Simon Glass [Mon, 3 Oct 2016 00:01:05 +0000 (18:01 -0600)]
README: i2c: Drop unused i2c CONFIG options
CONFIG_SYS_NUM_I2C_ADAPTERS and CONFIG_SYS_I2C_MULTI_NOPROBES are not used
in U-Boot, so drop them.
Signed-off-by: Simon Glass <sjg@chromium.org> Acked-by: Heiko Schocher <hs@denx.de> Reviewed-by: Bin Meng <bmeng.cn@gmail.com> Reviewed-by: York Sun <york.sun@nxp.com>
York Sun [Wed, 5 Oct 2016 20:19:08 +0000 (13:19 -0700)]
spi: fsl_qspi: Preserve endianness of QSPI MCR
The endianness can be changed by RCW + PBI sequence. It may have
other than power on reset value.
Signed-off-by: York Sun <york.sun@nxp.com> CC: Yuan Yao <yao.yuan@nxp.com> CC: Peng Fan <peng.fan@nxp.com> CC: Alison Wang <alison.wang@nxp.com> Reviewed-by: Jagan Teki <jteki@openedev.com>
Simon Glass [Sun, 25 Sep 2016 00:20:17 +0000 (18:20 -0600)]
spl: Make spl_boot_list a local variable
There is no need for this to be in the BSS region. By moving it we can delay
use of BSS in SPL. This is useful for machines where the BSS region is not
in writeable space. On 64-bit x86, SPL runs from SPI flash and it is easier
to eliminate BSS use than link SPL to run with BSS at a particular
cache-as-RAM (CAR) address.
Signed-off-by: Simon Glass <sjg@chromium.org> Reviewed-by: Tom Rini <trini@konsulko.com>
Simon Glass [Sun, 25 Sep 2016 00:20:13 +0000 (18:20 -0600)]
spl: Pass spl_image as a parameter to load_image() methods
Rather than having a global variable, pass the spl_image as a parameter.
This avoids BSS use, and makes it clearer what the function is actually
doing.
Signed-off-by: Simon Glass <sjg@chromium.org> Reviewed-by: Tom Rini <trini@konsulko.com>
Simon Glass [Sun, 25 Sep 2016 00:20:11 +0000 (18:20 -0600)]
spl: Convert spl_net_load_image() to use linker list
Add a linker list declaration for this method and remove the explicit
switch() code. We need two variants - one for BOOT_DEVICE_CPGMAC and one for
BOOT_DEVICE_USBETH.
Signed-off-by: Simon Glass <sjg@chromium.org> Reviewed-by: Tom Rini <trini@konsulko.com>
Simon Glass [Sun, 25 Sep 2016 00:20:10 +0000 (18:20 -0600)]
spi: Move freescale-specific code into a private header
At present there are two SPI functions only used by freescale which are
defined in the spi_flash.h header. One function name matches an existing
generic SPL function.
Move these into a private header to avoid confusion.
Arcturus looks like it does not actually support SPI, so drop the SPI code
from that board.
Signed-off-by: Simon Glass <sjg@chromium.org> Reviewed-by: Tom Rini <trini@konsulko.com>
Simon Glass [Sun, 25 Sep 2016 00:19:59 +0000 (18:19 -0600)]
spl: Convert spl_ram_load_image() to use linker list
Add a linker list declaration for this method and remove the explicit
switch() code.
Signed-off-by: Simon Glass <sjg@chromium.org> Reviewed-by: Tom Rini <trini@konsulko.com>
[trini: Include updating the DFU case] Signed-off-by: Tom Rini <trini@konsulko.com>
Simon Glass [Sun, 25 Sep 2016 00:19:51 +0000 (18:19 -0600)]
spl: Move spl_board_load_image() into a generic header
At present this is only used on ARM and sandbox, but it is just as
applicable to other architectures. Move the function prototype into the
generic SPL header.
Signed-off-by: Simon Glass <sjg@chromium.org> Reviewed-by: Tom Rini <trini@konsulko.com>
armv8/fsl-lsch2: Implement workaround for PIN MUX erratum A010539
Pin mux logic has 2 options in priority order, one is through RCW_SRC
and then through RCW_Fields. In case of QSPI booting, RCW_SRC logic
takes the priority for SPI pads and do not allow RCW_BASE and SPI_EXT
to control the SPI muxing. But actually those are DSPI controller's
pads instead of QSPI controller's, so this workaround allows RCW
fields SPI_BASE and SPI_EXT to control relevant pads muxing.
