]> git.dujemihanovic.xyz Git - u-boot.git/log
u-boot.git
5 years agopico-imx7d: Support distro boot for FIT image case
Jun Nie [Tue, 16 Jul 2019 07:42:57 +0000 (15:42 +0800)]
pico-imx7d: Support distro boot for FIT image case

Support distro boot for pico imx7d in FIT image case.

Signed-off-by: Jun Nie <jun.nie@linaro.org>
Tested-by: Joris Offouga <offougajoris@gmail.com>
5 years agopico-imx7d: add config to enable CAAM
Jun Nie [Tue, 16 Jul 2019 07:42:56 +0000 (15:42 +0800)]
pico-imx7d: add config to enable CAAM

Add config to enable CAAM so that hardware random number generator
works.

Signed-off-by: Jun Nie <jun.nie@linaro.org>
5 years agoARM: imx6q_logic: Remove legacy pinmuxing code from board file.
Adam Ford [Wed, 9 Oct 2019 17:53:57 +0000 (12:53 -0500)]
ARM: imx6q_logic: Remove legacy pinmuxing code from board file.

With the OCRAM expanded to 256KB and the SPL_PINCTRL enabled with
-u-boot.dtsi entries to include the pinmuxing in SPL, the manual
code setting up the pinmux can go away.

This patch removes the legacy pinmuxing code from the board file.

Signed-off-by: Adam Ford <aford173@gmail.com>
Reviewed-by: Fabio Estevam <festevam@gmail.com>
5 years agoARM: imx6q_logic: Enable Pin muxing in SPL
Adam Ford [Wed, 9 Oct 2019 17:53:56 +0000 (12:53 -0500)]
ARM: imx6q_logic: Enable Pin muxing in SPL

With the 256KB of OCRAM available to SPL now, there should be
enough room to enable the pinmuxing in SPL from the device tree.

This patch enables SPL_PINCTRL et al and adds the serial and
usdhc pin mux references to the -u-boot.dtsi file so the pins can be
configured from the device tree.

Signed-off-by: Adam Ford <aford173@gmail.com>
Reviewed-by: Fabio Estevam <festevam@gmail.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
5 years agoARM: imx6q_logic: Enable 256KB OCRAM
Adam Ford [Wed, 9 Oct 2019 17:53:55 +0000 (12:53 -0500)]
ARM: imx6q_logic: Enable 256KB OCRAM

Since this board config only supports, the i.MX6D/i.MX6Q configurations,
256K of on-chip RAM is available which increases the space available
to SPL to help further reduce board code and further enable device tree
functions.

This patch enables 256K of RAM.

Signed-off-by: Adam Ford <aford173@gmail.com>
Reviewed-by: Fabio Estevam <festevam@gmail.com>
5 years agoboard: tbs2910: Convert eth to driver model
Soeren Moch [Thu, 10 Oct 2019 22:59:50 +0000 (00:59 +0200)]
board: tbs2910: Convert eth to driver model

So we can remove lots of legacy config code.

Signed-off-by: Soeren Moch <smoch@web.de>
5 years agoboard: tbs2910: Convert pci to driver model
Soeren Moch [Thu, 10 Oct 2019 22:59:49 +0000 (00:59 +0200)]
board: tbs2910: Convert pci to driver model

So we can get rid of legacy pci config code.

PCI_PNP is not required for this board, remove it to reduce the size of the
u-boot binary.

DM_PCI does not start the host controller automatically anymore. Do so
in preboot instead.

Signed-off-by: Soeren Moch <smoch@web.de>
5 years agoARM: dts: imx6q-tbs2910: Sync devicetree with kernel 5.3
Soeren Moch [Thu, 10 Oct 2019 22:59:48 +0000 (00:59 +0200)]
ARM: dts: imx6q-tbs2910: Sync devicetree with kernel 5.3

Signed-off-by: Soeren Moch <smoch@web.de>
5 years agomx6: mx6sabre common: Use more conventional spelling of fdtfile
Sjoerd Simons [Thu, 20 Jun 2019 20:34:35 +0000 (22:34 +0200)]
mx6: mx6sabre common: Use more conventional spelling of fdtfile

Both the distro boot configuration as well as the general pxe support
standardized on fdtfile as the variable to hold the dtb filename.
Adjust to follow that such that e.g. "pxe boot" will just work.

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
5 years agoAdd i.MX7D based Meerkat96 board support
Shawn Guo [Sun, 7 Jul 2019 12:59:55 +0000 (20:59 +0800)]
Add i.MX7D based Meerkat96 board support

The Meerkat96 board, based on the NXP i.MX7D SoC, is a member of
96Boards community and complies with all Consumer Edition board
specifications.

https://www.novtech.com/products/meerkat96.html
https://www.96boards.org/product/imx7-96/

The initial supported/tested devices include:
 - Debug serial
 - SD
 - USB Host (with Ethernet)

With these support, it's good enough for loading Linux Kernel from SD or
Ethernet over USB.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Tested-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
5 years agoARM: dts: import meerkat96 board support
Shawn Guo [Sun, 7 Jul 2019 12:59:54 +0000 (20:59 +0800)]
ARM: dts: import meerkat96 board support

It imports device tree source of meerkat96 board from Linux Kernel.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
5 years agotravis: Exclude MIPS from the bcm job
Tom Rini [Sun, 13 Oct 2019 15:21:56 +0000 (11:21 -0400)]
travis: Exclude MIPS from the bcm job

We don't need to pull in anything from the MIPS job so exclude that from
the new bcm job and make it clear it's building only ARM.

Signed-off-by: Tom Rini <trini@konsulko.com>
5 years agoMerge https://gitlab.denx.de/u-boot/custodians/u-boot-x86
Tom Rini [Sat, 12 Oct 2019 14:52:48 +0000 (10:52 -0400)]
Merge https://gitlab.denx.de/u-boot/custodians/u-boot-x86

- Remember the device being emulated for Sandbox PCI
- Update Kconfig options for FSP 1.0
- Drop RESET_BASE and RESET_SEG_SIZE that are no longer used

5 years agoMerge branch '2019-10-11-master-imports'
Tom Rini [Sat, 12 Oct 2019 14:10:59 +0000 (10:10 -0400)]
Merge branch '2019-10-11-master-imports'

- Assorted cleanups
- FAT bugfixes
- mediatek platform updates

5 years agoMerge branch '2019-10-11-ti-imports'
Tom Rini [Sat, 12 Oct 2019 14:09:13 +0000 (10:09 -0400)]
Merge branch '2019-10-11-ti-imports'

- Various improvements to dra7xx, keystone 3, am65x SoCs
- Platform updates
- remoteproc improvements

5 years agocmd: itest: add support for .q size specifier
Marek Szyprowski [Wed, 2 Oct 2019 10:24:46 +0000 (12:24 +0200)]
cmd: itest: add support for .q size specifier

Add support for quad (64bits) memory access specifier.

Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Acked-by: Lukasz Majewski <lukma@denx.de>
5 years agofs: fat: get_contents() always returns -1 for errors
Heinrich Schuchardt [Thu, 12 Sep 2019 17:19:30 +0000 (19:19 +0200)]
fs: fat: get_contents() always returns -1 for errors

If out of memory, return -1 and not -ENOMEM from get_contents().

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
5 years agofs: fat: treat invalid FAT clusters as errors
Heinrich Schuchardt [Thu, 12 Sep 2019 17:19:29 +0000 (19:19 +0200)]
fs: fat: treat invalid FAT clusters as errors

When hitting an invalid FAT cluster while reading a file always print an
error message and return an error code.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
5 years agosandbox: Remove sandbox_noblk build
Tom Rini [Fri, 11 Oct 2019 20:28:47 +0000 (16:28 -0400)]
sandbox: Remove sandbox_noblk build

At this point, all drivers that do not use CONFIG_BLK are past their
migration deadlines, so remove this config as it's no longer helpful and
hinders enhancing block drivers.

Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
5 years agospl: Introduce SPL_DM_SPI Kconfig define
Lukasz Majewski [Mon, 9 Sep 2019 10:06:46 +0000 (12:06 +0200)]
spl: Introduce SPL_DM_SPI Kconfig define

This define indicates if DM_SPI shall be supported in SPL. This allows
proper operation of DM converted SPI drivers in SPL, which use
#if !CONFIG_IS_ENABLED(DM_SPI) to also support not yet DM/DTS converted
boards.

Signed-off-by: Lukasz Majewski <lukma@denx.de>
5 years agoarm64: print instructions leading to exception
Heinrich Schuchardt [Thu, 12 Sep 2019 17:09:26 +0000 (19:09 +0200)]
arm64: print instructions leading to exception

If an exception occurs in a loaded image and the relocation offset is
unknown, it is helpful to know the instructions pointed to by the
program counter. This patch adds the missing output.

A possible output is:
    Code: 910c4021 aa1303e0 f9400662 d63f0040 (e7f7defb)

The parentheses indicate the instruction causing the exception.

The output can be disassembled using scripts/decodecode:

echo 'Code: 90000360 9100b800 94002782 17ffff8f (e7f7defb)' | \
  ARCH=arm64 scripts/decodecode

Code: 90000360 9100b800 94002782 17ffff8f (e7f7defb)
All code
========
   0:   90000360    adrp    x0, 0x6c000
   4:   9100b800    add     x0, x0, #0x2e
   8:   94002782    bl      0x9e10
   c:   17ffff8f    b       0xfffffffffffffe48
  10:*  e7f7defb    .inst   0xe7f7defb ; undefined <-- trapping instruction

Code starting with the faulting instruction
===========================================
   0:   e7f7defb    .inst   0xe7f7defb ; undefined

We already have implemented the same for armv7.

For testing command 'exception undefined' can be used.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
5 years agocmd: avoid decimal conversion
T Karthik Reddy [Wed, 11 Sep 2019 13:39:53 +0000 (15:39 +0200)]
cmd: avoid decimal conversion

This patch uses auto instead of decimal in simple_strtoul().

Signed-off-by: T Karthik Reddy <t.karthik.reddy@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
5 years agoARM: asm/io.h: remove redundant #if !defined(readb) block
Rasmus Villemoes [Tue, 10 Sep 2019 08:51:54 +0000 (08:51 +0000)]
ARM: asm/io.h: remove redundant #if !defined(readb) block

readb is unconditionally defined earlier in io.h, so there's no point
checking whether it's undefined.

Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
5 years agoARM: asm/io.h: kill off confusing #ifdef __mem_pci block
Rasmus Villemoes [Tue, 10 Sep 2019 08:51:53 +0000 (08:51 +0000)]
ARM: asm/io.h: kill off confusing #ifdef __mem_pci block

No ARM board seems to define __mem_pci - and if it did, one would get tons of

  ./arch/arm/include/asm/io.h:307:0: warning: "readl" redefined

warnings, because readl and friends are unconditionally defined
earlier in io.h. Moreover, the redefinitions lack the memory barriers
that the first definitions have. So I'm guessing this is practically
dead code.

Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
5 years agospl: mmc: Add option to set eMMC HW boot partition
Mans Rullgard [Tue, 3 Sep 2019 14:16:50 +0000 (16:16 +0200)]
spl: mmc: Add option to set eMMC HW boot partition

This change allows setting pre-defined eMMC boot partition for SPL eMMC
booting. It is necessary in the case when one wants to boot (through falcon
boot) from eMMC after loading SPL from other memory (like SPI-NOR).

Signed-off-by: Mans Rullgard <mans@mansr.com>
[lukma: Edit the commit message]
Signed-off-by: Lukasz Majewski <lukma@denx.de>
Acked-by: Andreas Dannenberg <dannenberg@ti.com>
5 years agospl: mmc: Fix indentation in spl_mmc.c file
Mans Rullgard [Tue, 3 Sep 2019 14:17:17 +0000 (16:17 +0200)]
spl: mmc: Fix indentation in spl_mmc.c file

This fixes a wrongly indented block of code.

Signed-off-by: Mans Rullgard <mans@mansr.com>
[lukma: Make the commit message more verbose]
Signed-off-by: Lukasz Majewski <lukma@denx.de>
5 years agoarm: dra7xx: Hang on any failure during IOdelay recalibration
Lokesh Vutla [Tue, 1 Oct 2019 05:11:01 +0000 (10:41 +0530)]
arm: dra7xx: Hang on any failure during IOdelay recalibration

If there is any failure during IOdelay recalibration sequence, IOs are
not guaranteed to behave as expected. So hang on any failure during the
sequence.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoarm: dra7xx: Fix error path in iodelay recalibration
Lokesh Vutla [Tue, 1 Oct 2019 05:11:00 +0000 (10:41 +0530)]
arm: dra7xx: Fix error path in iodelay recalibration

When an error is reported in __recalibrate_iodelay_start(), de-isolation
of IO doesn't happen. Because of this, undefined behaviour is observed
on many peripherals without any error. So make sure io is out of isolation
at the end of iodelay recalibration.

Reported-by: Richard Woodruff <r-woodruff2@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agotao3530: Fix usage of mmc rescan
Jarkko Nikula [Mon, 30 Sep 2019 17:42:21 +0000 (20:42 +0300)]
tao3530: Fix usage of mmc rescan

Currently "bootcmd" does not work as intended but instead prints MMC usage
information and goes directly to "nandboot".

