]> git.dujemihanovic.xyz Git - u-boot.git/log
u-boot.git
2 years agotools: kwboot: Update references with public links
Pali Rohár [Wed, 2 Mar 2022 10:49:27 +0000 (11:49 +0100)]
tools: kwboot: Update references with public links

Public documents about BootROM of some Marvell SoCs are available in the
public Web Archive. Put this information into source code.

Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Stefan Roese <sr@denx.de>
Tested-by: Stefan Roese <sr@denx.de>
2 years agotools: kwboot: Update doc about Avanta
Pali Rohár [Wed, 2 Mar 2022 10:49:26 +0000 (11:49 +0100)]
tools: kwboot: Update doc about Avanta

Testes proved that current kwboot version supports also Avanta SoCs.
It looks like that Avanta SoCs are using same kwbimage format as Armada.

Signed-off-by: Pali Rohár <pali@kernel.org>
Tested-by: Tony Dinh <mibodhi@gmail.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Tested-by: Stefan Roese <sr@denx.de>
2 years agotools: kwboot: Update manpage
Pali Rohár [Wed, 2 Mar 2022 10:49:25 +0000 (11:49 +0100)]
tools: kwboot: Update manpage

Document -D, -b, -d, -q and -s options.

Add common examples how to use kwboot.

Add information about Armada 38x BootROM bug for debug console mode and how
to workaround it.

Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Stefan Roese <sr@denx.de>
Tested-by: Stefan Roese <sr@denx.de>
2 years agotools: kwboot: Update usage
Pali Rohár [Wed, 2 Mar 2022 10:49:24 +0000 (11:49 +0100)]
tools: kwboot: Update usage

Add all supported Armada SoCs and document -b and -d options in usage.

Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Stefan Roese <sr@denx.de>
Tested-by: Stefan Roese <sr@denx.de>
2 years agotools: kwboot: Add support for backspace key in mini terminal
Pali Rohár [Wed, 2 Mar 2022 10:49:23 +0000 (11:49 +0100)]
tools: kwboot: Add support for backspace key in mini terminal

Marvell BootROM recognize only '\b' byte as backspace. Use terminfo
for retrieving current backspace sequence and replace any occurrence of
backspace sequence by the '\b' byte.

Reading terminfo database is possible via tigetstr() function from system
library libtinfo.so.*. So link kwboot with -ltinfo.

Normally terminfo functions are in <term.h> system header file. But this
header file conflicts with U-Boot "termios_linux.h" header file. So declare
terminfo functions manually.

Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Stefan Roese <sr@denx.de>
Tested-by: Stefan Roese <sr@denx.de>
2 years agotools: kwboot: Fix sending and processing debug message pattern (-d option)
Pali Rohár [Wed, 2 Mar 2022 10:49:22 +0000 (11:49 +0100)]
tools: kwboot: Fix sending and processing debug message pattern (-d option)

-d option is currently broken. In most cases BootROM does not detect this
message pattern. For sending debug message pattern it is needed to do same
steps as for boot message pattern.

Implement sending debug message pattern via same separate thread like it is
for boot message pattern.

Checking if BootROM entered into UART debug mode is different than
detecting UART boot mode. When in boot mode, BootROM sends xmodem NAK
bytes. When in debug mode, BootROM activates console echo and reply back
every written byte (extept \r\n which is interpreted as executing command
and \b which is interpreting as removing the last sent byte).

So in kwboot, check that BootROM send back at least 4 debug message
patterns as a echo reply for debug message patterns which kwboot is sending
in the loop.

Then there is another observation, if host writes too many bytes (as
command) then BootROM command line buffer may overflow after trying to
execute such long command. To workaround this overflow, it is enough to
remove bytes from the input line buffer by sending 3 \b bytes for every
sent character. So do it.

With this change, it is possbile to enter into the UART debug mode with
kwboot -d option.

Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Stefan Roese <sr@denx.de>
Tested-by: Stefan Roese <sr@denx.de>
2 years agotools: kwboot: Use separate thread for sending boot message pattern
Pali Rohár [Wed, 2 Mar 2022 10:49:21 +0000 (11:49 +0100)]
tools: kwboot: Use separate thread for sending boot message pattern

After BootROM successfully detects boot message pattern on UART it waits
until host stop sending data on UART. For example Armada 385 BootROM
requires that host does not send anything on UART at least 24 ms. If host
is still sending something then BootROM waits (possibly infinitely).

BootROM successfully detects boot message pattern if it receives it in
small period of time after power on.

So to ensure that host put BootROM into UART boot mode, host must send
continuous stream of boot message pattern with a small gap (for A385 at
least 24 ms) after series of pattern. But this gap cannot be too often or
too long to ensure that it does not cover whole BootROM time window when it
is detecting for boot message pattern.

Therefore it is needed to do following steps in cycle without any delay:
1. send series of boot message pattern over UART
2. wait until kernel transmit all data
3. sleep small period of time

At the same time, host needs to monitor input queue, data received on the
UART and checking if it contains NAK byte by which BootROM informs that
xmodem transfer is ready.

But it is not possible to wait until kernel transmit all data on UART and
at the same time in the one process to also wait for input data. This is
limitation of POSIX tty API and also by linux kernel that it does not
provide asynchronous function for waiting until all data are transmitted.
There is only synchronous variant tcdrain().

So to correctly implement this handshake on systems with linux kernel, it
is needed to use tcdrain() in separate thread.

Implement sending of boot message pattern in one thread and reading of
reply in the main thread. Use pthread library for threads.

This change makes UART booting on Armada 385 more reliable. It is possible
to start kwboot and power on board after minute and kwboot correctly put
board into UART boot mode.

Old implementation without separate thread has an issue that it read just
one byte from UART input queue and then it send 128 message pattern to the
output queue. If some noise was on UART then kwboot was not able to read
BootROM response as its input queue was just overflowed and kwboot was
sending more data than receiving.

This change basically fixed above issue too.

Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Stefan Roese <sr@denx.de>
Tested-by: Stefan Roese <sr@denx.de>
2 years agotools: kwboot: Cleanup bootmsg and debugmsg variables
Pali Rohár [Wed, 2 Mar 2022 10:49:20 +0000 (11:49 +0100)]
tools: kwboot: Cleanup bootmsg and debugmsg variables

Function kwboot_debugmsg() is always called with kwboot_msg_debug as msg
and function kwboot_bootmsg() with kwboot_msg_debug as msg. Function
kwboot_bootmsg() is never called with NULL msg.

Simplify, cleanup and remove dead code.

No functional change.

Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Stefan Roese <sr@denx.de>
Tested-by: Stefan Roese <sr@denx.de>
2 years agotools: kwboot: Remove msg_req_delay
Pali Rohár [Wed, 2 Mar 2022 10:49:19 +0000 (11:49 +0100)]
tools: kwboot: Remove msg_req_delay

Variable msg_req_delay is set but never used. So completely remove it.

Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Stefan Roese <sr@denx.de>
Tested-by: Stefan Roese <sr@denx.de>
2 years agotools: kwboot: Check for return value of kwboot_tty_send() and tcflush()
Pali Rohár [Wed, 2 Mar 2022 10:49:18 +0000 (11:49 +0100)]
tools: kwboot: Check for return value of kwboot_tty_send() and tcflush()

Failure of kwboot_tty_send() and tcflush() functions is fatal, it does not
make sense to continue. So return error back to the caller like in other
places where are called these functions.

Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Stefan Roese <sr@denx.de>
Tested-by: Stefan Roese <sr@denx.de>
2 years agoARM: mvebu: x530: clearfog: Add ODT configuration
Chris Packham [Tue, 1 Mar 2022 00:53:23 +0000 (13:53 +1300)]
ARM: mvebu: x530: clearfog: Add ODT configuration

Commit 369e532691e0 ("ddr: marvell: a38x: allow board specific ODT
configuration") added the odt_config member to struct
mv_ddr_topology_map ahead of the clk_enable and ck_delay members. This
means that any boards that configured either of clk_enable or ck_delay
needed to have their board topology updated. This affects the x530 and
clearfog boards. Other A38x boards don't touch any of the trailing
members of mv_ddr_topology_map so don't need updating.

Fixes: 369e532691e0 ("ddr: marvell: a38x: allow board specific ODT configuration")
Signed-off-by: Chris Packham <judge.packham@gmail.com>
Acked-by: Baruch Siach <baruch@tkos.co.il>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agoarm64: a37xx: pinctrl: Fix PWM pins indexes
Marek Behún [Mon, 28 Feb 2022 14:59:37 +0000 (15:59 +0100)]
arm64: a37xx: pinctrl: Fix PWM pins indexes

Commit 5534fb4f4833 ("arm64: a37xx: pinctrl: Correct PWM pins
definitions") introduced bogus definitions os PWM pins: all 4 pins have
index 11, instead of having indexes 11, 12, 13, 14.

Fix this.

Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agodrivers: rtc: fix null pointer access in armada38x_rtc_reset
Francois Berder [Mon, 28 Feb 2022 09:31:45 +0000 (10:31 +0100)]
drivers: rtc: fix null pointer access in armada38x_rtc_reset

Replace null pointer by pointer to device registers when calling
armada38x_rtc_write.

Signed-off-by: Francois Berder <fberder@outlook.fr>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agoarm: a37xx: pci: Fix a3700_fdt_fix_pcie_regions() function again
Pali Rohár [Wed, 23 Feb 2022 12:52:32 +0000 (13:52 +0100)]
arm: a37xx: pci: Fix a3700_fdt_fix_pcie_regions() function again

The a3700_fdt_fix_pcie_regions() function still computes nonsense.

It computes the fixup offset from the PCI address taken from the first
row of the "ranges" array, which means that:
- PCI address must equal CPU address (otherwise the computed fix offset
  will be wrong),
- the first row must contain the lowest address.

This is the case for the default device-tree, which is why we didn't
notice it.

It also adds the fixup offset to all PCI and CPU addresses, which is
wrong.

Instead:
1) The fixup offset must be computed from the CPU address, not PCI
   address.

2) The fixup offset must be computed from the row containing the lowest
   CPU address, which is not necessarily contained in the first row.

3) The PCI address - the address to which the PCIe controller remaps the
   address space as seen from the point of view of the PCIe device -
   must be fixed by the fix offset in the same way as the CPU address
   only in the special case when the CPU adn PCI addresses are the same.
   Same addresses means that remapping is disabled, and thus if we
   change the CPU address, we need also to change the PCI address so
   that the remapping is still disabled afterwards.

Consider an example:
  The ranges entries contain:
    PCI address   CPU address
    70000000      EA000000
    E9000000      E9000000
    EB000000      EB000000

  By default CPU PCIe window is at:        E8000000 - F0000000
  Consider the case when TF-A moves it to: F2000000 - FA000000

  Until now the function would take the PCI address of the first entry:
  70000000, and the new base, F2000000, to compute the fix offset:
  F2000000 - 70000000 = 82000000, and then add 8200000 to all addresses,
  resulting in
    PCI address   CPU address
    F2000000      6C000000
    6B000000      6B000000
    6D000000      6D000000
  which is complete nonsense - none of the CPU addresses is in the
  requested window.

  Now it will take the lowest CPU address, which is in second row,
  E9000000, and compute the fix offset F2000000 - E9000000 = 09000000,
  and then add it to all CPU addresses and those PCI addresses which
  equal to their corresponding CPU addresses, resulting in
    PCI address   CPU address
    70000000      F3000000
    F2000000      F2000000
    F4000000      F4000000
  where all of the CPU addresses are in the needed window.

Fixes: 4a82fca8e330 ("arm: a37xx: pci: Fix a3700_fdt_fix_pcie_regions() function")
Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agoarm: mvebu: turris_omnia: Enable ext4 write support in defconfig
Marek Behún [Fri, 18 Feb 2022 16:46:25 +0000 (17:46 +0100)]
arm: mvebu: turris_omnia: Enable ext4 write support in defconfig

Enable ext4 write support in Turris Omnia's defconfig. Some users find
it useful.

Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agopci: pci_mvebu: Cleanup macro names
Pali Rohár [Fri, 18 Feb 2022 11:25:23 +0000 (12:25 +0100)]
pci: pci_mvebu: Cleanup macro names

Use "MVPCIE_" prefix instead of generic "PCIE_" prefix for pci_mvebu.c
specific macros. Define offset macros for Root Port registers and use
standard register macros from pci.h when accessing Root Port registers.

Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agopci: pci_mvebu: Remove unused SELECT and lane_mask
Pali Rohár [Fri, 18 Feb 2022 11:25:22 +0000 (12:25 +0100)]
pci: pci_mvebu: Remove unused SELECT and lane_mask

Macro SELECT() is unused and struct mvebu_pcie field lane_mask is unused
too. Remove them.

Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agotools: kwboot: Fix quitting terminal
Pali Rohár [Fri, 18 Feb 2022 11:24:13 +0000 (12:24 +0100)]
tools: kwboot: Fix quitting terminal

Sometimes kwboot after quitting terminal prints error message:

  terminal: Bad address

This is caused by trying to call write() syscall with count of (size_t)-1
bytes.

When quit sequence is split into more read() calls then number of input
bytes (nin) at the end of cycle can underflow and be negative. Fix it.

Fixes: de7514046ea5 ("tools: kwboot: Fix detection of quit esc sequence")
Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Stefan Roese <sr@denx.de>
2 years agoMerge tag 'u-boot-at91-fixes-2022.04-a' of https://source.denx.de/u-boot/custodians...
Tom Rini [Thu, 3 Mar 2022 13:24:37 +0000 (08:24 -0500)]
Merge tag 'u-boot-at91-fixes-2022.04-a' of https://source.denx.de/u-boot/custodians/u-boot-at91

First set of u-boot-atmel fixes for the 2022.04 cycle:

This fixes set includes only a single fix for the Ethernet on sama7g5ek
board which is broken at the moment.

2 years agoMerge https://gitlab.denx.de/u-boot/custodians/u-boot-fsl-qoriq
Tom Rini [Thu, 3 Mar 2022 13:24:13 +0000 (08:24 -0500)]
Merge https://gitlab.denx.de/u-boot/custodians/u-boot-fsl-qoriq

Update and fixes for sl28, lx2, pblimage generation for some powerpc
products

2 years agoMerge https://source.denx.de/u-boot/custodians/u-boot-usb
Tom Rini [Tue, 1 Mar 2022 12:48:39 +0000 (07:48 -0500)]
Merge https://source.denx.de/u-boot/custodians/u-boot-usb

2 years agophy: nop-phy: Fix phy reset if no reset-gpio defined
Tim Harvey [Mon, 28 Feb 2022 22:53:21 +0000 (14:53 -0800)]
phy: nop-phy: Fix phy reset if no reset-gpio defined

Ensure there is a valid reset-gpio defined before using it.

