Building U-Boot with unit tests on a non-sandbox systems fails:
ld.bfd: test/built-in.o: in function `compress_using_gzip':
test/compression.c:138: undefined reference to `gzip'
ld.bfd: test/built-in.o: in function `uncompress_using_bzip2':
test/compression.c:187: undefined reference to `BZ2_bzBuffToBuffDecompress'
ld.bfd: test/built-in.o: in function `uncompress_using_lzma':
test/compression.c:222: undefined reference to `lzmaBuffToBuffDecompress'
ld.bfd: test/built-in.o: in function `uncompress_using_lzo':
test/compression.c:257: undefined reference to `lzop_decompress'
ld.bfd: test/built-in.o: in function `uncompress_using_lz4':
test/compression.c:292: undefined reference to `ulz4fn
Add the missing dependencies.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de> Reviewed-by: Simon Glass <sjg@chromium.org>
cros_ec: Support keyboard scanning with EC_CMD_GET_NEXT_EVENT
The cros_ec_keyb driver currently uses EC_CMD_MKBP_STATE to scan the
keyboard, but this host command was superseded by EC_CMD_GET_NEXT_EVENT
and unavailable on more recent devices (including gru-kevin), as it was
removed in cros-ec commit 87a071941b89 ("mkbp: Add support for buttons
and switches.") dated 2016-07-06.
The EC_CMD_GET_NEXT_EVENT has been available since cros-ec commit d1ed75815efe ("MKBP event signalling implementation") dated 2014-10-20,
but it looks like it isn't included in firmware-* branches for at least
link, nyan-big, samus, snow, spring, panther and peach-pit which have
defconfigs in U-Boot. So this patch falls back to the old method if the
EC doesn't recognize the newer command.
The implementation is mostly adapted from Depthcharge commit f88af26b44fc ("cros_ec: Change keyboard scanning method.").
On a gru-kevin, the current driver before this patch fails to read the
pressed keys with:
Michal Simek [Wed, 26 Aug 2020 13:34:24 +0000 (15:34 +0200)]
binman: Fix typo in kernel-doc format for binman_symname()
Fix typo.
Fixes: 19790632648b ("binman: Support accessing binman tables at run time") Signed-off-by: Michal Simek <michal.simek@xilinx.com> Reviewed-by: Simon Glass <sjg@chromium.org> Reviewed-by: Simon Glass <sjg@chromium.org>
Up to now the sandbox would shutdown upon a cold reset request. Instead it
should be reset.
In our coding we use static variables like LIST_HEAD(efi_obj_list). A reset
can occur at any time, e.g. via an UEFI binary calling the reset service.
The only safe way to return to an initial state is to relaunch the U-Boot
binary.
The reset implementation uses execv() to relaunch U-Boot.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de> Reviewed-by: Simon Glass <sjg@chromium.org>
The command to shut down a device is 'poweroff'. It is a deficit of the
sandbox that it does not support resetting yet but shuts down upong seeing
the 'reset' command.
Once the sandbox properly supports reset we need the 'poweroff' command to
leave the sandbox.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de> Reviewed-by: Simon Glass <sjg@chromium.org>
Michal Simek [Wed, 4 Nov 2020 14:33:20 +0000 (15:33 +0100)]
arch: Move NEEDS_MANUAL_RELOC symbol to Kconfig
CONFIG_NEEDS_MANUAL_RELOC macro was out of Kconfig. Move it there to be
able to use compile-time checks to reduce the number of build paths.
Fixes: f9a882438966 ("dm: core: Convert #ifdef to if() in root.c") for Microblaze Reviewed-by: Tom Rini <trini@konsulko.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Heiko Schocher [Tue, 3 Nov 2020 14:22:36 +0000 (15:22 +0100)]
env: sf: fix init function behaviour
Michael wrote:
commit 92765f45bb95 ("env: Access Environment in SPI flashes before
relocation") at least breaks the Kontron sl28 board. I guess it also
breaks others which use a (late) SPI environment.
reason is, that env_init() sets the init bit, if there
is no init function defined in an environment driver,
and use default return value -ENOENT in this case
later for setting the default environment.
Change:
Environment driver can now implement an init
function and return, if this function does nothing,
simply -ENOENT.
env_init() now handles -ENOENT correct by setting the
inited bit for the environment driver. And if there
is no other environment driver whose init function
returns 0, load than the default environment.
This prevents that each environment driver needs to set the
default environment.
