GitLab

While our "extrawarns" logic has gotten out of sync with upstream
Kbuild, for now lets start by bringing in the latest set of DTC_FLAGS
from the Linux Kernel 5.0 to match their behavior in silencing warnings
from dtc.

Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Tom Rini <trini@konsulko.com>

The pointer checksum were used before checking that it
isn't NULL. We move the code that use it after the check.

Reported-by: Coverity (CID: 185835)
Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

- Don't attempt to write to ext4 volumes with the metadata_csum feature
  we don't support
- Correct how we build dtb files.
- Fix regression on SPL_FIT with FIT_IMAGE_TINY
- Minor TI platform fixes

U-Boot doesn't support metadata_csum feature. Writing to filesystem with
metadata_csum feature makes the filesystem corrupted and unbootable by
Linux:

[    2.527495] EXT4-fs (mmcblk0p2): ext4_check_descriptors: Checksum for group 0 failed (52188!=0)
[    2.537421] EXT4-fs (mmcblk0p2): ext4_check_descriptors: Checksum for group 1 failed (5262!=0)
...
[    2.653308] EXT4-fs (mmcblk0p2): ext4_check_descriptors: Checksum for group 14 failed (42611!=0)
[    2.662179] EXT4-fs (mmcblk0p2): ext4_check_descriptors: Checksum for group 15 failed (21527!=0)
[    2.687920] JBD2: journal checksum error
[    2.691982] EXT4-fs (mmcblk0p2): error loading journal
[    2.698292] VFS: Cannot open root device "mmcblk0p2" or unknown-block(179,2): error -74

Don't write to filesystem with meatadata_csum feature to not corrupt the
filesystem.
Signed-off-by: Sébastien Szymanski <sebastien.szymanski@armadeus.com>

When these were moved from mach-omap2 to board files they got placed
inside TI_SECURE_DEVICE ifdef block, they are not secure only, move
them up and out.

Fixes: 413b9077 ("ti: fastboot: Move weak overrides to board files")
Signed-off-by: Andrew F. Davis <afd@ti.com>
Reviewed-by: Lokesh Vutla <lokeshvutla@ti.com>

This patch adds a comment to the header with the PCI_foo macros related
to DEVFN to explain the difference in U-Boot vs Linux.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Simon Glass <sjg@chromium.org>
Cc: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

Currently, 'dtimg' allows users to check indexes equal to
dt_entry_count [1]. Forbid that [2].

[1] Behavior w/o the patch:

=> ext2load mmc 0:1 0x48000000 dtb.img
105695 bytes read in 5 ms (20.2 MiB/s)

=> dtimg dump 0x48000000
dt_table_header:
               magic = d7b7ab1e
          total_size = 105695
         header_size = 32
       dt_entry_size = 32
      dt_entry_count = 2
   dt_entries_offset = 32
           page_size = 4096
             version = 0
dt_table_entry[0]:
             dt_size = 105599
           dt_offset = 96
                  id = 0b779520
                 rev = 00000000
           custom[0] = 00000000
           custom[1] = 00000000
           custom[2] = 00000000
           custom[3] = 00000000
           (FDT)size = 105599
     (FDT)compatible = shimafuji,kingfisher
dt_table_entry[1]:
             dt_size = 105599
           dt_offset = 96
                  id = 0b779530
                 rev = 00000000
           custom[0] = 00000000
           custom[1] = 00000000
           custom[2] = 00000000
           custom[3] = 00000000
           (FDT)size = 105599
     (FDT)compatible = shimafuji,kingfisher

=> dtimg size 0x48000000 0 z; print z
z=19c7f
=> dtimg size 0x48000000 1 z; print z
z=19c7f
=> dtimg size 0x48000000 2 z; print z
z=d00dfeed
=> dtimg size 0x48000000 3 z
Error: index > dt_entry_count (3 > 2)

[2] Behavior with the patch:

=> dtimg size 0x48000000 0 z; print z
z=19c7f
=> dtimg size 0x48000000 1 z; print z
z=19c7f
=> dtimg size 0x48000000 2 z
Error: index >= dt_entry_count (2 >= 2)

Fixes: c0447334 ("common: Add support for Android DT image")
Signed-off-by: Eugeniu Rosca <erosca@de.adit-jv.com>

With DM_MMC and DM_USB enabled, enable CONFIG_BLK.
Signed-off-by: Faiz Abbas <faiz_abbas@ti.com>

This reverts commit 27cb7300.

I am not sure if I correctly understood the log of commit 27cb7300
("Ensure device tree DTS is compiled"), but the code-diff looks like
it was trying to solve the missed re-compilation when .dts was modified.

