The current bootflow-iteration algorithm reads the bootflow file into
an allocated memory buffer so it can be examined. This works well in
most cases.
However, while the common case is that the first bootflow is immediately
booted, it is also possible just to scan for available bootflows, perhaps
selecting one to boot later.
Even with the common case, EFI bootflows can be quite large. It doesn't
make sense to read it into an allocated buffer when we have kernel_addr_t
providing a suitable address for it. Even if we do have plenty of malloc()
space available, it is a violation of U-Boot's lazy-init principle to
read the bootflow before it is needed.
So overall it seems better to make a change.
Adjust the logic to read just the size of the EFI file at first. Later,
when the bootflow is booted, read the rest of the file into the designated
kernel buffer.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reported-by: Da Xue <da@libre.computer>
Reported-by: Vincent Stehlé <vincent.stehle@arm.com>
This checks for a network-based bootflow in two places, one of which is
less than ideal. Move the correct test into a function and use it in both
places.
Signed-off-by: Simon Glass <sjg@chromium.org>
If a driver cannot be bound, provide the driver name in the debug
message. Now the debug message may look like this:
(virtio-pci.l#0): virtio-rng driver not configured
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
The FAT file systems uses character '\xe5' to mark a deleted directory
entry. If a file name starts with this character, it is substituted by
'\x05' in the directory entry.
While (signed char)'\xe5' is a negative number 0xe5 is a positive integer
number. We therefore have define a constant DELETED_MARK which matches the
signedness of the characters in the directory entry.
Correct a comparison where we used the constant 0xe5 with the wrong sign.
Use the constant aRING instead of 0x05 like in the rest of the code.
Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Fixes: 57b745e238 ("fs: fat: call set_name() only once")
Fixes: 28cef9ca2e ("fs: fat: create correct short names")
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
As the defconfig files here have been removed we can also remove the
entries.
Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Given that we no longer have a configs/vexpress_aemv8a_defconfig file,
drop that and then include at least the aarch64-specific config.h file
here. Also move Linus and Peter up to the main entry as well so that
they'll get tagged for the board code too and not literally only the
defconfig.
Signed-off-by: Tom Rini <trini@konsulko.com>
Rajesh's email bounces. Remove his email from all boards he maintains.
Fortunately, he has co-maintainers on most boards. I have taken the
liberty of volunteering Pramod to maintain the LS1012AFRDM, since he
also maintains the LS1012AFRWY. Let me know if the board should be
orphaned instead.
Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Merge in a series for MediaTek update and another for Ten64.
To quote Weijie Gao for MediaTek:
This patch series add support for MediaTek MT7988 SoC with its reference
boards and related drivers.
This patch series add basic boot support on eMMC/SD/SPI-NOR/SPI-NAND for
these boards. The clock, pinctrl drivers and the SoC initializaton code
are also included.
Product spec for MT7988:
https://www.mediatek.com/products/broadband-wifi/mediatek-filogic-880
And to quote Mathew McBride for Ten64:
This is a series of updates for the Ten64 board,
that are part of our firmware releases but not yet upstreamed
into U-Boot.
Changes of note include:
- Turning on standard boot support
Standard boot improves the user experience over distroboot on Ten64,
as we had various hacks in our firmware to solve some corner-case
issues (e.g DTB handling) in distroboot, which are not needed with the
bootflow system.
- Recognition of the new 'RevD' board variant distributed to OEM
customers
- Fixing various boot issues related to FIT images and operating systems
running out of the NAND (OpenWrt, recovery environment).
- A better 'opt-out' solution for fsl_setenv_bootcmd for Layerscape
platforms booting from TF-A.
This was discussed when the Ten64 was upstreamed into U-Boot. I think
declaring fsl_setenv_bootcmd as __weak and allowing individual boards
to override is the best way to do this without significant rework.
(We actually depend on a similar feature for the DPAA2/MC firmware
loading)
Compared to our firmware branch, there is still a few features missing (e.g USB
Hub, fan controller and fixes for the VSC8514). Some of these depend on other
things (like sorting out device tree schemas) so may not appear in mainline
U-Boot for a while yet.
The u-boot version of the LS1088A device tree has
an extra compatible (simple-mfd) added to &fsl_mc
to facilitate usage with U-Boot's device model.
Unfortunately FreeBSD will only match the single
"fsl,qoriq-mc" exactly when the node is a "bus"
object, so we need to strip out the extra compatible
before presenting it to the operating system.
Signed-off-by: Mathew McBride <matt@traverse.com.au>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Our bootcmd is the same regardless of where the SoC
loaded it's code from, so we don't want
fsl_setenv_bootcmd to do anything.
