This patch removes the GLF_NOCACHE flag from the glocks associated with
flocks. There should be no good reason not to cache glocks for flocks:
they only force the glock to be demoted before they can be reacquired,
which can slow down performance and even cause glock hangs, especially
in cases where the flocks are held in Shared (SH) mode.
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
This patch allows flock glocks to use a non-blocking dequeue rather
than dq_wait. It also reverts the previous patch I had posted regarding
dq_wait. The reverted patch isn't necessarily a bad idea, but I decided
this might avoid unforeseen side effects, and was therefore safer.
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Normally GFP_KERNEL is ok here, but there is now a rarely used code path
relating to deallocation of unlinked inodes (in certain corner cases)
which if hit at times of memory shortage can cause recursion while
trying to free memory.
One solution would be to try and move the gfs2_glock_get() call so
that it is no longer called while another glock is held, but that
doesn't look at all easy, so GFP_NOFS is the best solution for the
time being.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
We must not leave items on the LRU list with GLF_LOCK set, since
they can be removed if the glock is brought back into use, which
may then potentially result in a hang, waiting for GLF_LOCK to
clear.
It doesn't happen very often, since it requires a glock that has
not been used for a long time to be brought back into use at the
same moment that the shrinker is part way through disposing of
glocks.
The fix is to set GLF_LOCK at a later time, when we already know
that the other locks can be obtained. Also, we now only release
the lru_lock in case a resched is needed, rather than on every
iteration.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Function gfs2_glock_dq_wait is supposed to dequeue a glock and then
wait for the lock to be demoted. The problem is, if this is a shared
lock, its demote will depend on the other holders, which means you
might end up waiting forever because the other process is blocked.
This problem is especially apparent when dealing with nested flocks.
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
The Zynq UART is Cadence IP and the driver has been renamed accordingly.
Migrate the DT to use the new binding for the UART driver.
Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com>
Acked-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Acked-by: Rob Herring <robh@kernel.org>
Tested-by: Michal Simek <michal.simek@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Add binding documentation for the Cadence UART.
Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com>
Acked-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Acked-by: Rob Herring <robh@kernel.org>
Tested-by: Michal Simek <michal.simek@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Add FlexCAN node for the two FlexCAN IP instances in Vybrid.
Signed-off-by: Stefan Agner <stefan@agner.ch>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
Add initial Rex Pro i.mx6dl board support. Ethernet, UART, USB, I2C, SPI, HDMI,
Audio, and SDHC cards are working. Currently no mainline u-boot, so boot with
cat zImage imx6dl-rex-basic.dtb > zImage.dtb, then using mkimage create uImage
Signed-off-by: Robert Nelson <robertcnelson@gmail.com>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
Add initial Rex Pro i.mx6q board support. Ethernet, UART, USB, I2C, SPI, HDMI,
Audio, and SDHC cards are working. Currently no mainline u-boot, so boot with
cat zImage imx6q-rex-pro.dtb > zImage.dtb, then using mkimage create uImage
Signed-off-by: Robert Nelson <robertcnelson@gmail.com>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
This patch splits the M53EVK device tree file into a common SoM
part and an EVK part. This is needed to make it easier for users
of the SoM to put it into different, non-reference baseboard.
Signed-off-by: Marek Vasut <marex@denx.de>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
Instead of relying on defaults or bootloader settings, explicitly define
all pad settings.
This resolves reported issues of no analogue audio output.
Signed-off-by: Iain Paton <ipaton0@gmail.com>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
The rev C1 Wandboard uses the Broadcom 4330 for WiFi and Bluetooth instead of
the 4329. This changes the PADS assigned for the control lines. Another
side effect of the change is that on the rev C1 board, usdhc driver can't
detect the chip presence correctly so usdhc2 now needs its 'non-removeable'
property removed.
