, the L1 cache must be flushed before, otherwise: * data cachelines are not empty and the other cpu may fetch data * cpu will lost some data leading to a memory corruption Note this bug is very difficult to reproduce and this test will not spot the issue everytime. Signed-off-by: Daniel Lezcano <daniel.lezcano(a)linaro.org> --- cpuidle/cpuidle-l1.c | 72 ++++++++++++++++++++++++++++++++++++++++++++++++ cpuidle/cpuidle_05.sh | 42 ++++++++++++++++++++++++++++ cpuidle/cpuidle_05.txt | 1 + 3 files changed, 115 insertions(+) create mode 100644 cpuidle/cpuidle-l1.c create mode 100755 cpuidle/cpuidle_05.sh create mode 100644 cpuidle/cpuidle_05.txt diff --git a/cpuidle/cpuidle-l1.c b/cpuidle/cpuidle-l1.c new file mode 100644 index 0000000..bbcde28 --- /dev/null +++ b/cpuidle/cpuidle-l1.c @@ -0,0 +1,72 @@ +#include <stdio.h> +#include <stdlib.h> +#include <sys/types.h> +#include <sys/wait.h> +#include <unistd.h> +#include <pthread.h> +#define BUFSIZE (4*1024) +#define DEADBEEF 0xDEADBEEF +static int buffer[BUFSIZE]; + +static pthread_t threads[64]; + +void *thread_routine(void *arg) +{ + int i, display = *(int *)arg; + int dummy; + + for (i = 0; i < 100; i++) { + + int j; + + for (j = 0; j < BUFSIZE * 1000; j++) { + dummy = buffer[j % BUFSIZE]; + dummy++; + } + + usleep(200000); + + if (buffer[i] != DEADBEEF) { + fprintf(stderr, "memory corruption\n"); + return (void *)-1; + } + + if (display == 0) + printf("%d%%%s", i, i < 10 ? "\b\b" : "\b\b\b"); + } + + if (display == 0) + printf(" \b\b\b\b"); + + return NULL; +} + +int main(int argc, char *argv[]) +{ + int i, ret = 0; + int nrcpus = sysconf(_SC_NPROCESSORS_ONLN); + + for (i = 0; i < BUFSIZE; i++) + buffer[i] = DEADBEEF; + + setbuf(stdout, NULL); + + for(i = 0; i < nrcpus; i++) { + + if (pthread_create(&threads[i], NULL, thread_routine, &i)) { + perror("pthread_create"); + return 1; + } + + } + + for (i = 0; i < nrcpus; i++) { + void *result; + pthread_join(threads[i], &result); + + if (result == (void *)-1) + ret = 1; + } + + return ret; +} diff --git a/cpuidle/cpuidle_05.sh b/cpuidle/cpuidle_05.sh new file mode 100755 index 0000000..679439d --- /dev/null +++ b/cpuidle/cpuidle_05.sh @@ -0,0 +1,42 @@ +#!/bin/bash +# +# PM-QA validation test suite for the power management on Linux +# +# Copyright (C) 2011, Linaro Limited. +# +# 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. +# +# You should have received a copy of the GNU General Public License +# along with this program; if not, write to the Free Software +# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. +# +# Contributors: +# Daniel Lezcano <daniel.lezcano(a)linaro.org> (IBM Corporation) +# - initial API and implementation +# + +# URL : https://wiki.linaro.org/WorkingGroups/PowerManagement/Resources/TestSuite/P… + +source ../include/functions.sh + +CPUIDLE_L1=./cpuidle-l1 + +if [ $(id -u) -ne 0 ]; then + log_skip "run as non-root" + exit 0 +fi + +check_cpuidle_l1() { + check "Fill L1 cache and sleep" "./$CPUIDLE_L1" +} + +check_cpuidle_l1 +test_status_show diff --git a/cpuidle/cpuidle_05.txt b/cpuidle/cpuidle_05.txt new file mode 100644 index 0000000..1f80e36 --- /dev/null +++ b/cpuidle/cpuidle_05.txt @@ -0,0 +1 @@ +Run cpuidle L1 test program to catch L1 flush missing vs cpu power down -- 1.7.9.5