Fixes: f9852acdce02 ("phy: nop-phy: Fix enabling reset")
Cc: Adam Ford <aford173@gmail.com>
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
2 years agoPrepare v2022.04-rc3
Tom Rini [Mon, 28 Feb 2022 23:02:33 +0000 (18:02 -0500)]
Prepare v2022.04-rc3

Signed-off-by: Tom Rini <trini@konsulko.com>
2 years agoMerge branch '2022-02-28-bugfixes'
Tom Rini [Mon, 28 Feb 2022 20:45:52 +0000 (15:45 -0500)]
Merge branch '2022-02-28-bugfixes'

- Assorted bugfixes

2 years agoboard: stemmy: Detect board variants and patch DTB
Linus Walleij [Sun, 20 Feb 2022 22:47:01 +0000 (23:47 +0100)]
board: stemmy: Detect board variants and patch DTB

This patch scans the cmdline from the Samsung SBL (second stage
bootloader) and stores the parameters board_id=N and lcdtype=N
in order to augment the DTB for different board and LCD types.

We then add a custom ft_board_setup() callback that will inspect
the DTB and patch it using the stored LCD type. At this point
we know which product we are dealing with, so using the passed
board_id we can also print the board variant for diagnostics.

We patch the Codina, Skomer and Kyle DTBs to use the right
LCD type as passed in lcdtype from the SBL.

This also creates an infrastructure for handling any other
Samsung U8500 board variants that may need a slightly augmented
DTB.

Cc: Markuss Broks <markuss.broks@gmail.com>
Cc: Stephan Gerhold <stephan@gerhold.net>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2 years agoarm: apple: Switch to fully dynamic mem layout
Janne Grunau [Sat, 19 Feb 2022 13:05:19 +0000 (14:05 +0100)]
arm: apple: Switch to fully dynamic mem layout

Support for Apple M1 Pro and Max will allow using a single binary for
all M1 SoCs. The M1 Pro/Max have a different memory layout. The RAM
start address is 0x100_0000_0000 instead of 0x8_0000_0000.
Replace the hardcoded memory layout with dynamic initialized
environment variables in board_late_init().

Tested on Mac Mini (2020) and Macbook Pro 14-inch (2021).

Signed-off-by: Janne Grunau <j@jannau.net>
Reviewed-by: Mark Kettenis <kettenis@openbsd.org>
2 years agoarm: pdu001: Setup pinmux for console UART as early as possible
Felix Brack [Mon, 21 Feb 2022 13:26:05 +0000 (14:26 +0100)]
arm: pdu001: Setup pinmux for console UART as early as possible

To make sure we get a working console as soon as possible in the SPL the
UART pins require to be configured earlier. This is especially
true for the pins of UART3, since the PDU001 board uses this UART for
the console by default.

Signed-off-by: Felix Brack <fb@ltec.ch>
2 years agoarm: pdu001: Fix early debugging UART
Felix Brack [Tue, 15 Feb 2022 14:27:23 +0000 (15:27 +0100)]
arm: pdu001: Fix early debugging UART

The changes from commit 0dba45864b2a ("arm: Init the debug UART")
prevent the early debug UART from being initialized correctly.
To fix this we not just configure the pin multiplexer but add setting up
early clocks.

Signed-off-by: Felix Brack <fb@ltec.ch>
Reviewed-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
2 years agotools: mkimage/dumpimage: Allow to use -l with -T
Pali Rohár [Sun, 13 Feb 2022 00:09:46 +0000 (01:09 +0100)]
tools: mkimage/dumpimage: Allow to use -l with -T

Currently -l option for mkimage and dumpimage ignores option -T and always
tries to autodetect image type.

With this change it is possible to tell mkimage and dumpimage to parse
image file as specific type (and not random autodetected type). This allows
to use mkimage -l or dumpimage -l as tool for validating image.

params.type for -l option is now by default initialized to zero
(IH_TYPE_INVALID) instead of IH_TYPE_KERNEL. imagetool_get_type() for
IH_TYPE_INVALID returns NULL, which is assigned to tparams. mkimage and
dumpimage code is extended to handle tparams with NULL for -l option. And
imagetool_verify_print_header() is extended to do validation via tparams if
is not NULL.

Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
2 years agotools: pblimage: fix image header verification function
Hou Zhiqiang [Thu, 17 Feb 2022 03:51:36 +0000 (11:51 +0800)]
tools: pblimage: fix image header verification function

The Layerscape platforms have different RCW header value from FSL
PowerPC platforms, the current image header verification callback
is only working on PowerPC, it will fail on Layerscape, this patch
is to fix this issue.

This is a historical problem and exposed by the following patch:
http://patchwork.ozlabs.org/project/uboot/patch/20220114173443.9877-1-pali@kernel.org

Signed-off-by: Hou Zhiqiang <Zhiqiang.Hou@nxp.com>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
2 years agolx2160a: Fix distroboot device list for configs without USB/SCSI/etc
Daniel Klauer [Wed, 9 Feb 2022 14:53:41 +0000 (15:53 +0100)]
lx2160a: Fix distroboot device list for configs without USB/SCSI/etc

The BOOT_TARGET_DEVICES list for distro_bootcmd was hard-coded to assume
that all boot devices are available/enabled in the configuration,
thus ignoring the actual config settings. The config_distro_bootcmd.h
header file specifically has compile-time checks to detect such problems.

To allow disabling USB, SCSI, etc. in custom lx2160a board configs,
make it depend on the config settings and use only the enabled features.

Signed-off-by: Daniel Klauer <daniel.klauer@gin.de>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
2 years agoboard: sl28: use fit image generator
Michael Walle [Fri, 25 Feb 2022 12:51:56 +0000 (18:21 +0530)]
board: sl28: use fit image generator

Simplify the binman config and fdt nodes by using the "@..-SEQ"
substitutions and CONFIG_OF_LIST.

Signed-off-by: Michael Walle <michael@walle.cc>
[Rebased]
Signed-off-by: Priyanka Jain <priyanka.jain@nxp.com>
2 years agoboard: sl28: disable random MAC address generation
Michael Walle [Fri, 25 Feb 2022 12:48:40 +0000 (18:18 +0530)]
board: sl28: disable random MAC address generation

Nowadays, u-boot (when CONFIG_NET_RANDOM_ETHADDR is set) will set
enetaddr to a random value if not set and then pass the randomly
generated MAC address to linux.

This is bad for the following reasons:
 (1) it makes it impossible for linux to detect this error
 (2) linux won't trigger any fallback mechanism for the case where
     it didn't find any valid MAC address
 (3) a saveenv will store this randomly generated MAC address in the
     environment

Probably, the user will also be unaware that something is wrong. He will
just get different MAC addresses on each reboot, asking himself why this
is the case.