So that rev B1 and earlier can continue to work, this patch splits the
board-specific definitions from imx6qdl-wandboard.dtsi into
imx6qdl-wandboard-revb1.dtsi and imx6qdl-wandboard-revc1.dtsi. The new files
include the original base imx6qdl-wandboard.dtsi which retains the common
definitions.
The existing imx6dl-wandboard.dts includes imx6qdl-wandboard-revc1.dtsi and
imx6dl-wandboard-revb1.dts (new) includes imx6qdl-wandboard-revb1.dtsi.
This makes the rev C1 board the new default. The same pattern is used for
imx6q-wandboard.dts.
So, from U-Boot on a WB-Quad you use imxq-wandboard-revb1.dtb for the older B1
board and imxq-wandboard.dtb for the current rev C1 board.
Signed-off-by: George Joseph <george.joseph@fairview5.com>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
Per the binding doc imx-sata.txt, the first entry of clock-names should
be "sata" than anything else. Correct it for imx53 SATA node.
It works for now only because SATA driver gets clock by index so far.
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
Since commit 98ea6ad2ed (ARM: dts: imx6: use imx51-ssi) the mx6 ssi is
compatible with imx51, so align all the mx6 variant ssi compatible strings as:
compatible = "fsl,<imx6-soc>-ssi", "fsl,imx51-ssi";
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
Use clock definitions in i.MX27 DTS files.
Additional changes included in this patch (imx27.dtsi):
- Fix IPG clock for UART6.
- Use EMI_AHB_GATE clock for WEIM.
- Added GPIO_IPG_GATE clock for GPIO nodes. Currently this clock is
not used by the driver, but it can be added in the future.
Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
The SDMA on imx6sl and imx6sx is more compatible with imx6q one than
imx35. Let's use "fsl,imx6q-sdma" instead of "fsl,imx35-sdma", so that
SDMA ROM script on imx6sl and imx6sx can work for audio driver just like
the case of imx6q.
Reported-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
Acked-by: Robin Gong <yibin.gong@freescale.com>
Like the other mx6 variants, we need to pass fsl,fifo-depth property in dtsi.
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
Use the correct compatible string for sdma and also provide the sdma firmware
path.
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
Add support for the PCI express bus available on MX6 Data Modul
edm-qmx6 board.
Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
Tested by loading the g_ether module and doing a ping between mx25pdk and the
host PC via USB.
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
The AHB to IP bridges (AIPSTZ) allow fine grained access rights management.
Add both bridges to the DT.
Signed-off-by: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
This patch splits the M28EVK device tree file into a common SoM
part and an EVK part. This is needed to make it easier for users
of the SoM to put it into different, non-reference baseboard.
Signed-off-by: Marek Vasut <marex@denx.de>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
Without that fix, the board freeze during boot.
This appeared after the following commit:
496f065 ARM: i.MX: Use of_clk_get_by_name() for timer clocks for DT case.
Signed-off-by: Denis Carikli <denis@eukrea.com>
Acked-by: Sascha Hauer <s.hauer@pengutronix.de>
Tested-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
When booting a board that does not have a keypad (such as imx6q-sabresd) the
following error is seen on boot:
imx-keypad 20b8000.kpp: OF: linux,keymap property not defined in /soc/aips-bus@02000000/kpp@020b8000
imx-keypad 20b8000.kpp: failed to build keymap
imx-keypad: probe of 20b8000.kpp failed with error -2
Let's disable the keypad functionality in the dtsi files and let each board dts
enable it when needed.
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
Spread-spectrum doesn't work with Cubox-i hardware. eSATA devices are
detected, but then fail on normal IO. Therefore, disable this feature.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
Add the transmit level, boost and attenuation parameters necessary for
the eSATA interface on Cubox-i to work.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
Fix usbmisc compatible string so that usbmisc node can be found and USB
functionality can be functional.
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
The imx6sx iomuxc-gpr syscon is compatible to imx6q, so let's add
compatible string 'fsl,imx6q-iomuxc-gpr' for imx6sx iomuxc-gpr syscon node.
