Pull request for efi-2022-07-rc4-3

UEFI:

* fix a problem in loading an image from a short-path
* fix building the bootmenu command for CONFIG_EFI_LOADER=n
* correct the bootefi command syntax
* add firmware management protocol to the documentation

Others:

* bootmenu: fix bootmenu title handling

Tested-by: Pali Rohár <pali@kernel.org> [n900, for bootmenu working as before]

The commit a3d0aa87 ("bootmenu: update bootmenu_entry structure")
changes the bootmenu title type from char to u16(UTF16 string)
to support EFI based system. If EFI_LOADER is not enabled,
printf("%ls") is not supported, so bootmenu does not appear
correctly.

This commit changes the type of menu title from u16(UTF16) to
utf-8 string and EFI strings is conveted into utf-8.

Fixes: a3d0aa87 ("bootmenu: update bootmenu_entry structure")
Signed-off-by: NMasahisa Kojima <masahisa.kojima@linaro.org>
Tested-by: NPali Rohar <pali@kernel.org>
Reviewed-by: NHeinrich Schuchardt <heinrich.schuchardt@canonical.com>

The firmware management protocol can be used to manage device firmware.
U-Boot can be configured to provide an implementation.

Document the related functions in the API section.
Signed-off-by: NVincent Stehlé <vincent.stehle@arm.com>
Reviewed-by: NHeinrich Schuchardt <heinrich.schuchardt@canonical.com>

This fixes the following warnings:

  ./lib/efi_loader/efi_firmware.c:283: warning: Function parameter or member 'package_version' not described in 'efi_firmware_fit_get_image_info'
  ./lib/efi_loader/efi_firmware.c:283: warning: Function parameter or member 'package_version_name' not described in 'efi_firmware_fit_get_image_info'
  ./lib/efi_loader/efi_firmware.c:369: warning: bad line: firmware image
  ./lib/efi_loader/efi_firmware.c:395: warning: Function parameter or member 'package_version' not described in 'efi_firmware_raw_get_image_info'
  ./lib/efi_loader/efi_firmware.c:395: warning: Function parameter or member 'package_version_name' not described in 'efi_firmware_raw_get_image_info'
Signed-off-by: NVincent Stehlé <vincent.stehle@arm.com>
Cc: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: NHeinrich Schuchardt <heinrich.schuchardt@canonical.com>

This commit fixes the following compile warnings
for the documentation.

./include/charset.h:276: warning: Function parameter or member 'size' not described in 'u16_strlcat'
./include/charset.h:276: warning: Excess function parameter 'count' description in 'u16_strlcat'
Signed-off-by: NMasahisa Kojima <masahisa.kojima@linaro.org>
Reviewed-by: NHeinrich Schuchardt <heinrich.schuchardt@canonical.com>

Booting from a short-form device path which starts with the first element
being a File Path Media Device Path failed because it doesn't contain
any valid device with simple file system protocol and efi_dp_find_obj()
in efi_load_image_from_path() will return NULL.
For instance,
/VenHw(e61d73b9-a384-4acc-aeab-82e828f3628b)/Scsi(0,0)/\helloworld.efi
-> shortened version: /\helloworld.efi

With this patch applied, all the media devices with simple file system
protocol are enumerated and the boot manager attempts to boot temporarily
generated device paths one-by-one.

This new implementation is still a bit incompatible with the UEFI
specification in terms of:
* not creating real boot options
* not try
  "If a device does not support the EFI_SIMPLE_FILE_SYSTEM_PROTOCOL, but
  supports the EFI_BLOCK_IO_PROTOCOL protocol, then the EFI Boot Service
  ConnectController must be called for this device with DriverImageHandle
  and RemainingDevicePath set to NULL and the Recursive flag is set to TRUE."
(See section 3.1.2 "Load Option Processing".)

But it still gives us a closer and better solution than the current.

