oweals/u-boot.git
10 months agoMerge tag 'efi-2020-07-rc6' of https://gitlab.denx.de/u-boot/custodians/u-boot-efi master
Tom Rini [Thu, 25 Jun 2020 17:33:15 +0000 (13:33 -0400)]
Merge tag 'efi-2020-07-rc6' of https://gitlab.denx.de/u-boot/custodians/u-boot-efi

Pull request for UEFI sub-system for efi-2020-07-rc6

Corrections for variable definitions are provided:

* Correct size of secure boot related UEFI variables.
* Do not use int for storing an enum.
* Replace fdt_addr by fdt_size where needed.

10 months agoARM: dts: imx6q-tbs2910: Fix Ethernet regression
Fabio Estevam [Thu, 25 Jun 2020 11:14:57 +0000 (08:14 -0300)]
ARM: dts: imx6q-tbs2910: Fix Ethernet regression

Since commit:

commit 6333cbb3817ed551cd7d4e92f7359c73ccc567fc
Author: Michael Walle <michael@walle.cc>
Date:   Thu May 7 00:11:58 2020 +0200

    phy: atheros: ar8035: remove static clock config

    We can configure the clock output in the device tree. Disable the
    hardcoded one in here. This is highly board-specific and should have
    never been enabled in the PHY driver.

    If bisecting shows that this commit breaks your board it probably
    depends on the clock output of your Atheros AR8035 PHY. Please have a
    look at doc/device-tree-bindings/net/phy/atheros.txt. You need to set
    "clk-out-frequency = <125000000>" because that value was the hardcoded
    value until this commit.

Signed-off-by: Michael Walle <michael@walle.cc>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
, the clock output setting for the AR803x driver is removed from being
hardcoded in the PHY driver and should be passed via device tree instead.

Update the device tree with the "qca,clk-out-frequency" property so that
Ethernet can work again.

Reported-by: Soeren Moch <smoch@web.de>
Signed-off-by: Fabio Estevam <festevam@gmail.com>
Tested-by: Soeren Moch <smoch@web.de>
10 months agoefi_loader: size of secure boot variables
Heinrich Schuchardt [Wed, 24 Jun 2020 10:38:00 +0000 (12:38 +0200)]
efi_loader: size of secure boot variables

The variables SetupMode, AuditMode, DeployedMode are explicitly defined as
UINT8 in the UEFI specification. The type of SecureBoot is UINT8 in EDK2.

Use variable name secure_boot instead of sec_boot for the value of the
UEFI variable SecureBoot.

Avoid abbreviations in function descriptions.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
10 months agoefi_loader: type of efi_secure_mode
Heinrich Schuchardt [Wed, 24 Jun 2020 10:14:49 +0000 (12:14 +0200)]
efi_loader: type of efi_secure_mode

Variable efi_secure_mode is meant to hold a value of enum efi_secure_mode.
So it should not be defined as int but as enum efi_secure_mode.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
10 months agocmd: bootefi: Fix fdt_size variable type in efi_carve_out_dt_rsv()
Bin Meng [Tue, 23 Jun 2020 06:50:50 +0000 (23:50 -0700)]
cmd: bootefi: Fix fdt_size variable type in efi_carve_out_dt_rsv()

Variable fdt_size should be of type 'fdt_size_t', not 'fdt_addr_t'.

Fixes 0d7c2913fdf7: ("cmd: bootefi: Honor the address & size cells properties correctly")
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
10 months agoMerge tag 'mmc-2020-6-24' of https://gitlab.denx.de/u-boot/custodians/u-boot-mmc
Tom Rini [Wed, 24 Jun 2020 13:05:35 +0000 (09:05 -0400)]
Merge tag 'mmc-2020-6-24' of https://gitlab.denx.de/u-boot/custodians/u-boot-mmc

- Fix fsl_esdhc_imx tunning mask
- Disable CMD CRC for normal tuning for fsl_esdhc_imx
- Retry CM1 until emmc ready
- Fix sdhci HISPD handling
- Cache-aligned extcsd reading

10 months agommc: sdhci: Fix HISPD bit handling
Jagan Teki [Thu, 18 Jun 2020 14:03:12 +0000 (19:33 +0530)]
mmc: sdhci: Fix HISPD bit handling

SDHCI HISPD bits need to be configured based on desired mmc
timings mode and some HISPD quirks.

So, handle the HISPD bit based on the mmc computed selected
mode(timing parameter) rather than fixed mmc card clock
frequency.

Linux handle the HISPD similar like this in below commit but no
SDHCI_QUIRK_BROKEN_HISPD_MODE,

commit <501639bf2173> ("mmc: sdhci: fix SDHCI_QUIRK_NO_HISPD_BIT handling")

This eventually fixed the mmc write issue observed in
rk3399 sdhci controller.

Bug log for refernece,
=> gpt write mmc 0 $partitions
Writing GPT: mmc write failed
** Can't write to device 0 **
** Can't write to device 0 **
error!

Cc: Kever Yang <kever.yang@rock-chips.com>
Cc: Peng Fan <peng.fan@nxp.com>
Peng Fan: added back "ctrl &= ~SDHCI_CTRL_HISPD;" per Jaehoon's suggestion
Tested-by: Suniel Mahesh <sunil@amarulasolutions.com> # roc-rk3399-pc
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
10 months agommc: retry CMD1 in mmc_send_op_cond() until the eMMC is ready
Haibo Chen [Mon, 15 Jun 2020 09:18:12 +0000 (17:18 +0800)]
mmc: retry CMD1 in mmc_send_op_cond() until the eMMC is ready

According to eMMC specification v5.1 section 6.4.3, we should issue
CMD1 repeatedly in the idle state until the eMMC is ready even if
mmc_send_op_cond() send CMD1 with argument = 0. Otherwise some eMMC
devices seems to enter the inactive mode after mmc_complete_op_cond()
issued CMD0 when the eMMC device is busy.

Signed-off-by: Haibo Chen <haibo.chen@nxp.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
10 months agommc: fsl_esdhc_imx: disable the CMD CRC check for standard tuning
Haibo Chen [Mon, 22 Jun 2020 11:38:04 +0000 (19:38 +0800)]
mmc: fsl_esdhc_imx: disable the CMD CRC check for standard tuning

In current code, we add 1ms dealy after each tuning command for standard
tuning method. Adding this 1ms dealy is because USDHC default check the
CMD CRC and DATA line. If detect the CMD CRC, USDHC standard tuning
IC logic do not wait for the tuning data sending out by the card, trigger
the buffer read ready interrupt immediately, and step to next cycle. So
when next time the new tuning command send out by USDHC, card may still
not send out the tuning data of the upper command,then some eMMC cards
may stuck, can't response to any command, block the whole tuning procedure.

If do not check the CMD CRC for tuning, then do not has this issue. USDHC
will wait for the tuning data of each tuning command and check them. If the
tuning data pass the check, it also means the CMD line also okay for tuning.

So this patch disable the CMD CRC check for tuning, save some time for the
whole tuning procedure.

Signed-off-by: Haibo Chen <haibo.chen@nxp.com>
10 months agommc: fsl_esdhc_imx: fix the mask for tuning start point
Haibo Chen [Mon, 22 Jun 2020 11:38:03 +0000 (19:38 +0800)]
mmc: fsl_esdhc_imx: fix the mask for tuning start point

According the RM, the bit[6~0] of register ESDHC_TUNING_CTRL is
TUNING_START_TAP, bit[7] of this register is to disable the command
CRC check for standard tuning. So fix it here.

