infra
/
linux-spica-e5-2600v4
1
1
Fork 0
Code Releases Activity

initial commit

This commit is contained in:
Sangbum Kim 2017-04-05 01:14:59 +09:00
commit a3fbee018e
48 changed files with 87784 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

47
.gitignore vendored Normal file
View File

@ -0,0 +1,47 @@
# Created by https://www.gitignore.io/api/rchlinuxpackages,archives,linux
#!! ERROR: rchlinuxpackages is undefined. Use list command to see defined gitignore types !!#
### Archives ###
# It's better to unpack these files and commit the raw source because
# git has its own built in compression methods.
*.7z
*.jar
*.rar
*.zip
*.gz
*.bzip
*.bz2
*.xz
*.lzma
*.cab
#packing-only formats
*.iso
*.tar
#package management formats
*.dmg
*.xpi
*.gem
*.egg
*.deb
*.rpm
*.msi
*.msm
*.msp
### Linux ###
*~
# temporary files which can be created if a process still has a handle open of a deleted file
.fuse_hidden*
# KDE directory preferences
.directory
# Linux trash folder which might appear on any partition or disk
.Trash-*

556
PKGBUILD Normal file
View File

@ -0,0 +1,556 @@
# Maintainer: xduugu
_pkgext=-spica
pkgbase=linux$_pkgext
pkgname=$pkgbase
# required by AUR
# comment the following line to build a single package containing the kernel and the headers
(( 1 )) && pkgname=("$pkgbase" "$pkgbase-headers" "$pkgbase-docs")
pkgdesc="The Linux Kernel and modules from Linus' git tree"
depends=('coreutils' 'linux-firmware' 'module-init-tools' 'mkinitcpio')
pkgver=4.11.rc5
pkgrel=1
url="http://www.kernel.org/"
arch=(i686 x86_64)
license=('GPL2')
makedepends=(git bc)
options=(!strip)
source=($pkgname::git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git)
#source=($pkgname::git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git)
md5sums=('SKIP')
sha256sums=('SKIP')
# set _gitrev to a git revision (man gitrevisions) like a tag, a commit sha1
# hash or a branch name to build from this tree instead of master
_gitrev="v4.10.8"
####################################################################
# KERNEL CONFIG FILES
#
# This PKGBUILD searches for config files in the current directory
# and will use the first one it finds from the following
# list as base configuration:
# config.local
# config.saved.$CARCH
# config.$CARCH
#
####################################################################
#############################################################
# PATCHES
#
# This package builds the vanilla git kernel by default,
# but it is possible to patch the source without modifying
# this PKGBUILD.
#
# Simply create a directory 'patches' in your PKGBUILD
# directory and _any_ file (dotfiles excluded) in this
# folder will be applied to the kernel source.
#
# Prefixing the patch file names with dots will obviously
# excluded them from the patching process.
#
#############################################################
#############################
# CONFIGURATION
#
# Uncomment desired options
#############################
#_make_modules=1
#######
# Skip the merge of Linus's kernel tree
#
# _skip_merge=1
MAKEFLAGS="-j $(expr $(cat /proc/cpuinfo |grep processor |wc -l) \* 2)"
#######
# Set to e.g. menuconfig, xconfig or gconfig
#
# For a full list of supported commands, please have a look
# at "Configuration targets" section of `make help`'s output
# or the help target in scripts/kconfig/Makefile
# http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=blob;f=scripts/kconfig/Makefile
#
# If unset or set to an empty or space-only string, the
# (manual) kernel configuration step will be skipped.
#
_config_cmd="${_config_cmd:-menuconfig}"
#_config_cmd="${_config_cmd:-oldconfig}"
#######
# Stop build process after kernel configuration
#
# This option enables _save_config implicitly.
#
# _configure_only=1
#######
# The directory where the kernel should be built
#
# Can be useful, for example, if you want to compile on a
# tmpfs mount, which can speed up the compilation process
#
#_build_dir="${_build_dir:-$srcdir}"
#######
# Append the date to the localversion
#
# e.g. -ARCH -> -ARCH-20090422
#
#_date_localversion=0
#######
# Set the pkgver to the kernel version
# rather than the build date
#
# _kernel_pkgver=1
#######
# Save the .config file to package directory
# as config.saved.$CARCH
#
_save_config=1
#######
# Do not compress kernel modules
#
_no_modules_compression=0
#######
# Make the kernel build process verbose
#
# _verbose=1
# internal variables
(( 1 )) && _kernel_src="$pkgname"
#(( 1 )) && _kernel_src="$BUILDDIR/$(find . -maxdepth 1 -type d -name "linux-*" -printf "%f\n" | head -1)"
#(( 1 )) && _kernel_src="$_build_dir/$pkgname_$"
#######
# define required functions
pkgver() {
cd "$_kernel_src"
git describe --always | sed 's/^v//;s/-/./g'
}
# single package
package() {
eval package_$pkgbase-headers
eval package_$pkgbase
}
# split package functions
eval "package_$pkgbase() { _generic_package_kernel; }"
eval "package_$pkgbase-headers() { _generic_package_kernel-headers; }"
eval "package_$pkgbase-docs() { _generic_package_kernel-docs; }"
##############################
# where the magic happens...
##############################
build() {
cd "$_kernel_src"
msg "Sanitizing source tree.."
[[ -n $_gitrev ]] && git reset --hard "$_gitrev"
# cleaning source trees
git clean -f
#################
# Apply patches
#################
msg "Applying patches..."
local i patches
for i in "${source[@]}"; do
i=${i##*/}
[[ $i =~ .*\.patch$ ]] && patches=("${patches[@]}" "$srcdir/$i")
[[ ${i%.*} =~ .*\.patch$ ]] && patches=("${patches[@]}" "$srcdir/${i%.*}")
done
shopt -s nullglob
for i in "${patches[@]}" "$startdir/patches/"*; do
msg2 "Applying ${i##*/}..."
patch -Np1 -i "$i" || (error "Applying ${i##*/} failed" && return 1)
done
shopt -u nullglob
#################
# CONFIGURATION
#################
#########################
# Loading configuration
#########################
msg "Loading configuration..."
for i in local "saved.$CARCH" "$CARCH"; do
if [[ -e $startdir/config.$i ]]; then
msg2 "Using kernel config file config.$i..."
cp -f "$startdir/config.$i" .config
break
fi
done
[[ ! -e .config ]] &&
warning "No suitable kernel config file was found. You'll have to configure the kernel from scratch."
###########################
# Start the configuration
###########################
msg "Updating configuration..."
yes "" | make config > /dev/null
if [[ -f "$startdir/config.saved.$CARCH" ]]; then
msg2 "migrating previous config..."
cp "$startdir/config.saved.$CARCH" .config
make oldconfig
else
msg2 "migrating default config..."
cp "$startdir/config.$CARCH" .config
make oldconfig
fi
if [[ -n ${_config_cmd// /} ]]; then
msg2 "Running make $_config_cmd..."
make $_config_cmd
else
warning "Unknown config command: $_config_cmd"
fi
##############################################
# Save the config file the package directory
##############################################
if [[ -n $_save_config || -n $_configure_only ]]; then
msg "Saving configuration..."
msg2 "Saving $_kernel_src/.config as $startdir/config.saved.$CARCH"
cp .config "$startdir/config.saved.$CARCH"
fi
#######################################
# Stop after configuration if desired
#######################################
if [[ -n $_configure_only ]]; then
rm -rf "$_kernel_src" "$srcdir" "$pkgdir"
return 1
fi
###############################
# Append date to localversion
###############################
if [[ -n $_date_localversion ]]; then
local _localversion="$(sed -rn 's/^CONFIG_LOCALVERSION="([^"]*)"$/\1/p' .config)"
[[ -n $_localversion ]] && msg2 "CONFIG_LOCALVERSION is set to: $_localversion"
# since this is a git package, the $pkgver is equal to $(date +%Y%m%d)
msg2 "Appending $pkgver to CONFIG_LOCALVERSION..."
sed -ri "s/^(CONFIG_LOCALVERSION=).*$/\1\"$_localversion-$pkgver\"/" .config
fi
#################
# BUILD PROCESS
#################
################################
# Build the kernel and modules
################################
msg "Building kernel and modules..."
if [[ -n $_make_modules ]]; then
make $MAKEFLAGS V="$_verbose" bzImage modules
else
make $MAKEFLAGS V="$_verbose" bzImage
fi
############
# CLEANUP
############
###################################
# Copy files from build directory
####################################
# if (( ! CLEANUP )) && [[ $_build_dir != $srcdir ]]; then
# msg "Saving $_kernel_src to $srcdir/${_kernel_src##*/}..."
# mv "$_kernel_src" "$srcdir"
# rm -rf "$_kernel_src"
# fi
}
_generic_package_initialization() {
cd "$_kernel_src"
_karch="x86"
######################
# Get kernel version
######################
_kernver=$(make kernelrelease)
_basekernel=${_kernver%%-*}
############################################################
# Use kernel version instead of the current date as pkgver
############################################################
if [[ -n $_kernel_pkgver ]]; then
pkgver=${_kernver//-/_}
msg "Setting pkgver to kernel version: $pkgver"
fi
}
_generic_package_kernel() {
pkgdesc="The Linux Kernel and modules from Linus' git tree"
depends=('coreutils' 'linux-firmware' 'module-init-tools' 'mkinitcpio')
backup=(etc/mkinitcpio.d/$pkgname.preset)
install=$pkgname.install
changelog=$pkgname.changelog
# set required variables
_generic_package_initialization
#############################################################
# Provide linux
# (probably someone wants to use this kernel exclusively?)
#############################################################
provides=("${provides[@]}" "linux=${_kernver//-/_}")
################
# INSTALLATION
################
#####################
# Install the image
#####################
msg "Installing kernel image..."
install -Dm644 arch/$_karch/boot/bzImage "$pkgdir/boot/vmlinuz-$pkgname"
##########################
# Install kernel modules
##########################
if [[ -n $_make_modules ]]; then
msg "Installing kernel modules..."
make INSTALL_MOD_PATH="$pkgdir" modules_install
[[ -z $_no_modules_compression ]] && find "$pkgdir" -name "*.ko"
-exec gzip -9 {} +
##################################
# Create important symlinks
##################################
msg "Creating important symlinks..."
# Create generic modules symlink
if [[ $_kernver != ${_basekernel}${_pkgext} ]]; then
cd "$pkgdir/lib/modules"
ln -s "$_kernver" "${_basekernel}${_pkgext}"
cd "$OLDPWD"
# remove header symlinks
cd "$pkgdir/lib/modules/$_kernver"
rm -rf source build
cd "$OLDPWD"
fi
fi
############################
# Install mkinitcpio files
############################
install -d "$pkgdir/etc/mkinitcpio.d"
msg "Generating $pkgname.preset..."
cat > "$pkgdir/etc/mkinitcpio.d/$pkgname.preset" <<EOF
# mkinitcpio preset file for $pkgname
ALL_config="/etc/mkinitcpio.conf"
ALL_kver="/boot/vmlinuz-$pkgname"
PRESETS=('default')
#default_config="/etc/mkinitcpio.conf"
default_image="/boot/initramfs-$pkgname.img"
COMPRESSION="lz4" # since kernel 2.6.34
EOF
msg "Generating $pkgname.kver..."
echo -e "# DO NOT EDIT THIS FILE\nALL_kver='$_kernver'" \
> "$pkgdir/etc/mkinitcpio.d/$pkgname.kver"
#######################
# Update install file
#######################
msg "Updating install file..."
sed -ri "s/^(pkgname=).*$/\1$pkgname/" "$startdir/$pkgname.install"
sed -ri "s/^(kernver=).*$/\1$_kernver/" "$startdir/$pkgname.install"
#######################
# Remove the firmware
#######################
# remove the firmware
rm -rf "${pkgdir}/lib/firmware"
if [[ -n $_make_modules ]]; then
# Now we call depmod...
depmod -b "${pkgdir}" -F System.map "${_kernver}"
# move module tree /lib -> /usr/lib
mkdir -p "${pkgdir}/usr"
mv "${pkgdir}/lib" "${pkgdir}/usr/"
fi
}
_generic_package_kernel-headers() {
pkgdesc="Header files and scripts for building modules for $pkgbase"
depends=("$pkgbase")
# set required variables
_generic_package_initialization
#############################################################
# Provide linux-headers
# (probably someone wants to use this kernel exclusively?)
#############################################################
provides=("${provides[@]}" "linux-headers=${_kernver//-/_}")
##############################
# Install fake kernel source
##############################
install -Dm644 Module.symvers "$pkgdir/usr/src/linux-$_kernver/Module.symvers"
install -Dm644 Makefile "$pkgdir/usr/src/linux-$_kernver/Makefile"
install -Dm644 kernel/Makefile "$pkgdir/usr/src/linux-$_kernver/kernel/Makefile"
install -Dm644 .config "$pkgdir/usr/lib/modules/$_kernver/.config"
#######################################################
# Install scripts directory and fix permissions on it
#######################################################
cp -a scripts "$pkgdir/usr/src/linux-$_kernver"
##########################
# Install header files
##########################
msg "Installing header files..."
for i in net/ipv4/netfilter/ipt_CLUSTERIP.c \
$(find include/ net/mac80211/ drivers/{md,media/video/} -iname "*.h") \
$(find include/config/ -type f) \
$(find . -name "Kconfig*")
do
mkdir -p "$pkgdir/usr/src/linux-$_kernver/${i%/*}"
cp -af "$i" "$pkgdir/usr/src/linux-$_kernver/$i"
done
# required by virtualbox and probably others
ln -s "../generated/autoconf.h" "$pkgdir/usr/src/linux-$_kernver/include/linux/"
########################################
# Install architecture dependent files
########################################
msg "Installing architecture files..."
mkdir -p "$pkgdir/usr/src/linux-$_kernver/arch/$_karch/kernel"
cp -a arch/$_karch/kernel/asm-offsets.s "$pkgdir/usr/src/linux-$_kernver/arch/$_karch/kernel/"
cp -a arch/$_karch/Makefile* "$pkgdir/usr/src/linux-$_kernver/arch/$_karch/"
cp -a arch/$_karch/configs "$pkgdir/usr/src/linux-$_kernver/arch/$_karch/"
# copy arch includes for external modules and fix the nVidia issue
mkdir -p "$pkgdir/usr/src/linux-$_kernver/arch/$_karch"
cp -a "arch/$_karch/include" "$pkgdir/usr/src/linux-$_kernver/arch/$_karch/"
# create a necessary symlink to the arch folder
cd "$pkgdir/usr/src/linux-$_kernver/arch"
if [[ $CARCH = "x86_64" ]]; then
ln -s $_karch x86_64
else
ln -s $_karch i386
fi
cd "$OLDPWD"
################################
# Remove unneeded architecures
################################
msg "Removing unneeded architectures..."
for i in "$pkgdir/usr/src/linux-$_kernver/arch/"*; do
[[ ${i##*/} =~ ($_karch|Kconfig) ]] || rm -rf "$i"
done
############################
# Remove .gitignore files
############################
msg "Removing .gitignore files from kernel source..."
find "$pkgdir/usr/src/linux-$_kernver/" -name ".gitignore" -delete
##################################
# Create important symlinks
##################################
msg "Creating important symlinks..."
# the build symlink needs to be relative
if [[ -n $_make_modules ]]; then
cd "$pkgdir/usr/lib/modules/$_kernver"
rm -rf source build
ln -s "/usr/src/linux-$_kernver" build
cd "$OLDPWD"
fi
if [[ $_kernver != ${_basekernver}${_pkgext} ]]; then
cd "$pkgdir/usr/src"
ln -s "linux-$_kernver" "linux-${_basekernel}${_pkgext}"
cd "$OLDPWD"
fi
}
_generic_package_kernel-docs() {
pkgdesc="Kernel hackers manual - HTML documentation that comes with the Linux kernel."
depends=("$pkgbase")
# set required variables
_generic_package_initialization
mkdir -p "$pkgdir/usr/src/linux-$_kernver"
cp -a Documentation "$pkgdir/usr/src/linux-$_kernver/"
}
# vim: set fenc=utf-8 ts=2 sw=2 noet:

3517
config.x86_64 Normal file
View File

File diff suppressed because it is too large Load Diff

163
linux-spica.changelog Normal file
View File

@ -0,0 +1,163 @@
2013-07-20
* added CONFIG_SND_HDA_I915=y to kernel config
* updated kernel config files to latest linux configs (3.10.1-1)
2013-04-15
* adjusted PKGBUILD to make use of pacman 4.1's vcs functionality
* updated kernel config files to latest linux configs (3.8.7-1)
2013-03-28
* added bc to makedepends
* updated kernel config files to latest linux configs (3.8.4-1)
2012-12-12
* updated kernel config files to latest linux configs (3.7-1)
2012-09-12
* added support for arch's kmod 10-1 package (thanks to gun26)
* updated kernel config files to latest linux configs (3.5.3-1)
2012-07-22
* unpatched kernels do not get the -dirty suffix appended to their version anymore
* updated kernel config files to latest linux configs (3.5.0-1)
2012-07-04
* updated kernel config files to latest linux configs (3.4.1-1)
* modules are now installed to /usr/lib instead of /lib
2012-03-19
* updated kernel config files to latest linux configs (3.3-1)
2012-01-05
* updated kernel config files to latest linux configs (3.2-1)
2011-11-05
* fixed kernel version in extramodules directory name (thanks to gun26)
2011-10-31
* removed System.map file (https://bugs.archlinux.org/task/25247)
* added extramodules directory (might not be very useful for git kernels,
but for stable ones)
* append pkgrel to kernel version for stable kernels
* updated kernel config files to latest linux configs (3.1-1)
2011-07-24
* new variable _gitrev that can be used to build a sepcific tag, commit or branch
* reworked mkinitcpio preset file
* removed kernel26{,-headers} from provides array because arch's linux
package conflicts with these
2011-07-23
* disabled file stripping
* renamed initramfs file from /boot/$pkgname.img to /boot/initramfs-$pkgname.img
* updated kernel config files to latest linux configs (3.0-1)
2011-07-22
* make sure the src directory is moved to $srcdir after finishing the build
step when the _build_dir config option is in use
* changed repository source url to linux.git (was linux-2.6.git) as a
result of the 3.0 release
2011-07-13
* fixed issue with mkinitcpio that led to broken images because the kernel
modules could not be found (solved by running depmod before mkinitcpio)
* removed kver file
* removed arch linux logo
* updated kernel config files to latest kernel26 configs (2.6.39.3-1)
2011-05-30
* renamed kernel26-git to linux-git
* worked around make 3.82 bug
2011-05-25
* make kernel26$_pkgext-{headers,docs} depend on $pkgbase and not on itself
2011-05-24
* updated kernel config files to latest kernel26 configs (2.6.39-1)
2011-01-31
* reworked kernel header package generation
* support for split (default) and single package kernels
* added -docs package
* honour MAKEFLAGS
2011-01-30
* compress kernel modules by default
* updated kernel config files to latest kernel26 configs (2.6.37-5)
* added crypto and xen headers
2010-10-05
* Kconfig file was unintentionally removed from /usr/src/linux-$kver/arch (thanks to Kariddi)
* updated kernel config files to latest kernel26 configs (2.6.35.7)
2010-08-18
* removed zc0301 from header copy list (driver was removed: 0d58cef664e01fb1848833455bfdbe1a3d91044c)
* updated kernel config files to latest kernel26 configs (2.6.35.2)
2010-06-17
* updated kernel config files to latest kernel26 configs (2.6.34)
* replaced kernel26-firmware-git dependency with linux-firmware-git
* introduced new variable _kernel_src that contains the directory where the
kernel is acutally built
2010-04-07
* make config options configurable via command line
* updated kernel config files to latest kernel26 configs (2.6.33.2)
* removed /boot/kconfig26$_pkgext
* replaced dynamic kernel version detection with a static string which is
updated at build time
* generate the *.preset file from the PKGBUILD
* introduced new variable _pkgext to make it easier to adapt this PKGBUILD
to other kernel sources
2010-01-12
* fixed abort when patches directory exists but is empty
2009-12-24
* included some new required kernel header files
* save build directory to $srcdir when building without makepkg's -c flag
and if $_build_dir != $srcdir
* fixed kernel version detection in install file (again... replaced `\s`
with `[:space:]`)
2009-12-22
* fixed kernel version detection in install file
* removed asm-$_karch references
* some code cleanup
2009-12-05
* added kernel26-headers to provides
2009-12-04
* updated kernel config files to latest kernel26 configs (2.6.32)
* reworked install scriptlet
2009-11-13
* added support for all available configuration commands of the kernel
* cleaned up the code
2009-10-25
* added changelog variable for pacman's next major release
* make use of new package function to reduce fakeroot usage
* added a warning about initial download size
2009-10-09
* removed .gitignore files from source tree
* updated kernel config files to latest kernel26 configs (2.6.31.3)
2009-09-10
* updated kernel config files to latest kernel26 configs (2.6.31)
2009-08-15
* added CONFIG_MEDIA_SUPPORT=m to config files
2009-08-14
* Since pacman 3.3.0, makepkg starts in $srcdir and not in $startdir
anymore. In order to get the kernel sources into $startdir again, it
is required to change to this directory
* updated kernel config files to latest kernel26 configs (2.6.30.4)
2009-06-11
* updated kernel config files to latest kernel26 configs (2.6.30)
2009-04-26
* added new configure option _build_dir
* moved the 'patches' folder from $srcdir to $startdir
* moved the git repository clone from $srcdir to $startdir
* made the linux-2.6 directory a bare repository, which reduces the folder
size by more than the half
* renamed cloned git repository from kernel26-git to linux-2.6.git
2009-04-24
* documented configure option _verbose which makes the kernel building
process verbose
* added a check for an existing package when using _kernel_pkgver to avoid
silent overwrites
* cleaned up the PKGBUILD
* remove the git changelog generation functionality
* replace dashes in kernel version with underscores instead of periods
* separate date from localversion by a dash when using _date_localversion
* provide kernel26
* some more documentation
2009-04-22
* use arch's kernel config files as default
* removed patches
* added possibility to patch the source without modifying the PKGBUILD
* started to make PKGBUILD more readable
* documented PKGBUILD configure options
* introduced new configure option '_configure_only' which stops the building
process after the kernel configuration
2009-04-21
* moved firmware from package into new dependency kernel26-firmware-git,
which makes it possible to install kernel26-git beside arch's stock kernel
* fetch latest git changes when there is already a local repository
* added some "|| return 1"
* kernel26-git.install: extract required kernel version from package files
rather than specify it explicitly
* renamed config to config.i686
* replaced $startdir with $srcdir and $pkgdir
* added quotes where necessary
* cleaned up the checkout part

37
linux-spica.install Normal file
View File

@ -0,0 +1,37 @@
pkgname=linux-spica
kernver=4.10.8spica-dirty
#bootdevice="BOOT_IMAGE=/boot/vmlinuz-$pkgname root=UUID=d670564f-2cb3-4981-9d51-6ed9c1327d47"
#option="rw quiet clocksource=hpet initrd=EFI/spi-ca/initrd intel_iommu=on pci-stub.ids=1002:683f,1002:aab0 vfio_iommu_type1.allow_unsafe_interrupts=1,kvm.ignore_msrs=1"
#option="rw quiet clocksource=hpet initrd=EFI/spi-ca/initrd quiet intremap=no_x2apic_optout zswap.enabled=1 zswap.max_pool_percent=25 zswap.compressor=lz4"
post_install () {
echo ">"
echo "> Updating module dependencies. Please wait ..."
depmod $kernver
echo ">"
echo "> Generating initramfs, using mkinitcpio. Please wait..."
echo ">"
mkinitcpio -p $pkgname
# echo "> Modifing efibootmgr..."
# efibootmgr|awk 'match($0,/^Boot([0-9a-fA-F]{4})\* spi-ca_v(.+)$/,m){printf "efibootmgr -b %s -B;echo \">> remove entry : %s\";",m[1],m[2]}'|sh
# echo "> Copy efistub from boot"
# cp -fv "boot/vmlinuz-$pkgname" "boot/efi/EFI/spi-ca/kernel.efi"
# cp -fv "boot/initramfs-$pkgname.img" "boot/efi/EFI/spi-ca/initrd"
# echo "> Registering efistub "
#echo 'efibootmgr -c -g -d /dev/sda -p 1 -L "spi-ca_v$kernver" -l "\EFI\spi-ca\kernel.efi" #-u "$bootdevice $option"'
# efibootmgr -c -g -d /dev/sde -p 1 -L "spi-ca_v$kernver" -l "\EFI\spi-ca\kernel" # -u "$bootdevice $option"
# echo "> Reordering Bootorder..."
# newentry=`efibootmgr|awk 'match($0,/^Boot([0-9a-fA-F]{4})\* spi-ca_v(.+)$/,m){print m[1]}'`
# prebootorder=`efibootmgr |grep BootOrder |cut -d : -f 2 |tr -d ' '`
# efibootmgr -O
# efibootmgr -o ${newentry},${prebootorder}
echo "> OK!"
}
post_upgrade() {
post_install
}
post_remove() {
rm -f -- "boot/initramfs-$pkgname.img"
}

103
patches-4.7/0001-block-cgroups-kconfig-build-bits-for-BFQ-v7r11-4.7.0.patch Normal file
View File

@ -0,0 +1,103 @@
From 22ee35ec82fa543b65c1b6d516a086a21f723846 Mon Sep 17 00:00:00 2001
From: Paolo Valente <paolo.valente@unimore.it>
Date: Tue, 7 Apr 2015 13:39:12 +0200
Subject: [PATCH 1/4] block: cgroups, kconfig, build bits for BFQ-v7r11-4.7.0
Update Kconfig.iosched and do the related Makefile changes to include
kernel configuration options for BFQ. Also increase the number of
policies supported by the blkio controller so that BFQ can add its
own.
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Signed-off-by: Arianna Avanzini <avanzini@google.com>
---
block/Kconfig.iosched | 32 ++++++++++++++++++++++++++++++++
block/Makefile | 1 +
include/linux/blkdev.h | 2 +-
3 files changed, 34 insertions(+), 1 deletion(-)
diff --git a/block/Kconfig.iosched b/block/Kconfig.iosched
index 421bef9..0ee5f0f 100644
--- a/block/Kconfig.iosched
+++ b/block/Kconfig.iosched
@@ -39,6 +39,27 @@ config CFQ_GROUP_IOSCHED
---help---
Enable group IO scheduling in CFQ.
+config IOSCHED_BFQ
+ tristate "BFQ I/O scheduler"
+ default n
+ ---help---
+ The BFQ I/O scheduler tries to distribute bandwidth among
+ all processes according to their weights.
+ It aims at distributing the bandwidth as desired, independently of
+ the disk parameters and with any workload. It also tries to
+ guarantee low latency to interactive and soft real-time
+ applications. If compiled built-in (saying Y here), BFQ can
+ be configured to support hierarchical scheduling.
+
+config CGROUP_BFQIO
+ bool "BFQ hierarchical scheduling support"
+ depends on CGROUPS && IOSCHED_BFQ=y
+ default n
+ ---help---
+ Enable hierarchical scheduling in BFQ, using the cgroups
+ filesystem interface. The name of the subsystem will be
+ bfqio.
+
choice
prompt "Default I/O scheduler"
default DEFAULT_CFQ
@@ -52,6 +73,16 @@ choice
config DEFAULT_CFQ
bool "CFQ" if IOSCHED_CFQ=y
+ config DEFAULT_BFQ
+ bool "BFQ" if IOSCHED_BFQ=y
+ help
+ Selects BFQ as the default I/O scheduler which will be
+ used by default for all block devices.
+ The BFQ I/O scheduler aims at distributing the bandwidth
+ as desired, independently of the disk parameters and with
+ any workload. It also tries to guarantee low latency to
+ interactive and soft real-time applications.
+
config DEFAULT_NOOP
bool "No-op"
@@ -61,6 +92,7 @@ config DEFAULT_IOSCHED
string
default "deadline" if DEFAULT_DEADLINE
default "cfq" if DEFAULT_CFQ
+ default "bfq" if DEFAULT_BFQ
default "noop" if DEFAULT_NOOP
endmenu
diff --git a/block/Makefile b/block/Makefile
index 9eda232..4a36683 100644
--- a/block/Makefile
+++ b/block/Makefile
@@ -18,6 +18,7 @@ obj-$(CONFIG_BLK_DEV_THROTTLING) += blk-throttle.o
obj-$(CONFIG_IOSCHED_NOOP) += noop-iosched.o
obj-$(CONFIG_IOSCHED_DEADLINE) += deadline-iosched.o
obj-$(CONFIG_IOSCHED_CFQ) += cfq-iosched.o
+obj-$(CONFIG_IOSCHED_BFQ) += bfq-iosched.o
obj-$(CONFIG_BLOCK_COMPAT) += compat_ioctl.o
obj-$(CONFIG_BLK_CMDLINE_PARSER) += cmdline-parser.o
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 3d9cf32..8d862a0 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -45,7 +45,7 @@ struct pr_ops;
* Maximum number of blkcg policies allowed to be registered concurrently.
* Defined here to simplify include dependency.
*/
-#define BLKCG_MAX_POLS 2
+#define BLKCG_MAX_POLS 3
struct request;
typedef void (rq_end_io_fn)(struct request *, int);
--
1.9.1

7097
patches-4.7/0002-block-introduce-the-BFQ-v7r11-I-O-sched-for-4.7.0.patch Normal file
View File

File diff suppressed because it is too large Load Diff

1101
patches-4.7/0003-block-bfq-add-Early-Queue-Merge-EQM-to-BFQ-v7r11-for.patch Normal file
View File

File diff suppressed because it is too large Load Diff

6492
patches-4.7/0004-block-bfq-turn-BFQ-v7r11-for-4.7.0-into-BFQ-v8r3-for.patch Normal file
View File

