arm64: zynqmp: Change firmware dependency

In case of mini U-Boot configurations there is no need to enable firmware
driver which just consume space for nothing. That's why add an option to
disable it.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Link: https://lore.kernel.org/r/d439399160ff3374f2b39f54f7dd70fa8c8bfea0.1642162121.git.michal.simek@xilinx.com

arm64: xilinx: dts: Add dma properties to fix dtbs_check warnings

Update dma name and add #dma-cells properties to fix dtbs_check
warnings.

Signed-off-by: Shravya Kumbham <shravya.kumbham@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Link: https://lore.kernel.org/r/fedbf83fd5c682b4d61905d2cb790d33c2f079d6.1642160644.git.michal.simek@xilinx.com

arm64: zynqmp: Update USB node handle from dwc3 to usb

The DWC3 bindings require all USB node handles to be '^usb@[0-9a-f]+$'.
Update the same in ZynqMP device tree.

Signed-off-by: Manish Narani <manish.narani@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Link: https://lore.kernel.org/r/26fc2777eeb92cb5cb1d558d7c19cfb54ac42d0c.1642160613.git.michal.simek@xilinx.com

arm64: zynqmp: Change compatible strings for cadence uart

Based on Linux kernel DT binding there should be different compatible
strings used that's why align zynqmp.dtsi with it.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Link: https://lore.kernel.org/r/e2404ccd22fd97fe6020be0b3b6eb3c8677f55b0.1642160583.git.michal.simek@xilinx.com

net: gem: Workaround gmii2rgmii bridge DT node issue

For configurations with gmii2rgmii and external phy the DT nodes link
should be gem->gmii2rgmii->phy. But due to limitation in Linux driver
the DT is mentioned as gem->phy and gmii2rgmii->phy as shown in below DT.

ethernet@ff0c0000 {
compatible = "cdns,zynqmp-gem\0cdns,gem";
status = "okay";
interrupt-parent = <0x04>;
interrupts = <0x00 0x3b 0x04 0x00 0x3b 0x04>;
reg = <0x00 0xff0c0000 0x00 0x1000>;
clock-names = "pclk\0hclk\0tx_clk\0rx_clk\0tsu_clk";
#address-cells = <0x01>;
#size-cells = <0x00>;
#stream-id-cells = <0x01>;
iommus = <0x0d 0x875>;
power-domains = <0x0c 0x1e>;
clocks = <0x03 0x1f 0x03 0x69 0x03 0x2e 0x03 0x32 0x03 0x2c>;
phy-handle = <0x0e>;
phy-mode = "gmii";
xlnx,ptp-enet-clock = <0x00>;
local-mac-address = [ff ff ff ff ff ff];
phandle = <0x4d>;

mdio {
#address-cells = <0x01>;
#size-cells = <0x00>;
phandle = <0x4e>;

ethernet-phy@1 {
reg = <0x01>;
rxc-skew-ps = <0x708>;
txc-skew-ps = <0x708>;
phandle = <0x0e>;
};

gmii_to_rgmii_0@8 {
compatible = "xlnx,gmii-to-rgmii-1.0";
phy-handle = <0x0e>;
reg = <0x08>;
phandle = <0x4f>;
};
};
};

Since same DT is used in Linux and U-Boot we need to workaround this
issue by using the gmii2rgmii node which points to phy and we should
ignore the gem pointing to phy directly.

Do this workaround by updating priv->phydev->node value with
priv->phy_of_node only if it is not valid node.

Signed-off-by: Ashok Reddy Soma <ashok.reddy.soma@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Link: https://lore.kernel.org/r/641eb13425ffe80e0743f60cf90d0f940577b9e9.1642162085.git.michal.simek@xilinx.com

Revert "net: gem: Disable broadcast setting"

This reverts commit eafdcda4a854932c0319656de7bf3f017f17ae67.

The main reason is that QEMU is using BOOTP protocol which is sending DHCP
Offer to a broadcast address that's why it can't be disabled.
DHCP protocol has no issue because it returns directly to client MAC
address.
Both of these options are described in RFC951
(https://datatracker.ietf.org/doc/html/rfc951#section-4)

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Link: https://lore.kernel.org/r/fc5f5e2aeca77847ed4ca6a263890375ab9f5163.1642162545.git.michal.simek@xilinx.com

phy: zynqmp: Add serdes/psgtr driver

Add PSGTR driver for Xilinx ZynqMP.
The most of configurations are taken from Linux kernel psgtr driver.

