The following error has been observed on i.MX25 with a high-speed SDSC
card:
Data Write Failed in PIO Mode.
It was caused by the timeout set on PRSSTAT.BWEN, which was triggered
because this bit takes 15 ms to be set after writing the first block to
DATPORT with this card. Without this timeout, all the blocks are
properly written.
This timeout was implemented by decrementing a variable, so it was
depending on the CPU frequency. Fix this issue by setting this timeout
to a long enough absolute duration (500 ms).
Fabio Estevam [Fri, 3 Nov 2017 15:40:10 +0000 (13:40 -0200)]
mx51: Select the ESDHC_A001 erratum
When a high speed card is connected to mx51evk the following error is seen:
U-Boot 2017.11-rc2 (Oct 18 2017 - 13:49:26 -0200)
CPU: Freescale i.MX51 rev3.0 at 800 MHz
Reset cause: POR
Board: MX51EVK
DRAM: 512 MiB
MMC: FSL_SDHC: 0, FSL_SDHC: 1
*** Warning - read failed, using default environment
In: serial
Out: serial
Err: serial
Net: FEC
Hit any key to stop autoboot: 0
=> saveenv
Saving Environment to MMC...
Writing to MMC(0)... failed
The root cause for the failure is the eSDHC-A001 erratum:
"eSDHC-A001 : Data timeout counter (SYSCTL[DTOCV]) is not reliable for
values of 0x4,0x8, and 0xC" that is listed
on some PowerArchitecture chips:
https://www.nxp.com/files-static/32bit/doc/errata/MPC8379ECE.pdf
Even though eSDHC-A001 is not documented on the i.MX51 errata document,
I have confirmed with the NXP design team that this erratum does affect
i.MX51, so fix the problem by selecting SYS_FSL_ERRATUM_ESDHC_A001
at SoC level.
The i.MX51 ts4800 board already selects this option, but it is better
to move this selection to the i.MX51 SoC level instead.
Successfully tested with a high speed SD card on a mx51evk board.
CPU: Freescale i.MX25 rev1.2 at 399 MHz
Reset cause: POR
Board: MX25PDK
I2C: ready
DRAM: 64 MiB
No arch specific invalidate_icache_all available!
MMC: FSL_SDHC: 0
*** Warning - read failed, using default environment
In: serial
Out: serial
Err: serial
Net: FEC
Hit any key to stop autoboot: 0
=> saveenv
Saving Environment to MMC...
Writing to MMC(0)... failed
, which prevents any usage of the SD card.
The root cause for the failure is the eSDHC-A001 erratum:
"eSDHC-A001 : Data timeout counter (SYSCTL[DTOCV]) is not reliable for
values of 0x4,0x8, and 0xC" that is listed
on some PowerArchitecture chips:
https://www.nxp.com/files-static/32bit/doc/errata/MPC8379ECE.pdf
Even though eSDHC-A001 is not documented on the i.MX25 errata document,
I have confirmed with the NXP design team that this erratum does affect
i.MX25, so fix the problem by selecting SYS_FSL_ERRATUM_ESDHC_A001
at SoC level.
Successfully tested with a high speed SD card on a mx25pdk board.
Fabio Estevam [Fri, 3 Nov 2017 15:40:08 +0000 (13:40 -0200)]
mx25: Move MX25 selection to Kconfig
The motivation for moving MX25 selection to Kconfig is to be
able to better handle MX25 specific errata, so that an errata option
can be selected at SoC level instead of board level.
This selection method also aligns with the way other i.MX SoCs are
selected in U-Boot.
Jan Kundrát [Fri, 3 Nov 2017 02:06:35 +0000 (03:06 +0100)]
Do not attempt to use the systemwide libfdt
U-Boot bundles a patched copy of libfdt, so it's wrong to attempt to
include it <like/this>. This breaks the build for me when I have dtc
fully installed in my host -- as happened earlier tonight with
Buildroot, for example.
There are several other occurrences throughout the code where '<libfdt'
matches. I'm not modifying these because I have no clue why the
<systemwide> include style is being used -- IMHO wrongly.