Fixes: 92765f45bb95 ("env: Access Environment in SPI flashes before relocation") Reported-by: Michael Walle <michael@walle.cc> Tested-by: Michael Walle <michael@walle.cc> [For the SF environment] Signed-off-by: Heiko Schocher <hs@denx.de>
Eugen Hristev [Mon, 26 Oct 2020 08:39:54 +0000 (10:39 +0200)]
ARM: dts: at91: sama5d3xmb_cmp: fix SPI bus unit address
w+arch/arm/dts/.at91sam9260ek.dtb.pre.tmp:119.21-123.7: Warning (spi_bus_reg): /ahb/apb/spi@fffc8000/mtd_dataflash@0: SPI bus unit address format error, expected "1"
Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
Eugen Hristev [Mon, 26 Oct 2020 08:39:53 +0000 (10:39 +0200)]
ARM: dts: at91: sam9260ek: fix SPI bus unit address
w+arch/arm/dts/.at91sam9260ek.dtb.pre.tmp:119.21-123.7: Warning (spi_bus_reg): /ahb/apb/spi@fffc8000/mtd_dataflash@0: SPI bus unit address format error, expected "1"
Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
Eugen Hristev [Mon, 26 Oct 2020 08:39:51 +0000 (10:39 +0200)]
ARM: dts: at91: gurnard: fix SPI bus unit address
w+arch/arm/dts/.at91sam9g45-gurnard.dtb.pre.tmp:118.21-122.7: Warning (spi_bus_reg): /ahb/apb/spi@fffa4000/mtd_dataflash@0: SPI bus unit address format error, expected "1"
Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
Eugen Hristev [Mon, 26 Oct 2020 08:39:50 +0000 (10:39 +0200)]
ARM: dts: at91: at91sam9g25ek: fix I2C bus unit address
w+arch/arm/dts/.at91sam9g25ek.dtb.pre.tmp:28.25-47.7: Warning (i2c_bus_reg): /ahb/apb/i2c@f8010000/camera@0x30: I2C bus unit address format error, expected "30"
Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
Eugen Hristev [Mon, 26 Oct 2020 08:39:49 +0000 (10:39 +0200)]
ARM: dts: at91: at91sam9g20ek_common: fix SPI bus unit address
w+arch/arm/dts/at91sam9g20ek_common.dtsi:100.21-104.7: Warning (spi_bus_reg): /ahb/apb/spi@fffc8000/mtd_dataflash@0: SPI bus unit address format error, expected "1"
Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
Eugen Hristev [Mon, 26 Oct 2020 08:39:48 +0000 (10:39 +0200)]
ARM: dts: at91: at91sam9g20-taurus: fix SPI bus unit address
w+arch/arm/dts/.at91sam9g20-taurus.dtb.pre.tmp:79.18-83.4: Warning (spi_bus_reg): /ahb/apb/spi@fffc8000/mtd_dataflash@0: SPI bus unit address format error, expected "1"
Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
Eugen Hristev [Mon, 26 Oct 2020 08:39:47 +0000 (10:39 +0200)]
ARM: dts: at91: at91sam9261ek: fix SPI unit address warning
w+arch/arm/dts/.at91sam9261ek.dtb.pre.tmp:124.15-144.7: Warning (spi_bus_reg): /ahb/apb/spi@fffc8000/tsc2046@0: SPI bus unit address format error, expected "2"
Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
Eugen Hristev [Mon, 26 Oct 2020 08:39:46 +0000 (10:39 +0200)]
ARM: dts: at91: vinco: fix I2C warning bus unit address
w+arch/arm/dts/.at91-vinco.dtb.pre.tmp:131.18-134.7: Warning (i2c_bus_reg): /ahb/apb/i2c@f8024000/rtc@64: I2C bus unit address format error, expected "32"
Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
Tom Rini [Sat, 31 Oct 2020 02:55:16 +0000 (22:55 -0400)]
Merge tag 'u-boot-rockchip-20201031' of https://gitlab.denx.de/u-boot/custodians/u-boot-rockchip
- New PX30 board: Engicam PX30.Core;
- Fix USB HID support for rock960;
- Remove host endianness dependency for rockchip mkimage;
- dts update for rk3288-tinker;
- Enable console MUX for some ROCKPi boards;
- Add config-based ddr selection for px30;
Jagan Teki [Wed, 28 Oct 2020 13:33:44 +0000 (19:03 +0530)]
board: engicam: Attach i.MX6 common code
The existing common code for Engicam boards uses i.MX6,
so attach that into i.MX6 Engicam boards so-that adding
new SoC variants of Engicam boards become meaningful.
