linux.git
7 years agoMLK-15001-4 arm64: dts: fsl-imx8qm-lpddr4-arm2: Remove it6263 support
Liu Ying [Mon, 5 Jun 2017 03:21:01 +0000 (11:21 +0800)]
MLK-15001-4 arm64: dts: fsl-imx8qm-lpddr4-arm2: Remove it6263 support

No driver works with the it6263 nodes, so let's remove them.

Signed-off-by: Liu Ying <victor.liu@nxp.com>
7 years agoMLK-15001-3 arm64: fsl-imx8qm.dtsi: Remove lvds nodes
Liu Ying [Fri, 2 Jun 2017 08:27:14 +0000 (16:27 +0800)]
MLK-15001-3 arm64: fsl-imx8qm.dtsi: Remove lvds nodes

No driver works with the lvds nodes, so let's remove them.

Signed-off-by: Liu Ying <victor.liu@nxp.com>
7 years agoMLK-15001-2 arm64: fsl-imx8qm.dtsi: Remove imxdpu nodes
Liu Ying [Fri, 2 Jun 2017 08:24:34 +0000 (16:24 +0800)]
MLK-15001-2 arm64: fsl-imx8qm.dtsi: Remove imxdpu nodes

No driver works with the imxdpu nodes, so let's remove them.

Signed-off-by: Liu Ying <victor.liu@nxp.com>
7 years agoMLK-15001-1 arm64: fsl-imx8qm.dtsi: Remove framebuffer nodes
Liu Ying [Fri, 2 Jun 2017 08:18:27 +0000 (16:18 +0800)]
MLK-15001-1 arm64: fsl-imx8qm.dtsi: Remove framebuffer nodes

No driver works with the framebuffer nodes, so let's remove them.

Signed-off-by: Liu Ying <victor.liu@nxp.com>
7 years agoMLK-15080 PCI: imx: pcie ep can't be probed properly
Richard Zhu [Thu, 15 Jun 2017 03:41:34 +0000 (11:41 +0800)]
MLK-15080 PCI: imx: pcie ep can't be probed properly

iMX7D Sabre SD board implement the GPIO expander
connected to a peripheral bus.
Probe deferral would be triggered when try to request
the expanded GPIO at the first time.
pcie ep can't be probed properly at the second probe,
because of the duplicated registration of the sysfs.

Change the registeration point of the sysfs to fix
this issue.

Signed-off-by: Richard Zhu <hongxing.zhu@nxp.com>
7 years agoMLK-15076: arm: dts: insulate ULP kernel boot from M4 boot
Han Xu [Thu, 15 Jun 2017 20:39:50 +0000 (15:39 -0500)]
MLK-15076: arm: dts: insulate ULP kernel boot from M4 boot

i.MX7ULP QSPI dtb was used to update the M4 images, it should not able
to boot the kernel even without the M4 image in QSPI.

Also fixed the typo in dtsi to correct the QSPI register address
mapping range.

Signed-off-by: Han Xu <han.xu@nxp.com>
7 years agoMLK-15098: ASoC: fsl: fix imx-pcm build error when no dma
Adrian Alonso [Thu, 15 Jun 2017 17:12:49 +0000 (12:12 -0500)]
MLK-15098: ASoC: fsl: fix imx-pcm build error when no dma

Fix build error when SOC_IMX_PCM_DMA is not enabled
error: expected identifier or ‘(’ before ‘{’ token})

Signed-off-by: Adrian Alonso <adrian.alonso@nxp.com>
7 years agoMLK-15053: ASoC: fsl: imx-cs42888: Reject unsupported sampling rates
Mihai Serban [Wed, 14 Jun 2017 12:40:32 +0000 (15:40 +0300)]
MLK-15053: ASoC: fsl: imx-cs42888: Reject unsupported sampling rates

Dynamic constraints for supported sampling rates cannot prevent aplay to
play audio files with higher rates. So we remove the constraints and hard
reject the unsupported samples.

Signed-off-by: Mihai Serban <mihai.serban@nxp.com>
Reviewed-by: Shengjiu Wang <shengjiu.wang@nxp.com>
7 years agoMLK-15096 arm64: dts: imx8qm/imx8qxp: enable enet2 port
Fugang Duan [Thu, 15 Jun 2017 09:33:46 +0000 (17:33 +0800)]
MLK-15096 arm64: dts: imx8qm/imx8qxp: enable enet2 port

Since the enet clock issue is fixed at SCU code by commit 07dcb2d71dd1
("Fix bug clock enable code. Ensure that the appropriate parent PLL is
enabled/disabled.") in scu mainline branch, then enable enet2 port for
imx8qm and imx8qxp lpddr4 arm2 boards.

Signed-off-by: Fugang Duan <fugang.duan@nxp.com>
7 years agoMLK-15095-02 ARM64: defconfig: add Bluetooth support
Fugang Duan [Thu, 15 Jun 2017 10:04:09 +0000 (18:04 +0800)]
MLK-15095-02 ARM64: defconfig: add Bluetooth support

Enable Bluetooth stack, driver and HCI interface config for
Bluetooth support.

Run "make savedefconfig" to change the defconfig.

Signed-off-by: Fugang Duan <fugang.duan@nxp.com>
7 years agoMLK-15095-01 arm64: dts: imx8qm: add modem reset for Murata 1FD BT on lpddr4 arm2...
Fugang Duan [Thu, 15 Jun 2017 09:04:15 +0000 (17:04 +0800)]
MLK-15095-01 arm64: dts: imx8qm: add modem reset for Murata 1FD BT on lpddr4 arm2 board

i.MX8QM lpddr4-arm2 support Murata 1FD BT module that need some reset
timing. Add modem reset to support the BT device.

Signed-off-by: Fugang Duan <fugang.duan@nxp.com>
7 years agoMLK-15094 tty: serial: fsl_lpuart: check dma_tx_in_progress in callback
Fugang Duan [Thu, 15 Jun 2017 08:48:31 +0000 (16:48 +0800)]
MLK-15094 tty: serial: fsl_lpuart: check dma_tx_in_progress in callback

There have a corner case that tx DMA .callback() is comming after
.flush_buffer(), then .callback() should check dma_tx_in_progress
flag and return in directly.

Signed-off-by: Fugang Duan <fugang.duan@nxp.com>
7 years agoMLK-15093 tty: serial: imx: enable bit TDMAEN in each DMA transfer
Fugang Duan [Mon, 12 Jun 2017 08:56:43 +0000 (16:56 +0800)]
MLK-15093 tty: serial: imx: enable bit TDMAEN in each DMA transfer

In below case:
write() -> flush() -> write() -> flush() ...

.imx_flush_buffer() _MAY_ clear UCR1_TDMAEN bit if the callback is not
comming or DMA transfer is not completed, to ensure DMA trigger is enabled
for the new DMA prep_sg, enable the UCR1_TDMAEN bit in .dma_tx_work().