Signed-off-by: Jan Kundrát <jan.kundrat@cesnet.cz>
Shawn Guo [Thu, 2 Nov 2017 08:46:34 +0000 (16:46 +0800)]
disk: part_dos: fix part_get_info_extended() function
The check in part_get_info_extended() for a successful partition
searching misses a condition for extended partition. In case of
(ext_part_sector == 0), we should anyway mark the partition as found,
even if it's an extended partition, i.e. (is_extended(pt->sys_ind) == 0).
Otherwise, the extended partition (type 0x0f) will never be identified,
and the following recursive call to part_get_info_extended() will get a
wrong 'part_num' and 'which_part' parameter. In the end, all those
partitions in extended table will not be identified.
Let's add the missing OR condition of (ext_part_sector == 0) for
is_extended() check to fix the problem.
The issue is discovered by running fastboot flash to an extended
partition on eMMC.
Lukasz Majewski [Fri, 27 Oct 2017 10:28:10 +0000 (12:28 +0200)]
gpt: Use cache aligned buffers for gpt_h and gpt_e
Before this patch one could receive following errors when executing
"gpt write" command on machine with cache enabled:
display5 factory > gpt write mmc ${mmcdev} ${partitions}
Writing GPT:
CACHE: Misaligned operation at range [4ef8f7f0, 4ef8f9f0]
CACHE: Misaligned operation at range [4ef8f9f8, 4ef939f8]
CACHE: Misaligned operation at range [4ef8f9f8, 4ef939f8]
CACHE: Misaligned operation at range [4ef8f7f0, 4ef8f9f0]
success!
To alleviate this problem - the calloc()s have been replaced with
malloc_cache_aligned() and memset().
After those changes the buffers are properly aligned (with both start
address and size) to SoC cache line.
Masahiro Yamada [Fri, 27 Oct 2017 06:04:20 +0000 (15:04 +0900)]
tools: image: allow to sign image nodes without -K option
If -K option is missing when you sign image nodes, it fails with
an unclear error message:
tools/mkimage Can't add hashes to FIT blob: -1
It is hard to figure out the cause of the failure.
In contrast, when you sign configuration nodes, -K is optional because
fit_config_process_sig() returns successfully if keydest is unset.
Probably this is a preferred behavior when you want to update FIT with
the same key; you do not have to update the public key in this case.
So, this commit changes fit_image_process_sig() to continue signing
without keydest. If ->add_verify_data() fails, show a clearer error
message, which has been borrowed from fit_config_process_sig().
Masahiro Yamada [Fri, 27 Oct 2017 04:25:21 +0000 (13:25 +0900)]
tools: image: fix "algo" property of public key for verified boot
The "algo_name" points to a property in a blob being edited. The
pointer becomes stale when fit_image_write_sig() inserts signatures.
Then crypto->add_verify_data() writes wrong data to the public key
destination.
Stephen Warren [Fri, 27 Oct 2017 00:23:35 +0000 (18:23 -0600)]
test/py: regenerate persistent GPT image if code changes
test_gpt generates a persistent disk image which can be re-used across
multiple test runs. Currently, if the Python code that generates the disk
image change, the image is not regenerated, which could cause test
failures e.g. if a test was updated to expect some new partition name or
size, yet the persistent disk image contained the old name or size. This
change introduces functionality to regenerate the disk image if the
instructions to generate the image have changed.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Stefan Agner [Thu, 26 Oct 2017 12:57:53 +0000 (14:57 +0200)]
tools: env: allow to print U-Boot version
The fw_env utility family has a default environment compiled in
which ties it quite strongly to the U-Boot source/config it has
been built with. Allow to display the U-Boot version it has been
built with using the -v/--version argument.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com> Reviewed-by: Lukasz Majewski <lukma@denx.de>
Tom Rini [Sat, 21 Oct 2017 16:54:02 +0000 (12:54 -0400)]
MAINTAINERS: Add missing boards and config entries
As part of my usual round of build testing, output about missing
MAINTAINERS information was not logged, and thus often overlooked.
Correct that mistake by ensuring that I log the output of
genboardscfg.py every time. As part of that, address a number of
missing MAINTAINERS entires. In the case of a missing file, I have put
the original submitter down. In the rest of the cases I have added the
config (and sometimes relevant header file) to the existing set of file
globs.
