On some STM32 SoC's package, GPIO bank may have hole in their GPIO bank
Example:
If GPIO bank have 16 GPIO pins [0-15].
In particular SoC's package case, some GPIO bank can have less GPIO pins:
- [0-10] => 11 pins;
- [2-7] => 6 pins.
Commit 41226e0dd954 ("gpio: stm32f7: Add gpio bank holes management")
proposed a first implementation by not counting GPIO "inside" hole. GPIO
are not displaying correctly using gpio or pinmux command when GPIO holes
are located at the beginning of GPIO bank.
To simplify, consider that all GPIO have 16 GPIO and use the gpio_ranges
struct to indicate if a GPIO is mapped or not. GPIO uclass offers several
GPIO functions ("input", "output", "unused", "unknown" and "func"), use
"unknown" GPIO function to indicate that a GPIO is not mapped.
stm32_offset_to_index() is no more needed and removed.
This must be reflected using the "gpio" command to indicate to user
that a particular GPIO is not mapped (marked as "unknown") as shown below:
Example for a 16 pins GPIO bank with the [2-7] mapping (only 6 pins
mapped):
GPIOI0 : unknown
GPIOI1 : unknown
GPIOI2 : analog
GPIOI3 : analog
GPIOI4 : alt function 0 push-pull pull-down
GPIOI5 : alt function 0 push-pull pull-down
GPIOI6 : alt function 0 push-pull pull-down
GPIOI7 : analog
GPIOI8 : unknown
GPIOI9 : unknown
GPIOI10 : unknown
GPIOI11 : unknown
GPIOI12 : unknown
GPIOI13 : unknown
GPIOI14 : unknown
GPIOI15 : unknown
Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com> Reviewed-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
ARM: stm32: Use CONFIG_TFTP_TSIZE on STMicroelectronics boards
Long TFTP transfers lead to a wall of # characters on UART, which in
the end may slow down the transfer itself. Use CONFIG_TFTP_TSIZE to
print progress in fewer # characters.
The Buck3 on DHCOR is used to supply IO voltage. It can output either 3V3
in the default DHCOR configuration, or 2V9 in case of AV96 DHCOR variant
which has extra Empirion DCDC converter in front of the 1V8 IO supply, or
outright 1V8 in case of 1V8 IO DHCOR without the Empirion DCDC converter.
The 2V9 mode in case of AV96 DHCOR variant is used to reduce unnecessarily
high input voltage to the Empirion DCDC converter, so move it into matching
DTSI to stop confusing users.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Patrice Chotard <patrice.chotard@foss.st.com> Cc: Patrick Delaunay <patrick.delaunay@foss.st.com> Reviewed-by: Patrice Chotard <patrice.chotard@foss.st.com> Reviewed-by: Patrick Delaunay <patrick.delaunay@foss.st.com> Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com>
Marek Vasut [Sun, 1 May 2022 16:43:29 +0000 (18:43 +0200)]
ARM: stm32: Use CONFIG_TFTP_TSIZE on DHSOM
Long TFTP transfers lead to a wall of # characters on UART, which in
the end may slow down the transfer itself. Use CONFIG_TFTP_TSIZE to
print progress in fewer # characters.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Patrick Delaunay <patrick.delaunay@foss.st.com> Cc: Patrice Chotard <patrice.chotard@foss.st.com> Cc: Ramon Fried <rfried.dev@gmail.com> Reviewed-by: Ramon Fried <rfried.dev@gmail.com> Reviewed-by: Patrick Delaunay <patrick.delaunay@foss.st.com> Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com>
Marek Vasut [Sun, 1 May 2022 16:43:28 +0000 (18:43 +0200)]
ARM: stm32: Use default CONFIG_TFTP_BLOCKSIZE on DHSOM
The DHCOM does ship with KS8851 with 1.5 kiB packet buffer. The DHSOM
may be extended with other MAC options connected to FMC2 bus, like the
DM9000, wih similar limitations. Use default CONFIG_TFTP_BLOCKSIZE of
1468 Bytes instead of 1536 Bytes, which always avoids overflowing the
packet buffers of such limited MACs, which leads to e.g. TFTP timeouts.
