Checks the pixel blending mode and plane alpha value when
do the plane_check. Mali DP supports blending the current plane
with the background either based on the pixel alpha blending
mode or by using the layer's alpha value, but not both at the
same time. If both case, plane_check will return failed.
Sets the HW when doing plane_update accordingly. If plane alpha
is the 0xffff, set the pixel blending bits accordingly. If not
we'd set ALPHA bit as zero and layer alpha value.
Changes since v1:
- Introduces to use it in the malidp driver, which depends on
the plane alpha patch
Changes since v2:
- Refines the comments of drm/mali-dp patchset
Changes since v3:
- Adds hardware limitation check
Changes since v4:
- Updates on drm/malidp, hardware limitation check only when
the format has alpha pixel.
- Rebases on drm-misc-next.
Signed-off-by: Lowry Li <lowry.li@arm.com>
Acked-by: Liviu Dudau <liviu.dudau@arm.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Smart layer enable rectangles is set to 1 when the driver is probed,
however when doing pm_suspend the value is lost and it's not set again
making the SMART_LAYER unusable, fix that by initializing the number
of rectangles everytime we do a plane update.
Signed-off-by: Alexandru Gheorghe <alexandru-cosmin.gheorghe@arm.com>
Acked-by: Liviu Dudau <liviu.dudau@arm.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
When we want to writeback to memory in NV12 format we need to program
the RGB2YUV coefficients. Currently, we don't program the coefficients
and NV12 doesn't work at all.
This patchset fixes that by programming a sane default(bt709, limited
range) as rgb2yuv coefficients.
In the long run, probably we need to think of a way for userspace to
be able to program that, but for now I think this is better than not
working at all or not advertising NV12 as a supported format for
memwrite.
Changes since v1:
- Write the rgb2yuv coefficients only once, since we don't change
them at all, just write them the first time NV12 is programmed,
suggested by Brian Starkey, here [1]
[1] https://lists.freedesktop.org/archives/dri-devel/2018-August/186819.html
Signed-off-by: Alexandru Gheorghe <alexandru-cosmin.gheorghe@arm.com>
Acked-by: Liviu Dudau <liviu.dudau@arm.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Currently, if userspace calls drm_wait_vblank before the crtc is
activated the crtc vblank_enable hook is called, which in case of
malidp driver triggers some warninngs. This happens because on
device init we don't inform the drm core about the vblank state
by calling drm_crtc_vblank_on/off/reset which together with
drm_vblank_get have some magic that prevents calling drm_vblank_enable
when crtc is off.
Signed-off-by: Alexandru Gheorghe <alexandru-cosmin.gheorghe@arm.com>
Acked-by: Liviu Dudau <liviu.dudau@arm.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Add plane alpha blending support with the different blend modes.
This has been tested on a icl to show the correct results,
on earlier platforms small rounding errors cause issues. But this
already happens case with fully transparant or fully opaque RGB8888
fb's.
The recommended HW workaround is to disable alpha blending when the
plane alpha is 0 (transparant, hide plane) or 0xff (opaque, disable blending).
This is easy to implement on any platform, so just do that.
The tests for userspace are also available, and pass on gen11.
Changes since v1:
- Change mistaken < 0xff0 to 0xff00.
- Only set PLANE_KEYMSK_ALPHA_ENABLE when plane alpha < 0xff00, ignore blend mode.
- Rework disabling FBC when per pixel alpha is used.
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
[mlankhorst: Change MISSING_CASE default to explicit alpha disable (mattrope)]
Link: https://patchwork.freedesktop.org/patch/msgid/20180815103405.22679-1-maarten.lankhorst@linux.intel.com
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Add pins, groups, and function for the Interrupt Controller for
External Devices (INTC-EX) on the R-Car E3 SoC.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Simon Horman <horms+renesas@verge.net.au>
This provides a pinctrl driver for the Renesas RZ/N1 device family.
Based on a patch originally written by Michel Pollet at Renesas.
Signed-off-by: Phil Edworthy <phil.edworthy@renesas.com>
Reviewed-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
The Renesas RZ/N1 device family PINCTRL node description.
Based on a patch originally written by Michel Pollet at Renesas.
Signed-off-by: Phil Edworthy <phil.edworthy@renesas.com>
Reviewed-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
platform_msi_create_device_domain() always creates a revmap-based
irqdomain, which has the drawback of requiring the number of MSIs
that can be allocated ahead of time. This is not always possible,
and we sometimes need to use a tree-based irqdomain instead.
Add a new platform_msi_create_device_tree_domain() helper to
that effect.
Reported-by: Miquel Raynal <miquel.raynal@bootlin.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
mvebu soc for 4.20 (part 1)
- use dt_fixup to provide fallback for enable-method for Armada XP
- document the marvell,prestera compatible string
- update Thomas Petazzoni email in MAINTAINERS file
* tag 'mvebu-soc-4.20-1' of git://git.infradead.org/linux-mvebu:
dt-bindings: marvell,prestera: Add common compatible string
MAINTAINERS: replace free-electrons.com by bootlin.com for Thomas Petazzoni
ARM: mvebu: use dt_fixup to provide fallback for enable-method
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
i.MX SoC update for 4.20:
- Add ipg clock support in MMDC driver for registers access, so that
we will be safe even if the clock is not turned on by firmware.
- Register pm_power_off handler to provide power off support for iMX6
based boards with external PMIC.
- Add platform code support for i.MX 6ULZ SoC which is a derivative of
i.MX6ULL with some modules removed.