Dongjin Kim [Sun, 29 Oct 2017 16:39:59 +0000 (12:39 -0400)]
ARM: dts: exynos: fix property values of LDO15/17 for ODROID-XU3/4
Looking at the schematic, LDO15 and LDO17 are tied as a power source of a
builtin network chipset. The voltage on LDO15 is corrected to 3.3V and the
name of LDO17 is corrected to "vdd_ldo17".
Signed-off-by: Dongjin Kim <tobetter@gmail.com> Reviewed-by: Lukasz Majewski <lukma@denx.de> Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
We move all instances of CONFIG_MTDIDS_DEFAULT and
CONFIG_MTDPARTS_DEFAULT from the header files to the defconfig files.
There's a few cases here where we need to expand upon what was in the
header file.
Tested-by: Adam Ford <aford173@gmail.com> #omap3_logic Signed-off-by: Tom Rini <trini@konsulko.com>
Michal Simek [Thu, 2 Nov 2017 14:45:34 +0000 (15:45 +0100)]
scsi: ceva: Start port in probe
The patch:
"dm: ahci: Unwind the confusing init code"
(sha1: 7cf1afce7fa3fe64189020fe14b93f7326dd0758)
introduce bug for ceva sata because port didn't start.
On the other hand the dwc_ahci.c was fixed correctly.
Do the same change for ceva too.
Signed-off-by: Michal Simek <michal.simek@xilinx.com> Tested-by: Mian Yousaf Kaukab <yousaf.kaukab@suse.com>
The RK3368-uQ7 uses a KSZ9031 PHY on-module. Enable PHY_MICREL_KSZ90X1
in the associated defconfig.
References: da3b9e7f ("Move PHY_MICREL and PHY_MICREL_KSZ90X1 to Kconfig") Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
The RK3368-uQ7 uses a KSZ9031 PHY on-module. Enable PHY_MICREL_KSZ90X1
in the associated defconfig (this somehow got lost with da3b9e7f).
References: da3b9e7f ("Move PHY_MICREL and PHY_MICREL_KSZ90X1 to Kconfig") Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Kever Yang [Thu, 26 Oct 2017 06:45:02 +0000 (14:45 +0800)]
rockchip: evb-rk3328: remove CONFIG_ENV_OFFSET
Remove CONFIG_ENV_OFFSET for there already have one in rockchip_common.h
Signed-off-by: Kever Yang <kever.yang@rock-chips.com> Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Kever Yang [Wed, 25 Oct 2017 08:24:48 +0000 (16:24 +0800)]
rockchip: rk3328: fix rockchip_get_cru api
The API for get priv pointer is wrong, fix it.
Signed-off-by: Kever Yang <kever.yang@rock-chips.com> Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Kever Yang [Thu, 12 Oct 2017 07:27:29 +0000 (15:27 +0800)]
rockchip: rk3399: init CPU clock when rkclk_init()
Init the CPU and its buses to speed up the boot time.
Move rkclk_init() to a place after rk3399_configure_cpu has defined
at the same time, or else there will be a warning.
Signed-off-by: Kever Yang <kever.yang@rock-chips.com> Reviewed-by: Simon Glass <sjg@chromium.org> Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Klaus Goger [Fri, 6 Oct 2017 17:24:08 +0000 (19:24 +0200)]
rockchip: configs: only add available BOOT_TARGET_DEVICES
BOOT_TARGET_DEVICES should only be added if the corresponding u-boot
command is enabled.
Signed-off-by: Klaus Goger <klaus.goger@theobroma-systems.com> Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Klaus Goger [Fri, 6 Oct 2017 17:24:09 +0000 (19:24 +0200)]
rockchip: configs: use rockchip-common.h for rk3368
rockchip-common.h already defines values that are missing from
rk3368_common.h
For example BOOT_TARGET_DEVICES was defined empty and therefore
distroboot had no boot targets.