10 years, 8 months

[dora][PATCH] eglibc: accept GNU make version 4 or later (backport from 2.19)

by Denys Dmytriyenko

From: Denys Dmytriyenko <denys(a)ti.com> Matches following commits from OE-Core: 9d59ceab6e58b422caf8ad7a306ed546316d8c3a bdc49826edf05233c6d506de404861a661043b1b Signed-off-by: Denys Dmytriyenko <denys(a)ti.com> --- This is only applicable to dora, since master already has eglibc-2.19 meta-linaro-toolchain/recipes-core/eglibc/eglibc_2.17.bbappend | 3 ++- meta-linaro-toolchain/recipes-core/eglibc/eglibc_2.18.bbappend | 1 + 2 files changed, 3 insertions(+), 1 deletion(-) diff --git a/meta-linaro-toolchain/recipes-core/eglibc/eglibc_2.17.bbappend b/meta-linaro-toolchain/recipes-core/eglibc/eglibc_2.17.bbappend index 9317ca1..14f28c8 100644 --- a/meta-linaro-toolchain/recipes-core/eglibc/eglibc_2.17.bbappend +++ b/meta-linaro-toolchain/recipes-core/eglibc/eglibc_2.17.bbappend @@ -1,6 +1,6 @@ MMYY = "13.07" RELEASE = "20${MMYY}-2" -PR = "r5" +PR = "r6" SRC_URI = "http://releases.linaro.org/${MMYY}/components/toolchain/eglibc-linaro/eglib… \ file://eglibc-svn-arm-lowlevellock-include-tls.patch \ @@ -26,6 +26,7 @@ SRC_URI = "http://releases.linaro.org/${MMYY}/components/toolchain/eglibc-linaro file://tzselect-awk.patch \ file://0001-eglibc-run-libm-err-tab.pl-with-specific-dirs-in-S.patch \ file://fix-tibetian-locales.patch \ + file://make-4.patch \ " SRC_URI[md5sum] = "6968b95121db1cffd147e07d6c55bffd" diff --git a/meta-linaro-toolchain/recipes-core/eglibc/eglibc_2.18.bbappend b/meta-linaro-toolchain/recipes-core/eglibc/eglibc_2.18.bbappend index d4f7db4..e6c9fa5 100644 --- a/meta-linaro-toolchain/recipes-core/eglibc/eglibc_2.18.bbappend +++ b/meta-linaro-toolchain/recipes-core/eglibc/eglibc_2.18.bbappend @@ -27,6 +27,7 @@ SRC_URI = "http://releases.linaro.org/${MMYY}/components/toolchain/eglibc-linaro file://0001-eglibc-run-libm-err-tab.pl-with-specific-dirs-in-S.patch \ file://fix-tibetian-locales.patch \ file://0001-ARM-Pass-dl_hwcap-to-IFUNC-resolver.patch \ + file://make-4.patch \ " SRC_URI[md5sum] = "3effe2d563db9dcf6f789db647d1b03d" -- 1.9.2

10 years, 8 months

Arndale 5250 crash when unplugging CPU1

by Daniel Lezcano

Hi all, I used the image: arndale-saucy_server_20140222-611.img I boots well but when I unplug CPU1, the kernel crashes. It seems related to the AFTR state (an idle state of cpuidle) and the u-boot version which is 2013-rc1. It seems this issue does not happen with a more recent u-boot (I was not yet able to test it). I don't know where to fill the bug for this release. Chander, any pointer to the right u-boot version you are using successfully ? 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

10 years, 8 months

[PATCH] idlestat: Add Makefile CROSS_COMPILE support

by Chris Muller

From: Christian Muller <christian.muller(a)linaro.org> Launchpad Bug #1306587: Makefile doesn't use CROSS_COMPILE flag. Add the CROSS_COMPILE flag support to allow cross compile and build of idlestat for other targets. Signed-off-by: Christian Muller <christian.muller(a)linaro.org> --- Makefile | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/Makefile b/Makefile index 5838e0f..f74b48c 100644 --- a/Makefile +++ b/Makefile @@ -22,14 +22,17 @@ # Zoran Markovic <zoran.markovic(a)linaro.org> # CFLAGS?=-g -Wall -CC?=gcc +CC=gcc OBJS = idlestat.o topology.o trace.o utils.o default: idlestat +%.o: %.c + $(CROSS_COMPILE)$(CC) -c -o $@ $< $(CFLAGS) + idlestat: $(OBJS) - $(CC) ${CFLAGS} $(OBJS) -o $@ + $(CROSS_COMPILE)$(CC) $(CFLAGS) $(OBJS) -o $@ clean: rm -f $(OBJS) idlestat -- 1.8.1.2