Signed-off-by: Fugang Duan <fugang.duan@nxp.com>
Tested-by: Fabio Estevam <fabio.estevam@nxp.com>
Tested-by: David Wolfe <david.wolfe@nxp.com>
7 years agoMLK-15092 reset: gpio-reset: add post reset delay
Fugang Duan [Thu, 15 Jun 2017 08:45:29 +0000 (16:45 +0800)]
MLK-15092 reset: gpio-reset: add post reset delay

Some devices need to wait for some milliseconds after reset, so add
post reset delay in the gpio-reset chip.

The post reset delay is optional.

Signed-off-by: Fugang Duan <fugang.duan@nxp.com>
7 years agoMLK-15042: ASoC: fsl_asrc: update supported format
Shengjiu Wang [Thu, 15 Jun 2017 07:32:17 +0000 (15:32 +0800)]
MLK-15042: ASoC: fsl_asrc: update supported format

The ASRC support 24 bit input width, but for S20_3LE the input width
is 20 bit, asrc will treat it as 24bit, which like a 24bit data shift
4 bit right, the result is the volume is lower than expected.
ASRC can't shift the 20bit data left 4 bit internally, so remove the
S20_3LE in supported list, add S24_3LE in supported list.

Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
(cherry picked from commit e65a014efe0a69c2db9de422fb780f0ba9c41dba)

7 years agoMLK-15063: ASoC: fsl: add dpcm_merged_chan for machine driver
Shengjiu Wang [Thu, 15 Jun 2017 05:19:34 +0000 (13:19 +0800)]
MLK-15063: ASoC: fsl: add dpcm_merged_chan for machine driver

Same as commit cfe36e2e7fce ("MLK-15043-2: ASoC: imx-cs42888: fix
noise issue with FE-BE case"). need to add same configuration
for imx-wm8960, imx-wm8962, imx-mqs.

Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
(cherry picked from commit 7195ad8ff56857028a6990296cc06e1c08a4e48e)

7 years agoMLK-15068: ASoC: fsl_ssi: fix the noise issue with S20_3LE Mono bitsream
Shengjiu Wang [Thu, 15 Jun 2017 05:07:20 +0000 (13:07 +0800)]
MLK-15068: ASoC: fsl_ssi: fix the noise issue with S20_3LE Mono bitsream

In master mode, clock of S20_3LE mono bistream is calculated by formula
"2 * params_width * params_rate", and this clock can't be divided from
clock soure, so switch to use the "2 * params_physical_width * params_rate"
formula to fix this issue.

Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
(cherry picked from commit 52a91e870b8cc95cbceeda205761134e25626863)

7 years agoMLK-15081 usb: gadget: utp: fix build warning with aarch64 compiler
Li Jun [Wed, 14 Jun 2017 15:16:30 +0000 (23:16 +0800)]
MLK-15081 usb: gadget: utp: fix build warning with aarch64 compiler

Compiler will generate below complain if using %d to print size_t, fix it
by using %zd.

drivers/usb/gadget/function/fsl_updater.c: In function ‘utp_do_read’:
./include/linux/kern_levels.h:4:18: warning: format ‘%d’ expects argument of type ‘int’, but argument 4 has type ‘size_t {aka long unsigned int}’ [-Wformat=]
 #define KERN_SOH "\001"  /* ASCII Start Of Header */
                  ^
./include/linux/kern_levels.h:13:19: note: in expansion of macro ‘KERN_SOH’
 #define KERN_INFO KERN_SOH "6" /* informational */
                   ^~~~~~~~
./include/linux/printk.h:284:9: note: in expansion of macro ‘KERN_INFO’
  printk(KERN_INFO pr_fmt(fmt), ##__VA_ARGS__)
         ^~~~~~~~~
drivers/usb/gadget/function/fsl_updater.c:245:3: note: in expansion of macro ‘pr_info’
   pr_info("Copied to %p, %d bytes started from %d\n",
   ^~~~~~~

Acked-by: Peter Chen <peter.chen@nxp.com>
Signed-off-by: Li Jun <jun.li@nxp.com>
7 years agoMLK-15074-7 ARM64: remove mfg_defconfig
Li Jun [Wed, 14 Jun 2017 14:10:19 +0000 (22:10 +0800)]
MLK-15074-7 ARM64: remove mfg_defconfig

As we can build in usb gadget function of utp for MFG tool, so we can
remove the dedicated menuconfig.

Signed-off-by: Li Jun <jun.li@nxp.com>
7 years agoMLK-15074-6 ARM64: defconfig: add USB gadget functions
Li Jun [Thu, 15 Jun 2017 15:18:50 +0000 (23:18 +0800)]
MLK-15074-6 ARM64: defconfig: add USB gadget functions

Build in CONFIG_USB_CONFIGFS, and add legency gadget function drivers
as module.

Acked-by: Peter Chen <peter.chen@nxp.com>
Signed-off-by: Li Jun <jun.li@nxp.com>
7 years agoMLK-17054-5 ARM64: defconfig: match the output of 'make savedefconfig'
Li Jun [Thu, 15 Jun 2017 15:14:43 +0000 (23:14 +0800)]
MLK-17054-5 ARM64: defconfig: match the output of 'make savedefconfig'

Update defconfig to match the output of make savedefconfig.

Acked-by: Peter Chen <peter.chen@nxp.com>
Signed-off-by: Li Jun <jun.li@nxp.com>
7 years agoMLK-15074-4 ARM: remove imx_v7_mfg_defconfig
Li Jun [Tue, 13 Jun 2017 14:22:19 +0000 (22:22 +0800)]
MLK-15074-4 ARM: remove imx_v7_mfg_defconfig

As we already build in USB mass storage for UTP via configfs
in kernel, we can use imx_v7_defconfig for MFG tool now, so
remove it.

Reviewed-by: Frank Li <frank.li@nxp.com>
Signed-off-by: Li Jun <jun.li@nxp.com>
7 years agoMLK-15074-3 ARM: imx_v7_defconfig: build in USB_CONFIGFS
Li Jun [Tue, 13 Jun 2017 13:57:14 +0000 (21:57 +0800)]
MLK-15074-3 ARM: imx_v7_defconfig: build in USB_CONFIGFS

Build in CONFIG_USB_CONFIGFS for MFG tool, with that, MFG tool can directly
use imx_v7_defconfig, and we can remove imx_v7_mfg_defconfig.

Acked-by: Peter Chen <peter.chen@nxp.com>
Signed-off-by: Li Jun <jun.li@nxp.com>
7 years agoMLK-15074-2 usb: gadget: utp: move the Kconfig out of legacy dir
Li Jun [Tue, 13 Jun 2017 13:49:44 +0000 (21:49 +0800)]
MLK-15074-2 usb: gadget: utp: move the Kconfig out of legacy dir

As the UTP is based on mass storage function, instead of legacy
g_mass_storage, so move it out of legacy folder.

Acked-by: Peter Chen <peter.chen@nxp.com>
Signed-off-by: Li Jun <jun.li@nxp.com>
7 years agoMLK-15074-1 usb: gadget: f_mass_storage: add utp device check
Li Jun [Tue, 13 Jun 2017 20:06:38 +0000 (04:06 +0800)]
MLK-15074-1 usb: gadget: f_mass_storage: add utp device check

Add utp device check for utp handling in f_mass_storage.