* tag 'imx-soc-4.20' of git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux:
ARM: imx: add i.mx6ulz msl support
ARM: imx6: register pm_power_off handler if "fsl,pmic-stby-poweroff" is set
ARM: imx: add mmdc ipg clock operation for mmdc
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Berlin SoC files has been moved from marvell dir to synaptics dir, but
commit bc52497a59 ("MAINTAINERS: update entry for ARM/berlin") didn't
update the dir accordingly. This patch fixes it.
From another side, new derivative SoCs from Synaptics may not be named
as berlin*, so let's update the entries accordingly.
Fixes: bc52497a59 ("MAINTAINERS: update entry for ARM/berlin")
Signed-off-by: Jisheng Zhang <Jisheng.Zhang@synaptics.com>
Reported-by: Joe Perches <joe@perches.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Drop include/dt-bindings/genpd/k2g.h which disappeared from kernel tree
some time back, however MAINTAINERS file was missed to be updated.
Fixes: d16645054d ("dt-bindings: Drop k2g genpd device ID macros")
Cc: Rob Herring <robh@kernel.org>
Cc: Dave Gerlach <d-gerlach@ti.com>
Cc: Santosh Shilimkar <ssantosh@kernel.org>
Cc: Tero Kristo <t-kristo@ti.com>
Reported-by: Joe Perches <joe@perches.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Acked-by: Santosh Shilimkar <ssantosh@kernel.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
This branch contains a series of improvments for omap3-gta04 phone,
and a series of clean-up for am335x to remove the deprecated phy_id
property.
The rest is to configure am57xx-idk boards for leds, load trigger,
and smps, am3517-evm audio configuration, beaglebone hdmi cec support,
coresight binding update, and fixes for i2c and spi warnings.
-----BEGIN PGP SIGNATURE-----
iQJFBAABCAAvFiEEkgNvrZJU/QSQYIcQG9Q+yVyrpXMFAlutJwIRHHRvbnlAYXRv
bWlkZS5jb20ACgkQG9Q+yVyrpXMMyw/+Oj8WK2Dej372d0m2I2401fLYD4L5B99U
H/It6JAAmumv2y/6IZEiPAC+TtBpTgbmMlQzyciaKRyWCjqpfg1DDb1mSNf+3RSq
VFsCkOx5cymGL2yKn8DzyIYGRejwkbfDVH6UbQTHv84mLtgX/8liqAdeSMOShx22
r7G1ismE4UwwNlVKVSd94Y/fJArDlanbRJqTXl+PVqPPDpC/MoMOu8KpPJ0hJyWi
pRYxJHSGTGttLxYv9w5i5S3N4C2b/A3vFfMo35EqHXr/t08Lfcqh1vhaMtcXRdQO
TrOMF4rvWZA4H8uVP+8sqQfwtcDrCnCBoh1VeFJrAHBdtfEJU2rqq2BoVvQXJYCi
WWu/iaa13C8xktS46NvBE4UODAxo52ciyQEzckDbQSjdUm7G82sfciSVRlWfrtjM
rBOyaNMWDTWEE7H99O75QlEA4E4J8xcgpmkHplZ3CuNEmtJx/OeknXD7F91g6FJt
13TmUj1/Wu72ZZLlkeazLLdVG1lj6M4WKdv3L3oDbaJkLAiRSQj/oIAi/R1gFbcW
i+SejX4NPqjA1f1i3tiKzmBY70ur0jPnR9iNNyMs+9VzqbRUQJQd/DOQG7TA9BYk
qE0jWBjnuWjo3+EAvel9LrIrJvrApXaJPvFC6eLaAjLbqkJ81lqwGQIivHrgHih5
vIWlcV1hWoA=
=bCXk
-----END PGP SIGNATURE-----
Merge tag 'omap-for-v4.20/dt-signed-v2' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap into next/dt
Devicetree changes for omap variants
This branch contains a series of improvments for omap3-gta04 phone,
and a series of clean-up for am335x to remove the deprecated phy_id
property.
The rest is to configure am57xx-idk boards for leds, load trigger,
and smps, am3517-evm audio configuration, beaglebone hdmi cec support,
coresight binding update, and fixes for i2c and spi warnings.
* tag 'omap-for-v4.20/dt-signed-v2' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap: (70 commits)
ARM: dts: add omap3-gta04a5one to Makefile
ARM: dts: omap3-gta04: add pulldown/up settings for twl4030 gpio
ARM: dts: am335x-boneblack: add cec support
ARM: dts: am3517-evm: Add support for UI board and Audio
ARM: dts: gta04: add serial console wakeup irq
ARM: dts: am57xx-idk-common: Hook smps12 regulator as cpu vdd-supply
ARM: dts: omap: Update coresight bindings for hardware ports
ARM: dts: ti: Fix SPI and I2C bus warnings
ARM: dts: dra62x-j5eco-evm: get rid of phy_id property
ARM: dts: dm8148-t410: get rid of phy_id property
ARM: dts: dm8148-evm: get rid of phy_id property
ARM: dts: am57xx-cl-som-am57x: get rid of phy_id property
ARM: dts: am57xx-idk-common: get rid of phy_id property
ARM: dts: dra7-evm: get rid of phy_id property
ARM: dts: dra71-evm: get rid of phy_id property
ARM: dts: dra72-evm-revc: get rid of phy_id property
ARM: dts: dra72-evm: get rid of phy_id property
ARM: dts: dra76-evm: get rid of phy_id property
ARM: dts: am437x-cm-t43: get rid of phy_id property
ARM: dts: am437x-gp-evm: get rid of phy_id property
...
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Document the R-Car E3 (R8A77990) SoC in the R-Car PCIe bindings.
