To quote the author:
This series makes a further attempt to get closer to having all tests
run in parallel. It introduces a new 'make pcheck' option which runs
tests in parallel, skipping those that are not compatible.

A number of fixes are included for existing tests. The vboot test is
updated to only run a single scenario in 'quick' mode.

This makes use of pytest's parallel-testing features. The resulting
times (including incremental building with LTO) on a 16-core machine are
as follows:

   make pcheck        - 1 minute 6 seconds
   make qcheck        - 3 minutes
   make check         - 5 minutes 15 seconds

Note that this is not a fair comparison, since 'make pcheck' omits a
number of tests, even more than 'make qcheck'.

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

Running tests in parallel is much faster, e.g. 15 seconds to run the tests
on sandbox (only), instead of 100 seconds (on a 16-core machine). Add a
'make pcheck' option to access this feature.

Note that the tools/ tests still run each tool's tests once after the
other, although within that, they do run in parallel. So for example,
the buildman tests run in parallel, then the binman tests run in
parallel. There would be a signiificant advantage to running them all
in parallel together, but that would require a large amount of
refactoring, e.g. with more use of pytest fixtures.

Update the documentation to represent the current state.

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

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

This breaks using pytest to run the tests. Drop it.

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

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

At present when -n is used, all workers try to build U-Boot at once.
Add a lock to ensure that only one of them builds, with the others using
the build that is produced.

The lock file is removed on startup.

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

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

This is a lot of code in a function that is too long. Split out the
building code.

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

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

Tidy up this code a little. Also use '-k' consistently, since -m is more
limited in what it can accept.

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

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

This assumes that the GPIO starts as 0 but it does not if
test_gpio_input() ran first and test_gpio_exit_statuses() was skipped.
This can happen when running tests in parallel.

Fix it.

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

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

Mark all the tests in this file as slow, since they take a while.

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

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

When doing a quick check we don't need to run all the vboot tests. Just
run the first one, which is enough to catch most problems.

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

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

This test seems to fail when run in parallel. Mark it single-threaded to
avoid any problems.

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

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

This test seems to interfere with the other test in this file. Mark it
single-threaded to avoid any problems.

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

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

Use a different temporary dir for each test, to allow them to run in
parallel.

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

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

This test seems to rely on the other test in this file. Mark it
single-threaded to avoid any problems.

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

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

Add a new 'singlethread' marker to allow tests to be skipped when running
in parallel.

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

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

This test relies on the silent_linux env variable being set. Add this
to the code so it can run without relying on other bootm tests having been
run first.

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

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

At present test_pinmux_status() assumes that test_pinmux_dev() has run
beforehand. Drop this assumption so we can run the tests in parallel.

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

- DM RTC improvements that should help in CI, allow disabling LTO from
  the make line, add extension (cape, etc) support to distro bootcmd,
  add a pause command and re-enable ARM v4T support.

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

At present the sandbox timer uses localtime() which can jump around twice
a year when daylight-saving time changes.

It would be tricky to make use of gmtime() since we still need to present
the time in local time, as seems to be required by U-Boot's RTC interface.

The problem can only happen once, so use a loop to detect it and try
again. This should be sufficient to detect either a change in the 'second'
value, or a daylight-saving change. We can assume that the latter also
incorporates a 'second' change, so there is no need to loop more than
twice.

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

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

It seems that the time can change in between getting it and reading the
offset. Check for this and try again if this happens.

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

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

Since resetting the RTC on sandbox causes it to read the base time from
the system, we cannot rely on this being unchanged since it was last read.
Allow for a one-second delay.

Reviewed-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Fixes: u-boot/u-boot#4


Reported-by: Bin Meng <bmeng.cn@gmail.com>
Reported-by: Tom Rini <trini@konsulko.com>
Suggested-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Signed-off-by: Simon Glass <sjg@chromium.org>

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

Some tests can have race conditions which are hard to detect on a single
one. Add a way to run tests more than once, to help with this.