Follow what the commit 669681104daa ("configs: Fix usage of mmc rescan")
does for other boards prior to support for TechNexion TAO3530 SoM was
added.

Signed-off-by: Jarkko Nikula <jarkko.nikula@bitmer.com>
5 years agoconfigs: am43xx_evm: Enable USB support
Lokesh Vutla [Mon, 30 Sep 2019 12:54:06 +0000 (18:24 +0530)]
configs: am43xx_evm: Enable USB support

spl_dm_init is failing as usb to be probed with the following error:
"omap_dwc3@483c0000: ret=-2"
Enable usb to make spl boot on am43xx devices

Reported-by: Tero Kristo <t-kristo@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoarm: dts: k3-j721e-common-proc-board: Mark main_uart0 as shared device
Lokesh Vutla [Sun, 29 Sep 2019 12:07:58 +0000 (17:37 +0530)]
arm: dts: k3-j721e-common-proc-board: Mark main_uart0 as shared device

Main uart0 is used as debug console by both R5SPL and A72 bootloader and
Linux. So mark it as shared device so that power-domain request is
successful by both cores.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoboard: ti: am335x-ice: Configure the CDCE913 clock synthesizer
Tero Kristo [Fri, 27 Sep 2019 16:14:29 +0000 (19:14 +0300)]
board: ti: am335x-ice: Configure the CDCE913 clock synthesizer

AM335x-ICE boards contain the CDCE913 clock synthesizer, and their
reset crystal capacitance load value of 10pF is wrong leading into
lost packets in certain networking tests. Add DT data for this
device, and probe it from the board file to program the crystal
capacitance load value to 0pF to avoid any problems.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
5 years agoboard: ti: am43xx-idk: Configure the CDCE913 clock synthesizer
Tero Kristo [Fri, 27 Sep 2019 16:14:28 +0000 (19:14 +0300)]
board: ti: am43xx-idk: Configure the CDCE913 clock synthesizer

AM43xx-IDK boards contain the CDCE913 clock synthesizer, and their
reset crystal capacitance load value of 10pF is wrong leading into
lost packets in certain networking tests. Add DT data for this
device, and probe it from the board file to program the crystal
capacitance load value to 0pF to avoid any problems.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
5 years agoboard: ti: am57xx-idk: Configure the CDCE913 clock synthesizer
Tero Kristo [Fri, 27 Sep 2019 16:14:27 +0000 (19:14 +0300)]
board: ti: am57xx-idk: Configure the CDCE913 clock synthesizer

AM57xx-IDK boards contain the CDCE913 clock synthesizer, and their
reset crystal capacitance load value of 10pF is wrong leading into
lost packets in certain networking tests. Add DT data for this
device, and probe it from the board file to program the crystal
capacitance load value to 0pF to avoid any problems.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
5 years agoclk: cdce9xx: add support for cdce9xx clock synthesizer
Tero Kristo [Fri, 27 Sep 2019 16:14:26 +0000 (19:14 +0300)]
clk: cdce9xx: add support for cdce9xx clock synthesizer

Add support for CDCE913/925/937/949 family of devices. These are modular
PLL-based low cost, high performance, programmable clock synthesizers,
multipliers and dividers. They generate up to 9 output clocks from a
single input frequency. The initial version of the driver does not
support programming of the PLLs, and thus they run in the bypass mode
only. The code is loosely based on the linux kernel cdce9xx driver.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
5 years agoarm: k3: Use driver_name to get ti_sci handle
Lokesh Vutla [Fri, 27 Sep 2019 08:02:15 +0000 (13:32 +0530)]
arm: k3: Use driver_name to get ti_sci handle

Use the driver name to get ti_sci handle rather than relying
on just the FIRMWARE uclass.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoconfigs: am65x_evm_a53: Enable DISPLAY_CPUINFO
Lokesh Vutla [Fri, 27 Sep 2019 08:02:14 +0000 (13:32 +0530)]
configs: am65x_evm_a53: Enable DISPLAY_CPUINFO

Enable CONFIG_DISPLAY_CPUINFO so that cpuinfo is printed during boot.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoconfigs: j721e_evm_a72: Enable DISPLAY_CPUINFO
Lokesh Vutla [Fri, 27 Sep 2019 08:02:13 +0000 (13:32 +0530)]
configs: j721e_evm_a72: Enable DISPLAY_CPUINFO

Enable CONFIG_DISPLAY_CPUINFO so that cpuinfo is printed during boot.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoboard: am65x: Print board name and version during boot
Lokesh Vutla [Fri, 27 Sep 2019 08:02:12 +0000 (13:32 +0530)]
board: am65x: Print board name and version during boot

Print the board name and ver along with the DT Model.
While at it print the ver for all the detected daughter cards.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoarm: k3: Add support for printing CPUINFO
Lokesh Vutla [Fri, 27 Sep 2019 08:02:11 +0000 (13:32 +0530)]
arm: k3: Add support for printing CPUINFO

Add support for printing CPU info for all K3 devices.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoconfigs: j721e_evm_a72: Disable PSCI_RESET
Lokesh Vutla [Wed, 25 Sep 2019 09:08:13 +0000 (14:38 +0530)]
configs: j721e_evm_a72: Disable PSCI_RESET

J721E uses TISCI protocol to reset the device and does not
support PSCI reset. So disable PSCI reset.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoconfigs: am65x_evm_a53: Disable PSCI_RESET
Lokesh Vutla [Wed, 25 Sep 2019 09:08:05 +0000 (14:38 +0530)]
configs: am65x_evm_a53: Disable PSCI_RESET

AM65x uses TISCI protocol to reset the device and does not
support PSCI reset. So disable PSCI reset.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agommc: am654_sdhci: Drop a redundant power_domain_on in probe
Lokesh Vutla [Tue, 24 Sep 2019 07:47:16 +0000 (13:17 +0530)]
mmc: am654_sdhci: Drop a redundant power_domain_on in probe

Power-domain is enabled by default in device_probe. am654 mmc driver
is enabling power-domain again in probe. As the second call is
redundant, drop power_domain_on from probe.

Tested-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
5 years agoconfigs: ti: Use addr_fit for run_fit environment variable
Andrew F. Davis [Tue, 17 Sep 2019 19:40:25 +0000 (15:40 -0400)]
configs: ti: Use addr_fit for run_fit environment variable

When running 'run_fit' the FIT file should have been loaded at
'addr_fit', although at this point they should be the same
use this variable instead of 'loadaddr'.

