Having an EFI_RNG_PROTOCOL without a backing RNG device leads to failure
to boot Linux 5.8.
Only install the EFI_RNG_PROTOCOL if we have a RNG device.
Reported-by: Scott K Logan <logans@cottsay.net> Cc: Neil Armstrong <narmstrong@baylibre.com> Cc: Ard Biesheuvel <ardb@kernel.org> Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
arm64: a37xx: pci: Disable link training when unloading driver
As required by PCI Express spec a delay for at least 100ms after
de-asserting PERST# signal is needed before link training is enabled.
Linux kernels prior to 5.8 version do not automatically disable link
training before de-asserting PERST# signal, therefore this requirement is
not fulfilled.
Above requirement is needed for proper detection of some Compex PCIe WiFi
cards. Otherwise Linux kernel cannot detect it.
To allow using those PCIe cards with older Linux kernel versions booted by
U-Boot compiled with U-Boot a37xx pci driver, disable link training in
U-Boot when unloading this pci driver.
Thanks to DM_FLAG_OS_PREPARE flag, U-Boot automatically unload this driver
when booting Linux kernel.
Signed-off-by: Pali Rohár <pali@kernel.org> Reviewed-by: Stefan Roese <sr@denx.de>
Testing on a DNS-325 NAS has shown that in order for the device to work
we need to set CONFIG_NR_DRAM_BANKS to 1 and not 2.
Tested-by: Dmitry N. Kolesnikov <dk.diklab@gmail.com> Reported-by: Dmitry N. Kolesnikov <dk.diklab@gmail.com> Cc: Stefan Herbrechtsmeier <stefan@herbrechtsmeier.net> Signed-off-by: Tom Rini <trini@konsulko.com>
Andre Heider [Fri, 11 Sep 2020 04:35:10 +0000 (06:35 +0200)]
arm: mvebu: Espressobin: Set environment variable fdtfile
Required for the generic distro mechanism.
Linux ships with 4 variants:
marvell/armada-3720-espressobin-v7-emmc.dtb
marvell/armada-3720-espressobin-v7.dtb
marvell/armada-3720-espressobin-emmc.dtb
marvell/armada-3720-espressobin.dtb
Use available information to determine the appropriate filename.
Fixes booting GRUB EFI arm64 on Fedora.
Reported-by: Dennis Gilmore <dennis@ausil.us> Signed-off-by: Andre Heider <a.heider@gmail.com>
- fix SPL image generation
- fix incorrect console output
- increase malloc_f and malloc_r space to fix LZMA decompression errors
- increase SPI flash clock
Signed-off-by: Mauro Condarelli <mc5686@mclink.it>
[squashed to one patch, fix commit subject and description] Signed-off-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
The GPIO bank name for banks J and K are not correct when using the
'gpio' command from the console.
The driver derives the bank name from the device tree instance string by
using the instance value and adding 'A': gpio0@xxaddrxx is Bank A,
gpio1@yyaddryy is Bank B and so on.
On the PIC32, there is no Bank I so instances 8 and 9 need to be
incremented as a minimum change.
An alternative (less opaque) implementation would be to use a bank-name
property instead but this would require modifying the driver code too.
Signed-off-by: John Robertson <john.robertson@simiatec.com>
GPIO state cannot be changed via the device tree (e.g. with gpio-hog) or
using the 'gpio' command from the console.
The root cause is a discrepancy between the driver and the device tree:
the driver code expects an absolute I/O address in the <reg> property,
while the device tree defines the address relative to a declaration in
the parent pinctrl node.
Changing the device tree to fix a driver issue would normally be wrong,
however:
- I have run the first version of U-Boot in which this driver appears
(v2016.03) and the same problem exists, so this is not a regression;
- There is no code that references a parent device tree node that might
suggest the intent of the author was to parse the DT as it exists now;
- The equivalent Linux PIC32 GPIO driver also uses absolute addresses
for the GPIO <reg> property. This change brings the U-Boot DT more
into line with Linux.