Fixes: fa33d207494c ("mmc: split fsl_esdhc driver for i.MX")
Signed-off-by: Haibo Chen <haibo.chen@nxp.com>
10 months agocmd: mmc: Cache-align extcsd read target
Marek Vasut [Sat, 20 Jun 2020 12:28:25 +0000 (14:28 +0200)]
cmd: mmc: Cache-align extcsd read target

The extcsd read target must be cache aligned in case the controller
uses DMA to read the extcsd register, make it so.

Signed-off-by: Marek Vasut <marex@denx.de>
Reviewed-by: Michael Trimarchi <michael@amarulasolutions.com>
10 months agoMerge branch '2020-06-23-misc-minor-fixes'
Tom Rini [Tue, 23 Jun 2020 18:43:47 +0000 (14:43 -0400)]
Merge branch '2020-06-23-misc-minor-fixes'

- Assorted minor fixes

10 months agopsci: add 'static inline' to invoke_psci_fn() stub
Masahiro Yamada [Wed, 20 May 2020 00:32:39 +0000 (09:32 +0900)]
psci: add 'static inline' to invoke_psci_fn() stub

Avoid potential multiple definitions when CONFIG_ARM_PSCI_FW
is disabled.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
10 months agodoc: sifive: Fix spelling of "environment".
Vagrant Cascadian [Thu, 4 Jun 2020 17:42:02 +0000 (10:42 -0700)]
doc: sifive: Fix spelling of "environment".

Signed-off-by: Vagrant Cascadian <vagrant@debian.org>
10 months agocmd: booti: Fix spelling of "environment".
Vagrant Cascadian [Thu, 4 Jun 2020 17:42:01 +0000 (10:42 -0700)]
cmd: booti: Fix spelling of "environment".

Signed-off-by: Vagrant Cascadian <vagrant@debian.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
10 months agonvme: Invalidate dcache before submitting admin cmd
Jagan Teki [Wed, 10 Jun 2020 19:18:23 +0000 (21:18 +0200)]
nvme: Invalidate dcache before submitting admin cmd

This patch try to avoids eviction of dirty lines during DMA
transfer. The code right now execute the following step:

- allocate the buffer
- start a dma operation using the non-coherent dma buffer
- invalidate cache lines associated with the buffer
- read the buffer

This can lead to reading back not valid information, because the cache
controller could evict dirty cache lines belonging to the buffer *after*
the DMA operation has started to fill the DRAM.
In order to avoid this, a new invalidation is required *before* starting
the DMA operation. The patch just adds an invalidation before submitting
the DMA command.

Example below shows the nvme disk scan result without the following
patch

=> nvme scan
nvme_get_info_from_identify: nn = 544502629, vwc = 100,
sn = dev_0T, mn = `�\�, fr = t_part, mdts = 105

So, invalidating the cache before submitting the admin command,
fix the cpu read.

Cc: André Przywara <andre.przywara@arm.com>
Reported-by: Suniel Mahesh <sunil@amarulasolutions.com>
Signed-off-by: Michael Trimarchi <michael@amarulasolutions.com>
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Tested-by: Suniel Mahesh <sunil@amarulasolutions.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
10 months agophy: atheros: Fix the "qca,clk-out-frequency" example
Fabio Estevam [Wed, 17 Jun 2020 18:07:04 +0000 (15:07 -0300)]
phy: atheros: Fix the "qca,clk-out-frequency" example

The correct name for the property is "qca,clk-out-frequency", so fix
it accordingly.

Signed-off-by: Fabio Estevam <festevam@gmail.com>
Reviewed-by: Michael Walle <michael@walle.cc>
10 months agozfs: fix missing include for disk_partition definition
Joel Johnson [Sat, 20 Jun 2020 04:45:47 +0000 (22:45 -0600)]
zfs: fix missing include for disk_partition definition

Commit 0528979fa7ab ("part: Drop disk_partition_t typedef") changed to
a struct. As a result it uncovered an apparent missing include in
zfs_common.h for part.h which actually contains the definition. The ZFS
handles the struct exclusively as pointers so it was only a warning.

    warning: ‘struct disk_partition’ declared inside parameter list
    will not be visible outside of this definition or declaration
 void zfs_set_blk_dev(struct blk_desc *rbdd, struct disk_partition *info);

Signed-off-by: Joel Johnson <mrjoel@lixil.net>
Series-CC: Simon Glass <sjg@chromium.org>

10 months agoMerge tag 'u-boot-imx-20200623' of https://gitlab.denx.de/u-boot/custodians/u-boot-imx
Tom Rini [Tue, 23 Jun 2020 12:20:55 +0000 (08:20 -0400)]
Merge tag 'u-boot-imx-20200623' of https://gitlab.denx.de/u-boot/custodians/u-boot-imx

Fixes for 2020.07
-----------------

Travis: https://travis-ci.org/github/sbabic/u-boot-imx/builds/701059103

- Fixes for atheros and cubox
- Toradex: mostly environment
- i.MX7: DDR fixes
- switch to DM
- sabrelite : fix MMC access

10 months agoconfigs: Resync with savedefconfig
Tom Rini [Tue, 23 Jun 2020 12:20:07 +0000 (08:20 -0400)]
configs: Resync with savedefconfig

Rsync all defconfig files using moveconfig.py

Signed-off-by: Tom Rini <trini@konsulko.com>
10 months agoPrepare v2020.07-rc5 v2020.07-rc5
Tom Rini [Tue, 23 Jun 2020 00:49:04 +0000 (20:49 -0400)]
Prepare v2020.07-rc5

Signed-off-by: Tom Rini <trini@konsulko.com>
10 months agomx6cuboxi: remove unused code
Walter Lozano [Tue, 19 May 2020 18:24:24 +0000 (15:24 -0300)]
mx6cuboxi: remove unused code

After enabling SPL_OF_CONTROL, SPL_DM and SPL_DM_MMC the MMC
initialization code is not longer needed.

This patch removes the unused code.

Signed-off-by: Walter Lozano <walter.lozano@collabora.com>
10 months agomx6cuboxi: enable OF_CONTROL and DM in SPL
Walter Lozano [Tue, 19 May 2020 18:24:23 +0000 (15:24 -0300)]
mx6cuboxi: enable OF_CONTROL and DM in SPL

In order to take the beneficts of DT and DM in SPL, like reusing the code
and avoid redundancy, enable SPL_OF_CONTROL, SPL_DM and SPL_DM_MMC.

With this new configuration SPL image is 50 KB, higher than the
38 KB from the previous version, but it still under the 68 KB limit.

Signed-off-by: Walter Lozano <walter.lozano@collabora.com>
10 months agomx6cuboxi: customize board_boot_order to access eMMC
Walter Lozano [Tue, 19 May 2020 18:24:22 +0000 (15:24 -0300)]
mx6cuboxi: customize board_boot_order to access eMMC

In SPL legacy code only one MMC device is created, based on BOOT_CFG
register, which can be either SD or eMMC. In this context
board_boot_order return always MMC1 when configure to boot from
SD/eMMC. After switching to DM both SD and eMMC devices are created
based on the information available on DT, but as board_boot_order
only returns MMC1 is not possible to boot from eMMC.

This patch customizes board_boot_order taking into account BOOT_CFG
register to point to correct MMC1 / MMC2 device. Additionally, handle
IO mux for the desired boot device.

