According to the dm_serial_ops documentation, pending() should:
> @return number of waiting characters, 0 for none, -ve on error
And:
> It is acceptable to return 1 if an indeterminant number
> of characters is waiting.
With the current implementation, we have:
* FIFO is full -> pending() returns 0
* FIFO is partially used -> pending() returns 1
* FIFO is empty -> pending() returns 1
This is not the same as what the documentation requires.
Moreover, since [1], arm reset now flushes all console devices
(including serial) before the cpu gets reset.
Because of the flawed logic:
=> reset # user calls reset
flush() is called
_serial_flush() is called
ops->pending(dev, false) # never returns false
# board hangs indefinitely without resetting.
Fix it by using AML_UART_TX_EMPTY instead of AML_UART_TX_FULL.
[1] commit c5f4cdb8eb ("console: Use flush() before panic and reset"),
Signed-off-by: Mattijs Korpershoek <mkorpershoek@baylibre.com>
Acked-by: Neil Armstrong <neil.armstrong@linaro.org>
Link: https://lore.kernel.org/r/20230606-fix-meson-serial-pending-v1-1-6a54d4a01f76@baylibre.com
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
i2c updates for v2023-07-rc4
Bugfixes:
- rockchip: De-initialize the bus after start bit failure
from Ondrej Jirman
- cdns: Fix broken retry mechanism on arbitration lost
The ubifsload command is truncating any address above 4GiB as it casts
this address to an u32, instead of using an unsigned long which most of
the other load commands do. Change this to an unsigned long to allow
loading into high memory for boards which use these areas.
Fixes the following error:
=> ubifsload 0x2100000000 /boot/Image.lzma
Loading file '/boot/Image.lzma' to addr 0x00000000...
Unhandled exception: Store/AMO access fault
Signed-off-by: Ben Dooks <ben.dooks@sifive.com>
Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
In the current implementation, in case of I2C arbitration lost, a retry is
attempted; the message counter and pointer are reset to the original values
and the I2C xfer process is restart from the beginning.
However the message counter and message pointer are respectively
decremented and incremented by one before attempting any transfer, causing
the 1st transfer not to be actually retried (in case of a single transfer,
nothing is actually retried at all).
This patch fixes this: in case of retry, the 1st transfer is also retried.
Tested on a ZynqMP Kria board, with upstream older u-boot, but the involved
file and underlying logic seem basically the same.
Signed-off-by: Andrea Merello <andrea.merello@iit.it>
Failure can happen when i2c is used without initializing pinctrl properly,
which U-Boot happily allows in SPL. Without this fix, further I2C access would
fail, even after proper pinctrl initialization.
Signed-off-by: Ondrej Jirman <megi@xff.cz>
Cc: Heiko Schocher <hs@denx.de>
Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Reviewed-by: Heiko Schocher <hs@denx.de>
Documentation:
* man pages for loadb, loadx
UEFI:
* check return value of part_get_info()
* improve unit test for RegisterProtocolNotify()
-----BEGIN PGP SIGNATURE-----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=rUPv
-----END PGP SIGNATURE-----
Merge tag 'efi-2023-07-rc4' of https://source.denx.de/u-boot/custodians/u-boot-efi
Pull request efi-2023-07-rc4
Documentation:
* man pages for loadb, loadx
UEFI:
* check return value of part_get_info()
* improve unit test for RegisterProtocolNotify()
Check that LocateHandleBuffer() return EFI_NOT_FOUND when called with
ByRegisterNotify and all handles already have been retrieved.
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
part_get_info() may return an error code. Check it.
Update function description of dp_part_node().
Addresses-Coverity-ID: 184067 ("Unchecked return value")
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
In the first silicon revision of the am62x family of SoCs, the hardware
wakeup event cannot be used if software is unable to unlock the RTC
device within one second after boot. To work around this limitation
unlock RTC as soon as possible in the boot flow to maximize our chance
of linux being able to use this device.
Add the erratum i2327 workaround to initialize the RTC.
Signed-off-by: Nishanth Menon <nm@ti.com>
[bb@ti.com: rebased from 2021.01 and expanded commit and code messages]
Signed-off-by: Bryan Brattlof <bb@ti.com>
Add support for j721s2-wiz-10g device to use clock-names interface
instead of explicitly defining clock nodes within device tree node.
Signed-off-by: Ravi Gunasekaran <r-gunasekaran@ti.com>
ATF and OPTEE regions may be firewalled from non-secure entities.
If we still map them for non-secure A53, speculative access may happen,
which will not cause any faults and related error response will be ignored,
but it's better to not to map those regions for non-secure A53 as there
will be no actual access at all.
Create separate table as ATF region is at different locations for am64
and am62/am62a.
Signed-off-by: Kamlesh Gurudasani <kamlesh@ti.com>
Add main_uart1 clocks in clk-data.c for J7200. Now,
main_uart1 clocks will be set up while booting the J7200 SoC.
Signed-off-by: Bhavya Kapoor <b-kapoor@ti.com>
Add device data for main_uart1 in dev-data.c for J7200. Now,
main_uart1 will be powered on while booting the J7200 SoC.