Signed-off-by: Tho Vu <tho.vu.wh@rvc.renesas.com>
Signed-off-by: Kazuya Mizuguchi <kazuya.mizuguchi.ks@renesas.com>
Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Simon Horman <horms+renesas@verge.net.au>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Cc: Geert Uytterhoeven <geert+renesas@glider.be>
Cc: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Cc: Phil Edworthy <phil.edworthy@renesas.com>
Cc: Rob Herring <robh@kernel.org>
Cc: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Cc: Simon Horman <horms+renesas@verge.net.au>
Cc: Wolfram Sang <wsa@the-dreams.de>
Cc: linux-renesas-soc@vger.kernel.org
- Add watchdog node on Armada 37xx
- Update PPv2 interrupts name
- Add support for the SolidRun Clearfog GT 8K (Aramda 8040 based)
- Add thermal-zone nodes for Aramda 7K/8K
-----BEGIN PGP SIGNATURE-----
iF0EABECAB0WIQQYqXDMF3cvSLY+g9cLBhiOFHI71QUCW7HpQwAKCRALBhiOFHI7
1a8HAJ4tUZNYmuo9AWVWLj+FJaTBgQSJYgCaA+tkbl2J+bsZfL7JFpHGVukKmzc=
=4cF/
-----END PGP SIGNATURE-----
Merge tag 'mvebu-dt64-4.20-1' of git://git.infradead.org/linux-mvebu into next/dt
mvebu dt64 for 4.20 (part 1)
- Add watchdog node on Armada 37xx
- Update PPv2 interrupts name
- Add support for the SolidRun Clearfog GT 8K (Aramda 8040 based)
- Add thermal-zone nodes for Aramda 7K/8K
* tag 'mvebu-dt64-4.20-1' of git://git.infradead.org/linux-mvebu:
arm64: dts: marvell: armada-37xx: add nodes to support watchdog
arm64: dts: marvell: armada-cp110: describe more PPv2 interrupts
arm64: dts: marvell: armada-cp110: change the PPv2 IRQ names
arm64: dts: add support for SolidRun Clearfog GT 8K
arm64: dts: marvell: add thermal-zone node in cp110 DTSI file
arm64: dts: marvell: add macro to make distinction between node names
arm64: dts: marvell: add thermal-zone node in ap806 DTSI file
arm64: dts: marvell: move AP806/CP110 thermal nodes into a new syscon
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
- updates the armada-xp-98dx3236 SoC and related boards to use the new
style dts bindings for nand
- add db-88f6820-amc board: plugin card for some of Marvell's switch
development kits
- fix SPI and I2C bus warnings coming with the new checks in dtc
- add new compatible string "marvell,prestera" to the armada-xp-98dx*
- fix sdhci supply property name on the clearfog (the '-supply' suffix
was missing)
-----BEGIN PGP SIGNATURE-----
iF0EABECAB0WIQQYqXDMF3cvSLY+g9cLBhiOFHI71QUCW7HlPQAKCRALBhiOFHI7
1WD8AKCdfJ2vxjsHeFJ5YEOA7JtdjJaaXwCffYZJ9IS0XblfyH9bsh9P/cy7nXI=
=CsQf
-----END PGP SIGNATURE-----
Merge tag 'mvebu-dt-4.20-1' of git://git.infradead.org/linux-mvebu into next/dt
mvebu dt for 4.20 (part 1)
- updates the armada-xp-98dx3236 SoC and related boards to use the new
style dts bindings for nand
- add db-88f6820-amc board: plugin card for some of Marvell's switch
development kits
- fix SPI and I2C bus warnings coming with the new checks in dtc
- add new compatible string "marvell,prestera" to the armada-xp-98dx*
- fix sdhci supply property name on the clearfog (the '-supply' suffix
was missing)
* tag 'mvebu-dt-4.20-1' of git://git.infradead.org/linux-mvebu:
ARM: dts: clearfog: fix sdhci supply property name
ARM: dts: mvebu: add "marvell,prestera" to PP nodes
ARM: dts: marvell: Fix SPI and I2C bus warnings
ARM: dts: mvebu: Add device tree for db-88f6820-amc board
ARM: dts: mvebu: db-xc3-24g4: use new style nand binding
ARM: dts: mvebu: db-dxbc2: use new style nand binding
ARM: dts: mvebu: 98dx3236: Rename nand controller node
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
- Add the second Dual UART device for LS208xA SoCs.
- Add necessary big-endian property for NOR device on LS104xA based
boards, remove unneeded big-endian property from IFC controller.
- DTC has new checks for I2C and SPI buses to land into 4.20. A patch
from Rob to fix the bus node names and warnings in unit-addresses.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQEcBAABAgAGBQJbscRcAAoJEFBXWFqHsHzOxVQIALdjNChTrg5K0GBRTavdl+Hq
sOjW0W6Hh5MN4r1wKpSsKAPD/XsPmjMpsreC/qlR+uFGleKzhrHWRtbmR/FG7Aut
Sv052Yew5Fc8mWzNGNwQes6rpagRioqnLD6L6bPFoelgGsVFxrysDhAWhBgp7Zj8
HhQvgVaP2t2FDJX6O3Vbve4OgUNrmg5fp6bm19X3BSDa8Bsm6+9QU1w0ulCeWA/R
HXiGtECO+N1td3ViVh2iZMoQgvXSxY0twn2srMlFP+TsfOxKvC8lN2/RhwqtrbQs
AIWTbCuJHukGUQfRuIhVYhPUVZomPuaBUIBm8Semjn/dhjmBK7eXytFKY7AQcKM=
=BFcP
-----END PGP SIGNATURE-----
Merge tag 'imx-dt64-4.20' of git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux into next/dt
Freescale arm64 device tree update for 4.20:
- Add the second Dual UART device for LS208xA SoCs.
