video: extend stdout video console work-around for 'vga'
cfb_console driver uses 'vga' console name and we still have board
environments defining this name. Re-use existing DM_VIDEO work-
around for console name to support 'vga' name in stdout environment.
Signed-off-by: Anatolij Gustschin <agust@denx.de> Tested-by: Soeren Moch <smoch@web.de> Reviewed-by: Simon Glass <sjg@chromium.org> Reviewed-by: Tom Rini <trini@konsulko.com>
video: ipuv3: remove some useless code to reduce binary size
To enable DM_VIDEO we must decrease binary size to fix build
breakage for some boards, so drop not needed code. Also add
!DM_VIDEO guards which can be later removed when last non DM
users will be converted.
Converting some boards to DM_VIDEO results in build breakage due
to increased code size. Make video console specific commands
optional to reduce binary size.
Signed-off-by: Anatolij Gustschin <agust@denx.de> Reviewed-by: Simon Glass <sjg@chromium.org>
Tom Rini [Fri, 12 Jun 2020 18:58:12 +0000 (14:58 -0400)]
Merge branch '2020-06-12-next-net' into next
- Merge tbs2910 distro boot support and associated clean-ups and size
reduction.
- Assorted networking corrections / bugfixes.
- Drop smc911x standalone API example as it was likely non-functional for a
long time.
- Enhanced support for TI PHYs
- rtl8139 DM conversion
Ravik Hasija [Tue, 19 May 2020 04:35:43 +0000 (21:35 -0700)]
net: tftp: fix option validation as per RFCs
RFC2348, RFC2349:
- Option string is case in-sensitive.
- Client must generate ERR pkt in case option value mismatch in server OACK
- Fix debug print for options
Signed-off-by: Ravik Hasija <rahasij@linux.microsoft.com> Reviewed-By: Ramon Fried <rfried.dev@gmail.com>
Marek Vasut [Sat, 9 May 2020 20:34:43 +0000 (22:34 +0200)]
net: rtl8139: Read HW address from EEPROM only on probe
Do not re-read the HW address from the EEPROM on every start of
transfer, otherwise the user will not be able to adjust ethaddr
as needed. Read the address only once, when the card is detected.
Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com> Cc: Joe Hershberger <joe.hershberger@ni.com>
Marek Vasut [Sat, 9 May 2020 20:34:41 +0000 (22:34 +0200)]
net: rtl8139: Split out common and non-DM functions
Split the driver into common and non-DM functionality, so that the
DM support can later re-use the common code, while we retain the
non-DM code until all the platforms are converted.
Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com> Cc: Joe Hershberger <joe.hershberger@ni.com>
Marek Vasut [Sat, 9 May 2020 20:34:40 +0000 (22:34 +0200)]
net: rtl8139: Pass private data into rtl8139_eeprom_delay()
Instead of always calling rtl8139_eeprom_delay() with priv->ioaddr,
call it with priv and let the function access priv->ioaddr. This
reduces code duplication and has no impact, since the compiler will
inline this as needed anyway.
Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com> Cc: Joe Hershberger <joe.hershberger@ni.com>
Marek Vasut [Sat, 9 May 2020 20:34:39 +0000 (22:34 +0200)]
net: rtl8139: Introduce device private data
Introduce rtl8139_pdata, which is a super-structure around eth_device
and tracks per-device state, here the device IO address, PCI BDF, RX
and TX ring position. Pass this structure around instead of the old
non-DM eth_device in preparation for DM conversion.
Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com> Cc: Joe Hershberger <joe.hershberger@ni.com>
Marek Vasut [Sat, 9 May 2020 20:34:38 +0000 (22:34 +0200)]
net: rtl8139: Clean up bus_to_phys()/phys_to_bus() macros
These macros depended on the dev variable being declared wherever
they were used. This is wrong and will not work with DM anyway, so
pass only the PCI BFD into these macros, which fixes the dependency
and prepares them for DM support as well.
Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com> Cc: Joe Hershberger <joe.hershberger@ni.com>
Ravik Hasija [Thu, 7 May 2020 21:55:32 +0000 (14:55 -0700)]
net: tftp: fix progress marker for file transfer
During packet sequence number wraparound the show_block_marker() API was
not called, as a result the progress marker doesn't stay within column
boundary. Use position in file instead of sequence number to align the
marker.