File diff suppressed because it is too large Load Diff

708
patches-4.7/BLD-4.7.patch Normal file
View File

@ -0,0 +1,708 @@
BLD changes for Linux kernel version 4.7
---
diff --git a/init/Kconfig b/init/Kconfig
index c02d897..edf8697 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -36,6 +36,15 @@ config BROKEN_ON_SMP
depends on BROKEN || !SMP
default y
+config BLD
+ bool "An alternate CPU load distribution technique for task scheduler"
+ depends on SMP
+ default y
+ help
+ This is an alternate CPU load distribution technique based for task
+ scheduler based on The Barbershop Load Distribution algorithm. Not
+ suitable for NUMA, should work well on SMP.
+
config INIT_ENV_ARG_LIMIT
int
default 32 if !UML
diff --git a/kernel/sched/bld.h b/kernel/sched/bld.h
new file mode 100644
index 0000000..f1f9fba
--- /dev/null
+++ b/kernel/sched/bld.h
@@ -0,0 +1,215 @@
+#ifdef CONFIG_BLD
+
+static DEFINE_RWLOCK(rt_list_lock);
+static LIST_HEAD(rt_rq_head);
+static LIST_HEAD(cfs_rq_head);
+static DEFINE_RWLOCK(cfs_list_lock);
+
+#ifdef CONFIG_FAIR_GROUP_SCHED
+static inline struct rq *rq_of_cfs(struct cfs_rq *cfs_rq)
+{
+ return cfs_rq->rq;
+}
+#else
+static inline struct rq *rq_of_cfs(struct cfs_rq *cfs_rq)
+{
+ return container_of(cfs_rq, struct rq, cfs);
+}
+#endif
+
+#ifdef CONFIG_RT_GROUP_SCHED
+static inline struct rq *rq_of_rt(struct rt_rq *rt_rq)
+{
+ return rt_rq->rq;
+}
+#else
+static inline struct rq *rq_of_rt(struct rt_rq *rt_rq)
+{
+ return container_of(rt_rq, struct rq, rt);
+}
+#endif
+
+static int select_cpu_for_wakeup(int task_type, struct cpumask *mask)
+{
+ int cpu = smp_processor_id(), i;
+ unsigned long load, varload;
+ struct rq *rq;
+
+ if (task_type) {
+ varload = ULONG_MAX;
+ for_each_cpu(i, mask) {
+ rq = cpu_rq(i);
+ load = rq->cfs.load.weight;
+ if (load < varload) {
+ varload = load;
+ cpu = i;
+ }
+ }
+ } else {
+ /* Here's an attempt to get a CPU within the mask where
+ * we can preempt easily. To achieve this we tried to
+ * maintain a lowbit, which indicate the lowest bit set on
+ * array bitmap. Since all CPUs contains high priority
+ * kernel threads therefore we eliminate 0, so it might not
+ * be right every time, but it's just an indicator.
+ */
+ varload = 1;
+
+ for_each_cpu(i, mask) {
+ rq = cpu_rq(i);
+ load = rq->rt.lowbit;
+ if (load >= varload) {
+ varload = load;
+ cpu = i;
+ }
+ }
+ }
+
+ return cpu;
+}
+
+static int bld_pick_cpu_cfs(struct task_struct *p, int sd_flags, int wake_flags)
+{
+ struct cfs_rq *cfs;
+ unsigned long flags;
+ unsigned int cpu = smp_processor_id();
+
+ read_lock_irqsave(&cfs_list_lock, flags);
+ list_for_each_entry(cfs, &cfs_rq_head, bld_cfs_list) {
+ cpu = cpu_of(rq_of_cfs(cfs));
+ if (cpu_online(cpu))
+ break;
+ }
+ read_unlock_irqrestore(&cfs_list_lock, flags);
+ return cpu;
+}
+
+static int bld_pick_cpu_rt(struct task_struct *p, int sd_flags, int wake_flags)
+{
+ struct rt_rq *rt;
+ unsigned long flags;
+ unsigned int cpu = smp_processor_id();
+
+ read_lock_irqsave(&rt_list_lock, flags);
+ list_for_each_entry(rt, &rt_rq_head, bld_rt_list) {
+ cpu = cpu_of(rq_of_rt(rt));
+ if (cpu_online(cpu))
+ break;
+ }
+ read_unlock_irqrestore(&rt_list_lock, flags);
+ return cpu;
+}
+
+static int bld_pick_cpu_domain(struct task_struct *p, int sd_flags, int wake_flags)
+{
+ unsigned int cpu = smp_processor_id(), want_affine = 0;
+ struct cpumask *tmpmask;
+
+ if (p->nr_cpus_allowed == 1)
+ return task_cpu(p);
+
+ if (sd_flags & SD_BALANCE_WAKE) {
+ if (cpumask_test_cpu(cpu, tsk_cpus_allowed(p))) {
+ want_affine = 1;
+ }
+ }
+
+ if (want_affine)
+ tmpmask = tsk_cpus_allowed(p);
+ else
+ tmpmask = sched_domain_span(cpu_rq(task_cpu(p))->sd);
+
+ if (rt_task(p))
+ cpu = select_cpu_for_wakeup(0, tmpmask);
+ else
+ cpu = select_cpu_for_wakeup(1, tmpmask);
+
+ return cpu;
+}
+
+static void track_load_rt(struct rq *rq, struct task_struct *p)
+{
+ unsigned long flag;
+ int firstbit;
+ struct rt_rq *first;
+ struct rt_prio_array *array = &rq->rt.active;
+
+ first = list_entry(rt_rq_head.next, struct rt_rq, bld_rt_list);
+ firstbit = sched_find_first_bit(array->bitmap);
+
+ /* Maintaining rt.lowbit */
+ if (firstbit > 0 && firstbit <= rq->rt.lowbit)
+ rq->rt.lowbit = firstbit;
+
+ if (rq->rt.lowbit < first->lowbit) {
+ write_lock_irqsave(&rt_list_lock, flag);
+ list_del(&rq->rt.bld_rt_list);
+ list_add_tail(&rq->rt.bld_rt_list, &rt_rq_head);
+ write_unlock_irqrestore(&rt_list_lock, flag);
+ }
+}
+
+static int bld_get_cpu(struct task_struct *p, int sd_flags, int wake_flags)
+{
+ unsigned int cpu;
+
+ if (sd_flags == SD_BALANCE_WAKE || (sd_flags == SD_BALANCE_EXEC && (get_nr_threads(p) > 1)))
+ cpu = bld_pick_cpu_domain(p, sd_flags, wake_flags);
+ else {
+ if (rt_task(p))
+ cpu = bld_pick_cpu_rt(p, sd_flags, wake_flags);
+ else
+ cpu = bld_pick_cpu_cfs(p, sd_flags, wake_flags);
+ }
+
+ return cpu;
+}
+
+static void bld_track_load_activate(struct rq *rq, struct task_struct *p)
+{
+ unsigned long flag;
+ if (rt_task(p)) {
+ track_load_rt(rq, p);
+ } else {
+ if (rq->cfs.pos != 2) {
+ struct cfs_rq *last;
+ last = list_entry(cfs_rq_head.prev, struct cfs_rq, bld_cfs_list);
+ if (rq->cfs.load.weight >= last->load.weight) {
+ write_lock_irqsave(&cfs_list_lock, flag);
+ list_del(&rq->cfs.bld_cfs_list);
+ list_add_tail(&rq->cfs.bld_cfs_list, &cfs_rq_head);
+ rq->cfs.pos = 2; last->pos = 1;
+ write_unlock_irqrestore(&cfs_list_lock, flag);
+ }
+ }
+ }
+}
+
+static void bld_track_load_deactivate(struct rq *rq, struct task_struct *p)
+{
+ unsigned long flag;
+ if (rt_task(p)) {
+ track_load_rt(rq, p);
+ } else {
+ if (rq->cfs.pos != 0) {
+ struct cfs_rq *first;
+ first = list_entry(cfs_rq_head.next, struct cfs_rq, bld_cfs_list);
+ if (rq->cfs.load.weight <= first->load.weight) {
+ write_lock_irqsave(&cfs_list_lock, flag);
+ list_del(&rq->cfs.bld_cfs_list);
+ list_add(&rq->cfs.bld_cfs_list, &cfs_rq_head);
+ rq->cfs.pos = 0; first->pos = 1;
+ write_unlock_irqrestore(&cfs_list_lock, flag);
+ }
+ }
+ }
+}
+#else
+static inline void bld_track_load_activate(struct rq *rq, struct task_struct *p)
+{
+}
+
+static inline void bld_track_load_deactivate(struct rq *rq, struct task_struct *p)
+{
+}
+#endif /* CONFIG_BLD */
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 97ee9ac..b2ddabc 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -24,6 +24,8 @@
* 2007-07-01 Group scheduling enhancements by Srivatsa Vaddagiri
* 2007-11-29 RT balancing improvements by Steven Rostedt, Gregory Haskins,
* Thomas Gleixner, Mike Kravetz
+ * 2012-Feb The Barbershop Load Distribution (BLD) algorithm - an alternate
+ * CPU load distribution technique for kernel scheduler by Rakib Mullick.
*/
#include <linux/kasan.h>
@@ -86,6 +88,7 @@
#include "sched.h"
#include "../workqueue_internal.h"
#include "../smpboot.h"
+#include "bld.h"
#define CREATE_TRACE_POINTS
#include <trace/events/sched.h>
@@ -750,6 +753,8 @@ static inline void enqueue_task(struct rq *rq, struct task_struct *p, int flags)
if (!(flags & ENQUEUE_RESTORE))
sched_info_queued(rq, p);
p->sched_class->enqueue_task(rq, p, flags);
+ if (!dl_task(p))
+ bld_track_load_activate(rq, p);
}
static inline void dequeue_task(struct rq *rq, struct task_struct *p, int flags)
@@ -758,6 +763,8 @@ static inline void dequeue_task(struct rq *rq, struct task_struct *p, int flags)
if (!(flags & DEQUEUE_SAVE))
sched_info_dequeued(rq, p);
p->sched_class->dequeue_task(rq, p, flags);
+ if (!dl_task(p))
+ bld_track_load_deactivate(rq, p);
}
void activate_task(struct rq *rq, struct task_struct *p, int flags)
@@ -1587,11 +1594,17 @@ int select_task_rq(struct task_struct *p, int cpu, int sd_flags, int wake_flags)
{
lockdep_assert_held(&p->pi_lock);
+#ifndef CONFIG_BLD
if (tsk_nr_cpus_allowed(p) > 1)
cpu = p->sched_class->select_task_rq(p, cpu, sd_flags, wake_flags);
else
cpu = cpumask_any(tsk_cpus_allowed(p));
-
+#else
+ if (dl_task(p))
+ cpu = dl_sched_class.select_task_rq(p, cpu, sd_flags, wake_flags);
+ else
+ cpu = bld_get_cpu(p, sd_flags, wake_flags);
+#endif
/*
* In order not to call set_task_cpu() on a blocking task we need
* to rely on ttwu() to place the task on a valid ->cpus_allowed
@@ -1794,7 +1807,11 @@ void scheduler_ipi(void)
*/
preempt_fold_need_resched();
+#ifndef CONFIG_BLD
if (llist_empty(&this_rq()->wake_list) && !got_nohz_idle_kick())
+#else
+ if (llist_empty(&this_rq()->wake_list))
+#endif
return;
/*
@@ -1816,13 +1833,16 @@ void scheduler_ipi(void)
/*
* Check if someone kicked us for doing the nohz idle load balance.
*/
+#ifndef CONFIG_BLD
if (unlikely(got_nohz_idle_kick())) {
this_rq()->idle_balance = 1;
raise_softirq_irqoff(SCHED_SOFTIRQ);
}
+#endif
irq_exit();
}
+#ifndef CONFIG_BLD
static void ttwu_queue_remote(struct task_struct *p, int cpu, int wake_flags)
{
struct rq *rq = cpu_rq(cpu);
@@ -1836,6 +1856,7 @@ static void ttwu_queue_remote(struct task_struct *p, int cpu, int wake_flags)
trace_sched_wake_idle_without_ipi(cpu);
}
}
+#endif /* CONFIG_BLD */
void wake_up_if_idle(int cpu)
{
@@ -1872,7 +1893,7 @@ static void ttwu_queue(struct task_struct *p, int cpu, int wake_flags)
struct rq *rq = cpu_rq(cpu);
struct pin_cookie cookie;
-#if defined(CONFIG_SMP)
+#if defined(CONFIG_SMP) && !defined(CONFIG_BLD)
if (sched_feat(TTWU_QUEUE) && !cpus_share_cache(smp_processor_id(), cpu)) {
sched_clock_cpu(cpu); /* sync clocks x-cpu */
ttwu_queue_remote(p, cpu, wake_flags);
@@ -2394,7 +2415,7 @@ int sched_fork(unsigned long clone_flags, struct task_struct *p)
* Silence PROVE_RCU.
*/
raw_spin_lock_irqsave(&p->pi_lock, flags);
- set_task_cpu(p, cpu);
+ __set_task_cpu(p, cpu);
raw_spin_unlock_irqrestore(&p->pi_lock, flags);
#ifdef CONFIG_SCHED_INFO
@@ -2941,7 +2962,14 @@ void sched_exec(void)
int dest_cpu;
raw_spin_lock_irqsave(&p->pi_lock, flags);
+#ifndef CONFIG_BLD
dest_cpu = p->sched_class->select_task_rq(p, task_cpu(p), SD_BALANCE_EXEC, 0);
+#else
+ if (dl_task(p))
+ dest_cpu = task_cpu(p);
+ else
+ dest_cpu = bld_get_cpu(p, SD_BALANCE_EXEC, 0);
+#endif
if (dest_cpu == smp_processor_id())
goto unlock;
@@ -3030,8 +3058,10 @@ void scheduler_tick(void)
#ifdef CONFIG_SMP
rq->idle_balance = idle_cpu(cpu);
+#ifndef CONFIG_BLD
trigger_load_balance(rq);
#endif
+#endif
rq_last_tick_reset(rq);
}
@@ -7262,7 +7292,9 @@ int sched_cpu_dying(unsigned int cpu)
raw_spin_unlock_irqrestore(&rq->lock, flags);
calc_load_migrate(rq);
update_max_interval();
+#ifndef CONFIG_BLD
nohz_balance_exit_idle(cpu);
+#endif
hrtick_clear(rq);
return 0;
}
@@ -7468,6 +7500,15 @@ void __init sched_init(void)
#endif /* CONFIG_SMP */
init_rq_hrtick(rq);
atomic_set(&rq->nr_iowait, 0);
+#ifdef CONFIG_BLD
+ INIT_LIST_HEAD(&rq->cfs.bld_cfs_list);
+ list_add_tail(&rq->cfs.bld_cfs_list, &cfs_rq_head);
+ rq->cfs.pos = 0;
+
+ INIT_LIST_HEAD(&rq->rt.bld_rt_list);
+ list_add_tail(&rq->rt.bld_rt_list, &rt_rq_head);
+ rq->rt.lowbit = INT_MAX;
+#endif
}
set_load_weight(&init_task);
@@ -7510,6 +7551,9 @@ void __init sched_init(void)
init_schedstats();
scheduler_running = 1;
+#ifdef CONFIG_BLD
+ printk(KERN_INFO "BLD: An Alternate CPU load distributor activated.\n");
+#endif
}
#ifdef CONFIG_DEBUG_ATOMIC_SLEEP
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index c8c5d2d..5b694b3 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -4880,6 +4880,7 @@ static unsigned long cpu_avg_load_per_task(int cpu)
return 0;
}
+#ifndef CONFIG_BLD
#ifdef CONFIG_FAIR_GROUP_SCHED
/*
* effective_load() calculates the load change as seen from the root_task_group
@@ -5411,6 +5412,7 @@ select_task_rq_fair(struct task_struct *p, int prev_cpu, int sd_flag, int wake_f
return new_cpu;
}
+#endif /* CONFIG_BLD */
/*
* Called immediately before a task is migrated to a new cpu; task_cpu(p) and
@@ -5741,6 +5743,7 @@ idle:
* further scheduler activity on it and we're being very careful to
* re-start the picking loop.
*/
+#ifndef CONFIG_BLD
lockdep_unpin_lock(&rq->lock, cookie);
new_tasks = idle_balance(rq);
lockdep_repin_lock(&rq->lock, cookie);
@@ -5754,7 +5757,7 @@ idle:
if (new_tasks > 0)
goto again;
-
+#endif /* CONFIG_BLD */
return NULL;
}
@@ -6415,8 +6418,9 @@ static unsigned long task_h_load(struct task_struct *p)
}
#endif
-/********** Helpers for find_busiest_group ************************/
+#ifndef CONFIG_BLD
+/********** Helpers for find_busiest_group ************************/
enum group_type {
group_other = 0,
group_imbalanced,
@@ -6507,6 +6511,7 @@ static inline int get_sd_load_idx(struct sched_domain *sd,
return load_idx;
}
+#endif /* CONFIG_BLD */
static unsigned long scale_rt_capacity(int cpu)
{
@@ -6615,6 +6620,7 @@ void update_group_capacity(struct sched_domain *sd, int cpu)
sdg->sgc->capacity = capacity;
}
+#ifndef CONFIG_BLD
/*
* Check whether the capacity of the rq has been noticeably reduced by side
* activity. The imbalance_pct is used for the threshold.
@@ -7848,6 +7854,7 @@ static inline int on_null_domain(struct rq *rq)
{
return unlikely(!rcu_dereference_sched(rq->sd));
}
+#endif /* CONFIG_BLD */
#ifdef CONFIG_NO_HZ_COMMON
/*
@@ -7856,12 +7863,39 @@ static inline int on_null_domain(struct rq *rq)
* needed, they will kick the idle load balancer, which then does idle
* load balancing for all the idle CPUs.
*/
+#ifndef CONFIG_BLD
static struct {
cpumask_var_t idle_cpus_mask;
atomic_t nr_cpus;
unsigned long next_balance; /* in jiffy units */
} nohz ____cacheline_aligned;
+void nohz_balance_exit_idle(unsigned int cpu)
+{
+ if (unlikely(test_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu)))) {
+ /*
+ * Completely isolated CPUs don't ever set, so we must test.
+ */
+ if (likely(cpumask_test_cpu(cpu, nohz.idle_cpus_mask))) {
+ cpumask_clear_cpu(cpu, nohz.idle_cpus_mask);
+ atomic_dec(&nohz.nr_cpus);
+ }
+ clear_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu));
+ }
+}
+
+static int sched_ilb_notifier(struct notifier_block *nfb,
+ unsigned long action, void *hcpu)
+{
+ switch (action & ~CPU_TASKS_FROZEN) {
+ case CPU_DYING:
+ nohz_balance_exit_idle(smp_processor_id());
+ return NOTIFY_OK;
+ default:
+ return NOTIFY_DONE;
+ }
+}
+
static inline int find_new_ilb(void)
{
int ilb = cpumask_first(nohz.idle_cpus_mask);
@@ -7900,20 +7934,6 @@ static void nohz_balancer_kick(void)
return;
}
-void nohz_balance_exit_idle(unsigned int cpu)
-{
- if (unlikely(test_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu)))) {
- /*
- * Completely isolated CPUs don't ever set, so we must test.
- */
- if (likely(cpumask_test_cpu(cpu, nohz.idle_cpus_mask))) {
- cpumask_clear_cpu(cpu, nohz.idle_cpus_mask);
- atomic_dec(&nohz.nr_cpus);
- }
- clear_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu));
- }
-}
-
static inline void set_cpu_sd_state_busy(void)
{
struct sched_domain *sd;
@@ -7930,6 +7950,8 @@ static inline void set_cpu_sd_state_busy(void)
unlock:
rcu_read_unlock();
}
+#endif /* CONFIG_BLD */
+#endif /* NO_HZ_COMMON */
void set_cpu_sd_state_idle(void)
{
@@ -7954,6 +7976,7 @@ unlock:
*/
void nohz_balance_enter_idle(int cpu)
{
+#ifndef CONFIG_BLD
/*
* If this cpu is going down, then nothing needs to be done.
*/
@@ -7972,10 +7995,8 @@ void nohz_balance_enter_idle(int cpu)
cpumask_set_cpu(cpu, nohz.idle_cpus_mask);
atomic_inc(&nohz.nr_cpus);
set_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu));
-}
#endif
-
-static DEFINE_SPINLOCK(balancing);
+}
/*
* Scale the max load_balance interval with the number of CPUs in the system.
@@ -7986,6 +8007,9 @@ void update_max_interval(void)
max_load_balance_interval = HZ*num_online_cpus()/10;
}
+#ifndef CONFIG_BLD
+static DEFINE_SPINLOCK(balancing);
+
/*
* It checks each scheduling domain to see if it is due to be balanced,
* and initiates a balancing operation if so.
@@ -8273,6 +8297,7 @@ void trigger_load_balance(struct rq *rq)
nohz_balancer_kick();
#endif
}
+#endif /* CONFIG_BLD */
static void rq_online_fair(struct rq *rq)
{
@@ -8288,7 +8313,6 @@ static void rq_offline_fair(struct rq *rq)
/* Ensure any throttled groups are reachable by pick_next_task */
unthrottle_offline_cfs_rqs(rq);
}
-
#endif /* CONFIG_SMP */
/*
@@ -8716,7 +8740,9 @@ const struct sched_class fair_sched_class = {
.put_prev_task = put_prev_task_fair,
#ifdef CONFIG_SMP
+#ifndef CONFIG_BLD
.select_task_rq = select_task_rq_fair,
+#endif
.migrate_task_rq = migrate_task_rq_fair,
.rq_online = rq_online_fair,
@@ -8777,6 +8803,7 @@ void show_numa_stats(struct task_struct *p, struct seq_file *m)
__init void init_sched_fair_class(void)
{
+#ifndef CONFIG_BLD
#ifdef CONFIG_SMP
open_softirq(SCHED_SOFTIRQ, run_rebalance_domains);
@@ -8785,5 +8812,5 @@ __init void init_sched_fair_class(void)
zalloc_cpumask_var(&nohz.idle_cpus_mask, GFP_NOWAIT);
#endif
#endif /* SMP */
-
+#endif /* BLD */
}
diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
index d5690b7..6f3589e 100644
--- a/kernel/sched/rt.c
+++ b/kernel/sched/rt.c
@@ -1375,6 +1375,7 @@ static void yield_task_rt(struct rq *rq)
#ifdef CONFIG_SMP
static int find_lowest_rq(struct task_struct *task);
+#ifndef CONFIG_BLD
static int
select_task_rq_rt(struct task_struct *p, int cpu, int sd_flag, int flags)
{
@@ -1430,6 +1431,7 @@ select_task_rq_rt(struct task_struct *p, int cpu, int sd_flag, int flags)
out:
return cpu;
}
+#endif /* CONFIG_BLD */
static void check_preempt_equal_prio(struct rq *rq, struct task_struct *p)
{
@@ -2335,7 +2337,9 @@ const struct sched_class rt_sched_class = {
.put_prev_task = put_prev_task_rt,
#ifdef CONFIG_SMP
+#ifndef CONFIG_BLD
.select_task_rq = select_task_rq_rt,
+#endif
.set_cpus_allowed = set_cpus_allowed_common,
.rq_online = rq_online_rt,
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index 898c0d2..720d524 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -415,9 +415,8 @@ struct cfs_rq {
#endif /* CONFIG_FAIR_GROUP_SCHED */
#endif /* CONFIG_SMP */
-#ifdef CONFIG_FAIR_GROUP_SCHED
struct rq *rq; /* cpu runqueue to which this cfs_rq is attached */
-
+#ifdef CONFIG_FAIR_GROUP_SCHED
/*
* leaf cfs_rqs are those that hold tasks (lowest schedulable entity in
* a hierarchy). Non-leaf lrqs hold other higher schedulable entities
@@ -441,6 +440,11 @@ struct cfs_rq {
struct list_head throttled_list;
#endif /* CONFIG_CFS_BANDWIDTH */
#endif /* CONFIG_FAIR_GROUP_SCHED */
+
+#ifdef CONFIG_BLD
+ struct list_head bld_cfs_list;
+ char pos;
+#endif
};
static inline int rt_bandwidth_enabled(void)
@@ -486,12 +490,16 @@ struct rt_rq {
/* Nests inside the rq lock: */
raw_spinlock_t rt_runtime_lock;
+ struct rq *rq;
#ifdef CONFIG_RT_GROUP_SCHED
unsigned long rt_nr_boosted;
- struct rq *rq;
struct task_group *tg;
#endif
+#ifdef CONFIG_BLD
+ struct list_head bld_rt_list;
+ int lowbit;
+#endif
};
/* Deadline class' related fields in a runqueue */

77
patches-4.7/allow_to_remove_administratively_set_mac_on_vfs.patch Normal file
View File

@ -0,0 +1,77 @@
From 33ceb32e7abc13e6a5f27916f49fa7188a13978f Mon Sep 17 00:00:00 2001
From: Corinna Vinschen <vinschen@redhat.com>
Date: Fri, 3 Jun 2016 09:40:11 -0400
Subject: [PATCH] igb: Allow to remove administratively set MAC on VFs
Before libvirt modifies the MAC address and vlan tag for an SRIOV VF
for use by a virtual machine (either using vfio device assignment or
macvtap passthru mode), it saves the current MAC address and vlan tag
so that it can reset them to their original value when the guest is
done. (libvirt can't leave the VF MAC set to the value used by the
now-defunct guest (since it may be started again later using a
different VF), but it certainly shouldn't just pick any random value,
either. So it saves the state of everything prior to using the VF, and
resets it to that.)
The igb driver initializes the MAC addresses of all VFs to
00:00:00:00:00:00, and reports that when asked (via an RTM_GETLINK
netlink message, also visible in the list of VFs in the output of "ip
link show"). But when libvirt attempts to restore the MAC address back
to 00:00:00:00:00:00 (using an RTM_SETLINK netlink message) the kernel
responds with "Invalid argument".
However, to allow libvirt to revert to the original state, we need a
way to remove the administrative set MAC on a VF, to allow normal host
operation again, and to reset/overwrite the VF MAC via VF netdev.
This patch implements the aforementioned scenario by allowing to set
the VF MAC to 00:00:00:00:00:00 via RTM_SETLINK on the PF.
igb_ndo_set_vf_mac resets the IGB_VF_FLAG_PF_SET_MAC flag to 0,
so it's possible to reset the VF MAC back to the original value via
the VF netdev.
Signed-off-by: Corinna Vinschen <vinschen@redhat.com>
---
drivers/net/ethernet/intel/igb/igb_main.c | 25 ++++++++++++++++++++-----
1 file changed, 20 insertions(+), 5 deletions(-)
diff --git a/drivers/net/ethernet/intel/igb/igb_main.c b/drivers/net/ethernet/intel/igb/igb_main.c
index 8e96c35..7be9279 100644
--- a/drivers/net/ethernet/intel/igb/igb_main.c
+++ b/drivers/net/ethernet/intel/igb/igb_main.c
@@ -7890,12 +7890,27 @@ static int igb_set_vf_mac(struct igb_adapter *adapter,
static int igb_ndo_set_vf_mac(struct net_device *netdev, int vf, u8 *mac)
{
struct igb_adapter *adapter = netdev_priv(netdev);
- if (!is_valid_ether_addr(mac) || (vf >= adapter->vfs_allocated_count))
+
+ if (vf >= adapter->vfs_allocated_count)
+ return -EINVAL;
+ /* Setting the VF MAC to 0 reverts the IGB_VF_FLAG_PF_SET_MAC
+ flag and allows to overwrite the MAC via VF netdev. This
+ is necessary to allow libvirt a way to restore the original
+ MAC after unbinding vfio-pci and reloading igbvf after shutting
+ down a VM. */
+ if (is_zero_ether_addr(mac)) {
+ adapter->vf_data[vf].flags &= ~IGB_VF_FLAG_PF_SET_MAC;
+ dev_info(&adapter->pdev->dev,
+ "remove administratively set MAC on VF %d\n",
+ vf);
+ } else if (is_valid_ether_addr (mac)) {
+ adapter->vf_data[vf].flags |= IGB_VF_FLAG_PF_SET_MAC;
+ dev_info(&adapter->pdev->dev, "setting MAC %pM on VF %d\n",
+ mac, vf);
+ dev_info(&adapter->pdev->dev,
+ "Reload the VF driver to make this change effective.");
+ } else
return -EINVAL;
- adapter->vf_data[vf].flags |= IGB_VF_FLAG_PF_SET_MAC;
- dev_info(&adapter->pdev->dev, "setting MAC %pM on VF %d\n", mac, vf);
- dev_info(&adapter->pdev->dev,
- "Reload the VF driver to make this change effective.");
if (test_bit(__IGB_DOWN, &adapter->state)) {
dev_warn(&adapter->pdev->dev,
"The VF MAC address has been set, but the PF device is not up.\n");
--
1.8.3.1

525
patches-4.7/enable_additional_cpu_optimizations_for_gcc_v4.9+_kernel_v3.15+.patch Normal file
View File