Signed-off-by: Mathew McBride <matt@traverse.com.au>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Allow individual Layerscape boards to opt-out of fsl_setenv_bootcmd
by declaring the original function as weak.
fsl_setenv_bootcmd is used to change the bootcmd based on the
TF-A boot source (e.g QSPI vs SD/MMC) for reasons including
secure boot / integrity measurements and DPAA2 configuration loading.
See previous discussion at [1].
On the Ten64 board, our bootcmd is the same across
all TF-A boot sources so we don't want this behaviour.
Signed-off-by: Mathew McBride <matt@traverse.com.au>
[1] https://patchwork.ozlabs.org/project/uboot/patch/20211110044639.7070-3-matt@traverse.com.au/#2790037
Reviewed-by: Peng Fan <peng.fan@nxp.com>
With the previous updates to the device tree, Ten64
can use Standard Boot 'out of the box'.
Signed-off-by: Mathew McBride <matt@traverse.com.au>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
The watchdog driver was previously enabled but not used
until U-Boot's fsl-ls1088a.dtsi was updated to describe them.
Some Linux distributions (e.g Debian 11) do not engage the
SP805 watchdogs, causing unexpected resets after boot.
To conserve the user experience, turn off the autostart,
and we will provide a mechanism to turn them on at boot
via env vars.
Signed-off-by: Mathew McBride <matt@traverse.com.au>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
The GE0 (first Gigabit Ethernet interface) is used as the
'serial number' for the board and appliance.
To ensure the 'true' board S/N is available regardless of how
the DPAA2 subsystem is configured, use serial# so it is passed in
the device tree.
Signed-off-by: Mathew McBride <matt@traverse.com.au>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
On Ten64 boards, the "serial number" is the MAC address of the
first Gigabit Ethernet interface (labelled GE0 on the appliance),
and counted up from there.
The previous logic did not take into account U-Boot's ordering
of the network interfaces. By setting aliases/ethernetX in the device
tree we can ensure the U-Boot 'ethX' is the same as the labelled
port order on the unit, as well as the one adopted by Linux.
Signed-off-by: Mathew McBride <matt@traverse.com.au>
Ensure nvme devices are scanned before reaching the shell,
otherwise extra user intervention ("nvme scan") is required
before they are visible to bootdev/bootflow.
Signed-off-by: Mathew McBride <matt@traverse.com.au>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
This restores NVMe functionality after PCI(e) NVMe support
was split out from the NVMe driver.
Signed-off-by: Mathew McBride <matt@traverse.com.au>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
The recovery-firmware and OpenWrt-NAND do not yet have bootflow
/bootstd entrypoints, so add bootmenu entries to make them
accessible.
Signed-off-by: Mathew McBride <matt@traverse.com.au>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
The default Ten64 MTD configuration reserves two ubifs partitions
for OpenWrt residing on NAND flash. Add the bootcmd for this system
into the default environment.
Signed-off-by: Mathew McBride <matt@traverse.com.au>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
The recovery environment[1] on the Ten64 is a OpenWrt-
based ramdisk stored on the NAND intended to help with
system setup tasks.
Before the bootargs were not being set for the recovery
command, relying instead on the existing bootargs variable.
Ensure the bootargs are set correctly prior to booting recovery.
Signed-off-by: Mathew McBride <matt@traverse.com.au>
[1] https://ten64doc.traverse.com.au/software/recovery/
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Change the firmware on microSD path to "firmware/traverse/ten64"
as per EBBR section 4.2[1].
The Traverse firmware tools now locate the DPAA2 firmware
and configuration files under that path on the rescue
SD card image.
If a user then installs a standard Linux
distribution over the top of that sdcard, (in theory)
it will be left alone by distribution boot tooling.
Signed-off-by: Mathew McBride <matt@traverse.com.au>
[1] https://arm-software.github.io/ebbr/index.html#firmware-partition-filesystem
Reviewed-by: Peng Fan <peng.fan@nxp.com>
The DPAA2 DPL (data plane layout) file was previously
being loaded into 0x80300000, and set to be applied
just before hand off to the kernel.
When a FIT image with a load_address of 0x80000000 was
booted with bootm, the DPL in memory was overwritten.
Move the DPL load to 0x8E000000 (196MiB away from 0x80000000,
and below the other typical load addr of 0x90000000).
Ideally in the future, the DPL lazyapply command
("fsl_mc lazyapply DPL $dpl_addr") should be set to
load the DPL contents into a memory area owned by U-Boot.