Fixes: commit 9cdf4702 ("efi_loader: support booting via short-form device-path")
Signed-off-by: NAKASHI Takahiro <takahiro.akashi@linaro.org>
Reviewed-by: NHeinrich Schuchardt <heinrich.schuchardt@canonical.com>

This helper function will be used to determine if the device is
removable media, initially for handling a short-path loading.
Signed-off-by: NAKASHI Takahiro <takahiro.akashi@linaro.org>
Reviewed-by: NHeinrich Schuchardt <xypron.glpk@gmx.de>

For indicating the address and size of a memory region other commands use a
<addr>[:<size>] syntax. Do the same for bootefi.

Fixes: 20589836 ("cmd: bootefi: restore ability to boot arbitrary blob")
Signed-off-by: NHeinrich Schuchardt <heinrich.schuchardt@canonical.com>

Changing the console timeout to 500 ms without restoring the original value
leads to failures in other tests. As the console timeout change is not
necessary for the text input protocol tests remove it.
Signed-off-by: NHeinrich Schuchardt <heinrich.schuchardt@canonical.com>

* change format of parameter documentation
* avoid superfluous assignments
Signed-off-by: NHeinrich Schuchardt <heinrich.schuchardt@canonical.com>

The commit 2158b0da ("bootmenu: add Kconfig option
not to enter U-Boot console") disables to enter U-Boot
console from bootmenu as default, this change affects the
existing bootmenu users.

This commit reverts the default behavior, the bootmenu can
enter U-Boot console same as before.
CMD_BOOTMENU_ENTER_UBOOT_CONSOLE is renamed
BOOTMENU_DISABLE_UBOOT_CONSOLE and depends on
AUTOBOOT_MENU_SHOW.

Fixes: 2158b0da ("bootmenu: add Kconfig option not to enter U-Boot console")
Signed-off-by: NMasahisa Kojima <masahisa.kojima@linaro.org>
Tested-by: NPali Rohar <pali@kernel.org>
Reviewed-by: NIlias Apalodimas <ilias.apalodimas@linaro.org>
Reviewed-by: NHeinrich Schuchardt <heinrich.schuchardt@canonical.com>

This documents some additional options which can be used with valgrind, as
well as directions for future work. It also fixes up inline literals to
actually be inline literals (and not italics). The content of this
documentation is primarily adapted from [1].

[1] https://lore.kernel.org/u-boot/57cb4b49-fa30-1194-9ac3-faa53e8033bd@gmail.com/Signed-off-by: NSean Anderson <seanga2@gmail.com>
Reviewed-by: NHeinrich Schuchardt <heinrich.schuchardt@canonical.com>

The following Python packages are used by U-Boot, too:

* python3-asteval
* python3-subunit
* python3-testtools
Signed-off-by: NHeinrich Schuchardt <heinrich.schuchardt@canonical.com>

Image created by LTO is not friendly to debugger, let's document this.
Signed-off-by: NBin Meng <bmeng.cn@gmail.com>

Mention CONFIG_CC_OPTIMIZE_FOR_DEBUG and LLDB.
Reviewed-by: NHeinrich Schuchardt <xypron.glpk@gmx.de>

It should be CONFIG_SANDBOX_RAM_SIZE_MB.
Signed-off-by: NBin Meng <bmeng.cn@gmail.com>
Reviewed-by: NHeinrich Schuchardt <heinrich.schuchardt@canonical.com>

There are no defined instruction sequences in include/valgrind.h for
Risc-V, so CONFIG_VALGRIND will do nothing on this arch (and possibly won't
compile?). Update Kconfig accordingly.
Signed-off-by: NSean Anderson <seanga2@gmail.com>
Reviewed-by: NHeinrich Schuchardt <heinrich.schuchardt@canonical.com>

In the bootmenu unit test the console timeout is set to 500 ms.
Other tests rely on the original timeout. Ensure that the original value
is restored.
Signed-off-by: NHeinrich Schuchardt <heinrich.schuchardt@canonical.com>

alloc_read_gpt_entries() writes differentiated error messages.
The caller is_gpt_valid() should not write an extra possibly wrong message.
Signed-off-by: NHeinrich Schuchardt <heinrich.schuchardt@canonical.com>

- Fixes for pytest timeout in CI, missing dependency on PCI for the
  e1000 driver, fix for CVE-2022-30767 (NFS), TI K3 AM642 DTS bugfix,
  MAINTAINERS updates, mksquashfs version check fix.