Additionally, the data sheet (Microchip ref. 60001361H) shows that the
register set to control a GPIO bank spans 0xE0 bytes, but the device
tree specified size is only 0x48 bytes.
Signed-off-by: John Robertson <john.robertson@simiatec.com>
John Robertson [Tue, 1 Sep 2020 02:55:42 +0000 (02:55 +0000)]
mips: pic32mzdask: disable SDHCI SDCD signal workaround
The PIC32MZ DA Starter Kit does not need the card detect workaround
because the SDCD signal line is connected properly. Disable the
workaround in this case.
Signed-off-by: John Robertson <john.robertson@simiatec.com>
John Robertson [Tue, 1 Sep 2020 02:55:14 +0000 (02:55 +0000)]
mmc: pic32: Refresh PIC32 MMC driver
The existing driver is not compatible with the Driver Model.
This patch makes the necessary changes while also removing obsolescent
calls/properties as follows:
- fdtdec_* calls replaced with dev_read_* equivalents;
- 'clock-freq-min-max' property replaced by querying the frequency of
the source clock 'base_clk';
- The card detect erratum workaround is applied during probe rather than
overriding get_cd.
The card detect workaround (Microchip ref. DS80000736E, erratum #15) is
not always needed and can be disabled using a vendor specific DT
property.
Signed-off-by: John Robertson <john.robertson@simiatec.com>
Chris Packham [Sun, 6 Sep 2020 22:37:35 +0000 (10:37 +1200)]
configs: powerpc: Don't set CONFIG_ENV_ADDR for QorIQ SPIFLASH
Setting CONFIG_ENV_ADDR to something other than 0 stops gd->env_addr
from being allocated dynamically. When the environment is in SPI we need
it to be allocated as we can't use a direct memory mapped address.
Signed-off-by: Chris Packham <judge.packham@gmail.com>
Somewhere between v2020.04 and v2020.07 the mpc8xxx_spi driver broke,
I'm guessing due to this hunk
@@ -559,6 +560,8 @@ int dm_gpio_set_dir_flags(struct gpio_desc *desc, ulong flags)
if (ret)
return ret;
+ /* combine the requested flags (for IN/OUT) and the descriptor flags */
+ flags |= desc->flags;
ret = _dm_gpio_set_dir_flags(desc, flags);
from commit 695e5fd5469a ("gpio: update dir_flags management"). But
the blame is mostly on the driver itself which seems rather confused:
The chip select gpios are requested with GPIOD_ACTIVE_LOW, but then in
each activate/deactivate, dm_gpio_set_dir_flags() is called with
merely GPIOD_IS_OUT, and then the driver call set_value(0) for
activate.
That used to work, but with the above hunk, the ACTIVE_LOW setting
from the request becomes persistent, so the gpio driver ends up being
asked to set the value to 1 in mpc8xxx_spi_cs_activate().
So drop the dm_gpio_set_dir_flags() calls in the activate/deactivate
functions, and use a value of 1 to mean "logically enabled".
Ideally, I think we should also drop the GPIOD_ACTIVE_LOW from the
request and make it up to the list of gpio cs in DT to indicate
whether that CS is enabled when driven low (as is of course usually
the case), but that requires changing
arch/powerpc/dts/gdsys/gazerbeam-base.dtsi among others, and I don't
have that hardware to test on. I have, however, tested our
own (mpc8309-based) hardware with this change, and I have also tested
that removing the GPIOD_ACTIVE_LOW from the request and updating our
DT as
Wolfgang Wallner [Wed, 16 Sep 2020 14:57:53 +0000 (16:57 +0200)]
x86: acpi: Add memset to initialize SPCR table
Add a missing memset to acpi_create_spcr().
The other acpi_create_xxxx() functions perform a memset on their
structures, acpi_create_spcr() does not and as a result the contents of
this table are partly uninitialized (and thus random after every reset).