@ -0,0 +1,525 @@
WARNING
This patch works with gcc versions 4.9+ and with kernel version 3.15+ and should
NOT be applied when compiling on older versions of gcc due to key name changes
of the march flags introduced with the version 4.9 release of gcc.[1]
Use the older version of this patch hosted on the same github for older
versions of gcc.
FEATURES
This patch adds additional CPU options to the Linux kernel accessible under:
Processor type and features --->
Processor family --->
The expanded microarchitectures include:
* AMD Improved K8-family
* AMD K10-family
* AMD Family 10h (Barcelona)
* AMD Family 14h (Bobcat)
* AMD Family 16h (Jaguar)
* AMD Family 15h (Bulldozer)
* AMD Family 15h (Piledriver)
* AMD Family 15h (Steamroller)
* AMD Family 15h (Excavator)
* AMD Family 17h (Zen)
* Intel Silvermont low-power processors
* Intel 1st Gen Core i3/i5/i7 (Nehalem)
* Intel 1.5 Gen Core i3/i5/i7 (Westmere)
* Intel 2nd Gen Core i3/i5/i7 (Sandybridge)
* Intel 3rd Gen Core i3/i5/i7 (Ivybridge)
* Intel 4th Gen Core i3/i5/i7 (Haswell)
* Intel 5th Gen Core i3/i5/i7 (Broadwell)
* Intel 6th Gen Core i3/i5.i7 (Skylake)
It also offers to compile passing the 'native' option which, "selects the CPU
to generate code for at compilation time by determining the processor type of
the compiling machine. Using -march=native enables all instruction subsets
supported by the local machine and will produce code optimized for the local
machine under the constraints of the selected instruction set."[3]
MINOR NOTES
This patch also changes 'atom' to 'bonnell' in accordance with the gcc v4.9
changes. Note that upstream is using the deprecated 'match=atom' flags when I
believe it should use the newer 'march=bonnell' flag for atom processors.[2]
It is not recommended to compile on Atom-CPUs with the 'native' option.[4] The
recommendation is to the 'atom' option instead.
BENEFITS
Small but real speed increases are measurable using a make endpoint comparing
a generic kernel to one built with one of the respective microarchs.
See the following experimental evidence supporting this statement:
https://github.com/graysky2/kernel_gcc_patch
REQUIREMENTS
linux version >=3.15
gcc version >=4.9
ACKNOWLEDGMENTS
This patch builds on the seminal work by Jeroen.[5]
REFERENCES
1. https://gcc.gnu.org/gcc-4.9/changes.html
2. https://bugzilla.kernel.org/show_bug.cgi?id=77461
3. https://gcc.gnu.org/onlinedocs/gcc/x86-Options.html
4. https://github.com/graysky2/kernel_gcc_patch/issues/15
5. http://www.linuxforge.net/docs/linux/linux-gcc.php
--- a/arch/x86/include/asm/module.h 2016-12-11 14:17:54.000000000 -0500
+++ b/arch/x86/include/asm/module.h 2016-12-20 15:17:16.567422788 -0500
@@ -15,6 +15,24 @@
#define MODULE_PROC_FAMILY "586MMX "
#elif defined CONFIG_MCORE2
#define MODULE_PROC_FAMILY "CORE2 "
+#elif defined CONFIG_MNATIVE
+#define MODULE_PROC_FAMILY "NATIVE "
+#elif defined CONFIG_MNEHALEM
+#define MODULE_PROC_FAMILY "NEHALEM "
+#elif defined CONFIG_MWESTMERE
+#define MODULE_PROC_FAMILY "WESTMERE "
+#elif defined CONFIG_MSILVERMONT
+#define MODULE_PROC_FAMILY "SILVERMONT "
+#elif defined CONFIG_MSANDYBRIDGE
+#define MODULE_PROC_FAMILY "SANDYBRIDGE "
+#elif defined CONFIG_MIVYBRIDGE
+#define MODULE_PROC_FAMILY "IVYBRIDGE "
+#elif defined CONFIG_MHASWELL
+#define MODULE_PROC_FAMILY "HASWELL "
+#elif defined CONFIG_MBROADWELL
+#define MODULE_PROC_FAMILY "BROADWELL "
+#elif defined CONFIG_MSKYLAKE
+#define MODULE_PROC_FAMILY "SKYLAKE "
#elif defined CONFIG_MATOM
#define MODULE_PROC_FAMILY "ATOM "
#elif defined CONFIG_M686
@@ -33,6 +51,26 @@
#define MODULE_PROC_FAMILY "K7 "
#elif defined CONFIG_MK8
#define MODULE_PROC_FAMILY "K8 "
+#elif defined CONFIG_MK8SSE3
+#define MODULE_PROC_FAMILY "K8SSE3 "
+#elif defined CONFIG_MK10
+#define MODULE_PROC_FAMILY "K10 "
+#elif defined CONFIG_MBARCELONA
+#define MODULE_PROC_FAMILY "BARCELONA "
+#elif defined CONFIG_MBOBCAT
+#define MODULE_PROC_FAMILY "BOBCAT "
+#elif defined CONFIG_MBULLDOZER
+#define MODULE_PROC_FAMILY "BULLDOZER "
+#elif defined CONFIG_MPILEDRIVER
+#define MODULE_PROC_FAMILY "STEAMROLLER "
+#elif defined CONFIG_MSTEAMROLLER
+#define MODULE_PROC_FAMILY "PILEDRIVER "
+#elif defined CONFIG_MJAGUAR
+#define MODULE_PROC_FAMILY "JAGUAR "
+#elif defined CONFIG_MEXCAVATOR
+#define MODULE_PROC_FAMILY "EXCAVATOR "
+#elif defined CONFIG_MZEN
+#define MODULE_PROC_FAMILY "ZEN "
#elif defined CONFIG_MELAN
#define MODULE_PROC_FAMILY "ELAN "
#elif defined CONFIG_MCRUSOE
--- a/arch/x86/Kconfig.cpu 2016-12-11 14:17:54.000000000 -0500
+++ b/arch/x86/Kconfig.cpu 2016-12-20 15:31:19.929002279 -0500
@@ -147,9 +147,8 @@ config MPENTIUM4
-Paxville
-Dempsey
-
config MK6
- bool "K6/K6-II/K6-III"
+ bool "AMD K6/K6-II/K6-III"
depends on X86_32
---help---
Select this for an AMD K6-family processor. Enables use of
@@ -157,7 +156,7 @@ config MK6
flags to GCC.
config MK7
- bool "Athlon/Duron/K7"
+ bool "AMD Athlon/Duron/K7"
depends on X86_32
---help---
Select this for an AMD Athlon K7-family processor. Enables use of
@@ -165,12 +164,83 @@ config MK7
flags to GCC.
config MK8
- bool "Opteron/Athlon64/Hammer/K8"
+ bool "AMD Opteron/Athlon64/Hammer/K8"
---help---
Select this for an AMD Opteron or Athlon64 Hammer-family processor.
Enables use of some extended instructions, and passes appropriate
optimization flags to GCC.
+config MK8SSE3
+ bool "AMD Opteron/Athlon64/Hammer/K8 with SSE3"
+ ---help---
+ Select this for improved AMD Opteron or Athlon64 Hammer-family processors.
+ Enables use of some extended instructions, and passes appropriate
+ optimization flags to GCC.
+
+config MK10
+ bool "AMD 61xx/7x50/PhenomX3/X4/II/K10"
+ ---help---
+ Select this for an AMD 61xx Eight-Core Magny-Cours, Athlon X2 7x50,
+ Phenom X3/X4/II, Athlon II X2/X3/X4, or Turion II-family processor.
+ Enables use of some extended instructions, and passes appropriate
+ optimization flags to GCC.
+
+config MBARCELONA
+ bool "AMD Barcelona"
+ ---help---
+ Select this for AMD Family 10h Barcelona processors.
+
+ Enables -march=barcelona
+
+config MBOBCAT
+ bool "AMD Bobcat"
+ ---help---
+ Select this for AMD Family 14h Bobcat processors.
+
+ Enables -march=btver1
+
+config MJAGUAR
+ bool "AMD Jaguar"
+ ---help---
+ Select this for AMD Family 16h Jaguar processors.
+
+ Enables -march=btver2
+
+config MBULLDOZER
+ bool "AMD Bulldozer"
+ ---help---
+ Select this for AMD Family 15h Bulldozer processors.
+
+ Enables -march=bdver1
+
+config MPILEDRIVER
+ bool "AMD Piledriver"
+ ---help---
+ Select this for AMD Family 15h Piledriver processors.
+
+ Enables -march=bdver2
+
+config MSTEAMROLLER
+ bool "AMD Steamroller"
+ ---help---
+ Select this for AMD Family 15h Steamroller processors.
+
+ Enables -march=bdver3
+
+config MEXCAVATOR
+ bool "AMD Excavator"
+ ---help---
+ Select this for AMD Family 15h Excavator processors.
+
+ Enables -march=bdver4
+
+config MZEN
+ bool "AMD Zen"
+ ---help---
+ Select this for AMD Family 17h Zen processors.
+
+ Enables -march=znver1
+
config MCRUSOE
bool "Crusoe"
depends on X86_32
@@ -252,6 +322,7 @@ config MVIAC7
config MPSC
bool "Intel P4 / older Netburst based Xeon"
+ select X86_P6_NOP
depends on X86_64
---help---
Optimize for Intel Pentium 4, Pentium D and older Nocona/Dempsey
@@ -261,8 +332,19 @@ config MPSC
using the cpu family field
in /proc/cpuinfo. Family 15 is an older Xeon, Family 6 a newer one.
+config MATOM
+ bool "Intel Atom"
+ select X86_P6_NOP
+ ---help---
+
+ Select this for the Intel Atom platform. Intel Atom CPUs have an
+ in-order pipelining architecture and thus can benefit from
+ accordingly optimized code. Use a recent GCC with specific Atom
+ support in order to fully benefit from selecting this option.
+
config MCORE2
- bool "Core 2/newer Xeon"
+ bool "Intel Core 2"
+ select X86_P6_NOP
---help---
Select this for Intel Core 2 and newer Core 2 Xeons (Xeon 51xx and
@@ -270,14 +352,79 @@ config MCORE2
family in /proc/cpuinfo. Newer ones have 6 and older ones 15
(not a typo)
-config MATOM
- bool "Intel Atom"
+ Enables -march=core2
+
+config MNEHALEM
+ bool "Intel Nehalem"
+ select X86_P6_NOP
---help---
- Select this for the Intel Atom platform. Intel Atom CPUs have an
- in-order pipelining architecture and thus can benefit from
- accordingly optimized code. Use a recent GCC with specific Atom
- support in order to fully benefit from selecting this option.
+ Select this for 1st Gen Core processors in the Nehalem family.
+
+ Enables -march=nehalem
+
+config MWESTMERE
+ bool "Intel Westmere"
+ select X86_P6_NOP
+ ---help---
+
+ Select this for the Intel Westmere formerly Nehalem-C family.
+
+ Enables -march=westmere
+
+config MSILVERMONT
+ bool "Intel Silvermont"
+ select X86_P6_NOP
+ ---help---
+
+ Select this for the Intel Silvermont platform.
+
+ Enables -march=silvermont
+
+config MSANDYBRIDGE
+ bool "Intel Sandy Bridge"
+ select X86_P6_NOP
+ ---help---
+
+ Select this for 2nd Gen Core processors in the Sandy Bridge family.
+
+ Enables -march=sandybridge
+
+config MIVYBRIDGE
+ bool "Intel Ivy Bridge"
+ select X86_P6_NOP
+ ---help---
+
+ Select this for 3rd Gen Core processors in the Ivy Bridge family.
+
+ Enables -march=ivybridge
+
+config MHASWELL
+ bool "Intel Haswell"
+ select X86_P6_NOP
+ ---help---
+
+ Select this for 4th Gen Core processors in the Haswell family.
+
+ Enables -march=haswell
+
+config MBROADWELL
+ bool "Intel Broadwell"
+ select X86_P6_NOP
+ ---help---
+
+ Select this for 5th Gen Core processors in the Broadwell family.
+
+ Enables -march=broadwell
+
+config MSKYLAKE
+ bool "Intel Skylake"
+ select X86_P6_NOP
+ ---help---
+
+ Select this for 6th Gen Core processors in the Skylake family.
+
+ Enables -march=skylake
config GENERIC_CPU
bool "Generic-x86-64"
@@ -286,6 +433,19 @@ config GENERIC_CPU
Generic x86-64 CPU.
Run equally well on all x86-64 CPUs.
+config MNATIVE
+ bool "Native optimizations autodetected by GCC"
+ ---help---
+
+ GCC 4.2 and above support -march=native, which automatically detects
+ the optimum settings to use based on your processor. -march=native
+ also detects and applies additional settings beyond -march specific
+ to your CPU, (eg. -msse4). Unless you have a specific reason not to
+ (e.g. distcc cross-compiling), you should probably be using
+ -march=native rather than anything listed below.
+
+ Enables -march=native
+
endchoice
config X86_GENERIC
@@ -310,7 +470,7 @@ config X86_INTERNODE_CACHE_SHIFT
config X86_L1_CACHE_SHIFT
int
default "7" if MPENTIUM4 || MPSC
- default "6" if MK7 || MK8 || MPENTIUMM || MCORE2 || MATOM || MVIAC7 || X86_GENERIC || GENERIC_CPU
+ default "6" if MK7 || MK8 || MK8SSE3 || MK10 || MBARCELONA || MBOBCAT || MBULLDOZER || MPILEDRIVER || MSTEAMROLLER || MEXCAVATOR || MZEN || MJAGUAR || MPENTIUMM || MCORE2 || MNEHALEM || MWESTMERE || MSILVERMONT || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MNATIVE || MATOM || MVIAC7 || X86_GENERIC || GENERIC_CPU
default "4" if MELAN || M486 || MGEODEGX1
default "5" if MWINCHIP3D || MWINCHIPC6 || MCRUSOE || MEFFICEON || MCYRIXIII || MK6 || MPENTIUMIII || MPENTIUMII || M686 || M586MMX || M586TSC || M586 || MVIAC3_2 || MGEODE_LX
@@ -341,45 +501,47 @@ config X86_ALIGNMENT_16
config X86_INTEL_USERCOPY
def_bool y
- depends on MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M586MMX || X86_GENERIC || MK8 || MK7 || MEFFICEON || MCORE2
+ depends on MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M586MMX || X86_GENERIC || MK8 || MK8SSE3 || MK7 || MEFFICEON || MCORE2 || MK10 || MBARCELONA || MNEHALEM || MWESTMERE || MSILVERMONT || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MNATIVE
config X86_USE_PPRO_CHECKSUM
def_bool y
- depends on MWINCHIP3D || MWINCHIPC6 || MCYRIXIII || MK7 || MK6 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MK8 || MVIAC3_2 || MVIAC7 || MEFFICEON || MGEODE_LX || MCORE2 || MATOM
+ depends on MWINCHIP3D || MWINCHIPC6 || MCYRIXIII || MK7 || MK6 || MK10 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MK8 || MK8SSE3 || MVIAC3_2 || MVIAC7 || MEFFICEON || MGEODE_LX || MCORE2 || MNEHALEM || MWESTMERE || MSILVERMONT || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MATOM || MNATIVE
config X86_USE_3DNOW
def_bool y
depends on (MCYRIXIII || MK7 || MGEODE_LX) && !UML
-#
-# P6_NOPs are a relatively minor optimization that require a family >=
-# 6 processor, except that it is broken on certain VIA chips.
-# Furthermore, AMD chips prefer a totally different sequence of NOPs
-# (which work on all CPUs). In addition, it looks like Virtual PC
-# does not understand them.
-#
-# As a result, disallow these if we're not compiling for X86_64 (these
-# NOPs do work on all x86-64 capable chips); the list of processors in
-# the right-hand clause are the cores that benefit from this optimization.
-#
config X86_P6_NOP
- def_bool y
+ default n
+ bool "Support for P6_NOPs on Intel chips"
depends on X86_64
- depends on (MCORE2 || MPENTIUM4 || MPSC)
+ depends on (MCORE2 || MPENTIUM4 || MPSC || MATOM || MNEHALEM || MWESTMERE || MSILVERMONT || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MNATIVE)
+ ---help---
+ P6_NOPs are a relatively minor optimization that require a family >=
+ 6 processor, except that it is broken on certain VIA chips.
+ Furthermore, AMD chips prefer a totally different sequence of NOPs
+ (which work on all CPUs). In addition, it looks like Virtual PC
+ does not understand them.
+
+ As a result, disallow these if we're not compiling for X86_64 (these
+ NOPs do work on all x86-64 capable chips); the list of processors in
+ the right-hand clause are the cores that benefit from this optimization.
+
+ Say Y if you have Intel CPU newer than Pentium Pro, N otherwise.
config X86_TSC
def_bool y
- depends on (MWINCHIP3D || MCRUSOE || MEFFICEON || MCYRIXIII || MK7 || MK6 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || M586MMX || M586TSC || MK8 || MVIAC3_2 || MVIAC7 || MGEODEGX1 || MGEODE_LX || MCORE2 || MATOM) || X86_64
+ depends on (MWINCHIP3D || MCRUSOE || MEFFICEON || MCYRIXIII || MK7 || MK6 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || M586MMX || M586TSC || MK8 || MK8SSE3 || MVIAC3_2 || MVIAC7 || MGEODEGX1 || MGEODE_LX || MCORE2 || MNEHALEM || MWESTMERE || MSILVERMONT || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MNATIVE || MATOM) || X86_64
config X86_CMPXCHG64
def_bool y
- depends on X86_PAE || X86_64 || MCORE2 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MATOM
+ depends on X86_PAE || X86_64 || MCORE2 || MNEHALEM || MWESTMERE || MSILVERMONT || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MATOM || MNATIVE
# this should be set for all -march=.. options where the compiler
# generates cmov.
config X86_CMOV
def_bool y
- depends on (MK8 || MK7 || MCORE2 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MVIAC3_2 || MVIAC7 || MCRUSOE || MEFFICEON || X86_64 || MATOM || MGEODE_LX)
+ depends on (MK8 || MK8SSE3 || MK10 || MBARCELONA || MBOBCAT || MBULLDOZER || MPILEDRIVER || MSTEAMROLLER || MEXCAVATOR || MZEN || MJAGUAR || MK7 || MCORE2 || MNEHALEM || MWESTMERE || MSILVERMONT || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MVIAC3_2 || MVIAC7 || MCRUSOE || MEFFICEON || X86_64 || MNATIVE || MATOM || MGEODE_LX)
config X86_MINIMUM_CPU_FAMILY
int
--- a/arch/x86/Makefile 2016-12-11 14:17:54.000000000 -0500
+++ b/arch/x86/Makefile 2016-12-20 15:26:01.142666491 -0500
@@ -104,13 +104,40 @@ else
KBUILD_CFLAGS += $(call cc-option,-mskip-rax-setup)
# FIXME - should be integrated in Makefile.cpu (Makefile_32.cpu)
+ cflags-$(CONFIG_MNATIVE) += $(call cc-option,-march=native)
cflags-$(CONFIG_MK8) += $(call cc-option,-march=k8)
+ cflags-$(CONFIG_MK8SSE3) += $(call cc-option,-march=k8-sse3,-mtune=k8)
+ cflags-$(CONFIG_MK10) += $(call cc-option,-march=amdfam10)
+ cflags-$(CONFIG_MBARCELONA) += $(call cc-option,-march=barcelona)
+ cflags-$(CONFIG_MBOBCAT) += $(call cc-option,-march=btver1)
+ cflags-$(CONFIG_MJAGUAR) += $(call cc-option,-march=btver2)
+ cflags-$(CONFIG_MBULLDOZER) += $(call cc-option,-march=bdver1)
+ cflags-$(CONFIG_MPILEDRIVER) += $(call cc-option,-march=bdver2)
+ cflags-$(CONFIG_MSTEAMROLLER) += $(call cc-option,-march=bdver3)
+ cflags-$(CONFIG_MEXCAVATOR) += $(call cc-option,-march=bdver4)
+ cflags-$(CONFIG_MZEN) += $(call cc-option,-march=znver1)
cflags-$(CONFIG_MPSC) += $(call cc-option,-march=nocona)
cflags-$(CONFIG_MCORE2) += \
- $(call cc-option,-march=core2,$(call cc-option,-mtune=generic))
- cflags-$(CONFIG_MATOM) += $(call cc-option,-march=atom) \
- $(call cc-option,-mtune=atom,$(call cc-option,-mtune=generic))
+ $(call cc-option,-march=core2,$(call cc-option,-mtune=core2))
+ cflags-$(CONFIG_MNEHALEM) += \
+ $(call cc-option,-march=nehalem,$(call cc-option,-mtune=nehalem))
+ cflags-$(CONFIG_MWESTMERE) += \
+ $(call cc-option,-march=westmere,$(call cc-option,-mtune=westmere))
+ cflags-$(CONFIG_MSILVERMONT) += \
+ $(call cc-option,-march=silvermont,$(call cc-option,-mtune=silvermont))
+ cflags-$(CONFIG_MSANDYBRIDGE) += \
+ $(call cc-option,-march=sandybridge,$(call cc-option,-mtune=sandybridge))
+ cflags-$(CONFIG_MIVYBRIDGE) += \
+ $(call cc-option,-march=ivybridge,$(call cc-option,-mtune=ivybridge))
+ cflags-$(CONFIG_MHASWELL) += \
+ $(call cc-option,-march=haswell,$(call cc-option,-mtune=haswell))
+ cflags-$(CONFIG_MBROADWELL) += \
+ $(call cc-option,-march=broadwell,$(call cc-option,-mtune=broadwell))
+ cflags-$(CONFIG_MSKYLAKE) += \
+ $(call cc-option,-march=skylake,$(call cc-option,-mtune=skylake))
+ cflags-$(CONFIG_MATOM) += $(call cc-option,-march=bonnell) \
+ $(call cc-option,-mtune=bonnell,$(call cc-option,-mtune=generic))
cflags-$(CONFIG_GENERIC_CPU) += $(call cc-option,-mtune=generic)
KBUILD_CFLAGS += $(cflags-y)
--- a/arch/x86/Makefile_32.cpu 2016-12-11 14:17:54.000000000 -0500
+++ b/arch/x86/Makefile_32.cpu 2016-12-20 15:27:05.357148298 -0500
@@ -23,7 +23,18 @@ cflags-$(CONFIG_MK6) += -march=k6
# Please note, that patches that add -march=athlon-xp and friends are pointless.
# They make zero difference whatsosever to performance at this time.
cflags-$(CONFIG_MK7) += -march=athlon
+cflags-$(CONFIG_MNATIVE) += $(call cc-option,-march=native)
cflags-$(CONFIG_MK8) += $(call cc-option,-march=k8,-march=athlon)
+cflags-$(CONFIG_MK8SSE3) += $(call cc-option,-march=k8-sse3,-march=athlon)
+cflags-$(CONFIG_MK10) += $(call cc-option,-march=amdfam10,-march=athlon)
+cflags-$(CONFIG_MBARCELONA) += $(call cc-option,-march=barcelona,-march=athlon)
+cflags-$(CONFIG_MBOBCAT) += $(call cc-option,-march=btver1,-march=athlon)
+cflags-$(CONFIG_MJAGUAR) += $(call cc-option,-march=btver2,-march=athlon)
+cflags-$(CONFIG_MBULLDOZER) += $(call cc-option,-march=bdver1,-march=athlon)
+cflags-$(CONFIG_MPILEDRIVER) += $(call cc-option,-march=bdver2,-march=athlon)
+cflags-$(CONFIG_MSTEAMROLLER) += $(call cc-option,-march=bdver3,-march=athlon)
+cflags-$(CONFIG_MEXCAVATOR) += $(call cc-option,-march=bdver4,-march=athlon)
+cflags-$(CONFIG_MZEN) += $(call cc-option,-march=znver1,-march=athlon)
cflags-$(CONFIG_MCRUSOE) += -march=i686 $(align)-functions=0 $(align)-jumps=0 $(align)-loops=0
cflags-$(CONFIG_MEFFICEON) += -march=i686 $(call tune,pentium3) $(align)-functions=0 $(align)-jumps=0 $(align)-loops=0
cflags-$(CONFIG_MWINCHIPC6) += $(call cc-option,-march=winchip-c6,-march=i586)
@@ -32,8 +43,16 @@ cflags-$(CONFIG_MCYRIXIII) += $(call cc-
cflags-$(CONFIG_MVIAC3_2) += $(call cc-option,-march=c3-2,-march=i686)
cflags-$(CONFIG_MVIAC7) += -march=i686
cflags-$(CONFIG_MCORE2) += -march=i686 $(call tune,core2)
-cflags-$(CONFIG_MATOM) += $(call cc-option,-march=atom,$(call cc-option,-march=core2,-march=i686)) \
- $(call cc-option,-mtune=atom,$(call cc-option,-mtune=generic))
+cflags-$(CONFIG_MNEHALEM) += -march=i686 $(call tune,nehalem)
+cflags-$(CONFIG_MWESTMERE) += -march=i686 $(call tune,westmere)
+cflags-$(CONFIG_MSILVERMONT) += -march=i686 $(call tune,silvermont)
+cflags-$(CONFIG_MSANDYBRIDGE) += -march=i686 $(call tune,sandybridge)
+cflags-$(CONFIG_MIVYBRIDGE) += -march=i686 $(call tune,ivybridge)
+cflags-$(CONFIG_MHASWELL) += -march=i686 $(call tune,haswell)
+cflags-$(CONFIG_MBROADWELL) += -march=i686 $(call tune,broadwell)
+cflags-$(CONFIG_MSKYLAKE) += -march=i686 $(call tune,skylake)
+cflags-$(CONFIG_MATOM) += $(call cc-option,-march=bonnell,$(call cc-option,-march=core2,-march=i686)) \
+ $(call cc-option,-mtune=bonnell,$(call cc-option,-mtune=generic))
# AMD Elan support
cflags-$(CONFIG_MELAN) += -march=i486

12
patches-4.7/explicitly_select_page_0_at_initialization.patch Normal file
View File

@ -0,0 +1,12 @@
diff --git a/drivers/net/ethernet/intel/igb/e1000_82575.c b/drivers/net/ethernet/intel/igb/e1000_82575.c
index a61447f..1264a36 100644
--- a/drivers/net/ethernet/intel/igb/e1000_82575.c
+++ b/drivers/net/ethernet/intel/igb/e1000_82575.c
@@ -246,6 +246,7 @@ static s32 igb_init_phy_params_82575(struct e1000_hw *hw)
E1000_STATUS_FUNC_SHIFT;
/* Set phy->phy_addr and phy->id. */
+ igb_write_phy_reg_82580(hw, I347AT4_PAGE_SELECT, 0);
ret_val = igb_get_phy_id_82575(hw);
if (ret_val)
return ret_val;

20
patches-4.7/i210at_nic_reset.patch Normal file
View File

@ -0,0 +1,20 @@
diff --git a/drivers/pci/msi.c b/drivers/pci/msi.c
index a080f4496fe2..142341f8331b 100644
--- a/drivers/pci/msi.c
+++ b/drivers/pci/msi.c
@@ -645,6 +645,15 @@ static int msi_capability_init(struct pci_dev *dev, int nvec)
return ret;
}
+ /*
+ * The mask can be ignored and PCI 2.3 does not specify mask bits for
+ * each MSI interrupt. So in case of hierarchical irqdomains we need
+ * to make sure that if masking is not available that the msi message
+ * is written prior to setting the MSI enable bit in the device.
+ */
+ if (pci_msi_ignore_mask || !entry->msi_attrib.maskbit)
+ irq_domain_activate_irq(irq_get_irq_data(entry->irq));
+
/* Set MSI enabled bits */
pci_intx_for_msi(dev, 0);
pci_msi_set_enable(dev, 1);

23
patches-4.7/reset_the_PHY_before_reading_the_phy_id.patch Normal file
View File

@ -0,0 +1,23 @@
diff --git a/drivers/net/ethernet/intel/igb/e1000_82575.c b/drivers/net/ethernet/intel/igb/e1000_82575.c
index a61447f..ee44398 100644
--- a/drivers/net/ethernet/intel/igb/e1000_82575.c
+++ b/drivers/net/ethernet/intel/igb/e1000_82575.c
@@ -245,6 +245,17 @@ static s32 igb_init_phy_params_82575(struct e1000_hw *hw)
hw->bus.func = (rd32(E1000_STATUS) & E1000_STATUS_FUNC_MASK) >>
E1000_STATUS_FUNC_SHIFT;
+ /* Make sure the PHY is in a good state. Several people have reported
+ * firmware leaving the PHY's page select register set to something
+ * other than the default of zero, which causes the PHY ID read to
+ * access something other than the intended register.
+ */
+ ret_val = hw->phy.ops.reset(hw);
+ if (ret_val) {
+ hw_dbg("Error resetting the PHY.\n");
+ goto out;
+ }
+
/* Set phy->phy_addr and phy->id. */
ret_val = igb_get_phy_id_82575(hw);
if (ret_val)

6894
patches-4.7/uksm-0.1.2.5-for-v4.7.patch Normal file
View File

File diff suppressed because it is too large Load Diff

16
patches-4.7/workaround_for_igb_i210_firmware_issue.patch Normal file
View File

@ -0,0 +1,16 @@
diff --git a/drivers/net/ethernet/intel/igb/e1000_phy.c b/drivers/net/ethernet/intel/igb/e1000_phy.c
index 5b54254..93ec2d0 100644
--- a/drivers/net/ethernet/intel/igb/e1000_phy.c
+++ b/drivers/net/ethernet/intel/igb/e1000_phy.c
@@ -77,6 +77,10 @@ s32 igb_get_phy_id(struct e1000_hw *hw)
s32 ret_val = 0;
u16 phy_id;
+ /* ensure phy page selection to fix misconfigured i210 */
+ if (hw->mac.type == e1000_i210)
+ phy->ops.write_reg(hw, I347AT4_PAGE_SELECT, 0);
+
ret_val = phy->ops.read_reg(hw, PHY_ID1, &phy_id);
if (ret_val)
goto out;

103
patches/0001-block-cgroups-kconfig-build-bits-for-BFQ-v7r11-4.10..patch Normal file
View File

@ -0,0 +1,103 @@
From 8500f47272575b4616beb487c483019248d8c501 Mon Sep 17 00:00:00 2001
From: Paolo Valente <paolo.valente@unimore.it>
Date: Tue, 7 Apr 2015 13:39:12 +0200
Subject: [PATCH 1/4] block: cgroups, kconfig, build bits for BFQ-v7r11-4.10.0
Update Kconfig.iosched and do the related Makefile changes to include
kernel configuration options for BFQ. Also increase the number of
policies supported by the blkio controller so that BFQ can add its
own.
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Signed-off-by: Arianna Avanzini <avanzini@google.com>
---
block/Kconfig.iosched | 32 ++++++++++++++++++++++++++++++++
block/Makefile | 1 +
include/linux/blkdev.h | 2 +-
3 files changed, 34 insertions(+), 1 deletion(-)
diff --git a/block/Kconfig.iosched b/block/Kconfig.iosched
index 421bef9..0ee5f0f 100644
--- a/block/Kconfig.iosched
+++ b/block/Kconfig.iosched
@@ -39,6 +39,27 @@ config CFQ_GROUP_IOSCHED
---help---
Enable group IO scheduling in CFQ.
+config IOSCHED_BFQ
+ tristate "BFQ I/O scheduler"
+ default n
+ ---help---
+ The BFQ I/O scheduler tries to distribute bandwidth among
+ all processes according to their weights.
+ It aims at distributing the bandwidth as desired, independently of
+ the disk parameters and with any workload. It also tries to
+ guarantee low latency to interactive and soft real-time
+ applications. If compiled built-in (saying Y here), BFQ can
+ be configured to support hierarchical scheduling.
+
+config CGROUP_BFQIO
+ bool "BFQ hierarchical scheduling support"
+ depends on CGROUPS && IOSCHED_BFQ=y
+ default n
+ ---help---
+ Enable hierarchical scheduling in BFQ, using the cgroups
+ filesystem interface. The name of the subsystem will be
+ bfqio.
+
choice
prompt "Default I/O scheduler"
default DEFAULT_CFQ
@@ -52,6 +73,16 @@ choice
config DEFAULT_CFQ
bool "CFQ" if IOSCHED_CFQ=y
+ config DEFAULT_BFQ
+ bool "BFQ" if IOSCHED_BFQ=y
+ help
+ Selects BFQ as the default I/O scheduler which will be
+ used by default for all block devices.
+ The BFQ I/O scheduler aims at distributing the bandwidth
+ as desired, independently of the disk parameters and with
+ any workload. It also tries to guarantee low latency to
+ interactive and soft real-time applications.
+
config DEFAULT_NOOP
bool "No-op"
@@ -61,6 +92,7 @@ config DEFAULT_IOSCHED
string
default "deadline" if DEFAULT_DEADLINE
default "cfq" if DEFAULT_CFQ
+ default "bfq" if DEFAULT_BFQ
default "noop" if DEFAULT_NOOP
endmenu
diff --git a/block/Makefile b/block/Makefile
index a827f98..3b14703 100644
--- a/block/Makefile
+++ b/block/Makefile
@@ -18,6 +18,7 @@ obj-$(CONFIG_BLK_DEV_THROTTLING) += blk-throttle.o
obj-$(CONFIG_IOSCHED_NOOP) += noop-iosched.o
obj-$(CONFIG_IOSCHED_DEADLINE) += deadline-iosched.o
obj-$(CONFIG_IOSCHED_CFQ) += cfq-iosched.o
+obj-$(CONFIG_IOSCHED_BFQ) += bfq-iosched.o
obj-$(CONFIG_BLOCK_COMPAT) += compat_ioctl.o
obj-$(CONFIG_BLK_CMDLINE_PARSER) += cmdline-parser.o
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 1ca8e8f..8e2d6ed 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -47,7 +47,7 @@ struct rq_wb;
* Maximum number of blkcg policies allowed to be registered concurrently.
* Defined here to simplify include dependency.
*/
-#define BLKCG_MAX_POLS 2
+#define BLKCG_MAX_POLS 3
typedef void (rq_end_io_fn)(struct request *, int);
--
2.10.0

7109
patches/0002-block-introduce-the-BFQ-v7r11-I-O-sched-for-4.10.0.patch Normal file
View File

File diff suppressed because it is too large Load Diff

1101
patches/0003-block-bfq-add-Early-Queue-Merge-EQM-to-BFQ-v7r11-for.patch Normal file
View File

File diff suppressed because it is too large Load Diff

9187
patches/0004-Turn-BFQ-v7r11-for-4.10.0-into-BFQ-v8r8-for-4.10.0.patch Normal file
View File

File diff suppressed because it is too large Load Diff

111
patches/0009-Implement-min-and-msec-hrtimeout-un-interruptible-sc.patch Normal file
View File

@ -0,0 +1,111 @@
From a7fb2842267fd275cae9cf44dd3037469f75eeef Mon Sep 17 00:00:00 2001
From: Con Kolivas <kernel@kolivas.org>
Date: Tue, 1 Nov 2016 12:54:20 +1100
Subject: [PATCH 09/25] Implement min and msec hrtimeout un/interruptible
schedule timeout variants with a lower resolution of 1ms to work around low
Hz time resolutions.
---
include/linux/sched.h | 6 +++++
kernel/time/hrtimer.c | 71 +++++++++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 77 insertions(+)
diff --git a/include/linux/sched.h b/include/linux/sched.h
index d752ef6..46544f4 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -457,6 +457,12 @@ extern signed long schedule_timeout_interruptible(signed long timeout);
extern signed long schedule_timeout_killable(signed long timeout);
extern signed long schedule_timeout_uninterruptible(signed long timeout);
extern signed long schedule_timeout_idle(signed long timeout);
+
+extern signed long schedule_msec_hrtimeout(signed long timeout);
+extern signed long schedule_min_hrtimeout(void);
+extern signed long schedule_msec_hrtimeout_interruptible(signed long timeout);
+extern signed long schedule_msec_hrtimeout_uninterruptible(signed long timeout);
+
asmlinkage void schedule(void);
extern void schedule_preempt_disabled(void);
diff --git a/kernel/time/hrtimer.c b/kernel/time/hrtimer.c
index c6ecedd..a47f5b3 100644
--- a/kernel/time/hrtimer.c
+++ b/kernel/time/hrtimer.c
@@ -1796,3 +1796,74 @@ int __sched schedule_hrtimeout(ktime_t *expires,
return schedule_hrtimeout_range(expires, 0, mode);
}
EXPORT_SYMBOL_GPL(schedule_hrtimeout);
+
+/*
+ * As per schedule_hrtimeout but taskes a millisecond value and returns how
+ * many milliseconds are left.
+ */
+signed long __sched schedule_msec_hrtimeout(signed long timeout)
+{
+ struct hrtimer_sleeper t;
+ int delta, secs, jiffs;
+ ktime_t expires;
+
+ if (!timeout) {
+ __set_current_state(TASK_RUNNING);
+ return 0;
+ }
+
+ jiffs = msecs_to_jiffies(timeout);
+ /*
+ * If regular timer resolution is adequate or hrtimer resolution is not
+ * (yet) better than Hz, as would occur during startup, use regular
+ * timers.
+ */
+ if (jiffs > 4 || hrtimer_resolution >= NSEC_PER_SEC / HZ)
+ return schedule_timeout(jiffs);
+
+ secs = timeout / 1000;
+ delta = (timeout % 1000) * NSEC_PER_MSEC;
+ expires = ktime_set(secs, delta);
+
+ hrtimer_init_on_stack(&t.timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
+ hrtimer_set_expires_range_ns(&t.timer, expires, delta);
+
+ hrtimer_init_sleeper(&t, current);
+
+ hrtimer_start_expires(&t.timer, HRTIMER_MODE_REL);
+
+ if (likely(t.task))
+ schedule();
+
+ hrtimer_cancel(&t.timer);
+ destroy_hrtimer_on_stack(&t.timer);
+
+ __set_current_state(TASK_RUNNING);
+
+ expires = hrtimer_expires_remaining(&t.timer);
+ timeout = ktime_to_ms(expires);
+ return timeout < 0 ? 0 : timeout;
+}
+
+EXPORT_SYMBOL(schedule_msec_hrtimeout);
+
+signed long __sched schedule_min_hrtimeout(void)
+{
+ return schedule_msec_hrtimeout(1);
+}
+
+EXPORT_SYMBOL(schedule_min_hrtimeout);
+
+signed long __sched schedule_msec_hrtimeout_interruptible(signed long timeout)
+{
+ __set_current_state(TASK_INTERRUPTIBLE);
+ return schedule_msec_hrtimeout(timeout);
+}
+EXPORT_SYMBOL(schedule_msec_hrtimeout_interruptible);
+
+signed long __sched schedule_msec_hrtimeout_uninterruptible(signed long timeout)
+{
+ __set_current_state(TASK_UNINTERRUPTIBLE);
+ return schedule_msec_hrtimeout(timeout);
+}
+EXPORT_SYMBOL(schedule_msec_hrtimeout_uninterruptible);
--
2.9.3

48
patches/0010-Special-case-calls-of-schedule_timeout-1-to-use-the-.patch Normal file
View File