- Add necessary big-endian property for NOR device on LS104xA based
boards, remove unneeded big-endian property from IFC controller.
- DTC has new checks for I2C and SPI buses to land into 4.20. A patch
from Rob to fix the bus node names and warnings in unit-addresses.
* tag 'imx-dt64-4.20' of git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux:
arm64: dts: fsl: Fix I2C and SPI bus warnings
arm64: dts: ls208xa: add second duart
arm64: dts: fsl: remove big-endian field from IFC controller
arm64: dts: Add big-endian in nor node for ls104xa
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
- New board support: Engicam's i.Core MX6 CPU module v1.5; ConnectCore
6UL Single Board Computer (SBC) Pro; i.MX6 ULZ based EVK board.
- Add Add SFF interface support for vf610-zii board.
- Disable unneeded devices like VPU and internal watchdog for imx51-zii
boards.
- Add 'no-sdio' and 'no-sd' property for vf610-zii-cfu1 board.
- Improve i.MX6 SLL GPIO support by adding gpio-ranges property and
clocks information.
- Update iomux header for i.MX7 Solo and i.MX6 ULL.
- Enable GPIO buttons as wakeup source for imx7d-sdb and imx6sx-sdb.
- Add GPIO keys and egalax touch screen support for imx6qdl-sabreauto.
- Switch to use SPDX-License-Identifier for more boards - vf610-twr,
imx7s-warp, Engicam boards.
- Add device tree bindings of 'fsl,pmic-stby-poweroff' property and add
the support for i.MX6 RIoTboard.
- DTC has new checks for SPI buses which will be landed on 4.20.
A patch from Rob to fix those 100+ warnings on i.MX boards. (Thanks!)
- Switch i.MX7 device tree to use updated coresight binding for
hardware ports.
- Misc small or random update and cleanup.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQEcBAABAgAGBQJbsb9RAAoJEFBXWFqHsHzObO8H/Al83K27UxL8XjKMOUsMooWX
BG5FfS9mDKOe8bN0BwZnUmrjuOOkNWooMxyahBXP9e0CWZnK/4G3OvTS83PNpA34
zE+Nm1oJia9zoUsWoCpBrvLrBJpjJqVajQiXzddUDgYLhbNXPXkiwBBEZqrwAJBK
CsgwRCj4qJA1WCetvwjYVVwtcVTCCnYcpx73YQMLK1sSbi0oVqtSMynSF3spzDVC
iduF+ur88lkjo0IK1PaWMOi8+G6E5bOV//4JJky5hNcTuMsmAAywmUEALoIYQjLr
EitPoTNi2FDnFDpGPqvsf51Cez6e+3iCAk0F5V6D+6nIvNDF9mGRdyqSnDhIyBY=
=B0Zy
-----END PGP SIGNATURE-----
Merge tag 'imx-dt-4.20' of git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux into next/dt
i.MX device tree update for 4.20:
- New board support: Engicam's i.Core MX6 CPU module v1.5; ConnectCore
6UL Single Board Computer (SBC) Pro; i.MX6 ULZ based EVK board.
- Add Add SFF interface support for vf610-zii board.
- Disable unneeded devices like VPU and internal watchdog for imx51-zii
boards.
- Add 'no-sdio' and 'no-sd' property for vf610-zii-cfu1 board.
- Improve i.MX6 SLL GPIO support by adding gpio-ranges property and
clocks information.
- Update iomux header for i.MX7 Solo and i.MX6 ULL.
- Enable GPIO buttons as wakeup source for imx7d-sdb and imx6sx-sdb.
- Add GPIO keys and egalax touch screen support for imx6qdl-sabreauto.
- Switch to use SPDX-License-Identifier for more boards - vf610-twr,
imx7s-warp, Engicam boards.
- Add device tree bindings of 'fsl,pmic-stby-poweroff' property and add
the support for i.MX6 RIoTboard.
- DTC has new checks for SPI buses which will be landed on 4.20.
A patch from Rob to fix those 100+ warnings on i.MX boards. (Thanks!)
- Switch i.MX7 device tree to use updated coresight binding for
hardware ports.
- Misc small or random update and cleanup.
* tag 'imx-dt-4.20' of git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux: (50 commits)
ARM: dts: imx: add i.mx6ulz and i.mx6ulz 14x14 evk support
dt-bindings: arm: add compatible for i.MX6ULZ 14x14 EVK board
ARM: dts: imx53-ppd: Remove 'num-chipselects' property
ARM: dts: vf610-twr: Switch to SPDX identifier
ARM: dts: vf: Switch to SPDX identifier
ARM: dts: imx6qdl-zii-rdu2: Disable the internal RTC
ARM: dts: imx51-zii-rdu1: Fix the rtc compatible string
ARM: dts: imx6ul: use nvmem-cells for cpu speed grading
ARM: dts: imx: Fix SPI bus warnings
ARM: dts: imx7: Update coresight binding for hardware ports
ARM: dts: vf610-zii-cfu1: Pass the 'no-sd' property
ARM: dts: vf610-zii-cfu1: Pass the 'no-sdio' property
ARM: dts: imx51-zii-scu2-mezz: Disable the internal watchdog
ARM: dts: imx51-zii-scu2-mezz: Disable VPU
ARM: dts: imx51-zii-scu3-esb: Disable VPU
ARM: dts: imx51: Add label for VPU node
ARM: dts: imx6ull: update vdd_soc voltage for 900MHz operating point
ARM: dts: imx6ul: Add DTS for ConnectCore 6UL SBC Pro
ARM: dts: imx6: RIoTboard provide standby on power off option
dt-bindings: imx6q-clock: add new fsl,pmic-stby-poweroff property
...