This also avoids receiving a short packet fragment at the end of each
TFTP block, which led to reduced performance.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Patrick Delaunay <patrick.delaunay@foss.st.com> Cc: Patrice Chotard <patrice.chotard@foss.st.com> Cc: Ramon Fried <rfried.dev@gmail.com> Reviewed-by: Ramon Fried <rfried.dev@gmail.com> Reviewed-by: Patrick Delaunay <patrick.delaunay@foss.st.com> Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com>
Marek Vasut [Tue, 26 Apr 2022 14:38:05 +0000 (16:38 +0200)]
stm32mp: psci: Retain MCUDIVR, PLL3CR, PLL4CR, MSSCKSELR across suspend
The SoC seems to lose the values of MCUDIVR, PLL3CR, PLL4CR, RCC_MSSCKSELR
during suspend/resume cycle, cache them and reinstate their values on resume.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Patrick Delaunay <patrick.delaunay@foss.st.com> Cc: Patrice Chotard <patrice.chotard@foss.st.com> Reviewed-by: Patrick Delaunay <patrick.delaunay@foss.st.com> Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com>
Marek Vasut [Tue, 26 Apr 2022 14:37:41 +0000 (16:37 +0200)]
ram: stm32mp1: Conditionally enable ASR
Enable DRAM ASR, auto self-refresh, conditionally, based on DT PWRCTL
register bits. While ASR does save considerable amount of power at
runtime automatically, it also causes LTDC underruns on large panels.
Let user select whether or not ASR is required or not, generally ASR
should be enabled on portable and battery operated devices.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Patrick Delaunay <patrick.delaunay@foss.st.com> Cc: Patrice Chotard <patrice.chotard@foss.st.com> Reviewed-by: Patrick Delaunay <patrick.delaunay@foss.st.com> Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com>
Patrick Delaunay [Tue, 26 Apr 2022 13:38:05 +0000 (15:38 +0200)]
arm: dts: stm32mp15: alignment with v5.18
Device tree alignment with Linux kernel v5.18-rc2:
- ARM: dts: stm32: Add support for the emtrion emSBC-Argon
(only the pincontrol part)
- ARM: dts: stm32: Drop duplicate status okay from DHCOM gpioc node
- ARM: dts: stm32: add st,stm32-sdmmc2 compatible on stm32mp151
- ARM: dts: stm32: fix AV96 board SAI2 pin muxing on stm32mp15
- ARM: dts: stm32: use exti 19 as main interrupt to support RTC wakeup on
stm32mp157
- ARM: dts: stm32: add DMA configuration to UART nodes on stm32mp151
- ARM: dts: stm32: keep uart4 behavior on *
- ARM: dts: stm32: Correct masks for GIC PPI interrupts on stm32mp15
Patrick Delaunay [Mon, 28 Mar 2022 17:25:32 +0000 (19:25 +0200)]
stm32mp: stm32prog: handle U-Boot script in flashlayout alternate
Update the stm32prog command to allow the reception of U-Boot script in
the FlashLayout alternate during the first USB enumeration.
This patch is aligned with the last TF-A behavior: the Flashlayout
is now loaded by U-Boot; it is no more present at STM32_DDR_BASE when
the stm32prog is launched after a serial boot, on UART or on USB.
The received script must be a U-Boot legacy image, no more need to add
a stm32image header.
Patrick Delaunay [Mon, 28 Mar 2022 17:25:31 +0000 (19:25 +0200)]
stm32mp: stm32prog: handle interruption during the first enumeration
When an interruption is received during the first USB enumeration
used to received the FlashLayout, with handle ctrl-c, the second
enumeration is not needed and the result for stm32prog_usb_loop
is false (reset is not needed).
This patch avoids the need of a second ctrl to interrupt the command
stm32prog.
Patrick Delaunay [Mon, 28 Mar 2022 17:25:30 +0000 (19:25 +0200)]
stm32mp: stm32prog: add support of UUID for FIP partition
Add support of UUID for FIP parttion, required by Firmware update
support in TF-A:
- UUID TYPE for FIP partition: 19d5df83-11b0-457b-be2c-7559c13142a5
- "fip-a" partition UUID: 4fd84c93-54ef-463f-a7ef-ae25ff887087
- "fip-b" partition UUID: 09c54952-d5bf-45af-acee-335303766fb3
This check is done with a new partition type "FIP" associated
at the FIP UUID.
The A/B partition UUID is detected by the partition name:
"fip-a", "fip-b".