PX30.Core is an EDIMM SOM based on Rockchip PX30 from Engicam.
General features:
- Rockchip PX30
- Up to 2GB DDR4
- eMMC 4 GB expandible
- rest of PX30 features
PX30.Core needs to mount on top of Engicam baseboards for creating
complete platform boards.
Possible baseboards are,
- EDIMM2.2
- C.TOUCH 2.0
Add support for it.
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com> Signed-off-by: Michael Trimarchi <michael@amarulasolutions.com> Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Engicam EDIMM2.2 Starter Kit is an EDIMM 2.2 Form Factor Capacitive
Evaluation Board.
Genaral features:
- LCD 7" C.Touch
- microSD slot
- Ethernet 1Gb
- Wifi/BT
- 2x LVDS Full HD interfaces
- 3x USB 2.0
- 1x USB 3.0
- HDMI Out
- Mini PCIe
- MIPI CSI
- 2x CAN
- Audio Out
SOM's like PX30.Core needs to mount on top of this Evaluation board
for creating complete PX30.Core EDIMM2.2 Starter Kit.
Add support for it.
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com> Signed-off-by: Michael Trimarchi <michael@amarulasolutions.com> Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
The Rockchip boot ROM expects little-endian values in the image header.
When running mkimage on a big-endian machine, these values need to be
byteswapped before writing or verifying the header.
This change fixes cross-compiling U-Boot SPL for the RK3399 SoC from a
big-endian ppc64 host machine.
Signed-off-by: Samuel Holland <samuel@sholland.org> Reviewed-by: Kever Yang<kever.yang@rock-chips.com>
Sean Anderson [Tue, 27 Oct 2020 23:55:41 +0000 (19:55 -0400)]
doc: Update logging documentation
This updates logging documentation with some examples of the new commands
added in the previous commits. It also removes some items from the to-do
list which have been implemented.
Signed-off-by: Sean Anderson <seanga2@gmail.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Sean Anderson [Tue, 27 Oct 2020 23:55:39 +0000 (19:55 -0400)]
test: Add a test for log filter-*
This exercises a few success and failure modes of the log filter-*
commands. log filter-list is not tested because it's purely informational.
I don't think there's a good way to test it except by testing if the output
of the command exactly matches a sample run.
Signed-off-by: Sean Anderson <seanga2@gmail.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Sean Anderson [Tue, 27 Oct 2020 23:55:38 +0000 (19:55 -0400)]
cmd: log: Add commands to manipulate filters
This adds several commands to add, list, and remove log filters. Due to the
complexity of adding a filter, `log filter-list` uses options instead of
positional arguments.
These commands have been added as subcommands to log by using a dash to
join the subcommand and subsubcommand. This is stylistic, and they could be
converted to proper subsubcommands if it is wished.
Signed-off-by: Sean Anderson <seanga2@gmail.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Sean Anderson [Tue, 27 Oct 2020 23:55:37 +0000 (19:55 -0400)]
test: Add a test for getopt
A few of these tests were inspired by those in glibc. The syntax for
invoking test_getopt is a bit funky, but it's necessary so that the CPP can
parse the arguments correctly.
Sean Anderson [Tue, 27 Oct 2020 23:55:36 +0000 (19:55 -0400)]
lib: Add getopt
Some commands can get very unweildy if they have too many positional
arguments. Adding options makes them easier to read, remember, and
understand.
This implementation of getopt has been taken from barebox, which has had
option support for quite a while. I have made a few modifications to their
version, such as the removal of opterr in favor of a separate getopt_silent
function. In addition, I have moved all global variables into struct
getopt_context.
The getopt from barebox also re-orders the arguments passed to it so that
non-options are placed last. This allows users to specify options anywhere.
For example, `ls -l foo/ -R` would be re-ordered to `ls -l -R foo/` as
getopt parsed the options. However, this feature conflicts with the const
argv in cmd_tbl->cmd. This was originally added in 54841ab50c ("Make sure
that argv[] argument pointers are not modified."). The reason stated in
that commit is that hush requires argv to stay unmodified. Has this
situation changed? Barebox also uses hush, and does not have this problem.
Perhaps we could use their fix?
I have assigned maintenance of getopt to Simon Glass, as it is currently
only used by the log command. I would also be fine maintaining it.