Acked-by: Peter Chen <peter.chen@nxp.com>
Signed-off-by: Li Jun <jun.li@nxp.com>
7 years agoMLK-15061-2: ASoC: fsl_asrc_m2m: fix asrc m2m can't work in imx8
Shengjiu Wang [Mon, 12 Jun 2017 07:19:42 +0000 (15:19 +0800)]
MLK-15061-2: ASoC: fsl_asrc_m2m: fix asrc m2m can't work in imx8

Power domain need to be enabled when asrc m2m start to work, and
disabled when it stop.
Switch back to use the pm_runtime_get_sync and pm_runtime_put_sync
for which is removed in commit 1a3d82e08fa2 ("ASoC: fsl: refine
the asrc driver for imx8qm").

Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
7 years agoMLK-15061-1: ASoC: fsl_asrc_m2m: fix dma_map_sg issue for arm64
Shengjiu Wang [Thu, 15 Jun 2017 07:50:50 +0000 (15:50 +0800)]
MLK-15061-1: ASoC: fsl_asrc_m2m: fix dma_map_sg issue for arm64

For arm64, the dma_map_sg and dma_unmap_sg need the device parameter,
otherwise, it will return error.

Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
7 years agoMLK-15043-2: ASoC: imx-cs42888: fix noise issue with FE-BE case
Shengjiu Wang [Wed, 14 Jun 2017 09:40:17 +0000 (17:40 +0800)]
MLK-15043-2: ASoC: imx-cs42888: fix noise issue with FE-BE case

The case is "aplay -Dhw:0,1 -d 5 -r 8000 -f S16_LE -c 9 audio8k16b9c.wav",
which is to playback 9 channel bitstream. But the maximum supported channel
of codec is 8, ALSA didn't return error for this case, but continue to
playback.

The reason is that in FE-BE case, ASLA only get the FE's hw parameter for
constraint list, omit the BE's parameter. This patch is to merge BE's
parameter to FE. in this situation with the 9 channel case, ASLA will
return error "aplay: set_params:1303: Channels count non available"

Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
(cherry picked from commit cfe36e2e7fce578eeac9a36eac04db2584e6e92e)

7 years agoMLK-15043-1: ASoC: soc-pcm: add dpcm_merged_chan in snd_soc_dai_link
Shengjiu Wang [Wed, 14 Jun 2017 09:35:20 +0000 (17:35 +0800)]
MLK-15043-1: ASoC: soc-pcm: add dpcm_merged_chan in snd_soc_dai_link

According to commit b073ed4e2126 ("ASoC: soc-pcm: DPCM cares BE format"),
Current DPCM only care FE channel, but it will set unsupported channel to
drivers.
So add dpcm_merged_chan, which is used to merge the BE's codec
channels configuration to FE if it exist in snd_soc_dai_link. And
dpcm_runtime_base_chan function is to get the channel configuration of BE,
which likes the dpcm_runtime_base_format function.

Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
(cherry picked from commit 19ff1095d4b68d17a5f68a2de5654be72b606809)

7 years agoMLK-15010 mmc: host: sdhci-esdhc-imx: add HS400_ES support for imx8
Haibo Chen [Tue, 13 Jun 2017 08:34:22 +0000 (16:34 +0800)]
MLK-15010 mmc: host: sdhci-esdhc-imx: add HS400_ES support for imx8

i.MX8QXP and i.MX8QM support Enhanced Strobe HS400 mode. This patch
add HS400_ES mode support, due to HS400_ES do not need tuning, select
HS400_ES mode should be faster than select HS400/HS200 mode.

Signed-off-by: Haibo Chen <haibo.chen@nxp.com>
7 years agoMLK-15047-2: ARM: imx: add IMX_SOC_IMX7
Octavian Purdila [Mon, 12 Jun 2017 08:58:44 +0000 (11:58 +0300)]
MLK-15047-2: ARM: imx: add IMX_SOC_IMX7

IMX_SOC_IMX7 is referenced in makefiles and kconfig but it is not
defined, so define it and select it for both IMX7D and IMX7ULP.

Fixes the following build errors:

arch/arm/mach-imx/built-in.o: In function `update_lpddr2_freq_smp':
platform-imx-dma.c:(.text+0xf7c): undefined reference to `imx_scu_base'
platform-imx-dma.c:(text+0xf88): undefined reference to `imx_scu_base'
arch/arm/mach-imx/built-in.o: In function `update_ddr_freq_imx_smp':
platform-imx-dma.c:(text+0x330c): undefined reference to `imx_scu_base'
platform-imx-dma.c:(text+0x3318): undefined reference to `imx_scu_base'
Makefile:969: recipe for target 'vmlinux' failed
make: *** [vmlinux] Error 1

Signed-off-by: Octavian Purdila <octavian.purdila@nxp.com>
7 years agoMLK-15047-1 media: platform: mxc v4l2 capture: add MXC_MIPI_CSI2 dependency
Octavian Purdila [Mon, 12 Jun 2017 08:55:39 +0000 (11:55 +0300)]
MLK-15047-1 media: platform: mxc v4l2 capture: add MXC_MIPI_CSI2 dependency

Fixes the following build errors:

ERROR: "mipi_csi2_reset" [drivers/media/platform/mxc/capture/ov5640_camera_mipi_int.ko] undefined!
ERROR: "mipi_csi2_set_datatype" [drivers/media/platform/mxc/capture/ov5640_camera_mipi_int.ko] undefined!
ERROR: "mipi_csi2_enable" [drivers/media/platform/mxc/capture/ov5640_camera_mipi_int.ko] undefined!
ERROR: "mipi_csi2_dphy_status" [drivers/media/platform/mxc/capture/ov5640_camera_mipi_int.ko] undefined!
ERROR: "mipi_csi2_disable" [drivers/media/platform/mxc/capture/ov5640_camera_mipi_int.ko] undefined!
ERROR: "mipi_csi2_set_lanes" [drivers/media/platform/mxc/capture/ov5640_camera_mipi_int.ko] undefined!
ERROR: "mipi_csi2_get_status" [drivers/media/platform/mxc/capture/ov5640_camera_mipi_int.ko] undefined!
ERROR: "mipi_csi2_get_info" [drivers/media/platform/mxc/capture/ov5640_camera_mipi_int.ko] undefined!
ERROR: "mipi_csi2_get_error1" [drivers/media/platform/mxc/capture/ov5640_camera_mipi_int.ko] undefined!

Signed-off-by: Octavian Purdila <octavian.purdila@nxp.com>
7 years agoMLK-15030: EPDC: Filled stride value of s0_param layer
Guoniu.Zhou [Sat, 10 Jun 2017 07:51:59 +0000 (15:51 +0800)]
MLK-15030: EPDC: Filled stride value of s0_param layer

When epdc driver use pxp, it didn't fill stride value of s0_parm, so it leads to
epdc can display picture normally.

So fill it in this patch.

Signed-off-by: Guoniu.Zhou <guoniu.zhou@nxp.com>
7 years agoMLK-15067: ASoC: fsl: imx-wm8958: Don't fail if gpr is missing
Daniel Baluta [Mon, 12 Jun 2017 14:21:26 +0000 (17:21 +0300)]
MLK-15067: ASoC: fsl: imx-wm8958: Don't fail if gpr is missing

There are boards without gpr setting, so it's better not to fail
in such cases and only print a warning.