Signed-off-by: Walter Lozano <walter.lozano@collabora.com>
10 months agomx6cuboxi: enable MMC and eMMC in DT for SPL
Walter Lozano [Tue, 19 May 2020 18:24:21 +0000 (15:24 -0300)]
mx6cuboxi: enable MMC and eMMC in DT for SPL

Signed-off-by: Walter Lozano <walter.lozano@collabora.com>
10 months agomx6ull_14x14_evk_plugin: Convert to DM_ETH
Fabio Estevam [Tue, 9 Jun 2020 13:44:00 +0000 (10:44 -0300)]
mx6ull_14x14_evk_plugin: Convert to DM_ETH

Convert to DM_ETH to avoid board removal from the project.

Signed-off-by: Fabio Estevam <festevam@gmail.com>
10 months agomx6slevk_spl: Convert to DM_ETH
Fabio Estevam [Tue, 9 Jun 2020 13:43:59 +0000 (10:43 -0300)]
mx6slevk_spl: Convert to DM_ETH

Convert to DM_ETH to avoid board removal from the project.

Signed-off-by: Fabio Estevam <festevam@gmail.com>
10 months agomx6slevk_spinor: Convert to DM_ETH
Fabio Estevam [Tue, 9 Jun 2020 13:43:58 +0000 (10:43 -0300)]
mx6slevk_spinor: Convert to DM_ETH

Convert to DM_ETH to avoid board removal from the project.

Signed-off-by: Fabio Estevam <festevam@gmail.com>
10 months agoarm: dts: imx: fsl-imx8qm.dtsi: fix gpio aliases
Ye Li [Wed, 10 Jun 2020 03:28:03 +0000 (20:28 -0700)]
arm: dts: imx: fsl-imx8qm.dtsi: fix gpio aliases

Current aliases missed gpio0 node, and this node shoud be
aliased to gpio index 0 to align with i.MX8QXP. Otherwise, we
will get below message when running "gpio status" command, and
see the reason by "dm uclass".

=> gpio status
Device 'gpio@5d090000': seq 0 is in use by 'gpio@5d080000'
Device 'gpio@5d0a0000': seq 1 is in use by 'gpio@5d090000'
Device 'gpio@5d0b0000': seq 2 is in use by 'gpio@5d0a0000'

=> dm uclass
uclass 36: gpio
0   * gpio@5d080000 @ fbaefb90, seq 0, (req -1)
1   * gpio@5d090000 @ fbaefc70, seq 1, (req 0)
2   * gpio@5d0a0000 @ fbaefd50, seq 2, (req 1)
3   * gpio@5d0b0000 @ fbaefe30, seq 5, (req 2)
4   * gpio@5d0c0000 @ fbaeff10, seq 3, (req 3)
5   * gpio@5d0d0000 @ fbaefff0, seq 4, (req 4)
6   * gpio@5d0e0000 @ fbaf00d0, seq 6, (req 5)
7   * gpio@5d0f0000 @ fbaf01b0, seq 7, (req 6)

Signed-off-by: Ye Li <ye.li@nxp.com>
10 months agogpio: mxc_gpio: change gpio index for i.MX8
Ye Li [Wed, 10 Jun 2020 03:28:02 +0000 (20:28 -0700)]
gpio: mxc_gpio: change gpio index for i.MX8

Since the i.MX8 GPIO banks are indexed from 0 not 1 on other i.MX
platforms, so we have to adjust the index accordingly.

Signed-off-by: Adrian Alonso <adrian.alonso@nxp.com>
Signed-off-by: Ye Li <ye.li@nxp.com>
10 months agomx6ul_14x14_evk: Avoid overlap of environment over U-Boot proper
Otavio Salvador [Tue, 9 Jun 2020 16:22:50 +0000 (13:22 -0300)]
mx6ul_14x14_evk: Avoid overlap of environment over U-Boot proper

We need to change the environment offset to avoid corrupting the U-Boot
binary when saving it.

Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Reviewed-by: Fabio Estevam <festevam@gmail.com>
10 months agomx6ul_14x14_evk: Enable SPL USB and SDP support
Otavio Salvador [Tue, 9 Jun 2020 16:22:49 +0000 (13:22 -0300)]
mx6ul_14x14_evk: Enable SPL USB and SDP support

This fixes the boot from USB loader, which is critical to easy the
manufacture process.

Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Reviewed-by: Fabio Estevam <festevam@gmail.com>
10 months agoARM: imx: soc: Select default TEXT_BASE for MX7
Marek Vasut [Thu, 21 May 2020 23:13:54 +0000 (01:13 +0200)]
ARM: imx: soc: Select default TEXT_BASE for MX7

Select default U-Boot and SPL text base for the MX7 SoC. The U-Boot
text base is picked as the one used by various MX7 boards. The SPL
text base however is different.

The SPL text base is set to 0x912000 instead of the usual 0x911000,
that is because the 0x911000 value cannot work. Using 0x911000 as a
SPL text base will result in the DCD header being placed below the
0x911000 address, which is a reserved SRAM area which must not be
used. This will actually trigger eMMC boot failure on MX7D at least.
Hence the increment.

Update all boards affected by this SPL problem to the new SPL_TEXT_BASE.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: NXP i.MX U-Boot Team <uboot-imx@nxp.com>
Cc: Peng Fan <peng.fan@nxp.com>
Cc: Stefano Babic <sbabic@denx.de>
10 months agoARM: imx: soc: Switch BOARD_EARLY_INIT_F to imply on MX7
Marek Vasut [Thu, 21 May 2020 23:13:00 +0000 (01:13 +0200)]
ARM: imx: soc: Switch BOARD_EARLY_INIT_F to imply on MX7

There are systems where board_early_init_f() is plain empty. Switch
the config option from "select" to "imply", to permit user to unset
the BOARD_EARLY_INIT_F if it were to be empty.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: NXP i.MX U-Boot Team <uboot-imx@nxp.com>
Cc: Peng Fan <peng.fan@nxp.com>
Cc: Stefano Babic <sbabic@denx.de>
10 months agoARM: imx: ddr: Fill in missing DDRC ZQCTLx on i.MX7
Marek Vasut [Thu, 21 May 2020 23:12:39 +0000 (01:12 +0200)]
ARM: imx: ddr: Fill in missing DDRC ZQCTLx on i.MX7

The iMX7 defines further DDRC ZQCTLx registers, however those were
thus far missing from the list of registers and not programmed. On
systems with LPDDR2 or DDR3, those registers must be programmed with
correct values, otherwise the DRAM may not work. However, existing
systems which worked without programming these registers before are
now setting those registers to 0, which is the default value, so no
functional change there.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: NXP i.MX U-Boot Team <uboot-imx@nxp.com>
Cc: Peng Fan <peng.fan@nxp.com>
Cc: Stefano Babic <sbabic@denx.de>
10 months agoimx: imx8qm_rom7720_a1: update README
Oliver Graute [Wed, 20 May 2020 11:47:06 +0000 (13:47 +0200)]
imx: imx8qm_rom7720_a1: update README

Update README to extract firmware from scripts

Signed-off-by: Oliver Graute <oliver.graute@kococonnector.com>
10 months agoARM: imx6: Fetch MAC address in board_init_late() on DH iMX6 PDK2
Marek Vasut [Thu, 14 May 2020 16:49:10 +0000 (18:49 +0200)]
ARM: imx6: Fetch MAC address in board_init_late() on DH iMX6 PDK2

This is needed to obtain the MAC from EEPROM/OTP only after the final
env is populated, otherwise the ethaddr might be overriden.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Claudius Heine <ch@denx.de>
Cc: Harald Seiler <hws@denx.de>
Cc: Ludwig Zenz <lzenz@dh-electronics.com>
Cc: Stefano Babic <sbabic@denx.de>
10 months agoarm: wandboard: move CONFIG_MXC_UART to defconfig
Heinrich Schuchardt [Sun, 7 Jun 2020 23:43:47 +0000 (01:43 +0200)]
arm: wandboard: move CONFIG_MXC_UART to defconfig

For using a debug UART on the Wandboard CONFIG_MXC_UART=y must be set in
the .config file.