As this board usually have a serial port, the user can just fix this by
setting the MAC address manually in the environment. Also disable the
netconsole just in case, because it cannot be guaranteed that it will
work in any case. After all, this was just a convenience option, because
the bootloader - right now - doesn't have the ability to read the MAC
address, which is stored in the OTP. But it is far more important to
have a clear view of whats wrong with a board and that means we can no
longer use this Kconfig option.

Signed-off-by: Michael Walle <michael@walle.cc>
[Rebased]
Signed-off-by: Priyanka Jain <priyanka.jain@nxp.com>
2 years agoboard: sl28: remove "Useful I2C tricks" section from docs
Michael Walle [Mon, 15 Nov 2021 22:45:50 +0000 (23:45 +0100)]
board: sl28: remove "Useful I2C tricks" section from docs

They are no longer needed, because we now have proper driver support for
the sl28cpld management controller.

Signed-off-by: Michael Walle <michael@walle.cc>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
2 years agoboard: sl28: disable recovery watchdog
Michael Walle [Mon, 15 Nov 2021 22:45:49 +0000 (23:45 +0100)]
board: sl28: disable recovery watchdog

This board has an internal watchdog which supervises the board startup.
Although, the initial state of the watchdog is configurable, it is
enabled by default. In board_late_init(), which means almost everything
worked as expected, disable the watchdog.

Signed-off-by: Michael Walle <michael@walle.cc>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
2 years agoboard: sl28: enable SoC watchdog support
Michael Walle [Mon, 15 Nov 2021 22:45:48 +0000 (23:45 +0100)]
board: sl28: enable SoC watchdog support

The SoC provides two additional watchdogs integrated in the SoC. Enable
support for these.

Signed-off-by: Michael Walle <michael@walle.cc>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
2 years agoboard: sl28: enable sl28cpld support
Michael Walle [Mon, 15 Nov 2021 22:45:47 +0000 (23:45 +0100)]
board: sl28: enable sl28cpld support

Enable the GPIO and watchdog driver. Don't start the watchdog
automatically, though.

Signed-off-by: Michael Walle <michael@walle.cc>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
2 years agoboard: sl28: print CPLD version on bootup
Michael Walle [Mon, 15 Nov 2021 22:45:46 +0000 (23:45 +0100)]
board: sl28: print CPLD version on bootup

Most of the time it is very useful to have the version of the board
management controller. Now that we have a driver, print it during
startup.

Signed-off-by: Michael Walle <michael@walle.cc>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
2 years agoboard: sl28: fix DRAM pretty print
Michael Walle [Mon, 15 Nov 2021 22:45:45 +0000 (23:45 +0100)]
board: sl28: fix DRAM pretty print

The current console output is:

DRAM:  4 GiB
DDR    4 GiB (DDR3, 32-bit, CL=11, ECC on)

The size is printed twice and we can save one line of console output if
we join both lines. The new output is as follows:

DRAM:  4 GiB (DDR3, 32-bit, CL=11, ECC on)

Signed-off-by: Michael Walle <michael@walle.cc>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
2 years agogpio: add sl28cpld driver
Michael Walle [Fri, 25 Feb 2022 12:40:24 +0000 (18:10 +0530)]
gpio: add sl28cpld driver

The gpio block is part of the sl28cpld sl28cpld management controller.
There are three different flavors: the usual input and output where the
direction is configurable, but also input only and output only variants.

Signed-off-by: Michael Walle <michael@walle.cc>
[Rebased]
Signed-off-by: Priyanka Jain <priyanka.jain@nxp.com>
2 years agowatchdog: add sl28cpld watchdog driver
Michael Walle [Mon, 15 Nov 2021 22:45:43 +0000 (23:45 +0100)]
watchdog: add sl28cpld watchdog driver

The watchdog timer is part of the sl28cpld management controller. The
watchdog timer usually supervises the bootloader boot-up and if it bites
the failsafe bootloader will be activated. Apart from that it supports
the usual board level reset and one SMARC speciality: driving the
WDT_TIMEOUT# signal.

Signed-off-by: Michael Walle <michael@walle.cc>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
2 years agomisc: add sl28cpld base driver
Michael Walle [Fri, 25 Feb 2022 12:36:24 +0000 (18:06 +0530)]
misc: add sl28cpld base driver

Add a multi-function device driver which will probe its children and
provides methods to access the device.

Signed-off-by: Michael Walle <michael@walle.cc>
[Rebased]
Signed-off-by: Priyanka Jain <priyanka.jain@nxp.com>
2 years agoMerge tag 'efi-2022-04-rc3' of https://source.denx.de/u-boot/custodians/u-boot-efi
Tom Rini [Sat, 26 Feb 2022 15:21:39 +0000 (10:21 -0500)]
Merge tag 'efi-2022-04-rc3' of https://source.denx.de/u-boot/custodians/u-boot-efi

Pull request for efi-2022-04-rc3

Documentation:

* add man-page for fatload
* add SMBIOS table page

UEFI:

* partial fix for UEFI secure boot with intermediate certs
* disable watchdog when returning to command line
* reset system after capsule update

2 years agoMerge https://source.denx.de/u-boot/custodians/u-boot-sh
Tom Rini [Sat, 26 Feb 2022 15:21:13 +0000 (10:21 -0500)]
Merge https://source.denx.de/u-boot/custodians/u-boot-sh

- rzg2_beacon updates

2 years agoefi_loader: update the timing of enabling and disabling EFI watchdog
Masahisa Kojima [Tue, 22 Feb 2022 00:58:30 +0000 (09:58 +0900)]
efi_loader: update the timing of enabling and disabling EFI watchdog

UEFI specification requires that 5 minutes watchdog timer is
armed before the firmware's boot manager invokes an EFI boot option.
This watchdog timer is updated as follows, according to the
UEFI specification.

 1) The EFI Image may reset or disable the watchdog timer as needed.
 2) If control is returned to the firmware's boot manager,
    the watchdog timer must be disabled.
 3) On successful completion of EFI_BOOT_SERVICES.ExitBootServices()
    the watchdog timer is disabled.

1) is up to the EFI image, and 3) is already implemented in U-Boot.
This patch implements 2), the watchdog is disabled when control is
returned to U-Boot.

In addition, current implementation arms the EFI watchdog at only
the first "bootefi" invocation. The EFI watchdog must be armed
in every EFI boot option invocation.

Signed-off-by: Masahisa Kojima <masahisa.kojima@linaro.org>
Reviewed-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
2 years agoefi_loader: test/py: Reset system after capsule update on disk
Masami Hiramatsu [Wed, 16 Feb 2022 06:16:12 +0000 (15:16 +0900)]
efi_loader: test/py: Reset system after capsule update on disk

Add a cold reset soon after processing capsule update on disk.
This is required in UEFI specification 2.9 Section 8.5.5
"Delivery of Capsules via file on Mass Storage device" as;

    In all cases that a capsule is identified for processing the system is
    restarted after capsule processing is completed.

This also reports the result of each capsule update so that the user can
notice that the capsule update has been succeeded or not from console log.

Signed-off-by: Masami Hiramatsu <masami.hiramatsu@linaro.org>
Reviewed-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
2 years agotest/py: Handle expected reboot while booting sandbox
Masami Hiramatsu [Wed, 16 Feb 2022 06:16:02 +0000 (15:16 +0900)]
test/py: Handle expected reboot while booting sandbox

Add expected_reset optional argument to ConsoleBase::ensure_spawned(),
ConsoleBase::restart_uboot() and ConsoleSandbox::restart_uboot_with_flags()
so that it can handle a reset while the 1st boot process after main
boot logo before prompt correctly.