Signed-off-by: Hou Zhiqiang <Zhiqiang.Hou@nxp.com>
[York Sun: Reformatted commit message] Reviewed-by: York Sun <york.sun@nxp.com>
York Sun [Tue, 13 Sep 2016 19:40:30 +0000 (12:40 -0700)]
armv8: fsl-layerscape: Fix "cpu status" command
The core position is not continuous for some SoCs. For example,
valid cores may present at position 0, 1, 4, 5, 8, 9, etc. Some
registers (including boot release register) only count existing
cores. Current implementation of cpu_mask() complies with the
continuous numbering. However, command "cpu status" queries the
spin table with actual core position. Add functions to calculate
core position from core number, to correctly calculate offsets.
fsl_sfp : Modify macros as per changes in SFP v3.4
SFP v3.4 supports 8 keys in SRK table which leads to corresponding
changes in OSPR key revocation field. So modify OSPR_KEY_REVOC_XXX
macros accordingly.
Signed-off-by: Sumit Garg <sumit.garg@nxp.com> Reviewed-by: York Sun <york.sun@nxp.com>
Xiaoliang Yang [Wed, 14 Sep 2016 03:36:14 +0000 (11:36 +0800)]
armv7: LS1021a: enable i-cache in start.S
Delete CONFIG_SKIP_LOWLEVEL_INIT define in ls1021atwr.h and
ls1021aqds.h can let it run cpu_init_cp15 to enable i-cache. First
stage of u-boot can run faster after that. There is a description
about skip lowlevel init in board/freescale/ls1021atwr/README.
Signed-off-by: Xiaoliang Yang <xiaoliang.yang@nxp.com> Reviewed-by: York Sun <york.sun@nxp.com>
Sumit Garg [Wed, 31 Aug 2016 12:54:15 +0000 (08:54 -0400)]
fsl_sec_mon: Update driver for Security Monitor
Update the API's for transition of Security Monitor states. Instead
of providing both initial and final states for transition, just
provide final state for transition as Security Monitor driver will
take care of it internally.
Signed-off-by: Sumit Garg <sumit.garg@nxp.com>
[York Sun: Reformatted commit message slightly] Reviewed-by: York Sun <york.sun@nxp.com>
Tang Yuantian [Mon, 8 Aug 2016 07:07:20 +0000 (15:07 +0800)]
armv8: fsl-lsch2: enable snoopable sata read and write
By default the SATA IP on the ls1043a/ls1046a SoCs does not
generating coherent/snoopable transactions. This patch enable
it in the SCFG_SNPCNFGCR register along with sata axicc register.
In addition, the dma-coherent property must be set on the SATA
controller nodes.
Signed-off-by: Tang Yuantian <yuantian.tang@nxp.com>
[York Sun: Reformatted commit message] Reviewed-by: York Sun <york.sun@nxp.com>
Merge 'spare' into 'bootloader' partition
Use same partition for ramdisk and rootfs boot scenarios.
Remove 'ramdisk' partition, use 'rootfs' for ramdisk
(ramdisk and nand boot scenarios are mutually exclusive).
Expand last partition to end of actual NAND size.
Adjust UBIFS rootfs boot kernel arguments.
Signed-off-by: Albert ARIBAUD (3ADEV) <albert.aribaud@3adev.fr>
Peng Fan [Wed, 28 Sep 2016 01:40:27 +0000 (09:40 +0800)]
imx: imx6ul: disable POR_B internal pull up
>From TO1.1, SNVS adds internal pull up control for POR_B,
the register filed is GPBIT[1:0], after system boot up,
it can be set to 2b'01 to disable internal pull up.
It can save about 30uA power in SNVS mode.
Signed-off-by: Peng Fan <peng.fan@nxp.com> Cc: Stefano Babic <sbabic@denx.de>
Stefan Agner [Thu, 15 Sep 2016 22:04:39 +0000 (15:04 -0700)]
imx: mx6: fix USB bmode to use reserved value
Currently the bmode "usb" uses BOOT_CFG1 to 0x01, -which means
BOOT_CFG1[7:4] is set to b0000. According to Table 8-7 Boot
Device Selection this is NOR/OneNAND and not Reserved.
Use 0x10 which leads to b0001, which is a Reserved boot device.
With that the SoC reliably falls back to the serial loader.
Peng Fan [Thu, 11 Aug 2016 06:02:57 +0000 (14:02 +0800)]
arm: imx: add i.MX6ULL 14x14 EVK board support
Add i.MX6ULL EVK board support:
Add device tree file, which is copied from NXP Linux.
Enabled DM_MMC, DM_GPIO, DM_I2C, DM_SPI, PINCTRL, DM_REGULATOR.
The uart iomux settings are still keeped in board file.