Signed-off-by: Bhavya Kapoor <b-kapoor@ti.com>
When selecting UDA partition for booting. MMC read
mode was selected as RAW.
Due to growing/changing size of u-boot and tispl
images.
It will be better change to FS in case of UDA FS instead of
adjusting offsets with new change.
Signed-off-by: Udit Kumar <u-kumar1@ti.com>
This patch adds documentation for j7200.
TRM link
https://www.ti.com/lit/pdf/spruiu1
Signed-off-by: Udit Kumar <u-kumar1@ti.com>
Tested-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Add main_uart5 clocks in clk-data.c for J721S2. Now,
main_uart5 clocks will be set up while booting the J721S2 SoC.
Signed-off-by: Bhavya Kapoor <b-kapoor@ti.com>
Reviewed-by: Bryan Brattlof <bb@ti.com>
Add device data for main_uart5 in dev-data.c for J721S2. Now,
main_uart5 will be powered on while booting the J721S2 SoC.
Signed-off-by: Bhavya Kapoor <b-kapoor@ti.com>
Reviewed-by: Bryan Brattlof <bb@ti.com>
Add main_uart2 clocks in clk-data.c for J721E. Now,
main_uart2 clocks will be set up while booting the J721E SoC.
Signed-off-by: Bhavya Kapoor <b-kapoor@ti.com>
Add device data for main_uart2 in dev-data.c for J721E. Now,
main_uart2 will be powered on while booting the J721E SoC.
Signed-off-by: Bhavya Kapoor <b-kapoor@ti.com>
Compiling with gcc 13 results in an error:
drivers/axi/axi-emul-uclass.c:16:5: warning: conflicting types for
‘axi_sandbox_get_emul’ due to enum/integer mismatch; have
‘int(struct udevice *, ulong, enum axi_size_t, struct udevice **)’
{aka ‘int(struct udevice *, long unsigned int, enum axi_size_t,
struct udevice **)’} [-Wenum-int-mismatch]
16 | int axi_sandbox_get_emul(struct udevice *bus, ulong address,
| ^~~~~~~~~~~~~~~~~~~~
In file included from drivers/axi/axi-emul-uclass.c:14:
./arch/sandbox/include/asm/axi.h:48:5: note: previous declaration of
‘axi_sandbox_get_emul’ with type ‘int(struct udevice *, ulong, uint,
struct udevice **)’ {aka ‘int(struct udevice *, long unsigned int,
unsigned int, struct udevice **)’}
48 | int axi_sandbox_get_emul(struct udevice *bus, ulong address, uint length,
| ^~~~~~~~~~~~~~~~~~~~
Adjust the header definition to match the implementation.
Define the size parameter as constant.
Fixes: 9a8bcabd8a ("axi: Add AXI sandbox driver and simple emulator")
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
The "Message not acknowledged" error message is missing a line feed,
leading to the console log getting garbled and joined together with
whatever the next output is in case this error happens:
"ti_sci system-controller@44043000: Message not acknowledgedAuthentication failed!"
Fix ths by adding the missing linefeed character.
Signed-off-by: Andreas Dannenberg <dannenberg@ti.com>
Acked-by: Andrew Davis <afd@ti.com>
This is required for UART boot flow where u-boot.img needs to be
downloaded via YMODEM.
Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Reviewed-by: Bryan Brattlof <bb@ti.com>
K3 devices have some firewalls set up by ROM that we usually remove so
that the development is easy in HS devices.
While removing the firewalls disabling a background region before
disabling the foreground regions keeps the firewall in a state where all
the transactions will be blacklisted until all the regions are disabled.
This causes a race for some other entity trying to access that memory
region before all the firewalls are disabled and causes an exception.
Since there is no guarantee on where the background regions lie based on
ROM configurations or no guarantee if the background regions will allow
all transactions across the memory spaces, iterate the loop twice removing
the foregrounds first and then backgrounds.
Signed-off-by: Manorit Chawdhry <m-chawdhry@ti.com>
This reverts commit b8ebf24e7f.
This patch seems to be fundamentally wrong and requires a different way
on how the background firewalls should be configured so revert the patch
Signed-off-by: Manorit Chawdhry <m-chawdhry@ti.com>
When build Xen target with Clang, the linker reports failure.
This patch adds the related info in the documentation as a known issue
and gives details for how to dismiss the building failure with Clang.
Signed-off-by: Leo Yan <leo.yan@linaro.org>
When build U-boot with clang with using commands:
$ make HOSTCC=clang xenguest_arm64_defconfig
$ make HOSTCC=clang CROSS_COMPILE=aarch64-linux-gnu- \
CC="clang -target aarch64-linux-gnueabi" -j8
The compiler reports error:
/tmp/start-acdf31.s:330:1: error: symbol '_start' is already defined
_start:
^
Because the symbol '_start' has been defined twice, one is defined in
arch/arm/cpu/armv8/start.S, another is defined in the header
boot0-linux-kernel-header.h.