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Document support for the Interrupt Controller for External Devices
(INTC-EX) in the Renesas E3 (r8a77990) SoC.
No driver update is needed.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
The PDC irqchp can convert a falling edge or level low interrupt to a
rising edge or level high interrupt at the GIC. We just need to setup
the GIC correctly. Set up the interrupt type for the IRQ_TYPE_EDGE_BOTH
as IRQ_TYPE_EDGE_RISING at the GIC.
Fixes: f55c73aef8 ("irqchip/pdc: Add PDC interrupt controller for QCOM SoCs")
Reported-by: Evan Green <evgreen@chromium.org>
Reviewed-by: Evan Green <evgreen@chromium.org>
Signed-off-by: Lina Iyer <ilina@codeaurora.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
If the LPI tables have been reserved with the EFI reservation
mechanism, we assume that these tables are safe to use even
when we find the redistributors to have LPIs enabled at
boot time, meaning that kexec can now work with GICv3.
You're welcome.
Tested-by: Jeremy Linton <jeremy.linton@arm.com>
Tested-by: Bhupesh Sharma <bhsharma@redhat.com>
Tested-by: Lei Zhang <zhang.lei@jp.fujitsu.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Upon enabling a redistributor, let's register the allocated tables
with the EFI table that tracks the memory reservations.
Tested-by: Jeremy Linton <jeremy.linton@arm.com>
Tested-by: Bhupesh Sharma <bhsharma@redhat.com>
Tested-by: Lei Zhang <zhang.lei@jp.fujitsu.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
If booting with LPIs enabled, all the redistributors must have the
exact same property table. No ifs, no buts.
Tested-by: Jeremy Linton <jeremy.linton@arm.com>
Tested-by: Bhupesh Sharma <bhsharma@redhat.com>
Tested-by: Lei Zhang <zhang.lei@jp.fujitsu.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
If using a kdump kernel, and that we cannot disable LPIs to install
our own tables, let's switch to using the already allocated tables.
This means that we'll change some of the initial kernel's memory,
but at least we'll be able to have LPIs in this secondary kernel.
Tested-by: Jeremy Linton <jeremy.linton@arm.com>
Tested-by: Bhupesh Sharma <bhsharma@redhat.com>
Tested-by: Lei Zhang <zhang.lei@jp.fujitsu.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
In order to cope with kexec and GICv3, let's try and spot when
we're booting with LPIs already enabled, and the tables already
programmed into the redistributors.
This code is currently guarded by a predicate that is always false,
meaning this is not functionnal just yet.
Reviewed-by: Julien Thierry <julien.thierry@arm.com>
Tested-by: Jeremy Linton <jeremy.linton@arm.com>
Tested-by: Bhupesh Sharma <bhsharma@redhat.com>
Tested-by: Lei Zhang <zhang.lei@jp.fujitsu.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
We're currently only tracking the page allocated to contain the
property table by its struct page. In the future, it is going to
be convenient to track both PA and VA for that page instead. Let's
do that.
Tested-by: Jeremy Linton <jeremy.linton@arm.com>
Tested-by: Bhupesh Sharma <bhsharma@redhat.com>
Tested-by: Lei Zhang <zhang.lei@jp.fujitsu.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Pending tables for the redistributors are currently allocated
one at a time as each CPU boots. This is causing some grief
for Linux/RT (allocation from within a CPU hotplug notifier is
frown upon).
Let's move this allocation to take place at init time, when we
only have a single CPU. It means we're allocating memory for CPUs
that are not online yet, but most system will boot all of their
CPUs anyway, so that's not completely wasted.
Tested-by: Jeremy Linton <jeremy.linton@arm.com>
Tested-by: Bhupesh Sharma <bhsharma@redhat.com>
Tested-by: Lei Zhang <zhang.lei@jp.fujitsu.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
As we're going to reuse some pre-allocated memory for the property
table, split out the zeroing of that table into a separate function
for later use.
Tested-by: Jeremy Linton <jeremy.linton@arm.com>
Tested-by: Bhupesh Sharma <bhsharma@redhat.com>
Tested-by: Lei Zhang <zhang.lei@jp.fujitsu.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
LPI_PENDING_SZ is always used in conjunction with a max(), which doesn't
make much sense, since we're guaranteed that LPI_PENDING_SZ is already
aligned to 64K. Let's remove it.
Tested-by: Jeremy Linton <jeremy.linton@arm.com>
Tested-by: Bhupesh Sharma <bhsharma@redhat.com>
Tested-by: Lei Zhang <zhang.lei@jp.fujitsu.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
We currently initialize the LPIs (and the ITS) fairly early, even
before the SMP support and the CPU interface. This is a bit odd
(as LPIs are not exactly crutial for the early boot process),
and is going to cause issues when reorganizing the probing code.
Let's move this initialization later.