USB3.0 and SGMII configurations are tested on SOM. In SGMII case also
IOU_SLCR reg is updated to get proper clock setup and signal detection
configuration.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Link: https://lore.kernel.org/r/36e6e9d3baf8511af1916e91e4887032ca2b6c20.1641458978.git.michal.simek@xilinx.com

serial: zynq: Add missing xlnx,zynqmp-uart compatible

Based on Linux kernel DT binding xlnx,zynqmp-uart is another compatible
string which can be used for this driver. That's why also list it here.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Link: https://lore.kernel.org/r/959a4cbbdd06a7fca2d9251ff0f535863a562b96.1641905717.git.michal.simek@xilinx.com

net: uclass: Save generated ethernet MAC addresses to the environment

When a MAC address is randomly generated we currently only update the
appropriate data structure.  For consistency and to re-align with
historic usage, it should be also saved to the appropriate environment
variable as well.

Cc: Wolfgang Denk <wd@denx.de>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
[trini: Update Kconfig, handle legacy networking case as well]
Signed-off-by: Tom Rini <trini@konsulko.com>
Acked-by: Ramon Fried <rfried.dev@gmail.com>
Link: https://lore.kernel.org/r/1a2518e3cc19c14a41875ef64c5acc1f16edc813.1641893287.git.michal.simek@xilinx.com

net: gem: Reduce timeout of mdio phy idle status check

Timeout for checking mdio phy idle status is 20seconds. In case of errors
this timeout will be too much. Reduce it to 100ms.

Signed-off-by: Ashok Reddy Soma <ashok.reddy.soma@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Link: https://lore.kernel.org/r/1b73aa57b77587391e1bcd6d9f0480163367ed1b.1637237121.git.michal.simek@xilinx.com

net: zynq: Add support for PHY configuration in SGMII mode

SGMII configuration depends on proper GT setting that's why when node has
phys property call PSGTR driver to configure it properly.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Sean Anderson <seanga2@gmail.com>
Link: https://lore.kernel.org/r/bbc8d7ed9d308199168e4455c7a3e3a5ac0890e7.1639562397.git.michal.simek@xilinx.com

net: zynq: Add support for GEM reset

Perform reset before core initialization.
Standard flow which close to 99% users are using getting all IPs out of
reset that there is no need to reset IP again. This is because of all low
level initialization is done in previous bootloader stage.
In SOM case these IPs are not touched by previous bootloader stage that's
why reset needs to be called before IP is accessed to make sure that it is
in correct state.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Link: https://lore.kernel.org/r/5ae1c85b282d632bb62030f1f24a0065661b9153.1638804318.git.michal.simek@xilinx.com

net: zynq: Add support for mdio bus address decoding

Xilinx DTS files are using two way how to describe ethernet phy.

The first (already supported) has phy as subnode of gem node.
eth {
        phy-handle = <&phy0>;
         phy0: ethernet-phy@21 {
                ...
        };
};

The second has mdio subnode (with mdio name) which has phy subnode. This
structure allow hadling MDIO reset signal (based on Linux mdio.yaml)
eth {
        phy-handle = <&phy0>;
        mdio {
                phy0: ethernet-phy@21 {
                        ...
                };
        };
};

This patch adds support for the second case where mdio subnode
is found driver will look at its parent to find out which gem is handling
MDIO bus.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Link: https://lore.kernel.org/r/6748007f0b6db9554d7a4b52352dce23ca403f9d.1638798796.git.michal.simek@xilinx.com

dt-bindings: versal: Add new PM_DEV_I2C_PMC macro

Add new macro for PMC I2C power domain.

Signed-off-by: Sandeep Gundlupet Raju <sandeep.gundlupet-raju@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Link: https://lore.kernel.org/r/0bf164f937df551d689dda2a35f9489c2e46b4ab.1638277017.git.michal.simek@xilinx.com

versal: Return ENVL_NOWHERE instead of ENVL_UNKNOWN

The system fails to boot without any environment location, so return
ENVL_NOWHERE when there's nowhere to store the environment instead
of ENVL_UNKNOWN.

The same change was also done by commit 50918d0df5cb ("xilinx: Return
ENVL_NOWHERE instead of ENVL_UNKNOWN") for zynq and zynqmp.

Signed-off-by: T Karthik Reddy <t.karthik.reddy@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Link: https://lore.kernel.org/r/e4ed36408f10159677ed8060bfd5289f3e0691fa.1637752614.git.michal.simek@xilinx.com

xilinx: Kconfig: add XILINX_OF_BOARD_DTB_ADDR default value for microblaze

The xilinx board_fdt_blob_setup() implementation makes use of
XILINX_OF_BOARD_DTB_ADDR, but no default value is currently defined for
microblaze. Add one so that microblaze could also work with
CONFIG_OF_SEPARATE.

Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com>
Link: https://lore.kernel.org/r/20211202195657.246723-1-ovidiu.panait@windriver.com
Signed-off-by: Michal Simek <michal.simek@xilinx.com>

microblaze: branch to base vector address on reset

Current code assumes that the vector base address is always at 0x0.
However, this value is configurable for MicroBlaze using the
CONFIG_XILINX_MICROBLAZE0_VECTOR_BASE_ADDR Kconfig option. Update the
reset routines to branch to this location instead.

Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com>
Link: https://lore.kernel.org/r/20211130163358.2531677-10-ovidiu.panait@windriver.com
Signed-off-by: Michal Simek <michal.simek@xilinx.com>

microblaze: start.S: add support for configurable vector base address

Current code assumes that the vector base address is always at 0x0.
However, this value is configurable for MicroBlaze, so update the
__setup_exceptions routine to work with any vector base address.

The r4 register is reserved for the vector base address inside
__setup_exceptions and the function prologe/epilogue are also updated to
save and restore r4.

Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com>
Link: https://lore.kernel.org/r/20211130163358.2531677-9-ovidiu.panait@windriver.com
Signed-off-by: Michal Simek <michal.simek@xilinx.com>

microblaze: add Kconfig symbol for the vector base address

MicroBlaze vector base address is configurable (hdl C_BASE_VECTORS
configuration parameter). Current code assumes that the reset vector
location is always 0x0.

Add the XILINX_MICROBLAZE0_VECTOR_BASE_ADDR Kconfig option so the user
can adjust the reset vector address.

Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com>
Link: https://lore.kernel.org/r/20211130163358.2531677-8-ovidiu.panait@windriver.com
Signed-off-by: Michal Simek <michal.simek@xilinx.com>

microblaze: migrate CONFIG_SYS_USR_EXCEP to Kconfig

Migrate CONFIG_SYS_USR_EXCEP to Kconfig. Also, rename it to
XILINX_MICROBLAZE0_USR_EXCEP in order to match the naming convention of
microblaze-generic Kconfig options.

Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com>
Link: https://lore.kernel.org/r/20211130163358.2531677-7-ovidiu.panait@windriver.com
Signed-off-by: Michal Simek <michal.simek@xilinx.com>

microblaze: drop CONFIG_SYS_RESET_ADDRESS macro

Microblaze is one the last two users of the CONFIG_SYS_RESET_ADDRESS
macro (the other is arch/powerpc/cpu/mpc8xx/cpu.c, but the macro is not
defined anywhere in powerpc code, so it should be removed there too).

Replace CONFIG_SYS_RESET_ADDRESS usage in start.S with
CONFIG_SYS_TEXT_BASE. If the reset address should really be
user-configurable, a new Kconfig option could be added.

Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com>
Link: https://lore.kernel.org/r/20211130163358.2531677-6-ovidiu.panait@windriver.com
Signed-off-by: Michal Simek <michal.simek@xilinx.com>

microblaze: start.S: use stack space as scratch memory for endian offset

To simpify the code, use stack space as scratch memory for endian offset
calculation, rather than saving/restoring the first unused MB vector.

Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com>
Link: https://lore.kernel.org/r/20211130163358.2531677-5-ovidiu.panait@windriver.com
Signed-off-by: Michal Simek <michal.simek@xilinx.com>

microblaze: Kconfig: SPL dependencies fixup

Enable SPL_LIBCOMMON_SUPPORT and SPL_LIBGENERIC_SUPPORT if CONFIG_SPL=y, in
order to fix the following link failures:
common/spl/spl.o: in function `board_init_r':
common/spl/spl.c:755: undefined reference to `puts'
...
common/spl/spl.o: in function `board_init_r':
common/spl/spl.c:756: undefined reference to `hang'
common/spl/spl.c:740: undefined reference to `memset'

Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com>
Link: https://lore.kernel.org/r/20211130163358.2531677-4-ovidiu.panait@windriver.com
Signed-off-by: Michal Simek <michal.simek@xilinx.com>

microblaze: spl: add board_boot_order() implementation

Microblaze has three boot modes defined in microblaze/include/asm/spl.h,
but only booting from NOR flash is currently useable. Add a custom
board_boot_order() implementation so that RAM and SPI boot modes can also
be selected if the corresponding load-image support is present.

Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com>
Link: https://lore.kernel.org/r/20211130163358.2531677-3-ovidiu.panait@windriver.com
Signed-off-by: Michal Simek <michal.simek@xilinx.com>

microblaze: u-boot.lds: replace __end symbol with _end

board_fdt_blob_setup() uses the _end symbol to find the dtb in the non-spl
case. In order to allow microblaze builds to compile successfully with
CONFIG_OF_SEPARATE, the _end symbol must be defined. Align microblaze with
the other architectures and use _end symbol rather than __end to mark the
end of the u-boot binary.

Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com>
Link: https://lore.kernel.org/r/20211130163358.2531677-2-ovidiu.panait@windriver.com
Signed-off-by: Michal Simek <michal.simek@xilinx.com>

xilinx: firmware: Move dcache handling directly to pmufw load config

Core function should make sure that data is stored properly that's why move
cache operations directly to zynqmp_pmufw_load_config_object() to be able
to call it from other functions.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Link: https://lore.kernel.org/r/8c92edd3650ce34a3cfd1c1e4e9103980830b1fa.1637236800.git.michal.simek@xilinx.com

arm64: zynqmp: Switch SOM to shared psu configuration

Previous psu init was targeting SOM + KV260 carrier card and also contain
configurations for other devices on carrier card. This config is removing
all expected configurations for CC and let U-Boot to handle all of it self.
This configuration is designed for SOM itself (and I would bet without
eMMC).

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Link: https://lore.kernel.org/r/5782131058dc372befd3fdb4dceabeea5ba56606.1637236693.git.michal.simek@xilinx.com

xilinx: versal: Fix sdhci node name as per DT

Fix the sdhci node name in versal board file as per the name in
device tree and also check for sdhci node as part of backward
compatibility.

Signed-off-by: T Karthik Reddy <t.karthik.reddy@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Link: https://lore.kernel.org/r/9110ecdabcabcef63fffd4719095acf4326a26e4.1637236638.git.michal.simek@xilinx.com

arm64: zynqmp: Remove clock-names from GEM in zynqmp-clk-ccf.dtsi

Remove clock-names from GEM nodes from clk-ccf because they should be only
present in zynqmp.dtsi. And as is visible both clock-names defined didn't
really match.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Link: https://lore.kernel.org/r/d6045d81b3e7e97df0ba3eeacb9f3f75ed7cff18.1637239345.git.michal.simek@xilinx.com

arm64: zynqmp: Add resets to all GEMs

There is a need to get IP out of reset to operate properly.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Link: https://lore.kernel.org/r/041362197e8de8e9c696da64429107505bdc0c73.1637239345.git.michal.simek@xilinx.com

arm64: zynqmp: Add u-boot,dm-pre-reloc to dpsub node

u-boot,dm-pre-reloc is necessary for DP driver to allocate enough space for
framebuffer before relocation.
Power domain driver is called when video console is used for example by
loading BMP image.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Link: https://lore.kernel.org/r/5e93acee1746913a6e42741e5e797f0b4ab98d44.1637239230.git.michal.simek@xilinx.com

clk: zynqmp: Fix gem tx/rx/ref clock handling

gemX_ref clock IDs starts at number 104. Till now it was at gemX_tx
location which wasn't correct.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Link: https://lore.kernel.org/r/d073e159b6316707306092a62bccb876cd89a602.1635506016.git.michal.simek@xilinx.com

clk: zynqmp: Add support for setting up clock for USB

USB range is not enabled but for setting up frequency it is needed.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Link: https://lore.kernel.org/r/c55c423f48ca8f953a2dfbdcb25068278d8e5ad6.1635506016.git.michal.simek@xilinx.com

zynqmp: gpio: Add support for zynqmp gpio modepin driver

ZynqMP modepin driver has capability to get/set/check status of modepin
gpios. These modepins are accessed using xilinx firmware. In modepin
register, [3:0] bits set direction, [7:4] bits read IO, [11:8] bits
set/clear IO.

Signed-off-by: T Karthik Reddy <t.karthik.reddy@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Link: https://lore.kernel.org/r/2d802d98fd56d95d764532a33e844d935e0cebb3.1635505900.git.michal.simek@xilinx.com

Prepare v2022.01-rc3

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

board: iot2050: update build documentation for OP-TEE

Set ta-target explicitly to correspond with OP-TEE recipe in
siemens/meta-iot2050.

Errors without explicit set of ta-target:
aarch64-linux-gnu-gcc: error: unrecognized command-line option ‘-mthumb’
aarch64-linux-gnu-gcc: error: unrecognized command-line option ‘-mno-unaligned-access’
aarch64-linux-gnu-gcc: error: unrecognized command-line option ‘-mfloat-abi=hard’
make: *** [mk/compile.mk:159: out/arm-plat-k3/ta_arm32-lib/libdl/dlfcn.o] Error 1

Signed-off-by: Ivan Mikhaylov <ivan.mikhaylov@siemens.com>
Reviewed-by: Jan Kiszka <jan.kiszka@siemens.com>

qemu: common: Fix build with update capsule

The common emulation Makefile has a dependency on a non-existent
qemu_capsule.o when building with support for capsule update enabled
(CONFIG_EFI_RUNTIME_UPDATE_CAPSULE=y).
The code which was in qemu_capsule.c has been completely moved to
lib/efi_loader/efi_capsule.c by commit 7a6fb28c8e4b ("efi_loader: capsule:
add back efi_get_public_key_data()").
Remove the false dependency.

