Merge branch 'master' of git://git.denx.de/u-boot-spi

- Fix a conflict in drivers/spi/atcspi200_spi.c related to the riscv
  tree fixing a warning.

Signed-off-by: Tom Rini <trini@konsulko.com>

Merge git://git.denx.de/u-boot-riscv

travis.yml: Support RISC-V 64-bit

Fix riscv: ax25-ae350 build fail problem
https://travis-ci.org/trini/u-boot/jobs/385147373

...
Building current source for 1 boards (1 thread, 2 jobs per thread)
     riscv:  +   ax25-ae350
+arch/riscv/cpu/ax25/start.S: Assembler messages:
+arch/riscv/cpu/ax25/start.S:48: Error: unrecognized opcode `sd a2,0(t0)'
+arch/riscv/cpu/ax25/start.S:112: Error: unrecognized opcode `ld t5,0(t0)'
...

After apply the commit
configs: ax25-ae350: Set 64-bit as default configuration

Toolchain shall be also setuped with 64-bit in .travis.yml.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Reviewed-by: Chih-Mao Chen <cmchen@andestech.com>
Cc: Greentime Hu <green.hu@gmail.com>

Merge git://git.denx.de/u-boot-marvell

Merge branch 'master' of git://git.denx.de/u-boot-ubi

mach-snapdragon: increase size of malloc pool

Pool size must be increased to support new additionals
drivers.

Signed-off-by: Ramon Fried <ramon.fried@gmail.com>

db820c: add dummy clock for uart node

The missing clock causes serial_msm driver probe to fail.
Added a dummy node so the probe succeeds, as the clock init
currently in db820c is empty.

Fixes: 11d59fe5374a ("serial: serial_msm: fail probe if settings clocks fails")

Signed-off-by: Ramon Fried <ramon.fried@gmail.com>

Import Android's dt_table.h for DT image format

Android documentation defines the recommended image format for storing
DTB/DTBO files in a single dtbo.img image. This patch includes the
latest header file with the struct definitions for this format from
AOSP.

The header was adapted to U-Boot's coding style and the function
declarations were removed.

Signed-off-by: Alex Deymo <deymo@google.com>
[trini: Change SDPX tag location]
Signed-off-by: Tom Rini <trini@konsulko.com>

configs: remove CONFIG_SYS_MVFS

This was being used by some Marvell boards to enable some file system
related features (many of which have already been moved to Kconfig).
Make the future migration of the final 2 or 3 config options easier by
expanding #define CONFIG_SYS_MVFS into the options that it enables and
remove CONFIG_SYS_MVFS.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Stefan Roese <sr@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>

mtd: ftsmc020: Drop unsed code

ftsmc020_init is not used anymore.
So it can be removed.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

board: ax25-ae350: Support cfi flash

Add smc_init() to get register base from dts and
deal with atfsmc020 controler initialzation job.

Write protect is enabled by default. So WP shall
be disabled when startup, then cfi flash can be
detected and erasing and writing can be executed.

Adp-ae3xx and adp-ag101p both do smc initilize job
in lowlevel_init.S and get register base fron
CONFIG_FTSMC020_BASE. They also can be moved those
codes to board stage. Remind them as todo jobs.
After that CONFIG_FTSMC020_BASE can be removed.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

riscv: dts: Support cfi flash

Add nor node for cfi-flash driver and smc node
for smc(aftsmc020) controller.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

configs: ax25-ae350: Support cfi flash

Enable cfi flash driver and setup flash
parameters to support parallel nor flash
which type is JS28F00A-M29EWH.

Verification:
Size detection, data read, erase and write are all ok.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

mmc: ftsdc010_mci: Sync compatible with DT mmc node

The compatible string of ftsdc010_mci.c is different from
the mmc driver in Linux Kernel. Modify it for consistency.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

riscv: dts: Sync DT with Linux Kernel

Use same dts to boot U-Boot and RISC-V
Linux Kernel v4.16-rc2 in ax25-ae350 platform.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

net: ftmac100: Fix compiler warning

Fix warnings as below when compile in 64-bit.

warning: cast from pointer to integer
of different size [-Wpointer-to-int-cast]

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

spi: atcspi200: Fix compiler warning

Fix warning as below when compile in 64-bit.

warning: format '%u' expects argument of type
'unsigned int', but argument 6 has type 'size_t
{aka long unsigned int}