This is related to commit ce72b6d2668049 ("MLK-13574-1: ASoC: imx-wm8960:
remove the gpr dependency").

Signed-off-by: Daniel Baluta <daniel.baluta@nxp.com>
7 years agoMLK-15056 arm: imx: use static mapping for imx7ulp AIPSx memory space
Bai Ping [Fri, 9 Jun 2017 10:10:35 +0000 (18:10 +0800)]
MLK-15056 arm: imx: use static mapping for imx7ulp AIPSx memory space

The AIPSx address space of i.MX7ULP need to be mapped as SZ_1M block
in iRAM tlb for suspend code use. If we use ioremap to map these
address region into kernel space, we can't make sure that the returned
virtual address is 1M alignment. So we can map this address regions
as static, then if we use the ioremap to map these memory regions, it will
always return the virtual address of static mapping. So we can make sure
the virtual address is 1M aligned.

Signed-off-by: Bai Ping <ping.bai@nxp.com>
7 years agoMLK-15031-08 tty: serial: fsl_lpuart: optimize the port.lock cover range
Fugang Duan [Tue, 6 Jun 2017 10:50:04 +0000 (18:50 +0800)]
MLK-15031-08 tty: serial: fsl_lpuart: optimize the port.lock cover range

In DMA tx callback, .uart_circ_chars_pending() also should be protected
by the port.lock, and in DMA rx callback, it can remove the lock protection
for .lpuart_copy_rx_to_tty() and .tty_flip_buffer_push() code range.

Signed-off-by: Fugang Duan <fugang.duan@nxp.com>
Reviewed-by: Robin Gong <yibin.gong@nxp.com>
Reviewed-by: Dong Aisheng <aisheng.dong@nxp.com>
7 years agoMLK-15031-07 tty: serial: fsl_lpuart: add port.icount for DMA rx path
Fugang Duan [Tue, 6 Jun 2017 10:44:16 +0000 (18:44 +0800)]
MLK-15031-07 tty: serial: fsl_lpuart: add port.icount for DMA rx path

Add port.icount to stat. the DMA RX received count.

Signed-off-by: Fugang Duan <fugang.duan@nxp.com>
Reviewed-by: Robin Gong <yibin.gong@nxp.com>
7 years agoMLK-15031-06 tty: serial: fsl_lpuart: re-calcaulate the minimum DMA transfer length
Fugang Duan [Tue, 6 Jun 2017 08:08:41 +0000 (16:08 +0800)]
MLK-15031-06 tty: serial: fsl_lpuart: re-calcaulate the minimum DMA transfer length

For lpuart32 loopback mode, DMA transfer length should be litter
than the tx/rx fifo free size to ensure there have no overrun.

Signed-off-by: Fugang Duan <fugang.duan@nxp.com>
Reviewed-by: Robin Gong <yibin.gong@nxp.com>
7 years agoMLK-15031-05 tty: serial: fsl_lpuart: enable dma eeop support
Fugang Duan [Fri, 26 May 2017 08:35:23 +0000 (16:35 +0800)]
MLK-15031-05 tty: serial: fsl_lpuart: enable dma eeop support

imx8qm lpuart support dma eeop:
- rx fifo > watermark level: watermark level RDRF triggler
- rx fifo <= watermark level: idle trigger
- rx fifo empty for some idle timer, eeop stop current DMA major loop

Signed-off-by: Fugang Duan <fugang.duan@nxp.com>
Reviewed-by: Robin Gong <yibin.gong@nxp.com>
7 years agoMLK-15031-04 tty: serial: fsl_lpuart: wait baud rate stable
Fugang Duan [Tue, 6 Jun 2017 07:14:42 +0000 (15:14 +0800)]
MLK-15031-04 tty: serial: fsl_lpuart: wait baud rate stable

After set the new baud rate, wait some time for the stable.

Signed-off-by: Fugang Duan <fugang.duan@nxp.com>
Reviewed-by: Robin Gong <yibin.gong@nxp.com>
7 years agoMLK-15031-03 tty: serial: fsl_lpuart: fix the typo in clock get failed path
Fugang Duan [Tue, 6 Jun 2017 05:57:13 +0000 (13:57 +0800)]
MLK-15031-03 tty: serial: fsl_lpuart: fix the typo in clock get failed path

Fix the typo in ipg clock get failed path.

Signed-off-by: Fugang Duan <fugang.duan@nxp.com>
Reviewed-by: Robin Gong <yibin.gong@nxp.com>
7 years agoMLK-15031-02 tty: serial: fsl_lpuart: free the rx dma buffer when port is closed
Fugang Duan [Tue, 6 Jun 2017 06:11:24 +0000 (14:11 +0800)]
MLK-15031-02 tty: serial: fsl_lpuart: free the rx dma buffer when port is closed

Free the rx dma buffer when the port is closed.

Signed-off-by: Fugang Duan <fugang.duan@nxp.com>
Reviewed-by: Robin Gong <yibin.gong@nxp.com>
7 years agoMLK-15031-01 tty: serial: fsl_lpuart: drop the error frame
Fugang Duan [Fri, 26 May 2017 06:09:27 +0000 (14:09 +0800)]
MLK-15031-01 tty: serial: fsl_lpuart: drop the error frame

Since the driver stats the error frames in port.icount.frame, it can
drop the error frame that no need to push it to the tty buffer.

Signed-off-by: Fugang Duan <fugang.duan@nxp.com>
Reviewed-by: Robin Gong <yibin.gong@nxp.com>
7 years agommc: sdhci: Do not use spin lock in set_ios paths
Adrian Hunter [Mon, 20 Mar 2017 17:50:37 +0000 (19:50 +0200)]
mmc: sdhci: Do not use spin lock in set_ios paths

The spin lock is not necessary in set_ios. Anything that is racing with
changes to the I/O state is already broken. The mmc core already provides
synchronization via "claiming" the host. So remove spin_lock and friends
from sdhci_set_ios and related callbacks.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Tested-by: Ludovic Desroches <ludovic.desroches@microchip.com>
(cherry picked from commit d1e4f74f911daee4d69b5fd0c81902f7d37de579)

Conflicts:
drivers/mmc/host/sdhci-msm.c
drivers/mmc/host/sdhci-of-at91.c
drivers/mmc/host/sdhci.c

7 years agoMLK-15027: arm: pxp: Fix uninitialized use of variables
Cristina Ciocan [Thu, 8 Jun 2017 13:47:38 +0000 (16:47 +0300)]
MLK-15027: arm: pxp: Fix uninitialized use of variables

This patch fixes build warning that 2 variables may be used uninitialized
in the pxp_fetch_config() function in drivers/dma/pxp/pxp_dma_v3.c .

The variables in_fmt and out_fmt are passed as parameters to
pxp_fetch_shift_calc() only if shift_bypass is false. This flag cannot be
false unless changed in a code block that also assigns in_fmt and out_fmt.

Since the compiler cannot detect this flow, it shows a warning that in_fmt
and out_fmt are not initialized. Fix this by changing the code flow such
that in_fmt and out_fmt are sent as parameters in the same code block where
they are assigned.