Fixes: b288cd960072 ("x86: acpi: Generate SPCR table") Signed-off-by: Wolfgang Wallner <wolfgang.wallner@br-automation.com> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Reviewed-by: Simon Glass <sjg@chromium.org> Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
[bmeng: fix the tags format in the commit message] Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Signed-off-by: Wolfgang Wallner <wolfgang.wallner@br-automation.com> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Reviewed-by: Simon Glass <sjg@chromium.org> Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Wolfgang Wallner [Fri, 11 Sep 2020 14:52:28 +0000 (16:52 +0200)]
x86: fsp: Replace e-mmc with emmc in devicetree bindings
The term eMMC is used inconsistently within the FSP devicetree
bindings (e-mmc and emmc), especially for "emmc-host-max-speed"
documentation and code disagree.
Change all eMMC instances within the FSP bindings to consistently
use "emmc". The term "emmc" is already used a lot within U-Boot,
while "e-mmc" is only used in the FSP bindings.
Signed-off-by: Wolfgang Wallner <wolfgang.wallner@br-automation.com> Reviewed-by: Simon Glass <sjg@chromium.org> Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
[bmeng: correct one typo in the commit message] Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Fixes: 0b885bcfd9b0 ("acpi: Add an acpi command") Cc: Wolfgang Wallner <wolfgang.wallner@br-automation.com> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Andy Shevchenko [Tue, 8 Sep 2020 13:57:22 +0000 (16:57 +0300)]
x86: edison: Move config SYS_MALLOC_LEN to Kconfig
This patch moves the the config SYS_MALLOC_LEN to Kconfig
as it is already done for zynq arch in commit 01aa5b8f0503
("Kconfig: Move config SYS_MALLOC_LEN to Kconfig for zynq").
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Reviewed-by: Simon Glass <sjg@chromium.org> Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tom Rini [Sat, 19 Sep 2020 12:31:52 +0000 (08:31 -0400)]
Merge branch '2020-09-18-assorted-bugfixes'
- SquashFS Coverity fixes
- bitflip fix in the alternate memtest command
- Disable networking on bcmstb boards where we didn't have any network
drivers enabled.
Maxim Uvarov [Fri, 28 Aug 2020 19:20:10 +0000 (22:20 +0300)]
efi: change 'env -e -i' usage syntax
'env -e -i' syntax was changed from "," to ":". Account for this also
in the documentation.
Fixes: 2b3fbcb59f41 ("efi_loader: use ':' as separator for setenv -i") Signed-off-by: Maxim Uvarov <maxim.uvarov@linaro.org>
Correct the usage description for setenv -e too. Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Allow specifying the node on which a property is searched.
Test the device tree consistency more rigorously.
Some efi_st_printf() calls have been converted to efi_st_error().
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Ralph Siemsen [Wed, 9 Sep 2020 16:10:00 +0000 (12:10 -0400)]
cmd: mem: fix range of bitflip test
The bitflip test uses two equal sized memory buffers. This is achieved
by splitting the range of memory into two pieces. The address of the
second buffer, as well as the length of each buffer, were not correctly
calculated. This caused bitflip test to access beyond the end of range.
This patch fixes the pointer arithmetic problem.
A second problem arises because u-boot "mtest" command expects the
ending address to be inclusive. When computing (end - start) this
results in missing 1 byte of the requested length. The bitflip test
expects a count rather than an "ending" address. Thus it fails to test
the last word of the requested range. Fixed by using (end - start + 1).
Added Kconfig option to optionally disable the bitflip test, since it
does add significantly to the time taken for "mtest".
Fixes: 8e434cb705d463bc8cff935160e4fb4c77cb99ab ("cmd: mem: Add bitflip
memory test to alternate mtest")
Signed-off-by: Ralph Siemsen <ralph.siemsen@linaro.org> Reviewed-by: Stefan Roese <sr@denx.de>
mmc: fsl_esdhc_imx: check the clock stable status after config the clock rate.
