In case when the environment on some location is malformed (CRC isn't
matching), there is a chance we won't be able to save the environment to
that location. For example, consider the case when we only have the
environment on eMMC, but it's zeroed out. In that case, we won't be able
to "env save" to it, because of "bad CRC" error. That's happening
because in env_load() function we consider malformed environment as
incorrect one, and defaulting to the location with highest (0)
priority, which can be different from one we are dealing with right now
(e.g., highest priority can be ENV_FAT on SD card, which is not
inserted, but we want to use ENV_MMC on eMMC, where we were booted
from).
This issue began to reproduce after commit d30ba2315a ("u-boot: remove
driver lookup loop from env_save()") on BeagleBone Black, but that
commit didn't introduce the wrong logic, it just changed the behavior
for default location to use, merely revealing this issue.
To fix that, let's implement next logic in env_load():
1. Try to find out correct environment; if found -- use it
2. If working environment wasn't found, but we found malformed one
(with bad CRC), let's use it for further "env save". But make sure
to use malformed environment location with highest priority.
3. If neither correct nor malformed environment was found, let's
default to environment location with highest priority (0)
Steps to reproduce mentioned issue on BeagleBone Black (fixed in this
patch):
1. Boot from SD card and erase eMMC in U-Boot shell:
=> mmc dev 1
=> mmc erase 0 100000
=> gpt write mmc 1 $partitions
2. Write new SPL and U-Boot to eMMC; the rest of eMMC will stay filled
with zeroes
3. Boot from eMMC; try to do:
=> env save
4. Observe the error (incorrect behavior). Correct behavior: environment
should be stored correctly on eMMC, in spite of it has "bad CRC"
Fixes: d30ba2315a ("u-boot: remove driver lookup loop from env_save()")
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Reviewed-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
Callers of env_import*() functions might want to check the case when we
have incorrect environment (with bad CRC). For example, when environment
location is being defined in env_load(), call chain may look like this:
env_load() -> drv->load() = env_mmc_load() -> env_import()
Return code will be passed from env_import() all way up to env_load().
Right now both env_mmc_load() and env_import() return -EIO error code,
so env_load() can't differentiate between two cases:
1. Driver reports the error, because device is not accessible
2. Device is actually accessible, but environment is broken
Let's return -ENOMSG in env_import(), so we can distinguish two cases
mentioned above. It will make it possible to continue working with "bad
CRC" environment (like doing "env save"), instead of considering it not
functional (implemented in subsequent patch).
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Reviewed-by: Simon Goldschmidt <simon.k.r.goldschmidt@gmail.com>
Join the party of some ARM boards and drop more
items from include/configs/xxx.h.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Michal Simek <michal.simek@xilinx.com>
Cc: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
Per Heiko the original changes were correct and something is misbehaving
on his hardware.
This reverts commit 3d5931e598.
Signed-off-by: Tom Rini <trini@konsulko.com>
commit 9a9d66f5ef ("env: add spi_flash_read_env function")
breaks Environment functionality, as it reads only
until 2 \0 are found, but fills the buffer with 0x0
instead 0xff which leads in an incorrect crc sum.
Fix: init the read buffer with 0xff instead 0x00
Signed-off-by: Heiko Schocher <hs@denx.de>
The spi_flash_read_env function is a wrapper over spi_flash_read, which
enables the env to read multiple flash page size from flash until '\0\0'
is read or the end of env partition is reached. Instead of reading the
entire env size. When it reads '\0\0', it stops reading further the env
and assumes that the rest of env is '\0'.
This is an optimization for large environments that contain few bytes
environment variables. In this case it doesn't need to read the entire
environment and only few pages.
Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Add the dollar_complete() function to auto-complete arguments starting
with a '$' and use it in the cmd_auto_complete() path such that all
args starting with a $ can be auto-completed based on the available env
vars.
Signed-off-by: Boris Brezillon <boris.brezillon@bootlin.com>
[trini: Fix some linking problems]
Signed-off-by: Tom Rini <trini@konsulko.com>
The env location label ENVL_ESATA is missed in location tables, so
when we configure the ENV in SATA, u-boot fails to get correct env
location and cause boot hang in board_f.
Signed-off-by: Ye Li <ye.li@nxp.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Defines env_sf_get_env_addr API to override sf environment address,
required to support multiple environment.
Signed-off-by: Rajesh Bhagat <rajesh.bhagat@nxp.com>
Reviewed-by: York Sun <york.sun@nxp.com>
Define env_ptr as static in flash and nand env driver to
allow these to compile together.
