It seems nobody tested the debug() option in spin_lock_irqsave().
Currently, when #define DEBUG, it spoils the compiler with
In file included from drivers/usb/dwc3/gadget.c:18:
drivers/usb/dwc3/gadget.c: In function ‘dwc3_gadget_set_selfpowered’:
include/log.h:235:4: warning: ‘flags’ is used uninitialized in this function [-Wuninitialized]
235 | printf(pr_fmt(fmt), ##args); \
| ^~~~~~
drivers/usb/dwc3/gadget.c:1347:17: note: ‘flags’ was declared here
1347 | unsigned long flags;
| ^~~~~
and so on...
Drop useless debug() call to make compiler happy.
Fixes: 0c06db5983 ("lib, linux: move linux specific defines to linux/compat.h")
Cc: Heiko Schocher <hs@denx.de>
Cc: Tom Rini <trini@ti.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Oleksandr Andrushchenko <oleksandr_andrushchenko@epam.com>
Reviewed-by: Heiko Schocher <hs@denx.de>
timeout should be in "millisecond" instead of second,
so divided it by 1000 when calculate the load value.
Signed-off-by: Zhao Qiang <qiang.zhao@nxp.com>
Address most of the checkpatch issues we found in km_arm and common km
code.
CC: Stefan Roese <sr@denx.de>
CC: Valentin Longchamp <valentin.longchamp@hitachi-powergrids.com>
Signed-off-by: Holger Brunck <holger.brunck@hitachi-powergrids.com>
Reviewed-by: Stefan Roese <sr@denx.de>
This commit add the support of signature check for config node
in spl/tpl when the function spl_load_simple_fit is used.
Signed-off-by: Philippe Reynes <philippe.reynes@softathome.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
USB clocks were completely forgotten as driver would always return 0 even if clock ID was unknown.
This behaviour changed with "IPQ40xx: clk: dont always return 0" and this will now causes the USB-s to fail probing as clock enable will return -EINVAL.
So to fix that lets add all of the USB clocks to the driver.
Fixes: 430e1dcf ("IPQ40xx: Add USB nodes")
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Cc: Luka Perkov <luka.perkov@sartura.hr>
Currently the driver will go through the clock ID-s and set/enable them as needed.
But if the ID is unknown it will fall through the switch case to the default case which will always return 0.
This is not correct and default cases should return a error code since clock ID is unknown.
So lets return -EINVAL instead.
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Cc: Luka Perkov <luka.perkov@sartura.hr>
There is no point in having break statements in the switch case as there is already a return before break.
So lets drop them from the driver.
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Cc: Luka Perkov <luka.perkov@sartura.hr>
Lets convert the driver to use dev_read_addr() instead of the devfdt_get_addr().
While we are here, lets also alphabetise the includes.
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Cc: Luka Perkov <luka.perkov@sartura.hr>
Move CONFIG_SYS_PROMPT_HUSH_PS2 in Kconfig, depending
on CONFIG_HUSH_PARSER, and remove the default value defined
in cli_hush.c under __U_BOOT__.
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
There is no reason to define default option for this macro which is
already done in common/cli_hush.c.
87 #ifndef CONFIG_SYS_PROMPT_HUSH_PS2
88 #define CONFIG_SYS_PROMPT_HUSH_PS2 "> "
89 #endif
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Acked-by: Oleksandr Andrushchenko <oleksandr_andrushchenko@epam.com>
CONFIG_TIMESTAMP is not related to the RTC drivers. It does not make any
sense to let the updating of the RTC by the sntp command depend on it.
Drop the CONFIG_TIMESTAMP checks.
Furthermore function dm_rtc_set() is enabled by CONFIG_DM_RTC. There is no
reason to require CONFIG_CMD_DATE when using a driver model RTC. The UEFI
sub-system can consume the RTC functions even if there is not date command.
Only check CONFIG_CMD_DATE when using a non-driver model RTC.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Add support for dealing with string operands, including reading a string
from memory into an environment variable and concatenating two strings.
Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Marek Behún <marek.behun@nic.cz>
At present a ulong is used to hold operand values. This means that
strings cannot be used. While most operations are not useful for strings,
concatenation is. As a starting point to supporting strings, convert the
code to use a struct instead of a ulong for operands.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present when more than one substitution is made this function
overwrites its buffers. Fix this bug and update the tests now that they
can pass.
Also update the debug code to show all substrings, since at present it
omits the final one.
Fixes: 855f18ea0e ("setexpr: add regex substring matching and substitution")
Signed-off-by: Simon Glass <sjg@chromium.org>
At present the 'nlen' variable increases with each loop. If the previous
loop had back references, then subsequent loops without back references
use the wrong value of nlen. The value is larger, meaning that the string
terminator from nbuf is copied along to the main buffer, thus terminating
the string prematurely.
This leads to the final result being truncated, e.g. missing the last
(unmatched) part of the string. So "match match tail" become
"replaced replaced" instead of "replaced replaced tail".
Fix this by resetting nlen to the correct value each time around the lop.