Sean Anderson [Tue, 27 Oct 2020 23:55:35 +0000 (19:55 -0400)]
cmd: log: Make "log level" print all log levels
This makes the log level command print all valid log levels. The default
log level is annotated. This provides an easy way to see which log levels
are compiled-in.
Signed-off-by: Sean Anderson <seanga2@gmail.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Sean Anderson [Tue, 27 Oct 2020 23:55:34 +0000 (19:55 -0400)]
cmd: log: Add commands to list categories and drivers
This allows users to query which categories and drivers are available on
their system. This allows them to construct filter-add commands without
(e.g.) adjusting the log format to show categories and drivers.
Signed-off-by: Sean Anderson <seanga2@gmail.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Sean Anderson [Tue, 27 Oct 2020 23:55:33 +0000 (19:55 -0400)]
cmd: log: Split off log level parsing
Move parsing of log level into its own function so it can be re-used. This
also adds support for using log level names instead of just the integer
equivalent.
Signed-off-by: Sean Anderson <seanga2@gmail.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Sean Anderson [Tue, 27 Oct 2020 23:55:30 +0000 (19:55 -0400)]
log: Add filter flag to match greater than a log level
This is the complement of the existing behavior to match only messages with
a log level less than a threshold. This is primarily useful in conjunction
with LOGFF_DENY.
Signed-off-by: Sean Anderson <seanga2@gmail.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Sean Anderson [Tue, 27 Oct 2020 23:55:27 +0000 (19:55 -0400)]
test: log: Convert log_test from python to C
When rebasing this series I had to renumber all my log tests because
someone made another log test in the meantime. This involved updaing a
number in several places (C and python), and it wasn't checked by the
compiler. So I though "how hard could it be to just rewrite in C?" And
though it wasn't hard, it *was* tedious. Tests are numbered the same as
before to allow for easier review.
A note that if a test fails, everything after it will probably also fail.
This is because that test won't clean up its filters. There's no easy way
to do the cleanup, except perhaps removing all filters in a wrapper
function.
Sean Anderson [Tue, 27 Oct 2020 23:55:26 +0000 (19:55 -0400)]
log: Add filter flag to deny on match
Without this flag, log filters can only explicitly accept messages.
Allowing denial makes it easier to filter certain subsystems. Unlike
allow-ing filters, deny-ing filters are added to the beginning of the
filter list. This should do the Right Thing most of the time, but it's
less-universal than allowing filters to be inserted anywhere. If this
becomes a problem, then perhaps log_filter_add* should take a filter number
to insert before/after.
Signed-off-by: Sean Anderson <seanga2@gmail.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Sean Anderson [Tue, 27 Oct 2020 23:55:21 +0000 (19:55 -0400)]
log: Fix incorrect documentation of log_filter.cat_list
Logging category lists are terminated by LOGC_END, not LOGC_NONE.
Fixes: e9c8d49d54 ("log: Add an implementation of logging") Signed-off-by: Sean Anderson <seanga2@gmail.com> Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Sean Anderson [Tue, 27 Oct 2020 23:55:20 +0000 (19:55 -0400)]
log: Fix missing negation of ENOMEM
Errors returned should be negative.
Fixes: 45fac9fc18 ("log: Correct missing free() on error in log_add_filter()") Signed-off-by: Sean Anderson <seanga2@gmail.com> Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Heiko Schocher [Sat, 10 Oct 2020 08:28:06 +0000 (10:28 +0200)]
imx6: enable early spi environment access on aristainetos boards
On aristianetos boards the display type is detected
through "panel" environment variable. Dependend on the
display type we detect the board type and this decides which
DTB we have to use for the board.
Heiko Schocher [Sat, 10 Oct 2020 08:28:05 +0000 (10:28 +0200)]
env: Access Environment in SPI flashes before relocation
Enable the new Kconfig option ENV_SPI_EARLY if you want
to use Environment in SPI flash before relocation.
Call env_init() and than you can use env_get_f() for
accessing Environment variables.
Signed-off-by: Heiko Schocher <hs@denx.de> Reviewed-by: Simon Glass <sjg@chromium.org>
Heiko Schocher [Sat, 10 Oct 2020 08:28:04 +0000 (10:28 +0200)]
env: split env_import_redund() into 2 functions
split from env_import_redund() the part which checks
which Environment is valid into a separate function
called env_check_redund() and call it from env_import_redund().
So env_check_redund() can be used from places which also
need to do this checks.