Reviewed-by: Julien Thierry <julien.thierry@arm.com>
Tested-by: Jeremy Linton <jeremy.linton@arm.com>
Tested-by: Bhupesh Sharma <bhsharma@redhat.com>
Tested-by: Lei Zhang <zhang.lei@jp.fujitsu.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
* Fix IRQ constants usage on MSM8974
* Add led, gpio-button, sdcc, and pcie nodes for IPQ8064
* Move/cleanup common nodes for IPQ8064
* Add i2c sensor nodes for MSM8974 Hammerhead
* Fixes for SAW, kpss, opp, pci range, and space/tab on IPQ4019
* Update coresight bindings
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1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=74RU
-----END PGP SIGNATURE-----
Merge tag 'qcom-dts-for-4.20' of git://git.kernel.org/pub/scm/linux/kernel/git/agross/linux into next/dt
Qualcomm Device Tree Changes for v4.20
* Fix IRQ constants usage on MSM8974
* Add led, gpio-button, sdcc, and pcie nodes for IPQ8064
* Move/cleanup common nodes for IPQ8064
* Add i2c sensor nodes for MSM8974 Hammerhead
* Fixes for SAW, kpss, opp, pci range, and space/tab on IPQ4019
* Update coresight bindings
* tag 'qcom-dts-for-4.20' of git://git.kernel.org/pub/scm/linux/kernel/git/agross/linux:
ARM: dts: qcom: Update coresight bindings for hardware ports
ARM: dts: qcom-msm8974: change invalid flag IRQ NONE to valid value
ARM: dts: qcom: ipq4019: fix space vs tab indenting inside qcom-ipq4019.dtsi
ARM: dts: qcom: ipq4019: fix PCI range
ARM: dts: qcom: ipq4019: fix cpu0's qcom,saw2 reg value
ARM: dts: qcom: ipq4019: add cpu operating points for cpufreq support
ARM: dts: qcom: ipq4019: use v2 of the kpss bringup mechanism
ARM: dts: qcom: msm8974-hammerhead: add device tree bindings for ALS / proximity
ARM: dts: qcom: msm8974-hammerhead: add device tree bindings for mpu6515
ARM: dts: qcom: Add led and gpio-button nodes to ipq8064 boards
ARM: dts: qcom: Move common nodes to ipq8064-v.1.0.dtsi
ARM: dts: qcom: Add sdcc nodes for ipq8064
ARM: dts: qcom: Add pcie nodes for ipq8064
ARM: dts: qcom-msm8974: change invalid flag IRQ NONE to valid value
ARM: dts: qcom-msm8974: use named constant for interrupt flag NONE
ARM: dts: qcom-msm8974: use named constant for interrupt flag LEVEL HIGH
ARM: dts: qcom-msm8974: use named constant for interrupt flag EDGE RISING
ARM: dts: qcom-msm8974: use named constant for interrupt type GIC_SPI
ARM: dts: qcom-msm8974: use named constant for interrupt type GIC_PPI
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Automatic NUMA Balancing uses a multi-stage pass to decide whether a page
should migrate to a local node. This filter avoids excessive ping-ponging
if a page is shared or used by threads that migrate cross-node frequently.
Threads inherit both page tables and the preferred node ID from the
parent. This means that threads can trigger hinting faults earlier than
a new task which delays scanning for a number of seconds. As it can be
load balanced very early in its lifetime there can be an unnecessary delay
before it starts migrating thread-local data. This patch migrates private
pages faster early in the lifetime of a thread using the sequence counter
as an identifier of new tasks.
With this patch applied, STREAM performance is the same as 4.17 even though
processes are not spread cross-node prematurely. Other workloads showed
a mix of minor gains and losses. This is somewhat expected most workloads
are not very sensitive to the starting conditions of a process.
4.19.0-rc5 4.19.0-rc5 4.17.0
numab-v1r1 fastmigrate-v1r1 vanilla
MB/sec copy 43298.52 ( 0.00%) 47335.46 ( 9.32%) 47219.24 ( 9.06%)
MB/sec scale 30115.06 ( 0.00%) 32568.12 ( 8.15%) 32527.56 ( 8.01%)
MB/sec add 32825.12 ( 0.00%) 36078.94 ( 9.91%) 35928.02 ( 9.45%)
MB/sec triad 32549.52 ( 0.00%) 35935.94 ( 10.40%) 35969.88 ( 10.51%)
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Reviewed-by: Rik van Riel <riel@surriel.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Jirka Hladky <jhladky@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Linux-MM <linux-mm@kvack.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/20181001100525.29789-3-mgorman@techsingularity.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Rate limiting of page migrations due to automatic NUMA balancing was
introduced to mitigate the worst-case scenario of migrating at high
frequency due to false sharing or slowly ping-ponging between nodes.
Since then, a lot of effort was spent on correctly identifying these
pages and avoiding unnecessary migrations and the safety net may no longer
be required.
Jirka Hladky reported a regression in 4.17 due to a scheduler patch that
avoids spreading STREAM tasks wide prematurely. However, once the task
was properly placed, it delayed migrating the memory due to rate limiting.
Increasing the limit fixed the problem for him.
Currently, the limit is hard-coded and does not account for the real
capabilities of the hardware. Even if an estimate was attempted, it would
not properly account for the number of memory controllers and it could
not account for the amount of bandwidth used for normal accesses. Rather
than fudging, this patch simply eliminates the rate limiting.
However, Jirka reports that a STREAM configuration using multiple
processes achieved similar performance to 4.16. In local tests, this patch
improved performance of STREAM relative to the baseline but it is somewhat
machine-dependent. Most workloads show little or not performance difference
implying that there is not a heavily reliance on the throttling mechanism
and it is safe to remove.