Signed-off-by: Masami Hiramatsu <masami.hiramatsu@linaro.org>
Reviewed-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
2 years agotest/py: Handle expected reset by command
Masami Hiramatsu [Wed, 16 Feb 2022 06:15:52 +0000 (15:15 +0900)]
test/py: Handle expected reset by command

Add wait_for_reboot optional argument to ConsoleBase::run_command()
so that it can handle an expected reset by command execution.

This is useful if a command will reset the sandbox while testing
such commands, e.g. run_command("reset", wait_for_reboot = True)

Signed-off-by: Masami Hiramatsu <masami.hiramatsu@linaro.org>
Reviewed-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
2 years agoefi_loader: use efi_update_capsule_firmware() for capsule on disk
Masami Hiramatsu [Wed, 16 Feb 2022 06:15:42 +0000 (15:15 +0900)]
efi_loader: use efi_update_capsule_firmware() for capsule on disk

Since the efi_update_capsule() represents the UpdateCapsule() runtime
service, it has to handle the capsule flags and update ESRT. However
the capsule-on-disk doesn't need to care about such things.

Thus, the capsule-on-disk should use the efi_capsule_update_firmware()
directly instead of calling efi_update_capsule().

This means the roles of the efi_update_capsule() and capsule-on-disk
are different. We have to keep the efi_update_capsule() for providing
runtime service API at boot time.

Suggested-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Signed-off-by: Masami Hiramatsu <masami.hiramatsu@linaro.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
2 years agoefi_loader: fix uefi secure boot with intermediate certs
Ilias Apalodimas [Mon, 14 Feb 2022 09:14:22 +0000 (11:14 +0200)]
efi_loader: fix uefi secure boot with intermediate certs

The general rule of accepting or rejecting an image is
 1. Is the sha256 of the image in dbx
 2. Is the image signed with a certificate that's found in db and
    not in dbx
 3. The image carries a cert which is signed by a cert in db (and
    not in dbx) and the image can be verified against the former
 4. Is the sha256 of the image in db

For example SHIM is signed by "CN=Microsoft Windows UEFI Driver Publisher",
which is issued by "CN=Microsoft Corporation UEFI CA 2011", which in it's
turn is issued by "CN=Microsoft Corporation Third Party Marketplace Root".
The latter is a self-signed CA certificate and with our current implementation
allows shim to execute if we insert it in db.

However it's the CA cert in the middle of the chain which usually ends up
in the system's db.  pkcs7_verify_one() might or might not return the root
certificate for a given chain.  But when verifying executables in UEFI,  the
trust anchor can be in the middle of the chain, as long as that certificate
is present in db.  Currently we only allow this check on self-signed
certificates,  so let's remove that check and allow all certs to try a
match an entry in db.

Open questions:
- Does this break any aspect of variable authentication since
  efi_signature_verify() is used on those as well?

Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
2 years agotools: mkeficapsule: remove duplicated code
AKASHI Takahiro [Wed, 16 Feb 2022 01:49:51 +0000 (10:49 +0900)]
tools: mkeficapsule: remove duplicated code

That code is mistakenly duplicated due to copy-and-paste error.
Just remove it.

Fixes: CID 348360
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Reviewed-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
2 years agodoc: describe fatload command
Heinrich Schuchardt [Fri, 18 Feb 2022 18:33:09 +0000 (19:33 +0100)]
doc: describe fatload command

Man-page for fatload command.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
2 years agodoc: describe generation of SMBIOS table
Heinrich Schuchardt [Wed, 16 Feb 2022 06:59:31 +0000 (07:59 +0100)]
doc: describe generation of SMBIOS table

SMBIOS is not x86 specific. So we should have an architecture independent
page describing it.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
2 years agoarm: rmobile: rzg2_beacon: Enable proper Ethernet PHY
Adam Ford [Fri, 25 Feb 2022 20:32:53 +0000 (14:32 -0600)]
arm: rmobile: rzg2_beacon: Enable proper Ethernet PHY

The wrong phy was being enabled, because it worked and the proper
PHY did not.  After the Renesas maintainer made some adjustments
to the device tree, Linux was able to use the proper driver, and
when that device tree was ported to Linux, the ethernet stopped
working due to the lack of rgmii-rxid support.  Now that
rgmii-rxid is supported, enable the proper driver to restore
ethernet function.

Fixes: 1eaf61c84db6 ("arm: dts: beacon-rzg2: Resync device trees with Linux 5.16-rc3")
Signed-off-by: Adam Ford <aford173@gmail.com>
2 years agonet: ravb: Add tx/rx delay flag checks and support for rgmii-rxid
Adam Ford [Fri, 25 Feb 2022 20:32:52 +0000 (14:32 -0600)]
net: ravb: Add tx/rx delay flag checks and support for rgmii-rxid

Some boards like the Beacon RZ/G2 SOM use either flags for
tx-internal-delay-ps, rx-internal-delay-ps or rgmii-rxid.

In Linux the APSR_RDM flag is set when either rx-internal-delay-ps
is set or the mode is rgmii-rxid, and the APSR_TDM is set when
tx-internal-delay-ps is found or rgmii-txid is set, and both
are set if rgmii-id is set.

The ravb driver in U-Boot driver was missing rgmii-rxid support,
so add that support in a similar fashion to what is done in Linux.

Signed-off-by: Adam Ford <aford173@gmail.com>
2 years agoMerge tag 'clk-2022.04-rc2' of https://source.denx.de/u-boot/custodians/u-boot-clk
Tom Rini [Fri, 25 Feb 2022 16:21:32 +0000 (11:21 -0500)]
Merge tag 'clk-2022.04-rc2' of https://source.denx.de/u-boot/custodians/u-boot-clk

Clock patches for v2022.04-rc2

This has an assortment of cleanups and the occasional bugfix. Also present
is the addition of the clock subsystem documentation to HTML docs.

CI: https://source.denx.de/u-boot/custodians/u-boot-clk/-/pipelines/11075

2 years agoscripts: Makefile.lib: generate dsdt_generated.c instead of dsdt.c
Philippe Reynes [Tue, 22 Feb 2022 13:54:39 +0000 (14:54 +0100)]
scripts: Makefile.lib: generate dsdt_generated.c instead of dsdt.c

There is a conflict between the static file
lib/acpi/dsdt.c and the file dsdt.c generated
dynamicaly by scripts/Makefile.lib. When a
mrproper is done, the static file dsdt.c is
removed. If a build with acpi enabled is
launched after, the following error is raised:

  CC      lib/acpi/acpi_table.o
make[2]: *** No rule to make target 'lib/acpi/dsdt.asl', needed by 'lib/acpi/dsdt.c'.  Stop.
scripts/Makefile.build:394: recipe for target 'lib/acpi' failed

To avoid such error, the generated file is named
dsdt_generated.c instead of dstdt.c.

Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
Tested-by: Heiko Thiery <heiko.thiery@gmail.com>
2 years agocmd: clk: fix long help message
Patrick Delaunay [Mon, 31 Jan 2022 16:21:40 +0000 (17:21 +0100)]
cmd: clk: fix long help message

Fix the long help message for "clk setfreq" command

Fixes: 7ab418fbe612 ("clk: add support for setting clk rate from cmdline")
Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Reviewed-by: Sean Anderson <seanga2@gmail.com>
Link: https://lore.kernel.org/r/20220131172131.4.Ic863c28ffdcc15b3f4616434c2efd88b4e45495c@changeid
2 years agocmd: clk: update result of do_clk_setfreq
Patrick Delaunay [Mon, 31 Jan 2022 16:21:39 +0000 (17:21 +0100)]
cmd: clk: update result of do_clk_setfreq

Update the result of do_clk_setfreq and always returns a CMD_RET_ value
(-EINVAL was a possible result).

This patch avoid the CLI output "exit not allowed from main input shell."

Fixes: 7ab418fbe612 ("clk: add support for setting clk rate from cmdline")
Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Reviewed-by: Sean Anderson <seanga2@gmail.com>
Link: https://lore.kernel.org/r/20220131172131.3.Iec2029edb7fc0b29e13bcb86058ad2f614f62779@changeid
2 years agocmd: clk: replace clk_lookup by uclass_get_device_by_name
Patrick Delaunay [Mon, 31 Jan 2022 16:21:38 +0000 (17:21 +0100)]
cmd: clk: replace clk_lookup by uclass_get_device_by_name

The function clk_lookup can be replaced by a direct call
to uclass_get_device_by_name for UCLASS_CLK.

This patch removes duplicated codes by the generic DM API and avoids
issue in clk_lookup because result of uclass_get_device wasn't tested;
when ret < 0, dev = NULL and dev->name is invalid, the next function
call strcmp(name, dev->name) causes a crash.

Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Link: https://lore.kernel.org/r/20220131172131.2.I7bc7762eff1e31ab7ff5b34c416ee03b8fe52200@changeid
2 years agocmd: clk: test the number of argument in setfreq command
Patrick Delaunay [Mon, 31 Jan 2022 16:21:37 +0000 (17:21 +0100)]
cmd: clk: test the number of argument in setfreq command

Test the number of argument in setfreq command to avoid a crash when
the command setfreq is called without argument:

  STM32MP> clk setfreq
  data abort
  pc : [<ddba3f18>]    lr : [<ddba3f89>]
  reloc pc : [<c018ff18>]    lr : [<c018ff89>]
  sp : dbaf45b8  ip : ddb1d859  fp : 00000002
  r10: dbb3fd80  r9 : dbb11e90  r8 : ddbf38cc
  r7 : ddb39725  r6 : 00000000  r5 : 00000000  r4 : dbb3fd84
  r3 : dbb3fd84  r2 : 0000000a  r1 : dbaf45bc  r0 : 00000011
  Flags: nzCv  IRQs off  FIQs off  Mode SVC_32 (T)
  Code: 4dd3 1062 85a3 ddbd (7803) 2b30
  Resetting CPU ...

Fixes: 7ab418fbe612 ("clk: add support for setting clk rate from cmdline")
Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Reviewed-by: Sean Anderson <seanga2@gmail.com>
Link: https://lore.kernel.org/r/20220131172131.1.I32a8f213d330dccd922f7aafc60d3d63fcbe8615@changeid
2 years agoclk: ccf: correct the test on the parent uclass in clk_enable/clk_disable
Patrick Delaunay [Mon, 24 Jan 2022 13:17:14 +0000 (14:17 +0100)]
clk: ccf: correct the test on the parent uclass in clk_enable/clk_disable

It is safe to check if the uclass id on the device is UCLASS_CLK
before to call the clk_ functions, but today this comparison is
not done on the device used in API: clkp->dev->parent
but on the device himself: clkp->dev.

This patch corrects this behavior and tests if the parent device
is a clock device before to call the clock API, clk_enable or
clk_disable, on this device.

Fixes: 0520be0f67e3 ("clk: prograte clk enable/disable to parent")
Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Reviewed-by: Sean Anderson <seanga2@gmail.com>
2 years agoclk: Add clk_get_by_name_optional
Sean Anderson [Sat, 15 Jan 2022 20:52:47 +0000 (15:52 -0500)]
clk: Add clk_get_by_name_optional

This adds a helper function for clk_get_by_name in cases where the clock is
optional. Hopefully this helps point driver writers in the right direction.
Also convert some existing users.

Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Neil Armstrong <narmstrong@baylibre.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Link: https://lore.kernel.org/r/20220115205247.566210-2-seanga2@gmail.com
2 years agoclk: Add driver API to HTML docs
Sean Anderson [Wed, 22 Dec 2021 17:11:13 +0000 (12:11 -0500)]
clk: Add driver API to HTML docs

This converts the existing driver API docs (clk-uclass.h) to kernel doc
format and adds them to the HTML documentation. Because the kernel doc
sphinx converter does not handle functions in structs very well, the
individual methods are documented separately. This is primarily inspired by
the phylink documentation [1], which uses this trick extensively.

[1] https://www.kernel.org/doc/html/latest/networking/kapi.html#c.phylink_mac_ops

Signed-off-by: Sean Anderson <seanga2@gmail.com>
Tested-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Link: https://lore.kernel.org/r/20211222171114.3091780-5-seanga2@gmail.com
2 years agoclk: Add client API to HTML docs
Sean Anderson [Wed, 22 Dec 2021 17:11:12 +0000 (12:11 -0500)]
clk: Add client API to HTML docs

This converts the existing client (aka clk.h) documentation to kernel doc
format, and adds it to the HTML docs. I have tried to preserve existing
comments as much as possible, refraining from semantic changes.

Signed-off-by: Sean Anderson <seanga2@gmail.com>
Tested-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Link: https://lore.kernel.org/r/20211222171114.3091780-4-seanga2@gmail.com
[rebased onto u-boot/master and resolved conflicts]
Signed-off-by: Sean Anderson <seanga2@gmail.com>
2 years agoclk: Inline clk_get_*_optional
Sean Anderson [Wed, 22 Dec 2021 17:11:11 +0000 (12:11 -0500)]
clk: Inline clk_get_*_optional

The optional varients of clk_get_* functions are just simple wrappers.
Reduce code size a bit by inlining them. On platforms where it is not used
(most of them), it will not be compiled in any more. On platforms where
they are used, the inlined branch should not cause any significant growth.

Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Link: https://lore.kernel.org/r/20211222171114.3091780-3-seanga2@gmail.com
2 years agoclk: Rename clk_get_optional_nodev
Sean Anderson [Wed, 22 Dec 2021 17:11:10 +0000 (12:11 -0500)]
clk: Rename clk_get_optional_nodev

This normalizes the name of this accessor function to put "_optional" last.

Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Link: https://lore.kernel.org/r/20211222171114.3091780-2-seanga2@gmail.com
2 years agoclk: cdce9xx: Convert .of_xlate to .request
Sean Anderson [Wed, 15 Dec 2021 16:47:17 +0000 (11:47 -0500)]
clk: cdce9xx: Convert .of_xlate to .request

This xlate function just performs some checking. We can do this in
request() instead and use the default xlate.

Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Tero Kristo <kristo@kernel.org>
Link: https://lore.kernel.org/r/20211215164718.2778664-1-seanga2@gmail.com
2 years agoclk: versaclock: Remove xlate function
Sean Anderson [Wed, 1 Dec 2021 20:13:17 +0000 (15:13 -0500)]
clk: versaclock: Remove xlate function

This function is the same as the default xlate. Remove it.

Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Adam Ford <aford173@gmail.com>
Link: https://lore.kernel.org/r/20211201201317.2174547-1-seanga2@gmail.com
2 years agoclk: Remove no-op request and rfree callbacks
Sean Anderson [Wed, 1 Dec 2021 19:51:00 +0000 (14:51 -0500)]
clk: Remove no-op request and rfree callbacks

These callbacks are optional. Remove ones which do nothing.

Signed-off-by: Sean Anderson <seanga2@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Link: https://lore.kernel.org/r/20211201195100.2173465-1-seanga2@gmail.com
2 years agoclk: Rename ICS8N3QV01 to CLK_ICS8N3QV01
Sean Anderson [Wed, 15 Dec 2021 16:36:20 +0000 (11:36 -0500)]
clk: Rename ICS8N3QV01 to CLK_ICS8N3QV01

This driver was missing a clock prefix. Add one.

Signed-off-by: Sean Anderson <seanga2@gmail.com>
Link: https://lore.kernel.org/r/20211215163620.2770126-4-seanga2@gmail.com
2 years agoclk: Alphabetize Kconfig
Sean Anderson [Wed, 15 Dec 2021 16:36:19 +0000 (11:36 -0500)]
clk: Alphabetize Kconfig

This alphabetizes the Kconfig for the clock subsystem. This will help
people find their clocks, and help prevent merge conflicts.

Signed-off-by: Sean Anderson <seanga2@gmail.com>
Link: https://lore.kernel.org/r/20211215163620.2770126-3-seanga2@gmail.com
2 years agoclk: Alphabetize Makefile
Sean Anderson [Wed, 15 Dec 2021 16:36:18 +0000 (11:36 -0500)]
clk: Alphabetize Makefile

This alphabetizes the clock makefile by Kconfig option. This will help
prevent merge conflicts.

Signed-off-by: Sean Anderson <seanga2@gmail.com>
Link: https://lore.kernel.org/r/20211215163620.2770126-2-seanga2@gmail.com
2 years agoMerge https://gitlab.denx.de/u-boot/custodians/u-boot-usb
Tom Rini [Wed, 23 Feb 2022 18:34:14 +0000 (13:34 -0500)]
Merge https://gitlab.denx.de/u-boot/custodians/u-boot-usb

- OMAP EHCI updates

2 years agoMerge tag 'dm-pull-22222' of https://gitlab.denx.de/u-boot/custodians/u-boot-dm
Tom Rini [Wed, 23 Feb 2022 17:28:54 +0000 (12:28 -0500)]
Merge tag 'dm-pull-22222' of https://gitlab.denx.de/u-boot/custodians/u-boot-dm

binman fixes/improvements to FIT generator
binman SPL fixes
moveconfig support regex matches

2 years agoconfigs: omap various: Remove OMAP_EHCI_PHY from defconfigs
Adam Ford [Sat, 19 Feb 2022 23:08:47 +0000 (17:08 -0600)]
configs: omap various: Remove OMAP_EHCI_PHY from defconfigs

With the Kconfig options being deleted, the references to
OMAP_EHCI_PHY are useless.  Remove them from the various
defconfigs.

Signed-off-by: Adam Ford <aford173@gmail.com>
2 years agousb: ehci-omap: Remove OMAP_EHCI_PHYx_RESET_GPIO from Kconfig
Adam Ford [Sat, 19 Feb 2022 23:08:46 +0000 (17:08 -0600)]
usb: ehci-omap: Remove OMAP_EHCI_PHYx_RESET_GPIO from Kconfig

With the omap-ehci driver now using the phy subsystem to enable
and disable reset, the driver no longer needs to know which
GPIO's are used, and they can be removed from Kconfig.

Signed-off-by: Adam Ford <aford173@gmail.com>
2 years agousb: ehci-omap: Use PHY system to manage phy resets
Adam Ford [Sat, 19 Feb 2022 23:08:45 +0000 (17:08 -0600)]
usb: ehci-omap: Use PHY system to manage phy resets

There are a few boards that use hard-coded GPIO definitions in
their respective defconfig files.  If the GPIO's are listed
in their device trees, the nop-phy can toggle the GPIO's,
so the EHCI driver does not need to know anything about the
GPIO's. Add functions for getting the phys and remove the GPIO
toggles since the phy will now do that.

Signed-off-by: Adam Ford <aford173@gmail.com>
2 years agousb: ehci-omap: Make Kconfig select PHY if USB_EHCI_OMAP
Adam Ford [Sat, 19 Feb 2022 23:08:44 +0000 (17:08 -0600)]
usb: ehci-omap: Make Kconfig select PHY if USB_EHCI_OMAP

The USB_EHCI_OMAP driver currently has a series of Kconfig options
which let users specify a GPIO for the reset pin.  Some devices
may have only one reset, while others might have more.

Since there is a nop phy driver, let's selct enable the PHY
system, and imply the nop phy driver.  The nop phy driver can now
toggle the reset pins when putting the phy in and out of reset.

If the gpio is listed under the phy, it will get toggled and
the hard-coded config options specifying the GPIO numbers can
eventually go away.

Signed-off-by: Adam Ford <aford173@gmail.com>
2 years agophy: nop-phy: Fix enabling reset
Adam Ford [Sat, 19 Feb 2022 23:08:43 +0000 (17:08 -0600)]
phy: nop-phy: Fix enabling reset

The reset function should place the phy into reset, while the
init function should take the phy out of reset.  Currently the
reset function takes it out of reset, and the init calls the
reset.

Signed-off-by: Adam Ford <aford173@gmail.com>
2 years agousb: ehci-omap: Move omap_ehci_hcd_init to omap_ehci_probe
Adam Ford [Sat, 19 Feb 2022 23:08:42 +0000 (17:08 -0600)]
usb: ehci-omap: Move omap_ehci_hcd_init to omap_ehci_probe

The OMAP3 hierarchy has the ehci node as a sub-node of the
usbhshost. The usbhshost node contains an ohci and an ehci
subnode.  The configuration of the ehci belongs in the
EHCI node and not its parent.  Move it to the proper probe.

usb start
  starting USB...
  Bus ehci@48064800: USB EHCI 1.00
  Bus usb_otg_hs@480ab000: Port not available.
  scanning bus ehci@48064800 for devices... 3 USB Device(s) found
       scanning usb for storage devices... 1 Storage Device(s) found

Signed-off-by: Adam Ford <aford173@gmail.com>
2 years agousb: ehci-omap: Drop dead code
Adam Ford [Sat, 19 Feb 2022 23:08:41 +0000 (17:08 -0600)]
usb: ehci-omap: Drop dead code

omap_ehci_hcd_stop appears to be dead code, and omap_ehci_hcd_init
is only called by the probe function, so it can be static to that
function.  Remove both from the header along with some additional
checking for DM_USB.