This is necessary to enable SW workaround for ERR007265,
please refer to imx6_pm_common_init of arch/arm/mach-imx/pm-imx6.c
for detail.
Signed-off-by: Anson Huang <b20788@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
This patch add support for the imx6dl based aristainetos board
with following configuration:
CPU: Freescale i.MX6DL rev1.1 at 792 MHz
DRAM: 1 GiB
NAND: 512 MiB
MMC: FSL_SDHC: 0, FSL_SDHC: 1
SF: Detected N25Q128 with page size 256 Bytes, erase size 64 KiB, total 16 MiB
As this board can used with 2 different display types, the
differences between them are extracted into 2 DTS files, and
the common settings are collected in a common file.
Signed-off-by: Heiko Schocher <hs@denx.de>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
The fixed-regulator's enable-active-high property
is needed to indicate that the GPIO regulator is
active high.
Before that the regulator state was inverted.
Signed-off-by: Denis Carikli <denis@eukrea.com>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
The DAI mode is and should be configured by the sound card driver as
codec and ssi have to be in the right modes to communicate with each
other. It is possible to operate the ssi unit or the codec in master mode,
sometimes even on the same board in different configurations.
With the latest changes in the fsl-ssi driver, the 'fsl,mode' property
is only handled as a fallback property. If the sound card sets the DAI
mode correctly, this fallback configuration is dropped.
Signed-off-by: Markus Pargmann <mpa@pengutronix.de>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
All dtsi files where already moved to generic dma bindings. Remove the
old non-generic DMA properties.
Signed-off-by: Markus Pargmann <mpa@pengutronix.de>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
All imx5*.dtsi files define the generic dma bindings. Drop the old
non-generic fsl,ssi-dma-events.
Signed-off-by: Markus Pargmann <mpa@pengutronix.de>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
Replace old ssi dma properties by the generic dma properties.
Signed-off-by: Markus Pargmann <mpa@pengutronix.de>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
The other i.mx6 variants use "vddarm" as the name of the internal regulator
that powers the core. Use the same convention here for consistency.
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
The S/PDIF rxtx4 and rxtx6 clock inputs are "ESAI_HCKT" and "MLB clock",
respectively, according to the SoC documentation, and they are currently
mapped to clocks "esai" and "mlb".
However, they do not seem to actually work correctly. Testing on a
Cubox-i system with fsl_spdif driver forced to select one of those as
input will result in I/O errors on audio playback, which I believe means
missing clock signal.
Possibly the "ESAI_HCKT" and "MLB clock" refer to some other clocks
related to ESAI and MLB, or we are missing something else.
Since audio playback will not work if fsl_spdif selects these clocks
(which happens rarely), set the inputs do dummy clocks, at least for
now.
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
Cc: Mark Brown <broonie@kernel.org>
Acked-by: Nicolin Chen <Guangyu.Chen@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
The rxtx2 clock of i.MX6 S/PDIF is currently set to "asrc" clock.
However, according to SoC documentation, rxtx2 is connected to
ASRC_EXT_CLK, a different external clock.
Testing on Cubox-i system seems to confirm that: when fsl_spdif is
forced to select rxtx2 as input clock, audio playback fails with an I/O
error.
Set rxtx2 to the dummy clock by default to prevent fsl_spdif from
selecting it.
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
Cc: Mark Brown <broonie@kernel.org>
Acked-by: Nicolin Chen <Guangyu.Chen@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
This patch adds support for the Ka-Ro electronics GmbH TX6 modules.
There are five distinct module types with either i.MX6Q or i.MX6DL and
LVDS or LCD display interface and one DTS file for a complete system
with an i.MX6DL based TX6 module and a baseboard mounted on the back
of a display (imx6dl-tx6dl-comtft.dts).
Signed-off-by: Lothar Waßmann <LW@KARO-electronics.de>
Signed-off-by: Shawn Guo <shawn.guo@freescale.com>