This driver depends on PCI. Update the Kconfig accordingly.
Signed-off-by: NSean Anderson <sean.anderson@seco.com>
Reviewed-by: NTim Harvey <tharvey@gateworks.com>
Reviewed-by: NStefan Roese <sr@denx.de>
Reviewed-by: NRamon Fried <rfried.dev@gmail.com>

Testing with mksquasshfs 4.5.1 results in an error

    ValueError: could not convert string to float: '4.5.1'

Version 4.10 would be considered to be lower than 4.4.

Fixes: 04c9813e ("test/py: rewrite common tools for SquashFS tests")
Signed-off-by: NHeinrich Schuchardt <heinrich.schuchardt@canonical.com>

Update MAINTAINERS List for LS2088ARDB and LS2088AQDS
platforms
Signed-off-by: NWasim Khan <wasim.khan@nxp.com>

Since commit dffdb1f8 ("board: ti: am64x: Use fdt functions
for ram and bank init") ddr_init() and dram_bank_init() have
switched to fdtdec for getting the memory configuration from
the am64xx dts files instead of using hardcoded values. This
requires an accessible memory node in SPL as we already have
in k3-am642-r5-evm.dts.

Make the memory node accessible in A53 SPL for both am642-sk
and am642-evm and in am642-sk R5 SPL.
Signed-off-by: NGeorgi Vlaev <g-vlaev@ti.com>
Reviewed-by: NTom Rini <trini@konsulko.com>

Add the rest of the ASPEED drivers that are in tree. Most are obvious,
except for ftgmac100 which matches the register layout used in the
ASPEED SoC.

I am the Linux maintainer for the ASPEED kernel port, and help maintain
the fork of u-boot used for OpenBMC, so add myself as a reviewer so I
can stay informed about u-boot changes.
Signed-off-by: NJoel Stanley <joel@jms.id.au>
Reviewed-by: NChia-Wei Wang <chiawei_wang@aspeedtech.com>

This patch mitigates the vulnerability identified via CVE-2019-14196.

The previous patch was bypassed/ineffective, and now the vulnerability
is identified via CVE-2022-30767. The patch removes the sanity check
introduced to mitigate CVE-2019-14196 since it's ineffective.
filefh3_length is changed to unsigned type integer, preventing negative
numbers from being used during comparison with positive values during
size sanity checks.
Signed-off-by: NAndrea zi0Black Cappa <zi0Black@protonmail.com>

Use test_part_types as the name instead of dm_compact.
Signed-off-by: NBin Meng <bmeng.cn@gmail.com>

Use test_fstypes as the name instead of test_dm_compact.
Signed-off-by: NBin Meng <bmeng.cn@gmail.com>

Reset the console timeout value as some tests may change its default
value during the execution.

This fixes the random case timeout issue seen in the U-Boot CI.
Signed-off-by: NBin Meng <bmeng.cn@gmail.com>

Set kernel_comp_addr_r and kernel_comp_size for compressed kernel.
Adjust existing addresses for ramdisk, so that kernel_comp_addr_r
comes before the ramdisk image, since the decompressed kernel size
is known to us. This way we can allow big ramdisk image to be loaded.
Signed-off-by: NBin Meng <bmeng.cn@gmail.com>
Reviewed-by: NLeo Yu-Chi Liang <ycliang@andestech.com>

Set kernel_comp_addr_r and kernel_comp_size for compressed kernel.
Adjust existing addresses for ramdisk, so that kernel_comp_addr_r
comes before the ramdisk image, since the decompressed kernel size
is known to us. This way we can allow big ramdisk image to be loaded.