Signed-off-by: Adam Ford <aford173@gmail.com>
2 years agoscripts: dtc: libfdt: fdt_ro.c: always define fdt_check_full
Philippe Reynes [Wed, 9 Feb 2022 17:01:24 +0000 (18:01 +0100)]
scripts: dtc: libfdt: fdt_ro.c: always define fdt_check_full

On some configs (like stm32mp15_dhcom_basic_defconfig), if configs
SPL_LOAD_FIT_FULL and SPL_FIT_FULL_CHECK are enabled. Then the compilatio
fails with the following error:

arm-linux-gnueabi-ld.bfd: boot/image-fit.o: in function `fit_check_format':
<PATH>/uboot/u-boot-stm/boot/image-fit.c:1641: undefined reference to `fdt_check_full'
scripts/Makefile.spl:509: recipe for target 'spl/u-boot-spl' failed

This issue happens because the function fdt_check_full is only defined if
"!defined(FDT_ASSUME_MASK) || FDT_ASSUME_MASK != 0xff". But this function
may be called even if this condition are not verified. To avoid this issue,
the function fdt_check_full is always defined.

Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
2 years agophy: phy-uclass: check the parents for phys
Angus Ainslie [Thu, 3 Feb 2022 18:08:38 +0000 (10:08 -0800)]
phy: phy-uclass: check the parents for phys

The port/hub leaf nodes don't contain the phy definitions in some dts
files so check the parents.

Signed-off-by: Angus Ainslie <angus@akkea.ca>
Reviewed-by: Simon Glass <sjg@chromium.org>
2 years agobinman: Allow different operations in FIT generator nodes
Simon Glass [Tue, 8 Feb 2022 18:50:03 +0000 (11:50 -0700)]
binman: Allow different operations in FIT generator nodes

At present we only support expanding out FDT nodes. Make the operation
into an @operation property, so that others can be supported.

Re-arrange and tidy up the documentation so that it has separate
headings for each topic.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agobinman: Tidy up the docs a little with fit
Simon Glass [Tue, 8 Feb 2022 18:50:02 +0000 (11:50 -0700)]
binman: Tidy up the docs a little with fit

Add a few quotes and clarify the data property.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agobinman: fit: Refactor to reduce function size
Simon Glass [Tue, 8 Feb 2022 18:50:01 +0000 (11:50 -0700)]
binman: fit: Refactor to reduce function size

Split subnode and property processing into separate functions to make
the _AddNode() function a little smaller. Tweak a few comments.

This does not change any functionality.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agobinman: Move entry-data collection into a Entry method
Simon Glass [Tue, 8 Feb 2022 18:50:00 +0000 (11:50 -0700)]
binman: Move entry-data collection into a Entry method

Collecting the data from a list of entries and putting it in a file is
a useful operation that will be needed by other entry types. Put this into
a method in the Entry class.

Add some documentation about how to collect data for an entry type.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agobinman: Add a ELF test file with disjoint text sections
Simon Glass [Tue, 8 Feb 2022 18:49:59 +0000 (11:49 -0700)]
binman: Add a ELF test file with disjoint text sections

Add a file that has two text sections at different addresses, so we can
test this behaviour in binman, once added.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agobinman: Support a list of strings with the mkimage etype
Simon Glass [Tue, 8 Feb 2022 18:49:58 +0000 (11:49 -0700)]
binman: Support a list of strings with the mkimage etype

At present the 'args' property of the mkimage entry type is a string. This
makes it difficult to include CONFIG options in that property. In
particular, this does not work:

   args = "-n CONFIG_SYS_SOC -E"

since the preprocessor does not operate within strings, nor does this:

   args = "-n" CONFIG_SYS_SOC" "-E"

since the device tree compiler does not understand string concatenation.

With this new feature, we can do:

   args = "-n", CONFIG_SYS_SOC, "-E";

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agobinman: Add to the TODO
Simon Glass [Tue, 8 Feb 2022 18:49:57 +0000 (11:49 -0700)]
binman: Add to the TODO

Add some ideas that have come to mind recently.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agobinman: Add support for TEE BL32
Roger Quadros [Sat, 19 Feb 2022 18:50:04 +0000 (20:50 +0200)]
binman: Add support for TEE BL32

Add an entry for OP-TEE Trusted OS 'BL32' payload.
This is required by platforms using Cortex-A cores with TrustZone
technology.

Signed-off-by: Roger Quadros <rogerq@kernel.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
Add missing-blob-help, renumber the test file, update entry-docs:
Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agoelf: Add a way to read segment information from an ELF file
Simon Glass [Tue, 8 Feb 2022 18:49:55 +0000 (11:49 -0700)]
elf: Add a way to read segment information from an ELF file

Add a function which reads the segments and the entry address.

Also fix a comment nit in the tests while we are here.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agobinman: Update docs to indicate mkimage is supported
Simon Glass [Tue, 8 Feb 2022 18:49:54 +0000 (11:49 -0700)]
binman: Update docs to indicate mkimage is supported

Now that there is a mkimage entry-type, update the docs to remove the
future reference.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agodtoc: Support reading a list of arguments
Simon Glass [Tue, 8 Feb 2022 18:49:53 +0000 (11:49 -0700)]
dtoc: Support reading a list of arguments

It is helpful to support a string or stringlist containing a list of
space-separated arguments, for example:

   args = "-n fred", "-a", "123";

This resolves to the list:

   -n fred -a 123

which can be passed to a program as arguments.

Add a helper to do the required processing.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agodtoc: Allow deleting nodes and adding them in the same sync
Simon Glass [Tue, 8 Feb 2022 18:49:52 +0000 (11:49 -0700)]
dtoc: Allow deleting nodes and adding them in the same sync

This does not work at present, since the current algorithm assumes that
either there are no nodes or all nodes have an offset. If a node is new,
but an old node is still in the tree, then syncing fails due to this
assumption.

Fix it and add a test.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agodtoc: Support deleting a node
Simon Glass [Tue, 8 Feb 2022 18:49:51 +0000 (11:49 -0700)]
dtoc: Support deleting a node

Add a function to delete a node. This is synced to the tree when
requested.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agodtoc: Support adding a string list to a device tree
Simon Glass [Tue, 8 Feb 2022 18:49:50 +0000 (11:49 -0700)]
dtoc: Support adding a string list to a device tree

Add a new function to add a string list.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agospl: Allow disabling binman symbols in SPL
Simon Glass [Tue, 8 Feb 2022 18:49:48 +0000 (11:49 -0700)]
spl: Allow disabling binman symbols in SPL

When CONFIG_SPL_FIT is enabled we do not access U-Boot directly in
the image, since it is embedded in a FIT which is parsed at runtime.

Provide a CONFIG option to drop the symbols in this case.

Signed-off-by: Simon Glass <sjg@chromium.org>
2 years agospl: x86: Correct the binman symbols for SPL
Simon Glass [Tue, 8 Feb 2022 18:49:47 +0000 (11:49 -0700)]
spl: x86: Correct the binman symbols for SPL

These symbols are incorrect, meaning that binman cannot find the
associated entry. This leads to errors like:

binman: Section '/binman/simple-bin': Symbol '_binman_spl_prop_size'
   in entry '/binman/simple-bin/u-boot-spl/u-boot-spl-nodtb':
   Entry 'spl' not found in list (mkimage,u-boot-spl-nodtb,
   u-boot-spl-bss-pad,u-boot-spl-dtb,u-boot-spl,u-boot-img,main-section)

Fix it.

Signed-off-by: Simon Glass <sjg@chromium.org>