Patrick Delaunay [Mon, 28 Mar 2022 17:25:27 +0000 (19:25 +0200)]
stm32mp: stm32prog: add CONFIG_CMD_STM32PROG_OTP
Add a configuration flag CONFIG_CMD_STM32PROG_OTP to enable the support of
OTP update in stm32prog command.
This new configuration allows to deactivate this feature for security reason
and it is a preliminary step for support of OPT update with the OP-TEE
provisioning TA.
Tom Rini [Sun, 8 May 2022 15:31:48 +0000 (11:31 -0400)]
Merge tag 'efi-2022-07-rc3-2' of https://source.denx.de/u-boot/custodians/u-boot-efi
Pull request for efi-2022-07-rc3-2
UEFI:
* Fix build errors due to
- using sed with non-standard extension for regular expression
- target architecture not recognized for CROSS_COMPILE=armv7a-*
- CONFIG_EVENT not selected
* add sha384/512 on certificate revocation
Others:
* factor out the user input handling in bootmenu command
efi_loader: add sha384/512 on certificate revocation
Currently we don't support sha384/512 for the X.509 certificate
in dbx. Moreover if we come across such a hash we skip the check
and approve the image, although the image might needs to be rejected.
Rework the code a bit and fix it by adding an array of structs with the
supported GUIDs, len and literal used in the U-Boot crypto APIs instead
of hardcoding the GUID types.
It's worth noting here that efi_hash_regions() can now be reused from
efi_signature_lookup_digest() and add sha348/512 support there as well
Mark Kettenis [Thu, 5 May 2022 14:42:23 +0000 (16:42 +0200)]
Makefile: Avoid non-portable GNU sed extension
Use [:space:] instead of \s and \S in regular expression that
determines the sandbox target architecture. Fixes the build
failure on OpenBSD introduced with commit 7b3569c68a46
("efi_loader: bootmgr: add booting from removable media").
Fixes: a990f3deb285 ("sandbox: allow cross-compiling sandbox") Signed-off-by: Mark Kettenis <kettenis@openbsd.org> Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
John Keeping [Wed, 20 Apr 2022 10:31:11 +0000 (11:31 +0100)]
fdt: don't set linux,phandle
This has been deprecated for over 10 years and everything now uses the
plain "phandle" property in preference. There's no need to set
linux,phandle when creating phandles for nodes that do not have one.
dtc changed the default to creating just phandle in version 1.4.5
released in September 2017 with the justification that the new style had
already been supported for 7 years by that point (see dtc commit 0016f8c
("dtc: change default phandles to ePAPR style instead of both")).
Philippe Reynes [Fri, 22 Apr 2022 15:46:49 +0000 (17:46 +0200)]
cmd: gpt: add subcommand repair
Adds a sub-command repair to the command gpt
that allow to repair a corrupted gpt table. If
the both gpt table (primary and backup) are
valid, then the command does nothing.
Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
Philippe Reynes [Fri, 22 Apr 2022 15:46:48 +0000 (17:46 +0200)]
disk: part_efi: add support to repair gpt table
If a gpt table is corrupted (after a power cut
for example), then the gpt table should repaired.
The function gpt_repair_headers check if at least
one gpt table is valid, and then only write the
corrupted gpt table.
Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
Patrice Chotard [Fri, 6 May 2022 06:22:35 +0000 (08:22 +0200)]
usb: host: ehci-generic: Remove DM_REGULATOR flag
Since commit 8b4a43deb9b0 ("power: regulator: add dummy helper")
regulator dummy helper are always available even if DM_REGULATOR
is not set.
DM_REGULATOR flag is no more needed to protect no DM core,
remove it.
- Assorted minor code cleanups.
- Clean-up the reset uclass code slightly and fix some issues with a
lack of handlers for a case in the driver.