STREAM on 2-socket machine
4.19.0-rc5 4.19.0-rc5
numab-v1r1 noratelimit-v1r1
MB/sec copy 43298.52 ( 0.00%) 44673.38 ( 3.18%)
MB/sec scale 30115.06 ( 0.00%) 31293.06 ( 3.91%)
MB/sec add 32825.12 ( 0.00%) 34883.62 ( 6.27%)
MB/sec triad 32549.52 ( 0.00%) 34906.60 ( 7.24%
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Reviewed-by: Rik van Riel <riel@surriel.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Jirka Hladky <jhladky@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Linux-MM <linux-mm@kvack.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/20181001100525.29789-2-mgorman@techsingularity.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Fix incorrect line number in example output
Signed-off-by: Andrew Murray <andrew.murray@arm.com>
Cc: Jiri Kosina <trivial@kernel.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Will Deacon <will.deacon@arm.com>
Cc: linux-doc@vger.kernel.org
Link: http://lkml.kernel.org/r/1538391663-54524-1-git-send-email-andrew.murray@arm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This code needs to select function #0, which is the first int in the
array of functions, not the number 0 which may or may not be the
function for "GPIO mode" per the enum mapping. We were getting lucky on
SDM845, where this was tested, because the function 0 matched the enum
value for "GPIO mode". On other platforms, e.g. MSM8996, the gpio enum
value is the last one in the list so this code doesn't work and we see a
warning at boot. Fix it by grabbing the first element out of the array
of functions.
Cc: Doug Anderson <dianders@chromium.org>
Cc: Bjorn Andersson <bjorn.andersson@linaro.org>
Cc: Niklas Cassel <niklas.cassel@linaro.org>
Reported-by: Niklas Cassel <niklas.cassel@linaro.org>
Fixes: 1de7ddb3a1 ("pinctrl: msm: Mux out gpio function with gpio_request()")
Signed-off-by: Stephen Boyd <swboyd@chromium.org>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Tested-by: Niklas Cassel <niklas.cassel@linaro.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
As DMC Package contain DMC FW for multiple steppings including default
stepping. This patch will help to load FW for that particular stepping,
if FW for that stepping is available, instead of loading default FW.
v2 : Fix formatting issue.
Signed-off-by: Jyoti Yadav <jyoti.r.yadav@intel.com>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1536169347-31326-1-git-send-email-jyoti.r.yadav@intel.com
Since 890658b7ab ("locking/mutex: Kill arch specific code"), there
are no mutex header files under arch/, so we can remove the redundant
entry from MAINTAINERS.
Reported-by: Joe Perches <joe@perches.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Cc: Jason Low <jason.low2@hpe.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/20181001142856.GC9716@arm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Our usual set of DT changes for the arm64 Allwinner SoCs.
The most notable things are:
- HDMI support on the A64
- New boards: OrangePi One Plus
-----BEGIN PGP SIGNATURE-----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=sPs9
-----END PGP SIGNATURE-----
Merge tag 'sunxi-dt64-for-4.20' of git://git.kernel.org/pub/scm/linux/kernel/git/sunxi/linux into next/dt
Allwinner arm64 DT changes for 4.20
Our usual set of DT changes for the arm64 Allwinner SoCs.
The most notable things are:
- HDMI support on the A64
- New boards: OrangePi One Plus
* tag 'sunxi-dt64-for-4.20' of git://git.kernel.org/pub/scm/linux/kernel/git/sunxi/linux: (28 commits)
arm64: dts: allwinner: a64: a64-olinuxino: set the PHY TX delay
arm64: dts: allwinner: a64: Enable HDMI output on A64 boards w/ HDMI
arm64: dts: allwinner: a64: Add display pipeline
arm64: dts: allwinner: h6: add system controller device tree node
arm64: dts: allwinner: h6: Add OrangePi One Plus initial support
arm64: dts: allwinner: a64: Rename r_i2c_pins_a label to r_i2c_pl89_pins
arm64: dts: allwinner: a64: Rename uart0_pins_a label to uart0_pb_pins
arm64: dts: allwinner: a64: Split out data strobe pin from mmc2 pinmux
arm64: dts: allwinner: a64: NanoPi-A64: Add blue status LED
arm64: dts: allwinner: a64: NanoPi-A64: Add Wifi chip
arm64: dts: allwinner: a64: NanoPi-A64: Add Ethernet
arm64: dts: allwinner: a64: NanoPi-A64: Fix DCDC1 voltage
arm64: dts: allwinner: a64: Olinuxino: enable USB
arm64: dts: allwinner: a64: Olinuxino: add Ethernet nodes
arm64: dts: allwinner: a64: Olinuxino: fix DRAM voltage
arm64: dts: allwinner: a64: Orange Pi Win: Adjust CSI power rails
arm64: dts: allwinner: a64: Orange Pi Win: Add SPI flash node
arm64: dts: allwinner: a64: Orange Pi Win: Add SDIO node
arm64: dts: allwinner: a64: Orange Pi Win: Add LED node
arm64: dts: allwinner: a64: Orange Pi Win: Add UARTs
...
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Amend the changes in commit:
1f03e8d291 ("locking/barriers: Replace smp_cond_acquire() with smp_cond_load_acquire()")
... by updating the documentation accordingly.
Also remove some obsolete information related to the implementation.
Signed-off-by: Andrea Parri <andrea.parri@amarulasolutions.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Acked-by: Will Deacon <will.deacon@arm.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Cc: Akira Yokosawa <akiyks@gmail.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Boqun Feng <boqun.feng@gmail.com>
Cc: Daniel Lustig <dlustig@nvidia.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Jade Alglave <j.alglave@ucl.ac.uk>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Luc Maranget <luc.maranget@inria.fr>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Cc: linux-arch@vger.kernel.org
Cc: parri.andrea@gmail.com
Link: http://lkml.kernel.org/r/20180926182920.27644-5-paulmck@linux.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
More than one kernel developer has expressed the opinion that the LKMM
should enforce ordering of writes by locking. In other words, given
the following code:
WRITE_ONCE(x, 1);
spin_unlock(&s):
spin_lock(&s);
WRITE_ONCE(y, 1);
the stores to x and y should be propagated in order to all other CPUs,
even though those other CPUs might not access the lock s. In terms of
the memory model, this means expanding the cumul-fence relation.