Currently, after config the clock rate, delay 10ms, this is quite a rough
method. Check the clock stable status in the present status register is
enough.
Tested-by: Ji Luo <ji.luo@nxp.com> Signed-off-by: Haibo Chen <haibo.chen@nxp.com>
Marek Vasut [Sat, 12 Sep 2020 23:35:08 +0000 (01:35 +0200)]
ARM: mx6: ddr: Add support for iMX6UL/ULL/SL/SDL
This patch adds support for iMX6UL/ULL/SL/SDL MMDC into the DDR calibration
code. The difference between MX6DQ and MX6UL/ULL/SL is that the later SoCs
have 2 SDQS registers, just like MX6SX, while the MX6DQ/MX6SDL has 8.
Fixes: 4f4c128c65 ("ARM: mx6: ddr: Add support for iMX6SX") Signed-off-by: Marek Vasut <marex@denx.de> Cc: Eric Nelson <eric@nelint.com> Cc: Fabio Estevam <fabio.estevam@nxp.com> Cc: Stefano Babic <sbabic@denx.de> Reviewed-by: Fabio Estevam <festevam@gmail.com>
The size of the binary created with the default U-boot config is much
greater than the default offset for environment `0x60000`.
In case if that binary is used for booting via MMC it is overlapped with
the environment stored on MMC.
This leads to U-Boot corruption while saving environment with `saveenv`
command and non-bootable SabreLite board.
The offset for environment `CONFIG_ENV_OFFSET=0x60000` was added in
commit a09fea1 but did not count in the change to `0xC0000` if option
`CONFIG_ENV_IS_IN_MMC` is used.
The offset is also used for variant with environment saving onto SPI NOR
flash (with enabled option `CONFIG_ENV_IS_IN_SPI_FLASH`). In that case the
U-Boot binary flashed on SPI NOR is also corrupted after environment
saving with the original 0x60000 offset.
Signed-off-by: Denis Pynkin <denis.pynkin@collabora.com> Reviewed-by: Tom Rini <trini@konsulko.com>
Peng Fan [Wed, 16 Sep 2020 07:17:18 +0000 (15:17 +0800)]
imx7: ccm: correct target interface num
According to i.MX 7Dual Applications Processor Reference Manual, Rev. 1
The target interface CCM root index ranges [0,124], so the number
should be 125.
Reported-by: Coverity 18045 Signed-off-by: Peng Fan <peng.fan@nxp.com> Reviewed-by: Ye Li <ye.li@nxp.com>
Igor Opaniuk [Mon, 14 Sep 2020 08:01:10 +0000 (11:01 +0300)]
colibri_imx7: use preboot for fdtfile evaluation
Enable and set preboot var with fdtfile evaluation.
preboot will be checked and run immediately before starting the
CONFIG_BOOTDELAY countdown and/or running the auto-boot command resp.
entering interactive mode.
This provides possibility to use different boot cmds in interactive mode
without manual setting fdtfile value, as it it's already evaluated
before entering interactive mode.
Igor Opaniuk [Mon, 14 Sep 2020 08:01:09 +0000 (11:01 +0300)]
colibri_vf: use preboot for fdtfile evaluation
Enable and set preboot var with fdtfile evaluation.
preboot will be checked and run immediately before starting the
CONFIG_BOOTDELAY countdown and/or running the auto-boot command resp.
entering interactive mode.
This provides possibility to use different boot cmds in interactive mode
without manual setting fdtfile value, as it it's already evaluated
before entering interactive mode.
Fixes: 304042c1f3("colibri_vf: set fdtfile for distroboot") Signed-off-by: Igor Opaniuk <igor.opaniuk@toradex.com>
Igor Opaniuk [Mon, 14 Sep 2020 08:01:08 +0000 (11:01 +0300)]
colibri-imx6ull: use preboot for fdtfile evaluation
Enable and set preboot var with fdtfile evaluation.
preboot will be checked and run immediately before starting the
CONFIG_BOOTDELAY countdown and/or running the auto-boot command resp.
entering interactive mode.