- Y2038 RTC fix
misc: fs_loader: Fix compile warnings when CONFIG_CMD_UBIFS is enabled
drivers/misc/fs_loader.c: In function ‘mount_ubifs’:
drivers/misc/fs_loader.c:46:12: warning: implicit declaration of function ‘ubi_part’ [-Wimplicit-function-declaration]
int ret = ubi_part(mtdpart, NULL);
^~~~~~~~
drivers/misc/fs_loader.c:53:9: warning: implicit declaration of function ‘cmd_ubifs_mount’ [-Wimplicit-function-declaration]
return cmd_ubifs_mount(ubivol);
^~~~~~~~~~~~~~~
drivers/misc/fs_loader.c: In function ‘umount_ubifs’:
drivers/misc/fs_loader.c:58:9: warning: implicit declaration of function ‘cmd_ubifs_umount’ [-Wimplicit-function-declaration]
return cmd_ubifs_umount();
^~~~~~~~~~~~~~~~
Marek Vasut [Tue, 26 Apr 2022 21:43:30 +0000 (23:43 +0200)]
reset: Return 0 if ops unimplemented and remove empty functions
In case the ops is not implemented, return 0 in the core right away.
This is better than having multiple copies of functions which just
return 0 in each reset driver. Drop all those empty functions.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Simon Glass <sjg@chromium.org> Cc: Tom Rini <trini@konsulko.com>
Local variable out.name lives on the stack and therefore cannot
be returned directly. Move the strdup() call into the function.
(Coverity 352460)
Fixes: 8f136b636a9f ("clk: scmi: register scmi clocks with CCF") Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com> Reviewed-by: Sean Anderson <seanga2@gmail.com>
Oliver Graute [Tue, 26 Apr 2022 07:26:12 +0000 (09:26 +0200)]
rtc: rv8803: fix off-by-one in month counting
tm_mon has a range from 0..11, but the RTC expects 1..12. So we adapt
the month accordingly. This was determined when comparing the driver
with the corresponding linux kernel driver.
Signed-off-by: Oliver Graute <oliver.graute@kococonnector.com> Reviewed-by: Michael Walle <michael@walle.cc> Reviewed-by: Heiko Schocher <hs@denx.de>
Fixes: 7eb53beeda16 ("cmd: adc: Add support for storing ADC result in env variable") Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
include: configs: am**x/j721e/j721s2_evm.h: Move the stack pointer init address in arm64
Currently, in case of arm64 bootloader and U-Boot the stack pointer is
initialized at an offset of NON_SECURE_MSRAM_SIZE from arm64 SPL's text
base address. After jumping to arm64, execution is done out of DDR.
Therefore, having an offset corresponding to the size of MSRAM does not
have any significance.
Instead, initialize the stack pointer after an offset of 4MB from the SPL
text base address. This helps in allocating larger memory for stack.
Sam Shih [Thu, 21 Apr 2022 06:23:53 +0000 (14:23 +0800)]
pinctrl: mediatek: add support for different types of IO pins
There are many pins in an SoC, and register usage may vary by pins.
This patch introduces a concept of "io type" and "io type group"
to mediatek pinctrl drivers. This can provide different pinconf
handlers implementation (eg: "bias-pull-up/down", "driving" and
"input-enable") for IO pins that belong to different types.
Sam Shih [Thu, 21 Apr 2022 06:23:52 +0000 (14:23 +0800)]
pinctrl: mediatek: introduce multiple memory bases support
Pinctrl design of some mediatek SoC need to access registers that
distribute in multiple memory base address. this patch introduce new
mechanism in mediatek pinctrl driver to support the chips which have
the new design.
This patch add a member 'base_calc' in pinctrl private data, and changed
original 'base' private data to an array of *iomem.
When 'base_calc' attribute is set, it will requests multiplue regs base
from the DT, if 'base_calc' attribute is not set, it only use legacy way
to request single reg resource from the DT.
Sam Shih [Thu, 21 Apr 2022 06:23:51 +0000 (14:23 +0800)]
pinctrl: mediatek: rewrite mtk_pinconf_set and related functions
There are many pins in a SoCs, and different pin may belong
to different "io_type", For example: some pins of MT7622 belongs
to "io_type A", the other belongs to "io_type B", and pinctrl "V0"
means handle pinconf via "io_type A" or "io_type B", so SoCs that
contain "io_type A" and "io_type B" pins, use "V0" in pinctrl driver.
This patch separates the implementation of register operations
(e.g: "bias-pull-up/down", "driving" and "input-enable") into
different functions, and lets the original V0/V1
ops to call the new functions.
Simon Glass [Wed, 27 Apr 2022 19:47:57 +0000 (13:47 -0600)]
sandbox: Avoid binman error when run without device tree
With sandbox, U-Boot can be run without a device tree (i.e. no -d or -T
parameter). In this case an empty device tree is created for convenience.
With a recent change this causes an error due to the missing '/binman'
node.