Signed-off-by: Rajesh Bhagat <rajesh.bhagat@nxp.com>
Reviewed-by: York Sun <york.sun@nxp.com>
Xilinx is introducing Versal, an adaptive compute acceleration platform
(ACAP), built on 7nm FinFET process technology. Versal ACAPs combine
Scalar Processing Engines, Adaptable Hardware Engines, and Intelligent
Engines with leading-edge memory and interfacing technologies to deliver
powerful heterogeneous acceleration for any application. The Versal AI
Core series has five devices, offering 128 to 400 AI Engines. The series
includes dual-core Arm Cortex™-A72 application processors, dual-core Arm
Cortex-R5 real-time processors, 256KB of on-chip memory with ECC, more
than 1,900 DSP engines optimized for high-precision floating point with
low latency.
The patch is adding necessary infrastructure in place without enabling
platform which is done in separate patch.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
For some reason the spi_flash_probe_bus_cs() is called
inside the setup_flash_device() with zero values in place
of configurated SPI flash mode and maximum flash speed.
This code causes HALT error during startup environment
relocation on some platforms - namely Armada-38x-GP board.
Fix the function call by replacing zeros with the appropriate
values - CONFIG_ENV_SPI_MAX_HZ and CONFIG_ENV_SPI_MODE.
Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
Cc: Igal Liberman <igall@marvell.com>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>
Add missing environment name for UBI, to prevent this NULL in output:
Loading Environment from <NULL>...
and rather have a valid UBI there:
Loading Environment from UBI...
Signed-off-by: Marek Vasut <marex@denx.de>
There is no reason to have the same Kconfig options for different SoCs
separately. The patch is merging them together.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
[trini: Fix ENV_SIZE around ENV_IS_NOWHERE]
Signed-off-by: Tom Rini <trini@konsulko.com>
Signed-off-by: Tom Rini <trini@konsulko.com>
"Failed" error message from env_load() only clutters the log with
unnecessary details, as we already have all needed warnings by that
time. Example:
Loading Environment from FAT... MMC: no card present
** Bad device mmc 0 **
Failed (-5)
Let's only print it in case when DEBUG is defined to keep log clear.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
When called with ENVOP_SAVE, env_get_location() only returns the
gd->env_load_location variable without actually checking for
the environment location and priority.
This behaviour causes env_save() to fall into an infinite loop when
the low-level drv->save() call fails.
The env_save() function should not loop through the environment
location list but it should save the environment into the location
stored in gd->env_load_location by the last env_load() call.
Signed-off-by: Nicholas Faustini <nicholas.faustini@azcomtech.com>
Reviewed-by: Simon Goldschmidt <sgoldschmidt@de.pepperl-fuchs.com>
Add Kconfig options SPL_ENV_* and TPL_ENV_* and simplify Makefile.
This allows SPL/TPL image has different environment setting from
full feature U-Boot.
Signed-off-by: York Sun <york.sun@nxp.com>
The error message should start with `## Error: ` so that it's easily
detectable by tests without needing to have a complex regexp for
matching all possible error message patterns.
Let's add the `## Error: ` prefix to the error messages since it's the
one already in use.
Suggested-by: Stephen Warren <swarren@wwwdotorg.org>
Signed-off-by: Quentin Schulz <quentin.schulz@bootlin.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
The function set_default_env() sets the hashtable flags for import_r().
Formally set_default_env() doesn't accept flags from its callers. In
practice the caller can (un)set the H_INTERACTIVE flag, but it has to be
done using the first character of the function's string argument. Other
flags like H_FORCE can't be set by the caller.
Change the function to accept flags argument. The benefits are:
1. The caller will have to explicitly set the H_INTERACTIVE flag,
instead of un-setting it using a special char in a string.
2. Add the ability to propagate flags from the caller to himport(),
especially the H_FORCE flag from do_env_default() in nvedit.c that
currently gets ignored for "env default -a -f" commands.
3. Flags and messages will not be coupled together. A caller will be
able to set flags without passing a string and vice versa.
Please note:
The propagation of H_FORCE from do_env_default() does not introduce any
functional changes, because currently himport_r() is set to destroy the
old environment regardless if H_FORCE flag is set or not. More changes
are needed to utilize the propagation of H_FORCE.
Signed-off-by: Yaniv Levinsky <yaniv.levinsky@compulab.co.il>
Acked-by: Igor Grinberg <grinberg@compulab.co.il>
The function set_default_vars() in common.c adds H_INTERACTIVE to the
h_import() flag, but the function has no way of telling if the command
actually was user directed like this flag suggest. The flag should be
set by the calling function do_env_default() in nvedit.c instead, where
the command is certainty user directed.
Move the H_INTERACTIVE flag from set_default_vars() to do_env_default().
Signed-off-by: Yaniv Levinsky <yaniv.levinsky@compulab.co.il>
Acked-by: Igor Grinberg <grinberg@compulab.co.il>
The env_flag in do_env_default() doesn't get propagated and therefore
gets ignored by himport_r(). This breaks to ability to "forcibly" reset
variables to their default values using the environment command.