Signed-off-by: Cristina Ciocan <cristina-mihaela.ciocan@nxp.com>
7 years agoMLK-15052-6: ARM: config: Disable the setting CONFIG_MTD_SPI_NOR_USE_4K_SECTORS
Han Xu [Mon, 8 May 2017 21:12:59 +0000 (16:12 -0500)]
MLK-15052-6: ARM: config: Disable the setting CONFIG_MTD_SPI_NOR_USE_4K_SECTORS

Disable the SECT_4K setting for UBIFS test

Signed-off-by: Han Xu <han.xu@nxp.com>
Acked-by: Frank Li <frank.li@nxp.com>
7 years agoMLK-15052-5: ARM64: defconfig: add flexspi in default config
Han Xu [Mon, 5 Jun 2017 21:23:48 +0000 (16:23 -0500)]
MLK-15052-5: ARM64: defconfig: add flexspi in default config

enable flexspi in default config file

Signed-off-by: Han Xu <han.xu@nxp.com>
Acked-by: Frank Li <frank.li@nxp.com>
7 years agoMLK-15052-4: mtd: spi-nor: fix the micron/st issue
Han Xu [Mon, 8 May 2017 20:12:33 +0000 (15:12 -0500)]
MLK-15052-4: mtd: spi-nor: fix the micron/st issue

Some MICRON related macros in spi-nor domain were ST, actually. We need
to add the REAL micron defination in header/source files for
mt35xu512aba Micron Octal Nor chip.

Signed-off-by: Han Xu <han.xu@nxp.com>
Acked-by: Frank Li <frank.li@nxp.com>
7 years agoMLK-15052-3: mtd: spi-nor: enable octal read mode in spi framework
Han Xu [Mon, 5 Jun 2017 21:19:39 +0000 (16:19 -0500)]
MLK-15052-3: mtd: spi-nor: enable octal read mode in spi framework

Enhanced spi-nor framework to support octal read mode

Signed-off-by: Han Xu <han.xu@nxp.com>
Acked-by: Frank Li <frank.li@nxp.com>
7 years agoMLK-15052-2: mtd: flexspi-nor: support flexspi-nor driver on i.MX8
Han Xu [Mon, 5 Jun 2017 21:20:52 +0000 (16:20 -0500)]
MLK-15052-2: mtd: flexspi-nor: support flexspi-nor driver on i.MX8

support the flexspi nor controller for i.MX8 platforms, read data
in octal ddr mode by default.

Signed-off-by: Han Xu <han.xu@nxp.com>
Acked-by: Frank Li <frank.li@nxp.com>
7 years agoMLK-15052-1: ARM64: dts: add flexspi in 8qxp device tree
Han Xu [Thu, 18 May 2017 15:46:53 +0000 (10:46 -0500)]
MLK-15052-1: ARM64: dts: add flexspi in 8qxp device tree

add the flexspi device tree node for i.mx8qxp

Signed-off-by: Han Xu <han.xu@nxp.com>
Acked-by: Frank Li <frank.li@nxp.com>
7 years agousb: chipdata: Replace the extcon API
Chanwoo Choi [Wed, 30 Nov 2016 05:57:33 +0000 (14:57 +0900)]
usb: chipdata: Replace the extcon API

This patch uses the resource-managed extcon API for extcon_register_notifier()
and replaces the deprecated extcon API as following:
- extcon_get_cable_state_() -> extcon_get_state()

Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Signed-off-by: Peter Chen <peter.chen@nxp.com>
(cherry picked from commit 3f991aa0b665c8e9bb702421a4e5005c3588fb62)

7 years agoextcon: usb-gpio: Don't miss event during suspend/resume
Roger Quadros [Wed, 15 Feb 2017 12:31:28 +0000 (14:31 +0200)]
extcon: usb-gpio: Don't miss event during suspend/resume

We must check for ID/VBUS changes during resume irrespective
of whether our device wakeup is enabled or not.

Without this we seem to be missing ID/VBUS events after
system suspend/resume.

Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
(cherry picked from commit 8680b4d1933fbe3349d51a4e1fd4513b12abffed)

7 years agoextcon: usb-gpio: Do not enable USB as wakeup source by default
Peter Chen [Mon, 13 Feb 2017 09:21:52 +0000 (17:21 +0800)]
extcon: usb-gpio: Do not enable USB as wakeup source by default

Whether the USB port as a wakeup source should be determined by user,
but not enabled by default.

Signed-off-by: Peter Chen <peter.chen@nxp.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
(cherry picked from commit 98fd079297dd274c15c926a337253675573c5832)

7 years agoextcon: usb-gpio: Add pinctrl operation during system PM
Peter Chen [Wed, 4 Jan 2017 07:19:51 +0000 (15:19 +0800)]
extcon: usb-gpio: Add pinctrl operation during system PM

At some systems, the pinctrl setting will be lost or needs to
set as "sleep" state to save power consumption. So, we need to
configure pinctrl as "sleep" state when system enters suspend,
and as "default" state after system resumes. In this way, the
pinctrl value can be recovered as "default" state after resuming.

Signed-off-by: Peter Chen <peter.chen@nxp.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
(cherry picked from commit bcb7440e76a96c8a244bd683142a38f7d5cecb93)

7 years agoextcon: usb-gpio: Add VBUS detection support
Roger Quadros [Tue, 20 Sep 2016 14:53:55 +0000 (17:53 +0300)]
extcon: usb-gpio: Add VBUS detection support

Driver can now work with both ID and VBUS pins or either one of
them.

There can be the following 3 cases

1) Both ID and VBUS GPIOs are available:

ID = LOW -> USB_HOST active, USB inactive
ID = HIGH -> USB_HOST inactive, USB state is same as VBUS.

2) Only ID GPIO is available:

ID = LOW -> USB_HOST active, USB inactive
ID = HIGH -> USB_HOST inactive, USB active

3) Only VBUS GPIO is available:

VBUS = LOW -> USB_HOST inactive, USB inactive
VBUS = HIGH -> USB_HOST inactive, USB active

Signed-off-by: Roger Quadros <rogerq@ti.com>
Reviewed-by: Peter Chen <peter.chen@nxp.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
(cherry picked from commit 541332a13b1ded42097ba96c52c7bc70931e528c)

7 years agoMLK-15032-3 usb: chipidea: core: remove the extcon change for imx_4.1.y
Li Jun [Fri, 9 Jun 2017 10:38:06 +0000 (18:38 +0800)]
MLK-15032-3 usb: chipidea: core: remove the extcon change for imx_4.1.y

commit 916e43e1d6fb ("MLK-13570-3 usb: chipidea: core: change extcon
usage for imx_4.1.y") is directly cherry-picked from 4.1.y, but which
is not valid anymore on 4.y kernel, so revert most part and only keep
the irq check after resume.

Reviewed-by: Peter Chen <peter.chen@nxp.com>
Signed-off-by: Li Jun <jun.li@nxp.com>
7 years agoMLK-15032-2 Revert "extcon: usb-gpio: add pinctrl operation during system PM"
Li Jun [Fri, 9 Jun 2017 10:29:22 +0000 (18:29 +0800)]
MLK-15032-2 Revert "extcon: usb-gpio: add pinctrl operation during system PM"

This reverts commit 4c7d332e3316 ("MLK-13638-3 extcon: usb-gpio: add
pinctrl operation during system PM"). We will use the upstream version.