Recently, commit 2737dfe0 ("kbuild: make arch-dtbs target PHONY")
fixed the issue in a more correct and more complete way.

Anyway, since the former commit, we see a clumsy log like this:

  make[2]: 'arch/sandbox/dts/sandbox.dtb' is up to date

Another problem is, it created multiple paths to descend into
arch/*/dts/, causing a race in parallel building.

So, let's revert it.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>

Since commit 27cb7300 ("Ensure device tree DTS is compiled"),
build succeeds irrespective of the correctness of Makefile.

In fact, you can compile any defconfig without adding any entry in
arch/*/dts/Makefile.

I am going to revert that commit, so device tree must be explicitly
listed in Makefile.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>

Since commit 27cb7300 ("Ensure device tree DTS is compiled"),
build succeeds irrespective of the correctness of Makefile.

I am going to revert that commit, so wrong code must be fixed.

CONFIG_MCR3000 is not defined anywhere. CONFIG_TARGET_MCR3000 is the
correct one.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>

Since commit 27cb7300 ("Ensure device tree DTS is compiled"),
build succeeds irrespective of the correctness of Makefile.

In fact, you can compile any defconfig without adding any entry in
arch/*/dts/Makefile.

As a result, a lot of wrong code have been merged unnoticed.

I am going to revert that commit, and lots of hidden issues have
come to light:

[1] Typos

  armada-3720-uDPU.dts, sun8i-a83t-tbs-a711.dts

  use the extension ".dts" instead of ".dtb"

[2] DTB is associated to undefined CONFIG option

  For example, mx6sllevk_defconfig defines CONFIG_MX6SLL, but
  associates its device tree to CONFIG_MX6SL, which is undefined.

[3] Lots of entries are missing
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Chris Packham <judge.packham@gmail.com>
[trini: add imx6ul pico dtbs]
Signed-off-by: Tom Rini <trini@konsulko.com>

When running mkimage with "-f auto", the loadable property
needs to be set in order to allow SPL FIT support to boot.
Signed-off-by: Abel Vesa <abel.vesa@nxp.com>
Tested-by: Fabio Estevam <festevam@gmail.com>

If FIT_IMAGE_TINY is enabled, spl_fit_image_get_os returns -ENOTSUPP.
In this case, we should default to IH_OS_U_BOOT not to IH_OS_INVALID.
Signed-off-by: Abel Vesa <abel.vesa@nxp.com>
Tested-by: Fabio Estevam <festevam@gmail.com>

Commit 1416e2d2 ("armv8: make SPL exception vectors optional") had a
typo in it which effectively disabled exception handling in SPL code always.

Since nobody complained, I guess we may as well disable exception handling
in SPL always by default.

So fix the bug to make the config option effective, but disable exception
handling in SPL by default. This gets us to the same functionality as before
by default, but with much less code included in the binary.
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Matthias Brugger <mbrugger@suse.com>
Reviewed-by: Andre Przywara <andre.przywara@arm.com>

When the SPL has full fitImage support, do not generate fitImage with
external data. The full fitImage code assumes the entire fitImage is
loaded in RAM, and the SPL uses fdt_totalsize() to determine the size
of the whole fitImage, which can not work with external data.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Tom Rini <trini@konsulko.com>

CONFIG_SYS_HELP_CMD_WIDTH is used to format the output of help without any
arguments.

CONFIG_SYS_HELP_CMD_WIDTH = 8 is too narrow to fit all our commands.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>

With U-boot supporting environment in multiple places, enable only
ENV_IS_IN_EMMC
Signed-off-by: Faiz Abbas <faiz_abbas@ti.com>
Reviewed-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>

Push the Starting kernel print to the end just before the
dm_remove_devices call.
Signed-off-by: Keerthy <j-keerthy@ti.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

- ARM: socfpga: Build sfp image only for Gen5 and Arria10 devices

Pull request for UEFI system for v2019.04-rc5

A bunch of small fixes. The major ones being

- avoid illegal memory access in efi_allocate_pool() on 32 bit systems
- avoid endless loop in HII protocol

- i2c: i2c_cdns: Fix below warnings with checker tool

This patch fixes below warnings found with checker tool.
The variable len in i2c_msg struct is of unsigned type
and it is received as recv_count which is unsigned type
but it is checked with < 0 which is always false, hence
removed it.
The local variable curr_recv_count is declared as signed
type and compared aginst unsigned recv_count which is
incorrect. This is fixed by declaring it as unsigned type.