To avoid duplicate definitions move the setting from
include/configs/wandboard.h to configs/wandboard_defconfig.

Document the debug UART settings in the README.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Fabio Estevam <festevam@gmail.com>
10 months agoFix MMC access on Sabrelite
Martyn Welch [Thu, 11 Jun 2020 17:53:50 +0000 (18:53 +0100)]
Fix MMC access on Sabrelite

It appears that MMC access on the Sabrelite has been broken since
cdcaee9518:

Loading Environment from MMC... Card did not respond to voltage select!
*** Warning - No block device, using default environment

Remove the board_mmc_init() and related entries now that we should be
using DM_MMC, add PINCTRL so that things work as expected.

Signed-off-by: Martyn Welch <martyn.welch@collabora.com>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
Reviewed-by: Fabio Estevam <festevam@gmail.com>
Tested-by: Troy Kisky <troy.kisky@boundarydevices.com>
Acked-by: Troy Kisky <troy.kisky@boundarydevices.com>
10 months agocolibri_imx6: boot env configuration updates
Igor Opaniuk [Tue, 16 Jun 2020 19:20:11 +0000 (22:20 +0300)]
colibri_imx6: boot env configuration updates

1. Drop legacy emmcboot wrapper from env.
2. Change the "boot try" order. Default one is: SD -> eMMC -> USB -> DHCP
3. Drop DFU defines

Signed-off-by: Igor Opaniuk <igor.opaniuk@toradex.com>
10 months agocolibri_imx7: boot env configuration updates
Igor Opaniuk [Tue, 16 Jun 2020 19:20:10 +0000 (22:20 +0300)]
colibri_imx7: boot env configuration updates

1. Drop legacy emmcboot wrapper from env.
2. Change the "boot try" order. Default one is: SD -> eMMC -> USB -> DHCP
3. Drop DFU defines

Signed-off-by: Igor Opaniuk <igor.opaniuk@toradex.com>
10 months agoapalis_imx6: boot env configuration updates
Igor Opaniuk [Tue, 16 Jun 2020 19:20:09 +0000 (22:20 +0300)]
apalis_imx6: boot env configuration updates

1. Drop legacy emmcboot wrapper from env.
2. Change the "boot try" order. Default one is: SD -> eMMC -> USB -> DHCP
3. Drop DFU defines

Signed-off-by: Igor Opaniuk <igor.opaniuk@toradex.com>
10 months agotoradex: imx: enable BOOTCOUNT feature
Igor Opaniuk [Tue, 16 Jun 2020 19:20:08 +0000 (22:20 +0300)]
toradex: imx: enable BOOTCOUNT feature

This introduces automatic boot counter that increases after every
reset.After a power-on reset, it will be initialized with 1,
and each reboot will increment the value by 1. By default it's
disabled if bootlimit isn't set.

To enable this feature you have set bootcount limit ("bootlimit"),
alternate boot action ("altbootcmd") that will be performed if
the new value of bootcount exceeds the value of bootlimit, and
"upgrade_available" to let U-Boot automatically increase and save
the counter value after every reset:

> setenv bootlimit 5
> setenv upgrade_available 1
> setenv altbootcmd "bootm ..."

In case the bootlimit exceeds, the message will be shown and
albootcmd executed:
Warning: Bootlimit (5) exceeded. Using altbootcmd.

To reset bootcount run:
> bootcount reset

Print current value:
> bootcount print

Signed-off-by: Igor Opaniuk <igor.opaniuk@toradex.com>
10 months agoapalis-tk1: fix setting fdtfile value
Igor Opaniuk [Tue, 16 Jun 2020 19:20:07 +0000 (22:20 +0300)]
apalis-tk1: fix setting fdtfile value

s/fdt-module/fdt_module/g, as we don't use dash in fdt_file anymore.

Fixes: 4c63a601("apalis-tk1: support v1.2 hardware revision")
Signed-off-by: Igor Opaniuk <igor.opaniuk@toradex.com>
10 months agoapalis-tk1: enable distroboot
Igor Opaniuk [Tue, 16 Jun 2020 19:20:06 +0000 (22:20 +0300)]
apalis-tk1: enable distroboot

1. Use distro_bootcmd as default boot command instead of
legacy emmcboot wrapper.
2. Drop emmcboot and sdboot wrappers.
3. Provide proper boot order for Distro Boot.

Signed-off-by: Igor Opaniuk <igor.opaniuk@toradex.com>
10 months agoconfigs/colibri_vf.h: drop sdboot in favour of distro_bootcmd
Max Krummenacher [Tue, 16 Jun 2020 19:20:05 +0000 (22:20 +0300)]
configs/colibri_vf.h: drop sdboot in favour of distro_bootcmd

The distro bootscript uses kernel_image to get the file name of
the kernel, so change that variable name.
UBI boot has precedence in the default boot command. If one wants
to boot from SD with a working NAND installation stop in U-Boot
and enter:

setenv fdtfile ${soc}-colibri-${fdt_board}.dtb && run distro_bootcmd

Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
10 months agocolibri_vf_defconfig: enable part cmd
Max Krummenacher [Tue, 16 Jun 2020 19:20:04 +0000 (22:20 +0300)]
colibri_vf_defconfig: enable part cmd

This allows to boot from SD/USB with passing the rootfs partition via UUID.

Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
10 months agocolibri-imx6ull/imx7: define bootubipart for distro boot
Stefan Agner [Tue, 16 Jun 2020 19:20:03 +0000 (22:20 +0300)]
colibri-imx6ull/imx7: define bootubipart for distro boot

When using distro boot to boot from UBI volumes the boot partition
has been hardcoded to "UBI" (capital letters). However, our default
MTD layout uses "ubi" (lower case letter). Define "ubi" as the
default UBI partition for distro boot for Toradex. This allows to
use distro boot without having to redefine the MTD partition layout
which is useful for TorizonCore.

Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
10 months agocolibri_imx7: add addresses required for distro boot
Stefan Agner [Tue, 16 Jun 2020 19:20:02 +0000 (22:20 +0300)]
colibri_imx7: add addresses required for distro boot

Define addresses required for full distro boot support.

Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
10 months agocolibri-imx8x: declare consoleargs
Oleksandr Suvorov [Tue, 16 Jun 2020 19:20:01 +0000 (22:20 +0300)]
colibri-imx8x: declare consoleargs

Store all console-related kernel parameters
in dedicated variable.

Signed-off-by: Oleksandr Suvorov <oleksandr.suvorov@toradex.com>
10 months agoapalis/colibri-imx8: re-enable CONFIG_IMX_SCU_THERMAL
Max Krummenacher [Tue, 16 Jun 2020 19:20:00 +0000 (22:20 +0300)]
apalis/colibri-imx8: re-enable CONFIG_IMX_SCU_THERMAL

This got dropped by a global 'make savedefconfig' resync as
required patches are still in flight.

Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
10 months agoapalis-imx8: enable of_system_setup
Marcel Ziswiler [Tue, 16 Jun 2020 19:19:59 +0000 (22:19 +0300)]
apalis-imx8: enable of_system_setup

Enable CONFIG_OF_SYSTEM_DEFAULT for Apalis iMX8.

Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
10 months agoARM: dts: imx6qdl-sabresd: Fix AR8031 phy-mode
Fabio Estevam [Wed, 17 Jun 2020 17:33:16 +0000 (14:33 -0300)]
ARM: dts: imx6qdl-sabresd: Fix AR8031 phy-mode

As per kernel commit 0672d22a1924 ("ARM: dts: imx: Fix the AR803X phy-mode)
the correct phy-mode should be "rgmii-id", so fix it accordingly
to fix the Ethernet regression.

This problem has been exposed by commit:

commit 13114f38e2ccea9386726d8b9831dfc310589548
Author: Vladimir Oltean <vladimir.oltean@nxp.com>
Date:   Thu May 7 00:11:51 2020 +0200

    phy: atheros: Explicitly disable RGMII delays

    To eliminate any doubts about the out-of-reset value of the PHY, that
    the driver previously relied on.

    If bisecting shows that this commit breaks your board you probably have
    a wrong PHY interface mode. You probably want the
    PHY_INTERFACE_MODE_RGMII_RXID or PHY_INTERFACE_MODE_RGMII_ID mode.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Signed-off-by: Fabio Estevam <festevam@gmail.com>
10 months agoARM: dts: imx6qdl-sabreauto: Fix AR8031 phy-mode
Fabio Estevam [Fri, 19 Jun 2020 01:14:42 +0000 (22:14 -0300)]
ARM: dts: imx6qdl-sabreauto: Fix AR8031 phy-mode

As per kernel commit 0672d22a1924 ("ARM: dts: imx: Fix the AR803X phy-mode)
the correct phy-mode should be "rgmii-id", so fix it accordingly
to fix the Ethernet regression.

This problem has been exposed by commit:

commit 13114f38e2ccea9386726d8b9831dfc310589548
Author: Vladimir Oltean <vladimir.oltean@nxp.com>
Date:   Thu May 7 00:11:51 2020 +0200

    phy: atheros: Explicitly disable RGMII delays

    To eliminate any doubts about the out-of-reset value of the PHY, that
    the driver previously relied on.

    If bisecting shows that this commit breaks your board you probably have
    a wrong PHY interface mode. You probably want the
    PHY_INTERFACE_MODE_RGMII_RXID or PHY_INTERFACE_MODE_RGMII_ID mode.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Fix the phy-mode accordingly to fix the regression.

Signed-off-by: Fabio Estevam <festevam@gmail.com>
10 months agomx6cuboxi: Convert to DM_ETH
Fabio Estevam [Thu, 18 Jun 2020 23:21:20 +0000 (20:21 -0300)]
mx6cuboxi: Convert to DM_ETH

Migration to DM_ETH is mandatory, so convert mx6cuboxi to Ethernet
Driver Model.

This also brings the benefit of restoring Ethernet functionality.

Reported-by: Tom Rini <trini@konsulko.com>
Signed-off-by: Fabio Estevam <festevam@gmail.com>
Tested-by: Tom Rini <trini@konsulko.com>
10 months agoARM: dts: imx6qdl-sr-som: Sync with kernel 5.8-rc1
Fabio Estevam [Thu, 18 Jun 2020 23:21:19 +0000 (20:21 -0300)]
ARM: dts: imx6qdl-sr-som: Sync with kernel 5.8-rc1

Sync the device tree with 5.8-rc1.

It basically contains the following extra kernel commit:

commit 86b08bd5b99480b79a25343f24c1b8c4ddcb5c09
Author: Russell King <rmk+kernel@armlinux.org.uk>
Date:   Wed Apr 15 16:44:17 2020 +0100

    ARM: dts: imx6-sr-som: add ethernet PHY configuration

    Add ethernet PHY configuration ahead of removing the quirk that
    configures the clocking mode for the PHY.  The RGMII delay is
    already set correctly.

Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: Fabio Estevam <festevam@gmail.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
, which passes the 'qca,clk-out-frequency' property and it is important
to specify the correct frequency generated by the AR8035.

Signed-off-by: Fabio Estevam <festevam@gmail.com>
Tested-by: Tom Rini <trini@konsulko.com>
10 months agonet: fec: Allow the PHY node to be retrieved
Fabio Estevam [Thu, 18 Jun 2020 23:21:18 +0000 (20:21 -0300)]
net: fec: Allow the PHY node to be retrieved

As we move towards driver model, it is required to let the FEC driver
know how to properly deal with an Ethernet PHY subnode in the device tree.

For example:

 &fec {
  pinctrl-names = "default";
  pinctrl-0 = <&pinctrl_microsom_enet_ar8035>;
phy-handle = <&phy>;
  phy-mode = "rgmii-id";
  phy-reset-duration = <2>;
  phy-reset-gpios = <&gpio4 15 GPIO_ACTIVE_LOW>;
  status = "okay";

mdio {
#address-cells = <1>;
#size-cells = <0>;

phy: ethernet-phy@0 {
reg = <0>;
qca,clk-out-frequency = <125000000>;
};
};
 };

Currently the PHY node pointer is incorrectly associated with the
Ethernel controller instead of the PHY node itself.

This causes the PHY properties, such as "qca,clk-out-frequency" in
the example above to not get parsed.

Fix this problem by populating the phy_of_node node.

Suggested-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Fabio Estevam <festevam@gmail.com>
Tested-by: Tom Rini <trini@konsulko.com>
10 months agophy: atheros: ar8035: Fix clock output calculation
Fabio Estevam [Thu, 18 Jun 2020 23:21:17 +0000 (20:21 -0300)]
phy: atheros: ar8035: Fix clock output calculation

The clock ouput frequency is calculated incorrectly for AR8035 due to
wrong masking of priv->clk_25m_reg and priv->clk_25m_mask.

This same issue has been already fixed in the kernel by:

commit b1f4c209d84057b6d40b939b6e4404854271d797
Author: Oleksij Rempel <o.rempel@pengutronix.de>
Date:   Wed Apr 1 11:57:32 2020 +0200

    net: phy: at803x: fix clock sink configuration on ATH8030 and ATH8035

    The masks in priv->clk_25m_reg and priv->clk_25m_mask are one-bits-set
    for the values that comprise the fields, not zero-bits-set.

    This patch fixes the clock frequency configuration for ATH8030 and
    ATH8035 Atheros PHYs by removing the erroneous "~".

    To reproduce this bug, configure the PHY  with the device tree binding
    "qca,clk-out-frequency" and remove the machine specific PHY fixups.

    Fixes: 2f664823a47021 ("net: phy: at803x: add device tree binding")
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Reported-by: Russell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: Russell King <rmk+kernel@armlinux.org.uk>
Tested-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
Apply the same fix in the U-Boot driver.

Tested on a i.MX6 Hummingboard.

Signed-off-by: Fabio Estevam <festevam@gmail.com>
Reviewed-by: Michael Walle <michael@walle.cc>
Tested-by: Tom Rini <trini@konsulko.com>
10 months agoMerge https://gitlab.denx.de/u-boot/custodians/u-boot-sh
Tom Rini [Sat, 20 Jun 2020 22:51:50 +0000 (18:51 -0400)]
Merge https://gitlab.denx.de/u-boot/custodians/u-boot-sh

10 months agoMerge tag 'efi-2020-07-rc5-2' of https://gitlab.denx.de/u-boot/custodians/u-boot-efi
Tom Rini [Fri, 19 Jun 2020 18:22:23 +0000 (14:22 -0400)]
Merge tag 'efi-2020-07-rc5-2' of https://gitlab.denx.de/u-boot/custodians/u-boot-efi

Pull request for UEFI sub-system for efi-2020-07-rc5 (2)

Fix memory reservations: do not use random bytes from the stack.