Reviewed-by: Peter Chen <peter.chen@nxp.com>
Signed-off-by: Li Jun <jun.li@nxp.com>
7 years agoMLK-15032-1 Revert "extcon: ext-usb-gpio: do not enable wakeup by default"
Li Jun [Fri, 9 Jun 2017 10:28:07 +0000 (18:28 +0800)]
MLK-15032-1 Revert "extcon: ext-usb-gpio: do not enable wakeup by default"

This reverts commit 358776f8c5d8 ("MLK-13912-1 extcon: ext-usb-gpio: do not
enable wakeup by default"), we will use the upstream patch version.

Reviewed-by: Peter Chen <peter.chen@nxp.com>
Signed-off-by: Li Jun <jun.li@nxp.com>
7 years agoMLK-15039: ASoC: fsl_esai: Fix channels swap when recording 3 channels audio
Mihai Serban [Thu, 8 Jun 2017 12:48:50 +0000 (15:48 +0300)]
MLK-15039: ASoC: fsl_esai: Fix channels swap when recording 3 channels audio

The change introduced by commit 00c174b3b28a
("MLK-14525: ASoC: fsl_esai: channel swap issue in 3 channels or 5 channels")
is no longer valid after improvements added with commit c35bc6ae5c48
("MLK-14778: ASoC: fsl: imx-cs42888: Improve support for odd number of channels")

Because we use TDM instead of I2S for 3,5 and 7 channels we must
initialize ESAI with the actual number of channels. There is no need
to count the additional channel required when I2S was used.

Signed-off-by: Mihai Serban <mihai.serban@nxp.com>
Reviewed-by: Daniel Baluta <daniel.baluta@nxp.com>
7 years agoMLK-15046 arm64: dts: fsl-imx8qxp: change CAN1 & 2 to use CAN0 clk and power domain
Dong Aisheng [Thu, 8 Jun 2017 13:22:40 +0000 (21:22 +0800)]
MLK-15046 arm64: dts: fsl-imx8qxp: change CAN1 & 2 to use CAN0 clk and power domain

Per information from Ranjani:
"Looks like all three CANs are controlled by one DSC clock slice
(SLSLICE[4]). Currently the SCFW is only allocating this clock to CAN0,
which explains why CAN0 works.  And once CAN0 is enabled, CAN1 and CAN2
access will also work."

This is a workaround patch to make CAN1 & CAN2 work temporarily.
Once SCFW supports shared clock management for all CAN, we can revert
this patch.

Cc: Ranjani Vaidyanathan <Ranjani.Vaidyanathan@nxp.com>
Reviewed-by: Frank Li <frank.li@nxp.com>
Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com>
7 years agoMLK-15046-11 arm64: dts: fsl-imx8qxp-lpddr4-arm2: add flexcan support
Dong Aisheng [Wed, 7 Jun 2017 05:58:13 +0000 (13:58 +0800)]
MLK-15046-11 arm64: dts: fsl-imx8qxp-lpddr4-arm2: add flexcan support

Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com>
7 years agoMLK-15046-10 arm64: defconfig: enable flexcan driver
Dong Aisheng [Tue, 6 Jun 2017 07:32:44 +0000 (15:32 +0800)]
MLK-15046-10 arm64: defconfig: enable flexcan driver

Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com>
7 years agoMLK-15046-9 arm64: dts: imx8qm: add flexcan support
Dong Aisheng [Mon, 19 Sep 2016 08:27:31 +0000 (16:27 +0800)]
MLK-15046-9 arm64: dts: imx8qm: add flexcan support

Add flexcan 1, 2 ,3 support.

Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com>
7 years agoMLK-15046-8 can: flexcan: enable flexcan support for arm64
Dong Aisheng [Wed, 17 May 2017 09:23:54 +0000 (17:23 +0800)]
MLK-15046-8 can: flexcan: enable flexcan support for arm64

Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com>
7 years agoMLK-15046-7 can: flexcan: add imx8qm support
Dong Aisheng [Sat, 3 Jun 2017 14:33:25 +0000 (22:33 +0800)]
MLK-15046-7 can: flexcan: add imx8qm support

The flexcan on MX8QM supports CAN FD prototol.

Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com>
7 years agoMLK-15046-6 can: flexcan: make MB mode store Remote frames
Dong Aisheng [Wed, 7 Jun 2017 03:19:27 +0000 (11:19 +0800)]
MLK-15046-6 can: flexcan: make MB mode store Remote frames

In MB transfer mode, the Remote Response Frame is generated by
default which will not store Remote Frames. That will cause MB
can't resceive Remote Frames. Let's make the Remote Request
Frame stored, then we can receive the Remote Frames and deliver
to userspaces.

Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com>
7 years agoMLK-15046-5 can: flexcan: add can fd bitrate switch support
Dong Aisheng [Sat, 3 Jun 2017 14:14:29 +0000 (22:14 +0800)]
MLK-15046-5 can: flexcan: add can fd bitrate switch support

Add can fd bitrate switch support

Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com>
7 years agoMLK-15046-4 can: flexcan: add can fd mode support
Dong Aisheng [Sat, 3 Jun 2017 07:34:23 +0000 (15:34 +0800)]
MLK-15046-4 can: flexcan: add can fd mode support

Add CAN FD protocol support which supports extended frames up to
64 bytes.

Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com>
7 years agoMLK-15046-3 can: flexcan: make Message Buffer size and number dynamically
Dong Aisheng [Sat, 3 Jun 2017 08:11:21 +0000 (16:11 +0800)]
MLK-15046-3 can: flexcan: make Message Buffer size and number dynamically

The Message Buffer size and number will change in new Flexcan IP
version supporting CAN FD protocol. Let's make them properties and
assigned dynamically.

And note that when CAN FD is enabled, the FlexCAN RAM is partitioned
in blocks of 512 bytes, there may be hole within the block.
Thus a help macro FLEXCAN_CANFD_MB_OFFSET is introduced to calculate
the internal MB address offset properly within the RAM block.

Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com>
7 years agoMLK-15046-2 can: flexcan: add message buffer rx support
Dong Aisheng [Thu, 1 Jun 2017 15:08:16 +0000 (23:08 +0800)]
MLK-15046-2 can: flexcan: add message buffer rx support

This patch adds the optional message buffer rx support which
is controlled by the flag FLEXCAN_QUIRK_DISABLE_RX_FIFO.

This can be used by the new Flexcan version support CAN FD which
can only work with message buffer mode.

Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com>
7 years agoMLK-15046-1 can: flexcan: re-structure to decouple the dependency on fifo mode
Dong Aisheng [Fri, 2 Jun 2017 08:16:20 +0000 (16:16 +0800)]
MLK-15046-1 can: flexcan: re-structure to decouple the dependency on fifo mode

Current driver is highly dependent on the struct flexcan_regs layout
which may vary once the Message Buffer is changed in the new flexcan
version supporting CAN FD protocol.