This fixes the following build error:

  make[1]: *** No rule to make target 'board/emulation/common/qemu_capsule.o', needed by 'board/emulation/common/built-in.o'.  Stop.

Fixes: commit 47a25e81d35c ("Revert "efi_capsule: Move signature from DTB to .rodata"")
Signed-off-by: Vincent Stehlé <vincent.stehle@arm.com>
Cc: Simon Glass <sjg@chromium.org>

fastboot: Add maintainers entry

Add an entry in maintainers for fastboot. It is starting off orphaned, but
hopefully someone can pick it up.

Signed-off-by: Sean Anderson <seanga2@gmail.com>

clk: Add myself as a maintainer for the clock subsystem

Lukasz has not been very responsive in reviewing clock patches. Add
myself as a maintainer.

Signed-off-by: Sean Anderson <seanga2@gmail.com>
Acked-by: Lukasz Majewski <lukma@denx.de>

Merge tag 'efi-2022-01-rc3-2' of https://source.denx.de/u-boot/custodians/u-boot-efi

Pull request for efi-2022-01-rc3-2

Test:
* fix pylint warnings

UEFI:
* disable interrupts before removing devices in ExitBootServices()
* implement poweroff in efi_system_reset() on sandbox
* allow booting via EFI even if some block device fails

test: fix pylint error in u_boot_console_exec_attach.py

* provide module docstring

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

test: fix pylint error in u_boot_console_sandbox.py

* provide module docstring

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

test: fix pylint errors in u_boot_utils.py

* there is no os.path.unlink() method
* don't inherit from object
* add module docstring
* move imports to the top
* avoid unused variable

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

test: fix pylint errors in u_boot_spawn.py

* don't inherit from object
* imports should be on the top level
* avoid unused variable names
* avoid unnecessary else after raise

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

test: fix pylint errors in multiplexed_log.py

* don't inherit from object
* remove superfluous comprehension
* add module docstring

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

efi_selftest: simplify endian conversion for FDT test

UEFI code is always little-endian. Remove a superfluous test.

Remove a superfluous type conversion.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>

efi_loader: segfault in efi_clear_os_indications()

If we call efi_clear_os_indications() before initializing the memory store
for UEFI variables a NULL pointer dereference occurs.

The error was observed on the sandbox with:

    usb start
    host bind 0 sandbox.img
    load host 0:1 $kernel_addr_r helloworld.efi
    bootefi $kernel_addr_r

Here efi_resister_disk() failed due to an error in the BTRFS implementation.

Move the logic to clear EFI_OS_INDICATIONS_FILE_CAPSULE_DELIVERY_SUPPORTED
to the rest of the capsule code.

If CONFIG_EFI_IGNORE_OSINDICATIONS=y, we should still clear the flag.
If OsIndications does not exist, we should not create it as it is owned by
the operating system.

Fixes: 149108a3eb59 ("efi_loader: clear OsIndications")
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Acked-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>

test: address some pylint warnings

* remove unused variables
* module description must precede import statements
* fix inconsistent return values

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>

sandbox: poweroff in efi_system_reset()

efi_system_reset() should exit if called with EFI_RESET_SHUTDOWN.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

efi_loader: efi_disk_register() should not fail

Our algorithm for creating USB device paths may lead to duplicate device
paths which result in efi_disk_register() failing. Instead we should just
skip devices that cannot be registered as EFI block devices.

Fix a memory leak in efi_disk_add_dev() caused by the duplicate device
path.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>

efi: Call bootm_disable_interrupts earlier in efi_exit_boot_services

If we look at the path that bootm/booti take when preparing to boot the
OS, we see that as part of (or prior to calling do_bootm_states,
explicitly) the process, bootm_disable_interrupts() is called prior to
announce_and_cleanup() which is where udc_disconnect() /
board_quiesce_devices() / dm_remove_devices_flags() are called from.  In
the EFI path, these are called afterwards.  In efi_exit_boot_services()
however we have been calling bootm_disable_interrupts() after the above
functions, as part of ensuring that we disable interrupts as required
by the spec.  However, bootm_disable_interrupts() is also where we go
and call usb_stop().  While this has been fine before, on the TI J721E
platform this leads us to an exception.  This exception seems likely to
be the case that we're trying to stop devices that we have already
disabled clocks for.  The most direct way to handle this particular
problem is to make EFI behave like the do_bootm_states() process and
ensure we call bootm_disable_interrupts() prior to ending up in
usb_stop().