Signed-off-by: Heiko Schocher <hs@denx.de> Reviewed-by: Simon Glass <sjg@chromium.org>
AKASHI Takahiro [Thu, 29 Oct 2020 04:47:46 +0000 (13:47 +0900)]
efi_loader: add efi_create_indexed_name()
This function will be used from several places in UEFI subsystem
to generate some specific form of utf-16 variable name.
For example, L"Capsule0001"
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Move function to separate module.
Use char * as argument instead of u16 *. Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de> Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
AKASHI Takahiro [Thu, 29 Oct 2020 04:47:41 +0000 (13:47 +0900)]
dfu: rename dfu_tftp_write() to dfu_write_by_name()
This function is essentially independent from tftp, and will also be
utilised in implementing UEFI capsule update in a later commit.
So just give it a more generic name.
In addition, a new configuration option, CONFIG_DFU_WRITE_ALT, was
introduced so that the file will be compiled with different options,
particularly one added in a later commit.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org> Reviewed-by: Tom Rini <trini@konsulko.com>
rtc: initialize emulated RTC from environment variable
Up to now the emulated RTC is initialized using the U-Boot build time.
With this patch the environment variable 'rtc_emul_epoch' can be used to
provide a better initial time. The variable is a decimal string with
the number of seconds since 1970-01-01. Here is an example where the RTC
had not been probed yet:
If the variable does not exist, the U-Boot build time is used as fallback.
The environment variable may be set when shutting down the operating system
if the U-Boot environment is exposed to the OS (cf. ENV_IS_IN_FAT and
ENV_IS_IN_EXT4).
Suggested-by: Pablo Sebastián Greco <pgreco@centosproject.org> Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
video: rockchip: Add missing dpcd_write() call to link_train_ce()
Found this by comparing it to the coreboot driver, a form of this call
was introduced there in their commit b9a7877568cf ("rockchip/*: refactor
edp driver"). This is copy-pasted from U-Boot's link_train_cr() slightly
above it.
Without this on a gru-kevin chromebook, I have:
clock recovery at voltage 0 pre-emphasis 0
requested signal parameters: lane 0 voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 1 voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 2 voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 3 voltage 0.4V pre_emph 3.5dB
using signal parameters: voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 0 voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 1 voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 2 voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 3 voltage 0.4V pre_emph 3.5dB
using signal parameters: voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 0 voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 1 voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 2 voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 3 voltage 0.4V pre_emph 3.5dB
using signal parameters: voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 0 voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 1 voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 2 voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 3 voltage 0.4V pre_emph 3.5dB
using signal parameters: voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 0 voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 1 voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 2 voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 3 voltage 0.4V pre_emph 3.5dB
using signal parameters: voltage 0.4V pre_emph 3.5dB
channel eq failed, ret=-5
link train failed!
rk_vop_probe() Device failed: ret=-5
With this, it looks like training succeeds:
clock recovery at voltage 0 pre-emphasis 0
requested signal parameters: lane 0 voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 1 voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 2 voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 3 voltage 0.4V pre_emph 3.5dB
using signal parameters: voltage 0.4V pre_emph 3.5dB
requested signal parameters: lane 0 voltage 0.4V pre_emph 6dB
requested signal parameters: lane 1 voltage 0.4V pre_emph 6dB
requested signal parameters: lane 2 voltage 0.4V pre_emph 6dB
requested signal parameters: lane 3 voltage 0.4V pre_emph 6dB
using signal parameters: voltage 0.4V pre_emph 6dB
requested signal parameters: lane 0 voltage 0.4V pre_emph 0dB
requested signal parameters: lane 1 voltage 0.4V pre_emph 0dB
requested signal parameters: lane 2 voltage 0.4V pre_emph 0dB
requested signal parameters: lane 3 voltage 0.4V pre_emph 0dB
using signal parameters: voltage 0.4V pre_emph 0dB
channel eq at voltage 0 pre-emphasis 0
config video failed
rk_vop_probe() Device failed: ret=-110
The "config video failed" error also goes away when I disable higher
log levels, and it claims to have successfully probed the device.
In order to correctly calculate the designware watchdog
timeouts, the watchdog clock is required. Implement required
clocks to facilitate this.
Signed-off-by: Jack Mitchell <ml@embed.me.uk> Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Simon Glass [Mon, 26 Oct 2020 23:40:26 +0000 (17:40 -0600)]
binman: Avoid calculated section data repeatedly
Refactor the implementation slightly so that section data is not
rebuilt when it is already available.
