Michal Simek [Tue, 30 Jul 2019 15:07:27 +0000 (17:07 +0200)]
arm64: zynqmp: Do not include pm_cfg_obj.o when SPL is disabled
xilinx_zynqmp_mini configuration is throwing build error:
readlink: missing operand
Try 'readlink --help' for more information.
because CONFIG_ZYNQMP_SPL_PM_CFG_OBJ_FILE is not defined at all and
Makefile pass ifneq condition. Add SPL_BUILD dependency which is also
reflected in Kconfig.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Alexander Dahl [Fri, 28 Jun 2019 12:41:24 +0000 (14:41 +0200)]
cmd: fpga: Change return value to avoid printing usage text
In cmd/fpga.c the commands should return enum command_ret_t, e.g.
CMD_RET_USAGE, CMD_RET_SUCCESS, or CMD_RET_FAILURE. What they actually
do is passing a return value from different 'fpga_' functions.
Passing on a return value of -1 from a called function leads to printing
out usage text. In case of actually correct usage with correctly
specified parameters but some fail at runtime printing out that usage
text is distracting.
The reason is most 'fpga_' functions return either FPGA_SUCCESS or
FPGA_FAIL, the latter was equal to -1 which is the same value as
CMD_RET_USAGE. So just passing on FPGA_FAIL lead to printing out usage.
We should only return CMD_RET_USAGE in cases, where the user sent wrong
input. Every other case should return CMD_RET_SUCCESS or
CMD_RET_FAILURE, and not simply pass an error code.
Simply changing FPGA_FAIL from -1 to 1 gets the job done.
Suggested-by: Michal Simek <michal.simek@xilinx.com> Signed-off-by: Alexander Dahl <ada@thorsis.com>
Alexander Dahl [Fri, 28 Jun 2019 12:41:20 +0000 (14:41 +0200)]
fpga: altera: Add some more device sizes
There seems to be only one place, where this is checked against:
`altera_validate()`. It should be non zero. Otherwise it is only used to
display it, so it probably does not really matter at the moment. But we
had the datasheet open anyway …
Sizes in datasheet are bit counts, display here is in bytes.
Luca Ceresoli [Thu, 20 Jun 2019 16:18:16 +0000 (18:18 +0200)]
arm64: zynqmp: add MAINTAINERS entry for Avnet UltraZed-EV
The board was added without adding a MAINTAINERS entry.
Fixes:
$ ./tools/genboardscfg.py -f
WARNING: no status info for 'avnet_ultrazedev_cc_v1_0_ultrazedev_som_v1_0'
WARNING: no maintainers for 'avnet_ultrazedev_cc_v1_0_ultrazedev_som_v1_0'
$
Reported-by: Michal Simek <michal.simek@xilinx.com> Signed-off-by: Luca Ceresoli <luca@lucaceresoli.net> Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Robert Hancock [Tue, 18 Jun 2019 15:47:16 +0000 (09:47 -0600)]
fpga: virtex2: Add slave serial programming support
This adds support for slave serial programming, in addition to the
previously supported slave SelectMAP mode. There are two ways that this
can be used:
-Using the clk and wdata callbacks in order to write image data one bit
at a time using pure bit-banging. This works, but is rather painfully
slow with typical image sizes.
-By specifying the wbulkdata callback instead, the image loading process
can be offloaded to SPI hardware. In this mode the clk and wdata
callbacks do not need to be specified. This allows the image to be
loaded much faster, taking only a few seconds with even relatively large
images.
Slave serial programming has been tested on the Kintex-7 series of
FPGAs.
Signed-off-by: Robert Hancock <hancock@sedsystems.ca> Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Robert Hancock [Tue, 18 Jun 2019 15:47:15 +0000 (09:47 -0600)]
fpga: virtex2: Add additional clock cycles after DONE assertion
Some Xilinx FPGA configuration options can result in the startup
sequence extending past the end of the FPGA bitstream. Continue applying
CCLK clock cycles for 8 cycles after DONE is asserted in order to ensure
the startup sequence is complete, as recommended by Xilinx.