@ -0,0 +1,48 @@
From a4f3820228ebab3d5d480d720fecebd3f7e71771 Mon Sep 17 00:00:00 2001
From: Con Kolivas <kernel@kolivas.org>
Date: Sat, 5 Nov 2016 09:27:36 +1100
Subject: [PATCH 10/25] Special case calls of schedule_timeout(1) to use the
min hrtimeout of 1ms, working around low Hz resolutions.
---
kernel/time/timer.c | 15 +++++++++++++--
1 file changed, 13 insertions(+), 2 deletions(-)
diff --git a/kernel/time/timer.c b/kernel/time/timer.c
index ef3128f..3f72c13 100644
--- a/kernel/time/timer.c
+++ b/kernel/time/timer.c
@@ -1750,6 +1750,17 @@ signed long __sched schedule_timeout(signed long timeout)
expire = timeout + jiffies;
+ if (timeout == 1 && hrtimer_resolution < NSEC_PER_SEC / HZ) {
+ /*
+ * Special case 1 as being a request for the minimum timeout
+ * and use highres timers to timeout after 1ms to workaround
+ * the granularity of low Hz tick timers.
+ */
+ if (!schedule_min_hrtimeout())
+ return 0;
+ goto out_timeout;
+ }
+
setup_timer_on_stack(&timer, process_timeout, (unsigned long)current);
__mod_timer(&timer, expire, false);
schedule();
@@ -1757,10 +1768,10 @@ signed long __sched schedule_timeout(signed long timeout)
/* Remove the timer from the object tracker */
destroy_timer_on_stack(&timer);
-
+out_timeout:
timeout = expire - jiffies;
- out:
+out:
return timeout < 0 ? 0 : timeout;
}
EXPORT_SYMBOL(schedule_timeout);
--
2.9.3

54
patches/0011-Convert-msleep-to-use-hrtimers-when-active.patch Normal file
View File

@ -0,0 +1,54 @@
From 534bc9d3e559420eaf57771f48d2c2f549dcc4d2 Mon Sep 17 00:00:00 2001
From: Con Kolivas <kernel@kolivas.org>
Date: Fri, 4 Nov 2016 09:25:54 +1100
Subject: [PATCH 11/25] Convert msleep to use hrtimers when active.
---
kernel/time/timer.c | 24 ++++++++++++++++++++++--
1 file changed, 22 insertions(+), 2 deletions(-)
diff --git a/kernel/time/timer.c b/kernel/time/timer.c
index 3f72c13..bc53598 100644
--- a/kernel/time/timer.c
+++ b/kernel/time/timer.c
@@ -1892,7 +1892,19 @@ void __init init_timers(void)
*/
void msleep(unsigned int msecs)
{
- unsigned long timeout = msecs_to_jiffies(msecs) + 1;
+ int jiffs = msecs_to_jiffies(msecs);
+ unsigned long timeout;
+
+ /*
+ * Use high resolution timers where the resolution of tick based
+ * timers is inadequate.
+ */
+ if (jiffs < 5 && hrtimer_resolution < NSEC_PER_SEC / HZ) {
+ while (msecs)
+ msecs = schedule_msec_hrtimeout_uninterruptible(msecs);
+ return;
+ }
+ timeout = msecs_to_jiffies(msecs) + 1;
while (timeout)
timeout = schedule_timeout_uninterruptible(timeout);
@@ -1906,7 +1918,15 @@ EXPORT_SYMBOL(msleep);
*/
unsigned long msleep_interruptible(unsigned int msecs)
{
- unsigned long timeout = msecs_to_jiffies(msecs) + 1;
+ int jiffs = msecs_to_jiffies(msecs);
+ unsigned long timeout;
+
+ if (jiffs < 5 && hrtimer_resolution < NSEC_PER_SEC / HZ) {
+ while (msecs && !signal_pending(current))
+ msecs = schedule_msec_hrtimeout_interruptible(msecs);
+ return msecs;
+ }
+ timeout = msecs_to_jiffies(msecs) + 1;
while (timeout && !signal_pending(current))
timeout = schedule_timeout_interruptible(timeout);
--
2.9.3

226
patches/0012-Replace-all-schedule-timeout-1-with-schedule_min_hrt.patch Normal file
View File

@ -0,0 +1,226 @@
From 8fef7b75352d874af02881de3493f2ce2d47a341 Mon Sep 17 00:00:00 2001
From: Con Kolivas <kernel@kolivas.org>
Date: Mon, 20 Feb 2017 13:28:30 +1100
Subject: [PATCH 12/25] Replace all schedule timeout(1) with
schedule_min_hrtimeout()
---
drivers/block/swim.c | 6 +++---
drivers/char/ipmi/ipmi_msghandler.c | 2 +-
drivers/char/ipmi/ipmi_ssif.c | 2 +-
drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c | 2 +-
drivers/gpu/drm/vmwgfx/vmwgfx_irq.c | 2 +-
drivers/mfd/ucb1x00-core.c | 2 +-
drivers/misc/sgi-xp/xpc_channel.c | 2 +-
drivers/net/caif/caif_hsi.c | 2 +-
drivers/ntb/test/ntb_perf.c | 2 +-
drivers/staging/comedi/drivers/ni_mio_common.c | 2 +-
fs/afs/vlocation.c | 2 +-
fs/btrfs/extent-tree.c | 2 +-
fs/btrfs/inode-map.c | 2 +-
sound/usb/line6/pcm.c | 2 +-
14 files changed, 16 insertions(+), 16 deletions(-)
diff --git a/drivers/block/swim.c b/drivers/block/swim.c
index b5afd49..7d09955 100644
--- a/drivers/block/swim.c
+++ b/drivers/block/swim.c
@@ -332,7 +332,7 @@ static inline void swim_motor(struct swim __iomem *base,
if (swim_readbit(base, MOTOR_ON))
break;
current->state = TASK_INTERRUPTIBLE;
- schedule_timeout(1);
+ schedule_min_hrtimeout();
}
} else if (action == OFF) {
swim_action(base, MOTOR_OFF);
@@ -351,7 +351,7 @@ static inline void swim_eject(struct swim __iomem *base)
if (!swim_readbit(base, DISK_IN))
break;
current->state = TASK_INTERRUPTIBLE;
- schedule_timeout(1);
+ schedule_min_hrtimeout();
}
swim_select(base, RELAX);
}
@@ -375,7 +375,7 @@ static inline int swim_step(struct swim __iomem *base)
for (wait = 0; wait < HZ; wait++) {
current->state = TASK_INTERRUPTIBLE;
- schedule_timeout(1);
+ schedule_min_hrtimeout();
swim_select(base, RELAX);
if (!swim_readbit(base, STEP))
diff --git a/drivers/char/ipmi/ipmi_msghandler.c b/drivers/char/ipmi/ipmi_msghandler.c
index 92e53ac..a2418e7 100644
--- a/drivers/char/ipmi/ipmi_msghandler.c
+++ b/drivers/char/ipmi/ipmi_msghandler.c
@@ -2953,7 +2953,7 @@ static void cleanup_smi_msgs(ipmi_smi_t intf)
/* Current message first, to preserve order */
while (intf->curr_msg && !list_empty(&intf->waiting_rcv_msgs)) {
/* Wait for the message to clear out. */
- schedule_timeout(1);
+ schedule_min_hrtimeout();
}
/* No need for locks, the interface is down. */
diff --git a/drivers/char/ipmi/ipmi_ssif.c b/drivers/char/ipmi/ipmi_ssif.c
index cca6e5b..fd3c7da 100644
--- a/drivers/char/ipmi/ipmi_ssif.c
+++ b/drivers/char/ipmi/ipmi_ssif.c
@@ -1185,7 +1185,7 @@ static int ssif_remove(struct i2c_client *client)
/* make sure the driver is not looking for flags any more. */
while (ssif_info->ssif_state != SSIF_NORMAL)
- schedule_timeout(1);
+ schedule_min_hrtimeout();
ssif_info->stopping = true;
del_timer_sync(&ssif_info->retry_timer);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
index b6a0806..b5b02cf 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
@@ -235,7 +235,7 @@ static int vmw_fifo_wait_noirq(struct vmw_private *dev_priv,
DRM_ERROR("SVGA device lockup.\n");
break;
}
- schedule_timeout(1);
+ schedule_min_hrtimeout();
if (interruptible && signal_pending(current)) {
ret = -ERESTARTSYS;
break;
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c b/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c
index 0c7e172..4c1555c 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c
@@ -156,7 +156,7 @@ int vmw_fallback_wait(struct vmw_private *dev_priv,
break;
}
if (lazy)
- schedule_timeout(1);
+ schedule_min_hrtimeout();
else if ((++count & 0x0F) == 0) {
/**
* FIXME: Use schedule_hr_timeout here for
diff --git a/drivers/mfd/ucb1x00-core.c b/drivers/mfd/ucb1x00-core.c
index d6fb2e1..7ac951b 100644
--- a/drivers/mfd/ucb1x00-core.c
+++ b/drivers/mfd/ucb1x00-core.c
@@ -253,7 +253,7 @@ unsigned int ucb1x00_adc_read(struct ucb1x00 *ucb, int adc_channel, int sync)
break;
/* yield to other processes */
set_current_state(TASK_INTERRUPTIBLE);
- schedule_timeout(1);
+ schedule_min_hrtimeout();
}
return UCB_ADC_DAT(val);
diff --git a/drivers/misc/sgi-xp/xpc_channel.c b/drivers/misc/sgi-xp/xpc_channel.c
index 128d561..38e68e9 100644
--- a/drivers/misc/sgi-xp/xpc_channel.c
+++ b/drivers/misc/sgi-xp/xpc_channel.c
@@ -837,7 +837,7 @@ xpc_allocate_msg_wait(struct xpc_channel *ch)
atomic_inc(&ch->n_on_msg_allocate_wq);
prepare_to_wait(&ch->msg_allocate_wq, &wait, TASK_INTERRUPTIBLE);
- ret = schedule_timeout(1);
+ ret = schedule_min_hrtimeout();
finish_wait(&ch->msg_allocate_wq, &wait);
atomic_dec(&ch->n_on_msg_allocate_wq);
diff --git a/drivers/net/caif/caif_hsi.c b/drivers/net/caif/caif_hsi.c
index ddabce7..67fb5ce 100644
--- a/drivers/net/caif/caif_hsi.c
+++ b/drivers/net/caif/caif_hsi.c
@@ -944,7 +944,7 @@ static void cfhsi_wake_down(struct work_struct *work)
break;
set_current_state(TASK_INTERRUPTIBLE);
- schedule_timeout(1);
+ schedule_min_hrtimeout();
retry--;
}
diff --git a/drivers/ntb/test/ntb_perf.c b/drivers/ntb/test/ntb_perf.c
index 434e1d4..2f9543b 100644
--- a/drivers/ntb/test/ntb_perf.c
+++ b/drivers/ntb/test/ntb_perf.c
@@ -308,7 +308,7 @@ static int perf_move_data(struct pthr_ctx *pctx, char __iomem *dst, char *src,
if (unlikely((jiffies - last_sleep) > 5 * HZ)) {
last_sleep = jiffies;
set_current_state(TASK_INTERRUPTIBLE);
- schedule_timeout(1);
+ schedule_min_hrtimeout();
}
if (unlikely(kthread_should_stop()))
diff --git a/drivers/staging/comedi/drivers/ni_mio_common.c b/drivers/staging/comedi/drivers/ni_mio_common.c
index b2e3828..beae38b 100644
--- a/drivers/staging/comedi/drivers/ni_mio_common.c
+++ b/drivers/staging/comedi/drivers/ni_mio_common.c
@@ -4655,7 +4655,7 @@ static int cs5529_wait_for_idle(struct comedi_device *dev)
if ((status & NI67XX_CAL_STATUS_BUSY) == 0)
break;
set_current_state(TASK_INTERRUPTIBLE);
- if (schedule_timeout(1))
+ if (schedule_min_hrtimeout())
return -EIO;
}
if (i == timeout) {
diff --git a/fs/afs/vlocation.c b/fs/afs/vlocation.c
index 45a8639..855d08e 100644
--- a/fs/afs/vlocation.c
+++ b/fs/afs/vlocation.c
@@ -129,7 +129,7 @@ static int afs_vlocation_access_vl_by_id(struct afs_vlocation *vl,
if (vl->upd_busy_cnt > 1) {
/* second+ BUSY - sleep a little bit */
set_current_state(TASK_UNINTERRUPTIBLE);
- schedule_timeout(1);
+ schedule_min_hrtimeout();
}
continue;
}
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index dcd2e79..16bf891 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -5952,7 +5952,7 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes)
if (flush != BTRFS_RESERVE_NO_FLUSH &&
btrfs_transaction_in_commit(fs_info))
- schedule_timeout(1);
+ schedule_min_hrtimeout();
if (delalloc_lock)
mutex_lock(&BTRFS_I(inode)->delalloc_mutex);
diff --git a/fs/btrfs/inode-map.c b/fs/btrfs/inode-map.c
index 144b119..03d2e8e 100644
--- a/fs/btrfs/inode-map.c
+++ b/fs/btrfs/inode-map.c
@@ -89,7 +89,7 @@ static int caching_kthread(void *data)
btrfs_release_path(path);
root->ino_cache_progress = last;
up_read(&fs_info->commit_root_sem);
- schedule_timeout(1);
+ schedule_min_hrtimeout();
goto again;
} else
continue;
diff --git a/sound/usb/line6/pcm.c b/sound/usb/line6/pcm.c
index fab53f5..fda1ab5 100644
--- a/sound/usb/line6/pcm.c
+++ b/sound/usb/line6/pcm.c
@@ -131,7 +131,7 @@ static void line6_wait_clear_audio_urbs(struct snd_line6_pcm *line6pcm,
if (!alive)
break;
set_current_state(TASK_UNINTERRUPTIBLE);
- schedule_timeout(1);
+ schedule_min_hrtimeout();
} while (--timeout > 0);
if (alive)
dev_err(line6pcm->line6->ifcdev,
--
2.9.3

397
patches/0013-Change-all-schedule_timeout-with-msecs_to_jiffies-po.patch Normal file
View File

@ -0,0 +1,397 @@
From 56e8b01452fbb6c1aa85b0a52fbd352fddf7e959 Mon Sep 17 00:00:00 2001
From: Con Kolivas <kernel@kolivas.org>
Date: Mon, 20 Feb 2017 13:29:16 +1100
Subject: [PATCH 13/25] Change all schedule_timeout with msecs_to_jiffies
potentially under 50ms to use schedule_msec_hrtimeout.
---
drivers/bluetooth/hci_qca.c | 2 +-
drivers/char/snsc.c | 4 ++--
drivers/media/pci/ivtv/ivtv-ioctl.c | 2 +-
drivers/media/pci/ivtv/ivtv-streams.c | 2 +-
drivers/net/can/usb/peak_usb/pcan_usb.c | 2 +-
drivers/net/usb/lan78xx.c | 2 +-
drivers/net/usb/usbnet.c | 2 +-
drivers/scsi/fnic/fnic_scsi.c | 4 ++--
drivers/scsi/snic/snic_scsi.c | 2 +-
drivers/staging/lustre/lnet/lnet/lib-eq.c | 2 +-
drivers/staging/rts5208/rtsx.c | 2 +-
drivers/staging/speakup/speakup_acntpc.c | 4 ++--
drivers/staging/speakup/speakup_apollo.c | 2 +-
drivers/staging/speakup/speakup_decext.c | 2 +-
drivers/staging/speakup/speakup_decpc.c | 2 +-
drivers/staging/speakup/speakup_dectlk.c | 2 +-
drivers/staging/speakup/speakup_dtlk.c | 4 ++--
drivers/staging/speakup/speakup_keypc.c | 4 ++--
drivers/staging/speakup/synth.c | 2 +-
drivers/staging/unisys/visornic/visornic_main.c | 6 +++---
drivers/target/target_core_user.c | 2 +-
drivers/video/fbdev/omap/hwa742.c | 2 +-
drivers/video/fbdev/pxafb.c | 2 +-
23 files changed, 30 insertions(+), 30 deletions(-)
diff --git a/drivers/bluetooth/hci_qca.c b/drivers/bluetooth/hci_qca.c
index 05c2307..6954d29 100644
--- a/drivers/bluetooth/hci_qca.c
+++ b/drivers/bluetooth/hci_qca.c
@@ -880,7 +880,7 @@ static int qca_set_baudrate(struct hci_dev *hdev, uint8_t baudrate)
* then host can communicate with new baudrate to controller
*/
set_current_state(TASK_UNINTERRUPTIBLE);
- schedule_timeout(msecs_to_jiffies(BAUDRATE_SETTLE_TIMEOUT_MS));
+ schedule_msec_hrtimeout((BAUDRATE_SETTLE_TIMEOUT_MS));
set_current_state(TASK_INTERRUPTIBLE);
return 0;
diff --git a/drivers/char/snsc.c b/drivers/char/snsc.c
index ec07f0e..3410b46 100644
--- a/drivers/char/snsc.c
+++ b/drivers/char/snsc.c
@@ -198,7 +198,7 @@ scdrv_read(struct file *file, char __user *buf, size_t count, loff_t *f_pos)
add_wait_queue(&sd->sd_rq, &wait);
spin_unlock_irqrestore(&sd->sd_rlock, flags);
- schedule_timeout(msecs_to_jiffies(SCDRV_TIMEOUT));
+ schedule_msec_hrtimeout((SCDRV_TIMEOUT));
remove_wait_queue(&sd->sd_rq, &wait);
if (signal_pending(current)) {
@@ -294,7 +294,7 @@ scdrv_write(struct file *file, const char __user *buf,
add_wait_queue(&sd->sd_wq, &wait);
spin_unlock_irqrestore(&sd->sd_wlock, flags);
- schedule_timeout(msecs_to_jiffies(SCDRV_TIMEOUT));
+ schedule_msec_hrtimeout((SCDRV_TIMEOUT));
remove_wait_queue(&sd->sd_wq, &wait);
if (signal_pending(current)) {
diff --git a/drivers/media/pci/ivtv/ivtv-ioctl.c b/drivers/media/pci/ivtv/ivtv-ioctl.c
index 2dc4b20..8e061cf 100644
--- a/drivers/media/pci/ivtv/ivtv-ioctl.c
+++ b/drivers/media/pci/ivtv/ivtv-ioctl.c
@@ -1151,7 +1151,7 @@ void ivtv_s_std_dec(struct ivtv *itv, v4l2_std_id std)
TASK_UNINTERRUPTIBLE);
if ((read_reg(IVTV_REG_DEC_LINE_FIELD) >> 16) < 100)
break;
- schedule_timeout(msecs_to_jiffies(25));
+ schedule_msec_hrtimeout((25));
}
finish_wait(&itv->vsync_waitq, &wait);
mutex_lock(&itv->serialize_lock);
diff --git a/drivers/media/pci/ivtv/ivtv-streams.c b/drivers/media/pci/ivtv/ivtv-streams.c
index d27c6df..e9ffc4e 100644
--- a/drivers/media/pci/ivtv/ivtv-streams.c
+++ b/drivers/media/pci/ivtv/ivtv-streams.c
@@ -834,7 +834,7 @@ int ivtv_stop_v4l2_encode_stream(struct ivtv_stream *s, int gop_end)
while (!test_bit(IVTV_F_I_EOS, &itv->i_flags) &&
time_before(jiffies,
then + msecs_to_jiffies(2000))) {
- schedule_timeout(msecs_to_jiffies(10));
+ schedule_msec_hrtimeout((10));
}
/* To convert jiffies to ms, we must multiply by 1000
diff --git a/drivers/net/can/usb/peak_usb/pcan_usb.c b/drivers/net/can/usb/peak_usb/pcan_usb.c
index 838545c..34f8972 100644
--- a/drivers/net/can/usb/peak_usb/pcan_usb.c
+++ b/drivers/net/can/usb/peak_usb/pcan_usb.c
@@ -250,7 +250,7 @@ static int pcan_usb_write_mode(struct peak_usb_device *dev, u8 onoff)
} else {
/* the PCAN-USB needs time to init */
set_current_state(TASK_INTERRUPTIBLE);
- schedule_timeout(msecs_to_jiffies(PCAN_USB_STARTUP_TIMEOUT));
+ schedule_msec_hrtimeout((PCAN_USB_STARTUP_TIMEOUT));
}
return err;
diff --git a/drivers/net/usb/lan78xx.c b/drivers/net/usb/lan78xx.c
index 08f8703..3b3bc86 100644
--- a/drivers/net/usb/lan78xx.c
+++ b/drivers/net/usb/lan78xx.c
@@ -2544,7 +2544,7 @@ static void lan78xx_terminate_urbs(struct lan78xx_net *dev)
while (!skb_queue_empty(&dev->rxq) &&
!skb_queue_empty(&dev->txq) &&
!skb_queue_empty(&dev->done)) {
- schedule_timeout(msecs_to_jiffies(UNLINK_TIMEOUT_MS));
+ schedule_msec_hrtimeout((UNLINK_TIMEOUT_MS));
set_current_state(TASK_UNINTERRUPTIBLE);
netif_dbg(dev, ifdown, dev->net,
"waited for %d urb completions\n", temp);
diff --git a/drivers/net/usb/usbnet.c b/drivers/net/usb/usbnet.c
index 3de65ea..f8a4b18 100644
--- a/drivers/net/usb/usbnet.c
+++ b/drivers/net/usb/usbnet.c
@@ -767,7 +767,7 @@ static void wait_skb_queue_empty(struct sk_buff_head *q)
spin_lock_irqsave(&q->lock, flags);
while (!skb_queue_empty(q)) {
spin_unlock_irqrestore(&q->lock, flags);
- schedule_timeout(msecs_to_jiffies(UNLINK_TIMEOUT_MS));
+ schedule_msec_hrtimeout((UNLINK_TIMEOUT_MS));
set_current_state(TASK_UNINTERRUPTIBLE);
spin_lock_irqsave(&q->lock, flags);
}
diff --git a/drivers/scsi/fnic/fnic_scsi.c b/drivers/scsi/fnic/fnic_scsi.c
index adb3d58..de73e78 100644
--- a/drivers/scsi/fnic/fnic_scsi.c
+++ b/drivers/scsi/fnic/fnic_scsi.c
@@ -217,7 +217,7 @@ int fnic_fw_reset_handler(struct fnic *fnic)
/* wait for io cmpl */
while (atomic_read(&fnic->in_flight))
- schedule_timeout(msecs_to_jiffies(1));
+ schedule_msec_hrtimeout((1));
spin_lock_irqsave(&fnic->wq_copy_lock[0], flags);
@@ -2201,7 +2201,7 @@ static int fnic_clean_pending_aborts(struct fnic *fnic,
}
}
- schedule_timeout(msecs_to_jiffies(2 * fnic->config.ed_tov));
+ schedule_msec_hrtimeout((2 * fnic->config.ed_tov));
/* walk again to check, if IOs are still pending in fw */
if (fnic_is_abts_pending(fnic, lr_sc))
diff --git a/drivers/scsi/snic/snic_scsi.c b/drivers/scsi/snic/snic_scsi.c
index abada16..0bf30dc 100644
--- a/drivers/scsi/snic/snic_scsi.c
+++ b/drivers/scsi/snic/snic_scsi.c
@@ -2356,7 +2356,7 @@ snic_reset(struct Scsi_Host *shost, struct scsi_cmnd *sc)
/* Wait for all the IOs that are entered in Qcmd */
while (atomic_read(&snic->ios_inflight))
- schedule_timeout(msecs_to_jiffies(1));
+ schedule_msec_hrtimeout((1));
ret = snic_issue_hba_reset(snic, sc);
if (ret) {
diff --git a/drivers/staging/lustre/lnet/lnet/lib-eq.c b/drivers/staging/lustre/lnet/lnet/lib-eq.c
index d05c6cc..3f62b6f 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-eq.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-eq.c
@@ -328,7 +328,7 @@ __must_hold(&the_lnet.ln_eq_wait_lock)
schedule();
} else {
now = jiffies;
- schedule_timeout(msecs_to_jiffies(tms));
+ schedule_msec_hrtimeout((tms));
tms -= jiffies_to_msecs(jiffies - now);
if (tms < 0) /* no more wait but may have new event */
tms = 0;
diff --git a/drivers/staging/rts5208/rtsx.c b/drivers/staging/rts5208/rtsx.c
index 68d75d0..aef88c4 100644
--- a/drivers/staging/rts5208/rtsx.c
+++ b/drivers/staging/rts5208/rtsx.c
@@ -537,7 +537,7 @@ static int rtsx_polling_thread(void *__dev)
for (;;) {
set_current_state(TASK_INTERRUPTIBLE);
- schedule_timeout(msecs_to_jiffies(POLLING_INTERVAL));
+ schedule_msec_hrtimeout((POLLING_INTERVAL));
/* lock the device pointers */
mutex_lock(&dev->dev_mutex);
diff --git a/drivers/staging/speakup/speakup_acntpc.c b/drivers/staging/speakup/speakup_acntpc.c
index efb791b..fd02fb2 100644
--- a/drivers/staging/speakup/speakup_acntpc.c
+++ b/drivers/staging/speakup/speakup_acntpc.c
@@ -204,7 +204,7 @@ static void do_catch_up(struct spk_synth *synth)
full_time_val = full_time->u.n.value;
spin_unlock_irqrestore(&speakup_info.spinlock, flags);
if (synth_full()) {
- schedule_timeout(msecs_to_jiffies(full_time_val));
+ schedule_msec_hrtimeout((full_time_val));
continue;
}
set_current_state(TASK_RUNNING);
@@ -232,7 +232,7 @@ static void do_catch_up(struct spk_synth *synth)
jiffy_delta_val = jiffy_delta->u.n.value;
delay_time_val = delay_time->u.n.value;
spin_unlock_irqrestore(&speakup_info.spinlock, flags);
- schedule_timeout(msecs_to_jiffies(delay_time_val));
+ schedule_msec_hrtimeout((delay_time_val));
jiff_max = jiffies+jiffy_delta_val;
}
}
diff --git a/drivers/staging/speakup/speakup_apollo.c b/drivers/staging/speakup/speakup_apollo.c
index 3cbc8a7..3c17854 100644
--- a/drivers/staging/speakup/speakup_apollo.c
+++ b/drivers/staging/speakup/speakup_apollo.c
@@ -172,7 +172,7 @@ static void do_catch_up(struct spk_synth *synth)
outb(UART_MCR_DTR, speakup_info.port_tts + UART_MCR);
outb(UART_MCR_DTR | UART_MCR_RTS,
speakup_info.port_tts + UART_MCR);
- schedule_timeout(msecs_to_jiffies(full_time_val));
+ schedule_msec_hrtimeout((full_time_val));
continue;
}
if (time_after_eq(jiffies, jiff_max) && (ch == SPACE)) {
diff --git a/drivers/staging/speakup/speakup_decext.c b/drivers/staging/speakup/speakup_decext.c
index 1a5cf3d..fa2b4e1 100644
--- a/drivers/staging/speakup/speakup_decext.c
+++ b/drivers/staging/speakup/speakup_decext.c
@@ -186,7 +186,7 @@ static void do_catch_up(struct spk_synth *synth)
if (ch == '\n')
ch = 0x0D;
if (synth_full() || !spk_serial_out(ch)) {
- schedule_timeout(msecs_to_jiffies(delay_time_val));
+ schedule_msec_hrtimeout((delay_time_val));
continue;
}
set_current_state(TASK_RUNNING);
diff --git a/drivers/staging/speakup/speakup_decpc.c b/drivers/staging/speakup/speakup_decpc.c
index d6479bd..f7554bf 100644
--- a/drivers/staging/speakup/speakup_decpc.c
+++ b/drivers/staging/speakup/speakup_decpc.c
@@ -403,7 +403,7 @@ static void do_catch_up(struct spk_synth *synth)
if (ch == '\n')
ch = 0x0D;
if (dt_sendchar(ch)) {
- schedule_timeout(msecs_to_jiffies(delay_time_val));
+ schedule_msec_hrtimeout((delay_time_val));
continue;
}
set_current_state(TASK_RUNNING);
diff --git a/drivers/staging/speakup/speakup_dectlk.c b/drivers/staging/speakup/speakup_dectlk.c
index 7646567..639192e 100644
--- a/drivers/staging/speakup/speakup_dectlk.c
+++ b/drivers/staging/speakup/speakup_dectlk.c
@@ -251,7 +251,7 @@ static void do_catch_up(struct spk_synth *synth)
if (ch == '\n')
ch = 0x0D;
if (synth_full_val || !spk_serial_out(ch)) {
- schedule_timeout(msecs_to_jiffies(delay_time_val));
+ schedule_msec_hrtimeout((delay_time_val));
continue;
}
set_current_state(TASK_RUNNING);
diff --git a/drivers/staging/speakup/speakup_dtlk.c b/drivers/staging/speakup/speakup_dtlk.c
index 38aa401..1640519 100644
--- a/drivers/staging/speakup/speakup_dtlk.c
+++ b/drivers/staging/speakup/speakup_dtlk.c
@@ -217,7 +217,7 @@ static void do_catch_up(struct spk_synth *synth)
delay_time_val = delay_time->u.n.value;
spin_unlock_irqrestore(&speakup_info.spinlock, flags);
if (synth_full()) {
- schedule_timeout(msecs_to_jiffies(delay_time_val));
+ schedule_msec_hrtimeout((delay_time_val));
continue;
}
set_current_state(TASK_RUNNING);
@@ -233,7 +233,7 @@ static void do_catch_up(struct spk_synth *synth)
delay_time_val = delay_time->u.n.value;
jiffy_delta_val = jiffy_delta->u.n.value;
spin_unlock_irqrestore(&speakup_info.spinlock, flags);
- schedule_timeout(msecs_to_jiffies(delay_time_val));
+ schedule_msec_hrtimeout((delay_time_val));
jiff_max = jiffies + jiffy_delta_val;
}
}
diff --git a/drivers/staging/speakup/speakup_keypc.c b/drivers/staging/speakup/speakup_keypc.c
index 5e2170b..30b5df7 100644
--- a/drivers/staging/speakup/speakup_keypc.c
+++ b/drivers/staging/speakup/speakup_keypc.c
@@ -206,7 +206,7 @@ spin_lock_irqsave(&speakup_info.spinlock, flags);
full_time_val = full_time->u.n.value;
spin_unlock_irqrestore(&speakup_info.spinlock, flags);
if (synth_full()) {
- schedule_timeout(msecs_to_jiffies(full_time_val));
+ schedule_msec_hrtimeout((full_time_val));
continue;
}
set_current_state(TASK_RUNNING);
@@ -239,7 +239,7 @@ spin_lock_irqsave(&speakup_info.spinlock, flags);
jiffy_delta_val = jiffy_delta->u.n.value;
delay_time_val = delay_time->u.n.value;
spin_unlock_irqrestore(&speakup_info.spinlock, flags);
- schedule_timeout(msecs_to_jiffies(delay_time_val));
+ schedule_msec_hrtimeout((delay_time_val));
jiff_max = jiffies+jiffy_delta_val;
}
}
diff --git a/drivers/staging/speakup/synth.c b/drivers/staging/speakup/synth.c
index a61c02b..14299e5 100644
--- a/drivers/staging/speakup/synth.c
+++ b/drivers/staging/speakup/synth.c
@@ -120,7 +120,7 @@ void spk_do_catch_up(struct spk_synth *synth)
if (ch == '\n')
ch = synth->procspeech;
if (!spk_serial_out(ch)) {
- schedule_timeout(msecs_to_jiffies(full_time_val));
+ schedule_msec_hrtimeout((full_time_val));
continue;
}
if (time_after_eq(jiffies, jiff_max) && (ch == SPACE)) {
diff --git a/drivers/staging/unisys/visornic/visornic_main.c b/drivers/staging/unisys/visornic/visornic_main.c
index c1f674f..4f30a7a 100644
--- a/drivers/staging/unisys/visornic/visornic_main.c
+++ b/drivers/staging/unisys/visornic/visornic_main.c
@@ -468,7 +468,7 @@ visornic_disable_with_timeout(struct net_device *netdev, const int timeout)
}
set_current_state(TASK_INTERRUPTIBLE);
spin_unlock_irqrestore(&devdata->priv_lock, flags);
- wait += schedule_timeout(msecs_to_jiffies(10));
+ wait += schedule_msec_hrtimeout((10));
spin_lock_irqsave(&devdata->priv_lock, flags);
}
@@ -479,7 +479,7 @@ visornic_disable_with_timeout(struct net_device *netdev, const int timeout)
while (1) {
set_current_state(TASK_INTERRUPTIBLE);
spin_unlock_irqrestore(&devdata->priv_lock, flags);
- schedule_timeout(msecs_to_jiffies(10));
+ schedule_msec_hrtimeout((10));
spin_lock_irqsave(&devdata->priv_lock, flags);
if (atomic_read(&devdata->usage))
break;
@@ -611,7 +611,7 @@ visornic_enable_with_timeout(struct net_device *netdev, const int timeout)
}
set_current_state(TASK_INTERRUPTIBLE);
spin_unlock_irqrestore(&devdata->priv_lock, flags);
- wait += schedule_timeout(msecs_to_jiffies(10));
+ wait += schedule_msec_hrtimeout((10));
spin_lock_irqsave(&devdata->priv_lock, flags);
}
diff --git a/drivers/target/target_core_user.c b/drivers/target/target_core_user.c
index 8041710..f907a81 100644
--- a/drivers/target/target_core_user.c
+++ b/drivers/target/target_core_user.c
@@ -451,7 +451,7 @@ tcmu_queue_cmd_ring(struct tcmu_cmd *tcmu_cmd)
pr_debug("sleeping for ring space\n");
spin_unlock_irq(&udev->cmdr_lock);
- ret = schedule_timeout(msecs_to_jiffies(TCMU_TIME_OUT));
+ ret = schedule_msec_hrtimeout((TCMU_TIME_OUT));
finish_wait(&udev->wait_cmdr, &__wait);
if (!ret) {
pr_warn("tcmu: command timed out\n");
diff --git a/drivers/video/fbdev/omap/hwa742.c b/drivers/video/fbdev/omap/hwa742.c
index a4ee65b..cf38bcb 100644
--- a/drivers/video/fbdev/omap/hwa742.c
+++ b/drivers/video/fbdev/omap/hwa742.c
@@ -926,7 +926,7 @@ static void hwa742_resume(void)
if (hwa742_read_reg(HWA742_PLL_DIV_REG) & (1 << 7))
break;
set_current_state(TASK_UNINTERRUPTIBLE);
- schedule_timeout(msecs_to_jiffies(5));
+ schedule_msec_hrtimeout((5));
}
hwa742_set_update_mode(hwa742.update_mode_before_suspend);
}
diff --git a/drivers/video/fbdev/pxafb.c b/drivers/video/fbdev/pxafb.c
index ef73f14..7b5483b 100644
--- a/drivers/video/fbdev/pxafb.c
+++ b/drivers/video/fbdev/pxafb.c
@@ -1287,7 +1287,7 @@ static int pxafb_smart_thread(void *arg)
mutex_unlock(&fbi->ctrlr_lock);
set_current_state(TASK_INTERRUPTIBLE);
- schedule_timeout(msecs_to_jiffies(30));
+ schedule_msec_hrtimeout((30));
}
pr_debug("%s(): task ending\n", __func__);
--
2.9.3