Cc: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Cc: Heinrich Schuchardt <xypron.glpk@gmx.de>
Cc: Simon Glass <sjg@chromium.org>
Suggested-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>

Merge branch 'efi-2022-01' of https://source.denx.de/u-boot/custodians/u-boot-efi

Scripts:
* Update spelling.txt

LMB:
* remove extern keyword in lmb.h
* drop unused lmb_size_bytes()

Test:
* test truncation in snprintf()

Documentation:
* add include/lmb.h to HTML documentation

UEFI:
* reduce non-debug logging output for measured boot
* fix use after free in measured boot
* startup the tpm device when installing the protocol
* implement EFI_EVENT_GROUP_BEFORE_EXIT_BOOT_SERVICES
* record capsule result only if capsule is read

efi_loader: startup the tpm device when installing the protocol

Due to U-Boot's lazy binding mentality the TPM is probed but not properly
initialized.  The user can startup the device from the command line
e.g 'tpm2 startup TPM2_SU_CLEAR'.  However we can initialize the TPM during
the TCG protocol installation,  which is easier to use overall.

Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Reviewed-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>

efi_loader: fix FinalEvents table if an EFI app invoked GetEventLog

As described in the TCG spec [1] in sections 7.1.1 and 7.1.2 the FinalEvent
table should include events after GetEventLog has been called.  This
currently works for us as long as the kernel is the only EFI application
calling that.  Specifically we only implement what's described in 7.1.1.

So refactor the code a bit and support EFI application(s) calling
GetEventLog.  Events will now be logged in both the EventLog and FinalEvent
table as long as ExitBootServices haven't been invoked.

[1] https://trustedcomputinggroup.org/wp-content/uploads/EFI-Protocol-Specification-rev13-160330final.pdf

Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>

efi_loader: bump EFI_SPECIFICATION_VERSION to 2.9

We have implemented all what is new in UEFI specification 2.9 and relevant
for U-Boot.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>

efi_selftest: unit test for EFI_GROUP_BEFORE_EXIT_BOOT_SERVICE

Add a test for the EFI_GROUP_BEFORE_EXIT_BOOT_SERVICE event group.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>

efi_loader: EFI_EVENT_GROUP_BEFORE_EXIT_BOOT_SERVICES

Implement the EFI_EVENT_GROUP_BEFORE_EXIT_BOOT_SERVICES event group
handling.

Add the definition of EFI_EVENT_GROUP_AFTER_READY_TO_BOOT.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>

efi_loader: capsule: Record capsule result only if capsule is read

Record capsule update result only if the capsule file is
successfully read, because the capsule GUID is not sure when
the file can not be read or the file is not a capsule.
Without this fix, if user puts a dummy (non-capsule) file
under (ESP)EFI/UpdateCapsule, U-Boot causes a synchronous
abort.

This also fixes use-after-free bug of the 'capsule' variable.

Fixes: c74cd8bd08d1 ("efi_loader: capsule: add capsule_on_disk support")
Signed-off-by: Masami Hiramatsu <masami.hiramatsu@linaro.org>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>

efi_loader: Reduce efi_tcg2 logging statement

log_info() is used for the debug level logging statement
which should use log_debug() instead. Convert it to reduce the
log output.

Signed-off-by: Masahisa Kojima <masahisa.kojima@linaro.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>

efi_loader: Sphinx comments in efi_api.h

Fix incorrect Sphinx comments in efi_api.h:

* add missing 'struct'
* correct indentation

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>

efi: add comment for efi_system_table and efi_configuration_table

This commit adds the comment for efi_system_table and
efi_configuration_table structure.

Signed-off-by: Masahisa Kojima <masahisa.kojima@linaro.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>

doc: fix typos in trace.rst

Fix obvious typos. Use US spelling consistently.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>

scripts: update spelling.txt from upstream Linux

This list is used by checkpatch.pl. The Linux v5.15 version has several
words that where mispelled in U-Boot too.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>

test: test truncation in snprintf()

Test that the return value of snprintf() is correct in the case of
truncation.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

lmb: fix typo 'commun'

%s/commun/common/

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>

lmb: drop unused lmb_size_bytes()

lmb_size_bytes() is unused.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>

lmb: remove extern keyword in lmb.h

The extern keyword is not needed in include/lmb.h to declare functions.
Remove it.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>

doc: add include/lmb.h to the HTML documentation

Correct Sphinx style comments in include/lmb.h

Add the logical memory block API to the HTML documentation.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>

lmb: remove lmb_is_nomap() from include

Defining static functions in includes should be avoided.
Function lmb_is_nomap() is only used in the unit test.
So move it to the unit test.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>

Merge tag 'u-boot-amlogic-20211119' of https://source.denx.de/u-boot/custodians/u-boot-amlogic

- pinctrl: Correct the driver GPIO declaration
- meson64_android: handle errors on boot and run fastboot on boot failure

Merge tag 'tpm-19112021' of https://source.denx.de/u-boot/custodians/u-boot-tpm

TPM1.2 and Atmel fixes

# gpg verification failed.

configs: meson64_android: add PANIC stage for SYSTEM fails

If bootloader was updated without running oem format, reboot will cause
boot loop because the SYSTEM stage fails.