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

configs: ax25-ae350: Set 64-bit as default configuration

Set 64-bit as default configuration for ax25-ae350.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

doc: ae250: Rename as ae350

Rename
 nx25 as ax25
 ae250 as ae350
 nx25-ae250 as ax25-ae350
 including filename, variable, string and definition.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

configs: nx25-ae250:Rename as ax25-ae350

Rename
 nx25 as ax25
 ae250 as ae350
 nx25-ae250 as ax25-ae350
 including filename, variable, string and definition.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

board: nx25-ae250: Rename as ax25-ae350

Rename
 nx25 as ax25
 ae250 as ae350
 nx25-ae250 as ax25-ae350
 including filename, variable, string and definition.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

riscv: cpu: nx25: Rename as ax25

Andes has rearranged the product combinations.
nx25 and ax25 both are RISC-V architecture cpu core.
But ax25 has MMU unit inside, and nx25 is not.

Cpu nx25 and platform ae250 are arranged in pairs.
Cpu ax25 and platform ae350 are arranged in pairs.

This patch will rename
nx25 as ax25
ae250 as ae350
nx25-ae250 as ax25-ae350
including filename, variable, string and definition.

Then u-boot can boot linux kernel in ae350
platform reasonably.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

SPDX: Convert single license tags to Linux Kernel style

Fix license tags problem after apply patchs about
riscv: Enable efi_loader support.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

efi_loader: Enable RISC-V support

We have almost all pieces needed to support RISC-V UEFI binaries in place already.
The only missing piece are ELF relocations for runtime code and
data.

This patch adds respective support in the linker script and the runtime
relocation code. It also allows users to enable the EFI_LOADER configuration
switch on RISC-V platforms.

Signed-off-by: Alexander Graf <agraf@suse.de>

riscv: nx25: Enable distro boot

Distro boot allows for a common boot path on systems that allow distributions
to easily boot from a default configuration.

This patch enables distro boot for the nx25-ae250. Hopefully this can serve
as a good example for new boards, so they enable it as well.

Signed-off-by: Alexander Graf <agraf@suse.de>

distro: Extend with RISC-V defines

While we don't have UEFI naming conventions for RISC-V file paths yet,
we need to search for something. So let's copy the removable file paths
from the RISC-V edk2 port.

Also add the official VCI strings that contain the standardized RISC-V
architecture ID fields.

Signed-off-by: Alexander Graf <agraf@suse.de>

efi_loader: Use EFI_CACHELINE_SIZE in the image loader too

We were using our EFI_CACHELINE_SIZE define only in the runtime service
code, but left the image loader to use plain CONFIG_SYS_CACHELINE_SIZE.

This patch moves EFI_CACHELINE_SIZE into efi_loader.h and converts
the image loader to use it.

Signed-off-by: Alexander Graf <agraf@suse.de>

riscv: Add board_quiesce_devices stub

This patch adds an empty stub for board_quiesce_devices() which allows boards
to quiesce their devices before we boot into an OS in a platform agnostic way.

Signed-off-by: Alexander Graf <agraf@suse.de>

riscv: Add EFI application infrastructure

The hello world binary and a few selftests require to build EFI target
binaries, not just the EFI host environment.

This patch adds all required files to generate an EFI binary for
RISC-V.

Signed-off-by: Alexander Graf <agraf@suse.de>

riscv: Enable function sections

The linker can remove sections that are never addressed, so it makes a lot
of sense to declare every function as an individual section.

This reduces the output U-Boot code size by ~30kb for me.

Signed-off-by: Alexander Graf <agraf@suse.de>

riscv: Add setjmp/longjmp code

To support efi_loader we need to have platform support for setjmp/longjmp.
Add it here.

Signed-off-by: Alexander Graf <agraf@suse.de>

configs: remove redundant comment sections

mv-common.h and mv-plug-common.h still had comments delimiting sections
of configuration options that have all been moved to Kconfig by previous
treewide efforts. Remove the redundant comment sections.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>

configs: remove empty #ifdef block from mv-common.h

The last option guarded by this ifdef was removed in commit 68d534201733
("sf: Move SPI flash drivers to defconfig"). Remove the now empty
ifdef/endif block and the associated comment.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>

arm64: mvebu: defconfig: enable CONFIG_AHCI_MVEBU

This patch enables the new ahci mvebu driver for marvell arm64 platform
SOCs(A3k and A8k). And since AHCI_MVEBU selects SCSI_AHCI, so
"CONFIG_SCSI_AHCI=y" is removed from those default config files.