Signed-off-by: Robert Hancock <hancock@sedsystems.ca> Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Robert Hancock [Tue, 18 Jun 2019 15:47:14 +0000 (09:47 -0600)]
fpga: virtex2: Split out image writing from pre/post operations
This is in preparation for adding slave serial programming support,
which uses the same pre/post operations as slave SelectMAP, to avoid
duplicating code.
Signed-off-by: Robert Hancock <hancock@sedsystems.ca> Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Robert Hancock [Tue, 18 Jun 2019 15:47:13 +0000 (09:47 -0600)]
fpga: virtex2: added Kconfig option
Add an option to allow this driver to be selected with Kconfig. As noted
in the description, this driver should also work with many newer Xilinx
FPGA families as the programming methods are essentially the same.
Also added a missing FPGA_XILINX dependency to the similar Spartan 3
driver.
Signed-off-by: Robert Hancock <hancock@sedsystems.ca> Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Luca Ceresoli [Tue, 11 Jun 2019 16:39:41 +0000 (18:39 +0200)]
arm64: zynqmp: add support for Avnet UltraZed-EV Starter Kit
Avnet UltraZed-EV Starter Kit is composed by the UltraZed-EV SoM and the
only publicly-available compatible carrier card. The SoM is based on the EV
version of the Xilinx ZynqMP SoC+FPGA.
The psu_init_gpl.c file has been generated from the board definition files
at [0] using Vivado 2018.3 and then minimized by
tools/zynqmp_psu_init_minimize.sh. Manually removed serdes init code since
it is not mentioned in device tree and fixed a checkpatch error.
Luca Ceresoli [Fri, 24 May 2019 13:40:02 +0000 (15:40 +0200)]
arm64: zynqmp: add tool to minimize psu_init_gpl.c files
This script transforms a pair of psu_init_gpl.c and .h files produced by
the Xilinx Vivado tool for ZynqMP into a smaller psu_init_gpl.c file that
is almost checkpatch compliant.
Based on a script by Michal Simek.
Signed-off-by: Luca Ceresoli <luca@lucaceresoli.net> Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Luca Ceresoli [Fri, 24 May 2019 13:40:01 +0000 (15:40 +0200)]
arm64: zynqmp: xil_io.h: declare functions as static
Fixes sparse warnings when building zynqmp defconfigs:
./board/xilinx/zynqmp/xil_io.h:12:6: warning: symbol 'Xil_Out32' was not declared. Should it be static?
./board/xilinx/zynqmp/xil_io.h:17:5: warning: symbol 'Xil_In32' was not declared. Should it be static?
./board/xilinx/zynqmp/xil_io.h:22:6: warning: symbol 'usleep' was not declared. Should it be static?
Also add __maybe_unused to usleep() since it is not used by minimized
psu_init_gpl.c files, so it would warn as "defined but not used".
Signed-off-by: Luca Ceresoli <luca@lucaceresoli.net> Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Luca Ceresoli [Tue, 21 May 2019 16:06:44 +0000 (18:06 +0200)]
arm64: zynqmp: add tool to convert PMU config object .c to binary
The recently-added ZYNQMP_SPL_PM_CFG_OBJ_FILE option allows SPL to load a
PMUFW configuration object from a binary blob. However the configuration
object is produced by Xilinx proprietary tools as a C source file and no
tool exists to easily convert it to a binary blob in an embedded Linux
build system for U-Boot to use.
Add a simple Python script to do the conversion.
It is definitely not a complete C language parser, but it is enough to
parse the known patterns generated by Xilinx tools, including:
- defines
- literal integers, optionally with a 'U' suffix
- bitwise OR between them
Signed-off-by: Luca Ceresoli <luca@lucaceresoli.net> Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Luca Ceresoli [Tue, 21 May 2019 16:06:43 +0000 (18:06 +0200)]
arm64: zynqmp: spl: install a PMU firmware config object at runtime
Optionally allow U-Boot to load a configuration object into the Power
Management Unit (PMU) firmware on Xilinx ZynqMP.