Signed-off-by: Klaus Goger <klaus.goger@theobroma-systems.com> Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com> Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
CPU: Freescale i.MX6SL rev1.0 792 MHz (running at 396 MHz)
CPU: Commercial temperature grade (0C to 95C) at 33C
Reset cause: POR
Board: MX6SLEVK
I2C: ready
DRAM: 1 GiB
MMC: FSL_SDHC: 0
MMC Device 1 not found
*** Warning - No MMC card found, using default environment
As mx6slevk_spl_defconfig does not use CONFIG_DM_MMC and its
board file does not register the mmc controller for U-Boot proper,
let's fix this by adding CONFIG_DM_MMC=y and device tree support.
While at it, add more DM drivers, so that it becomes closer to
mx6slevk_defconfig.
Stefan Agner [Tue, 17 Oct 2017 11:59:19 +0000 (13:59 +0200)]
configs: vf610: increase maximum size and enforce correct limit
On Vybrid SoCs U-Boot gets loaded into GFX SRAM which is 512KiB.
Currently 32KiB is reserved for the IMX header. However, this is
not reflected in the size limit. In v2017.11-rc2 the actual size
limit (512KiB-32KiB) has been reached for Colibri VF61, which
lead to a successful build of U-Boot but not a working binary.
The IMX header is much smaller than 32KiB, typically around 1KiB.
Decrease the reserved size to 4KiB and specify the correct U-Boot
size limit. Apply this new base address and limit for all Vybrid
based boards.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
Fabio Estevam [Sun, 15 Oct 2017 13:01:48 +0000 (11:01 -0200)]
net: fec_mxc: Change "error frame" message to debug level
As reported by Jonathan Gray:
"After the recent changes to add SimpleNetworkProtocol to efi_loader
when booting off mmc via an efi payload that doesn't use
SimpleNetworkProtocol U-Boot's fec_mxc driver will now display
various "error frame" messages.
....
MMC Device 1 not found
MMC Device 2 not found
MMC Device 3 not found
Scanning disks on sata...
Found 6 disks
reading efi/boot/bootarm.efi
67372 bytes read in 32 ms (2 MiB/s)
## Starting EFI application at 12000000 ...
>> OpenBSD/armv7 BOOTARM 1.0
error frame: 0x8f57ec40 0x00003d74
error frame: 0x8f57ec40 0x00007079
error frame: 0x8f57ec40 0x00006964
error frame: 0x8f57ec40 0x00006f6f
error frame: 0x8f57ec40 0x0000726f
error frame: 0x8f57ec40 0x00002074
error frame: 0x8f57ec40 0x00006f6f"
Heinrich Schuchardt explains:
"A receive FIFO overrun can be expected if network packages are not
processed.
With the network patches we check if a package is available quite often."
Move the "error frame" messages to debug level so that a clean output
log can be seen.
Reported-by: Jonathan Gray <jsg@jsg.id.au> Suggested-by: Heinrich Schuchardt <xypron.glpk@gmx.de> Signed-off-by: Fabio Estevam <fabio.estevam@nxp.com>
Alexander Graf [Thu, 19 Oct 2017 21:23:50 +0000 (23:23 +0200)]
efi_loader: Disable env_save() call on boot
With the introduction of EFI variable support, we also wanted to persist
these EFI variables. However, the way it was implemented we ended up
persisting all U-Boot environment variables on every EFI boot.
That could potentially lead to unexpected side effects because variables
that were not supposed to be written to persisted env get written. It also
means we may end up writing the environment more often than we should.
For this release, let's just disable EFI variable persistence and instead
implement it properly for the next one.
Reported-by: Heinrich Schuchardt <xypron.glpk@gmx.de> Fixes: ad644e7c182 ("efi_loader: efi variable support") Signed-off-by: Alexander Graf <agraf@suse.de> Acked-by: Rob Clark <robdclark@gmail.com> Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
armv8: sec_firmware: Add support for loadables in FIT
Enable support for loadables in SEC firmware FIT image. Currently
support is added for single loadable image.
Brief description of implementation:
Add two more address pointers (loadable_h, loadable_l) as arguments to
sec_firmware_init() api.
Create new api: sec_firmware_checks_copy_loadable() to check if loadables
node is present in SEC firmware FIT image. If present, verify loadable
image and copies it to secure DDR memory.
Populate address pointers with secure DDR memory addresses where loadable
is copied.
Example use-case could be trusted OS (tee.bin) as loadables node in SEC
firmware FIT image.