Add a final PANIC stage running fastboot to permit recovery.

Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Mattijs Korpershoek <mkorpershoek@baylibre.com>

configs: meson64_android: bypass other checks on run_fastboot=1

This can lead to GPT and BCB errors even if fastboot was selected early
by usb rom boot and the eMMC is blank/invalid.

Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Mattijs Korpershoek <mkorpershoek@baylibre.com>

pinctrl: meson: Correct the driver GPIO declaration

This should use the provided U_BOOT_DRIVER() macro so that the driver gets
added to the appropriate linker list. Fix it.

Signed-off-by: Simon Glass <sjg@chromium.org>
Fixes: 7c9dcfed50f ("pinctrl: meson: rework gx pmx function")
Reported-by: Tom Rini <trini@konsulko.com>
Tested-by: Tom Rini <trini@konsulko.com> on libretech-cc
Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>

Merge branch '2021-11-18-regression-fixes'

- An assortment of fixes related to GD, GD_FLG_SKIP_RELOC, and the lmb
- Environment fix on synquacer developmentbox
- Fix for get_info is not valid in partition code

arm64: Add missing GD_FLG_SKIP_RELOC handling

In case U-Boot enters relocation with GD_FLG_SKIP_RELOC, skip the
relocation. The code still has to set up new_gd pointer and new
stack pointer.

Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Tom Rini <trini@konsulko.com>

lmb: Reserve U-Boot separately if relocation is disabled

In case U-Boot starts with GD_FLG_SKIP_RELOC, the U-Boot code is
not relocated, however the stack and heap is at the end of DRAM
after relocation. Reserve a LMB area for the non-relocated U-Boot
code so it won't be overwritten.

Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Tom Rini <trini@konsulko.com>

board_f: Copy GD to new GD even if relocation disabled

Even if U-Boot has relocation disabled via GD_FLG_SKIP_RELOC , the
relocated stage of U-Boot still picks GD from new_gd location. The
U-Boot itself is not relocated, but GD might be, so copy the GD to
new GD location even if relocation is disabled.

Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com>
Cc: Peng Fan <peng.fan@oss.nxp.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Tom Rini <trini@konsulko.com>

part: return -ENOSYS when get_info not valid.

In some case, get_info() interface can be NULL, add this check to stop
from crash.

Signed-off-by: schspa <schspa@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

Revert "board: synquacer: developerbox: Don't set gd->env_addr to default_environment"

Without default setting of gd->env_addr, U-Boot will cause
a synchronous abort if the env-variables on the SPI flash is
broken or not saved corectly. Set gd->env_addr correctly.

This reverts commit 535870f3b0fb09ee9b2885409f05304111464643.

Signed-off-by: Masami Hiramatsu <masami.hiramatsu@linaro.org>
Reviewed-by: Marek Behún <marek.behun@nic.cz>

drivers: tpm: atmel_twi: fix printf specifier compile warning

%d was being used as the specifier for size_t, leading to a
compiler warning

Signed-off-by: Mathew McBride <matt@traverse.com.au>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>

drivers: tpm: atmel_twi: implement get_desc operation

Without get_desc, the tpm command will not provide a
description of the device in 'tpm device' or 'tpm info'.

Due to the characteristics of the Atmel TPM it isn't
possible to determine certain attributes (e.g open/close
status) without using the TPM stack (compare Infineon
and ST TPM drivers), so just print out the chip model
and udevice name as the identifier.

Signed-off-by: Mathew McBride <matt@traverse.com.au>
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>

drivers: tpm: atmel_twi: do not use an offset byte

This driver was broken due to an empty offset byte being prepended
at the start of every transmission.

The hardware does not mimic an EEPROM device with registers so
an offset byte is not required.

Signed-off-by: Mathew McBride <matt@traverse.com.au>
Acked-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>

drivers: tpm: atmel_twi: drop non-DM_I2C compatibility

There are no users of this driver without DM_I2C

Signed-off-by: Mathew McBride <matt@traverse.com.au>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>

cmd: tpm-v1: fix load_key_by_sha1 compile errors

This command is not compiled by default and has not been updated alongside
changes to the tpmv1 API, such as passing the TPM udevice to the relevant
functions.

Signed-off-by: Mathew McBride <matt@traverse.com.au>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>

cmd: tpm-v1: fix compile error in TPMv1 list resources command

This command is not compiled by default and was not
updated to pass the udevice to tpm_get_capability.