This patch tends to decouple the dependency on both register layout
and underlying transfer mechanism (FIFO mode or Message buffer).
The Message Buffer mode is still not supported which will be added
later.

To achieve this, struct flexcan_regs is removed and changed to use
reg offset instead and flexcan_{read|write} is then re-prototyped.

Besides that, two new designed API flexcan_mb_read/write are introduced
for raw message buffer read/write which is independent on underlying
transfer mechanism.

This patch has no function change but a preparation for the later
CAN FD support.

Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com>
7 years agoMLK-15029 ARM: imx_v7_defconfig: remove CONFIG_MXC_GPU_VIV=y
Octavian Purdila [Wed, 7 Jun 2017 08:22:58 +0000 (11:22 +0300)]
MLK-15029 ARM: imx_v7_defconfig: remove CONFIG_MXC_GPU_VIV=y

Commit 827acb06eaf00b83 ("MGS-2947 enable gpu subsystem for imx8qm and
imx8qxp") sets MXC_GPU_VIV's default to yes so we should remove
CONFIG_MXC_GPU_VIV=y from the defconfig.

Signed-off-by: Octavian Purdila <octavian.purdila@nxp.com>
7 years agodrm/imx: imx-ldb: remove unnecessary double disable check
Philipp Zabel [Mon, 8 Aug 2016 13:56:39 +0000 (15:56 +0200)]
drm/imx: imx-ldb: remove unnecessary double disable check

Since the atomic modeset conversion, this should not be an issue
anymore.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
(cherry picked from commit cdda2df7e05d84fee0a8298076941928af036c73)

7 years agoMLK-14859 nohz: Fix buggy tick delay on IRQ storms
Frederic Weisbecker [Thu, 1 Jun 2017 14:47:09 +0000 (16:47 +0200)]
MLK-14859 nohz: Fix buggy tick delay on IRQ storms

When the tick is stopped and we reach the dynticks evaluation code on
IRQ exit, we perform a soft tick restart if we observe an expired timer
from there. It means we program the nearest possible tick but we stay in
dynticks mode (ts->tick_stopped = 1) because we may need to stop the tick
again after that expired timer is handled.

Now this solution works most of the time but if we suffer an IRQ storm
and those interrupts trigger faster than the hardware clockevents min
delay, our tick won't fire until that IRQ storm is finished.

Here is the problem: on IRQ exit we reprog the timer to at least
NOW() + min_clockevents_delay. Another IRQ fires before the tick so we
reschedule again to NOW() + min_clockevents_delay, etc... The tick
is eternally rescheduled min_clockevents_delay ahead.

A solution is to simply remove this soft tick restart. After all
the normal dynticks evaluation path can handle 0 delay just fine. And
by doing that we benefit from the optimization branch which avoids
clock reprogramming if the clockevents deadline hasn't changed since
the last reprog. This fixes our issue because we don't do repetitive
clock reprog that always add hardware min delay.

As a side effect it should even optimize the 0 delay path in general.

Reported-and-tested-by: Octavian Purdila <octavian.purdila@nxp.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Rik van Riel <riel@redhat.com>
Cc: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
7 years agoMLK-14999 arm: dts: imx7ulp: add ipg clk for i2c device node
Gao Pan [Fri, 2 Jun 2017 02:09:10 +0000 (10:09 +0800)]
MLK-14999 arm: dts: imx7ulp: add ipg clk for i2c device node

The lpi2c needs two clks, per clk and ipg clk. This patch adds ipg
clk for lpi2c device node.

Signed-off-by: Gao Pan <pandy.gao@nxp.com>
7 years agoMLK-15016-2: arm64: defconfig: added mfgtools defconfig
Frank Li [Tue, 6 Jun 2017 17:04:13 +0000 (12:04 -0500)]
MLK-15016-2: arm64: defconfig: added mfgtools defconfig

Added new config for mfgtools

Signed-off-by: Frank Li <Frank.Li@nxp.com>
7 years agoMLK-15016-1 arm64: dts: imx8qxp: enable chipidea otg port
Frank Li [Tue, 6 Jun 2017 16:56:54 +0000 (11:56 -0500)]
MLK-15016-1 arm64: dts: imx8qxp: enable chipidea otg port

Enable OTG port

Signed-off-by: Frank Li <Frank.Li@nxp.com>
7 years agoMLK-15014 dma: fsl-edma-v3: clear DONE before E_SG enabled
Robin Gong [Tue, 6 Jun 2017 08:56:49 +0000 (16:56 +0800)]
MLK-15014 dma: fsl-edma-v3: clear DONE before E_SG enabled

Below described in RM, otherwise, channel error status(CHa_ES)
may be triggered:
The user must clear the CHa_CSR[DONE] bit before writing the
TCDa_CSR[MAJORELINK] or TCDa_CSR[ESG] bits.

Signed-off-by: Robin Gong <yibin.gong@nxp.com>
7 years agoMLK-14748 clocksource: imx-tpm: Increase the min_delta
Shenwei Wang [Mon, 24 Apr 2017 18:24:32 +0000 (13:24 -0500)]
MLK-14748 clocksource: imx-tpm: Increase the min_delta

The current min_delta for TPM clock event is 2 ticks which
is too small. As the TPM is running at 3MHz, 2 ticks equal
2/3 us. According to our testing, the interrupt latency will
be longer than this min_delta, especially when GPU is running.

This patch changed the min_delta to 300 which give the system
around 100us for interrupt handling in case the "set_next_event"
call is interrupted by other signals.

Also a simple validation code is added before the function returns.

Signed-off-by: Shenwei Wang <shenwei.wang@nxp.com>
Signed-off-by: Bai Ping <ping.bai@nxp.com>
(cherry picked from commit 4f882165cc31672f3c98de74ab02b757cb96ad26)

7 years agoMGS-2842 ARM: imx: correct PFD setting rate flow
Anson Huang [Fri, 21 Apr 2017 22:47:55 +0000 (06:47 +0800)]
MGS-2842 ARM: imx: correct PFD setting rate flow

According to design, PFD needs to be gated before
setting rate, this patch adds warning for PFD when
there is any try to set PFD rate with gate open;

Since PFD may be enabled during kernel boot up,
here doing enable and disable before setting APLL_PFD2
rate is to make sure it is gated by clock framework
before setting rate.

Signed-off-by: Anson Huang <Anson.Huang@nxp.com>
(cherry picked from commit bc731e14dc8401efa55fee65948c3ec31c9e5483)

7 years agoMLK-14972-04 dts: imx8: add thermal nodes for imx8qm/qxp
Bai Ping [Tue, 6 Jun 2017 06:02:21 +0000 (14:02 +0800)]
MLK-14972-04 dts: imx8: add thermal nodes for imx8qm/qxp

Add thermal device related dts node for i.MX8QM/QXP.

Signed-off-by: Bai Ping <ping.bai@nxp.com>
7 years agoMLK-14972-03 ARM64: cofnigs: Enable i.MX8QM/QXP thermal driver by default
Bai Ping [Tue, 6 Jun 2017 06:00:38 +0000 (14:00 +0800)]
MLK-14972-03 ARM64: cofnigs: Enable i.MX8QM/QXP thermal driver by default

Enable the i.MX8QM/QXP thermal driver support in defconfig.