Signed-off-by: Andrew F. Davis <afd@ti.com>
5 years agoam335x, guardian: adapt guardian board to DM
Moses Christopher [Tue, 17 Sep 2019 14:25:38 +0000 (14:25 +0000)]
am335x, guardian: adapt guardian board to DM

  - update partition table - remove env partitions
  - dts: add new interfaces (uart2, extra gpio-key)
         remove unneeded entries
         update nand timings for performance improvement
  - defconfig: adapt configurations to suit DM
               remove unneeded configs
  - am335x_guardian.h: remove mmc boot

Signed-off-by: Moses Christopher <BollavarapuMoses.Christopher@in.bosch.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
5 years agoam335x, guardian: update guardian board
Moses Christopher [Tue, 17 Sep 2019 14:25:37 +0000 (14:25 +0000)]
am335x, guardian: update guardian board

  - add BOARD_LATE_INIT function calls in board.c
  - add swi_status detection in board.c
  - mux: add guardian interfaces to single pinmux structure
  - am33xx, kconfig: add BOARD_LATE_INIT for GUARDIAN board

Signed-off-by: Moses Christopher <BollavarapuMoses.Christopher@in.bosch.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
5 years agocmd: ti: ddr3: Move the print statement after test
Krunal Bhargav [Mon, 16 Sep 2019 08:17:19 +0000 (13:47 +0530)]
cmd: ti: ddr3: Move the print statement after test

If the ECC is enabled over the entire memory region, we need to ensure
the printf/put calls do not modify the stack after ECC is disabled.
Moved the printf/put statements after ECC is enabled.

Signed-off-by: Krunal Bhargav <k-bhargav@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoarm: omap: emif-common: Fix memory priming for ECC
Krunal Bhargav [Mon, 16 Sep 2019 08:17:18 +0000 (13:47 +0530)]
arm: omap: emif-common: Fix memory priming for ECC

Before the priming begins, we need to disable RMW (Read Modify Write)
and disable ECC verification for read accesses. By default, the EMIF
tool enables RMW and read accesses in the EMIF_ECC_CTRL_REG.

Signed-off-by: Krunal Bhargav <k-bhargav@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoarm: omap: emif-common: Disable interleaving
Krunal Bhargav [Mon, 16 Sep 2019 08:17:17 +0000 (13:47 +0530)]
arm: omap: emif-common: Disable interleaving

If ECC is enabled, we need to ensure interleaving is disabled for higher
address space.

Signed-off-by: Krunal Bhargav <k-bhargav@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agocmd: ti: ddr3: Fix ecc address calculation
Lokesh Vutla [Mon, 16 Sep 2019 08:17:16 +0000 (13:47 +0530)]
cmd: ti: ddr3: Fix ecc address calculation

ecc_address_range registers contains the start address and end address
of the DDR address space. But the ddr cmd driver is assuming the register
contains the start address and size of the DDR address space. Because
of this some valid ecc addresses are errored out as invalid address.
Fix this calculation.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoarm: omap: emif-common: Fix ecc address calculation
Lokesh Vutla [Mon, 16 Sep 2019 08:17:15 +0000 (13:47 +0530)]
arm: omap: emif-common: Fix ecc address calculation

ecc_address_range registers contains the start address and end address
of the DDR address space. But the ddr driver is assuming the register
contains the start address and size of the DDR address space. Because
of this the ecc enabling is failing for the 2nd range of ecc addresses.
Fix this calculation.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoregulator: fixed: Modify enable-active-high behavior
Patrice Chotard [Mon, 26 Aug 2019 11:50:31 +0000 (13:50 +0200)]
regulator: fixed: Modify enable-active-high behavior

Regulator should not be enabled at probe time if regulator-boot-on
property is not in the dt node.

"enable-active-high" property is only used to indicate the GPIO
polarity.

See kernel documentation :
 - Documentation/devicetree/bindings/regulator/fixed-regulator.yaml
 - Documentation/devicetree/bindings/regulator/gpio-regulator.yaml

Signed-off-by: Christophe Kerello <christophe.kerello@st.com>
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
5 years agoata: ahci: Don't forget to clear upper address regs.
Oleksandr Rybalko [Thu, 22 Aug 2019 10:26:56 +0000 (12:26 +0200)]
ata: ahci: Don't forget to clear upper address regs.

In 32bits mode upper bits need to be set to 0, otherwise controller will
try to DMA into not existing memory and stops with error.

Tested-by: Frank Wunderlich <frank-w@public-files.de>
Signed-off-by: Frank Wunderlich <frank-w@public-files.de>
Signed-off-by: Oleksandr Rybalko <ray@ddteam.net>
5 years agoahci-pci: ASM1061 report wrong class, but support AHCI.
Oleksandr Rybalko [Thu, 22 Aug 2019 10:26:55 +0000 (12:26 +0200)]
ahci-pci: ASM1061 report wrong class, but support AHCI.

Tested-by: Frank Wunderlich <frank-w@public-files.de>
Signed-off-by: Frank Wunderlich <frank-w@public-files.de>
Signed-off-by: Oleksandr Rybalko <ray@ddteam.net>
5 years agodt-bindings: phy: add a document for MediaTek tphy
Ryder Lee [Thu, 22 Aug 2019 10:26:54 +0000 (12:26 +0200)]
dt-bindings: phy: add a document for MediaTek tphy

This adds a document for tphy which supports physical layer
functionality for a number of controllers on MediaTek SoCs,
such as, USB2.0, USB3.0, PCIe, and SATA.

Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Frank Wunderlich <frank-w@public-files.de>
5 years agodt-bindings: pcie: add a document for MT7623 PCIe controller
Ryder Lee [Thu, 22 Aug 2019 10:26:53 +0000 (12:26 +0200)]
dt-bindings: pcie: add a document for MT7623 PCIe controller

This adds a document for MT7623 PCIe controller.

Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Frank Wunderlich <frank-w@public-files.de>
5 years agoarm: dts: split mtk-reset.h into per-chip header
Ryder Lee [Thu, 22 Aug 2019 10:26:52 +0000 (12:26 +0200)]
arm: dts: split mtk-reset.h into per-chip header

This follows the linux header rules to avoid conflict bitfields.

Tested-by: Frank Wunderlich <frank-w@public-files.de>
Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Frank Wunderlich <frank-w@public-files.de>
5 years agoarm: dts: add PCIe controller for MT7623 SoC
Ryder Lee [Thu, 22 Aug 2019 10:26:51 +0000 (12:26 +0200)]
arm: dts: add PCIe controller for MT7623 SoC

This adds PCIe and its PHY nodes for MT7623.

Tested-by: Frank Wunderlich <frank-w@public-files.de>
Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Frank Wunderlich <frank-w@public-files.de>
5 years agophy: mediatek: add MediaTek T-PHY support for PCIe
Ryder Lee [Thu, 22 Aug 2019 10:26:50 +0000 (12:26 +0200)]
phy: mediatek: add MediaTek T-PHY support for PCIe

The driver provides PHY for USB2, USB3.0, PCIe and SATA, and now
we just enable PCIe. As for the other functionalities will be
added gradually in upcoming days.