Each individual test is run the requested number of times before moving
to the next test. If any runs failed, a message is shown.

This is most useful when running a single test, since running all tests
multiple times can take a while.

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

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

Use this (newish) macro since it is designed for the purpose of making
sure things are non-NULL.

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

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

Check that sandbox builds and runs tests OK with LTO disabled.

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

Simon Glass authored 2 years ago and Tom Rini committed 2 years ago

LTO (Link-Time Optimisation) is an very useful feature which can
significantly reduce the size of U-Boot binaries. So far it has been
made available for selected ARM boards and sandbox.

However, incremental builds are much slower when LTO is used. For example,
an incremental build of sandbox takes 2.1 seconds on my machine, but 6.7
seconds with LTO enabled.

Add a NO_LTO parameter to the build, similar to NO_SDL, so it can be
disabled during development if needed, for faster builds.

Add some documentation about LTO while we are here.

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

Sergei Antonov authored 2 years ago and Tom Rini committed 2 years ago

There is currently a problem that U-Boot can not work on ARMv4
because assembly imlementations of memcpy() and some other functions
use "bx lr" instruction that is not available on ARMv4 ("mov pc, lr"
should be used instead).

A working preprocessor-based solution to this problem is found in
arch/arm/lib/relocate.S. Move it to the "ret" macro in
arch/arm/include/asm/assembler.h and change all "bx lr" code
to "ret lr" in functions that may run on ARMv4. Linux source code
deals with this problem in the same manner.

v1 -> v2:
Comment update. Pointed out by Andre Przywara.

Signed-off-by: Sergei Antonov <saproj@gmail.com>
CC: Samuel Holland <samuel@sholland.org>
CC: Ye Li <ye.li@nxp.com>
CC: Simon Glass <sjg@chromium.org>
CC: Andre Przywara <andre.przywara@arm.com>
CC: Marek Vasut <marex@denx.de>
CC: Sean Anderson <sean.anderson@seco.com>
CC: Tom Rini <trini@konsulko.com>

Samuel Dionne-Riel authored 2 years ago and Tom Rini committed 2 years ago

This command is being introduced with the goal of allowing user-friendly
"generic use case" U-Boot builds to pause until user input under some
situations.

The main use case would be when a boot failure happens, to pause until
the user has had time to acknowledge the current state.

Tested using:

    make && ./u-boot -v -T -c 'ut lib lib_test_hush_pause'

Signed-off-by: Samuel Dionne-Riel <samuel@dionne-riel.com>
Cc: Simon Glass <sjg@chromium.org>

Matwey V. Kornilov authored 2 years ago and Tom Rini committed 2 years ago

Try to load required DTB overlays if the board supports extensions and
CONFIG_CMD_EXTENSION is enabled.

Signed-off-by: Matwey V. Kornilov <matwey.kornilov@gmail.com>

- Assorted Kconfig migrations

Marek Vasut authored 2 years ago and Tom Rini committed 2 years ago

Enable d-cache early in SPL right after DRAM is started up.
This reduces U-Boot proper load time by 650ms when loaded
from SPI NOR.

Signed-off-by: Marek Vasut <marex@denx.de>
Signed-off-by: Philip Oberfichtner <pro@denx.de>

Philip Oberfichtner authored 2 years ago and Tom Rini committed 2 years ago

Introduce the new Kconfig symbol CONFIG_SPL_SYS_L2_PL310 to allow the
SPL to build cache-pl310.c.

Before this commit, the SPL could enable the PL310 L2 cache [1], but the
cache maintenance functions from cache-pl310.c were only useable for
non-SPL builds.

After enabling the cache one must be able to flush it, too. Thus this
commit allows cache-pl310.c to be included in the SPL build.