10 months agoMerge tag 'u-boot-stm32-20200619' of https://gitlab.denx.de/u-boot/custodians/u-boot-stm
Tom Rini [Fri, 19 Jun 2020 16:41:24 +0000 (12:41 -0400)]
Merge tag 'u-boot-stm32-20200619' of https://gitlab.denx.de/u-boot/custodians/u-boot-stm

- fix SD card cart detect on DHCOM and ST boards

10 months agoMerge https://gitlab.denx.de/u-boot/custodians/u-boot-tegra
Tom Rini [Fri, 19 Jun 2020 16:41:05 +0000 (12:41 -0400)]
Merge https://gitlab.denx.de/u-boot/custodians/u-boot-tegra

10 months agoARM: dts: stm32: Reinstate card detect behavior on ST boards
Patrick Delaunay [Fri, 19 Jun 2020 09:19:45 +0000 (11:19 +0200)]
ARM: dts: stm32: Reinstate card detect behavior on ST boards

The cd-gpios with (GPIO_ACTIVE_LOW | GPIO_PULL_UP) gpio is thus far
unsupported, reinstate the old cd-gpios behavior until this handling
is fully implemented. This avoid potential issue for SDCard boot:
the card detect fails with floating gpio.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
10 months agoARM: dts: stm32: Reinstate card detect behavior on DHSOM
Marek Vasut [Thu, 18 Jun 2020 18:35:17 +0000 (20:35 +0200)]
ARM: dts: stm32: Reinstate card detect behavior on DHSOM

The cd-gpios with (GPIO_ACTIVE_LOW | GPIO_PULL_UP) gpio is thus far
unsupported, reinstate the old cd-gpios behavior until this handling
is fully implemented. This permits the DHSOM to boot from SD again,
without this patch the card detect fails.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com>
10 months agocmd: bootefi: Honor the address & size cells properties correctly
Atish Patra [Fri, 19 Jun 2020 01:51:50 +0000 (18:51 -0700)]
cmd: bootefi: Honor the address & size cells properties correctly

fdtdec_get_addr_size reads the uses a fixed value for address & size
cell properties which may not be correct always.

Use the auto variant of the function which automatically reads
 #address-cells & #size-cells from parent and uses to read the "reg"
property.

Signed-off-by: Atish Patra <atish.patra@wdc.com>
On 32 bit systems fdt_addr_t may be 4 bytes long but size is defined as
u64. Avoid filling the upper 4 bytes of the sizep parameter of
fdtdec_get_addr_size_auto_parent() with random bytes from the stack by
defining a separate variable fdt_size.

Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
10 months agot210: Nano: Add NVME support
Tom Warren [Mon, 11 May 2020 23:38:58 +0000 (16:38 -0700)]
t210: Nano: Add NVME support

This allows Nano to use the key M.2 on the CVB to connect and use a M.2
NVMe SSD stick. Works on my Nano B00, WD SN750 NVMe SSD shows up
w/'nvme' commands. Will add booting from NVMe via distro cmds in a future
patch.

Signed-off-by: Tom Warren <twarren@nvidia.com>
10 months agoMerge tag 'fixes-for-v2020.07-rc4' of https://gitlab.denx.de/u-boot/custodians/u...
Tom Rini [Thu, 18 Jun 2020 22:00:05 +0000 (18:00 -0400)]
Merge tag 'fixes-for-v2020.07-rc4' of https://gitlab.denx.de/u-boot/custodians/u-boot-video

- fix stdout in default environment for apalis_imx6, colibri_imx6
  and cm_fx6 boards

10 months agocolibri_t20: change maintainer
Igor Opaniuk [Mon, 13 Apr 2020 15:11:42 +0000 (18:11 +0300)]
colibri_t20: change maintainer

Take over maintainership for Colibri T20 module.

Signed-off-by: Igor Opaniuk <igor.opaniuk@toradex.com>
Acked-by: Lucas Stach <dev@lynxeye.de>
10 months agoARM: tegra: Enable PSCI support for Tegra210 and Tegra186
Jon Hunter [Thu, 18 Jun 2020 11:54:39 +0000 (12:54 +0100)]
ARM: tegra: Enable PSCI support for Tegra210 and Tegra186

The PSCI nodes are currently not populated for the Tegra210 and Tegra186
devices. This prevents the PSCI driver from being able to identify the
PSCI method used by these devices and causes the probe of the PSCI
driver to fail.

Since commit 81ea00838c68 ("efi_loader: PSCI reset and shutdown") was
added, which moves the PSCI EFI system reset handler into the PSCI
driver, this has prevented the EFI system reset from working for
Tegra210 and Tegra186. Therefore, populating these nodes is necessary
to fix the EFI system reset for Tegra210 and Tegra186.

Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Tested-by: Peter Robinson <pbrobinson@gmail.com>
10 months agofirmware: PSCI: Fix PSCI support for OF live trees
Jon Hunter [Thu, 18 Jun 2020 11:54:38 +0000 (12:54 +0100)]
firmware: PSCI: Fix PSCI support for OF live trees

When CONFIG_OF_LIVE is enabled, dev_of_offset() cannot be used and
if used returns an invalid offset. This causes the call to
fdt_stringlist_get() in the psci_probe() to fail to read the 'method'
property from the PSCI node for the device and hence prevents PSCI
from working. Fix this by using the ofnode_read_string() API instead
of the fdt_stringlist_get() because this will handle reading the
property both when CONFIG_OF_LIVE is enabled or disabled.

Due to the above problem and since commit 81ea00838c68 ("efi_loader:
PSCI reset and shutdown") was added, the EFI system reset has been
broken for Tegra210 and Tegra196 platforms. This also fixes the EFI
system reset for these Tegra platforms.

Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Tested-by: Peter Robinson <pbrobinson@gmail.com>
10 months agoAzure/GitLab: Move to latest Docker image
Tom Rini [Thu, 18 Jun 2020 17:51:39 +0000 (13:51 -0400)]
Azure/GitLab: Move to latest Docker image

- Add guestfstools, efitools
- Latest Ubuntu/bionic snapshot

Signed-off-by: Tom Rini <trini@konsulko.com>
10 months agoARM: rmobile: Increase PHY auto-negotiation timeout to 20s
Marek Vasut [Sun, 14 Jun 2020 12:41:07 +0000 (14:41 +0200)]
ARM: rmobile: Increase PHY auto-negotiation timeout to 20s

The ethernet PHY on all known Gen3 boards takes a while to come out
of reset, increase the auto-negotiation timeout to prevent it from
timing out in case the ethernet is used right after the board was
reset.

Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com>
Cc: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>
10 months agocolibri_imx6: fix video stdout in default environment
Anatolij Gustschin [Tue, 26 May 2020 08:50:17 +0000 (10:50 +0200)]
colibri_imx6: fix video stdout in default environment

After migration to DM 'vga' name is not longer supported,
change it to 'vidconsole' in the default environment.

Signed-off-by: Anatolij Gustschin <agust@denx.de>
Acked-by: Igor Opaniuk <igor.opaniuk@toradex.com>
10 months agoapalis_imx6: fix video stdout in default environment
Anatolij Gustschin [Tue, 26 May 2020 08:43:44 +0000 (10:43 +0200)]
apalis_imx6: fix video stdout in default environment

After migration to DM 'vga' name is not longer supported,
change it to 'vidconsole' in the default environment.

Signed-off-by: Anatolij Gustschin <agust@denx.de>
Acked-by: Igor Opaniuk <igor.opaniuk@toradex.com>
10 months agocm_fx6: fix video stdout in default environment
Anatolij Gustschin [Tue, 26 May 2020 08:52:00 +0000 (10:52 +0200)]
cm_fx6: fix video stdout in default environment

After migration to DM 'vga' name is not longer supported,
change it to 'vidconsole' in the default environment.