Signed-off-by: Mathew McBride <matt@traverse.com.au>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Board revision C (production) and later require the SFP+
retimer to be turned on (or reset) on boot, by way of issuing
a command to the board's microcontroller (via I2C).
The comparison statement here was incorrect, as the board
ID decrements every revision (from 0xFF downwards),
so this was matching board RevA,B,C instead of Rev >= C.
Another oops that transpired when working on this issue,
is that if the board controller is not called (such as
CONFIG_TEN64_CONTROLLER=n or earlier board rev), then
the retimer udevice was not obtained. So the board
version check has to be moved inside board_cycle_retimer
(which probes/fetches the retimer device) as well.
Signed-off-by: Mathew McBride <matt@traverse.com.au>
Ten64 board revision D is a variant that removes the USB hub
and PCIe expander/switch, but is otherwise compatible with the
main production "C" version.
At the same time, revise the printf specifiers (PCB version
"1064-0201%s") to reduce the number of string characters related
to the boot printout.
Signed-off-by: Mathew McBride <matt@traverse.com.au>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
This patch adds general board files based on MT7988 SoCs.
MT7988 uses one mmc controller for booting from both SD and eMMC,
and the pins of mmc controller booting from SD are also shared with
one of spi controllers.
So two configs are need for these boot types:
1. mt7988_rfb_defconfig - SPI-NOR, SPI-NAND and eMMC
2. mt7988_sd_rfb_defconfig - SPI-NAND and SD
Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
This patch adds basic support for MediaTek MT7988 SoC.
This includes files that will initialize the SoC after boot and
its device tree.
Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
This patch converts magic and version fields of ghf common header
to one field with the type of uint32_t to make this header flexible
for futher updates.
Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
This patch adds support for MediaTek MT7988.
MT7988 features MediaTek NETSYS v3, including three GMACs, and two
of them supports 10Gbps USXGMII.
MT7988 embeds a MT7531 switch (not MCM) which supports accessing
internal registers through MMIO instead of MDIO.
Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
MT7981 has its GMAC2 PHY shared with USB3. To enable GMAC2, mux
register must be set to connect the SGMII phy to GMAC2.
Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
Now that individual 2.5Gbps SGMII support has been added to
mtk-eth, all boards that use 2.5Gbps link with mt7531 must be
converted to use "2500base-x" instead of "sgmii".
Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
Existing SGMII support of mtk-eth is actually a MediaTek-specific
2.5Gbps high-speed SGMII (HSGMII) which does not support
auto-negotiation mode.
This patch adds SGMII 1Gbps auto-negotiation mode and rename the
existing HSGMII to 2500basex.
Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
mt7531_mmd_ind_read and mt753x_switch_init are defined without static.
Since they're not used outside this file, we should add them back.
Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
fixup to add static qualifier
The original direct MDIO clause 45 access via SoC is missing the
data output. This patch adds it back to ensure MDIO clause 45 can
work properly for external PHYs.
Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
Not all switches requires 1 second delay after deasserting reset.
MT7531 requires only maximum 200ms.
This patch defines dedicated reset wait time for each switch chip, and will
significantly improve the boot time for boards using MT7531.
Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
So far the switch is initialized in probe stage and is connected to PSE
unconditionally. This will cause all packets being flooded to PSE and may
cause PSE hang before entering linux.
This patch changes the connection between switch and PSE:
- Still initialize switch in probe stage, but disconnect it with PSE
- Connect switch with PSE on eth start
- Disconnect on eth stop
Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
The original mediatek pinctrl functions for driving configuration
'mtk_pinconf_drive_set_*' do not return -ENOSUPP even if input
parameters are not supported.
This patch fixes the return value in those functions.
Signed-off-by: Sam Shih <sam.shih@mediatek.com>
Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
There exists a situation of the mediatek pinctrl driver that may return
wrong pin function value for the pinmux driver:
- All pin function arrays are defined without const
- Some pin function arrays contain all-zero value, e.g.:
static int mt7622_spi_funcs[] = { 0, 0, 0, 0, 0, 0, };
- These arrays will be put into .bss section during compilation
- .bss section has no "a" attribute and does not exist in the final binary
file after objcopy.
- FDT binary blob is appended to the u-boot binary, which occupies the
.bss section.
- During board_f stage, .bss has not been initialized, and contains the
data of FDT, which is not full-zero data.
- pinctrl driver is initialized in board_f stage, and it will get wrong
data if another driver is going to set default pinctrl.
Since pinmux information and soc data are only meant to be read-only, thus
should be declared as const. This will force all pinctrl data being put
into .rodata section. Since .rodata has "a" attribute, even the all-zero
data will be allocated and filled with correct value in to u-boot binary.
Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