Add this node to avoid the problem, as well as a test that U-Boot can
be run without a device tree.
Fixes: f117b6b175 ("arch: Kconfig: imply BINMAN for SANDBOX") Fixes: https://source.denx.de/u-boot/u-boot/-/issues/11 Signed-off-by: Simon Glass <sjg@chromium.org>
Sean Anderson [Thu, 28 Apr 2022 16:39:49 +0000 (12:39 -0400)]
ls10xxx: Use a sane SPL_SYS_MALLOC_F_LEN default
SPL_SYS_MALLOC_F_LEN defaults to SYS_MALLOC_F_LEN. 0x10000 is 64 KiB, or
around half of the total OCRAM size. Revert to the default of 0x2000. This
fixes SPL boot.
Fixes: 4f76d38c02 ("imx8/ls10xx: Use a sane SYS_MALLOC_F_LEN default") Signed-off-by: Sean Anderson <sean.anderson@seco.com> Reviewed-by: Fabio Estevam <festevam@denx.de>
Marek Behún [Wed, 27 Apr 2022 10:42:02 +0000 (12:42 +0200)]
net: mvneta: Drop unneeded macro
Macro MVNETA_GMAC_FORCE_LINK_UP can be dropped from value assignment in
fixed link case, since it's value is written into the register later in
the function for link-down-to-link-up case. The value is written as
MVNETA_GMAC_FORCE_LINK_DOWN | MVNETA_GMAC_FORCE_LINK_PASS, and so the
macro definition can also be dropped.
Signed-off-by: Marek Behún <marek.behun@nic.cz> Reviewed-by: Ramon Fried <rfried.dev@gmail.com> Reviewed-by: Stefan Roese <sr@denx.de>
Marek Behún [Wed, 27 Apr 2022 10:41:48 +0000 (12:41 +0200)]
arm: mvebu: Espressobin: Use DM registered MDIO to configure switch
In order to be able to get rid of the non-DM MDIO bus registered in
mvneta driver, we need to stop using board_network_enable() and instead
use the DM registered MDIO device to configure switch in
last_stage_init().
Signed-off-by: Marek Behún <marek.behun@nic.cz> Reviewed-by: Stefan Roese <sr@denx.de>
Stefan Roese [Thu, 7 Apr 2022 07:11:54 +0000 (09:11 +0200)]
mips: octeon: nic23: Enable ethernet support
This patch enables the Kconfig symbols needed for full ethernet support
on the NIC23. Additionally board specific setup is done, mostly GPIOs
related to SFP / GPIO configuration. With this, ethernet can be used on
this board. Here an example of a tftp load:
=> tftp ffffffff81000000 big
Using ethernet-mac-nexus@11800e2000000 device
TFTP from server 192.168.1.5; our IP address is 192.168.1.247
Filename 'big'.
Load address: 0xffffffff81000000
Loading: ################################################## 10 MiB
9.7 MiB/s
done
Bytes transferred = 10485760 (a00000 hex)
This patch enables the Kconfig symbols needed for full ethernet support
on the EBB7304. Also the PHY autonegotiation timeout is increased, as
the default 5 seconds are sometime a bit short. With this, ethernet can
be used on this board. Here an example of a tftp load:
=> tftp ffffffff81000000 big
ethernet-mac-nexus@11800e0000000 Waiting for PHY auto negotiation to complete....... done
Using ethernet-mac-nexus@11800e0000000 device
TFTP from server 192.168.1.5; our IP address is 192.168.1.243
Filename 'big'.
Load address: 0xffffffff81000000
Loading: ################################################## 10 MiB
13.2 MiB/s
done
Bytes transferred = 10485760 (a00000 hex)
Stefan Roese [Thu, 7 Apr 2022 07:11:52 +0000 (09:11 +0200)]
net: Add ethernet support for MIPS Octeon
This patchs adds the ethernet & MDIO driver for the MIPS Octeon II / III
SoC platform. Please note that these drivers are based on the 2013
U-Boot version from Marvell and make use of the platform supported
helper functions for the ethernet functionality, including stuff like
SFP handling.
Signed-off-by: Stefan Roese <sr@denx.de> Cc: Ramon Fried <rfried.dev@gmail.com> Cc: Joe Hershberger <joe.hershberger@ni.com>
projects / uboot.git / log