Signed-off-by: Ken Ma <make@marvell.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Stefan Roese <sr@denx.de>

ata: ahci_mvebu: add scsi support

Mvebu AHCI is AHCI driver which uses SCSI under the hood.
This patch adjusts AHCI setup to support SCSI by creating
a SCSI device as a child. Since the functions of creating
SCSI device need the kconfig option DM_SCSI, so let
AHCI_MVEBU select DM_SCSI.

Signed-off-by: Ken Ma <make@marvell.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Stefan Roese <sr@denx.de>

ata: ahci_mvebu: a8040 a0: remove bad port register offsets workarounds

This workaround was added for A8040/7040 A0.
A8040/7040 A0 is no longer supported so this workaround
can be removed.

Signed-off-by: David Sniatkiwicz <davidsn@marvell.com>
Signed-off-by: Ken Ma <make@marvell.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Stefan Roese <sr@denx.de>

ata: mvebu: move mvebu sata driver to drivers/ata directory

Currently mvebu sata driver is in arch/arm/mach_mvebu directory, this
patch moves it to drivers/ata directory with renaming "sata.c" to
"ahci_mvebu.c" which is aligned to Linux.
New ahci driver's kconfig option is added as AHCI_MVEBU which selects
SCSI_AHCI and is based on AHCI.

Signed-off-by: Ken Ma <make@marvell.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Stefan Roese <sr@denx.de>

dts: mvebu: a80x0: Enable SD/eMMC interfaces

Enable SDHCI interface on AP and CP0 in A80x0 DTS files

Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Signed-off-by: Evan Wang <xswang@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>

mvebu: pinctrl: Add SD/eMMC PHY selector to the driver

When the pin control driver selects SD/eMMC function for
a pin group, there is additional configuration to be done
for this case - switch the PHY to work with SDHCI interface.
This patch adds the missing functionality into the pin
control driver.

Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Signed-off-by: Evan Wang <xswang@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>

mvebu: pinctrl: sync compatible string with Linux 4.17-rc4

For pinctrl driver of mvebu, the compatible strings
supported are defined differently from Linux version.
The patch aligned the compatible string with
Linux 4.17-rc4.

Signed-off-by: Evan Wang <xswang@marvell.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>

ARM: kirkwood: Add device-tree for keymile

Import the dts files from Linux 4.17 and enable CONFIG_OF_CONTROL.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>

ARM: kirkwood: Add device-tree for netspace & inetspace

Import the dts files from Linux 4.17 and enable CONFIG_OF_CONTROL.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>

ARM: kirkwood: Add device-tree for lschlv2 & lsxhl

Import the dts files from Linux 4.17 and enable CONFIG_OF_CONTROL.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>

ARM: kirkwood: Add device-tree for ds109

Import the dts files from Linux 4.17 and enable CONFIG_OF_CONTROL.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>

ARM: kirkwood: Add device-tree for dreamplug

Import the dts file from Linux 4.17 and enable CONFIG_OF_CONTROL.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>

ARM: kirkwood: Add device-tree for d2net_v2 & net2big_v2

Import the dts files from Linux 4.17 and enable CONFIG_OF_CONTROL.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>

board: turris_mox: Add fixed regulator support to defconfig

Without this USB3 won't work in U-Boot.

Signed-off-by: Marek Behun <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>

sf: Add Macronix MX25U25635F ID

Add ID for the Macronix MX25U25635F flash.

Signed-off-by: Marek Vasut <marex@denx.de>
Reviewed-by: Jagan Teki <jagan@openedev.com>

sf: Add Winbond W25Q256 ID

Add ID for the Winbond W25Q256 flash.

Signed-off-by: Marek Vasut <marex@denx.de>
Reviewed-by: Jagan Teki <jagan@openedev.com>

add FIT data-position & data-offset property support

Add FIT data-position & data-offset property support for bootm,
which were already supported in SPL.

Signed-off-by: Kelvin Cheung <keguang.zhang@gmail.com>

malloc: Use malloc simple before malloc is fully initialized in memalign()

Follow implementation in mALLOc(). Check GD_FLG_FULL_MALLOC_INIT flag and use
malloc_simple if GD_FLG_FULL_MALLOC_INIT is unset. Adjust the malloc bytes
to align with the requested alignment.

The original memalign() function will access mchunkptr struct to adjust the
alignment if there is misalignment happen, but mchunkptr struct is not being
initialized before full malloc is initialized. This cause the system crash.