The configuration object is required by the PMU FW to enable most SoC
peripherals. So far the only way to boot using U-Boot SPL was to hard-code
the configuration object in the PMU firmware. Allow a different boot
process, where the PMU FW is equal for any ZynqMP chip and its
configuration is passed at runtime by U-Boot SPL.
All the code for Inter-processor communication with the PMU is isolated in
a new file (pmu_ipc.c). The code is inspired by the same feature as
implemented in the Xilinx First Stage Bootloader (FSBL) and Arm Trusted
Firmware:
Andrew F. Davis [Tue, 16 Jul 2019 13:49:38 +0000 (09:49 -0400)]
arm: mach-keystone: Use appended original image size in image processing
When an image is signed/encrypted on K2 devices the image may get padded
to meet alignment requirements for the signature/encryption type. The
original size is appended as 4 bytes little-endian to the end of the
final image.
Normally the trailing extra bytes are ignored and so restoring the
exact original size is not important. In the case of initrd the
original size is important as the kernel uses it to look for
additional filesystem data and can do the wrong thing when the
size is not correct.
Read off the original size and report it back from the image post
processing stage.
Signed-off-by: Andrew F. Davis <afd@ti.com> Reviewed-by: Lokesh Vutla <lokeshvutla@ti.com>
Extend the davinci NAND driver to support the driver model. For now this
doesn't add any device-tree parsing due to the fact that we can't access
the actual nand node on the device-tree - it's a subnode of the aemif
device and we don't have an aemif driver on davinci at the moment.
omapl138_lcdk_defconfig: don't build support for SPI
The following warning is emited when building u-boot for da850-lcdk:
===================== WARNING ======================
This board does not use CONFIG_DM_SPI. Please update
the board before v2019.04 for no dm conversion
and v2019.07 for partially dm converted drivers.
Failure to update can lead to driver/board removal
See doc/driver-model/MIGRATION.txt for more info.
====================================================
===================== WARNING ======================
This board does not use CONFIG_DM_SPI_FLASH. Please update
the board to use CONFIG_SPI_FLASH before the v2019.07 release.
Failure to update by the deadline may result in board removal.
See doc/driver-model/MIGRATION.txt for more info.
====================================================
While we could fix it by selecting CONFIG_DM_SPI, there's no need to
build it at all as SPI is disabled on da850-lcdk. Remove all unneeded
options from the defconfig.
davinci: remove CONFIG_DM_I2C_COMPAT from defconfigs
This option is no longer used on any davinci board but still selected
in defconfigs which causes the following warning:
===================== WARNING ======================
This board uses CONFIG_DM_I2C_COMPAT. Please remove
(possibly in a subsequent patch in your series)
before sending patches to the mailing list.
====================================================
Remove all references to CONFIG_DM_I2C_COMPAT from davinci.
Adam Ford [Mon, 29 Jul 2019 13:28:39 +0000 (08:28 -0500)]
ARM: da850_am18xxevm: Remove config and references
This board has been abandonded for a while. I do not have this hardware
and nobody has appeared to notice and/or care that it's orphaned.
Since much of this code is inserted into the da850evm code base, removing
it from support will make da850evm support easier going forward.
This patch removes the da850_am18xxevm_defconfig, README references to this
board and the reference to this board from the MAINTAINERS list.
Stephen Warren [Fri, 19 Jul 2019 17:21:17 +0000 (11:21 -0600)]
Makefile: fix implementation of BINMAN_DEBUG
binman only accepts the -D argument early on the command-line, yet the
Makefile currently passes it near the end. This causes the build to fail
if this feature is used. Re-order the command-line to fix this.
Signed-off-by: Stephen Warren <swarren@nvidia.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Simon Glass [Sat, 20 Jul 2019 18:24:10 +0000 (12:24 -0600)]
binman: Clean up all output directories in tests
At present some tests leave behind output directories. This happens
because some tests call binman, which sets up an output directory, then
call it again, which sets up another output directory and leaves the
original one behind.
Fix this by using a separate temporary directory when binman is called
twice, or by manually removing the output directory.
Simon Glass [Sat, 20 Jul 2019 18:24:08 +0000 (12:24 -0600)]
binman: Update state when replacing device-tree entries
Since the state module holds references to all the device trees used by
binman, it must be updated when the device trees are updated. Add support
for this.