Signed-off-by: Anatolij Gustschin <agust@denx.de>
Cc: Nikita Kiryanov <nikita@compulab.co.il>
10 months agoMerge tag 'u-boot-amlogic-20200618' of https://gitlab.denx.de/u-boot/custodians/u...
Tom Rini [Thu, 18 Jun 2020 13:46:55 +0000 (09:46 -0400)]
Merge tag 'u-boot-amlogic-20200618' of https://gitlab.denx.de/u-boot/custodians/u-boot-amlogic

- move amlogic board documentation into doc/board in RsT format

10 months agoboard: amlogic: add board doc files to MAINTAINERS
Neil Armstrong [Thu, 11 Jun 2020 08:57:39 +0000 (10:57 +0200)]
board: amlogic: add board doc files to MAINTAINERS

Add the recently added reStructuredText board documentation in the
appropriate MAINTAINERS files.

Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
10 months agoboard: amlogic: move boards doc into doc/board/amlogic
Neil Armstrong [Thu, 11 Jun 2020 08:53:49 +0000 (10:53 +0200)]
board: amlogic: move boards doc into doc/board/amlogic

Move the natural text Amlogic board README files to doc/board/amlogic
into reStructuredText and :
- add reStructuredText markup for bash code
- fix secondary titles markup
- move board support into global support matrix

Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
10 months agoMerge tag 'u-boot-stm32-20200616' of https://gitlab.denx.de/u-boot/custodians/u-boot-stm
Tom Rini [Tue, 16 Jun 2020 13:18:56 +0000 (09:18 -0400)]
Merge tag 'u-boot-stm32-20200616' of https://gitlab.denx.de/u-boot/custodians/u-boot-stm

- fix boot with OP-TEE for stm32mp15 boards

10 months agodts: ARM: stm32mp15: add OP-TEE node in u-boot DTSI
Etienne Carriere [Fri, 5 Jun 2020 07:24:30 +0000 (09:24 +0200)]
dts: ARM: stm32mp15: add OP-TEE node in u-boot DTSI

Add OP-TEE firmware node in stm32mp15 U-Boot DTSI. This node is
needed since commit [1] that changed U-Boot/stm32mp15 to detect
OP-TEE availability by probing the resource instead of relying on
U-Boot configuration. The software sequence implemented by [1] is
fine but U-Boot DTS/DTSI files were not updated accordingly since,
hence OP-TEE presence is never detected by U-Boot, preventing Linux
kernel from using OP-TEE resources.

For consistency and to synchronize stm32mp15 DTSI files (excluding
U-Boot specific DTSI files) with the Linux kernel ones, this change
also moves the OP-TEE reserved memory nodes from board generic DTSI
files to U-Boot specific board DTSI files.

Link: [1] commit 43df0a159df6 ("stm32mp1: dynamically detect op-tee presence")
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
10 months agoboard: stm32mp1: fix handling of DT OP-TEE reserved memory nodes
Etienne Carriere [Fri, 5 Jun 2020 07:24:29 +0000 (09:24 +0200)]
board: stm32mp1: fix handling of DT OP-TEE reserved memory nodes

Fix the sequence in stm32mp1 fdt.c that disables OP-TEE resources
defined in FDT when U-boot detects OP-TEE firmware is not present.

Before this change, helper function stm32_fdt_disable_optee()
set property status to "disabled" for the OP-TEE reserved memory
nodes but this has no impact since Linux kernel does not consider
the status property for reserved-memory subnodes. This change
make U-Boot to attempt to delete the node instead.

Fixes: 4a1b975dac02 ("board: stm32mp1: reserve memory for OP-TEE in device tree")
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
10 months agoMerge tag 'efi-2020-07-rc5' of https://gitlab.denx.de/u-boot/custodians/u-boot-efi
Tom Rini [Mon, 15 Jun 2020 20:44:58 +0000 (16:44 -0400)]
Merge tag 'efi-2020-07-rc5' of https://gitlab.denx.de/u-boot/custodians/u-boot-efi

Pull request for UEFI sub-system for efi-2020-07-rc5

Use correct printf code in efi_image_parse().
Add random number generation to HTML documentation.

10 months agoregulator: fix: enable gpio when requested
Mark Kettenis [Mon, 27 Apr 2020 09:09:47 +0000 (11:09 +0200)]
regulator: fix: enable gpio when requested

The fix in commit b7adcdd073c0 has the side-effect that the regulator
will be disabled when requesting the relevant gpio in
regulator_common_ofdata_to_platdata() and enabled in
regulator_pre_probe() when the regulator was already enabled.
This leads to a short interruption in the 3.3V power to the PCIe
slot on the firefly-rk3399 which makes an ADATA SX8000NP NVMe SSD
unhappy.

Fix this by setting the GPIOD_IS_OUT_ACTIVE flag again when the
'regulator-boot-on' property is set, but check for this property
explicitly instead of relying on the "boot_on" member of
the uclass platdata.

Signed-off-by: Mark Kettenis <kettenis@openbsd.org>
Tested-by: Patrice Chotard <patrice.chotard@st.com>
10 months agoMerge branch '2020-06-15-misc-bugfixes'
Tom Rini [Mon, 15 Jun 2020 15:24:42 +0000 (11:24 -0400)]
Merge branch '2020-06-15-misc-bugfixes'

- Assorted bug fixes

10 months agoenv: Kconfig: cosmetics: update comment for SYS_RELOC_GD_ENV_ADDR
Patrick Delaunay [Wed, 10 Jun 2020 17:28:42 +0000 (19:28 +0200)]
env: Kconfig: cosmetics: update comment for SYS_RELOC_GD_ENV_ADDR

Update the comment for SYS_RELOC_GD_ENV_ADDR as gd->env_addr is updated
in board_r.c::initr_reloc_global_data() under the compilation flags
CONFIG_SYS_RELOC_GD_ENV_ADDR

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
10 months agoconfig: qemu: increase SYS_MALLOC_F_LEN
Heinrich Schuchardt [Sun, 7 Jun 2020 09:54:00 +0000 (11:54 +0200)]
config: qemu: increase SYS_MALLOC_F_LEN

Several configutation options require additional memory before relocation:

* CONSOLE_RECORD
* LOG
* RSA

The current default of 0x400 is too small to encompass them all. Increase
the value of SYS_MALLOC_F_LEN to 0x2000 for ARCH_QEMU.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
10 months agoRevert "lib: Improve _parse_integer_fixup_radix base 16 detection"
Sean Anderson [Sun, 7 Jun 2020 05:36:45 +0000 (01:36 -0400)]
Revert "lib: Improve _parse_integer_fixup_radix base 16 detection"

This reverts commit 0486497e2b5f4d36fa968a1a60fea358cbf70b65.

The strtoul has well-defined semantics. It is defined by the C standard and
POSIX. To quote the relevant section of the man pages,

> If base is zero or 16, the string may then include a "0x" prefix, and the
> number will be read in base 16; otherwise, a zero base is taken as 10
> (decimal) unless the next character is '0', in which case it is taken as
> 8 (octal).

Keeping these semantics is important for several reasons. First, it is very
surprising for standard library functions to behave differently than usual.
Every other implementation of strtoul has different semantics than the
implementation in U-Boot at the moment. Second, it can result in very
surprising results from small changes. For example, changing the string
"1f" to "20" causes the parsed value to *decrease*. Forcing use of the "0x"
prefix to specify hexidecimal numbers is a feature, not a bug. Lastly, this
is slightly less performant, since the entire number is parsed twice.