CPU: Freescale i.MX6ULL rev1.0 at 396MHz
CPU: Commercial temperature grade (0C to 95C) at 15C
Reset cause: POR
Model: Freescale i.MX6 ULL 14x14 EVK Board
Board: MX6ULL 14x14 EVK
DRAM: 512 MiB
MMC: initialized IMX pinctrl driver
FSL_SDHC: 0, FSL_SDHC: 1
In: serial
Out: serial
Err: serial
Net: CPU Net Initialization Failed
No ethernet found.
Hit any key to stop autoboot: 0
=> mmc dev 1
switch to partitions #0, OK
mmc1 is current device
Signed-off-by: Peng Fan <peng.fan@nxp.com> Cc: Stefano Babic <sbabic@denx.de>
Peng Fan [Thu, 11 Aug 2016 06:02:53 +0000 (14:02 +0800)]
arm: dts: imx6ull: add pinctrl defines
Add pinctrl defines for NXP i.MX 6ULL.
Since i.MX6ULL reuses some definitions of i.MX6UL,
also add i.MX6UL pinctrl defines from linux kernel commit (29b4817d401).
Signed-off-by: Peng Fan <peng.fan@nxp.com> Cc: Simon Glass <sjg@chromium.org> Cc: Stefano Babic <sbabic@denx.de>
Peng Fan [Thu, 11 Aug 2016 06:02:50 +0000 (14:02 +0800)]
imx: imx6ull: adjust the ldo 1.2v bandgap voltage
Per to design team, on i.MX6UL, the LDO 1.2V bandgap voltage
is 30mV higher, so we need to adjust the REFTOP_VBGADJ(anatop
MISC0 bit[6:4]) setting to 2b'110.
Signed-off-by: Peng Fan <peng.fan@nxp.com> Signed-off-by: Bai Ping <ping.bai@nxp.com> Cc: Stefano Babic <sbabic@denx.de>
Peng Fan [Thu, 11 Aug 2016 06:02:41 +0000 (14:02 +0800)]
imx: ocotp: support i.MX6ULL
i.MX6ULL has two 128 bits fuse banks, bank 7 and bank 8,
while other banks use 256 bits. So we have to adjust the
word and bank index when accessing the bank 8.
When in command line `fuse read 8 0 1`, you can image
`fuse read 7 4 1` in the ocotp driver implementation for 6ULL.
When programming, we use word index, so need to fix bank7/8 programming
for i.mx6ull.
For example: fuse prog 8 3 1; The word index is (8 << 3 | 3) --> 67.
But actully it should be (7 << 3 | 7) ---> 63.
So fix it.
Signed-off-by: Peng Fan <peng.fan@nxp.com> Cc: Stefano Babic <sbabic@denx.de>
There is no stable mmcblk device numbering over different linux versions.
Enable CMD_PART to be able to query the UUID of the root filesystem partition.
So we can pass root=PARTUUID=XXX instead of root=/dev/mmcblkXpY in bootargs.
Leave the default environment as is for now to stay compatible with original
TBS settings.
Filip Brozovic [Wed, 14 Sep 2016 11:50:39 +0000 (13:50 +0200)]
imx: iomux-v3: fix pad setup on i.MX6DQP when CONFIG_MX6QDL is defined
The CPU detection macro is_mx6dq returns 0 on an i.MX6DQP, so we need to
check for it explicitly in order to correctly initialize the pads when
CONFIG_MX6QDL is defined.
Signed-off-by: Filip Brozovic <fbrozovic@gmail.com>
The linux kernel imx_v6_v7_defconfig sets the user/kernel memory split
to 3G/1G now (was 2G/2G before). We have to adapt the BOOTMAPSZ so that
the decompressor finds zImage and dtb in lowmem.
mx6sabresd: Make SPL DDR configuration to match the DCD table
When using SPL on i.mx6 we frequently notice some DDR initialization
mismatches between the SPL code and the non-SPL code.
This causes stability issues like the ones reported at 7dbda25ecd6d7c
("mx6ul_14x14_evk: Pass refsel and refr fields to avoid hang") and also:
http://lists.denx.de/pipermail/u-boot/2016-September/266355.html .
As the non-SPL code have been tested for long time and proves to be reliable,
let's configure the DDR in the exact same way as the non-SPL case.
The idea is simple: just use the DCD table and write directly to the DDR
registers.
Retrieved the DCD tables from:
board/freescale/mx6sabresd/mx6q_4x_mt41j128.cfg
and
board/freescale/mx6sabresd/mx6qp.cfg
(NXP U-Boot branch imx_v2015.04_4.1.15_1.0.0_ga)
This method makes it easier for people converting from non-SPL to SPL code.
Other benefit is that the SPL binary size is reduced from 44 kB to 39.9 kB.