Fixes: 855f18ea0e ("setexpr: add regex substring matching and substitution")
Signed-off-by: Simon Glass <sjg@chromium.org>
Add tests to check for buffer overflow using simple replacement as well
as back references. At present these don't fully pass.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present this function always allocates a large amount of stack, and
selects its own size for buffers. This makes it hard to test the code
for buffer overflow.
Separate out the inner logic of the substitution so that tests can call
this directly. This will allow checking that the algorithm does not
overflow the buffer.
Fix up one of the error lines at the same time, since it should be
printing nbuf_size, not data_size.
Signed-off-by: Simon Glass <sjg@chromium.org>
Add tests for the setexpr regex commands.
Note that these tests currently crash on sandbox due to an existing bug in
the setexpr implementation, so two of the tests are commented out.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present this function assumes that a size of 4 refers to a ulong. This
is true on 32-bit machines but not commonly on 64-bit machines.
This means that the 'l' specify does not work correctly with setexpr.
Add an explicit case for 32-bit values so that 64-bit machines can still
use the 'l' specifier. On 32-bit machines, 64-bit is still not supported.
This corrects the operation of the default size (which is 4 for setexpr),
so update the tests accordingly.
The original code for reading from memory was included in 47ab5ad145
("cmd_setexpr: allow memory addresses in expressions") but I am not adding
a Fixes: tag since that code was not written with 64-bit machines in mind.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present these values are open-coded in a few places. Add constants so
the meaning is clear.
Also add a comment to cmd_get_data_size()
Signed-off-by: Simon Glass <sjg@chromium.org>
This command currently has no tests. Add some for basic assignment and the
integer operations.
Note that the default size for setexpr is ulong, which varies depending on
the build machine. So for sandbox on a 64-bit host, this means that the
default size is 64 bits.
Signed-off-by: Simon Glass <sjg@chromium.org>
On more places is used pattern 'command > $@ || rm -f $@'. But it does not
propagate failure from 'command' as 'rm -f' returns success.
Fix it by calling 'false' to correctly propagate failure after 'rm -f'.
Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
Use ofnode_ or dev_ APIs instead of fdt_ and fdtdec_ APIs so that the
driver can support live DT.
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Heiko Schocher <hs@denx.de>
- ProxyDHCP allows a second DHCP server to exist alongside your main
DHCP server and supply additional BOOTP related options
- When u-boot sends out a DHCP request, the real DHCP server will
respond with a normal response containing the new client IP address
while simultaneously the ProxyDHCP server will respond with a blank
client IP address and a `bootfile` option
- This patch adds CONFIG_SERVERIP_FROM_PROXYDHCP (default false) to
enable this behavior and CONFIG_SERVERIP_FROM_PROXYDHCP_DELAY_MS
(default 100) which tells u-boot to wait additional time after
receiving the main DHCP response to give the ProxyDHCP response time
to arrive
- The PXE spec for ProxyDHCP is more complicated than the solution
added here as diagramed on page 16:
http://www.pix.net/software/pxeboot/archive/pxespec.pdf:
```
DHCP Discover will be retried four times. The four timeouts are 4, 8, 16
and 32 seconds respectively. If a DHCPOFFER is received without an Option
timeouts in an attempt to receive a PXE response.
```
- Adding a simple delay worked for my purposes but let me know if a
more robust solution is required
Signed-off-by: Lyle Franklin <lylejfranklin@gmail.com>
As FDTDIR label doesn't specify exact file to be loaded, it should
not fail if no file exists in the directory. In this case try to boot
with internal FDT if it exists.
Signed-off-by: Anton Leontiev <aleontiev@elvees.com>
The origin patch didn't have this change and it was caused by manual
resolution where additional backslash was added.
Fixes: 6723b4c6ca ("binman: Call helper function binman_set_rom_offset() to fill offset")
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
The test should be checking whether $flags are non-zero and $drv_flags
contain specific flags, however these two sets of flags are separate,
and the two tests should be logically ANDed, not bitwise ANDed.
Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com>
Cc: Simon Glass <sjg@chromium.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
Since commit 6900797678 ("cros_ec: Support keyboard scanning with
EC_CMD_GET_NEXT_EVENT") the cros-ec-keyb driver has started using this
command, but the sandbox EC emulator does not recognize it and
continuously prints:
** Unknown EC command 0x67
This patch makes the sandbox driver send basic responses to the command,
but the response only supports keyboard scans for now.
The EC side of this command stores and returns events from a queue, and
returns -EC_RES_UNAVAILABLE when there are no new events. This should be
possible to implement by hooking into the SDL event queue (perhaps via
sandbox_sdl_poll_events). Implementing that is a bit harder to do since
the existing sandbox code is discarding pending keyboard events, then
reading the current keyboard state.
Since the EC emulator never explicitly fails to work on this command,
the fallback to the older command will not trigger and will not be
tested anymore.
Fixes: 6900797678 ("cros_ec: Support keyboard scanning with EC_CMD_GET_NEXT_EVENT")
Reported-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alper Nebi Yasak <alpernebiyasak@gmail.com>
Tested-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
At present if CROSS_COMPILE contains a tilde, such as
~/.buildman-toolchains/gcc-7.3.0-nolibc/i386-linux/bin/i386-linux-gcc
then binman gives a confusing error:
binman: Error 255 running '~/..buildman-toolchains/gcc-7.3.0- ...