Signed-off-by: Ravik Hasija <rahasij@linux.microsoft.com> Reviewed-By: Ramon Fried <rfried.dev@gmail.com>
Dan Murphy [Mon, 4 May 2020 21:14:39 +0000 (16:14 -0500)]
net: phy: Add support for TI PHY init
ti_phy_init function was allocated to the DP83867 PHY. This function
name is to generic for a specific PHY. The function can be moved to a
TI specific file that can register all TI PHYs that are defined in the
defconfig. The ti_phy_init file will contain all TI PHYs initialization
so that only phy_ti_init can be called from the framework.
In addition to the above the config flag for the DP83867 needs to be changed
in the Kconfig and dependent defconfig files. The config flag that was
used for the DP83867 was also generic in nature so a more specific
config flag for the DP83867 was created.
Acked-by: Michal Simek <michal.simek@xilinx.com> Signed-off-by: Dan Murphy <dmurphy@ti.com>
net: phy: add phyid search in vendor specific space
There are devices accesible through mdio clause-45, such as
retimers, that do not have PMA or PCS blocks.
This patch adds MDIO_MMD_VEND1 on the list of device addresses
where phyid is searched. Previous order of devices was kept.
Tom Warren [Thu, 26 Mar 2020 22:59:13 +0000 (15:59 -0700)]
net: rt8169: WAR for DHCP not getting IP after kernel boot/reboot
This is a WAR for DHCP failure after rebooting from the L4T kernel. The
r8169.c kernel driver is setting bit 19 of the rt816x HW register 0xF0,
which goes by FuncEvent and MISC in various driver source/datasheets.
That bit is called RxDv_Gated_En in the r8169.c kernel driver. Clear it
here at the end of probe to ensure that U-Boot can get an IP assigned
via DHCP.
Marek Vasut [Sat, 14 Mar 2020 23:18:57 +0000 (00:18 +0100)]
net: smc911x: Drop the standalone EEPROM example
Drop the example, for two reasons. First, it is tapping directly into
the IO accessors of the SMC911x, while it should instead go through
the net device API. Second, this makes conversion of the SMC911x driver
to DM real hard.
Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com> Cc: Joe Hershberger <joe.hershberger@ni.com> Cc: Tom Rini <trini@konsulko.com> Acked-by: Joe Hershberger <joe.hershberger@ni.com>
This keeps the compatibility with the old bootcmd.
The fdtfile environment variable also needed to be set to
imx6q-tbs2910.dtb to enable booting mainline kernels
otherwise with extlinux.conf it tries to load
mx6-tbs2910.dtb instead.
With arm-linux-gnueabi-gcc 9.2.0-1 from the Parabola
GNU/Linux distribution, we have the following size
differences:
- text: +2041 bytes
- data: 0 bytes
- bss: 0 bytes
- total: +2041 bytes
board: tbs2910_defconfig: Add support for booting general purpose distributions
This is needed to enable distro_boot support later on.
Adding the necessary configuration to enable booting general
purpose distributions is typically done by enabling
CONFIG_DISTRO_DEFAULTS.
However the resulting image size is is already very close to
the size limit and PXE isn't strictly needed, so instead we
made sure that all the configuration selected by
CONFIG_DISTRO_DEFAULTS were selected, with the exception of
PXE related configuration (CONFIG_BOOTP_PXE and
CONFIG_CMD_PXE) that are both disabled.
With arm-linux-gnueabi-gcc 9.3.0-1 from the Parabola
GNU/Linux distribution, we have the following size
differences:
- text: +7526 bytes
- data: +28 bytes
- bss: -12 bytes
- total: +7542 bytes
board: tbs2910: move CONFIG_BOOTCOMMAND from header to defconfig
This doesn't affect the size of the image: with
arm-linux-gnueabi-gcc 9.2.0-1 from the Parabola GNU/Linux
distribution, the text, data, bss and total sizes remain
unchanged.
As that the resulting image is already very close to the
size limit, and that CONFIG_GZIP is not strictly required,
removing it shouldn't hurt.
With arm-linux-gnueabi-gcc 9.2.0-1 from the Parabola
GNU/Linux distribution we have the following size
reduction:
- text: 9752
- data: 0
- bss: 16
- total: 9768
The loadb and loads commands are not needed for booting.
There are also more reliable and faster alternatives to
loadb and loads that can be used with the current configuration.
As that the resulting image is already very close to the
size limit, removing the loadb and loads commands shouldn't
hurt.