This provides possibility to use different boot cmds in interactive mode
without manual setting fdtfile value, as it it's already evaluated
before entering interactive mode.
I also followed the development of the SquashFS support in U-Boot
as part of Joao Marcos internship, so I would also appreciate receiving
new contributions and bug reports related to this topic.
Thomas Petazzoni [Mon, 14 Sep 2020 11:52:39 +0000 (13:52 +0200)]
MAINTAINERS: add myself as reviewer for SquashFS
As I have followed the development of the SquashFS support in U-Boot
as part of Joao Marcos work, it makes sense to get Cc'ed on
contributions/bug reports related to the squashfs support.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Marek Vasut [Sat, 12 Sep 2020 10:42:52 +0000 (12:42 +0200)]
lib: fdt: Fix fdtdec_setup_mem..() conversion to livetree API
Repair incorrectly negated condition in the original patch which broke
DT memory node parsing on everything which has more than one DT memory
node, e.g. R-Car3.
In case multiple valid memory nodes are present in the DT, the original
patch would complete parsing cycle for the first memory node, then move
on to the next one, identify it as a valid, and end the parsing. The fix
is to invert the condition, to make the code behave as it did before the
livetree conversion, so it would continue parsing the subsequent memory
nodes as well.
Fixes: c2f0950c33 ("lib: fdt: Convert fdtdes_setup_mem..() to livetree API") Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com> Cc: Michal Simek <michal.simek@xilinx.com> Cc: Simon Glass <sjg@chromium.org> Cc: Tom Rini <trini@konsulko.com> Tested-by: Biju Das <biju.das.jz@bp.renesas.com> Reviewed-by: Michal Simek <michal.simek@xilinx.com>
Tom Rini [Wed, 16 Sep 2020 12:08:06 +0000 (08:08 -0400)]
Merge tag 'efi-2020-10-rc5' of https://gitlab.denx.de/u-boot/custodians/u-boot-efi
Pull request for UEFI sub-system for efi-2020-10-rc5
The following bugs are fixed:
* unaligned access in br_i32_decode()
* missing restore of global data pointer in UEFI selftest
* missing restore of global data pointer on RISC-V in UEfI subsystem
* efi_var_mem_notify_exit_boot_services() should not be __efi_runtime
The linker sometimes creates PT_LOAD segments with length (p_filesz) zero
as described in https://man7.org/linux/man-pages/man5/elf.5.html. This
leads to build failures. We should ignore empty segments.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de> Tested-by: Tom Rini <trini@konsulko.com> Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Kever Yang [Mon, 13 Apr 2020 01:38:30 +0000 (09:38 +0800)]
rockchip: rv1108: use correct API for board callback
Use board_early_init_f() instead of mach_cpu_init() for board, the
board_early_init_f() is used for board init and after dm_initf, while
the mach_cpu_init() is used for CPU/SOC and before dm_initf()(not able
to use syscon API).
Fixes: 9cec336708 ("rockchip: evb-rv1108: Use syscon API to get grf base") Fixes: 4aa33690fc {"rockchip: elgin-rv1108: Use syscon API to get grf base") Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
mmc: msm_sdhci: Use mmc_of_parse for setting host_caps
Since the introduction of 'get_cd' callback in sdhci core,
dragonboard410c's MMC interface is broken. It turns out that 'get_cd'
callback checks for the host_caps for validating the chip select. And
since the msm_sdhci driver is not parsing the host_caps from DT, not
all of the cababilities are parsed properly. This results in the MMC
interfaces to be broken.
Hence, fix this by adding a call to 'mmc_of_parse' during driver probe.
efi_var_mem_notify_exit_boot_services() is invoked when ExitBootServices()
is called by the UEFI payload.
efi_var_mem_notify_exit_boot_services() should not be defined as
__efi_runtime as it is invoking EFI_ENTRY() and EFI_EXIT() which themselves
are not __efi_runtime.