This is adapted from the Linux version.

Tested-by: Frank Wunderlich <frank-w@public-files.de>
Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Frank Wunderlich <frank-w@public-files.de>
5 years agopci: mediatek: add PCIe controller support for MT7623
Ryder Lee [Thu, 22 Aug 2019 10:26:49 +0000 (12:26 +0200)]
pci: mediatek: add PCIe controller support for MT7623

This adds PCIe controller support for MT7623.
This is adapted from the Linux version.

Tested-by: Frank Wunderlich <frank-w@public-files.de>
Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Frank Wunderlich <frank-w@public-files.de>
5 years agospl: add a generic function board_init_f
Philippe Reynes [Thu, 19 Sep 2019 14:18:39 +0000 (16:18 +0200)]
spl: add a generic function board_init_f

This commit add a generic function board_init_f that
only initialize some device (for example serial). It
avoid to define a board function only to launch the
serial configuration.

Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
5 years agobcm968580xref: enable spi-nor support
Philippe Reynes [Wed, 14 Aug 2019 13:18:41 +0000 (15:18 +0200)]
bcm968580xref: enable spi-nor support

This commit enable the support of the spi-nor for the
broadcom reference board bcm968580xref.

Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
Reviewed-by: Kursad Oney <kursad.oney@broadcom.com>
5 years agodt: bcm968580xref: add a spi-nor device
Philippe Reynes [Wed, 14 Aug 2019 13:18:40 +0000 (15:18 +0200)]
dt: bcm968580xref: add a spi-nor device

This commit add a spi-nor device in the bcm96850xref device tree.

Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
Reviewed-by: Kursad Oney <kursad.oney@broadcom.com>
5 years agodt: bcm6858: add hsspi controller
Philippe Reynes [Wed, 14 Aug 2019 13:18:39 +0000 (15:18 +0200)]
dt: bcm6858: add hsspi controller

This commit add a hsspi controller in the bcm6858 device tree.

Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
Reviewed-by: Kursad Oney <kursad.oney@broadcom.com>
5 years agoconfigs: Add hsspi/spi support to bcm963158.
Kursad Oney [Wed, 14 Aug 2019 13:18:38 +0000 (15:18 +0200)]
configs: Add hsspi/spi support to bcm963158.

This commit enable the support of the spi-nor for the
broadcom reference board bcm963158.

Signed-off-by: Kursad Oney <kursad.oney@broadcom.com>
Reviewed-by: Philippe Reynes <philippe.reynes@softathome.com>
5 years agodt: bcm963158: add a spi-nor device
Kursad Oney [Wed, 14 Aug 2019 13:18:37 +0000 (15:18 +0200)]
dt: bcm963158: add a spi-nor device

This change adds a spi nor flash device to the bcm963158 board.

Signed-off-by: Kursad Oney <kursad.oney@broadcom.com>
Reviewed-by: Philippe Reynes <philippe.reynes@softathome.com>
5 years agodt: bcm63158: Add hsspi controller
Kursad Oney [Wed, 14 Aug 2019 13:18:36 +0000 (15:18 +0200)]
dt: bcm63158: Add hsspi controller

This change adds the hsspi controller to the 63158 dtsi.

Signed-off-by: Kursad Oney <kursad.oney@broadcom.com>
Reviewed-by: Philippe Reynes <philippe.reynes@softathome.com>
5 years agospi: bcm63xx_hsspi: Continue init when using no reset and fixed-clock.
Kursad Oney [Wed, 14 Aug 2019 13:18:35 +0000 (15:18 +0200)]
spi: bcm63xx_hsspi: Continue init when using no reset and fixed-clock.

The Broadcom ARM implementations do not yet have a clock framework so
one can use a fixed clock as the root clock of the hsspi block. The
fixed clock does not have an "enable" routine, since it's always
enabled. So when we hit this issue, getting an ENOSYS return, do not
bail but continue initialization.

Similarly the block might already have been out of reset, say, when
we are booting from a SPI device. So if the reset signal is not configured
in the device tree, do not bail out and instead skip deasserting the reset.

Signed-off-by: Kursad Oney <kursad.oney@broadcom.com>
Reviewed-by: Philippe Reynes <philippe.reynes@softathome.com>
5 years agospi: bcm63xx_hsspi: switch to raw I/O functions.
Kursad Oney [Wed, 14 Aug 2019 13:18:34 +0000 (15:18 +0200)]
spi: bcm63xx_hsspi: switch to raw I/O functions.

Make the driver compatible with both big and little endian SOCs.
Replace big-endian calls with their raw equivalents, expect for
writing the command to FIFO. That still has to be in big-endian
format.

Signed-off-by: Kursad Oney <kursad.oney@broadcom.com>
Reviewed-by: Philippe Reynes <philippe.reynes@softathome.com>
Reviewed-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
5 years agowaitbit: Add the generic wait_for_bit macros for 16 and 32 bits.
Kursad Oney [Wed, 14 Aug 2019 13:18:33 +0000 (15:18 +0200)]
waitbit: Add the generic wait_for_bit macros for 16 and 32 bits.

wait_for_bit_le32 and wait_for_bit_le16 use the raw I/O functions
which would default to big-endian on BE systems. Create the generic
equivalents to use the native endianness.

Signed-off-by: Kursad Oney <kursad.oney@broadcom.com>
Reviewed-by: Philippe Reynes <philippe.reynes@softathome.com>
Reviewed-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
5 years agospi: hsspi: allow to be used on bcm6858 and bcm63158
Kursad Oney [Wed, 14 Aug 2019 13:18:32 +0000 (15:18 +0200)]
spi: hsspi: allow to be used on bcm6858 and bcm63158

This IP exists in both MIPS and ARM cores, so we also
allow to use this driver on bcm6858 and bcm63158.

Signed-off-by: Kursad Oney <kursad.oney@broadcom.com>
Reviewed-by: Philippe Reynes <philippe.reynes@softathome.com>
Reviewed-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
5 years agoboard: j721e: Add README
Lokesh Vutla [Wed, 4 Sep 2019 10:31:51 +0000 (16:01 +0530)]
board: j721e: Add README

Add README file explaining the build and boot procedure for J721E evm.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoarmv8: K3: j721e: Updated ddr address regions in MMU table
Kedar Chitnis [Wed, 4 Sep 2019 10:31:50 +0000 (16:01 +0530)]
armv8: K3: j721e: Updated ddr address regions in MMU table

The A72 U-Boot code loads and boots a number of remote processors
including the C71x DSP, both the C66_0 and C66_1 DSPs, and the various
Main R5FSS Cores. In order to view the code loaded by the U-Boot by
remote cores, U-Boot should configure the memory region with right
memory attributes. Right now U-Boot carves out a memory region which
is not sufficient for all the images to be loaded. So, increase this
carve out region by 256MB.