325
patches/0014-Replace-all-calls-to-schedule_timeout_interruptible-.patch Normal file
View File

@ -0,0 +1,325 @@
From af94b59651831b7e176ce8cb98441bdccb87eac0 Mon Sep 17 00:00:00 2001
From: Con Kolivas <kernel@kolivas.org>
Date: Mon, 20 Feb 2017 13:30:07 +1100
Subject: [PATCH 14/25] Replace all calls to schedule_timeout_interruptible of
potentially under 50ms to use schedule_msec_hrtimeout_interruptible.
---
drivers/hwmon/fam15h_power.c | 2 +-
drivers/iio/light/tsl2563.c | 6 +-----
drivers/media/i2c/msp3400-driver.c | 4 ++--
drivers/media/pci/ivtv/ivtv-gpio.c | 6 +++---
drivers/media/radio/radio-mr800.c | 2 +-
drivers/media/radio/radio-tea5777.c | 2 +-
drivers/media/radio/tea575x.c | 2 +-
drivers/misc/panel.c | 2 +-
drivers/parport/ieee1284.c | 2 +-
drivers/parport/ieee1284_ops.c | 2 +-
drivers/platform/x86/intel_ips.c | 8 ++++----
net/core/pktgen.c | 2 +-
sound/soc/codecs/wm8350.c | 12 ++++++------
sound/soc/codecs/wm8900.c | 2 +-
sound/soc/codecs/wm9713.c | 4 ++--
15 files changed, 27 insertions(+), 31 deletions(-)
diff --git a/drivers/hwmon/fam15h_power.c b/drivers/hwmon/fam15h_power.c
index 15aa49d..991e8a7 100644
--- a/drivers/hwmon/fam15h_power.c
+++ b/drivers/hwmon/fam15h_power.c
@@ -238,7 +238,7 @@ static ssize_t acc_show_power(struct device *dev,
prev_ptsc[cu] = data->cpu_sw_pwr_ptsc[cu];
}
- leftover = schedule_timeout_interruptible(msecs_to_jiffies(data->power_period));
+ leftover = schedule_msec_hrtimeout_interruptible((data->power_period));
if (leftover)
return 0;
diff --git a/drivers/iio/light/tsl2563.c b/drivers/iio/light/tsl2563.c
index 04598ae..a8c095d 100644
--- a/drivers/iio/light/tsl2563.c
+++ b/drivers/iio/light/tsl2563.c
@@ -282,11 +282,7 @@ static void tsl2563_wait_adc(struct tsl2563_chip *chip)
default:
delay = 402;
}
- /*
- * TODO: Make sure that we wait at least required delay but why we
- * have to extend it one tick more?
- */
- schedule_timeout_interruptible(msecs_to_jiffies(delay) + 2);
+ schedule_msec_hrtimeout_interruptible(delay + 1);
}
static int tsl2563_adjust_gainlevel(struct tsl2563_chip *chip, u16 adc)
diff --git a/drivers/media/i2c/msp3400-driver.c b/drivers/media/i2c/msp3400-driver.c
index 201a9800..5cebabc 100644
--- a/drivers/media/i2c/msp3400-driver.c
+++ b/drivers/media/i2c/msp3400-driver.c
@@ -184,7 +184,7 @@ static int msp_read(struct i2c_client *client, int dev, int addr)
break;
dev_warn(&client->dev, "I/O error #%d (read 0x%02x/0x%02x)\n", err,
dev, addr);
- schedule_timeout_interruptible(msecs_to_jiffies(10));
+ schedule_msec_hrtimeout_interruptible((10));
}
if (err == 3) {
dev_warn(&client->dev, "resetting chip, sound will go off.\n");
@@ -225,7 +225,7 @@ static int msp_write(struct i2c_client *client, int dev, int addr, int val)
break;
dev_warn(&client->dev, "I/O error #%d (write 0x%02x/0x%02x)\n", err,
dev, addr);
- schedule_timeout_interruptible(msecs_to_jiffies(10));
+ schedule_msec_hrtimeout_interruptible((10));
}
if (err == 3) {
dev_warn(&client->dev, "resetting chip, sound will go off.\n");
diff --git a/drivers/media/pci/ivtv/ivtv-gpio.c b/drivers/media/pci/ivtv/ivtv-gpio.c
index f752f39..23372af6 100644
--- a/drivers/media/pci/ivtv/ivtv-gpio.c
+++ b/drivers/media/pci/ivtv/ivtv-gpio.c
@@ -117,7 +117,7 @@ void ivtv_reset_ir_gpio(struct ivtv *itv)
curout = (curout & ~0xF) | 1;
write_reg(curout, IVTV_REG_GPIO_OUT);
/* We could use something else for smaller time */
- schedule_timeout_interruptible(msecs_to_jiffies(1));
+ schedule_msec_hrtimeout_interruptible((1));
curout |= 2;
write_reg(curout, IVTV_REG_GPIO_OUT);
curdir &= ~0x80;
@@ -137,11 +137,11 @@ int ivtv_reset_tuner_gpio(void *dev, int component, int cmd, int value)
curout = read_reg(IVTV_REG_GPIO_OUT);
curout &= ~(1 << itv->card->xceive_pin);
write_reg(curout, IVTV_REG_GPIO_OUT);
- schedule_timeout_interruptible(msecs_to_jiffies(1));
+ schedule_msec_hrtimeout_interruptible((1));
curout |= 1 << itv->card->xceive_pin;
write_reg(curout, IVTV_REG_GPIO_OUT);
- schedule_timeout_interruptible(msecs_to_jiffies(1));
+ schedule_msec_hrtimeout_interruptible((1));
return 0;
}
diff --git a/drivers/media/radio/radio-mr800.c b/drivers/media/radio/radio-mr800.c
index c2927fd..bdee269 100644
--- a/drivers/media/radio/radio-mr800.c
+++ b/drivers/media/radio/radio-mr800.c
@@ -382,7 +382,7 @@ static int vidioc_s_hw_freq_seek(struct file *file, void *priv,
retval = -ENODATA;
break;
}
- if (schedule_timeout_interruptible(msecs_to_jiffies(10))) {
+ if (schedule_msec_hrtimeout_interruptible((10))) {
retval = -ERESTARTSYS;
break;
}
diff --git a/drivers/media/radio/radio-tea5777.c b/drivers/media/radio/radio-tea5777.c
index 83fe7ab..aaae5fa 100644
--- a/drivers/media/radio/radio-tea5777.c
+++ b/drivers/media/radio/radio-tea5777.c
@@ -249,7 +249,7 @@ static int radio_tea5777_update_read_reg(struct radio_tea5777 *tea, int wait)
}
if (wait) {
- if (schedule_timeout_interruptible(msecs_to_jiffies(wait)))
+ if (schedule_msec_hrtimeout_interruptible((wait)))
return -ERESTARTSYS;
}
diff --git a/drivers/media/radio/tea575x.c b/drivers/media/radio/tea575x.c
index 4dc2067..29f4416 100644
--- a/drivers/media/radio/tea575x.c
+++ b/drivers/media/radio/tea575x.c
@@ -416,7 +416,7 @@ int snd_tea575x_s_hw_freq_seek(struct file *file, struct snd_tea575x *tea,
for (;;) {
if (time_after(jiffies, timeout))
break;
- if (schedule_timeout_interruptible(msecs_to_jiffies(10))) {
+ if (schedule_msec_hrtimeout_interruptible((10))) {
/* some signal arrived, stop search */
tea->val &= ~TEA575X_BIT_SEARCH;
snd_tea575x_set_freq(tea);
diff --git a/drivers/misc/panel.c b/drivers/misc/panel.c
index 6030ac5..f0c1a101 100644
--- a/drivers/misc/panel.c
+++ b/drivers/misc/panel.c
@@ -760,7 +760,7 @@ static void long_sleep(int ms)
if (in_interrupt())
mdelay(ms);
else
- schedule_timeout_interruptible(msecs_to_jiffies(ms));
+ schedule_msec_hrtimeout_interruptible((ms));
}
/*
diff --git a/drivers/parport/ieee1284.c b/drivers/parport/ieee1284.c
index f9fd4b3..00ad2f3 100644
--- a/drivers/parport/ieee1284.c
+++ b/drivers/parport/ieee1284.c
@@ -215,7 +215,7 @@ int parport_wait_peripheral(struct parport *port,
/* parport_wait_event didn't time out, but the
* peripheral wasn't actually ready either.
* Wait for another 10ms. */
- schedule_timeout_interruptible(msecs_to_jiffies(10));
+ schedule_msec_hrtimeout_interruptible((10));
}
}
diff --git a/drivers/parport/ieee1284_ops.c b/drivers/parport/ieee1284_ops.c
index c0e7d21..e1b4fd4 100644
--- a/drivers/parport/ieee1284_ops.c
+++ b/drivers/parport/ieee1284_ops.c
@@ -536,7 +536,7 @@ size_t parport_ieee1284_ecp_read_data (struct parport *port,
/* Yield the port for a while. */
if (count && dev->port->irq != PARPORT_IRQ_NONE) {
parport_release (dev);
- schedule_timeout_interruptible(msecs_to_jiffies(40));
+ schedule_msec_hrtimeout_interruptible((40));
parport_claim_or_block (dev);
}
else
diff --git a/drivers/platform/x86/intel_ips.c b/drivers/platform/x86/intel_ips.c
index 55663b3..0363fed 100644
--- a/drivers/platform/x86/intel_ips.c
+++ b/drivers/platform/x86/intel_ips.c
@@ -812,7 +812,7 @@ static int ips_adjust(void *data)
ips_gpu_lower(ips);
sleep:
- schedule_timeout_interruptible(msecs_to_jiffies(IPS_ADJUST_PERIOD));
+ schedule_msec_hrtimeout_interruptible((IPS_ADJUST_PERIOD));
} while (!kthread_should_stop());
dev_dbg(&ips->dev->dev, "ips-adjust thread stopped\n");
@@ -991,7 +991,7 @@ static int ips_monitor(void *data)
seqno_timestamp = get_jiffies_64();
old_cpu_power = thm_readl(THM_CEC);
- schedule_timeout_interruptible(msecs_to_jiffies(IPS_SAMPLE_PERIOD));
+ schedule_msec_hrtimeout_interruptible((IPS_SAMPLE_PERIOD));
/* Collect an initial average */
for (i = 0; i < IPS_SAMPLE_COUNT; i++) {
@@ -1018,7 +1018,7 @@ static int ips_monitor(void *data)
mchp_samples[i] = mchp;
}
- schedule_timeout_interruptible(msecs_to_jiffies(IPS_SAMPLE_PERIOD));
+ schedule_msec_hrtimeout_interruptible((IPS_SAMPLE_PERIOD));
if (kthread_should_stop())
break;
}
@@ -1045,7 +1045,7 @@ static int ips_monitor(void *data)
* us to reduce the sample frequency if the CPU and GPU are idle.
*/
old_cpu_power = thm_readl(THM_CEC);
- schedule_timeout_interruptible(msecs_to_jiffies(IPS_SAMPLE_PERIOD));
+ schedule_msec_hrtimeout_interruptible((IPS_SAMPLE_PERIOD));
last_sample_period = IPS_SAMPLE_PERIOD;
setup_deferrable_timer_on_stack(&timer, monitor_timeout,
diff --git a/net/core/pktgen.c b/net/core/pktgen.c
index 8e69ce4..0227415 100644
--- a/net/core/pktgen.c
+++ b/net/core/pktgen.c
@@ -1992,7 +1992,7 @@ static void pktgen_mark_device(const struct pktgen_net *pn, const char *ifname)
mutex_unlock(&pktgen_thread_lock);
pr_debug("%s: waiting for %s to disappear....\n",
__func__, ifname);
- schedule_timeout_interruptible(msecs_to_jiffies(msec_per_try));
+ schedule_msec_hrtimeout_interruptible((msec_per_try));
mutex_lock(&pktgen_thread_lock);
if (++i >= max_tries) {
diff --git a/sound/soc/codecs/wm8350.c b/sound/soc/codecs/wm8350.c
index 2efc5b4..3e3248c 100644
--- a/sound/soc/codecs/wm8350.c
+++ b/sound/soc/codecs/wm8350.c
@@ -236,10 +236,10 @@ static void wm8350_pga_work(struct work_struct *work)
out2->ramp == WM8350_RAMP_UP) {
/* delay is longer over 0dB as increases are larger */
if (i >= WM8350_OUTn_0dB)
- schedule_timeout_interruptible(msecs_to_jiffies
+ schedule_msec_hrtimeout_interruptible(
(2));
else
- schedule_timeout_interruptible(msecs_to_jiffies
+ schedule_msec_hrtimeout_interruptible(
(1));
} else
udelay(50); /* doesn't matter if we delay longer */
@@ -1123,7 +1123,7 @@ static int wm8350_set_bias_level(struct snd_soc_codec *codec,
(platform->dis_out4 << 6));
/* wait for discharge */
- schedule_timeout_interruptible(msecs_to_jiffies
+ schedule_msec_hrtimeout_interruptible(
(platform->
cap_discharge_msecs));
@@ -1139,7 +1139,7 @@ static int wm8350_set_bias_level(struct snd_soc_codec *codec,
WM8350_VBUFEN);
/* wait for vmid */
- schedule_timeout_interruptible(msecs_to_jiffies
+ schedule_msec_hrtimeout_interruptible(
(platform->
vmid_charge_msecs));
@@ -1190,7 +1190,7 @@ static int wm8350_set_bias_level(struct snd_soc_codec *codec,
wm8350_reg_write(wm8350, WM8350_POWER_MGMT_1, pm1);
/* wait */
- schedule_timeout_interruptible(msecs_to_jiffies
+ schedule_msec_hrtimeout_interruptible(
(platform->
vmid_discharge_msecs));
@@ -1208,7 +1208,7 @@ static int wm8350_set_bias_level(struct snd_soc_codec *codec,
pm1 | WM8350_OUTPUT_DRAIN_EN);
/* wait */
- schedule_timeout_interruptible(msecs_to_jiffies
+ schedule_msec_hrtimeout_interruptible(
(platform->drain_msecs));
pm1 &= ~WM8350_BIASEN;
diff --git a/sound/soc/codecs/wm8900.c b/sound/soc/codecs/wm8900.c
index c77b49a..fc50456 100644
--- a/sound/soc/codecs/wm8900.c
+++ b/sound/soc/codecs/wm8900.c
@@ -1112,7 +1112,7 @@ static int wm8900_set_bias_level(struct snd_soc_codec *codec,
/* Need to let things settle before stopping the clock
* to ensure that restart works, see "Stopping the
* master clock" in the datasheet. */
- schedule_timeout_interruptible(msecs_to_jiffies(1));
+ schedule_msec_hrtimeout_interruptible((1));
snd_soc_write(codec, WM8900_REG_POWER2,
WM8900_REG_POWER2_SYSCLK_ENA);
break;
diff --git a/sound/soc/codecs/wm9713.c b/sound/soc/codecs/wm9713.c
index 7e48221..0c85a20 100644
--- a/sound/soc/codecs/wm9713.c
+++ b/sound/soc/codecs/wm9713.c
@@ -199,7 +199,7 @@ static int wm9713_voice_shutdown(struct snd_soc_dapm_widget *w,
/* Gracefully shut down the voice interface. */
snd_soc_update_bits(codec, AC97_HANDSET_RATE, 0x0f00, 0x0200);
- schedule_timeout_interruptible(msecs_to_jiffies(1));
+ schedule_msec_hrtimeout_interruptible((1));
snd_soc_update_bits(codec, AC97_HANDSET_RATE, 0x0f00, 0x0f00);
snd_soc_update_bits(codec, AC97_EXTENDED_MID, 0x1000, 0x1000);
@@ -868,7 +868,7 @@ static int wm9713_set_pll(struct snd_soc_codec *codec,
wm9713->pll_in = freq_in;
/* wait 10ms AC97 link frames for the link to stabilise */
- schedule_timeout_interruptible(msecs_to_jiffies(10));
+ schedule_msec_hrtimeout_interruptible((10));
return 0;
}
--
2.9.3

160
patches/0015-Replace-all-calls-to-schedule_timeout_uninterruptibl.patch Normal file
View File

@ -0,0 +1,160 @@
From 1137ff2bfa5eb63b53747fe303fdb3937c5e1077 Mon Sep 17 00:00:00 2001
From: Con Kolivas <kernel@kolivas.org>
Date: Mon, 20 Feb 2017 13:30:32 +1100
Subject: [PATCH 15/25] Replace all calls to schedule_timeout_uninterruptible
of potentially under 50ms to use schedule_msec_hrtimeout_uninterruptible
---
drivers/media/pci/cx18/cx18-gpio.c | 4 ++--
drivers/net/wireless/intel/ipw2x00/ipw2100.c | 4 ++--
drivers/rtc/rtc-wm8350.c | 6 +++---
drivers/scsi/lpfc/lpfc_scsi.c | 2 +-
sound/pci/maestro3.c | 4 ++--
sound/soc/codecs/rt5631.c | 4 ++--
sound/soc/soc-dapm.c | 2 +-
7 files changed, 13 insertions(+), 13 deletions(-)
diff --git a/drivers/media/pci/cx18/cx18-gpio.c b/drivers/media/pci/cx18/cx18-gpio.c
index 38dc6b8..3cd3098 100644
--- a/drivers/media/pci/cx18/cx18-gpio.c
+++ b/drivers/media/pci/cx18/cx18-gpio.c
@@ -95,11 +95,11 @@ static void gpio_reset_seq(struct cx18 *cx, u32 active_lo, u32 active_hi,
/* Assert */
gpio_update(cx, mask, ~active_lo);
- schedule_timeout_uninterruptible(msecs_to_jiffies(assert_msecs));
+ schedule_msec_hrtimeout_uninterruptible((assert_msecs));
/* Deassert */
gpio_update(cx, mask, ~active_hi);
- schedule_timeout_uninterruptible(msecs_to_jiffies(recovery_msecs));
+ schedule_msec_hrtimeout_uninterruptible((recovery_msecs));
}
/*
diff --git a/drivers/net/wireless/intel/ipw2x00/ipw2100.c b/drivers/net/wireless/intel/ipw2x00/ipw2100.c
index 356aba9..d2cc761 100644
--- a/drivers/net/wireless/intel/ipw2x00/ipw2100.c
+++ b/drivers/net/wireless/intel/ipw2x00/ipw2100.c
@@ -830,7 +830,7 @@ static int ipw2100_hw_send_command(struct ipw2100_priv *priv,
* doesn't seem to have as many firmware restart cycles...
*
* As a test, we're sticking in a 1/100s delay here */
- schedule_timeout_uninterruptible(msecs_to_jiffies(10));
+ schedule_msec_hrtimeout_uninterruptible((10));
return 0;
@@ -1281,7 +1281,7 @@ static int ipw2100_start_adapter(struct ipw2100_priv *priv)
IPW_DEBUG_FW("Waiting for f/w initialization to complete...\n");
i = 5000;
do {
- schedule_timeout_uninterruptible(msecs_to_jiffies(40));
+ schedule_msec_hrtimeout_uninterruptible((40));
/* Todo... wait for sync command ... */
read_register(priv->net_dev, IPW_REG_INTA, &inta);
diff --git a/drivers/rtc/rtc-wm8350.c b/drivers/rtc/rtc-wm8350.c
index fa247de..f1a28d8 100644
--- a/drivers/rtc/rtc-wm8350.c
+++ b/drivers/rtc/rtc-wm8350.c
@@ -121,7 +121,7 @@ static int wm8350_rtc_settime(struct device *dev, struct rtc_time *tm)
/* Wait until confirmation of stopping */
do {
rtc_ctrl = wm8350_reg_read(wm8350, WM8350_RTC_TIME_CONTROL);
- schedule_timeout_uninterruptible(msecs_to_jiffies(1));
+ schedule_msec_hrtimeout_uninterruptible((1));
} while (--retries && !(rtc_ctrl & WM8350_RTC_STS));
if (!retries) {
@@ -204,7 +204,7 @@ static int wm8350_rtc_stop_alarm(struct wm8350 *wm8350)
/* Wait until confirmation of stopping */
do {
rtc_ctrl = wm8350_reg_read(wm8350, WM8350_RTC_TIME_CONTROL);
- schedule_timeout_uninterruptible(msecs_to_jiffies(1));
+ schedule_msec_hrtimeout_uninterruptible((1));
} while (retries-- && !(rtc_ctrl & WM8350_RTC_ALMSTS));
if (!(rtc_ctrl & WM8350_RTC_ALMSTS))
@@ -227,7 +227,7 @@ static int wm8350_rtc_start_alarm(struct wm8350 *wm8350)
/* Wait until confirmation */
do {
rtc_ctrl = wm8350_reg_read(wm8350, WM8350_RTC_TIME_CONTROL);
- schedule_timeout_uninterruptible(msecs_to_jiffies(1));
+ schedule_msec_hrtimeout_uninterruptible((1));
} while (retries-- && rtc_ctrl & WM8350_RTC_ALMSTS);
if (rtc_ctrl & WM8350_RTC_ALMSTS)
diff --git a/drivers/scsi/lpfc/lpfc_scsi.c b/drivers/scsi/lpfc/lpfc_scsi.c
index ad350d9..69a58a8 100644
--- a/drivers/scsi/lpfc/lpfc_scsi.c
+++ b/drivers/scsi/lpfc/lpfc_scsi.c
@@ -5109,7 +5109,7 @@ lpfc_reset_flush_io_context(struct lpfc_vport *vport, uint16_t tgt_id,
tgt_id, lun_id, context);
later = msecs_to_jiffies(2 * vport->cfg_devloss_tmo * 1000) + jiffies;
while (time_after(later, jiffies) && cnt) {
- schedule_timeout_uninterruptible(msecs_to_jiffies(20));
+ schedule_msec_hrtimeout_uninterruptible((20));
cnt = lpfc_sli_sum_iocb(vport, tgt_id, lun_id, context);
}
if (cnt) {
diff --git a/sound/pci/maestro3.c b/sound/pci/maestro3.c
index cafea6d..d374514 100644
--- a/sound/pci/maestro3.c
+++ b/sound/pci/maestro3.c
@@ -2016,7 +2016,7 @@ static void snd_m3_ac97_reset(struct snd_m3 *chip)
outw(0, io + GPIO_DATA);
outw(dir | GPO_PRIMARY_AC97, io + GPIO_DIRECTION);
- schedule_timeout_uninterruptible(msecs_to_jiffies(delay1));
+ schedule_msec_hrtimeout_uninterruptible((delay1));
outw(GPO_PRIMARY_AC97, io + GPIO_DATA);
udelay(5);
@@ -2024,7 +2024,7 @@ static void snd_m3_ac97_reset(struct snd_m3 *chip)
outw(IO_SRAM_ENABLE | SERIAL_AC_LINK_ENABLE, io + RING_BUS_CTRL_A);
outw(~0, io + GPIO_MASK);
- schedule_timeout_uninterruptible(msecs_to_jiffies(delay2));
+ schedule_msec_hrtimeout_uninterruptible((delay2));
if (! snd_m3_try_read_vendor(chip))
break;
diff --git a/sound/soc/codecs/rt5631.c b/sound/soc/codecs/rt5631.c
index 0e41808..611cb9f 100644
--- a/sound/soc/codecs/rt5631.c
+++ b/sound/soc/codecs/rt5631.c
@@ -419,7 +419,7 @@ static void onebit_depop_mute_stage(struct snd_soc_codec *codec, int enable)
hp_zc = snd_soc_read(codec, RT5631_INT_ST_IRQ_CTRL_2);
snd_soc_write(codec, RT5631_INT_ST_IRQ_CTRL_2, hp_zc & 0xf7ff);
if (enable) {
- schedule_timeout_uninterruptible(msecs_to_jiffies(10));
+ schedule_msec_hrtimeout_uninterruptible((10));
/* config one-bit depop parameter */
rt5631_write_index(codec, RT5631_SPK_INTL_CTRL, 0x307f);
snd_soc_update_bits(codec, RT5631_HP_OUT_VOL,
@@ -529,7 +529,7 @@ static void depop_seq_mute_stage(struct snd_soc_codec *codec, int enable)
hp_zc = snd_soc_read(codec, RT5631_INT_ST_IRQ_CTRL_2);
snd_soc_write(codec, RT5631_INT_ST_IRQ_CTRL_2, hp_zc & 0xf7ff);
if (enable) {
- schedule_timeout_uninterruptible(msecs_to_jiffies(10));
+ schedule_msec_hrtimeout_uninterruptible((10));
/* config depop sequence parameter */
rt5631_write_index(codec, RT5631_SPK_INTL_CTRL, 0x302f);
diff --git a/sound/soc/soc-dapm.c b/sound/soc/soc-dapm.c
index 27dd02e..7ba49f4 100644
--- a/sound/soc/soc-dapm.c
+++ b/sound/soc/soc-dapm.c
@@ -134,7 +134,7 @@ static void dapm_assert_locked(struct snd_soc_dapm_context *dapm)
static void pop_wait(u32 pop_time)
{
if (pop_time)
- schedule_timeout_uninterruptible(msecs_to_jiffies(pop_time));
+ schedule_msec_hrtimeout_uninterruptible((pop_time));
}
static void pop_dbg(struct device *dev, u32 pop_time, const char *fmt, ...)
--
2.9.3

84
patches/0016-Fix-build-for-disabled-highres-timers-with-hrtimeout.patch Normal file
View File

@ -0,0 +1,84 @@
From 37496baeea800e745a77620e90660496135f7fa5 Mon Sep 17 00:00:00 2001
From: Con Kolivas <kernel@kolivas.org>
Date: Mon, 20 Feb 2017 13:31:42 +1100
Subject: [PATCH 16/25] Fix build for disabled highres timers with hrtimeout
code.
---
include/linux/freezer.h | 1 +
include/linux/sched.h | 22 ++++++++++++++++++++++
kernel/time/timer.c | 2 ++
3 files changed, 25 insertions(+)
diff --git a/include/linux/freezer.h b/include/linux/freezer.h
index dd03e83..2fda682 100644
--- a/include/linux/freezer.h
+++ b/include/linux/freezer.h
@@ -296,6 +296,7 @@ static inline void set_freezable(void) {}
#define wait_event_freezekillable_unsafe(wq, condition) \
wait_event_killable(wq, condition)
+#define pm_freezing (false)
#endif /* !CONFIG_FREEZER */
#endif /* FREEZER_H_INCLUDED */
diff --git a/include/linux/sched.h b/include/linux/sched.h
index 46544f4..680494d 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -458,10 +458,32 @@ extern signed long schedule_timeout_killable(signed long timeout);
extern signed long schedule_timeout_uninterruptible(signed long timeout);
extern signed long schedule_timeout_idle(signed long timeout);
+#ifdef CONFIG_HIGH_RES_TIMERS
extern signed long schedule_msec_hrtimeout(signed long timeout);
extern signed long schedule_min_hrtimeout(void);
extern signed long schedule_msec_hrtimeout_interruptible(signed long timeout);
extern signed long schedule_msec_hrtimeout_uninterruptible(signed long timeout);
+#else
+static inline signed long schedule_msec_hrtimeout(signed long timeout)
+{
+ return schedule_timeout(msecs_to_jiffies(timeout));
+}
+
+static inline signed long schedule_min_hrtimeout(void)
+{
+ return schedule_timeout(1);
+}
+
+static inline signed long schedule_msec_hrtimeout_interruptible(signed long timeout)
+{
+ return schedule_timeout_interruptible(msecs_to_jiffies(timeout));
+}
+
+static inline signed long schedule_msec_hrtimeout_uninterruptible(signed long timeout)
+{
+ return schedule_timeout_uninterruptible(msecs_to_jiffies(timeout));
+}
+#endif
asmlinkage void schedule(void);
extern void schedule_preempt_disabled(void);
diff --git a/kernel/time/timer.c b/kernel/time/timer.c
index bc53598..0e22641 100644
--- a/kernel/time/timer.c
+++ b/kernel/time/timer.c
@@ -1750,6 +1750,7 @@ signed long __sched schedule_timeout(signed long timeout)
expire = timeout + jiffies;
+#ifdef CONFIG_HIGH_RES_TIMERS
if (timeout == 1 && hrtimer_resolution < NSEC_PER_SEC / HZ) {
/*
* Special case 1 as being a request for the minimum timeout
@@ -1760,6 +1761,7 @@ signed long __sched schedule_timeout(signed long timeout)
return 0;
goto out_timeout;
}
+#endif
setup_timer_on_stack(&timer, process_timeout, (unsigned long)current);
__mod_timer(&timer, expire, false);
--
2.9.3

61
patches/0018-Make-threaded-IRQs-optionally-the-default-which-can-.patch Normal file
View File

@ -0,0 +1,61 @@
From a0d29f014dbcb29649dff1a9b8df58bad7be3926 Mon Sep 17 00:00:00 2001
From: Con Kolivas <kernel@kolivas.org>
Date: Wed, 7 Dec 2016 21:13:16 +1100
Subject: [PATCH 18/25] Make threaded IRQs optionally the default which can be
disabled.
---
kernel/irq/Kconfig | 14 ++++++++++++++
kernel/irq/manage.c | 10 ++++++++++
2 files changed, 24 insertions(+)
diff --git a/kernel/irq/Kconfig b/kernel/irq/Kconfig
index 3bbfd6a..351bf16 100644
--- a/kernel/irq/Kconfig
+++ b/kernel/irq/Kconfig
@@ -95,6 +95,20 @@ config IRQ_DOMAIN_DEBUG
config IRQ_FORCED_THREADING
bool
+config FORCE_IRQ_THREADING
+ bool "Make IRQ threading compulsory"
+ depends on IRQ_FORCED_THREADING
+ default y
+ ---help---
+
+ Make IRQ threading mandatory for any IRQ handlers that support it
+ instead of being optional and requiring the threadirqs kernel
+ parameter. Instead they can be optionally disabled with the
+ nothreadirqs kernel parameter.
+
+ Enable if you are building for a desktop or low latency system,
+ otherwise say N.
+
config SPARSE_IRQ
bool "Support sparse irq numbering" if MAY_HAVE_SPARSE_IRQ
---help---
diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c
index 6b66959..6b3fb17 100644
--- a/kernel/irq/manage.c
+++ b/kernel/irq/manage.c
@@ -22,7 +22,17 @@
#include "internals.h"
#ifdef CONFIG_IRQ_FORCED_THREADING
+#ifdef CONFIG_FORCE_IRQ_THREADING
+__read_mostly bool force_irqthreads = true;
+#else
__read_mostly bool force_irqthreads;
+#endif
+static int __init setup_noforced_irqthreads(char *arg)
+{
+ force_irqthreads = false;
+ return 0;
+}
+early_param("nothreadirqs", setup_noforced_irqthreads);
static int __init setup_forced_irqthreads(char *arg)
{
--
2.9.3