To fix building failure, this patch removes the symbol '_start' from
boot0-linux-kernel-header.h.
Signed-off-by: Leo Yan <leo.yan@linaro.org>
Buffers created through DEFINE_(CACHE_)ALIGN_BUFFER are actually
pointers to the real underlying buffer. Using sizeof(...) is
not appropriate in this case.
Signed-off-by: Sam Edwards <CFSworks@gmail.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
As per [1], dfu_alt_info is mentioned to be as semicolon separated
string of information on each alternate and the parsing logic in
the dfu.c is based on this.
Typically, the dfu_alt_info_* is defined in .h files as preprocessor
macros with 'alt' info separated by semicolon.
But when dfu_alt_info_* is added in the environment files(.env)
the script at "scripts/env2string.awk" converts a newline to space.
Thus adding a space character after semicolon. This results in
incorrect parsing in dfu.c which is based on the information that
'alt' info are only semicolon separated.
One option is to add dfu_alt_info_* variable in .env in single line.
But there is possiblity for it to exceed the line length limit.
So update the parsing logic to remove leading space characters
before adding to the dfu list.
[1]: https://u-boot.readthedocs.io/en/latest/usage/dfu.html
Signed-off-by: Ravi Gunasekaran <r-gunasekaran@ti.com>
For sf commands, when '0' length is passed for erase, update, write or
read, there might be undesired results. Ideally '0' length means nothing to
do.
So print 'ERROR: Invalid size 0' and return cmd failure when length '0' is
passed to sf commands. Same thing applies for nand commands also.
Example:
ZynqMP> sf erase 0 0
ERROR: Invalid size 0
ZynqMP> sf write 10000 0 0
ERROR: Invalid size 0
ZynqMP> sf read 10000 0 0
ERROR: Invalid size 0
ZynqMP> sf update 1000 10000 0
ERROR: Invalid size 0
ZynqMP>
Signed-off-by: Ashok Reddy Soma <ashok.reddy.soma@amd.com>
Currently, all in-tree .dts files (apart from some under test/ and
tools/), reside in arch/$ARCH/dts. However, in the linux kernel tree,
dts files for arm64 boards, and probably in the not too distant
future [1], arm boards as well, live in subdirectories of that.
For private forks, using a vendor or project subdirectory is also more
convenient to clearly separate private code from upstream - in the
same way that code under board/ is also split and easy to maintain.
In order to prepare for us to follow suit and do the splitting of the
in-tree .dts files, and to make life a little easier for private forks
that already place dts files not directly in arch/$ARCH/dts, change
the $(srctree)/arch/$(ARCH)/dts path to instead refer to the directory of
the .dts file being compiled. This should be a no-op for all existing
cases.
[1] https://lore.kernel.org/lkml/20220328000915.15041-1-ansuelsmth@gmail.com/
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
If a file does not exist, it should be created.
Fixes: f676b45151 ("fs: Add semihosting filesystem")
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
ARM requires a 4-byte alignment on all ARM code (though this
requirement is relaxed to 2-byte for some THUMB code) and we
should be explicit about that here.
GAS has its own fix for this[1] that forces proper alignment
on any section containing assembled instructions, but this is
not universal: Clang's and other gaslike assemblers lack this
implicit alignment. Whether or not this is considered a bug in
those assemblers, it is better to ask directly for what we want.
[1]: https://sourceware.org/bugzilla/show_bug.cgi?id=12931
Signed-off-by: Sam Edwards <CFSworks@gmail.com>
Make it a little bit easier for the user to utilize the 'size' command.
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Reviewed-by: Simon Glass <sjg@chromium.org>
Invoking the sandbox with
/u-boot -c ⧵0xef⧵0xbf⧵0xbd
results in a segmentation fault.
Function b_getch() retrieves a character from the input stream. This
character may be > 0x7f. If type char is signed, static_get() will
return a negative number and in parse_stream() we will use that
negative number as an index for array map[] resulting in a buffer
overflow.
Reported-by: Harry Lockyer <harry_lockyer@tutanota.com>
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Per the fit_conf_get_node() API doc, it returns configuration node
offset when found (>=0).
Signed-off-by: Bin Meng <bmeng@tinylab.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
These files should have both 'always' and 'targets' so that dependencies
are detected correctly.
When only 'always' is used, the target is built every time, although I am
not quite sure why.
Make sure each has both 'always' and 'targets' to avoid this problem.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present two acpi files are built every time since they use a version
number from version.h
This is not necessary. Make use of the same technique as for the version
string, so that they are build only when they change.
Signed-off-by: Simon Glass <sjg@chromium.org>
These rules run on every build even if nothing has changed. The FORCE
dependency is only needed for if_changed, not for cmd. Drop it.
Signed-off-by: Simon Glass <sjg@chromium.org>
This fixes an error with trying to link against do_bootm() when
CONFIG_CMD_BOOTM is disabled.
Signed-off-by: Samuel Holland <samuel@sholland.org>
Reviewed-by: Mattijs Korpershoek <mkorpershoek@baylibre.com>
Reviewed-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