Fix this by expanding it out before running the tool.
Signed-off-by: Simon Glass <sjg@chromium.org>
Add a tool to update or insert an Octeon specific header into the U-Boot
image. This is needed e.g. for booting via SPI NOR, eMMC and NAND.
While working on this, move enum cvmx_board_types_enum and
cvmx_board_type_to_string() to cvmx-bootloader.h and remove the
unreferenced (unsupported) board definition.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Aaron Williams <awilliams@marvell.com>
Cc: Chandrakala Chavva <cchavva@marvell.com>
Cc: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
Reviewed-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
Using ram_get_info() is complicated and does not work after relocation.
Now that gd->ram_size holds the full RAM size, let's use it instead and
remove the ram_get_size logic completely.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Aaron Williams <awilliams@marvell.com>
Cc: Chandrakala Chavva <cchavva@marvell.com>
Cc: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
Reviewed-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
With this patch, gd->ram_size now holds to full RAM size detected by the
DDR init code. It introduces the get_effective_memsize() function to
report the maximum usable RAM size in U-Boot to the system instead.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Aaron Williams <awilliams@marvell.com>
Cc: Chandrakala Chavva <cchavva@marvell.com>
Cc: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
Reviewed-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
Don't use "platdata_auto_alloc_size" but "priv_auto_alloc_size" instead
to auto allocate the private data struct, which is referenced via
dev_get_priv() in this driver. This fixes an ugly bug detected while
trying to boot via SPI NOR.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Aaron Williams <awilliams@marvell.com>
Cc: Chandrakala Chavva <cchavva@marvell.com>
Cc: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
Octeon has a specific boot header, when booted via SPI NOR, NAND or MMC.
Here the only 2 instructions are allowed in the first few bytes of the
image. And these instructions need to be one branch and a nop. This
patch adds the necessary nop after the nop, to that the common MIPS
image is compatible with this Octeon header.
The tool to patch the Octeon boot header into the image will be send in
a follow-up patch.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Aaron Williams <awilliams@marvell.com>
Cc: Chandrakala Chavva <cchavva@marvell.com>
Cc: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
Reviewed-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
The following errors are corrected:
* Linux crash when accessing UEFI variables at runtime.
* UEFI variable using standalone MM on 32 bit systems
not working due to missing packing of communication
structure
* NULL dereference when FAT16 root directory is full
* FAT files with a short file name starting with 0xE5 (0x05 in directory
entry) where treated as deleted.
The UEFI SetTime() service is enabled on ARM QEMU.
-----BEGIN PGP SIGNATURE-----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=CgpH
-----END PGP SIGNATURE-----
Merge tag 'efi-2021-01-rc3-3' of https://gitlab.denx.de/u-boot/custodians/u-boot-efi
Pull request for UEFI sub-system for efi-2021-01-rc3 (3)
The following errors are corrected:
* Linux crash when accessing UEFI variables at runtime.
* UEFI variable using standalone MM on 32 bit systems
not working due to missing packing of communication
structure
* NULL dereference when FAT16 root directory is full
* FAT files with a short file name starting with 0xE5 (0x05 in directory
entry) where treated as deleted.
The UEFI SetTime() service is enabled on ARM QEMU.
commit 1fabfeef506c ("efi_loader: parameter check in GetNextVariableName()")
introduces a check using u16_strnlen(). This code is used on EFI
runtime variables as well, so unless we mark it as runtime, the kernel
will crash trying to access it.
Fixes: 1fabfeef506c ("efi_loader: parameter check in GetNextVariableName()")
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Reviewed-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Pack struct efi_mm_communicate_header as done in EDK2 as seen in
release 201808 [1]. If not packed sizeof() for the structure adds
4 additional bytes on 32bit targets which breaks the ABI.
Link: [1] https://github.com/tianocore/edk2/blob/edk2-stable201808/MdePkg/Include/Protocol/MmCommunication.h#L21
Fixes: 23a397d2e2 ("efi_loader: Add headers for EDK2 StandAloneMM communication")
Signed-off-by: Etienne Carriere <etienne.carriere@linaro.org>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Enable EFI_SET_TIME on the sandbox and QEMU ARM to ensure that we compile
and test the relevant code.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Using constants instead of anonymous numbers increases code readability.
Fixes: 704df6aa0a ("fs: fat: refactor write interface for a file offset")
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
0x05 is used as replacement letter for 0xe5 at the first position of short
file names. We must not skip over directory entries starting with 0x05.
Cf. Microsoft FAT Specification, August 30 2005
Fixes: 39606d462c ("fs: fat: handle deleted directory entries correctly")
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
When trying to create a file in the full root directory of a FAT32
filesystem a NULL dereference can be observed.
When the root directory of a FAT16 filesystem is full fill_dir_slot() must
return -1 to signal that a new directory entry could not be allocated.
Fixes: cd2d727fff ("fs: fat: allocate a new cluster for root directory of fat32")
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