This fixes the str_simple_strtoul test failing with

test/str_ut.c:29, run_strtoul(): expect_val == val: Expected 0x44b (1099), got 0x1099ab (1087915)
test/str_ut.c:46, str_simple_strtoul(): 0 == run_strtoul(uts, str2, 0, 1099, 4): Expected 0x0 (0), got 0x1 (1)

Signed-off-by: Sean Anderson <seanga2@gmail.com>
CC: Michal Simek <michal.simek@xilinx.com>
CC: Shiril Tichkule <shirilt@xilinx.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
10 months agotrivial: Fix booot occurences
Michal Simek [Fri, 5 Jun 2020 09:45:12 +0000 (11:45 +0200)]
trivial: Fix booot occurences

s/booot/boot/g

The first ase is booot instead of boot and second u-booot instead of
u-boot.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
10 months agooptee: fix copy of optee reserved-memory node
Etienne Carriere [Fri, 5 Jun 2020 07:22:11 +0000 (09:22 +0200)]
optee: fix copy of optee reserved-memory node

Fix the loop that parses FDT for a reserved memory node named "optee".

Before this change, if at least one subnode was found in the
reserved-memory node, the function endlessly looped since instruction
continue returned back in the loop without updating variable subnode.
This change fixes the issue by using a for loop.

Fixes: 6ccb05eae01b ("image: fdt: copy possible optee nodes to a loaded devicetree")
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
10 months agoMerge tag 'mmc-2020-6-15' of https://gitlab.denx.de/u-boot/custodians/u-boot-mmc
Tom Rini [Mon, 15 Jun 2020 14:20:35 +0000 (10:20 -0400)]
Merge tag 'mmc-2020-6-15' of https://gitlab.denx.de/u-boot/custodians/u-boot-mmc

- fsl_esdhc sdr104 and hs200 fix and error path fix
- fsl_esdhc workaround 3.3v io issue
- ca_dw_mmc cleanup
- presidio-asic emmc DT update.

11 months agommc: fsl_esdhc: Gracefully fail on unsupported voltage switch
Marek Vasut [Fri, 22 May 2020 16:28:33 +0000 (18:28 +0200)]
mmc: fsl_esdhc: Gracefully fail on unsupported voltage switch

Unsupported voltage on voltage switch is not an error, do not
print error message in such a case. This happens e.g. if the
eMMC is already in 1V8 mode or when testing 1V2 mode operation
on systems which only do 3V3/1V8 switching.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <fabio.estevam@nxp.com>
Cc: Jaehoon Chung <jh80.chung@samsung.com>
Cc: Peng Fan <peng.fan@nxp.com>
Cc: Stefano Babic <sbabic@denx.de>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
11 months agommc: fsl_esdhc: Fix SDR104 and HS200 support
Marek Vasut [Fri, 22 May 2020 16:19:08 +0000 (18:19 +0200)]
mmc: fsl_esdhc: Fix SDR104 and HS200 support

The 3V3/1V8 switching could never have worked on any of the iMXes
ever since 51313b49f2 ("mmc: fsl_esdhc: support SDR104 and HS200"),
because that commit uses priv->vqmmc_dev when switching voltages on
mode switch, while local vqmmc_dev in probe to store the regulator
pointer. Those are two different variables with the same name. So
the priv->vqmmc_dev was always NULL and thus voltage switch between
modes never really suceeded.

Fix this by assigning priv->vqmmc_dev with value of the vqmmc_dev
in probe.

Fixes: 51313b49f2 ("mmc: fsl_esdhc: support SDR104 and HS200")
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <fabio.estevam@nxp.com>
Cc: Jaehoon Chung <jh80.chung@samsung.com>
Cc: Peng Fan <peng.fan@nxp.com>
Cc: Stefano Babic <sbabic@denx.de>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
11 months agoboard: presidio-asic: update eMMC DT information
Alex Nemirovsky [Fri, 22 May 2020 00:24:23 +0000 (17:24 -0700)]
board: presidio-asic: update eMMC DT information

Change DT compatibility name to match change in driver's name.
Remove unused io_ds and fifo_mode fields from DT.

Signed-off-by: Alex Nemirovsky <alex.nemirovsky@cortina-access.com>
CC: Peng Fan <peng.fan@nxp.com>
CC: Jaehoon Chung <jh80.chung@samsung.com>
CC: Tom Rini <trini@konsulko.com>
11 months agommc: ca_dw_mmc: Misc cleanup of driver
Arthur Li [Fri, 22 May 2020 00:24:22 +0000 (17:24 -0700)]
mmc: ca_dw_mmc: Misc cleanup of driver

- Rename DT compatible name
- Remove uneccessary if-statement to support 8-bit buswidth
- Remove redundant error msg
- Use symbolic constants in switch statement

Signed-off-by: Arthur Li <arthur.li@cortina-access.com>
Signed-off-by: Alex Nemirovsky <alex.nemirovsky@cortina-access.com>
CC: Peng Fan <peng.fan@nxp.com>
CC: Jaehoon Chung <jh80.chung@samsung.com>
CC: Tom Rini <trini@konsulko.com>
11 months agommc: fsl_esdhc: workaround for hardware 3.3v IO reliability issue
Yangbo Lu [Tue, 19 May 2020 03:06:44 +0000 (11:06 +0800)]
mmc: fsl_esdhc: workaround for hardware 3.3v IO reliability issue

When eSDHC operates at 3.3v, damage can accumulate in an internal
level shifter at a higher than expected rate. The faster the interface
runs, the more damage accumulates. This issue now is found on LX2160A
eSDHC1 for only SD card.

The hardware workaround is recommended to use an on-board level shifter
that is 1.8v on SoC side and 3.3v on SD card side.

For boards without hardware workaround, this option could be enabled,
ensuring 1.8v IO voltage and disabling eSDHC if no card.
This option assumes no hotplug, and u-boot has to make all the way to
to linux to use 1.8v UHS-I speed mode if has card.
If you do not want the workaround for better user experience, of course
you can choose to not select it running eSDHC in unsafe mode.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
Acked-by: Peng Fan <peng.fan@nxp.com>
11 months agommc: fsl_esdhc: read register once for card inserted status
Yangbo Lu [Tue, 19 May 2020 03:06:43 +0000 (11:06 +0800)]
mmc: fsl_esdhc: read register once for card inserted status

No need to poll register for card inserted status.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
Acked-by: Peng Fan <peng.fan@nxp.com>
11 months agoefi_loader: printf code in efi_image_parse()
Heinrich Schuchardt [Sat, 6 Jun 2020 11:17:48 +0000 (13:17 +0200)]
efi_loader: printf code in efi_image_parse()

For size_t we have to use %zu for printing not %lu.

Fixes: 4540dabdcaca ("efi_loader: image_loader: support image
authentication")
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
11 months agodoc: random number generation
Heinrich Schuchardt [Sat, 13 Jun 2020 10:29:52 +0000 (12:29 +0200)]
doc: random number generation

Add random number generation APIs to the HTML documentation.
Fix style issues.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
11 months agoMerge https://gitlab.denx.de/u-boot/custodians/u-boot-usb
Tom Rini [Sun, 14 Jun 2020 16:01:14 +0000 (12:01 -0400)]
Merge https://gitlab.denx.de/u-boot/custodians/u-boot-usb

- eth/r8152: update the firmware

11 months agoMerge https://gitlab.denx.de/u-boot/custodians/u-boot-socfpga
Tom Rini [Sun, 14 Jun 2020 16:00:50 +0000 (12:00 -0400)]
Merge https://gitlab.denx.de/u-boot/custodians/u-boot-socfpga

- cyclone5 bugfix