Simon Glass [Sat, 20 Jul 2019 18:24:06 +0000 (12:24 -0600)]
binman: Support replacing data in a cbfs
At present binman cannot replace data within a CBFS since it does not
allow rewriting of the files in that CBFS. Implement this by using the
new WriteData() method to handle the case.
Add a header to compressed data so that the amount of compressed data can
be determined without reference to the size of the containing entry. This
allows the entry to be larger that the contents, without causing errors in
decompression. This is necessary to cope with a compressed device tree
being updated in such a way that it shrinks after the entry size is
already set (an obscure case). It is not used with CBFS since it has its
own metadata for this. Increase the number of passes allowed to resolve
the position of entries, to handle this case.
Simon Glass [Sat, 20 Jul 2019 18:24:05 +0000 (12:24 -0600)]
binman: Update Entry.WriteData() to handle special sections
At present this method assumes that the parent section does not need
to recalculate its position or adjust any metadata it may contain. But
when the entry changes size this may not be true. Also if the parent
section is more than just a container (e.g. it is a CBFS) then the
section may need to regenerate its output.
Add a new WriteChildData() method to sections and call this from the
WriteData() method, to handle this situation.
Simon Glass [Sat, 20 Jul 2019 18:24:04 +0000 (12:24 -0600)]
binman: Update Entry.ReadEntry() to work through classes
At present we simply extract the data directly from entries using the
image_pos information. This happens to work on current entry types, but
cannot work if the entry type encodes the data in some way. Update the
ReadData() method to provide the data by calling a new ReadChildData()
method in the parent. This allows the entry_Section class, or possibly
any other container class, to return the correct data in all cases.
Simon Glass [Sat, 20 Jul 2019 18:24:01 +0000 (12:24 -0600)]
binman: Place Intel descriptor at image start
The Intel descriptor must always appear at the start of an (x86) image,
so it is supposed to position itself there always. However there is no
explicit test for this. Add one and fix a bug introduced by the recent
change to adjust Entry to read the node in a separate call.
Simon Glass [Sat, 20 Jul 2019 18:24:00 +0000 (12:24 -0600)]
binman: Adjust fmap to ignore CBFS files
The FMAP is not intended to show the files inside a CBFS. The FMAP can be
used to locate the CBFS itself, but then the CBFS must be read to find out
what is in it.
Update the FMAP to work this way and add some debugging while we are here.
Simon Glass [Sat, 20 Jul 2019 18:23:59 +0000 (12:23 -0600)]
libfdt: Copy the struct region in fdt_resize()
At present this function appears to copy only the data before the struct
region and the data in the string region. It does not seem to copy the
struct region itself.
>From the arguments of this function it seems that it should support fdt
and buf being different. This patch attempts to fix this problem.
Upstream commit: c72fa77 libfdt: Copy the struct region in fdt_resize()
Simon Glass [Sat, 20 Jul 2019 18:23:58 +0000 (12:23 -0600)]
binman: Support shrinking a entry after packing
Sometimes an entry may shrink after it has already been packed. In that
case we must repack the items. Of course it is always possible to just
leave the entry at its original size and waste space at the end. This is
what binman does by default, since there is the possibility of the entry
changing size every time binman calculates its contents, thus causing a
loop.
Simon Glass [Sat, 20 Jul 2019 18:23:56 +0000 (12:23 -0600)]
binman: Allow updating entries that change size
So far we don't allow entries to change size when repacking. But this is
not very useful since it is common for entries to change size after an
updated binary is built, etc.
Add support for this, respecting the original offset/size/alignment
constraints of the image layout. For this to work the original image
must have been created with the 'allow-repack' property.
This does not support entry types with sub-entries such as files and
CBFS, but it does support sections.
Simon Glass [Sat, 20 Jul 2019 18:23:55 +0000 (12:23 -0600)]
binman: Add more tests for image header position
The positioning does not currently work correctly if at the end of an
image with no fixed size. Also if the header is in the middle of an image
it can cause a gap in the image since the header position is normally at
the image end, so entries after it are placed after the end of the image.