69
patches/0020-Don-t-use-hrtimer-overlay-when-pm_freezing-since-som.patch Normal file
View File

@ -0,0 +1,69 @@
From a278cad439033005610ddda23882f2c681c669d1 Mon Sep 17 00:00:00 2001
From: Con Kolivas <kernel@kolivas.org>
Date: Mon, 20 Feb 2017 13:32:58 +1100
Subject: [PATCH 20/25] Don't use hrtimer overlay when pm_freezing since some
drivers still don't correctly use freezable timeouts.
---
kernel/time/hrtimer.c | 2 +-
kernel/time/timer.c | 9 +++++----
2 files changed, 6 insertions(+), 5 deletions(-)
diff --git a/kernel/time/hrtimer.c b/kernel/time/hrtimer.c
index 26ac1f8..25e1555 100644
--- a/kernel/time/hrtimer.c
+++ b/kernel/time/hrtimer.c
@@ -1818,7 +1818,7 @@ signed long __sched schedule_msec_hrtimeout(signed long timeout)
* (yet) better than Hz, as would occur during startup, use regular
* timers.
*/
- if (jiffs > 4 || hrtimer_resolution >= NSEC_PER_SEC / HZ)
+ if (jiffs > 4 || hrtimer_resolution >= NSEC_PER_SEC / HZ || pm_freezing)
return schedule_timeout(jiffs);
delta = (timeout % 1000) * NSEC_PER_MSEC;
diff --git a/kernel/time/timer.c b/kernel/time/timer.c
index 0e22641..45a6e1f 100644
--- a/kernel/time/timer.c
+++ b/kernel/time/timer.c
@@ -42,6 +42,7 @@
#include <linux/sched/sysctl.h>
#include <linux/slab.h>
#include <linux/compat.h>
+#include <linux/freezer.h>
#include <linux/uaccess.h>
#include <asm/unistd.h>
@@ -1901,12 +1902,12 @@ void msleep(unsigned int msecs)
* Use high resolution timers where the resolution of tick based
* timers is inadequate.
*/
- if (jiffs < 5 && hrtimer_resolution < NSEC_PER_SEC / HZ) {
+ if (jiffs < 5 && hrtimer_resolution < NSEC_PER_SEC / HZ && !pm_freezing) {
while (msecs)
msecs = schedule_msec_hrtimeout_uninterruptible(msecs);
return;
}
- timeout = msecs_to_jiffies(msecs) + 1;
+ timeout = jiffs + 1;
while (timeout)
timeout = schedule_timeout_uninterruptible(timeout);
@@ -1923,12 +1924,12 @@ unsigned long msleep_interruptible(unsigned int msecs)
int jiffs = msecs_to_jiffies(msecs);
unsigned long timeout;
- if (jiffs < 5 && hrtimer_resolution < NSEC_PER_SEC / HZ) {
+ if (jiffs < 5 && hrtimer_resolution < NSEC_PER_SEC / HZ && !pm_freezing) {
while (msecs && !signal_pending(current))
msecs = schedule_msec_hrtimeout_interruptible(msecs);
return msecs;
}
- timeout = msecs_to_jiffies(msecs) + 1;
+ timeout = jiffs + 1;
while (timeout && !signal_pending(current))
timeout = schedule_timeout_interruptible(timeout);
--
2.9.3

34
patches/0021-Make-writeback-throttling-default-enabled.patch Normal file
View File

@ -0,0 +1,34 @@
From da915e0f3abeb61f6a132bb77b7d0a9bf0573233 Mon Sep 17 00:00:00 2001
From: Con Kolivas <kernel@kolivas.org>
Date: Mon, 20 Feb 2017 13:38:23 +1100
Subject: [PATCH 21/25] Make writeback throttling default enabled.
---
block/Kconfig | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/block/Kconfig b/block/Kconfig
index 8bf114a..83e6f9d 100644
--- a/block/Kconfig
+++ b/block/Kconfig
@@ -123,7 +123,7 @@ config BLK_CMDLINE_PARSER
config BLK_WBT
bool "Enable support for block device writeback throttling"
- default n
+ default y
---help---
Enabling this option enables the block layer to throttle buffered
background writeback from the VM, making it more smooth and having
@@ -133,7 +133,7 @@ config BLK_WBT
config BLK_WBT_SQ
bool "Single queue writeback throttling"
- default n
+ default y
depends on BLK_WBT
---help---
Enable writeback throttling by default on legacy single queue devices
--
2.9.3

43
patches/0022-Swap-sucks.patch Normal file
View File

@ -0,0 +1,43 @@
From 2f96168f72bbd431c0e6d28b44393e98b49ca787 Mon Sep 17 00:00:00 2001
From: Con Kolivas <kernel@kolivas.org>
Date: Mon, 20 Feb 2017 13:48:54 +1100
Subject: [PATCH 22/25] Swap sucks.
---
include/linux/swap.h | 6 +-----
mm/vmscan.c | 2 +-
2 files changed, 2 insertions(+), 6 deletions(-)
diff --git a/include/linux/swap.h b/include/linux/swap.h
index 7f47b70..1c2ed28 100644
--- a/include/linux/swap.h
+++ b/include/linux/swap.h
@@ -361,11 +361,7 @@ extern struct page *swapin_readahead(swp_entry_t, gfp_t,
extern atomic_long_t nr_swap_pages;
extern long total_swap_pages;
-/* Swap 50% full? Release swapcache more aggressively.. */
-static inline bool vm_swap_full(void)
-{
- return atomic_long_read(&nr_swap_pages) * 2 < total_swap_pages;
-}
+#define vm_swap_full() 1
static inline long get_nr_swap_pages(void)
{
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 532a2a7..15e4260 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -141,7 +141,7 @@ struct scan_control {
/*
* From 0 .. 100. Higher means more swappy.
*/
-int vm_swappiness = 60;
+int vm_swappiness = 33;
/*
* The total number of pages which are beyond the high watermark within all
* zones.
--
2.9.3

533
patches/enable_additional_cpu_optimizations_for_gcc_v4.9+_kernel_v3.15+.patch Normal file
View File

@ -0,0 +1,533 @@
WARNING
This patch works with gcc versions 4.9+ and with kernel version 3.15+ and should
NOT be applied when compiling on older versions of gcc due to key name changes
of the march flags introduced with the version 4.9 release of gcc.[1]
Use the older version of this patch hosted on the same github for older
versions of gcc.
FEATURES
This patch adds additional CPU options to the Linux kernel accessible under:
Processor type and features --->
Processor family --->
The expanded microarchitectures include:
* AMD Improved K8-family
* AMD K10-family
* AMD Family 10h (Barcelona)
* AMD Family 14h (Bobcat)
* AMD Family 16h (Jaguar)
* AMD Family 15h (Bulldozer)
* AMD Family 15h (Piledriver)
* AMD Family 15h (Steamroller)
* AMD Family 15h (Excavator)
* AMD Family 17h (Zen)
* Intel Silvermont low-power processors
* Intel 1st Gen Core i3/i5/i7 (Nehalem)
* Intel 1.5 Gen Core i3/i5/i7 (Westmere)
* Intel 2nd Gen Core i3/i5/i7 (Sandybridge)
* Intel 3rd Gen Core i3/i5/i7 (Ivybridge)
* Intel 4th Gen Core i3/i5/i7 (Haswell)
* Intel 5th Gen Core i3/i5/i7 (Broadwell)
* Intel 6th Gen Core i3/i5.i7 (Skylake)
It also offers to compile passing the 'native' option which, "selects the CPU
to generate code for at compilation time by determining the processor type of
the compiling machine. Using -march=native enables all instruction subsets
supported by the local machine and will produce code optimized for the local
machine under the constraints of the selected instruction set."[3]
MINOR NOTES
This patch also changes 'atom' to 'bonnell' in accordance with the gcc v4.9
changes. Note that upstream is using the deprecated 'match=atom' flags when I
believe it should use the newer 'march=bonnell' flag for atom processors.[2]
It is not recommended to compile on Atom-CPUs with the 'native' option.[4] The
recommendation is use to the 'atom' option instead.
BENEFITS
Small but real speed increases are measurable using a make endpoint comparing
a generic kernel to one built with one of the respective microarchs.
See the following experimental evidence supporting this statement:
https://github.com/graysky2/kernel_gcc_patch
REQUIREMENTS
linux version >=3.15
gcc version >=4.9
ACKNOWLEDGMENTS
This patch builds on the seminal work by Jeroen.[5]
REFERENCES
1. https://gcc.gnu.org/gcc-4.9/changes.html
2. https://bugzilla.kernel.org/show_bug.cgi?id=77461
3. https://gcc.gnu.org/onlinedocs/gcc/x86-Options.html
4. https://github.com/graysky2/kernel_gcc_patch/issues/15
5. http://www.linuxforge.net/docs/linux/linux-gcc.php
--- a/arch/x86/include/asm/module.h 2016-12-11 14:17:54.000000000 -0500
+++ b/arch/x86/include/asm/module.h 2017-01-06 20:44:36.602227264 -0500
@@ -15,6 +15,24 @@
#define MODULE_PROC_FAMILY "586MMX "
#elif defined CONFIG_MCORE2
#define MODULE_PROC_FAMILY "CORE2 "
+#elif defined CONFIG_MNATIVE
+#define MODULE_PROC_FAMILY "NATIVE "
+#elif defined CONFIG_MNEHALEM
+#define MODULE_PROC_FAMILY "NEHALEM "
+#elif defined CONFIG_MWESTMERE
+#define MODULE_PROC_FAMILY "WESTMERE "
+#elif defined CONFIG_MSILVERMONT
+#define MODULE_PROC_FAMILY "SILVERMONT "
+#elif defined CONFIG_MSANDYBRIDGE
+#define MODULE_PROC_FAMILY "SANDYBRIDGE "
+#elif defined CONFIG_MIVYBRIDGE
+#define MODULE_PROC_FAMILY "IVYBRIDGE "
+#elif defined CONFIG_MHASWELL
+#define MODULE_PROC_FAMILY "HASWELL "
+#elif defined CONFIG_MBROADWELL
+#define MODULE_PROC_FAMILY "BROADWELL "
+#elif defined CONFIG_MSKYLAKE
+#define MODULE_PROC_FAMILY "SKYLAKE "
#elif defined CONFIG_MATOM
#define MODULE_PROC_FAMILY "ATOM "
#elif defined CONFIG_M686
@@ -33,6 +51,26 @@
#define MODULE_PROC_FAMILY "K7 "
#elif defined CONFIG_MK8
#define MODULE_PROC_FAMILY "K8 "
+#elif defined CONFIG_MK8SSE3
+#define MODULE_PROC_FAMILY "K8SSE3 "
+#elif defined CONFIG_MK10
+#define MODULE_PROC_FAMILY "K10 "
+#elif defined CONFIG_MBARCELONA
+#define MODULE_PROC_FAMILY "BARCELONA "
+#elif defined CONFIG_MBOBCAT
+#define MODULE_PROC_FAMILY "BOBCAT "
+#elif defined CONFIG_MBULLDOZER
+#define MODULE_PROC_FAMILY "BULLDOZER "
+#elif defined CONFIG_MPILEDRIVER
+#define MODULE_PROC_FAMILY "PILEDRIVER "
+#elif defined CONFIG_MSTEAMROLLER
+#define MODULE_PROC_FAMILY "STEAMROLLER "
+#elif defined CONFIG_MJAGUAR
+#define MODULE_PROC_FAMILY "JAGUAR "
+#elif defined CONFIG_MEXCAVATOR
+#define MODULE_PROC_FAMILY "EXCAVATOR "
+#elif defined CONFIG_MZEN
+#define MODULE_PROC_FAMILY "ZEN "
#elif defined CONFIG_MELAN
#define MODULE_PROC_FAMILY "ELAN "
#elif defined CONFIG_MCRUSOE
--- a/arch/x86/Kconfig.cpu 2016-12-11 14:17:54.000000000 -0500
+++ b/arch/x86/Kconfig.cpu 2017-01-06 20:46:14.004109597 -0500
@@ -115,6 +115,7 @@ config MPENTIUMM
config MPENTIUM4
bool "Pentium-4/Celeron(P4-based)/Pentium-4 M/older Xeon"
depends on X86_32
+ select X86_P6_NOP
---help---
Select this for Intel Pentium 4 chips. This includes the
Pentium 4, Pentium D, P4-based Celeron and Xeon, and
@@ -147,9 +148,8 @@ config MPENTIUM4
-Paxville
-Dempsey
-
config MK6
- bool "K6/K6-II/K6-III"
+ bool "AMD K6/K6-II/K6-III"
depends on X86_32
---help---
Select this for an AMD K6-family processor. Enables use of
@@ -157,7 +157,7 @@ config MK6
flags to GCC.
config MK7
- bool "Athlon/Duron/K7"
+ bool "AMD Athlon/Duron/K7"
depends on X86_32
---help---
Select this for an AMD Athlon K7-family processor. Enables use of
@@ -165,12 +165,83 @@ config MK7
flags to GCC.
config MK8
- bool "Opteron/Athlon64/Hammer/K8"
+ bool "AMD Opteron/Athlon64/Hammer/K8"
---help---
Select this for an AMD Opteron or Athlon64 Hammer-family processor.
Enables use of some extended instructions, and passes appropriate
optimization flags to GCC.
+config MK8SSE3
+ bool "AMD Opteron/Athlon64/Hammer/K8 with SSE3"
+ ---help---
+ Select this for improved AMD Opteron or Athlon64 Hammer-family processors.
+ Enables use of some extended instructions, and passes appropriate
+ optimization flags to GCC.
+
+config MK10
+ bool "AMD 61xx/7x50/PhenomX3/X4/II/K10"
+ ---help---
+ Select this for an AMD 61xx Eight-Core Magny-Cours, Athlon X2 7x50,
+ Phenom X3/X4/II, Athlon II X2/X3/X4, or Turion II-family processor.
+ Enables use of some extended instructions, and passes appropriate
+ optimization flags to GCC.
+
+config MBARCELONA
+ bool "AMD Barcelona"
+ ---help---
+ Select this for AMD Family 10h Barcelona processors.
+
+ Enables -march=barcelona
+
+config MBOBCAT
+ bool "AMD Bobcat"
+ ---help---
+ Select this for AMD Family 14h Bobcat processors.
+
+ Enables -march=btver1
+
+config MJAGUAR
+ bool "AMD Jaguar"
+ ---help---
+ Select this for AMD Family 16h Jaguar processors.
+
+ Enables -march=btver2
+
+config MBULLDOZER
+ bool "AMD Bulldozer"
+ ---help---
+ Select this for AMD Family 15h Bulldozer processors.
+
+ Enables -march=bdver1
+
+config MPILEDRIVER
+ bool "AMD Piledriver"
+ ---help---
+ Select this for AMD Family 15h Piledriver processors.
+
+ Enables -march=bdver2
+
+config MSTEAMROLLER
+ bool "AMD Steamroller"
+ ---help---
+ Select this for AMD Family 15h Steamroller processors.
+
+ Enables -march=bdver3
+
+config MEXCAVATOR
+ bool "AMD Excavator"
+ ---help---
+ Select this for AMD Family 15h Excavator processors.
+
+ Enables -march=bdver4
+
+config MZEN
+ bool "AMD Zen"
+ ---help---
+ Select this for AMD Family 17h Zen processors.
+
+ Enables -march=znver1
+
config MCRUSOE
bool "Crusoe"
depends on X86_32
@@ -252,6 +323,7 @@ config MVIAC7
config MPSC
bool "Intel P4 / older Netburst based Xeon"
+ select X86_P6_NOP
depends on X86_64
---help---
Optimize for Intel Pentium 4, Pentium D and older Nocona/Dempsey
@@ -261,8 +333,19 @@ config MPSC
using the cpu family field
in /proc/cpuinfo. Family 15 is an older Xeon, Family 6 a newer one.
+config MATOM
+ bool "Intel Atom"
+ select X86_P6_NOP
+ ---help---
+
+ Select this for the Intel Atom platform. Intel Atom CPUs have an
+ in-order pipelining architecture and thus can benefit from
+ accordingly optimized code. Use a recent GCC with specific Atom
+ support in order to fully benefit from selecting this option.
+
config MCORE2
- bool "Core 2/newer Xeon"
+ bool "Intel Core 2"
+ select X86_P6_NOP
---help---
Select this for Intel Core 2 and newer Core 2 Xeons (Xeon 51xx and
@@ -270,14 +353,79 @@ config MCORE2
family in /proc/cpuinfo. Newer ones have 6 and older ones 15
(not a typo)
-config MATOM
- bool "Intel Atom"
+ Enables -march=core2
+
+config MNEHALEM
+ bool "Intel Nehalem"
+ select X86_P6_NOP
---help---
- Select this for the Intel Atom platform. Intel Atom CPUs have an
- in-order pipelining architecture and thus can benefit from
- accordingly optimized code. Use a recent GCC with specific Atom
- support in order to fully benefit from selecting this option.
+ Select this for 1st Gen Core processors in the Nehalem family.
+
+ Enables -march=nehalem
+
+config MWESTMERE
+ bool "Intel Westmere"
+ select X86_P6_NOP
+ ---help---
+
+ Select this for the Intel Westmere formerly Nehalem-C family.
+
+ Enables -march=westmere
+
+config MSILVERMONT
+ bool "Intel Silvermont"
+ select X86_P6_NOP
+ ---help---
+
+ Select this for the Intel Silvermont platform.
+
+ Enables -march=silvermont
+
+config MSANDYBRIDGE
+ bool "Intel Sandy Bridge"
+ select X86_P6_NOP
+ ---help---
+
+ Select this for 2nd Gen Core processors in the Sandy Bridge family.
+
+ Enables -march=sandybridge
+
+config MIVYBRIDGE
+ bool "Intel Ivy Bridge"
+ select X86_P6_NOP
+ ---help---
+
+ Select this for 3rd Gen Core processors in the Ivy Bridge family.
+
+ Enables -march=ivybridge
+
+config MHASWELL
+ bool "Intel Haswell"
+ select X86_P6_NOP
+ ---help---
+
+ Select this for 4th Gen Core processors in the Haswell family.
+
+ Enables -march=haswell
+
+config MBROADWELL
+ bool "Intel Broadwell"
+ select X86_P6_NOP
+ ---help---
+
+ Select this for 5th Gen Core processors in the Broadwell family.
+
+ Enables -march=broadwell
+
+config MSKYLAKE
+ bool "Intel Skylake"
+ select X86_P6_NOP
+ ---help---
+
+ Select this for 6th Gen Core processors in the Skylake family.
+
+ Enables -march=skylake
config GENERIC_CPU
bool "Generic-x86-64"
@@ -286,6 +434,19 @@ config GENERIC_CPU
Generic x86-64 CPU.
Run equally well on all x86-64 CPUs.
+config MNATIVE
+ bool "Native optimizations autodetected by GCC"
+ ---help---
+
+ GCC 4.2 and above support -march=native, which automatically detects
+ the optimum settings to use based on your processor. -march=native
+ also detects and applies additional settings beyond -march specific
+ to your CPU, (eg. -msse4). Unless you have a specific reason not to
+ (e.g. distcc cross-compiling), you should probably be using
+ -march=native rather than anything listed below.
+
+ Enables -march=native
+
endchoice
config X86_GENERIC
@@ -310,7 +471,7 @@ config X86_INTERNODE_CACHE_SHIFT
config X86_L1_CACHE_SHIFT
int
default "7" if MPENTIUM4 || MPSC
- default "6" if MK7 || MK8 || MPENTIUMM || MCORE2 || MATOM || MVIAC7 || X86_GENERIC || GENERIC_CPU
+ default "6" if MK7 || MK8 || MK8SSE3 || MK10 || MBARCELONA || MBOBCAT || MBULLDOZER || MPILEDRIVER || MSTEAMROLLER || MEXCAVATOR || MZEN || MJAGUAR || MPENTIUMM || MCORE2 || MNEHALEM || MWESTMERE || MSILVERMONT || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MNATIVE || MATOM || MVIAC7 || X86_GENERIC || GENERIC_CPU
default "4" if MELAN || M486 || MGEODEGX1
default "5" if MWINCHIP3D || MWINCHIPC6 || MCRUSOE || MEFFICEON || MCYRIXIII || MK6 || MPENTIUMIII || MPENTIUMII || M686 || M586MMX || M586TSC || M586 || MVIAC3_2 || MGEODE_LX
@@ -341,45 +502,46 @@ config X86_ALIGNMENT_16
config X86_INTEL_USERCOPY
def_bool y
- depends on MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M586MMX || X86_GENERIC || MK8 || MK7 || MEFFICEON || MCORE2
+ depends on MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M586MMX || X86_GENERIC || MK8 || MK8SSE3 || MK7 || MEFFICEON || MCORE2 || MK10 || MBARCELONA || MNEHALEM || MWESTMERE || MSILVERMONT || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MNATIVE
config X86_USE_PPRO_CHECKSUM
def_bool y
- depends on MWINCHIP3D || MWINCHIPC6 || MCYRIXIII || MK7 || MK6 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MK8 || MVIAC3_2 || MVIAC7 || MEFFICEON || MGEODE_LX || MCORE2 || MATOM
+ depends on MWINCHIP3D || MWINCHIPC6 || MCYRIXIII || MK7 || MK6 || MK10 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MK8 || MK8SSE3 || MVIAC3_2 || MVIAC7 || MEFFICEON || MGEODE_LX || MCORE2 || MNEHALEM || MWESTMERE || MSILVERMONT || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MATOM || MNATIVE
config X86_USE_3DNOW
def_bool y
depends on (MCYRIXIII || MK7 || MGEODE_LX) && !UML
-#
-# P6_NOPs are a relatively minor optimization that require a family >=
-# 6 processor, except that it is broken on certain VIA chips.
-# Furthermore, AMD chips prefer a totally different sequence of NOPs
-# (which work on all CPUs). In addition, it looks like Virtual PC
-# does not understand them.
-#
-# As a result, disallow these if we're not compiling for X86_64 (these
-# NOPs do work on all x86-64 capable chips); the list of processors in
-# the right-hand clause are the cores that benefit from this optimization.
-#
config X86_P6_NOP
- def_bool y
- depends on X86_64
- depends on (MCORE2 || MPENTIUM4 || MPSC)
+ default n
+ bool "Support for P6_NOPs on Intel chips"
+ depends on (MCORE2 || MPENTIUM4 || MPSC || MATOM || MNEHALEM || MWESTMERE || MSILVERMONT || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MNATIVE)
+ ---help---
+ P6_NOPs are a relatively minor optimization that require a family >=
+ 6 processor, except that it is broken on certain VIA chips.
+ Furthermore, AMD chips prefer a totally different sequence of NOPs
+ (which work on all CPUs). In addition, it looks like Virtual PC
+ does not understand them.
+
+ As a result, disallow these if we're not compiling for X86_64 (these
+ NOPs do work on all x86-64 capable chips); the list of processors in
+ the right-hand clause are the cores that benefit from this optimization.
+
+ Say Y if you have Intel CPU newer than Pentium Pro, N otherwise.
config X86_TSC
def_bool y
- depends on (MWINCHIP3D || MCRUSOE || MEFFICEON || MCYRIXIII || MK7 || MK6 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || M586MMX || M586TSC || MK8 || MVIAC3_2 || MVIAC7 || MGEODEGX1 || MGEODE_LX || MCORE2 || MATOM) || X86_64
+ depends on (MWINCHIP3D || MCRUSOE || MEFFICEON || MCYRIXIII || MK7 || MK6 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || M586MMX || M586TSC || MK8 || MK8SSE3 || MVIAC3_2 || MVIAC7 || MGEODEGX1 || MGEODE_LX || MCORE2 || MNEHALEM || MWESTMERE || MSILVERMONT || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MNATIVE || MATOM) || X86_64
config X86_CMPXCHG64
def_bool y
- depends on X86_PAE || X86_64 || MCORE2 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MATOM
+ depends on X86_PAE || X86_64 || MCORE2 || MNEHALEM || MWESTMERE || MSILVERMONT || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MATOM || MNATIVE
# this should be set for all -march=.. options where the compiler
# generates cmov.
config X86_CMOV
def_bool y
- depends on (MK8 || MK7 || MCORE2 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MVIAC3_2 || MVIAC7 || MCRUSOE || MEFFICEON || X86_64 || MATOM || MGEODE_LX)
+ depends on (MK8 || MK8SSE3 || MK10 || MBARCELONA || MBOBCAT || MBULLDOZER || MPILEDRIVER || MSTEAMROLLER || MEXCAVATOR || MZEN || MJAGUAR || MK7 || MCORE2 || MNEHALEM || MWESTMERE || MSILVERMONT || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MVIAC3_2 || MVIAC7 || MCRUSOE || MEFFICEON || X86_64 || MNATIVE || MATOM || MGEODE_LX)
config X86_MINIMUM_CPU_FAMILY
int
--- a/arch/x86/Makefile 2016-12-11 14:17:54.000000000 -0500
+++ b/arch/x86/Makefile 2017-01-06 20:44:36.603227283 -0500
@@ -104,13 +104,40 @@ else
KBUILD_CFLAGS += $(call cc-option,-mskip-rax-setup)
# FIXME - should be integrated in Makefile.cpu (Makefile_32.cpu)
+ cflags-$(CONFIG_MNATIVE) += $(call cc-option,-march=native)
cflags-$(CONFIG_MK8) += $(call cc-option,-march=k8)
+ cflags-$(CONFIG_MK8SSE3) += $(call cc-option,-march=k8-sse3,-mtune=k8)
+ cflags-$(CONFIG_MK10) += $(call cc-option,-march=amdfam10)
+ cflags-$(CONFIG_MBARCELONA) += $(call cc-option,-march=barcelona)
+ cflags-$(CONFIG_MBOBCAT) += $(call cc-option,-march=btver1)
+ cflags-$(CONFIG_MJAGUAR) += $(call cc-option,-march=btver2)
+ cflags-$(CONFIG_MBULLDOZER) += $(call cc-option,-march=bdver1)
+ cflags-$(CONFIG_MPILEDRIVER) += $(call cc-option,-march=bdver2)
+ cflags-$(CONFIG_MSTEAMROLLER) += $(call cc-option,-march=bdver3)
+ cflags-$(CONFIG_MEXCAVATOR) += $(call cc-option,-march=bdver4)
+ cflags-$(CONFIG_MZEN) += $(call cc-option,-march=znver1)
cflags-$(CONFIG_MPSC) += $(call cc-option,-march=nocona)
cflags-$(CONFIG_MCORE2) += \
- $(call cc-option,-march=core2,$(call cc-option,-mtune=generic))
- cflags-$(CONFIG_MATOM) += $(call cc-option,-march=atom) \
- $(call cc-option,-mtune=atom,$(call cc-option,-mtune=generic))
+ $(call cc-option,-march=core2,$(call cc-option,-mtune=core2))
+ cflags-$(CONFIG_MNEHALEM) += \
+ $(call cc-option,-march=nehalem,$(call cc-option,-mtune=nehalem))
+ cflags-$(CONFIG_MWESTMERE) += \
+ $(call cc-option,-march=westmere,$(call cc-option,-mtune=westmere))
+ cflags-$(CONFIG_MSILVERMONT) += \
+ $(call cc-option,-march=silvermont,$(call cc-option,-mtune=silvermont))
+ cflags-$(CONFIG_MSANDYBRIDGE) += \
+ $(call cc-option,-march=sandybridge,$(call cc-option,-mtune=sandybridge))
+ cflags-$(CONFIG_MIVYBRIDGE) += \
+ $(call cc-option,-march=ivybridge,$(call cc-option,-mtune=ivybridge))
+ cflags-$(CONFIG_MHASWELL) += \
+ $(call cc-option,-march=haswell,$(call cc-option,-mtune=haswell))
+ cflags-$(CONFIG_MBROADWELL) += \
+ $(call cc-option,-march=broadwell,$(call cc-option,-mtune=broadwell))
+ cflags-$(CONFIG_MSKYLAKE) += \
+ $(call cc-option,-march=skylake,$(call cc-option,-mtune=skylake))
+ cflags-$(CONFIG_MATOM) += $(call cc-option,-march=bonnell) \
+ $(call cc-option,-mtune=bonnell,$(call cc-option,-mtune=generic))
cflags-$(CONFIG_GENERIC_CPU) += $(call cc-option,-mtune=generic)
KBUILD_CFLAGS += $(cflags-y)
--- a/arch/x86/Makefile_32.cpu 2016-12-11 14:17:54.000000000 -0500
+++ b/arch/x86/Makefile_32.cpu 2017-01-06 20:44:36.603227283 -0500
@@ -23,7 +23,18 @@ cflags-$(CONFIG_MK6) += -march=k6
# Please note, that patches that add -march=athlon-xp and friends are pointless.
# They make zero difference whatsosever to performance at this time.
cflags-$(CONFIG_MK7) += -march=athlon
+cflags-$(CONFIG_MNATIVE) += $(call cc-option,-march=native)
cflags-$(CONFIG_MK8) += $(call cc-option,-march=k8,-march=athlon)
+cflags-$(CONFIG_MK8SSE3) += $(call cc-option,-march=k8-sse3,-march=athlon)
+cflags-$(CONFIG_MK10) += $(call cc-option,-march=amdfam10,-march=athlon)
+cflags-$(CONFIG_MBARCELONA) += $(call cc-option,-march=barcelona,-march=athlon)
+cflags-$(CONFIG_MBOBCAT) += $(call cc-option,-march=btver1,-march=athlon)
+cflags-$(CONFIG_MJAGUAR) += $(call cc-option,-march=btver2,-march=athlon)
+cflags-$(CONFIG_MBULLDOZER) += $(call cc-option,-march=bdver1,-march=athlon)
+cflags-$(CONFIG_MPILEDRIVER) += $(call cc-option,-march=bdver2,-march=athlon)
+cflags-$(CONFIG_MSTEAMROLLER) += $(call cc-option,-march=bdver3,-march=athlon)
+cflags-$(CONFIG_MEXCAVATOR) += $(call cc-option,-march=bdver4,-march=athlon)
+cflags-$(CONFIG_MZEN) += $(call cc-option,-march=znver1,-march=athlon)
cflags-$(CONFIG_MCRUSOE) += -march=i686 $(align)-functions=0 $(align)-jumps=0 $(align)-loops=0
cflags-$(CONFIG_MEFFICEON) += -march=i686 $(call tune,pentium3) $(align)-functions=0 $(align)-jumps=0 $(align)-loops=0
cflags-$(CONFIG_MWINCHIPC6) += $(call cc-option,-march=winchip-c6,-march=i586)
@@ -32,8 +43,16 @@ cflags-$(CONFIG_MCYRIXIII) += $(call cc-
cflags-$(CONFIG_MVIAC3_2) += $(call cc-option,-march=c3-2,-march=i686)
cflags-$(CONFIG_MVIAC7) += -march=i686
cflags-$(CONFIG_MCORE2) += -march=i686 $(call tune,core2)
-cflags-$(CONFIG_MATOM) += $(call cc-option,-march=atom,$(call cc-option,-march=core2,-march=i686)) \
- $(call cc-option,-mtune=atom,$(call cc-option,-mtune=generic))
+cflags-$(CONFIG_MNEHALEM) += -march=i686 $(call tune,nehalem)
+cflags-$(CONFIG_MWESTMERE) += -march=i686 $(call tune,westmere)
+cflags-$(CONFIG_MSILVERMONT) += -march=i686 $(call tune,silvermont)
+cflags-$(CONFIG_MSANDYBRIDGE) += -march=i686 $(call tune,sandybridge)
+cflags-$(CONFIG_MIVYBRIDGE) += -march=i686 $(call tune,ivybridge)
+cflags-$(CONFIG_MHASWELL) += -march=i686 $(call tune,haswell)
+cflags-$(CONFIG_MBROADWELL) += -march=i686 $(call tune,broadwell)
+cflags-$(CONFIG_MSKYLAKE) += -march=i686 $(call tune,skylake)
+cflags-$(CONFIG_MATOM) += $(call cc-option,-march=bonnell,$(call cc-option,-march=core2,-march=i686)) \
+ $(call cc-option,-mtune=bonnell,$(call cc-option,-mtune=generic))
# AMD Elan support
cflags-$(CONFIG_MELAN) += -march=i486