We still have GetData() set up to rebuild the section, since we don't
currently track when things change that might affect a section. For
example, if a blob is updated within a section, we must rebuild it.
Tracking that would be possible but is more complex, so it left for
another time.
Simon Glass [Mon, 26 Oct 2020 23:40:25 +0000 (17:40 -0600)]
binman: Support compression of sections
With the previous changes, it is now possible to compress entire
sections. Add some tests to check that compression works correctly,
including updating the metadata.
Simon Glass [Mon, 26 Oct 2020 23:40:24 +0000 (17:40 -0600)]
binman: Use the actual contents in CheckSize()
At present this function adds up the total size of entries to work out the
size of a section's contents. With compression this is no-longer enough.
We may as well bite the bullet and build the section contents instead.
Call _BuildSectionData() to get the (possibly compressed) contents and
GetPaddedData() to get the same but with padding added.
Note that this is inefficient since the section contents is calculated
twice. Future work will improve this.
This affects testPackOverlapMap() since the error is reported with a
different section size now (enough to hold the contents). Update that at
the same time.
Simon Glass [Mon, 26 Oct 2020 23:40:23 +0000 (17:40 -0600)]
binman: Update CheckEntries() for compressed sections
At present this function assumes that the size of a section is at least as
large as its contents. With compression this is often not the case. Relax
this constraint by using the uncompressed size, if available.
Simon Glass [Mon, 26 Oct 2020 23:40:22 +0000 (17:40 -0600)]
binman: Drop CheckEntries()
This method introduces a separation between packing and checking that is
different for sections. In order to handle compression properly, we need
to be able to deal with a section's size being smaller than the
uncompressed size of its contents. It is easier to make this work if
everything happens in the Pack() method.
The only real user of CheckEntries() is entry_Section and it can call it
directly. Drop the call from 'control' and handle it locally.
Simon Glass [Mon, 26 Oct 2020 23:40:21 +0000 (17:40 -0600)]
binman: Call CheckSize() from the section's Pack() method
At present CheckSize() is called from the function that packs the entries.
Move it up to the main Pack() function so that _PackEntries() can just
do the packing.
Simon Glass [Mon, 26 Oct 2020 23:40:19 +0000 (17:40 -0600)]
binman: Move sort and expand to the main Pack() function
At present sorting and expanding entries are side-effects of the
CheckEntries() function. This is a bit confusing, as 'checking' would
not normally involve making changes.
Move these steps into the Pack() function instead.
Simon Glass [Mon, 26 Oct 2020 23:40:17 +0000 (17:40 -0600)]
binman: Avoid reporting image-pos with compression
When a section is compressed, all entries within it are grouped together
into a compressed block of data. This obscures the start of each
individual child entry.
Avoid reporting bogus 'image-pos' properties in this case, since it is
not possible to access the entry at the location provided. The entire
section must be decompressed first.
CBFS does not support compressing whole sections, only individual files,
so needs no special handling here.
Simon Glass [Mon, 26 Oct 2020 23:40:16 +0000 (17:40 -0600)]
binman: Set section contents in GetData()
Section contents is not set up when ObtainContents() is called, since
packing often changes the layout of the contents. Ensure that the contents
are correctly recorded by making this function regenerate the section. It
is normally only called by the parent section (when packing) or by the
top-level image code, when writing out the image. So the performance
impact is fairly small.
Now that sections have their contents in their 'data' property, update
testSkipAtStartSectionPad() to check it.
Simon Glass [Mon, 26 Oct 2020 23:40:14 +0000 (17:40 -0600)]
binman: Make section padding consistent with other entries
At present padding of sections is inconsistent with other entry types, in
that different pad bytes are used.
When a normal entry is padded by its parent, the parent's pad byte is
used. But for sections, the section's pad byte is used.
Adjust logic to always do this the same way.
Note there is still a special case in entry_Section.GetPaddedData() where
an image is padded with the pad byte of the top-level section. This is
necessary since otherwise there would be no way to set the pad byte of
the image, without adding a top-level section to every image.
Simon Glass [Mon, 26 Oct 2020 23:40:13 +0000 (17:40 -0600)]
binman: Move section padding to the parent
Each section is padded up to its size, if the contents are not large
enough. Move this logic from _BuildSectionData() to
GetPaddedDataForEntry() so that all the padding is in one place.
With this, the testDual test is working again, so enable it.
projects / u-boot.git / log