Fix these problems and add more tests to cover these cases.
Simon Glass [Sat, 20 Jul 2019 18:23:53 +0000 (12:23 -0600)]
binman: Write the original input fdtmap to a file
When reading an image in, write its fdtmap to a file in the output
directory. This is useful for debugging. Update the 'ls' command to set up
the output directory; otherwise it will fail.
Simon Glass [Sat, 20 Jul 2019 18:23:51 +0000 (12:23 -0600)]
binman: Add info to allow safely repacking an image later
At present it is not possible to discover the contraints to repacking an
image (e.g. maximum section size) since this information is not preserved
from the original image description.
Add new 'orig-offset' and 'orig-size' properties to hold this. Add them to
the main device tree in the image.
Simon Glass [Sat, 20 Jul 2019 18:23:50 +0000 (12:23 -0600)]
binman: Support updating entries in an existing image
While it is useful and efficient to build images in a single pass from a
unified description, it is sometimes desirable to update the image later.
Add support for replace an existing file with one of the same size. This
avoids needing to repack the file. Support for more advanced updates will
come in future patches.
Simon Glass [Sat, 20 Jul 2019 18:23:49 +0000 (12:23 -0600)]
binman: Tidy up _SetupDtb() to use its own temporary file
At present EnsureCompiled() uses an file from the 'output' directory (in
the tools module) when compiling the device tree. This is fine in most
cases, allowing useful inspection of the output files from binman.
However in functional tests, _SetupDtb() creates an output directory and
immediately removes it afterwards. This serves no benefit and just
confuses things, since the 'official' output directory is supposed to be
created and destroyed in control.Binman().
Add a new parameter for the optional temporary directory to use, and use a
separate temporary directory in _SetupDtb().
Simon Glass [Sat, 20 Jul 2019 18:23:48 +0000 (12:23 -0600)]
binman: Allow the fdtmap to remain unchanged
When updating an existing image where the size of all entries remains the
same, we should not need to regenerate the fdtmap. Update the entry to
return the same fdtmap as was read from the image.
Simon Glass [Sat, 20 Jul 2019 18:23:46 +0000 (12:23 -0600)]
binman: Add a function to obtain the image for an Entry
At present we have an 'image' property in the entry for this purpose, but
this is not necessary and seems error-prone in the presence of
inheritance. Add a function instead. The Entry_section class overrides
this with a special version, since top-level sections are in fact images,
since Image inherits Entry_section.
Simon Glass [Sat, 20 Jul 2019 18:23:45 +0000 (12:23 -0600)]
binman: Adjust Entry to read the node in a separate call
At present the Entry constructor sets up the object and then immediately
reads its device-tree node to obtain its properties.
This breaks a convention that constructors should not do any processing.
A consequence is that we must pass all arguments to the constructor and
cannot have the node-reading proceed in a different way unless we pass
flags to that constructor. We already have a 'test' flag in a few cases,
and now need to control whether the 'orig_offset' and 'orig_size'
properties are set or not.
Adjust the code to require a separate call to ReadNode() after
construction. The Image class remains as it was.
Simon Glass [Sat, 20 Jul 2019 18:23:44 +0000 (12:23 -0600)]
binman: Add an image name into the fdtmap
Since binman supports multiple images it is useful to know which one
created the image that has been read. Then it is possible to look up that
name in the 'master' device tree (containing the description of all
images).
Simon Glass [Sat, 20 Jul 2019 18:23:43 +0000 (12:23 -0600)]
binman: Store the entry in output_fdt_files
In some cases we want to access the Entry object for a particular device
tree. This allows us to read its contents or update it. Add this
information to output_fdt_files and provide a function to read it.
Also rename output_fdt_files since its name is no-longer descriptive.
Simon Glass [Sat, 20 Jul 2019 18:23:42 +0000 (12:23 -0600)]
binman: Allow state functions to fail to return data
At present these state functions raise an exception if they cannot find
what is requested. But in some cases the information is optional (e.g. an
fdtmap in a coming patch) so it is better to return gracefully.