Scenario example of the problem:
# setenv kernel uImage
# setenv .flags kernel:so
# env default -f kernel
## Error: Can't overwrite "kernel"
himport_r: can't insert "kernel=zImage" into hash table
Change the call path so it will pass the flag correctly.
Signed-off-by: Yaniv Levinsky <yaniv.levinsky@compulab.co.il>
Acked-by: Igor Grinberg <grinberg@compulab.co.il>
This patch added support to enable CONFIG_ENV_SIZE, CONFIG_ENV_OFFSET
and CONFIG_ENV_SECT_SIZE through Kconfig for Zynq and Zynqmp.
Signed-off-by: Vipul Kumar <vipul.kumar@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
If ENV_IS_IN_EXT4 is set you shouldn't be able to select ENV_IS_NOWHERE.
Signed-off-by: Alex Kiernan <alex.kiernan@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested-by: Petr Vorel <petr.vorel@gmail.com>
When U-Boot started using SPDX tags we were among the early adopters and
there weren't a lot of other examples to borrow from. So we picked the
area of the file that usually had a full license text and replaced it
with an appropriate SPDX-License-Identifier: entry. Since then, the
Linux Kernel has adopted SPDX tags and they place it as the very first
line in a file (except where shebangs are used, then it's second line)
and with slightly different comment styles than us.
In part due to community overlap, in part due to better tag visibility
and in part for other minor reasons, switch over to that style.
This commit changes all instances where we have a single declared
license in the tag as both the before and after are identical in tag
contents. There's also a few places where I found we did not have a tag
and have introduced one.
Signed-off-by: Tom Rini <trini@konsulko.com>
We have a large number of places where while we historically referenced
gd in the code we no longer do, as well as cases where the code added
that line "just in case" during development and never dropped it.
Signed-off-by: Tom Rini <trini@konsulko.com>
For CONFIG_ENV_FAT_INTERFACE != 'mmc' a link error
env/fat.c:93: undefined reference to `mmc_initialize'
occurs if CONFIG_MMC_SUPPORT is not enabled.
Fixes: 26862b4a40 ("env: mmc/fat/ext4: make sure that the MMC sub-system
is initialized before using it")
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Relocate env drivers if manual relocation is enabled. This
patch fixes the issue of u-boot hang incase if env is
present in any of the flash devices.
Signed-off-by: Siva Durga Prasad Paladugu <sivadur@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Convert CONFIG_ENV_VARS_UBOOT_RUNTIME_CONFIG to Kconfig
Signed-off-by: Alex Kiernan <alex.kiernan@gmail.com>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
Reviewed-by: Petr Vorel <petr.vorel@gmail.com>
Reviewed-by: Petr Vorel <petr.vorel at gmail.com>
Use CONFIG_IS_ENABLED to see if CONFIG_BLK is enabled. Otherwise
SPL compilation breaks on boards which do have CONFIG_BLK enabled but
not DM_MMC for the SPL as follows:
env/mmc.c: In function ‘init_mmc_for_env’:
env/mmc.c:164:6: warning: implicit declaration of function ‘blk_get_from_parent’; did you mean ‘efi_get_ram_base’? [-Wimplicit-function-declaration]
if (blk_get_from_parent(mmc->dev, &dev))
^~~~~~~~~~~~~~~~~~~
efi_get_ram_base
env/mmc.c:164:29: error: ‘struct mmc’ has no member named ‘dev’
if (blk_get_from_parent(mmc->dev, &dev))
^~
Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Reviewed-by: Simon Glass <sjg@chromium.org>
Modifying the default environment via CONFIG_EXTRA_ENV_SETTINGS is
somewhat inflexible, partly because the cpp language does not allow
appending to an existing macro. This prevents reuse of "environment
fragments" for different boards, which in turn makes maintaining that
environment consistently tedious and error-prone.
This implements a Kconfig option for allowing one to define the entire
default environment in an external file, which can then, for example, be
generated programmatically as part of a Yocto recipe, or simply be kept
in version control separately from the U-boot repository.
Tested-by: Sean Nyekjaer <sean.nyekjaer@prevas.dk>
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
Commit 7653942b10 ("common/env_embedded.c: drop support for
CONFIG_SYS_USE_PPCENV") dropped the .ppcenv section which was
used in linking scripts to allow fine placement of embedded
environment sections.