Signed-off-by: Bai Ping <ping.bai@nxp.com>
7 years agoMLK-14972-02 driver: thermal: Add i.MX8QM/QXP thermal support
Bai Ping [Tue, 6 Jun 2017 05:53:24 +0000 (13:53 +0800)]
MLK-14972-02 driver: thermal: Add i.MX8QM/QXP thermal support

Add i.MX8QM/QXP thermal driver support.

Signed-off-by: Bai Ping <ping.bai@nxp.com>
7 years agoMLK-14972-01 doc: dt-bindings: add imx sc temp sensor binding doc
Bai Ping [Wed, 24 May 2017 07:38:45 +0000 (15:38 +0800)]
MLK-14972-01 doc: dt-bindings: add imx sc temp sensor binding doc

Add i.MX8QM/QXP temp sensor binding doc.

Signed-off-by: Bai Ping <ping.bai@nxp.com>
7 years agoMLK-15006-2: ARM64: dts: enable esai and cs42888 in imx8qm dts
Shengjiu Wang [Mon, 5 Jun 2017 03:29:57 +0000 (11:29 +0800)]
MLK-15006-2: ARM64: dts: enable esai and cs42888 in imx8qm dts

Enable ESAI, ASRC, CS42888 in imx8qxp validation board.

Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
7 years agoMLK-15006-1: clk: imx8qm: fix AUD_MLCKOUT0 and AUD_MLCKOUT1 parent issue
Shengjiu Wang [Mon, 5 Jun 2017 03:12:03 +0000 (11:12 +0800)]
MLK-15006-1: clk: imx8qm: fix AUD_MLCKOUT0 and AUD_MLCKOUT1 parent issue

Correct these two audio clock's parent.

Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
7 years agoMGS-2954 arm64: dts: mx8: GPU: Update Frequencies to match with design
Prabhu Sundararaj [Mon, 5 Jun 2017 23:55:18 +0000 (18:55 -0500)]
MGS-2954 arm64: dts: mx8: GPU: Update Frequencies to match with design

Update GPU frequencies for 8QM and 8QXP to match with design.

Signed-off-by: Prabhu Sundararaj <prabhu.sundararaj@nxp.com>
7 years agoMLK-15011: ARM64: dts: increase the cma size for imx8qxp
Shengjiu Wang [Tue, 6 Jun 2017 03:28:18 +0000 (11:28 +0800)]
MLK-15011: ARM64: dts: increase the cma size for imx8qxp

Current the CMA size is 128M, after GPU enabled, there is no
space left for other drivers. So increase it to 640M, which is
align with imx8qm

Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
7 years agoMLK-15004-8: ARM64: defconfig: built-in audio drivers
Shengjiu Wang [Mon, 5 Jun 2017 03:06:56 +0000 (11:06 +0800)]
MLK-15004-8: ARM64: defconfig: built-in audio drivers

built-in the CS42888 sound card, SPDIF and MQS.

Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
7 years agoMLK-15004-7: ARM64: dts: enable esai and cs42888 in imx8qxp dts
Shengjiu Wang [Mon, 5 Jun 2017 03:30:11 +0000 (11:30 +0800)]
MLK-15004-7: ARM64: dts: enable esai and cs42888 in imx8qxp dts

Enable ESAI, ASRC, CS42888 in imx8qxp validation board.

Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
7 years agoMLK-15004-6: Document: sound: update document for audio
Shengjiu Wang [Mon, 5 Jun 2017 08:01:04 +0000 (16:01 +0800)]
MLK-15004-6: Document: sound: update document for audio

update compatible string for imx-audio-cs42888.txt and fsl,esai.txt

Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
7 years agoMLK-15004-5: clk: imx: fix AUD_MLCKOUT0 and AUD_MLCKOUT1 parent issue
Shengjiu Wang [Mon, 5 Jun 2017 03:12:29 +0000 (11:12 +0800)]
MLK-15004-5: clk: imx: fix AUD_MLCKOUT0 and AUD_MLCKOUT1 parent issue

Correct these two audio clock's parent.

Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
7 years agoMLK-15004-4: ASoC: fsl_esai: esai workaround for imx8qxp Rev1
Shengjiu Wang [Mon, 5 Jun 2017 03:04:02 +0000 (11:04 +0800)]
MLK-15004-4: ASoC: fsl_esai: esai workaround for imx8qxp Rev1

In imx8qxp rev1, there is hardware issue (TKT331800). ESAI
dma request signal connection issue in SS_ADMA top level
integration, The ESAI dma request signal are active_low, the EDMA
input is high active, but there is no polarity convert logic
between them.

This patch is to add a workaround for this issue. It use the
GPT to convert dma request signal to EDMA, and use anther GPT
to clear the dma request.

Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
7 years agoMLK-15004-3: ASoC: fsl_esai: switch to use imx-pcm-dma-v2
Shengjiu Wang [Mon, 5 Jun 2017 03:09:25 +0000 (11:09 +0800)]
MLK-15004-3: ASoC: fsl_esai: switch to use imx-pcm-dma-v2

The difference of imx-pcm-dma and imx-pcm-dma-v2 is that first
one will request dma channel in probe, the second one request
dma channel when substream is opened.
When the case is ASRC+ ESAI, the FE+BE is working, which need
to reconfigure the dma channel, so use the imx-pcm-dma-v2 is
more flexible

Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
7 years agoMLK-15004-2: ASoC: fsl_acm: add acm header file
Shengjiu Wang [Mon, 5 Jun 2017 07:16:29 +0000 (15:16 +0800)]
MLK-15004-2: ASoC: fsl_acm: add acm header file

This header file define the offset for control registers
and the GPT capture event.

Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
7 years agoMLK-15004-1: ASoC: codec: cs42xx8: Add reset gpio of codec
Shengjiu Wang [Thu, 23 Jun 2016 10:49:00 +0000 (18:49 +0800)]
MLK-15004-1: ASoC: codec: cs42xx8: Add reset gpio of codec

Need to set the reset pin high when running, otherwise the chip
will remain in reset state.

Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
7 years agoMLK-15003-3: ARM64: dts: add one more cell in edma note
Shengjiu Wang [Mon, 5 Jun 2017 03:17:05 +0000 (11:17 +0800)]
MLK-15003-3: ARM64: dts: add one more cell in edma note

Add one more cell in edma note, which is the property of
local/remote access. The default access is local.

Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
7 years agoMLK-15003-2: Document: fsl_edma_v3: update document
Shengjiu Wang [Mon, 5 Jun 2017 03:37:57 +0000 (11:37 +0800)]
MLK-15003-2: Document: fsl_edma_v3: update document

update fsl_edma_v3 document for #dma-cell is changed
one more cell is added, which is for local/remote access.

Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
7 years agoMLK-15003-1: DMA: fsl-edma-v3: add one more parameter for xlate
Shengjiu Wang [Mon, 5 Jun 2017 03:05:52 +0000 (11:05 +0800)]
MLK-15003-1: DMA: fsl-edma-v3: add one more parameter for xlate

The parameter is "is_remote", which is to use remote access for
edma, the default access is local access.

Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>