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>
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>
Commit 0528979fa7 ("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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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 13114f38e2
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>
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 13114f38e2
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>
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>
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>
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>
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>
Fix memory reservations: do not use random bytes from the stack.
-----BEGIN PGP SIGNATURE-----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=j2MV
-----END PGP SIGNATURE-----
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.
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>