Signed-off-by: Sumit Garg <sumit.garg@nxp.com> Reviewed-by: York Sun <york.sun@nxp.com>
armv8: layerscape: Allocate 66 MB DDR for secure memory
Change DDR allocated for secure memory from 2 MB to 66 MB. This
additional 64 MB secure memory is required for trusted OS running
in Trusted Execution Environment using ARMv8 TrustZone.
Signed-off-by: Sumit Garg <sumit.garg@nxp.com> Reviewed-by: York Sun <york.sun@nxp.com>
Ran Wang [Mon, 23 Oct 2017 02:09:25 +0000 (10:09 +0800)]
armv8: ls1088aqds: Enable USB command on QDS for qspi-boot
Signed-off-by: Amrita Kumari <amrita.kumari@nxp.com> Signed-off-by: Ashish Kumar <Ashish.Kumar@nxp.com> Signed-off-by: Ran Wang <ran.wang_1@nxp.com> Reviewed-by: Bin Meng <bmeng.cn@gmail.com> Reviewed-by: York Sun <york.sun@nxp.com>
Ran Wang [Mon, 23 Oct 2017 02:09:24 +0000 (10:09 +0800)]
armv8: ls1088ardb: Enable USB command RDB qspi-boot
Signed-off-by: Ashish Kumar <Ashish.Kumar@nxp.com> Signed-off-by: Amrita Kumari <amrita.kumari@nxp.com> Signed-off-by: Ran Wang <ran.wang_1@nxp.com> Reviewed-by: Bin Meng <bmeng.cn@gmail.com> Reviewed-by: York Sun <york.sun@nxp.com>
Masahiro Yamada [Thu, 26 Oct 2017 12:24:17 +0000 (21:24 +0900)]
i2c: fti2c010: remove unused/unmaintained driver
CONFIG_SYS_I2C_FTI2C010 is not enabled by anyone.
Commit 2852709676c8 ("dm: i2c: Add a note to I2C drivers which need
conversion") prompted to convert this driver to DM before June 2017,
but not converted yet.
Bin Meng [Thu, 19 Oct 2017 01:20:59 +0000 (18:20 -0700)]
x86: fsp: graphics: Add some notes about the graphics info hob
On some platforms (eg: Braswell), the FSP will not produce the
graphics info HOB unless you plug some cables to the display
interface (eg: HDMI) on the board. Add such notes in the FSP
video driver.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Bin Meng [Thu, 19 Oct 2017 01:20:57 +0000 (18:20 -0700)]
spi: ich: Lock down controller settings if required
Some Intel FSP (like Braswell) does SPI lock-down during the call
to fsp_notify(INIT_PHASE_BOOT). But before SPI lock-down is done,
it's bootloader's responsibility to configure the SPI controller's
opcode registers properly otherwise SPI controller driver doesn't
know how to communicate with the SPI flash device.
Rather than passively doing the opcode configuration, let's add a
simple DTS property "intel,spi-lock-down" and let the driver call
the opcode configuration function if required by such FSP.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com> Reviewed-by: Simon Glass <sjg@chromium.org>
It's not appropriate to call the Intel SPI driver specific stuff in
the FSP codes. We may add a simple DTS property "intel,spi-lock-down"
and let the Intel SPI driver call these stuff instead.
Bin Meng [Thu, 19 Oct 2017 01:20:53 +0000 (18:20 -0700)]
x86: braswell: Fix unexpected crash during Linux kernel boot
It was observed that when booting Linux kernel on Intel Cherry Hill
board, unexpected crash happens quite randomly. Sometimes kernel
just oops, while sometimes kernel throws MCE errors and hangs:
This looks like a hardware error per mcelog. After debugging, it
seems turning off turbo mode on the processor does not expose this
behavior, although U-Boot runs OK with turbo mode on. Suspect it is
related to an errata of Braswell processor.
To fix this, remove the Braswell cpu driver which does the turbo
mode configuration, and switch to use the generic cpu-x86 driver.
Also there is a configuration option in the FSP that turns on the
turbo mode and that has been turned off too.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com> Reviewed-by: Simon Glass <sjg@chromium.org>
projects / u-boot.git / log