Signed-off-by: Ley Foon Tan <ley.foon.tan@intel.com>
Reviewed-by: Marek Vasut <marek.vasut@gmail.com>

mmc: stm32_sdmmc2: Fix stm32_sdmmc2_start_cmd()

SDMMC_CMD_CPSMEN bit is wrongly check and set in
SDMMC_ARG register instead of SDMMC_CMD register.

Signed-off-by: Patrice Chotard <patrice.chotard@st.com>

stm32mp1: use OTP to configure MAC address and serial number

Use OTP57 and 58 for MAC address
- OTP57 = MAC address  bits [31:0]
- OTP58 = MAC address  bit  [47:32] stored in OTP  LSB's

Use manufacture information in OTP13 to OTP15 to build unique
chip id saved in env variable "serial#"
(used for USB device enumeration)

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>

stm32mp1: add FUSE command support

Add support of fuse command (read/write/program/sense)
on bank 0 to access to BSEC SAFMEM (4096 OTP bits).

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>

stm32mp1: add bsec driver

Add a MISC driver with read and write access to BSEC IP
(Boot and Security and OTP control)
- offset 0: shadowed values
- offset 0x80000000: OTP fuse box values (SAFMEM)

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>

stm32mp1: remove the second TAMP_BOOT_CONTEXT update

The register TAMP_BOOT_CONTEXT is already updated in
get_bootmode() in cpu.c and no need to be done
twice.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>

stm32mp1: Allow to activate CONFIG_DEBUG_UART

Add the needed information to enable the debug uart
to have printf before the serial driver probe
(so before probe for clock, pincontrol and reset drivers)

To enable the debug on uart 4 (default console):
+ CONFIG_DEBUG_UART=y
+ CONFIG_DEBUG_UART_STM32=y

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>

serial: stm32: Add setparity support

Add possibility to update the serial parity used.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>

dm: serial: Add setparity

Implements serial setparity ops to allow uart parity change.
It allows to select ODD, EVEN or NONE parity.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>

serial: stm32: Fix bits defines name

Rename USART_ISR_FLAG_xxx bits to USART_ISR_xxx bits and
USART_ICR_OREF to USART_ICR_ORECF in order to match datasheets.
Sort defines by descendant order.

Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

serial: stm32: Add debug uart support

Add support for early debug printf, before the availability of
driver model and device tree support.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

ram: stm32_sdram: Adds stm32f429-disco fixes for HardFault at booting

- adds reading FMC swap setting from DTB to SDRAM driver
- sets FMC swap for stm32f429-disco board
- changes ram start address to 0x90000000

Signed-off-by: Radoslaw Pietrzyk <radoslaw.pietrzyk@gmail.com>
Acked-by: Patrice Chotard <patrice.chotard@st.com>

serial: serial_msm: added pinmux & config

Serial port configuration was missing from previous implementation.
It only worked because it was preconfigured by LK.
This patch configures the uart for 115200 8N1.
It also configures the pin mux for uart pins using DT bindings.

Signed-off-by: Ramon Fried <ramon.fried@gmail.com>

db410: added pinctrl node and serial bindings

Added TLMM pinctrl node for pin muxing & config.
Additionally, added a serial node for uart.

Signed-off-by: Ramon Fried <ramon.fried@gmail.com>

mach-snapdragon: Introduce pinctrl driver

This patch adds pinmux and pinctrl driver for TLMM
subsystem in snapdragon chipsets.
Currently, supporting only 8016, but implementation is
generic and 8096 can be added easily.

Driver is using the generic dt-bindings and doesn't
introduce any new bindings (yet).

Signed-off-by: Ramon Fried <ramon.fried@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

mach-snapdragon: Fix UART clock flow

UART clock enabling flow was wrong.
Changed the flow according to downstream implementation in LK.

Signed-off-by: Ramon Fried <ramon.fried@gmail.com>

serial: serial_msm: initialize uart only before relocation

The uart is already initialized prior to relocation,
reinitialization after relocation is unnecessary.

Signed-off-by: Ramon Fried <ramon.fried@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

serial: serial_msm: fail probe if settings clocks fails

Failure to set the clocks will causes data abort exception when
trying to write to AHB uart registers.
This patch ensures that we don't touch these registers if clock
setting failed.

