The ahead of 3 patches removed the load_idx of cpu_load.
The 4th patch fixed a incorrect cpu_load assign for nohz_full
performance testing result is comming soon.
Any comments are appreciated!
Regards
Alex
Hi,
Here is version 4 of fix to armv7-m build failure in
sigreturn_codes.S. The difference with version 3 is
fixed formating and renamed internally used macro
as per Dave's comments on [3].
Solution is based on .org directive Dave's suggestion
on last email in [1].
It uses conditional compilation and it uses .org
directive to keep sigreturn_codes layout.
Note I did not use ARM and THUMB macros because
those switch between CONFIG_THUMB2_KERNEL and not.
On v7a kernel we need both arm and thumb snipets
regardless of CONFIG_THUMB2_KERNEL setting.
And conditional compilation only kicks in with
CONFIG_CPU_THUMBONLY, for that local ARM_INSTR
macro is created.
Version 3 [3] conditional compilation and .org directive.
Pretty much the same code but incorrectly formated. Please
look at [3] for details.
Version 2 [2] tried to use '.acrh armv4t' directive
to allow both arm and thumb2 opcodes, but solution
deemed to be too fragile.
Version 1 [1] used conditional compilation and added
thumb2 nop instructions in CONFIG_CPU_THUMBONLY
Fix was tested
linux-next with efm32_defconfig build (along with few other fixes)
rmk-next BE/LE vexpress build/boot and LTP rt_sigaction0? tests run
Uwe tested version 1 and 3 with efm32
Thanks,
Victor
[1] http://lists.infradead.org/pipermail/linux-arm-kernel/2013-November/210393.…
[2] http://lists.infradead.org/pipermail/linux-arm-kernel/2013-November/210949.…
[3] http://lists.infradead.org/pipermail/linux-arm-kernel/2013-November/212307.…
Victor Kamensky (1):
ARM: signal: fix armv7-m build issue in sigreturn_codes.S
arch/arm/kernel/sigreturn_codes.S | 40 ++++++++++++++++++++++++++++++---------
1 file changed, 31 insertions(+), 9 deletions(-)
--
1.8.1.4
Hi Rafael,
These are V2 of my CPUIdle cleanup series.. Few patches are dropped as they
required further modifications. Last one is rewritten as suggested by Daniel.
Most of them are already Acked by Daniel.
Viresh Kumar (16):
cpuidle: fix indentation of cpumask
cpuidle: Fix comments in cpuidle core
cpuidle: make __cpuidle_get_cpu_driver() inline
cpuidle: make __cpuidle_device_init() return void
cpuidle: make __cpuidle_driver_init() return void
cpuidle: rearrange code in __cpuidle_driver_init()
cpuidle: rearrange __cpuidle_register_device() to keep minimal exit
points
cpuidle: merge two if() statements for checking error cases
cpuidle: reduce code duplication inside cpuidle_idle_call()
cpuidle: replace multiline statements with single line in
cpuidle_idle_call()
cpuidle: call cpuidle_get_driver() from after taking
cpuidle_driver_lock
cpuidle: use drv instead of cpuidle_driver in show_current_driver()
cpuidle: free all state kobjects from cpuidle_free_state_kobj()
cpuidle: don't calculate time-diff if entered_state < 0
cpuidle: don't call poll_idle_init() for every cpu
cpuidle: remove cpuidle_unregister_governor()
Documentation/cpuidle/governor.txt | 1 -
drivers/cpuidle/coupled.c | 2 +-
drivers/cpuidle/cpuidle.c | 95 ++++++++++----------------------------
drivers/cpuidle/driver.c | 69 ++++++++++++++++++++-------
drivers/cpuidle/governor.c | 43 -----------------
drivers/cpuidle/sysfs.c | 30 ++++++------
include/linux/cpuidle.h | 8 +---
7 files changed, 94 insertions(+), 154 deletions(-)
--
1.7.12.rc2.18.g61b472e
This script bisects kernel for build/boot/func regressions on
appointed board, run it under target kernel source tree.
If kernel friendly for network, it is a fully automatic bisect
script, otherwise it is a semi-automatic scripts.
It will be better if has some way to automatic hardware reset board
power. Or has some way send command via serial.
Usage:
$./bisect.sh -g v3.11~2 -b v3.11 -t func 2>&1 | tee func.log
Any comments are appreciated!
--
Thanks
Alex