This implies that GCC randomly places objects from env/embedded.o
and environment is not guaranteed to be located at the correct address:
04003df8 g F .text 00000038 mii_init
04004000 g O .text 00000004 env_size
04004004 g O .text 00002000 environment
04006004 g F .text 00000040 .hidden __lshrdi3
This patch restores this capability by allocating each object marked
with __UBOOT_ENV_SECTION__ into a different section. Hence
'environment' will be alone in .text.environment, allowing a
fine placement in u-boot.lds with:
. = DEFINED(env_offset) ? env_offset : .;
env/embedded.o (.text.environment)
Fixes: 7653942b10 ("common/env_embedded.c: drop support for CONFIG_SYS_USE_PPCENV")
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Since the dawn of time for the Allwinner support in mainline U-Boot
we store the environment to the SD card and write directly at
544KB from the beginning of the device. This leads to problems when
the U-Boot proper image grows beyond 504KB and eventually overlaps.
With one release of having the environment preferably in a FAT
partition, let's now turn off the MMC variant fallback, so we get back
all the space we need to implement features.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
Introduce another difference from upstream (kernel) source in
fs/ubifs/super.c: adding preprocessor condition as y variable in
mount_ubifs() depends on CONFIG_UBIFS_SILENCE_MSG:
fs/ubifs/super.c:1337:15: error: variable ?y? set but not used [-Werror=unused-but-set-variable]
long long x, y;
Not setting CONFIG_UBIFS_SILENCE_MSG in am335x_igep003x_defconfig and
igep0032_defconfig. Although it was defined in their config headers, it
depends on CMD_UBIFS which is not set for them.
Signed-off-by: Petr Vorel <petr.vorel@gmail.com>
Cc: Joe Hershberger <joe.hershberger@ni.com>
Cc: Heiko Schocher <hs@denx.de>
When booting from a non-MMC device, the MMC sub-system may not be
initialized when the environment is first accessed.
We need to make sure that the MMC sub-system is ready in even a non-MMC
boot case.
Therefore, initialize mmc before loading environment from it.
Signed-off-by: Faiz Abbas <faiz_abbas@ti.com>
With multiple environments, the 'get_char' callback for env
drivers does not really make sense any more because it is
only supported by two drivers (eeprom and nvram).
To restore single character loading for these drivers,
override 'env_get_char_spec'.
Signed-off-by: Simon Goldschmidt <sgoldschmidt@de.pepperl-fuchs.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
Commit 7d714a24d7 ("env: Support multiple environments") added
static variable env_load_location. When saving environmental
variables, this variable is presumed to have the value set before.
In case the value was set before relocation and U-Boot runs from a
NOR flash, this variable wasn't writable. This causes failure when
saving the environment. To save this location, global data must be
used instead.
Signed-off-by: York Sun <york.sun@nxp.com>
CC: Maxime Ripard <maxime.ripard@free-electrons.com>
When saving the environment on a platform which has DMA alignment
larger than the natural alignment, env_fat_save triggers a debug
message in file_fat_write:
Saving Environment to FAT... writing uboot.env
FAT: Misaligned buffer address (9df1c8e0)
OK
Signed-off-by: Alex Kiernan <alex.kiernan@gmail.com>
For the redundant environment configuration, env_sf_load still
contained duplicate code instead of using env_import_redund().
Simplify the code by only executing the load twice and delegating
everything else to env_import_redund.
Signed-off-by: Simon Goldschmidt <sgoldschmidt@de.pepperl-fuchs.com>
Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
For multiple env drivers to correctly implement fallback when
one environment fails to load (e.g. crc error), the return value
of env_import has to be propagated by all env driver's load
function.
Without this change, the first driver that succeeds to load an
environment with an invalid CRC return 0 (success) and no other
drivers are checked.
Signed-off-by: Simon Goldschmidt <sgoldschmidt@de.pepperl-fuchs.com>
Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
There is more common code in mmc, nand and ubi env drivers that
can be shared by moving to env_import_redund.
For this, a status/error value whether the buffers were loaded
are passed as additional parameters to env_import_redund.
Ideally, these are already returned to the env driver by the
storage driver. This is the case for mmc, nand and ubi, so for
this change, code deduplicated.
Signed-off-by: Simon Goldschmidt <sgoldschmidt@de.pepperl-fuchs.com>
Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
env_import (and env_import_redund) currently return 1 on success
and 0 on error. However, they are only used from functions
returning 0 on success or a negative value on error.
Let's clean this up by making env_import and env_import_redund
return 0 on success and -EIO on error (as was the case for all
users before).
Users that cared for the return value are also updated. Funny
enough, this only affects onenand.c and sf.c
Signed-off-by: Simon Goldschmidt <sgoldschmidt@de.pepperl-fuchs.com>
Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Now that we have everything in place to implement the transition scheme,
let's enable it by default.
Reviewed-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Allow boards and architectures to override the default environment lookup
code by overriding env_get_location.
Reviewed-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Now that we have everything in place in the code, let's allow to build
multiple environments backend through Kconfig.
Reviewed-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