Update these two functions to return None when the data cannot be found.
Simon Glass [Sat, 20 Jul 2019 18:23:41 +0000 (12:23 -0600)]
binman: Support loading entry data from a file
When modifying an image it is convenient to load the data from the file
into each entry so that it can be reprocessed. Add a new LoadData() method
to handle this.
Simon Glass [Sat, 20 Jul 2019 18:23:40 +0000 (12:23 -0600)]
binman: Store image fdtmap when loading from a file
This data provides all the information about the position and size of each
entry. Store it for later use when loading an image. It can be reused as
is if the image is modified without changing offsets/sizes.
Simon Glass [Sat, 20 Jul 2019 18:23:38 +0000 (12:23 -0600)]
dtoc: Update Fdt.FromData() to allow a name
It is confusing when something goes wrong with a device tree which was
created from data rather than a file, since there is no identifying
filename. Add an option to provide this. Use the filename as the name,
where available
Simon Glass [Sat, 20 Jul 2019 18:23:33 +0000 (12:23 -0600)]
binman: Simplify state.fdt_subset
At present this excludes the device tree passed in to binman since it
is always returned first by GetAllFdts(). However, this is easy to ensure
by adding a check in that function. Change this dict to includes all
device trees, and rename it to fdt_set.
Simon Glass [Sat, 20 Jul 2019 18:23:32 +0000 (12:23 -0600)]
binman: Adjust state.fdt_files to be keyed by entry type
It makes more sense to use entry type as the key for this dictionary,
since the filename can in principle be anything. Make this change and also
rename fdt_files and add a comment to explain it better.
Simon Glass [Sat, 20 Jul 2019 18:23:31 +0000 (12:23 -0600)]
binman: Adjust GetFdt() to be keyed by etype
At present the FDTs are keyed by their default filename (not their actual
filename). It seems easier to key by the entry type, since this is always
the same for each FDT type.
To do this, add a new Entry method called GetFdtEtype(). This is necessary
since some entry types contain a device tree which are not the simple
three entry types 'u-boot-dtb', 'u-boot-spl' or 'u-boot-tpl'.
The code already returns a dict for GetFdt(). Update the value of that
dict to include the filename so that existing code can work.
Simon Glass [Sat, 20 Jul 2019 18:23:29 +0000 (12:23 -0600)]
binman: Rename state.GetFdts()
This function name conflicts with Entry.GetFdts() which has a different
purpose. Rename it to avoid confusion. Also update a stale comment
relating to this function.
Simon Glass [Sat, 20 Jul 2019 18:23:28 +0000 (12:23 -0600)]
binman: Convert GetFdtSet() to use a dict
At present this function returns a set of device-tree filenames. It has no
way of returning the actual device-tree object. Change it to a dictionary
so that we can add this feature in a future patch.
Simon Glass [Sat, 20 Jul 2019 18:23:26 +0000 (12:23 -0600)]
binman: Use print() to print output
At present tout writes directly to stdout. This is not necessary and it
prevents tests from redirecting output. Change it to use print() for the
non-progress output.
Simon Glass [Sat, 20 Jul 2019 18:23:23 +0000 (12:23 -0600)]
dtoc: Return a non-zero exit code when tests fail
At present 'dtoc -t' return a success code even if some of the tests fail.
Fix this by checking the test result and setting the exit code. This
allows 'make qcheck' to function as expected.
Packed structures like struct uuid are not aligned. GCC 9.1 therefore
throws an error when trying to compile gen_rand_uuid().
lib/uuid.c: In function ‘gen_rand_uuid’:
lib/uuid.c:244:2: error: converting a packed ‘struct uuid’ pointer
(alignment 1) to a ‘unsigned int’ pointer (alignment 4) may result in
an unaligned pointer value [-Werror=address-of-packed-member]
244 | unsigned int *ptr = (unsigned int *)&uuid;
| ^~~~~~~~
Generate the uuid in a properly aligned buffer.
The byte order of a random number should not matter. Do not call
cpu_to_be32() to change the byte order.
Reported-by: Ramon Fried <rfried.dev@gmail.com> Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
projects / u-boot.git / log