With arm-linux-gnueabi-gcc 9.2.0-1 from the Parabola
GNU/Linux distribution we have the following size
reduction:
- text: 6733 bytes
- data: 116 bytes
- bss: 1172 bytes
- total: 8021 bytes
The current recommendation for best security practice from the US government
is to use SHA384 for TOP SECRET [1].
This patch adds support for SHA384 and SHA512 in the hash command, and also
allows FIT images to be hashed with these algorithms, and signed with
sha384,rsaXXXX and sha512,rsaXXXX
The SHA implementation is adapted from the linux kernel implementation.
[1] Commercial National Security Algorithm Suite
http://www.iad.gov/iad/programs/iad-initiatives/cnsa-suite.cfm
Michael Walle [Mon, 1 Jun 2020 23:47:09 +0000 (01:47 +0200)]
dm: uclass: don't assign aliased seq numbers
If there are aliases for an uclass, set the base for the "dynamically"
allocated numbers next to the highest alias.
Please note, that this might lead to holes in the sequences, depending
on the device tree. For example if there is only an alias "ethernet1",
the next device seq number would be 2.
In particular this fixes a problem with boards which are using ethernet
aliases but also might have network add-in cards like the E1000. If the
board is started with the add-in card and depending on the order of the
drivers, the E1000 might occupy the first ethernet device and mess up
all the hardware addresses, because the devices are now shifted by one.
Also adapt the test cases to the new handling and add test cases
checking the holes in the seq numbers.
Signed-off-by: Michael Walle <michael@walle.cc> Reviewed-by: Alex Marginean <alexandru.marginean@nxp.com> Tested-by: Alex Marginean <alexandru.marginean@nxp.com> Acked-by: Vladimir Oltean <olteanv@gmail.com> Reviewed-by: Simon Glass <sjg@chromium.org> Tested-by: Michal Simek <michal.simek@xilinx.com> [on zcu102-revA]
Michael Walle [Mon, 1 Jun 2020 23:47:08 +0000 (01:47 +0200)]
dm: core: fix dev_read_alias_highest_id() without libfdt
If CONFIG_DM_DEV_READ_INLINE is set, dev_read_alias_highest_id() calls
libfdt_get_highest_id(). But this function is only available if we have
libfdt compiled in. If its not available return -1, which matches the
return code for no alias found.
This fixes the following error on omapl138_lcdk:
arm-linux-gnueabi-ld.bfd: drivers/built-in.o: in function `dev_read_alias_highest_id':
/home/mw/repo/u-boot/include/dm/read.h:986: undefined reference to `fdtdec_get_alias_highest_id'
Signed-off-by: Michael Walle <michael@walle.cc> Reviewed-by: Simon Glass <sjg@chromium.org>
Michael Walle [Mon, 1 Jun 2020 23:47:07 +0000 (01:47 +0200)]
usb: provide a device tree node to USB devices
It is possible to specify a device tree node for an USB device. This is
useful if you have a static USB setup and want to use aliases which
point to these nodes, like on the Raspberry Pi.
The nodes are matched against their hub port number, the compatible
strings are not matched for now.
Signed-off-by: Michael Walle <michael@walle.cc> Reviewed-by: Marek Vasut <marex@denx.de> Reviewed-by: Simon Glass <sjg@chromium.org>
Patrick Delaunay [Wed, 27 May 2020 07:43:31 +0000 (09:43 +0200)]
doc: log: correct option name CONFIG_LOG_MAX_LEVEL
Replace CONFIG_(SPL_)MAX_LOG_LEVEL by the correct name as defined in
common/Kconfig:
line 668:config LOG_MAX_LEVEL
line 688:config SPL_LOG_MAX_LEVEL
line 708:config TPL_LOG_MAX_LEVEL
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Pratyush Yadav [Tue, 26 May 2020 12:05:57 +0000 (17:35 +0530)]
regmap: Check for out-of-range offsets before mapping them
In regmap_raw_{read,write}_range(), offsets are checked to make sure
they aren't out of range. But this check happens _after_ the address is
mapped from physical memory. Input should be sanity-checked before using
it. Mapping the address before validating it leaves the door open to
passing an invalid address to map_physmem(). So check for out of range
offsets _before_ mapping them.
This fixes a segmentation fault in sandbox when -1 is used as an offset
to regmap_{read,write}().