Signed-off-by: Kedar Chitnis <kedarc@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoarmv8: K3: am65x: Update DDR address regions in MMU table
Suman Anna [Wed, 4 Sep 2019 10:31:49 +0000 (16:01 +0530)]
armv8: K3: am65x: Update DDR address regions in MMU table

The A53 U-Boot code can load and boot the MCU domain R5F cores (either a
single core in LockStep mode or 2 cores in Split mode) to achieve various
early system functionalities. Change the memory attributes for the DDR
regions used by the remote processors so that the cores can see and
execute the proper code loaded by U-Boot.

These regions are currently limited to 0xa0000000 to 0xa2100000 as per
the DDR carveouts assigned for these R5F cores in the overall DDR memory
map.

Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoconfigs: am65x_evm_a53: Enhance bootcmd to start remoteprocs
Suman Anna [Wed, 4 Sep 2019 10:31:48 +0000 (16:01 +0530)]
configs: am65x_evm_a53: Enhance bootcmd to start remoteprocs

The A53 U-boot can support early booting of the MCU R5F remote processor(s)
from U-boot prompt to achieve various system usecases before booting the
Linux kernel. Update the default BOOTCOMMAND to provide an automatic and
easier way to start the MCU R5F cores through added environment variables.

Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoconfigs: am65x_evm_a53: Enable R5F remoteproc driver
Suman Anna [Wed, 4 Sep 2019 10:31:47 +0000 (16:01 +0530)]
configs: am65x_evm_a53: Enable R5F remoteproc driver

Enable the R5F remoteproc driver for the AM65x GP EVM so that the
MCU domain R5F cores can be booted from A53 U-boot.

Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoconfigs: j721e_evm_a72: Enhance bootcmd to start remoteprocs
Suman Anna [Wed, 4 Sep 2019 10:31:46 +0000 (16:01 +0530)]
configs: j721e_evm_a72: Enhance bootcmd to start remoteprocs

The A72 U-boot can support early booting of any of the R5F or C66x
or C71x remote processors from U-boot prompt to achieve various system
usecases before booting the Linux kernel. Update the default BOOTCOMMAND
to provide an automatic and easier way to start various remote processors
through added environment variables.

Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
Signed-off-by: Suman Anna <s-anna@ti.com>
5 years agoconfigs: j721e_evm_a72: Enable R5F and DSP remoteproc driver
Lokesh Vutla [Wed, 4 Sep 2019 10:31:45 +0000 (16:01 +0530)]
configs: j721e_evm_a72: Enable R5F and DSP remoteproc driver

Enable R5F and DSP remoteproc drivers for j721e running on a72.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoenv: ti: am65x_evm: Add env support to boot the MCU R5F rprocs
Suman Anna [Wed, 4 Sep 2019 10:31:44 +0000 (16:01 +0530)]
env: ti: am65x_evm: Add env support to boot the MCU R5F rprocs

Add support to boot the MCU domain R5F Core0 remoteproc at U-boot prompt
on the AM65x EVM boards by using the 'boot_rprocs' and other env variables
defined in the common environment file k3_rproc.h, and updating the
'DEFAULT_RPROCS' macro.

The default configuration is to use the MCU R5F in Split mode, so both
the R5F Core0 and Core1 are started before loading and booting the Linux
kernel using the following firmware:
   MCU R5FSS0 Core0 (Split) : 0 /lib/firmware/am65x-mcu-r5f0_0-fw
   MCU R5FSS0 Core1 (Split) : 1 /lib/firmware/am65x-mcu-r5f0_1-fw

The MCU R5FSS was initially running the R5 SPL in LockStep mode with ATCM
disabled, and is actually shutdown to enable it to be reconfigured and
booted by either A53 U-Boot or Linux kernel in remoteproc mode and using
ATCM.

The MCU R5FSS would need to be reconfigured for Lockstep mode through
DT if a fault-tolerant/safety application were to be run on the cluster
with the DEFAULT_RPROCS macro updated to remove the Core1 firmware.

Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoenv: ti: j721e-evm: Add support to boot rprocs including R5Fs and DSPs
Suman Anna [Wed, 4 Sep 2019 10:31:43 +0000 (16:01 +0530)]
env: ti: j721e-evm: Add support to boot rprocs including R5Fs and DSPs

Add support to boot some remoteprocs at U-boot prompt on the J721E EVM
boards by using the 'boot_rprocs' and other env variables defined in the
common environment file k3_rproc.h, and updating the 'DEFAULT_RPROCS'
macro.

The list of R5F cores to be started before loading and booting the Linux
kernel are as follows, and in this order:
   Main R5FSS0 (Split) Core1 : 3 /lib/firmware/j7-main-r5f0_1-fw
   Main R5FSS1 (LockStep)    : 4 /lib/firmware/j7-main-r5f1_0-fw

The MCU R5FSS0 and Main R5FSS1 are currently in LockStep mode, so the
equivalent Core1 rprocs (rproc #1 and #5) are not included. The Main
R5FSS0 Core0 (rproc #2) is already started by R5 SPL, so is not included
in the list either.

The DSP cores are started in the following order before loading and
booting the Linux kernel:
   C66_0: 6 /lib/firmware/j7-c66_0-fw
   C66_1: 7 /lib/firmware/j7-c66_1-fw
   C71_0: 8 /lib/firmware/j7-c71_0-fw

The order of the rprocs to boot can be changed at runtime if desired by
overwriting the 'rproc_fw_binaries' environment variable at U-boot prompt.

Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoenv: ti: k3_rproc: Add common rproc environment variables
Suman Anna [Wed, 4 Sep 2019 10:31:42 +0000 (16:01 +0530)]
env: ti: k3_rproc: Add common rproc environment variables

Add a new file include/environment/ti/k3_rproc.h that defines
common environment variables useful for booting various remote
processors from U-Boot. This file is expected to be included in
the board config files with the EXTRA_ENV_RPROC_SETTINGS added
to CONFIG_EXTRA_ENV_SETTINGS and DEFAULT_RPROCS macro overwritten
to include the actual list of processors to be booted.

The 'boot_rprocs' variable just needs to be added to the board's
bootcmd to automatically boot the processors, and runtime control
can be achieved through the 'dorprocboot' variable.

The variables are currently defined to use MMC as the boot media,
and can be expanded in the future to include other boot media.
The immediate usage is intended for K3 J721E SoCs.

Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
Signed-off-by: Suman Anna <s-anna@ti.com>
5 years agoarm: dts: k3-am65-mcu: Add MCU domain R5F DT nodes
Suman Anna [Wed, 4 Sep 2019 10:31:41 +0000 (16:01 +0530)]
arm: dts: k3-am65-mcu: Add MCU domain R5F DT nodes

The AM65x SoCs has a single dual-core Arm Cortex-R5F processor
subsystem/cluster (MCU_R5FSS0) within the MCU domain. This cluster
can be configured at boot time to be either run in a LockStep mode
or in an Asymmetric Multi Processing (AMP) fashion in Split-mode.
This subsystem has 64 KB each Tightly-Coupled Memory (TCM) internal
memories for each core split between two banks - ATCM and BTCM
(further interleaved into two banks). There are some IP integration
differences from standard Arm R5 clusters such as the absence of
an ACP port, presence of an additional TI-specific Region Address
Translater (RAT) module for translating 32-bit CPU addresses into
larger system bus addresses etc.

Add the DT node for the MCU domain R5F cluster/subsystem, the two
R5 cores are added as child nodes to the main cluster/subsystem node.
The cluster is configured to run in Split-mode by default, with the
ATCMs enabled to allow the R5 cores to execute code from DDR with
boot-strapping code from ATCM. The inter-processor communication
between the main A72 cores and these processors is achieved through
shared memory and Mailboxes.

Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoarm: dts: k3-j721e-main: Add C71x DSP node
Lokesh Vutla [Wed, 4 Sep 2019 10:31:40 +0000 (16:01 +0530)]
arm: dts: k3-j721e-main: Add C71x DSP node

The J721E SoCs have a single TMS320C71x DSP Subsystem in the MAIN
voltage domain containing the next-generation C711 CPU core. The
subsystem has 32 KB of L1D configurable SRAM/Cache and 512 KB of
L2 configurable SRAM/Cache. This subsystem has a CMMU but is not
used currently. The inter-processor communication between the main
A72 cores and the C711 processor is achieved through shared memory
and a Mailbox. Add the DT node for this DSP processor sub-system
in the common k3-j721e-main.dtsi file.

Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoarm: dts: k3-j721e-main: Add C66x DSP nodes
Lokesh Vutla [Wed, 4 Sep 2019 10:31:39 +0000 (16:01 +0530)]
arm: dts: k3-j721e-main: Add C66x DSP nodes

The J721E SoCs have two TMS320C66x DSP Core Subsystems (C66x CorePacs)
in the MAIN voltage domain, each with a C66x Fixed/Floating-Point DSP
Core, and 32 KB of L1P & L1D configurable SRAMs/Cache and an additional
288 KB of L2 configurable SRAM/Cache. These subsystems do not have
an MMU but contain a Region Address Translator (RAT) sub-module for
translating 32-bit processor addresses into larger bus addresses.
The inter-processor communication between the main A72 cores and
these processors is achieved through shared memory and Mailboxes.
Add the DT nodes for these DSP processor sub-systems in the common
k3-j721e-main.dtsi file.

Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoarm: dts: k3-j721e-main: Add MAIN domain R5F cluster nodes
Lokesh Vutla [Wed, 4 Sep 2019 10:31:38 +0000 (16:01 +0530)]
arm: dts: k3-j721e-main: Add MAIN domain R5F cluster nodes

The J721E SoCs have 3 dual-core Arm Cortex-R5F processor (R5FSS)
subsystems/clusters. One R5F cluster (MCU_R5FSS0) is present within
the MCU domain, and the remaining two clusters are present in the
MAIN domain (MAIN_R5FSS0 & MAIN_R5FSS1). Each of these can be
configured at boot time to be either run in a LockStep mode or in
an Asymmetric Multi Processing (AMP) fashion in Split-mode. These
subsystems have 64 KB each Tightly-Coupled Memory (TCM) internal
memories for each core split between two banks - ATCM and BTCM
(further interleaved into two banks). There are some IP integration
differences from standard Arm R5 clusters such as the absence of
an ACP port, presence of an additional TI-specific Region Address
Translater (RAT) module for translating 32-bit CPU addresses into
larger system bus addresses etc.

Add the DT nodes for these two MAIN domain R5F cluster/subsystems,
the two R5 cores are each added as child nodes to the corresponding
main cluster node. Configure SS0 in split mode an SS1 in lockstep mode,
with the ATCMs enabled to allow the R5 cores to execute code from DDR
with boot-strapping code from ATCM.

Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoarm: dts: k3-j721e-mcu: Add MCU domain R5F cluster node
Lokesh Vutla [Wed, 4 Sep 2019 10:31:37 +0000 (16:01 +0530)]
arm: dts: k3-j721e-mcu: Add MCU domain R5F cluster node

The J721E SoCs have 3 dual-core Arm Cortex-R5F processor (R5FSS)
subsystems/clusters. One R5F cluster (MCU_R5FSS0) is present within
the MCU domain, and the remaining two clusters are present in the
MAIN domain (MAIN_R5FSS0 & MAIN_R5FSS1). Each of these can be
configured at boot time to be either run in a LockStep mode or in
an Asymmetric Multi Processing (AMP) fashion in Split-mode. These
subsystems have 64 KB each Tightly-Coupled Memory (TCM) internal
memories for each core split between two banks - ATCM and BTCM
(further interleaved into two banks). There are some IP integration
differences from standard Arm R5 clusters such as the absence of
an ACP port, presence of an additional TI-specific Region Address
Translater (RAT) module for translating 32-bit CPU addresses into
larger system bus addresses etc.

Add the DT node for the MCU domain R5F cluster/subsystem, the two
R5 cores are added as child nodes to the main cluster/subsystem node.
The cluster is configured to run in LockStep mode by default, with the
ATCMs enabled to allow the R5 cores to execute code from DDR with
boot-strapping code from ATCM. The inter-processor communication
between the main A72 cores and these processors is achieved through
shared memory and Mailboxes.

Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoremoteproc: Introduce K3 C66 and C71 remoteproc driver
Lokesh Vutla [Wed, 4 Sep 2019 10:31:36 +0000 (16:01 +0530)]
remoteproc: Introduce K3 C66 and C71 remoteproc driver

Certain SoCs with K3 architecture have integrated a C66 Corepac DSP
subsystem and an advanced C71 DSPs. Introduce a remoteproc driver
that that does take care of loading an elf to any of the specified
DSPs and start it.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Signed-off-by: Suman Anna <s-anna@ti.com>
5 years agodt-bindings: remoteproc: Add bindings for DSP C66x clusters on TI K3 SoCs
Suman Anna [Wed, 4 Sep 2019 10:31:35 +0000 (16:01 +0530)]
dt-bindings: remoteproc: Add bindings for DSP C66x clusters on TI K3 SoCs

Some Texas Instruments K3 family of SoCs have one of more Digital Signal
Processor (DSP) subsystems that are comprised of either a TMS320C66x
CorePac and/or a next-generation TMS320C71x CorePac processor subsystem.
Add the device tree bindings document for the C66x DSP devices on these
SoCs. The added example illustrates the DT nodes for the first C66x DSP
device present on the K3 J721E family of SoCs.

Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoremoteproc: Introduce K3 remoteproc driver for R5F subsystem
Lokesh Vutla [Wed, 4 Sep 2019 10:31:34 +0000 (16:01 +0530)]
remoteproc: Introduce K3 remoteproc driver for R5F subsystem

SoCs with K3 architecture have an integrated Arm Cortex-R5F subsystem
that is comprised of dual-core Arm Cortex-R5F processor cores. This R5
subsytem can be configured at boot time to be either run in a LockStep
mode or in an Asymmetric Multi Processing (AMP) fashion in Split-mode.
This subsystem has each Tightly-Coupled Memory (TCM) internal memories
for each core split between two banks - TCMA and TCMB.

Add a remoteproc driver to support this subsystem to be able to load
and boot the R5 cores primarily in LockStep mode or split mode.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Signed-off-by: Suman Anna <s-anna@ti.com>
5 years agodt-bindings: remoteproc: Add bindings for R5F subsystem on TI K3 SoCs
Suman Anna [Wed, 4 Sep 2019 10:31:33 +0000 (16:01 +0530)]
dt-bindings: remoteproc: Add bindings for R5F subsystem on TI K3 SoCs

The Texas Instruments K3 family of SoCs have one of more dual-core
Arm Cortex R5F processor subsystems/clusters (R5FSS). Add the device
tree bindings document for these R5F subsystem devices. These R5F
processors do not have an MMU, and so require fixed memory carveout
regions matching the firmware image addresses. The nodes require more
than one memory region, with the first memory region used for DMA
allocations at runtime. The remaining memory regions are reserved
and are used for the loading and running of the R5F remote processors.

The added example illustrates the DT nodes for the single R5FSS device
present on K3 AM65x family of SoCs.

Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoremoteproc: tisci_proc: Add helper api for controlling core power domain
Lokesh Vutla [Wed, 4 Sep 2019 10:31:32 +0000 (16:01 +0530)]
remoteproc: tisci_proc: Add helper api for controlling core power domain

Power domain for the remote cores needs to be handled in a right
sequence as mandated by the spec. Introduce tisci helper apis
that can control power-domains of remote cores. TISCI clients
can use this api and control the remote cores power domain instead
of hooking it to power-domain layer.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoremoteproc: elf_loader: Introduce rproc_elf_get_boot_addr() api
Lokesh Vutla [Wed, 4 Sep 2019 10:31:31 +0000 (16:01 +0530)]
remoteproc: elf_loader: Introduce rproc_elf_get_boot_addr() api

Introduce rproc_elf_get_boot_addr() that returns the entry point of
the elf file. This api auto detects the 64/32 bit elf file and returns
the boot addr accordingly.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoremoteproc: elf_loader: Introduce a common elf loader and checker functions
Lokesh Vutla [Wed, 4 Sep 2019 10:31:30 +0000 (16:01 +0530)]
remoteproc: elf_loader: Introduce a common elf loader and checker functions

Introduce a common remoteproc elf loader and checker functions that
automatically detects the 64 bit elf file or 32 bit elf file and
loads/checks the sections accordingly.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Fabien Dessenne <fabien.dessenne@st.com>
5 years agoremoteproc: elf-loader: Add 64 bit elf loading support
Lokesh Vutla [Wed, 4 Sep 2019 10:31:29 +0000 (16:01 +0530)]
remoteproc: elf-loader: Add 64 bit elf loading support

The current rproc-elf-loader supports loading of only 32 bit elf files.
Introduce support for loading of 64 bit elf files in rproc-elf-loader.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Fabien Dessenne <fabien.dessenne@st.com>
5 years agoremoteproc: elf_loader: Always check the validity of the image before loading
Lokesh Vutla [Wed, 4 Sep 2019 10:31:28 +0000 (16:01 +0530)]
remoteproc: elf_loader: Always check the validity of the image before loading

rproc_elf32_load_image() rely on user to send a valid address for elf loading.
Instead do a sanity check on the address passed by user. This will help
all rproc elf users to not call sanity_check explicitly before calling
elf_loading.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Fabien Dessenne <fabien.dessenne@st.com>
5 years agoremoteproc: ops: Add elf section size as input parameter to device_to_virt api
Lokesh Vutla [Wed, 4 Sep 2019 10:31:27 +0000 (16:01 +0530)]
remoteproc: ops: Add elf section size as input parameter to device_to_virt api

Introduce a new parameter "size" that accepts size of the region to
remoteproc ops callback device_to_virt(). This can enforce more checks
on the region that device_to_virt() is dealing with.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Tested-by: Fabien Dessenne <fabien.dessenne@st.com>
Reviewed-by: Fabien Dessenne <fabien.dessenne@st.com>
5 years agodm: core: Add a function to count the children of a device
Lokesh Vutla [Wed, 4 Sep 2019 10:31:26 +0000 (16:01 +0530)]
dm: core: Add a function to count the children of a device

Add a function to count the available children of a device.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
5 years agoboard: ti: am65x: Add UART boot procedure in README
Andreas Dannenberg [Thu, 15 Aug 2019 20:55:32 +0000 (15:55 -0500)]
board: ti: am65x: Add UART boot procedure in README

am65x ROM support booting over UART. And U-Boot built for am65x EVM
supports UART boot as well. Add the UART boot procedure into the README
also providing a corresponding example command sequence for execution
on a host PC.

Signed-off-by: Andreas Dannenberg <dannenberg@ti.com>
5 years agoconfigs: am65x_evm_r5: Activate early console functionality
Andreas Dannenberg [Thu, 15 Aug 2019 20:55:31 +0000 (15:55 -0500)]
configs: am65x_evm_r5: Activate early console functionality

Activate early console functionality on AM654x devices to allow for an
alternate serial port to be used to support UART-based boot. This is so
that System Firmware (SYSFW) can get loaded over the serial port prior
to the main console being brought up and made available.

Signed-off-by: Andreas Dannenberg <dannenberg@ti.com>
5 years agoarmv7R: dts: k3: am654: Add MCU_UART0 related definitions
Andreas Dannenberg [Thu, 15 Aug 2019 20:55:30 +0000 (15:55 -0500)]
armv7R: dts: k3: am654: Add MCU_UART0 related definitions

Although we currently use the MAIN_UART0 for R5 SPL console output there
are cases where we require access to the MCU_UART0 as well for example in
case of UART-based Y-Modem boot. To support these scenarios add related
DTS definitions to be able to use that UART early on.

Signed-off-by: Andreas Dannenberg <dannenberg@ti.com>