Signed-off-by: Ramon Fried <ramon.fried@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

db820c: set clk node to be probed before relocation

The clock and serial nodes are needed before relocation.
This patch ensures that the msm-serial driver will probe
and provide uart output before relocation.

Signed-off-by: Ramon Fried <ramon.fried@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

Remove CONFIG_MVGBE from config_whitelist.txt

Now that there are more boards defining this it can be removed from the
whitelist.

Signed-off-by: Chris Packham <judge.packham@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

PCI: Document pciauto_region_allocate()

Add a doc comment for pciauto_region_allocate().

Signed-off-by: Tuomas Tynkkynen <tuomas.tynkkynen@iki.fi>
Reviewed-by: Simon Glass <sjg@chromium.org>

PCI: autoconfig: Don't allocate 64-bit addresses to 32-bit only resources

Currently, if we happen to allocate an address requiring 64 bits to a
device only supporting 32-bit BARs, the address eventually gets silently
truncated to 32 bits. Avoid this by adding a new flag to
pciauto_region_allocate() to bail out in such situations.

Signed-off-by: Tuomas Tynkkynen <tuomas.tynkkynen@iki.fi>
Reviewed-by: Simon Glass <sjg@chromium.org>

PCI: Add newlines to debug prints in pci_auto_common.c

All of the debug output from this file is squished to one line. Fix
it.

Signed-off-by: Tuomas Tynkkynen <tuomas.tynkkynen@iki.fi>
Reviewed-by: Simon Glass <sjg@chromium.org>

doc: qemu-arm: Drop highmem=off references

Now that U-Boot works fine with highmem enabled, there is no need to
tell users to disable highmem.

Signed-off-by: Tuomas Tynkkynen <tuomas.tynkkynen@iki.fi>
Reviewed-by: Simon Glass <sjg@chromium.org>

ARM: qemu-arm: Bump RAM size in AArch64 MMU table

Now that PCI devices work with highmem-enabled QEMU emulation, bump up
the RAM size in the MMU tables to gain access to the full 255 GB of RAM
potential instead of the puny 3 GB.

Signed-off-by: Tuomas Tynkkynen <tuomas.tynkkynen@iki.fi>
Reviewed-by: Simon Glass <sjg@chromium.org>

PCI: dm: Ignore 64-bit memory regions if CONFIG_SYS_PCI_64BIT not set

Currently, qemu_arm_defconfig and qemu_arm64_defconfig only work with
the 'highmem=off' parameter passed to QEMU's virt machine. The reason is
that when 'highmem' is not disabled, QEMU appends 64-bit a memory
resource to the PCI controller's regions property in DT in addition to
the 32-bit PCI memory window in low memory. And the current DT parsing
code picks the last (thus the 64-bit one) memory resource, whose address
eventually gets silently truncated to 32 bits because
CONFIG_SYS_PCI_64BIT is not set, which obviously causes PCI to break.

Avoid this problem by ignoring memory regions whose addresses are above
the 32-bit boundary when CONFIG_SYS_PCI_64BIT is not set.

Signed-off-by: Tuomas Tynkkynen <tuomas.tynkkynen@iki.fi>
Reviewed-by: Simon Glass <sjg@chromium.org>

configs: add TPMv2.x support in Sandbox

Enable the Sandbox TPMv2 driver in all possible configurations.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>

sandbox: dts: add Sandbox TPMv2.x node

This node declares the presence of the Sandbox TPMv2.x emulated chip,
available for testing.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>

doc: device-tree-bindings: add Sandbox TPMv2.0 module info

Add Sandbox TPMv2.0 module bindings.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>

tpm: add a Sandbox TPMv2.x driver

This driver can emulate all the basic functionalities of a TPMv2.x
chip and should behave like them during regular testing.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>

test/py: add TPMv2.x test suite

Add tests for the TPMv2.x commands.
These commands may run both on a physical TPM and with the sandbox
driver.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>

doc: device-tree-bindings: add TIS TPMv2.0 SPI module info

Add TIS TPMv2.0 SPI module bindings.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>

tpm2: tis_spi: add the possibility to reset the chip with a gpio

On some designs, the reset line could not be connected to the SoC reset
line, in this case, request the GPIO and ensure the chip gets reset.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

tpm: add support for TPMv2.x SPI modules

Add the tpm2_tis_spi driver that should support any TPMv2 compliant
(SPI) module.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

tpm: add PCR authentication commands support

Add support for the TPM2_PCR_SetAuthPolicy and
TPM2_PCR_SetAuthValue commands.