Marcus Comstedt [Fri, 22 May 2020 16:41:36 +0000 (18:41 +0200)]
libfdt: Remove superfluous NUL character from overlay symbols
The symbol path already ends with a NUL character (something which is
actually checked by the code) and this NUL is included in
rel_path_len, so there is no need to add a second one.
This change fixes incorrect display in "fdt list /__symbols" after
applying an overlay with symbols.
Signed-off-by: Marcus Comstedt <marcus@mc.pp.se> Reviewed-by: Simon Glass <sjg@chromium.org>
Ovidiu Panait [Fri, 15 May 2020 06:30:12 +0000 (09:30 +0300)]
buildman: Remove _of_#_ from results directory paths
Currently, the following scenario will rebuild the first commit even
though it is not really necessary - the commit sha or the position in the
patchset did not change:
$ git am <local-patch-0001>
$ tools/buildman/buildman -P -E -W -b master mx6
<do some more development work>
$ git am <local-patch-0002>
$ tools/buildman/buildman -P -E -W -b master mx6 <- will rebuild the first
commit as well, even
though nothing has
changed about it.
This is due to the fact that previous results directories get removed
when the number of commits change. By removing the _of_#_ part of the
directory path, the commits will be rebuilt only if the commit sha or the
position in the patchset changes. Also, update the testcase to reflect this
change.
Simon Glass [Wed, 6 May 2020 22:29:06 +0000 (16:29 -0600)]
patman: Don't try to process checkpatch lines twice
Once we have determined what the line refers to there is no point in
processing it further. Update the logic to continue to the next line in
these cases.
Simon Glass [Wed, 6 May 2020 22:29:05 +0000 (16:29 -0600)]
patman: Support emacs mode with checkpatch
If checkpatch is run in 'emacs' mode it shows the filename at the
start of each line. Add support for this so that the warnings and errors
are correctly detected.
1. Removing the image_offset will break secondary (redundant) boot support for sd and emmc.
2. When booting from emmc boot partition, the image_offset is 0. But the flash.bin
generated by mkimage with imximage-8mp-lpddr4.cfg is for sd. It expects to be burn at 32KB offset.
The fit offset 0x60000 has already included the 32KB offset. So when you burn this flash.bin
to emmc boot partition at offset 0, the fit offset should subtract the 32KB (0x60000 - 0x8000).
Marek Vasut [Sat, 30 May 2020 20:44:46 +0000 (22:44 +0200)]
ARM: dts: imx: m53menlo: Convert to DM_ETH
Convert the board to DM_ETH instead of legacy networking. This requires
a minor addition to the DT to satisfy the requirement for specifying a
PHY node. No functional change from board user perspective.
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>
Marek Vasut [Sat, 30 May 2020 20:44:45 +0000 (22:44 +0200)]
ARM: imx: m53menlo: Do not fail boot on invalid splash screen
None of these splash screen loading errors are so critical as to
justify complete failure to boot, so just print error message as
needed and return 0, the boot can very likely continue without
the splash.
Fix a couple of missing free(dst) instances as well.
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>
Marek Vasut [Sat, 30 May 2020 00:14:48 +0000 (02:14 +0200)]
ARM: imx: ddr: Add missing PHY reset
The iMX7D RM 9.2.4.9.3 Power removal flow Table 9-11. Re-enabling power
explicitly says both the DDR controller and the PHY must be reset in the
correct sequence. Currently the code only resets the controller. This
leads to a misbehavior where the system brings the DRAM up after reboot,
but the DRAM is unstable. Add the missing reset.
The easiest way to trigger this is by triggering WDT without having the
WDT assert WDOG_B signal, i.e. mw.w 0x30280000 0x25 .
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>
Peng Fan [Tue, 26 May 2020 23:33:50 +0000 (20:33 -0300)]
imx: imx8mp_evk: fix boot issue
The u-boot-spl.bin pad with ddr firmware conflicts with the
CONFIG_MALLOC_F_ADDR area, the ddr firmware will be overwritten
by malloc in SPL stage and cause ddr initialization not able
to finish. So update the related addresses to fix the issue.
When booting from SD card or eMMC user partition, image_offset is
0x8000. It is useless to add and remove 0x8000.
When booting from other device, image_offset is 0 so this computation is wrong.
Simplfy this computation to work on all booting devices.