Locks should also provide read-read (and read-write) ordering in a
similar way. Given:
READ_ONCE(x);
spin_unlock(&s);
spin_lock(&s);
READ_ONCE(y); // or WRITE_ONCE(y, 1);
the load of x should be executed before the load of (or store to) y.
The LKMM already provides this ordering, but it provides it even in
the case where the two accesses are separated by a release/acquire
pair of fences rather than unlock/lock. This would prevent
architectures from using weakly ordered implementations of release and
acquire, which seems like an unnecessary restriction. The patch
therefore removes the ordering requirement from the LKMM for that
case.
There are several arguments both for and against this change. Let us
refer to these enhanced ordering properties by saying that the LKMM
would require locks to be RCtso (a bit of a misnomer, but analogous to
RCpc and RCsc) and it would require ordinary acquire/release only to
be RCpc. (Note: In the following, the phrase "all supported
architectures" is meant not to include RISC-V. Although RISC-V is
indeed supported by the kernel, the implementation is still somewhat
in a state of flux and therefore statements about it would be
premature.)
Pros:
The kernel already provides RCtso ordering for locks on all
supported architectures, even though this is not stated
explicitly anywhere. Therefore the LKMM should formalize it.
In theory, guaranteeing RCtso ordering would reduce the need
for additional barrier-like constructs meant to increase the
ordering strength of locks.
Will Deacon and Peter Zijlstra are strongly in favor of
formalizing the RCtso requirement. Linus Torvalds and Will
would like to go even further, requiring locks to have RCsc
behavior (ordering preceding writes against later reads), but
they recognize that this would incur a noticeable performance
degradation on the POWER architecture. Linus also points out
that people have made the mistake, in the past, of assuming
that locking has stronger ordering properties than is
currently guaranteed, and this change would reduce the
likelihood of such mistakes.
Not requiring ordinary acquire/release to be any stronger than
RCpc may prove advantageous for future architectures, allowing
them to implement smp_load_acquire() and smp_store_release()
with more efficient machine instructions than would be
possible if the operations had to be RCtso. Will and Linus
approve this rationale, hypothetical though it is at the
moment (it may end up affecting the RISC-V implementation).
The same argument may or may not apply to RMW-acquire/release;
see also the second Con entry below.
Linus feels that locks should be easy for people to use
without worrying about memory consistency issues, since they
are so pervasive in the kernel, whereas acquire/release is
much more of an "experts only" tool. Requiring locks to be
RCtso is a step in this direction.
Cons:
Andrea Parri and Luc Maranget think that locks should have the
same ordering properties as ordinary acquire/release (indeed,
Luc points out that the names "acquire" and "release" derive
from the usage of locks). Andrea points out that having
different ordering properties for different forms of acquires
and releases is not only unnecessary, it would also be
confusing and unmaintainable.
Locks are constructed from lower-level primitives, typically
RMW-acquire (for locking) and ordinary release (for unlock).
It is illogical to require stronger ordering properties from
the high-level operations than from the low-level operations
they comprise. Thus, this change would make
while (cmpxchg_acquire(&s, 0, 1) != 0)
cpu_relax();
an incorrect implementation of spin_lock(&s) as far as the
LKMM is concerned. In theory this weakness can be ameliorated
by changing the LKMM even further, requiring
RMW-acquire/release also to be RCtso (which it already is on
all supported architectures).
As far as I know, nobody has singled out any examples of code
in the kernel that actually relies on locks being RCtso.
(People mumble about RCU and the scheduler, but nobody has
pointed to any actual code. If there are any real cases,
their number is likely quite small.) If RCtso ordering is not
needed, why require it?
A handful of locking constructs (qspinlocks, qrwlocks, and
mcs_spinlocks) are built on top of smp_cond_load_acquire()
instead of an RMW-acquire instruction. It currently provides
only the ordinary acquire semantics, not the stronger ordering
this patch would require of locks. In theory this could be
ameliorated by requiring smp_cond_load_acquire() in
combination with ordinary release also to be RCtso (which is
currently true on all supported architectures).
On future weakly ordered architectures, people may be able to
implement locks in a non-RCtso fashion with significant
performance improvement. Meeting the RCtso requirement would
necessarily add run-time overhead.
Overall, the technical aspects of these arguments seem relatively
minor, and it appears mostly to boil down to a matter of opinion.
Since the opinions of senior kernel maintainers such as Linus,
Peter, and Will carry more weight than those of Luc and Andrea, this
patch changes the model in accordance with the maintainers' wishes.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Reviewed-by: Will Deacon <will.deacon@arm.com>
Reviewed-by: Andrea Parri <andrea.parri@amarulasolutions.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Cc: akiyks@gmail.com
Cc: boqun.feng@gmail.com
Cc: dhowells@redhat.com
Cc: j.alglave@ucl.ac.uk
Cc: linux-arch@vger.kernel.org
Cc: luc.maranget@inria.fr
Cc: npiggin@gmail.com
Cc: parri.andrea@gmail.com
Link: http://lkml.kernel.org/r/20180926182920.27644-2-paulmck@linux.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>