Change the command file and the help accordingly.

Note: These commands could not be tested because the TPMs available
do not support them, however they could be useful for someone else.
The user is warned by the command help.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>

tpm: add TPM2_HierarchyChangeAuth command support

Add support for the TPM2_HierarchyChangeAuth command.

Change the command file and the help accordingly.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>

tpm: add dictionary attack mitigation commands support

Add support for the TPM2_DictionaryAttackParameters and
TPM2_DictionaryAttackLockReset commands.

Change the command file and the help accordingly.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>

tpm: add TPM2_GetCapability command support

Add support for the TPM2_GetCapability command.

Change the command file and the help accordingly.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>

tpm: add TPM2_PCR_Read command support

Add support for the TPM2_PCR_Read command.

Change the command file and the help accordingly.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>

tpm: add TPM2_PCR_Extend command support

Add support for the TPM2_PCR_Extend command.

Change the command file and the help accordingly.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>

tpm: add TPM2_Clear command support

Add support for the TPM2_Clear command.

Change the command file and the help accordingly.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>

tpm: add TPM2_SelfTest command support

Add support for the TPM2_Selftest command.

Change the command file and the help accordingly.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>

tpm: add TPM2_Startup command support

Add support for the TPM2_Startup command.

Change the command file and the help accordingly.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>

tpm: report driver error code to upper layer

Instead of returning a generic 'library' error, report back the actual
error code so it can be displayed to the user by the regular error path.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>

tpm: add possible traces to analyze buffers returned by the TPM

When debugging, it is welcome to get more information about what the TPM
returns. Add the possibility to print the packets received to show their
exact content.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>

tpm: add macros to enhance TPM commands readability

TPM commands are much easier to read/write with these macros that will
transform words or integers into byte strings. This way, there is no
need to call pack_byte_string() while all variable length in a command
are known (and at must 4 bytes, which is a lot of them).

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>

tpm: prepare support for TPMv2.x commands

Choice between v1 and v2 compliant functions is done with the
configuration.

Create the various files that will receive TPMv2-only code on the same
scheme as for the TPMv1 code.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>

tpm: add missing parameter in private data structure description

Both parameters 'duration_ms' and 'retry_time_ms' of the tpm_chip_priv
structure are documented is the comment above the declaration but 'buf'
was forgotten. Add the missing description.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

tpm: disociate TPMv1.x specific and generic code

There are no changes in this commit but a new organization of the code
as follow.

* cmd/ directory:
        > move existing code from cmd/tpm.c in cmd/tpm-common.c
> move specific code in cmd/tpm-v1.c
> create a specific header file with generic definitions for
  commands only called cmd/tpm-user-utils.h

* lib/ directory:
        > move existing code from lib/tpm.c in lib/tpm-common.c
> move specific code in lib/tpm-v1.c
> create a specific header file with generic definitions for
  the library itself called lib/tpm-utils.h

* include/ directory:
        > move existing code from include/tpm.h in include/tpm-common.h
> move specific code in include/tpm-v1.h

Code designated as 'common' is compiled if TPM are used. Code designated
as 'specific' is compiled only if the right specification has been
selected.

All files include tpm-common.h.
Files in cmd/ include tpm-user-utils.h.
Files in lib/ include tpm-utils.h.
Depending on the specification, files may include either (not both)
tpm-v1.h or tpm-v2.h.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
[trini: Fix a few more cases of tpm.h -> tpm-v1.h, some Kconfig logic]
Signed-off-by: Tom Rini <trini@konsulko.com>

tpm: prepare introduction of TPMv2.x support in Kconfig

Because both major revisions are not compatible at all, let's make them
mutually exclusive in Kconfig. This way we will be sure, when using a
command or a library function that it is supported by the right
revision.

Current drivers are currently prefixed by "tpm_", we will prefix TPMv2.x
files by "tpm2_" to make the distinction without moving everything.

The Kconfig menu about TPM drivers is now divided into two sections, one
for each specification. Compliant drivers with one specification will
only show up if this specification _only_ has been selected, otherwise a
comment is displayed.

Once a driver is selected by the user, it selects automatically a
boolean value, that is needed in order to activate the TPM commands.
Selecting the TPM commands will automatically select the right
command/library files.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
[trini: Rework deps as TPM_V1 and TPM_V2 depend on TPM,
        drop TPM_DRIVER_SELECTED]
Signed-off-by: Tom Rini <trini@konsulko.com>