Fixes: f1f990a8c958 ("efi_loader: memory buffer for variables") Fixes: e01aed47d6a0 ("efi_loader: Enable run-time variable support for tee based variables") Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de> Acked-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Robert Reither [Mon, 14 Sep 2020 11:12:02 +0000 (13:12 +0200)]
rsa: crash in br_i32_decode() called from rsa_gen_key_prop()
Fixes problem for unaligned 32bit big-endian access in
lib/rsa/rsa-keyprop.c.
Exchanges br_i32_decode() with get_unaligned_be32().
This will keep the unaligned access for architectures capable and will do
some byte-shift magic for the not so capable ones.
Reported-by: Heinrich Schuchardt <xypron.glpk@gmx.de> Signed-by: Robert Reither <robert.reither@external.thalesgroup.com>
Remove unused include. Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
test: do no assume hush parser in validate_empty()
The environment variable test uses function validate_empty() to check that
a variable is not defined. If the hush parser is not enabled, we cannot
refer to a variable by $var_name but only by ${var_name}.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de> Acked-by: Stephen Warren <swarren@nvidia.com>
Attempting to place device tree immediately after an image in memory can lead
to mis-aligned data accesses if that image size is not divisible by the
alignment requirements of the architecture.
Data aborts caused by this were observed on a custom Marvel A388 based system,
where the image was a uboot FIT file. The total size varies depending on the
uboot device tree size, which does not always lead to correct alignment.
The minimum alignment specified for ARM [1] and ARM64 [2] linux booting has been
used
T Karthik Reddy [Mon, 31 Aug 2020 12:27:37 +0000 (14:27 +0200)]
mtd: nand: Fix nand write error with bad block addresses above 32-bit
Nand writes should skip the bad blocks with "nand write" command.
In case of bad blocks with above 32-bit address, nand_block_isbad()
returns false due to truncated bad block address.
In below code segment,
if (nand_block_isbad(mtd, offset & ~(mtd->erasesize - 1)))
offset is 64-bit and mtd->erasesize is 32-bit, hence the truncation is
happening. Cast 'mtd->erasesize' with loff_t to fix this issue.
Signed-off-by: T Karthik Reddy <t.karthik.reddy@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Pedro Aguilar [Mon, 31 Aug 2020 09:01:41 +0000 (11:01 +0200)]
env: Crash in 'env import' when using checksum and a specific size
This patch adds a sanity check that avoids 'size' to overflow and crash when
importing an environment that contains a checksum. Example with the wrong size
that causes the crash:
=> env import -c 0x4100000 3 v1
This assumes that v1 has already been successfully exported with
'env export -c -s 0x100 0x4100000 v1'
Signed-off-by: Pedro Aguilar <pedro.aguilar@vimar.com>
bootm: update image OS image size when decompressing
In bootm_load_os() the OS image is decompressed. In later stages of the
boot process we need the decompressed size of the image.
Update images->os.image_len after decompression.
Passing the correct size is necessary if we want to check loaded EFI
binararies for file truncation by comparing the loaded size to the header
field SizeOfImage.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de> Reviewed-by: Simon Glass <sjg@chromium.org>
Michal Simek [Thu, 27 Aug 2020 16:01:28 +0000 (18:01 +0200)]
common: Kconfig: Add dependency for default variables strings
Kconfig provides several config options for setting up default variables
but these are unused when variables are passed to U-Boot via file.
That's why cover this dependency in Kconfig.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Chuanjia Liu [Mon, 31 Aug 2020 07:53:12 +0000 (15:53 +0800)]
PCI: mediatek: Release the resource when PCIe enable port fail
On the mt7623 platform, if one port enable fail and other port
enable succeed. It will hang on when using pci enum
because the resource was not released correctly.
Signed-off-by: Chuanjia Liu <Chuanjia.Liu@mediatek.com> Tested-by: Frank Wunderlich <frank-w@public-files.de>
projects / u-boot.git / log