Update unleashed.rst to remove the manual environment configuration
for compressed kernel boot.
Signed-off-by: NBin Meng <bmeng.cn@gmail.com>
Reviewed-by: NLeo Yu-Chi Liang <ycliang@andestech.com>

Move kernel_comp_addr_r to an address that comes before the ramdisk
image, since the decompressed kernel size is known to us. This way
we can allow big ramdisk image to be loaded.
Signed-off-by: NBin Meng <bmeng.cn@gmail.com>
Reviewed-by: NLeo Yu-Chi Liang <ycliang@andestech.com>

Clean up asm/io.h by
- removing commented code
- removing outdated comments
- removing unused definitions (for mem_isa, mem_pci)

This massively improves the readability of the file.

Suggested by commits:
7ab2e47d ("arm: Clean up asm/io.h")
909d0399 ("ARM: asm/io.h: kill off confusing #ifdef __mem_pci block")
Signed-off-by: NLeo Yu-Chi Liang <ycliang@andestech.com>
Reviewed-by: NBin Meng <bmeng.cn@gmail.com>

Commit ba1ed5b0 ("Convert CONFIG_ARCH_MAP_SYSMEM to Kconfig") clearly
defined that this option is available for SANDBOX (was also for already
removed NDS32). That's why there is no way how this code can be enabled
with current Kconfig layout for riscv.
Based on this removing this code.
Signed-off-by: NMichal Simek <michal.simek@amd.com>
Reviewed-by: NBin Meng <bmeng.cn@gmail.com>

The sbi command displays the ID of the implementation of the RISC-V
Supervisor Binary Interface Specification. A new ID for Coffer has recently
been added.
Signed-off-by: NHeinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: NLeo Yu-Chi Liang <ycliang@andestech.com>

Disable BINMAN_FDT for ae350 boards which don't actually use it.

Fixes: 836eac7c ("fdt: Make OF_BOARD a bool option")
Signed-off-by: NRick Chen <rick@andestech.com>
Reviewed-by: NLeo Yu-Chi Liang <ycliang@andestech.com>

Enable OF_HAS_PRIOR_STAGE for ae350 boards with OF_BOARD

Fixes: 239d22c7 ("fdt: Enable OF_HAS_PRIOR_STAGE for most boards with OF_BOARD")
Signed-off-by: NRick Chen <rick@andestech.com>
Reviewed-by: NLeo Yu-Chi Liang <ycliang@andestech.com>

Xilinx changes for v2022.07-rc4

zynqmp:
- Fix DP PLL configuration for zcu102/zcu106 and SOM
- Fix split mode for starting R5s
- DT fixes
- Remove firmware node for mini configurations
- Wire TEE for multi DTB fit image

xilinx:
- Handle board_get_usable_ram_top(0) properly

phy:
- Extend psgtr timeout

mmc:
- Fix mini configuration which misses zynqmp_pm_is_function_supported()

The bulk of it is (finally!) some DT sync from the kernel. We refrained
from syncing one incompatible change, as this would spoil booting Linux
kernels before v5.13 with U-Boot's DT (via UEFI, for instance).
I test booted Linux v5.18 and v5.4 with that new DT on some boards, and
the headless peripherals (MMC, USB, Ethernet) seemed to work.
The rest are fixes:
- silencing missing clock warnings due to the new pinctrl driver
- fixing "UART0 on PortF", allowing UART access through the SD card pins
- add an F1C100s clock driver, to enable MMC support (SPI comes later)
- some cleanups for CONS_INDEX_n in Kconfig

Tested on BananaPi-M1, Pine64-LTS, Pine-H64, X96-Mate (H616) and
OrangePi-Zero.

Add support for i2c devices