[1] See for example arch/arm/mach-imx/cache.c: v7_outer_cache_enable()

Signed-off-by: Philip Oberfichtner <pro@denx.de>

Philip Oberfichtner authored 2 years ago and Tom Rini committed 2 years ago

This converts CONFIG_SYS_L2_PL310 to Kconfig.

For omap2 and mvebu the 'select SYS_L2_PL310' locations were
determined using ./tools/moveconfig -i CONFIG_SYS_L2_PL310.

For mx6 I manually chose ARCH_MX6 as 'select' location. The
correctness has been verified using

	$ ./tools/moveconfig.py -f ARCH_MX6 ~SYS_L2_PL310 ~SYS_L2CACHE_OFF
	0 matches

That means whenever an ARCH_MX6 board had SYS_L2_PL310 disabled, this
was correctly reflected in SYS_L2CACHE_OFF. Thus it's safe to insert
the 'select' statement under ARCH_MX6.

Signed-off-by: Philip Oberfichtner <pro@denx.de>

This removes the following symbols:
   CONFIG_SYS_I2C_EEPROM_PAGE_WRITE_BITS
   CONFIG_SYS_I2C_EEPROM_PAGE_WRITE_DELAY_MS
   CONFIG_SYS_I2C_LDI_ADDR
   CONFIG_SYS_I2C_DVI_ADDR
   CONFIG_SYS_I2C_DVI_BUS_NUM

They are unused by any code in tree at this time.

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

This converts the following to Kconfig:
   CONFIG_SYS_I2C_EEPROM_CCID
   CONFIG_SYS_I2C_EEPROM_NXID
   CONFIG_SYS_EEPROM_BUS_NUM

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

- Assorted Arm, TI and Qualcomm platform updates

Andrew Davis authored 2 years ago and Tom Rini committed 2 years ago

We can check if the message was acknowledged in the common
ti_sci_do_xfer() which lets us remove it from after each call to this
function. This simplifies the code and reduces binary size.

Signed-off-by: Andrew Davis <afd@ti.com>

Andrew Davis authored 2 years ago and Tom Rini committed 2 years ago

The inline hint is not needed here, the compiler will do the right thing
based on if we are compiling for speed or for code size. In this case the
inline causes this function to be placed inside each callsite which is
not the right thing to do for either speed nor size. There is no
performance benefit to this due to the larger function size reducing
cache locality, but there is a huge size penalty. Remove inline keyword.

Signed-off-by: Andrew Davis <afd@ti.com>

Andrew Davis authored 2 years ago and Tom Rini committed 2 years ago

We don't need to print the same message in every location, just
print it in the function that fails and remove all the extra
message printouts.

Signed-off-by: Andrew Davis <afd@ti.com>

Andrew Davis authored 2 years ago and Tom Rini committed 2 years ago

This ti_sci_do_xfer() function already prints out the reason for the
failure, and the caller of each of these functions should also notify
the user of the failed task. Remove this extra level of error message.

Signed-off-by: Andrew Davis <afd@ti.com>

Sumit Garg authored 2 years ago and Tom Rini committed 2 years ago

Currently for all Qcom SoCs/boards there are separate compatibles for
GPIO and pinctrl. But this is inconsistent with official (upstream) Linux
bindings which requires only a single compatible "qcom,<SoC name>-pinctrl"
and there is no such compatible property as "qcom,tlmm-<SoC name>".

So fix this inconsistency for Qcom SoCs in order to comply with upstream
DT bindings. This is done via removing compatibles from "msm_gpio" driver
and via binding to "msm_gpio" driver from pinctrl driver in case
"gpio-controller" property is specified for pinctrl node.

Suggested-by: Stephan Gerhold <stephan@gerhold.net>
Signed-off-by: Sumit Garg <sumit.garg@linaro.org>

Sumit Garg authored 2 years ago and Tom Rini committed 2 years ago

DT compatible is sufficient to make platform specific differentiation,
so remove redundant CONFIG_SDM845 check.

Signed-off-by: Sumit Garg <sumit.garg@linaro.org>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>