Signed-off-by: Mathew McBride <matt@traverse.com.au>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>

Merge branch '2021-11-15-assorted-fixes'

- Rename "tqc" to "tq" and related updates
- Assorted minor ARM updates, build updates and documentation updates

Merge tag 'xilinx-for-v2022.01-rc3' of https://source.denx.de/u-boot/custodians/u-boot-microblaze

Xilinx changes for v2022.01-rc3

sdhci:
- Fix emmc mini case with missing firmware interface

zynqmp:
- Restore JTAG interface if required
- Allow overriding board name
- Add support for DLC21
- Fix one fallthrought statement description
- Use config macro instead of name duplication
- Save multiboot to variable

firmware:
- Handle ipi_req errors better
- Use local buffer in case user doesn't need it instead of NULL/0 location

spi:
- gqsi: Fix write issue at low frequencies

net:
- gem: Disable broadcasts

image: Explicitly declare do_bdinfo()

Compiler is not happy:

common/image-board.c: In function ‘boot_get_kbd’:
common/image-board.c:902:17: warning: implicit declaration of function ‘do_bdinfo’ [-Wimplicit-function-declaration]
  902 |                 do_bdinfo(NULL, 0, 0, NULL);
      |                 ^~~~~~~~~

Move the forward declaration to a header.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

usb: doc: Fix spelling issues in README.usb

Fix spelling issues in README.usb.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

arm: Fix bad memcpy.S str8w macro argument count

Remove the extra (empty) argument passed to str8w, causing the following
error:

   <instantiation>:40:47: error: too many positional arguments
     str8w r0, r3, r4, r5, r6, r7, r8, r9, ip, , abort=19f
                                                 ^
   u-boot/arch/arm/lib/memcpy.S:240:5: note: while in macro instantiation
   17: forward_copy_shift pull=16 push=16
       ^

Note: no functional change intended.

Fixes: d8834a1323af ("arm: Use optimized memcpy and memset from linux")
Signed-off-by: Pierre-Clément Tosi <ptosi@google.com>

configs: synquacer: Fix dfu_alt_info to use nor1

Fix dfu_alt_info to use nor1 instead of the device name.
This reverts a part of commit 59bd18d4c4d7 ("configs: synquacer:
Remove mtdparts settings and update DFU setting") because the
commit a4f2d8341455 ("mtd: spi: nor: force mtd name to "nor%d"")
changed the mtd device naming scheme to nor%d.

Signed-off-by: Masami Hiramatsu <masami.hiramatsu@linaro.org>
Acked-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>

makefile: add missing semicolons

On some distributions, as Debian GNU 11, this targets fails
with errors.

Signed-off-by: Angelo Dureghello <angelo.dureghello@timesys.com>

board: tq: fix spelling of "TQ-Systems"

"TQ-Systems" is written with a dash.

Signed-off-by: Matthias Schiffer <matthias.schiffer@ew.tq-group.com>

board: rename "tqc" vendor to "tq"

The subdivision name "TQ Components" hasn't been in use for a long time.
Rename the vendor directory to "tq", which also matches our Device Tree
vendor prefix.

Signed-off-by: Matthias Schiffer <matthias.schiffer@ew.tq-group.com>

arm64: relocate-rela: Add support for ld.lld

Cap end of relocations by the binary size.

Linkers like to insert some auxiliary sections between .rela.dyn and
.bss_start. These sections don't make their way to the final binary, but
reloc_rela still tries to relocate them, resulting in attempted read
past the end of file.

When linking U-Boot with ld.lld, the STATIC_RELA feature (enabled by
default on arm64) breaks the build. After this patch, U-Boot can be
linked successfully with and without CONFIG_STATIC_RELA.

Originally-from: Elena Petrova <lenaptr@google.com>
Signed-off-by: Alistair Delva <adelva@google.com>
Cc: David Brazdil <dbrazdil@google.com>
Cc: Scott Wood <scottwood@freescale.com>
Cc: Tom Rini <trini@konsulko.com>

Prepare v2022.01-rc2

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

usb: Make USB_MUSB_PIO_ONLY selected by USB_MUSB_SUNXI

This ensures the USB_MUSB_PIO_ONLY config is set to an apppropriate
value from the changes enabling USB_MUSB_GADGET does.

Namely, USB_MUSB_PIO_ONLY default to =y on USB_MUSB_SUNXI being y.

Signed-off-by: Samuel Dionne-Riel <samuel@dionne-riel.com>

xilinx: zynqmp: fix ZYNQMP_RESTORE_JTAG check

Config check should be done without the SPL_ prefix.

Signed-off-by: Ricardo Salveti <ricardo@foundries.io>
Reviewed-by: Oleksandr Suvorov <oleksandr.suvorov@foundries.io>
Link: https://lore.kernel.org/r/20211104192802.3093811-1-ricardo@foundries.io
Signed-off-by: Michal Simek <michal.simek@xilinx.com>

Dockerfile, CI: Update to latest "focal" tag

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