6933
patches/uksm-0.1.2.6-for-v4.10.patch Normal file
View File

File diff suppressed because it is too large Load Diff

148
patches_disabled/05-intel-pstate-backport.patch Normal file
View File

@ -0,0 +1,148 @@
--- linux-4.8/drivers/cpufreq/intel_pstate.c.orig 2016-10-02 19:24:33.000000000 -0400
+++ linux-4.8/drivers/cpufreq/intel_pstate.c 2016-10-09 19:32:01.073141319 -0400
@@ -181,6 +181,8 @@
* @cpu: CPU number for this instance data
* @update_util: CPUFreq utility callback information
* @update_util_set: CPUFreq utility callback is set
+ * @iowait_boost: iowait-related boost fraction
+ * @last_update: Time of the last update.
* @pstate: Stores P state limits for this CPU
* @vid: Stores VID limits for this CPU
* @pid: Stores PID parameters for this CPU
@@ -206,6 +208,7 @@
struct vid_data vid;
struct _pid pid;
+ u64 last_update;
u64 last_sample_time;
u64 prev_aperf;
u64 prev_mperf;
@@ -216,6 +219,7 @@
struct acpi_processor_performance acpi_perf_data;
bool valid_pss_table;
#endif
+ unsigned int iowait_boost;
};
static struct cpudata **all_cpu_data;
@@ -229,6 +233,7 @@
* @p_gain_pct: PID proportional gain
* @i_gain_pct: PID integral gain
* @d_gain_pct: PID derivative gain
+ * @boost_iowait: Whether or not to use iowait boosting.
*
* Stores per CPU model static PID configuration data.
*/
@@ -240,6 +245,7 @@
int p_gain_pct;
int d_gain_pct;
int i_gain_pct;
+ bool boost_iowait;
};
/**
@@ -1029,7 +1035,7 @@
},
};
-static struct cpu_defaults silvermont_params = {
+static const struct cpu_defaults silvermont_params = {
.pid_policy = {
.sample_rate_ms = 10,
.deadband = 0,
@@ -1037,6 +1043,7 @@
.p_gain_pct = 14,
.d_gain_pct = 0,
.i_gain_pct = 4,
+ .boost_iowait = true,
},
.funcs = {
.get_max = atom_get_max_pstate,
@@ -1050,7 +1057,7 @@
},
};
-static struct cpu_defaults airmont_params = {
+static const struct cpu_defaults airmont_params = {
.pid_policy = {
.sample_rate_ms = 10,
.deadband = 0,
@@ -1058,6 +1065,7 @@
.p_gain_pct = 14,
.d_gain_pct = 0,
.i_gain_pct = 4,
+ .boost_iowait = true,
},
.funcs = {
.get_max = atom_get_max_pstate,
@@ -1071,7 +1079,7 @@
},
};
-static struct cpu_defaults knl_params = {
+static const struct cpu_defaults knl_params = {
.pid_policy = {
.sample_rate_ms = 10,
.deadband = 0,
@@ -1091,7 +1099,7 @@
},
};
-static struct cpu_defaults bxt_params = {
+static const struct cpu_defaults bxt_params = {
.pid_policy = {
.sample_rate_ms = 10,
.deadband = 0,
@@ -1099,6 +1107,7 @@
.p_gain_pct = 14,
.d_gain_pct = 0,
.i_gain_pct = 4,
+ .boost_iowait = true,
},
.funcs = {
.get_max = core_get_max_pstate,
@@ -1222,36 +1231,18 @@
static inline int32_t get_target_pstate_use_cpu_load(struct cpudata *cpu)
{
struct sample *sample = &cpu->sample;
- u64 cummulative_iowait, delta_iowait_us;
- u64 delta_iowait_mperf;
- u64 mperf, now;
- int32_t cpu_load;
+ int32_t busy_frac, boost;
- cummulative_iowait = get_cpu_iowait_time_us(cpu->cpu, &now);
+ busy_frac = div_fp(sample->mperf, sample->tsc);
- /*
- * Convert iowait time into number of IO cycles spent at max_freq.
- * IO is considered as busy only for the cpu_load algorithm. For
- * performance this is not needed since we always try to reach the
- * maximum P-State, so we are already boosting the IOs.
- */
- delta_iowait_us = cummulative_iowait - cpu->prev_cummulative_iowait;
- delta_iowait_mperf = div64_u64(delta_iowait_us * cpu->pstate.scaling *
- cpu->pstate.max_pstate, MSEC_PER_SEC);
-
- mperf = cpu->sample.mperf + delta_iowait_mperf;
- cpu->prev_cummulative_iowait = cummulative_iowait;
+ boost = cpu->iowait_boost;
+ cpu->iowait_boost >>= 1;
- /*
- * The load can be estimated as the ratio of the mperf counter
- * running at a constant frequency during active periods
- * (C0) and the time stamp counter running at the same frequency
- * also during C-states.
- */
- cpu_load = div64_u64(int_tofp(100) * mperf, sample->tsc);
- cpu->sample.busy_scaled = cpu_load;
+ if (busy_frac < boost)
+ busy_frac = boost;
- return get_avg_pstate(cpu) - pid_calc(&cpu->pid, cpu_load);
+ sample->busy_scaled = busy_frac * 100;
+ return get_avg_pstate(cpu) - pid_calc(&cpu->pid, sample->busy_scaled);
}
static inline int32_t get_target_pstate_use_performance(struct cpudata *cpu)

94
patches_disabled/06-dirty_cow.patch Normal file
View File

@ -0,0 +1,94 @@
From 19be0eaffa3ac7d8eb6784ad9bdbc7d67ed8e619 Mon Sep 17 00:00:00 2001
From: Linus Torvalds <torvalds@linux-foundation.org>
Date: Thu, 13 Oct 2016 13:07:36 -0700
Subject: mm: remove gup_flags FOLL_WRITE games from __get_user_pages()
This is an ancient bug that was actually attempted to be fixed once
(badly) by me eleven years ago in commit 4ceb5db9757a ("Fix
get_user_pages() race for write access") but that was then undone due to
problems on s390 by commit f33ea7f404e5 ("fix get_user_pages bug").
In the meantime, the s390 situation has long been fixed, and we can now
fix it by checking the pte_dirty() bit properly (and do it better). The
s390 dirty bit was implemented in abf09bed3cce ("s390/mm: implement
software dirty bits") which made it into v3.9. Earlier kernels will
have to look at the page state itself.
Also, the VM has become more scalable, and what used a purely
theoretical race back then has become easier to trigger.
To fix it, we introduce a new internal FOLL_COW flag to mark the "yes,
we already did a COW" rather than play racy games with FOLL_WRITE that
is very fundamental, and then use the pte dirty flag to validate that
the FOLL_COW flag is still valid.
Reported-and-tested-by: Phil "not Paul" Oester <kernel@linuxace.com>
Acked-by: Hugh Dickins <hughd@google.com>
Reviewed-by: Michal Hocko <mhocko@suse.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Willy Tarreau <w@1wt.eu>
Cc: Nick Piggin <npiggin@gmail.com>
Cc: Greg Thelen <gthelen@google.com>
Cc: stable@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
---
include/linux/mm.h | 1 +
mm/gup.c | 14 ++++++++++++--
2 files changed, 13 insertions(+), 2 deletions(-)
diff --git a/include/linux/mm.h b/include/linux/mm.h
index e9caec6..ed85879 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -2232,6 +2232,7 @@ static inline struct page *follow_page(struct vm_area_struct *vma,
#define FOLL_TRIED 0x800 /* a retry, previous pass started an IO */
#define FOLL_MLOCK 0x1000 /* lock present pages */
#define FOLL_REMOTE 0x2000 /* we are working on non-current tsk/mm */
+#define FOLL_COW 0x4000 /* internal GUP flag */
typedef int (*pte_fn_t)(pte_t *pte, pgtable_t token, unsigned long addr,
void *data);
diff --git a/mm/gup.c b/mm/gup.c
index 96b2b2f..22cc22e 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -60,6 +60,16 @@ static int follow_pfn_pte(struct vm_area_struct *vma, unsigned long address,
return -EEXIST;
}
+/*
+ * FOLL_FORCE can write to even unwritable pte's, but only
+ * after we've gone through a COW cycle and they are dirty.
+ */
+static inline bool can_follow_write_pte(pte_t pte, unsigned int flags)
+{
+ return pte_write(pte) ||
+ ((flags & FOLL_FORCE) && (flags & FOLL_COW) && pte_dirty(pte));
+}
+
static struct page *follow_page_pte(struct vm_area_struct *vma,
unsigned long address, pmd_t *pmd, unsigned int flags)
{
@@ -95,7 +105,7 @@ retry:
}
if ((flags & FOLL_NUMA) && pte_protnone(pte))
goto no_page;
- if ((flags & FOLL_WRITE) && !pte_write(pte)) {
+ if ((flags & FOLL_WRITE) && !can_follow_write_pte(pte, flags)) {
pte_unmap_unlock(ptep, ptl);
return NULL;
}
@@ -412,7 +422,7 @@ static int faultin_page(struct task_struct *tsk, struct vm_area_struct *vma,
* reCOWed by userspace write).
*/
if ((ret & VM_FAULT_WRITE) && !(vma->vm_flags & VM_WRITE))
- *flags &= ~FOLL_WRITE;
+ *flags |= FOLL_COW;
return 0;
}
--
cgit v0.12

8876
patches_disabled/3.12-sched-bfs-444.patch Normal file
View File

File diff suppressed because it is too large Load Diff

9153
patches_disabled/4.4-sched-bfs-467.patch Normal file
View File

File diff suppressed because it is too large Load Diff

708
patches_disabled/BLD-4.6.patch Normal file
View File

@ -0,0 +1,708 @@
diff --git a/init/Kconfig b/init/Kconfig
index 0dfd09d..8d704e5 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -36,6 +36,15 @@ config BROKEN_ON_SMP
depends on BROKEN || !SMP
default y
+config BLD
+ bool "An alternate CPU load distribution technique for task scheduler"
+ depends on SMP
+ default y
+ help
+ This is an alternate CPU load distribution technique based for task
+ scheduler based on The Barbershop Load Distribution algorithm. Not
+ suitable for NUMA, should work well on SMP.
+
config INIT_ENV_ARG_LIMIT
int
default 32 if !UML
diff --git a/kernel/sched/bld.h b/kernel/sched/bld.h
new file mode 100644
index 0000000..f1f9fba
--- /dev/null
+++ b/kernel/sched/bld.h
@@ -0,0 +1,215 @@
+#ifdef CONFIG_BLD
+
+static DEFINE_RWLOCK(rt_list_lock);
+static LIST_HEAD(rt_rq_head);
+static LIST_HEAD(cfs_rq_head);
+static DEFINE_RWLOCK(cfs_list_lock);
+
+#ifdef CONFIG_FAIR_GROUP_SCHED
+static inline struct rq *rq_of_cfs(struct cfs_rq *cfs_rq)
+{
+ return cfs_rq->rq;
+}
+#else
+static inline struct rq *rq_of_cfs(struct cfs_rq *cfs_rq)
+{
+ return container_of(cfs_rq, struct rq, cfs);
+}
+#endif
+
+#ifdef CONFIG_RT_GROUP_SCHED
+static inline struct rq *rq_of_rt(struct rt_rq *rt_rq)
+{
+ return rt_rq->rq;
+}
+#else
+static inline struct rq *rq_of_rt(struct rt_rq *rt_rq)
+{
+ return container_of(rt_rq, struct rq, rt);
+}
+#endif
+
+static int select_cpu_for_wakeup(int task_type, struct cpumask *mask)
+{
+ int cpu = smp_processor_id(), i;
+ unsigned long load, varload;
+ struct rq *rq;
+
+ if (task_type) {
+ varload = ULONG_MAX;
+ for_each_cpu(i, mask) {
+ rq = cpu_rq(i);
+ load = rq->cfs.load.weight;
+ if (load < varload) {
+ varload = load;
+ cpu = i;
+ }
+ }
+ } else {
+ /* Here's an attempt to get a CPU within the mask where
+ * we can preempt easily. To achieve this we tried to
+ * maintain a lowbit, which indicate the lowest bit set on
+ * array bitmap. Since all CPUs contains high priority
+ * kernel threads therefore we eliminate 0, so it might not
+ * be right every time, but it's just an indicator.
+ */
+ varload = 1;
+
+ for_each_cpu(i, mask) {
+ rq = cpu_rq(i);
+ load = rq->rt.lowbit;
+ if (load >= varload) {
+ varload = load;
+ cpu = i;
+ }
+ }
+ }
+
+ return cpu;
+}
+
+static int bld_pick_cpu_cfs(struct task_struct *p, int sd_flags, int wake_flags)
+{
+ struct cfs_rq *cfs;
+ unsigned long flags;
+ unsigned int cpu = smp_processor_id();
+
+ read_lock_irqsave(&cfs_list_lock, flags);
+ list_for_each_entry(cfs, &cfs_rq_head, bld_cfs_list) {
+ cpu = cpu_of(rq_of_cfs(cfs));
+ if (cpu_online(cpu))
+ break;
+ }
+ read_unlock_irqrestore(&cfs_list_lock, flags);
+ return cpu;
+}
+
+static int bld_pick_cpu_rt(struct task_struct *p, int sd_flags, int wake_flags)
+{
+ struct rt_rq *rt;
+ unsigned long flags;
+ unsigned int cpu = smp_processor_id();
+
+ read_lock_irqsave(&rt_list_lock, flags);
+ list_for_each_entry(rt, &rt_rq_head, bld_rt_list) {
+ cpu = cpu_of(rq_of_rt(rt));
+ if (cpu_online(cpu))
+ break;
+ }
+ read_unlock_irqrestore(&rt_list_lock, flags);
+ return cpu;
+}
+
+static int bld_pick_cpu_domain(struct task_struct *p, int sd_flags, int wake_flags)
+{
+ unsigned int cpu = smp_processor_id(), want_affine = 0;
+ struct cpumask *tmpmask;
+
+ if (p->nr_cpus_allowed == 1)
+ return task_cpu(p);
+
+ if (sd_flags & SD_BALANCE_WAKE) {
+ if (cpumask_test_cpu(cpu, tsk_cpus_allowed(p))) {
+ want_affine = 1;
+ }
+ }
+
+ if (want_affine)
+ tmpmask = tsk_cpus_allowed(p);
+ else
+ tmpmask = sched_domain_span(cpu_rq(task_cpu(p))->sd);
+
+ if (rt_task(p))
+ cpu = select_cpu_for_wakeup(0, tmpmask);
+ else
+ cpu = select_cpu_for_wakeup(1, tmpmask);
+
+ return cpu;
+}
+
+static void track_load_rt(struct rq *rq, struct task_struct *p)
+{
+ unsigned long flag;
+ int firstbit;
+ struct rt_rq *first;
+ struct rt_prio_array *array = &rq->rt.active;
+
+ first = list_entry(rt_rq_head.next, struct rt_rq, bld_rt_list);
+ firstbit = sched_find_first_bit(array->bitmap);
+
+ /* Maintaining rt.lowbit */
+ if (firstbit > 0 && firstbit <= rq->rt.lowbit)
+ rq->rt.lowbit = firstbit;
+
+ if (rq->rt.lowbit < first->lowbit) {
+ write_lock_irqsave(&rt_list_lock, flag);
+ list_del(&rq->rt.bld_rt_list);
+ list_add_tail(&rq->rt.bld_rt_list, &rt_rq_head);
+ write_unlock_irqrestore(&rt_list_lock, flag);
+ }
+}
+
+static int bld_get_cpu(struct task_struct *p, int sd_flags, int wake_flags)
+{
+ unsigned int cpu;
+
+ if (sd_flags == SD_BALANCE_WAKE || (sd_flags == SD_BALANCE_EXEC && (get_nr_threads(p) > 1)))
+ cpu = bld_pick_cpu_domain(p, sd_flags, wake_flags);
+ else {
+ if (rt_task(p))
+ cpu = bld_pick_cpu_rt(p, sd_flags, wake_flags);
+ else
+ cpu = bld_pick_cpu_cfs(p, sd_flags, wake_flags);
+ }
+
+ return cpu;
+}
+
+static void bld_track_load_activate(struct rq *rq, struct task_struct *p)
+{
+ unsigned long flag;
+ if (rt_task(p)) {
+ track_load_rt(rq, p);
+ } else {
+ if (rq->cfs.pos != 2) {
+ struct cfs_rq *last;
+ last = list_entry(cfs_rq_head.prev, struct cfs_rq, bld_cfs_list);
+ if (rq->cfs.load.weight >= last->load.weight) {
+ write_lock_irqsave(&cfs_list_lock, flag);
+ list_del(&rq->cfs.bld_cfs_list);
+ list_add_tail(&rq->cfs.bld_cfs_list, &cfs_rq_head);
+ rq->cfs.pos = 2; last->pos = 1;
+ write_unlock_irqrestore(&cfs_list_lock, flag);
+ }
+ }
+ }
+}
+
+static void bld_track_load_deactivate(struct rq *rq, struct task_struct *p)
+{
+ unsigned long flag;
+ if (rt_task(p)) {
+ track_load_rt(rq, p);
+ } else {
+ if (rq->cfs.pos != 0) {
+ struct cfs_rq *first;
+ first = list_entry(cfs_rq_head.next, struct cfs_rq, bld_cfs_list);
+ if (rq->cfs.load.weight <= first->load.weight) {
+ write_lock_irqsave(&cfs_list_lock, flag);
+ list_del(&rq->cfs.bld_cfs_list);
+ list_add(&rq->cfs.bld_cfs_list, &cfs_rq_head);
+ rq->cfs.pos = 0; first->pos = 1;
+ write_unlock_irqrestore(&cfs_list_lock, flag);
+ }
+ }
+ }
+}
+#else
+static inline void bld_track_load_activate(struct rq *rq, struct task_struct *p)
+{
+}
+
+static inline void bld_track_load_deactivate(struct rq *rq, struct task_struct *p)
+{
+}
+#endif /* CONFIG_BLD */
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index d1f7149..c3236de 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -24,6 +24,8 @@
* 2007-07-01 Group scheduling enhancements by Srivatsa Vaddagiri
* 2007-11-29 RT balancing improvements by Steven Rostedt, Gregory Haskins,
* Thomas Gleixner, Mike Kravetz
+ * 2012-Feb The Barbershop Load Distribution (BLD) algorithm - an alternate
+ * CPU load distribution technique for kernel scheduler by Rakib Mullick.
*/
#include <linux/kasan.h>
@@ -86,6 +88,7 @@
#include "sched.h"
#include "../workqueue_internal.h"
#include "../smpboot.h"
+#include "bld.h"
#define CREATE_TRACE_POINTS
#include <trace/events/sched.h>
@@ -713,6 +716,8 @@ static inline void enqueue_task(struct rq *rq, struct task_struct *p, int flags)
if (!(flags & ENQUEUE_RESTORE))
sched_info_queued(rq, p);
p->sched_class->enqueue_task(rq, p, flags);
+ if (!dl_task(p))
+ bld_track_load_activate(rq, p);
}
static inline void dequeue_task(struct rq *rq, struct task_struct *p, int flags)
@@ -721,6 +726,8 @@ static inline void dequeue_task(struct rq *rq, struct task_struct *p, int flags)
if (!(flags & DEQUEUE_SAVE))
sched_info_dequeued(rq, p);
p->sched_class->dequeue_task(rq, p, flags);
+ if (!dl_task(p))
+ bld_track_load_deactivate(rq, p);
}
void activate_task(struct rq *rq, struct task_struct *p, int flags)
@@ -1515,8 +1522,16 @@ int select_task_rq(struct task_struct *p, int cpu, int sd_flags, int wake_flags)
{
lockdep_assert_held(&p->pi_lock);
- if (p->nr_cpus_allowed > 1)
+ if (p->nr_cpus_allowed > 1) {
+#ifndef CONFIG_BLD
cpu = p->sched_class->select_task_rq(p, cpu, sd_flags, wake_flags);
+#else
+ if(dl_task(p))
+ cpu = dl_sched_class.select_task_rq(p, cpu, sd_flags, wake_flags);
+ else
+ cpu = bld_get_cpu(p, sd_flags, wake_flags);
+#endif
+ }
/*
* In order not to call set_task_cpu() on a blocking task we need
@@ -1706,7 +1721,11 @@ void scheduler_ipi(void)
*/
preempt_fold_need_resched();
+#ifndef CONFIG_BLD
if (llist_empty(&this_rq()->wake_list) && !got_nohz_idle_kick())
+#else
+ if (llist_empty(&this_rq()->wake_list))
+#endif
return;
/*
@@ -1728,13 +1747,16 @@ void scheduler_ipi(void)
/*
* Check if someone kicked us for doing the nohz idle load balance.
*/
+#ifndef CONFIG_BLD
if (unlikely(got_nohz_idle_kick())) {
this_rq()->idle_balance = 1;
raise_softirq_irqoff(SCHED_SOFTIRQ);
}
+#endif
irq_exit();
}
+#ifndef CONFIG_BLD
static void ttwu_queue_remote(struct task_struct *p, int cpu)
{
struct rq *rq = cpu_rq(cpu);
@@ -1747,6 +1769,13 @@ static void ttwu_queue_remote(struct task_struct *p, int cpu)
}
}
+#endif
+
+bool cpus_share_cache(int this_cpu, int that_cpu)
+{
+ return per_cpu(sd_llc_id, this_cpu) == per_cpu(sd_llc_id, that_cpu);
+}
+
void wake_up_if_idle(int cpu)
{
struct rq *rq = cpu_rq(cpu);
@@ -1770,18 +1799,13 @@ void wake_up_if_idle(int cpu)
out:
rcu_read_unlock();
}
-
-bool cpus_share_cache(int this_cpu, int that_cpu)
-{
- return per_cpu(sd_llc_id, this_cpu) == per_cpu(sd_llc_id, that_cpu);
-}
#endif /* CONFIG_SMP */
static void ttwu_queue(struct task_struct *p, int cpu)
{
struct rq *rq = cpu_rq(cpu);
-#if defined(CONFIG_SMP)
+#if defined(CONFIG_SMP) && !defined(CONFIG_BLD)
if (sched_feat(TTWU_QUEUE) && !cpus_share_cache(smp_processor_id(), cpu)) {
sched_clock_cpu(cpu); /* sync clocks x-cpu */
ttwu_queue_remote(p, cpu);
@@ -2292,7 +2316,7 @@ int sched_fork(unsigned long clone_flags, struct task_struct *p)
* Silence PROVE_RCU.
*/
raw_spin_lock_irqsave(&p->pi_lock, flags);
- set_task_cpu(p, cpu);
+ __set_task_cpu(p, cpu);
raw_spin_unlock_irqrestore(&p->pi_lock, flags);
#ifdef CONFIG_SCHED_INFO
@@ -2837,7 +2861,14 @@ void sched_exec(void)
int dest_cpu;
raw_spin_lock_irqsave(&p->pi_lock, flags);
+#ifndef CONFIG_BLD
dest_cpu = p->sched_class->select_task_rq(p, task_cpu(p), SD_BALANCE_EXEC, 0);
+#else
+ if (dl_task(p))
+ dest_cpu = task_cpu(p);
+ else
+ dest_cpu = bld_get_cpu(p, SD_BALANCE_EXEC, 0);
+#endif
if (dest_cpu == smp_processor_id())
goto unlock;
@@ -2926,8 +2957,10 @@ void scheduler_tick(void)
#ifdef CONFIG_SMP
rq->idle_balance = idle_cpu(cpu);
+#ifndef CONFIG_BLD
trigger_load_balance(rq);
#endif
+#endif
rq_last_tick_reset(rq);
}
@@ -7359,6 +7392,15 @@ void __init sched_init(void)
#endif
init_rq_hrtick(rq);
atomic_set(&rq->nr_iowait, 0);
+#ifdef CONFIG_BLD
+ INIT_LIST_HEAD(&rq->cfs.bld_cfs_list);
+ list_add_tail(&rq->cfs.bld_cfs_list, &cfs_rq_head);
+ rq->cfs.pos = 0;
+
+ INIT_LIST_HEAD(&rq->rt.bld_rt_list);
+ list_add_tail(&rq->rt.bld_rt_list, &rt_rq_head);
+ rq->rt.lowbit = INT_MAX;
+#endif
}
set_load_weight(&init_task);
@@ -7399,6 +7441,9 @@ void __init sched_init(void)
init_sched_fair_class();
scheduler_running = 1;
+#ifdef CONFIG_BLD
+ printk(KERN_INFO "BLD: An Alternate CPU load distributor activated.\n");
+#endif
}
#ifdef CONFIG_DEBUG_ATOMIC_SLEEP
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index e7dd0ec..555572f 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -4746,6 +4746,7 @@ static void task_waking_fair(struct task_struct *p)
record_wakee(p);
}
+#ifndef CONFIG_BLD
#ifdef CONFIG_FAIR_GROUP_SCHED
/*
* effective_load() calculates the load change as seen from the root_task_group
@@ -5248,6 +5249,7 @@ select_task_rq_fair(struct task_struct *p, int prev_cpu, int sd_flag, int wake_f
return new_cpu;
}
+#endif /* CONFIG_BLD */
/*
* Called immediately before a task is migrated to a new cpu; task_cpu(p) and
@@ -5552,6 +5554,7 @@ idle:
* further scheduler activity on it and we're being very careful to
* re-start the picking loop.
*/
+#ifndef CONFIG_BLD
lockdep_unpin_lock(&rq->lock);
new_tasks = idle_balance(rq);
lockdep_pin_lock(&rq->lock);
@@ -5565,7 +5568,7 @@ idle:
if (new_tasks > 0)
goto again;
-
+#endif
return NULL;
}
@@ -6226,8 +6229,9 @@ static unsigned long task_h_load(struct task_struct *p)
}
#endif
-/********** Helpers for find_busiest_group ************************/
+#ifndef CONFIG_BLD
+/********** Helpers for find_busiest_group ************************/
enum group_type {
group_other = 0,
group_imbalanced,
@@ -6318,6 +6322,7 @@ static inline int get_sd_load_idx(struct sched_domain *sd,
return load_idx;
}
+#endif /* CONFIG_BLD */
static unsigned long scale_rt_capacity(int cpu)
{
@@ -6426,6 +6431,7 @@ void update_group_capacity(struct sched_domain *sd, int cpu)
sdg->sgc->capacity = capacity;
}
+#ifndef CONFIG_BLD
/*
* Check whether the capacity of the rq has been noticeably reduced by side
* activity. The imbalance_pct is used for the threshold.
@@ -7659,6 +7665,8 @@ static inline int on_null_domain(struct rq *rq)
return unlikely(!rcu_dereference_sched(rq->sd));
}
+#endif /* CONFIG_BLD */
+
#ifdef CONFIG_NO_HZ_COMMON
/*
* idle load balancing details
@@ -7666,12 +7674,39 @@ static inline int on_null_domain(struct rq *rq)
* needed, they will kick the idle load balancer, which then does idle
* load balancing for all the idle CPUs.
*/
+#ifndef CONFIG_BLD
static struct {
cpumask_var_t idle_cpus_mask;
atomic_t nr_cpus;
unsigned long next_balance; /* in jiffy units */
} nohz ____cacheline_aligned;
+static inline void nohz_balance_exit_idle(int cpu)
+{
+ if (unlikely(test_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu)))) {
+ /*
+ * Completely isolated CPUs don't ever set, so we must test.
+ */
+ if (likely(cpumask_test_cpu(cpu, nohz.idle_cpus_mask))) {
+ cpumask_clear_cpu(cpu, nohz.idle_cpus_mask);
+ atomic_dec(&nohz.nr_cpus);
+ }
+ clear_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu));
+ }
+}
+
+static int sched_ilb_notifier(struct notifier_block *nfb,
+ unsigned long action, void *hcpu)
+{
+ switch (action & ~CPU_TASKS_FROZEN) {
+ case CPU_DYING:
+ nohz_balance_exit_idle(smp_processor_id());
+ return NOTIFY_OK;
+ default:
+ return NOTIFY_DONE;
+ }
+}
+
static inline int find_new_ilb(void)
{
int ilb = cpumask_first(nohz.idle_cpus_mask);
@@ -7709,20 +7744,7 @@ static void nohz_balancer_kick(void)
smp_send_reschedule(ilb_cpu);
return;
}
-
-static inline void nohz_balance_exit_idle(int cpu)
-{
- if (unlikely(test_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu)))) {
- /*
- * Completely isolated CPUs don't ever set, so we must test.
- */
- if (likely(cpumask_test_cpu(cpu, nohz.idle_cpus_mask))) {
- cpumask_clear_cpu(cpu, nohz.idle_cpus_mask);
- atomic_dec(&nohz.nr_cpus);
- }
- clear_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu));
- }
-}
+#endif /* CONFIG_BLD */
static inline void set_cpu_sd_state_busy(void)
{
@@ -7764,6 +7786,7 @@ unlock:
*/
void nohz_balance_enter_idle(int cpu)
{
+#ifndef CONFIG_BLD
/*
* If this cpu is going down, then nothing needs to be done.
*/
@@ -7782,23 +7805,10 @@ void nohz_balance_enter_idle(int cpu)
cpumask_set_cpu(cpu, nohz.idle_cpus_mask);
atomic_inc(&nohz.nr_cpus);
set_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu));
-}
-
-static int sched_ilb_notifier(struct notifier_block *nfb,
- unsigned long action, void *hcpu)
-{
- switch (action & ~CPU_TASKS_FROZEN) {
- case CPU_DYING:
- nohz_balance_exit_idle(smp_processor_id());
- return NOTIFY_OK;
- default:
- return NOTIFY_DONE;
- }
+#endif
}
#endif
-static DEFINE_SPINLOCK(balancing);
-
/*
* Scale the max load_balance interval with the number of CPUs in the system.
* This trades load-balance latency on larger machines for less cross talk.
@@ -7808,6 +7818,9 @@ void update_max_interval(void)
max_load_balance_interval = HZ*num_online_cpus()/10;
}
+#ifndef CONFIG_BLD
+static DEFINE_SPINLOCK(balancing);
+
/*
* It checks each scheduling domain to see if it is due to be balanced,
* and initiates a balancing operation if so.
@@ -8095,6 +8108,7 @@ void trigger_load_balance(struct rq *rq)
nohz_balancer_kick();
#endif
}
+#endif /* CONFIG_BLD */
static void rq_online_fair(struct rq *rq)
{
@@ -8531,7 +8545,9 @@ const struct sched_class fair_sched_class = {
.put_prev_task = put_prev_task_fair,
#ifdef CONFIG_SMP
+#ifndef CONFIG_BLD
.select_task_rq = select_task_rq_fair,
+#endif
.migrate_task_rq = migrate_task_rq_fair,
.rq_online = rq_online_fair,
@@ -8593,6 +8609,7 @@ void show_numa_stats(struct task_struct *p, struct seq_file *m)
__init void init_sched_fair_class(void)
{
+#ifndef CONFIG_BLD
#ifdef CONFIG_SMP
open_softirq(SCHED_SOFTIRQ, run_rebalance_domains);
@@ -8602,5 +8619,5 @@ __init void init_sched_fair_class(void)
cpu_notifier(sched_ilb_notifier, 0);
#endif
#endif /* SMP */
-
+#endif /* BLD */
}
diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
index ec4f538d..4462bed 100644
--- a/kernel/sched/rt.c
+++ b/kernel/sched/rt.c
@@ -1375,6 +1375,7 @@ static void yield_task_rt(struct rq *rq)
#ifdef CONFIG_SMP
static int find_lowest_rq(struct task_struct *task);
+#ifndef CONFIG_BLD
static int
select_task_rq_rt(struct task_struct *p, int cpu, int sd_flag, int flags)
{
@@ -1430,6 +1431,7 @@ select_task_rq_rt(struct task_struct *p, int cpu, int sd_flag, int flags)
out:
return cpu;
}
+#endif /* CONFIG_BLD */
static void check_preempt_equal_prio(struct rq *rq, struct task_struct *p)
{
@@ -2335,7 +2337,9 @@ const struct sched_class rt_sched_class = {
.put_prev_task = put_prev_task_rt,
#ifdef CONFIG_SMP
+#ifndef CONFIG_BLD
.select_task_rq = select_task_rq_rt,
+#endif
.set_cpus_allowed = set_cpus_allowed_common,
.rq_online = rq_online_rt,
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index ec2e8d2..aaab735 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -408,9 +408,8 @@ struct cfs_rq {
#endif /* CONFIG_FAIR_GROUP_SCHED */
#endif /* CONFIG_SMP */
-#ifdef CONFIG_FAIR_GROUP_SCHED
struct rq *rq; /* cpu runqueue to which this cfs_rq is attached */
-
+#ifdef CONFIG_FAIR_GROUP_SCHED
/*
* leaf cfs_rqs are those that hold tasks (lowest schedulable entity in
* a hierarchy). Non-leaf lrqs hold other higher schedulable entities
@@ -434,6 +433,11 @@ struct cfs_rq {
struct list_head throttled_list;
#endif /* CONFIG_CFS_BANDWIDTH */
#endif /* CONFIG_FAIR_GROUP_SCHED */
+
+#ifdef CONFIG_BLD
+ struct list_head bld_cfs_list;
+ char pos;
+#endif
};
static inline int rt_bandwidth_enabled(void)
@@ -479,12 +483,16 @@ struct rt_rq {
/* Nests inside the rq lock: */
raw_spinlock_t rt_runtime_lock;
+ struct rq *rq;
#ifdef CONFIG_RT_GROUP_SCHED
unsigned long rt_nr_boosted;
- struct rq *rq;
struct task_group *tg;
#endif
+#ifdef CONFIG_BLD
+ struct list_head bld_rt_list;
+ int lowbit;
+#endif
};
/* Deadline class' related fields in a runqueue */