drivers/i2c/i2c-cdns.c: In function ‘cdns_i2c_read_data’:
drivers/i2c/i2c-cdns.c:317:18: warning: comparison of
unsigned expression < 0 is always false [-Wtype-limits]
  if ((recv_count < 0))
                  ^
drivers/i2c/i2c-cdns.c:340:24: warning: comparison of
integer expressions of different signedness:
‘u32’ {aka ‘unsigned int’} and ‘int’ [-Wsign-compare]
  updatetx = recv_count > curr_recv_count;
                        ^
drivers/i2c/i2c-cdns.c:361:39: warning: comparison of
integer expressions of different signedness:
‘u32’ {aka ‘unsigned int’} and ‘int’ [-Wsign-compare]
    while (readl(&regs->transfer_size) !=
Signed-off-by: Siva Durga Prasad Paladugu <siva.durga.paladugu@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Heiko Schocher <hs@denx.de>

The sfp file is only valid for Gen5 (Cyclone5 & Arria5) and Arria10
devices.  The file should only be built for these devices.
Signed-off-by: Dalon Westergreen <dalon.westergreen@intel.com>

The check testing the string result of get_string() returned the wrong
result. The result was ignored.

Use efi_st_strcmp_16_8() for the string comparison.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>

printf("%ls", ..) expects u16 * as argument to print. There is not need for
a conversion to wchar_t *.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>

It does not make any sense to check if a pointer is NULL if we have
dereferenced it before.

Reported-by: Coverity (CID 185827)
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>

A misplaced return statement lead to a memory leak in
efi_dump_single_var().

Reported-by: Coverity (CID 185829)
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>

Add a missing return statement in efi_get_next_variable_name().

Reported-by: Coverity (CID 185834)
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>

Avoid an endless loop in add_strings_package().
Suggested-by: Takahiro Akashi <takahiro.akashi@linaro.org>
Reported-by: Coverity (CID 185833)
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>

efi_allocate_pages() expects a (uint64_t *) pointer to pass the address of
the assigned memory. If we pass the address of a pointer here, an illegal
memory access occurs on 32bit systems.

Fixes: 282a06cb ("efi_loader: Expose U-Boot addresses in memory map
for sandbox")
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>

- Enable network interface on clearfog_gt_8k (Baruch)
- Fix dreamplug boot by adding an spi0 alias to the DT (Chris)
- Fix / enhance Marvell ddr3 setup / parameters (Chris)
- Change CONFIG_SYS_MALLOC_F_LEN to 0x2000 on db-88f6820-amc (Chris)
- Enable SPL_FLASH_BAR on db-88f6820-amc (Chris)
- Use correct pcie controller name in Armada-38x dts files (Chris)
- Disable d-cache on Kirkwood platforms as currently needed (Chris)
- Add a more descriptive comment to pci_mvebu.c (Stefan)
- Update Marvell maintainers entry (Stefan)

- Fastboot fixes

Prior to commit 93b283d4 ("ARM: CPU: arm926ejs: Consolidate cache
routines to common file") the kirkwood boards didn't have and dcache
support. The network and usb drivers rely on this. Set
CONFIG_SYS_DCACHE_OFF in the Kirkwood specific config.h.
Reported-by: Leigh Brown <leigh@solinno.co.uk>
Signed-off-by: Chris Packham <judge.packham@gmail.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>

When armada-385.dtsi was sync'd from Linux the name of the node
describing the pcie controller was changed from pcie-controller to pcie.
Some of the boards that include armada-385.dtsi were missed in the
update retaining the old name. This updates the affected boards.
Reported-by: Влад Мао <vlaomao@gmail.com>
Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>

This board uses Micron N25Q256A SPI flash. Enable SPI_FLASH_BAR to allow
us to access the whole chip.
Signed-off-by: Chris Packham <judge.packham@gmail.com>
Cc: Vignesh R <vigneshr@ti.com>
Signed-off-by: Stefan Roese <sr@denx.de>

Set CONFIG_SYS_MALLOC_F_LEN to match the rest of the mvebu boards. The
default of 0x400 is not enough when booting from SPI.
Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>

Inactive chip-selects will give invalid values for read_sample so don't
consider them when trying to determine the overall min/max read sample.
Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>

[https://github.com/MarvellEmbeddedProcessors/mv-ddr-marvell/pull/18]
Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>

Based on the JEDEC standard JESD79-3F. The tRAS timings should include
the highest speed bins at a given frequency. This is similar to commit
683c67b ("mv_ddr: ddr3: fix tfaw timimg parameter") where the wrong
comparison was used in the initial implementation.
Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>

[https://github.com/MarvellEmbeddedProcessors/mv-ddr-marvell/pull/15]
Signed-off-by: Chris Packham <judge.packham@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>