10 years, 8 months

Google hangout not working in debian sid/testing

by Nicolas Dechesne

hi, first of all, this isn't meant to be a distro trolling discussion, there are better places for that ;-) several days ago google HO started to fail after updating my debian/sid setup. as discussed on G+ [1] and on googlegroups [2], that seemed to be impacting anyone on debian/sid. when i realized that it was working on debian/testing (Jessie) i decided just use 'testing'... however since last evening update on 'testing', it started to fail too.. leaving me without any working HO at all, which is quite inconvenient at Linaro.. after bisecting my system using snapshot.debian.org, i figured out which package was the culprit: # apt-cache policy libcairo2 libcairo2: Installed: 1.12.14-4 Candidate: 1.12.14-4 Package pin: 1.12.14-4 Version table: 1.12.16-2 1001 500 http://ftp.fr.debian.org/debian/ testing/main amd64 Packages *** 1.12.14-4 1001 500 http://snapshot.debian.org/archive/debian/20130927T214600Z/testing/main amd64 Packages 100 /var/lib/dpkg/status Simply downgrading libcairo2 [3] 'fixes' the problem. I know it's not 'fixing' it really, but i hope that this will be enough information so that the real fix is uploaded. In the mean time, if any of use use Debian , you can use that as a workaround.. cheers nico [1] https://plus.google.com/u/0/108790110407014289037/posts/5q22d1ReavG [2] https://productforums.google.com/d/msg/hangouts/vYsaeEnXJXs/hOy2HaW5LKoJ [3] http://snapshot.debian.org/archive/debian/20130927T214600Z/pool/main/c/cair…

10 years, 8 months

UEFI for QEMU/KVM v0.1 releases

by Michael Casadevall

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 I'm pleased to announced that the first version of UEFI for QEMU/KVM has bee released, and is available for public consumption. There's still a fair number of bugs in this release, but feel free to grab it, play with it, or try and break it The (sparse) documentation is available here https://wiki.linaro.org/LEG/UEFIforQEMU Michael -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.14 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQEcBAEBAgAGBQJTQvQmAAoJEGKVRgSEnX1QrgYH/iNJAwyPX3noTuqo/695yhs/ DlHVxfjMi3Woec0laIm8dPXUxWpI8ObQnAXqVoj3sLcu8x8lZyyH/qXWdS2XiVYa NxNP7B4z8sQ2oKTBw9AH/lgNrfjq9kJwrFQdZnGnU5bBFiF52c4huVqv6kH0Uzsu k3QpYWqLVJ+0WQvH9t/TjdEI2xaLw1r9vAbp3IIv8nydHRNUOaujSO2qYNQDqpu6 ydxp8FqF++jyW7Yy2MbcAk+s32ODxhMHQUHQ1CcucrxI0ezOBmBfD1zOgaOtTgEA nlRAwId8rK5NpduLVs4LqUI9I0a4Oe9kvNv4WNyj/9N1LsGdsqKuh0pPb2rLstI= =sI5b -----END PGP SIGNATURE-----

10 years, 8 months

ARM Trusted Firmware and UEFI for FVP models

by Ryan Harkin

Hello all, Just so you know, both ARM Trusted Firmware (TF) and UEFI have changed. The 13.03 release has TF code that is pre v0.3. Since v0.3, TF expects to load all other bootloader stages from a "Firmware Image Package" (aka FIP). This includes UEFI. Also, the upstream UEFI code for FVP has changed. It is no longer built to run from NOR flash at address zero, but instead, it runs from RAM. This means that you cannot use the lastest TF with an old UEFI binary and you cannot use the latest UEFI binary with the old TF. I've updated the CI job for FVP's TF so that it now builds UEFI, then builds TF, then build the FIP blob with the UEFI binary inside it. Currently the job builds from the release-prep branch of linaro-edk2.git. This actually contains the upstream Tianocore code at the moment, so it will be missing a lot of the features you might want, like ethernet, or Linaro's local-FDT mods. I'll run another job after release-prep has been updated. And I'll have to update the hwpack to include fip.bin and so on. You can find the CI job here: https://ci.linaro.org/jenkins/view/components/job/fvp-pre-boot/ With the output binaries here: http://snapshots.linaro.org/components/kernel/fvp-pre-boot/38 And you can find the script here: http://git.linaro.org/ci/fvp-pre-boot.git The basic commands to build it all distil down to something like: git clone git://git.linaro.org/uefi/linaro-edk2.git cd linaro-edk2 uefi-build.sh fvp_minimal export BL33=`pwd`/`find . -name FVP_AARCH64_EFI.fd` cd .. git clone git://git.linaro.org/arm/models/arm-trusted-firmware.git cd arm-trusted-firmware CROSS_COMPILE=aarch64-linux-gnu- make PLAT=${platform} all fip Then you will have bl1.bin and fip.bin in build/fvp/release/. When you launch the FVP model, you specify the bl1.bin as normal. However, instead of passing the UEFI binary to the NOR Flash arg, you pass fip.bin, eg: -C bp.flashloader0.fname=<path to>/fip.bin Hope that all helps. Cheers, Ryan.