Fabio Estevam [Mon, 11 May 2020 14:00:27 +0000 (11:00 -0300)]
imx8mm_evk: Select the watchdog driver
Currently the watchdog driver is not selected, which causes the following
warnings in both SPL and U-Boot proper:
U-Boot SPL 2020.07-rc1-00387-g67887903af (May 07 2020 - 23:49:27 -0300)
Normal Boot
WDT: Started without servicing (60s timeout)
Trying to boot from MMC1
Patrick Wildt [Fri, 8 May 2020 09:59:47 +0000 (11:59 +0200)]
imx: move ATF to the back of the FIT to fix loading over yModem
With yModem the FIT Image is only supplied once, so we can only
seek forward in the yModem supplied image and never backwards.
With the recent changes to the SPL mechanism, including loading
U-Boot first, FDT after, then the loadables, we must also reorder
the FIT image script to make sure that the loadables are last in
the FIT image.
Signed-off-by: Patrick Wildt <patrick@blueri.se> Tested-by: Sébastien Szymanski <sebastien.szymanski@armadeus.com>
Heiko Stuebner [Fri, 5 Jun 2020 10:06:43 +0000 (12:06 +0200)]
rockchip: puma: enable new usb config options
With recently added changes we get support for usb3 including handling
of the phys (type-c and inno-usb2), so enable the necessary config
options on puma.
Signed-off-by: Heiko Stuebner <heiko.stuebner@theobroma-systems.com> Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Heiko Stuebner [Fri, 5 Jun 2020 10:06:42 +0000 (12:06 +0200)]
rockchip: puma: drop special handling of usb host regulator
With the current usb stack in u-boot, all host ports on puma work
flawlessly without any additional special handling, so drop that
usb hub hacking from the puma board.
Tested with mass-storage and usb-ethernet on both usb3 and usb2 ports.
Signed-off-by: Heiko Stuebner <heiko.stuebner@theobroma-systems.com> Reviewed-by: Kever Yang <kever.yang@rock-chips.com> Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Heiko Stuebner [Fri, 5 Jun 2020 10:06:41 +0000 (12:06 +0200)]
rockchip: puma: remove separate fit generator
The introduction of the puma-specific generator was mainly a way
to split the pmu firmware from the ATF binary and not having to
distribute that 4GB (sparse) image that was created before moving
to the bl31.elf as base.
Looking at the publically available repository for that separate
pmu firmware
https://git.theobroma-systems.com/rk3399-cortex-m0.git/
there is also no activity for 3 years and apart from some build
customizations no other changes were done.
And even then, if changes need to be made, this can very well also
happen in the atf context itself, so there is no real need to
diverge from the established build procedure and we can just go
back to using the main make_fit_atf.py script.
Signed-off-by: Heiko Stuebner <heiko.stuebner@theobroma-systems.com> Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Heiko Stuebner [Fri, 5 Jun 2020 10:06:40 +0000 (12:06 +0200)]
rockchip: puma: reorganize devicetrees to actually work and match upstream
So far the puma dts files only just included the main puma dtsi without
handling the actual baseboard and rk3399-puma.dtsi was very much
detached from the variant in the mainline Linux kernel.
Recent changes resulted in a strange situation with nonworking puma boards.
Commit ab800e5a6f28 ("arm: dts: rockchip: puma: move U-Boot specific bits to u-boot.dtsi")
moved the sdram include from rk3399-puma-ddrX.dts to new files
rk3399-puma-ddrx-u-boot.dtsi which were never included anywhere though.
Commit 167efc2c7a46 ("arm64: dts: rk3399: Sync v5.7-rc1 from Linux")
replaced the rk3399-puma.dtsi nearly completely, but in the kernel
it definitly depends on a baseboard dts to actually enable peripherals
like sd-slot, uarts, etc.
So to untagle this and bring the whole thing more in line with mainline
Linux, bring the rk3399-puma-haikou.dts over as well, drop the separate
DDR-option devicetrees and instead replace them with a puma Kconfig option
to select and include the needed DDR variant.
Signed-off-by: Heiko Stuebner <heiko.stuebner@theobroma-systems.com> Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Heiko Stuebner [Fri, 5 Jun 2020 10:06:39 +0000 (12:06 +0200)]
rockchip: puma: fix indentation of misc_init_r
The commit moving puma to the generic cpuid/macaddr helpers used 7 spaces
as indentation, so correct that by moving to the required tabs.
Fixes: fa177ff0208b ("board: puma: Use rockchip_* helpers to setup cpuid and macaddr") Signed-off-by: Heiko Stuebner <heiko.stuebner@theobroma-systems.com> Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
projects / u-boot.git / log