695
patches_disabled/BLD-4.8.patch Normal file
View File

@ -0,0 +1,695 @@
diff --git a/init/Kconfig b/init/Kconfig
index cac3f09..4e49d16 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -36,6 +36,15 @@ config BROKEN_ON_SMP
depends on BROKEN || !SMP
default y
+config BLD
+ bool "An alternate CPU load distribution technique for task scheduler"
+ depends on SMP
+ default y
+ help
+ This is an alternate CPU load distribution technique based for task
+ scheduler based on The Barbershop Load Distribution algorithm. Not
+ suitable for NUMA, should work well on SMP.
+
config INIT_ENV_ARG_LIMIT
int
default 32 if !UML
diff --git a/kernel/sched/bld.h b/kernel/sched/bld.h
new file mode 100644
index 0000000..f1f9fba
--- /dev/null
+++ b/kernel/sched/bld.h
@@ -0,0 +1,215 @@
+#ifdef CONFIG_BLD
+
+static DEFINE_RWLOCK(rt_list_lock);
+static LIST_HEAD(rt_rq_head);
+static LIST_HEAD(cfs_rq_head);
+static DEFINE_RWLOCK(cfs_list_lock);
+
+#ifdef CONFIG_FAIR_GROUP_SCHED
+static inline struct rq *rq_of_cfs(struct cfs_rq *cfs_rq)
+{
+ return cfs_rq->rq;
+}
+#else
+static inline struct rq *rq_of_cfs(struct cfs_rq *cfs_rq)
+{
+ return container_of(cfs_rq, struct rq, cfs);
+}
+#endif
+
+#ifdef CONFIG_RT_GROUP_SCHED
+static inline struct rq *rq_of_rt(struct rt_rq *rt_rq)
+{
+ return rt_rq->rq;
+}
+#else
+static inline struct rq *rq_of_rt(struct rt_rq *rt_rq)
+{
+ return container_of(rt_rq, struct rq, rt);
+}
+#endif
+
+static int select_cpu_for_wakeup(int task_type, struct cpumask *mask)
+{
+ int cpu = smp_processor_id(), i;
+ unsigned long load, varload;
+ struct rq *rq;
+
+ if (task_type) {
+ varload = ULONG_MAX;
+ for_each_cpu(i, mask) {
+ rq = cpu_rq(i);
+ load = rq->cfs.load.weight;
+ if (load < varload) {
+ varload = load;
+ cpu = i;
+ }
+ }
+ } else {
+ /* Here's an attempt to get a CPU within the mask where
+ * we can preempt easily. To achieve this we tried to
+ * maintain a lowbit, which indicate the lowest bit set on
+ * array bitmap. Since all CPUs contains high priority
+ * kernel threads therefore we eliminate 0, so it might not
+ * be right every time, but it's just an indicator.
+ */
+ varload = 1;
+
+ for_each_cpu(i, mask) {
+ rq = cpu_rq(i);
+ load = rq->rt.lowbit;
+ if (load >= varload) {
+ varload = load;
+ cpu = i;
+ }
+ }
+ }
+
+ return cpu;
+}
+
+static int bld_pick_cpu_cfs(struct task_struct *p, int sd_flags, int wake_flags)
+{
+ struct cfs_rq *cfs;
+ unsigned long flags;
+ unsigned int cpu = smp_processor_id();
+
+ read_lock_irqsave(&cfs_list_lock, flags);
+ list_for_each_entry(cfs, &cfs_rq_head, bld_cfs_list) {
+ cpu = cpu_of(rq_of_cfs(cfs));
+ if (cpu_online(cpu))
+ break;
+ }
+ read_unlock_irqrestore(&cfs_list_lock, flags);
+ return cpu;
+}
+
+static int bld_pick_cpu_rt(struct task_struct *p, int sd_flags, int wake_flags)
+{
+ struct rt_rq *rt;
+ unsigned long flags;
+ unsigned int cpu = smp_processor_id();
+
+ read_lock_irqsave(&rt_list_lock, flags);
+ list_for_each_entry(rt, &rt_rq_head, bld_rt_list) {
+ cpu = cpu_of(rq_of_rt(rt));
+ if (cpu_online(cpu))
+ break;
+ }
+ read_unlock_irqrestore(&rt_list_lock, flags);
+ return cpu;
+}
+
+static int bld_pick_cpu_domain(struct task_struct *p, int sd_flags, int wake_flags)
+{
+ unsigned int cpu = smp_processor_id(), want_affine = 0;
+ struct cpumask *tmpmask;
+
+ if (p->nr_cpus_allowed == 1)
+ return task_cpu(p);
+
+ if (sd_flags & SD_BALANCE_WAKE) {
+ if (cpumask_test_cpu(cpu, tsk_cpus_allowed(p))) {
+ want_affine = 1;
+ }
+ }
+
+ if (want_affine)
+ tmpmask = tsk_cpus_allowed(p);
+ else
+ tmpmask = sched_domain_span(cpu_rq(task_cpu(p))->sd);
+
+ if (rt_task(p))
+ cpu = select_cpu_for_wakeup(0, tmpmask);
+ else
+ cpu = select_cpu_for_wakeup(1, tmpmask);
+
+ return cpu;
+}
+
+static void track_load_rt(struct rq *rq, struct task_struct *p)
+{
+ unsigned long flag;
+ int firstbit;
+ struct rt_rq *first;
+ struct rt_prio_array *array = &rq->rt.active;
+
+ first = list_entry(rt_rq_head.next, struct rt_rq, bld_rt_list);
+ firstbit = sched_find_first_bit(array->bitmap);
+
+ /* Maintaining rt.lowbit */
+ if (firstbit > 0 && firstbit <= rq->rt.lowbit)
+ rq->rt.lowbit = firstbit;
+
+ if (rq->rt.lowbit < first->lowbit) {
+ write_lock_irqsave(&rt_list_lock, flag);
+ list_del(&rq->rt.bld_rt_list);
+ list_add_tail(&rq->rt.bld_rt_list, &rt_rq_head);
+ write_unlock_irqrestore(&rt_list_lock, flag);
+ }
+}
+
+static int bld_get_cpu(struct task_struct *p, int sd_flags, int wake_flags)
+{
+ unsigned int cpu;
+
+ if (sd_flags == SD_BALANCE_WAKE || (sd_flags == SD_BALANCE_EXEC && (get_nr_threads(p) > 1)))
+ cpu = bld_pick_cpu_domain(p, sd_flags, wake_flags);
+ else {
+ if (rt_task(p))
+ cpu = bld_pick_cpu_rt(p, sd_flags, wake_flags);
+ else
+ cpu = bld_pick_cpu_cfs(p, sd_flags, wake_flags);
+ }
+
+ return cpu;
+}
+
+static void bld_track_load_activate(struct rq *rq, struct task_struct *p)
+{
+ unsigned long flag;
+ if (rt_task(p)) {
+ track_load_rt(rq, p);
+ } else {
+ if (rq->cfs.pos != 2) {
+ struct cfs_rq *last;
+ last = list_entry(cfs_rq_head.prev, struct cfs_rq, bld_cfs_list);
+ if (rq->cfs.load.weight >= last->load.weight) {
+ write_lock_irqsave(&cfs_list_lock, flag);
+ list_del(&rq->cfs.bld_cfs_list);
+ list_add_tail(&rq->cfs.bld_cfs_list, &cfs_rq_head);
+ rq->cfs.pos = 2; last->pos = 1;
+ write_unlock_irqrestore(&cfs_list_lock, flag);
+ }
+ }
+ }
+}
+
+static void bld_track_load_deactivate(struct rq *rq, struct task_struct *p)
+{
+ unsigned long flag;
+ if (rt_task(p)) {
+ track_load_rt(rq, p);
+ } else {
+ if (rq->cfs.pos != 0) {
+ struct cfs_rq *first;
+ first = list_entry(cfs_rq_head.next, struct cfs_rq, bld_cfs_list);
+ if (rq->cfs.load.weight <= first->load.weight) {
+ write_lock_irqsave(&cfs_list_lock, flag);
+ list_del(&rq->cfs.bld_cfs_list);
+ list_add(&rq->cfs.bld_cfs_list, &cfs_rq_head);
+ rq->cfs.pos = 0; first->pos = 1;
+ write_unlock_irqrestore(&cfs_list_lock, flag);
+ }
+ }
+ }
+}
+#else
+static inline void bld_track_load_activate(struct rq *rq, struct task_struct *p)
+{
+}
+
+static inline void bld_track_load_deactivate(struct rq *rq, struct task_struct *p)
+{
+}
+#endif /* CONFIG_BLD */
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 44817c6..f0f3321 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -24,6 +24,8 @@
* 2007-07-01 Group scheduling enhancements by Srivatsa Vaddagiri
* 2007-11-29 RT balancing improvements by Steven Rostedt, Gregory Haskins,
* Thomas Gleixner, Mike Kravetz
+ * 2012-Feb The Barbershop Load Distribution (BLD) algorithm - an alternate
+ * CPU load distribution technique for kernel scheduler by Rakib Mullick.
*/
#include <linux/kasan.h>
@@ -87,6 +89,7 @@
#include "sched.h"
#include "../workqueue_internal.h"
#include "../smpboot.h"
+#include "bld.h"
#define CREATE_TRACE_POINTS
#include <trace/events/sched.h>
@@ -751,6 +754,8 @@ static inline void enqueue_task(struct rq *rq, struct task_struct *p, int flags)
if (!(flags & ENQUEUE_RESTORE))
sched_info_queued(rq, p);
p->sched_class->enqueue_task(rq, p, flags);
+ if (!dl_task(p))
+ bld_track_load_activate(rq, p);
}
static inline void dequeue_task(struct rq *rq, struct task_struct *p, int flags)
@@ -759,6 +764,8 @@ static inline void dequeue_task(struct rq *rq, struct task_struct *p, int flags)
if (!(flags & DEQUEUE_SAVE))
sched_info_dequeued(rq, p);
p->sched_class->dequeue_task(rq, p, flags);
+ if (!dl_task(p))
+ bld_track_load_deactivate(rq, p);
}
void activate_task(struct rq *rq, struct task_struct *p, int flags)
@@ -1588,11 +1595,17 @@ int select_task_rq(struct task_struct *p, int cpu, int sd_flags, int wake_flags)
{
lockdep_assert_held(&p->pi_lock);
+#ifndef CONFIG_BLD
if (tsk_nr_cpus_allowed(p) > 1)
cpu = p->sched_class->select_task_rq(p, cpu, sd_flags, wake_flags);
else
cpu = cpumask_any(tsk_cpus_allowed(p));
-
+#else
+ if (dl_task(p))
+ cpu = dl_sched_class.select_task_rq(p, cpu, sd_flags, wake_flags);
+ else
+ cpu = bld_get_cpu(p, sd_flags, wake_flags);
+#endif
/*
* In order not to call set_task_cpu() on a blocking task we need
* to rely on ttwu() to place the task on a valid ->cpus_allowed
@@ -1795,7 +1808,11 @@ void scheduler_ipi(void)
*/
preempt_fold_need_resched();
+#ifndef CONFIG_BLD
if (llist_empty(&this_rq()->wake_list) && !got_nohz_idle_kick())
+#else
+ if (llist_empty(&this_rq()->wake_list))
+#endif
return;
/*
@@ -1817,13 +1834,16 @@ void scheduler_ipi(void)
/*
* Check if someone kicked us for doing the nohz idle load balance.
*/
+#ifndef CONFIG_BLD
if (unlikely(got_nohz_idle_kick())) {
this_rq()->idle_balance = 1;
raise_softirq_irqoff(SCHED_SOFTIRQ);
}
+#endif
irq_exit();
}
+#ifndef CONFIG_BLD
static void ttwu_queue_remote(struct task_struct *p, int cpu, int wake_flags)
{
struct rq *rq = cpu_rq(cpu);
@@ -1837,6 +1857,7 @@ static void ttwu_queue_remote(struct task_struct *p, int cpu, int wake_flags)
trace_sched_wake_idle_without_ipi(cpu);
}
}
+#endif /* CONFIG_BLD */
void wake_up_if_idle(int cpu)
{
@@ -1873,7 +1894,7 @@ static void ttwu_queue(struct task_struct *p, int cpu, int wake_flags)
struct rq *rq = cpu_rq(cpu);
struct pin_cookie cookie;
-#if defined(CONFIG_SMP)
+#if defined(CONFIG_SMP) && !defined(CONFIG_BLD)
if (sched_feat(TTWU_QUEUE) && !cpus_share_cache(smp_processor_id(), cpu)) {
sched_clock_cpu(cpu); /* sync clocks x-cpu */
ttwu_queue_remote(p, cpu, wake_flags);
@@ -2971,7 +2992,14 @@ void sched_exec(void)
int dest_cpu;
raw_spin_lock_irqsave(&p->pi_lock, flags);
+#ifndef CONFIG_BLD
dest_cpu = p->sched_class->select_task_rq(p, task_cpu(p), SD_BALANCE_EXEC, 0);
+#else
+ if (dl_task(p))
+ dest_cpu = task_cpu(p);
+ else
+ dest_cpu = bld_get_cpu(p, SD_BALANCE_EXEC, 0);
+#endif
if (dest_cpu == smp_processor_id())
goto unlock;
@@ -3078,8 +3106,10 @@ void scheduler_tick(void)
#ifdef CONFIG_SMP
rq->idle_balance = idle_cpu(cpu);
+#ifndef CONFIG_BLD
trigger_load_balance(rq);
#endif
+#endif
rq_last_tick_reset(rq);
}
@@ -7313,7 +7343,9 @@ int sched_cpu_dying(unsigned int cpu)
raw_spin_unlock_irqrestore(&rq->lock, flags);
calc_load_migrate(rq);
update_max_interval();
+#ifndef CONFIG_BLD
nohz_balance_exit_idle(cpu);
+#endif
hrtick_clear(rq);
return 0;
}
@@ -7519,6 +7551,15 @@ void __init sched_init(void)
#endif /* CONFIG_SMP */
init_rq_hrtick(rq);
atomic_set(&rq->nr_iowait, 0);
+#ifdef CONFIG_BLD
+ INIT_LIST_HEAD(&rq->cfs.bld_cfs_list);
+ list_add_tail(&rq->cfs.bld_cfs_list, &cfs_rq_head);
+ rq->cfs.pos = 0;
+
+ INIT_LIST_HEAD(&rq->rt.bld_rt_list);
+ list_add_tail(&rq->rt.bld_rt_list, &rt_rq_head);
+ rq->rt.lowbit = INT_MAX;
+#endif
}
set_load_weight(&init_task);
@@ -7561,6 +7602,9 @@ void __init sched_init(void)
init_schedstats();
scheduler_running = 1;
+#ifdef CONFIG_BLD
+ printk(KERN_INFO "BLD: An Alternate CPU load distributor activated.\n");
+#endif
}
#ifdef CONFIG_DEBUG_ATOMIC_SLEEP
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 039de34..f823e5b 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -4924,6 +4924,7 @@ static unsigned long cpu_avg_load_per_task(int cpu)
return 0;
}
+#ifndef CONFIG_BLD
#ifdef CONFIG_FAIR_GROUP_SCHED
/*
* effective_load() calculates the load change as seen from the root_task_group
@@ -5455,6 +5456,7 @@ select_task_rq_fair(struct task_struct *p, int prev_cpu, int sd_flag, int wake_f
return new_cpu;
}
+#endif /* CONFIG_BLD */
/*
* Called immediately before a task is migrated to a new cpu; task_cpu(p) and
@@ -5785,6 +5787,7 @@ idle:
* further scheduler activity on it and we're being very careful to
* re-start the picking loop.
*/
+#ifndef CONFIG_BLD
lockdep_unpin_lock(&rq->lock, cookie);
new_tasks = idle_balance(rq);
lockdep_repin_lock(&rq->lock, cookie);
@@ -5798,7 +5801,7 @@ idle:
if (new_tasks > 0)
goto again;
-
+#endif /* CONFIG_BLD */
return NULL;
}
@@ -6459,8 +6462,9 @@ static unsigned long task_h_load(struct task_struct *p)
}
#endif
-/********** Helpers for find_busiest_group ************************/
+#ifndef CONFIG_BLD
+/********** Helpers for find_busiest_group ************************/
enum group_type {
group_other = 0,
group_imbalanced,
@@ -6551,6 +6555,7 @@ static inline int get_sd_load_idx(struct sched_domain *sd,
return load_idx;
}
+#endif /* CONFIG_BLD */
static unsigned long scale_rt_capacity(int cpu)
{
@@ -6659,6 +6664,7 @@ void update_group_capacity(struct sched_domain *sd, int cpu)
sdg->sgc->capacity = capacity;
}
+#ifndef CONFIG_BLD
/*
* Check whether the capacity of the rq has been noticeably reduced by side
* activity. The imbalance_pct is used for the threshold.
@@ -7892,6 +7898,7 @@ static inline int on_null_domain(struct rq *rq)
{
return unlikely(!rcu_dereference_sched(rq->sd));
}
+#endif /* CONFIG_BLD */
#ifdef CONFIG_NO_HZ_COMMON
/*
@@ -7900,12 +7907,39 @@ static inline int on_null_domain(struct rq *rq)
* needed, they will kick the idle load balancer, which then does idle
* load balancing for all the idle CPUs.
*/
+#ifndef CONFIG_BLD
static struct {
cpumask_var_t idle_cpus_mask;
atomic_t nr_cpus;
unsigned long next_balance; /* in jiffy units */
} nohz ____cacheline_aligned;
+void nohz_balance_exit_idle(unsigned int cpu)
+{
+ if (unlikely(test_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu)))) {
+ /*
+ * Completely isolated CPUs don't ever set, so we must test.
+ */
+ if (likely(cpumask_test_cpu(cpu, nohz.idle_cpus_mask))) {
+ cpumask_clear_cpu(cpu, nohz.idle_cpus_mask);
+ atomic_dec(&nohz.nr_cpus);
+ }
+ clear_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu));
+ }
+}
+
+static int sched_ilb_notifier(struct notifier_block *nfb,
+ unsigned long action, void *hcpu)
+{
+ switch (action & ~CPU_TASKS_FROZEN) {
+ case CPU_DYING:
+ nohz_balance_exit_idle(smp_processor_id());
+ return NOTIFY_OK;
+ default:
+ return NOTIFY_DONE;
+ }
+}
+
static inline int find_new_ilb(void)
{
int ilb = cpumask_first(nohz.idle_cpus_mask);
@@ -7944,20 +7978,6 @@ static void nohz_balancer_kick(void)
return;
}
-void nohz_balance_exit_idle(unsigned int cpu)
-{
- if (unlikely(test_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu)))) {
- /*
- * Completely isolated CPUs don't ever set, so we must test.
- */
- if (likely(cpumask_test_cpu(cpu, nohz.idle_cpus_mask))) {
- cpumask_clear_cpu(cpu, nohz.idle_cpus_mask);
- atomic_dec(&nohz.nr_cpus);
- }
- clear_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu));
- }
-}
-
static inline void set_cpu_sd_state_busy(void)
{
struct sched_domain *sd;
@@ -7974,6 +7994,8 @@ static inline void set_cpu_sd_state_busy(void)
unlock:
rcu_read_unlock();
}
+#endif /* CONFIG_BLD */
+#endif /* NO_HZ_COMMON */
void set_cpu_sd_state_idle(void)
{
@@ -7998,6 +8020,7 @@ unlock:
*/
void nohz_balance_enter_idle(int cpu)
{
+#ifndef CONFIG_BLD
/*
* If this cpu is going down, then nothing needs to be done.
*/
@@ -8016,10 +8039,8 @@ void nohz_balance_enter_idle(int cpu)
cpumask_set_cpu(cpu, nohz.idle_cpus_mask);
atomic_inc(&nohz.nr_cpus);
set_bit(NOHZ_TICK_STOPPED, nohz_flags(cpu));
-}
#endif
-
-static DEFINE_SPINLOCK(balancing);
+}
/*
* Scale the max load_balance interval with the number of CPUs in the system.
@@ -8030,6 +8051,9 @@ void update_max_interval(void)
max_load_balance_interval = HZ*num_online_cpus()/10;
}
+#ifndef CONFIG_BLD
+static DEFINE_SPINLOCK(balancing);
+
/*
* It checks each scheduling domain to see if it is due to be balanced,
* and initiates a balancing operation if so.
@@ -8317,6 +8341,7 @@ void trigger_load_balance(struct rq *rq)
nohz_balancer_kick();
#endif
}
+#endif /* CONFIG_BLD */
static void rq_online_fair(struct rq *rq)
{
@@ -8332,7 +8357,6 @@ static void rq_offline_fair(struct rq *rq)
/* Ensure any throttled groups are reachable by pick_next_task */
unthrottle_offline_cfs_rqs(rq);
}
-
#endif /* CONFIG_SMP */
/*
@@ -8791,7 +8815,9 @@ const struct sched_class fair_sched_class = {
.put_prev_task = put_prev_task_fair,
#ifdef CONFIG_SMP
+#ifndef CONFIG_BLD
.select_task_rq = select_task_rq_fair,
+#endif
.migrate_task_rq = migrate_task_rq_fair,
.rq_online = rq_online_fair,
@@ -8852,6 +8878,7 @@ void show_numa_stats(struct task_struct *p, struct seq_file *m)
__init void init_sched_fair_class(void)
{
+#ifndef CONFIG_BLD
#ifdef CONFIG_SMP
open_softirq(SCHED_SOFTIRQ, run_rebalance_domains);
@@ -8860,5 +8887,5 @@ __init void init_sched_fair_class(void)
zalloc_cpumask_var(&nohz.idle_cpus_mask, GFP_NOWAIT);
#endif
#endif /* SMP */
-
+#endif /* BLD */
}
diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
index d5690b7..6f3589e 100644
--- a/kernel/sched/rt.c
+++ b/kernel/sched/rt.c
@@ -1375,6 +1375,7 @@ static void yield_task_rt(struct rq *rq)
#ifdef CONFIG_SMP
static int find_lowest_rq(struct task_struct *task);
+#ifndef CONFIG_BLD
static int
select_task_rq_rt(struct task_struct *p, int cpu, int sd_flag, int flags)
{
@@ -1430,6 +1431,7 @@ select_task_rq_rt(struct task_struct *p, int cpu, int sd_flag, int flags)
out:
return cpu;
}
+#endif /* CONFIG_BLD */
static void check_preempt_equal_prio(struct rq *rq, struct task_struct *p)
{
@@ -2335,7 +2337,9 @@ const struct sched_class rt_sched_class = {
.put_prev_task = put_prev_task_rt,
#ifdef CONFIG_SMP
+#ifndef CONFIG_BLD
.select_task_rq = select_task_rq_rt,
+#endif
.set_cpus_allowed = set_cpus_allowed_common,
.rq_online = rq_online_rt,
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index c64fc51..a1d329b 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -416,9 +416,8 @@ struct cfs_rq {
#endif /* CONFIG_FAIR_GROUP_SCHED */
#endif /* CONFIG_SMP */
-#ifdef CONFIG_FAIR_GROUP_SCHED
struct rq *rq; /* cpu runqueue to which this cfs_rq is attached */
-
+#ifdef CONFIG_FAIR_GROUP_SCHED
/*
* leaf cfs_rqs are those that hold tasks (lowest schedulable entity in
* a hierarchy). Non-leaf lrqs hold other higher schedulable entities
@@ -442,6 +441,11 @@ struct cfs_rq {
struct list_head throttled_list;
#endif /* CONFIG_CFS_BANDWIDTH */
#endif /* CONFIG_FAIR_GROUP_SCHED */
+
+#ifdef CONFIG_BLD
+ struct list_head bld_cfs_list;
+ char pos;
+#endif
};
static inline int rt_bandwidth_enabled(void)
@@ -487,12 +491,16 @@ struct rt_rq {
/* Nests inside the rq lock: */
raw_spinlock_t rt_runtime_lock;
+ struct rq *rq;
#ifdef CONFIG_RT_GROUP_SCHED
unsigned long rt_nr_boosted;
- struct rq *rq;
struct task_group *tg;
#endif
+#ifdef CONFIG_BLD
+ struct list_head bld_rt_list;
+ int lowbit;
+#endif
};
/* Deadline class' related fields in a runqueue */

24
patches_disabled/ck2-version.patch Normal file
View File

@ -0,0 +1,24 @@
Add -ck version in inconspicuous place where it will merge relatively easily
with later kernel versions.
-ck
---
Makefile | 4 ++++
1 file changed, 4 insertions(+)
Index: linux-3.12-ck2/Makefile
===================================================================
--- linux-3.12-ck2.orig/Makefile 2013-12-03 20:12:24.295109675 +1100
+++ linux-3.12-ck2/Makefile 2013-12-03 20:12:24.285109800 +1100
@@ -10,6 +10,10 @@ NAME = One Giant Leap for Frogkind
# Comments in this file are targeted only to the developer, do not
# expect to learn how to build the kernel reading this file.
+CKVERSION = -ck2
+CKNAME = BFS Powered
+EXTRAVERSION := $(EXTRAVERSION)$(CKVERSION)
+
# Do not:
# o use make's built-in rules and variables
# (this increases performance and avoids hard-to-debug behaviour);

21
patches_disabled/hz-default_1000.patch Normal file
View File

@ -0,0 +1,21 @@
Set default HZ to 1000 which is what most desktop users should still be using.
-ck
---
kernel/Kconfig.hz | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
Index: linux-3.12-ck2/kernel/Kconfig.hz
===================================================================
--- linux-3.12-ck2.orig/kernel/Kconfig.hz 2013-12-03 20:12:23.401120851 +1100
+++ linux-3.12-ck2/kernel/Kconfig.hz 2013-12-03 20:12:23.390120989 +1100
@@ -4,7 +4,7 @@
choice
prompt "Timer frequency"
- default HZ_250
+ default HZ_1000
help
Allows the configuration of the timer frequency. It is customary
to have the timer interrupt run at 1000 Hz but 100 Hz may be more

53
patches_disabled/hz-no_default_250.patch Normal file
View File

@ -0,0 +1,53 @@
Make 250HZ not be the default to discourage desktop users from choosing this
option since 1000 will provide better latencies with only miniscule amounts
of extra overhead and power consumption.
-ck
---
kernel/Kconfig.hz | 17 ++++++++++-------
1 file changed, 10 insertions(+), 7 deletions(-)
Index: linux-3.12-ck2/kernel/Kconfig.hz
===================================================================
--- linux-3.12-ck2.orig/kernel/Kconfig.hz 2013-12-03 20:12:22.956126414 +1100
+++ linux-3.12-ck2/kernel/Kconfig.hz 2013-12-03 20:12:22.946126539 +1100
@@ -23,13 +23,14 @@ choice
with lots of processors that may show reduced performance if
too many timer interrupts are occurring.
- config HZ_250
+ config HZ_250_NODEFAULT
bool "250 HZ"
help
- 250 Hz is a good compromise choice allowing server performance
- while also showing good interactive responsiveness even
- on SMP and NUMA systems. If you are going to be using NTSC video
- or multimedia, selected 300Hz instead.
+ 250 HZ is a lousy compromise choice allowing server interactivity
+ while also showing desktop throughput and no extra power saving on
+ laptops. No good for anything.
+
+ Recommend 100 or 1000 instead.
config HZ_300
bool "300 HZ"
@@ -43,14 +44,16 @@ choice
bool "1000 HZ"
help
1000 Hz is the preferred choice for desktop systems and other
- systems requiring fast interactive responses to events.
+ systems requiring fast interactive responses to events. Laptops
+ can also benefit from this choice without sacrificing battery life
+ if dynticks is also enabled.
endchoice
config HZ
int
default 100 if HZ_100
- default 250 if HZ_250
+ default 250 if HZ_250_NODEFAULT
default 300 if HZ_300
default 1000 if HZ_1000

46
patches_disabled/kconfig-expose_vmsplit_option.patch Normal file
View File

@ -0,0 +1,46 @@
The options to alter the vmsplit to enable more lowmem are hidden behind the
expert option. Make it more exposed for -ck users and make the help menu
more explicit about what each option means.
-ck
---
arch/x86/Kconfig | 12 ++++++------
1 file changed, 6 insertions(+), 6 deletions(-)
Index: linux-3.12-ck2/arch/x86/Kconfig
===================================================================
--- linux-3.12-ck2.orig/arch/x86/Kconfig 2013-12-03 20:12:23.848115263 +1100
+++ linux-3.12-ck2/arch/x86/Kconfig 2013-12-03 20:12:23.838115388 +1100
@@ -1175,7 +1175,7 @@ config HIGHMEM64G
endchoice
choice
- prompt "Memory split" if EXPERT
+ prompt "Memory split"
default VMSPLIT_3G
depends on X86_32
---help---
@@ -1195,17 +1195,17 @@ choice
option alone!
config VMSPLIT_3G
- bool "3G/1G user/kernel split"
+ bool "Default 896MB lowmem (3G/1G user/kernel split)"
config VMSPLIT_3G_OPT
depends on !X86_PAE
- bool "3G/1G user/kernel split (for full 1G low memory)"
+ bool "1GB lowmem (3G/1G user/kernel split)"
config VMSPLIT_2G
- bool "2G/2G user/kernel split"
+ bool "2GB lowmem (2G/2G user/kernel split)"
config VMSPLIT_2G_OPT
depends on !X86_PAE
- bool "2G/2G user/kernel split (for full 2G low memory)"
+ bool "2GB lowmem (2G/2G user/kernel split)"
config VMSPLIT_1G
- bool "1G/3G user/kernel split"
+ bool "3GB lowmem (1G/3G user/kernel split)"
endchoice
config PAGE_OFFSET

40
patches_disabled/preempt-desktop-tune.patch Normal file
View File

@ -0,0 +1,40 @@
Enable preempt by default and make people steer away from voluntary.
-ck
---
kernel/Kconfig.preempt | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
Index: linux-3.12-ck2/kernel/Kconfig.preempt
===================================================================
--- linux-3.12-ck2.orig/kernel/Kconfig.preempt 2013-12-03 20:12:22.511131977 +1100
+++ linux-3.12-ck2/kernel/Kconfig.preempt 2013-12-03 20:12:22.500132115 +1100
@@ -1,7 +1,7 @@
choice
prompt "Preemption Model"
- default PREEMPT_NONE
+ default PREEMPT
config PREEMPT_NONE
bool "No Forced Preemption (Server)"
@@ -17,7 +17,7 @@ config PREEMPT_NONE
latencies.
config PREEMPT_VOLUNTARY
- bool "Voluntary Kernel Preemption (Desktop)"
+ bool "Voluntary Kernel Preemption (Nothing)"
help
This option reduces the latency of the kernel by adding more
"explicit preemption points" to the kernel code. These new
@@ -31,7 +31,8 @@ config PREEMPT_VOLUNTARY
applications to run more 'smoothly' even when the system is
under load.
- Select this if you are building a kernel for a desktop system.
+ Select this for no system in particular (choose Preemptible
+ instead on a desktop if you know what's good for you).
config PREEMPT
bool "Preemptible Kernel (Low-Latency Desktop)"

7064
patches_disabled/uksm-0.1.2.2-for-v3.12.patch Normal file
View File

File diff suppressed because it is too large Load Diff

6896
patches_disabled/uksm-0.1.2.5-for-v4.8.patch Normal file
View File

File diff suppressed because it is too large Load Diff