10 years, 8 months

TianoCore KVM Port: Status Update

by Michael Casadevall

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 As I've made a fair bit of headway since LinaroConnect, I wanted to drop a line on my current progress with porting TianoCore to KVM Summary (tl;dr version): KVM can start TianoCore, and boot all the way to shell, and access HDDs via VirtioBlk. We can start grub and successfully retrieve files from ext partitions, load a device tree, and start the kernel. The kernel runs through most of the EFI stub, but falls over during ExitBootServices() Long Version: So, after much blood sweat and tears, we're finally at the point of trying to actually start a kernel, though this (for the moment) remains an elusive goal. The current problem is that once we call EBS(), we get an exception from EFI with no Image information, which means the exception handler doesn't know where it came from. After several seconds, we get a second exception from within DxeCore, and then EFI falls over. Debugging EFI is difficult and error prone, combined with limited debug facilities from the gdb-stub in QEMU (no breakpoints), and no decent way to load all of EFI itself (you have to run add-symbol-file manually with the output of commands printed on the console; supposedly its possible to generate a giant GdbSyms.dll file to import in a single go, but I haven't succeeded at this). This is further complicated that it appears we're asserting somewhere in a driver, and short of adding printfs to *every* driver, its impossible to know which is asseting. Previous attempts to debug assets shows that EFI does "odd" things to the stack when we hit an exception, making walking it with GDB impossible. I need to figure out what madness EFI does with my SP so I can get the entire stack on an explosion, but this remains at best hopeful thinking. Further complicating things is that during EBS, my print debugging goes away. I might just cheat and roll a simple assembly function to bang out messages through serial without calling anything else. Ugly as sin, but this should let me get useful debug output through the EBS framework. Complicating matters is that I need to locate each and all EBS() event functions, which are spread *everywhere* in TianoCore, and then debug them each individually. I'm open to ideas on how best to accomplish this. On a larger scale, there are a couple of other bugs and odds and ends which currently affect us: * wfi doesn't work THis is probably the biggest w.r.t. to functionality that should work, but doesn't. The EFI event loop is built on checking the timer, then calling wfi to check the timer later. The problem here is we call wfi ... and UEFI never comes back despite events firing (I can put print code in the interrupt handler to confirm this). This may be related to the VGIC errors I get running kvm under foundation, but haven't taken the time to properly nail down the bug here. This was worked around by commenting out the wfi, turning event loop into a busy loop, but this has to be resolved before we can ever consider merging it * No RTC I looked through virt.c in KVM, and as best I can tell, I've got no RTC at all (no PL031). It also appears that the kernel can't get RTC as running a kernel gets me a 1970 clock. I'm not sure if this is by design or not, but it causes GetTime() to return EFI_ERROR, and I suspect may be one of the exceptions I'm getting avoid (Shell prints a ton of warnings that GetTime is busted). * No terminfo support (not ARM specific) EFI assumes its working against a real console or terminal. As such, it doesn't respect anything like termcap or such, and output gets jumbled due to incorrect escape sequencies and such. This isn't specific to ARM as I get identical behavior with OVMF running on stdio or over telnet. Backspace for instance requires me to type ^H into the console manually As we expect to have this usable remotely, we need to determine how to handle the terminal and escape sequences properly so we're not printing garbage to the screen. A "dumb" mode might just be the best way to handle this, but something like grub do display fancy graphics. Ideas welcome. If anyone wants to play with it, my code is on git, but requires a bit of setup to get running, and you need a patch to KVM to successfully start UEFI at the moment. Michael -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.14 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQEcBAEBAgAGBQJTNTJTAAoJEGKVRgSEnX1Qm2YH/jzWQjB/DPSHdv33hUlxVuIm j2AGElid7HM+3gp5U/ILdXvCyP9LFg8RATsvUa7bSHdwFZxRdAqYUWzm/Jxg6ZYV kyTzTDuPhbXp8koVLF14H/oGlpwnEH5rhzNa3UeRZTNYnNp3UMeZ1NXgGu9ziCfk hMGGteTMfV/6fGD6AETtlpD99KKwFiprD198QAil2rpVRi4WZBcgp7Y9/c/+cLfy sJD7p+zN8jBh9r9njpxh0o5rrQoG3N1NAemPGbSAc+qMMYCDOmjasGLwYiJuuMDy apRBM9+pbYsEgBGxM1z3XGywBJWBaLGzizk8vfPN2kN70/F7RW89X1rqG/95Ais= =OwC0 -----END PGP SIGNATURE-----

10 years, 8 months

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

linaro-dev April 2014