1 .. SPDX-License-Identifier: GPL-2.0+
2 .. Copyright (c) 2016 Google, Inc
7 Firmware often consists of several components which must be packaged together.
8 For example, we may have SPL, U-Boot, a device tree and an environment area
9 grouped together and placed in MMC flash. When the system starts, it must be
10 able to find these pieces.
12 Building firmware should be separate from packaging it. Many of the complexities
13 of modern firmware build systems come from trying to do both at once. With
14 binman, you build all the pieces that are needed, using whatever assortment of
15 projects and build systems are needed, then use binman to stitch everything
22 Binman reads your board's device tree and finds a node which describes the
23 required image layout. It uses this to work out what to place where.
25 Binman provides a mechanism for building images, from simple SPL + U-Boot
26 combinations, to more complex arrangements with many parts. It also allows
27 users to inspect images, extract and replace binaries within them, repacking if
34 Apart from basic padding, alignment and positioning features, Binman supports
35 hierarchical images, compression, hashing and dealing with the binary blobs
36 which are a sad trend in open-source firmware at present.
38 Executable binaries can access the location of other binaries in an image by
39 using special linker symbols (zero-overhead but somewhat limited) or by reading
40 the devicetree description of the image.
42 Binman is designed primarily for use with U-Boot and associated binaries such
43 as ARM Trusted Firmware, but it is suitable for use with other projects, such
44 as Zephyr. Binman also provides facilities useful in Chromium OS, such as CBFS,
47 Binman provides a way to process binaries before they are included, by adding a
50 Binman is intended for use with U-Boot but is designed to be general enough
51 to be useful in other image-packaging situations.
57 As mentioned above, packaging of firmware is quite a different task from
58 building the various parts. In many cases the various binaries which go into
59 the image come from separate build systems. For example, ARM Trusted Firmware
60 is used on ARMv8 devices but is not built in the U-Boot tree. If a Linux kernel
61 is included in the firmware image, it is built elsewhere.
63 It is of course possible to add more and more build rules to the U-Boot
64 build system to cover these cases. It can shell out to other Makefiles and
65 build scripts. But it seems better to create a clear divide between building
66 software and packaging it.
68 At present this is handled by manual instructions, different for each board,
69 on how to create images that will boot. By turning these instructions into a
70 standard format, we can support making valid images for any board without
71 manual effort, lots of READMEs, etc.
75 - Each binary can have its own build system and tool chain without creating
76 any dependencies between them
77 - Avoids the need for a single-shot build: individual parts can be updated
78 and brought in as needed
79 - Provides for a standard image description available in the build and at
81 - SoC-specific image-signing tools can be accommodated
82 - Avoids cluttering the U-Boot build system with image-building code
83 - The image description is automatically available at run-time in U-Boot,
84 SPL. It can be made available to other software also
85 - The image description is easily readable (it's a text file in device-tree
86 format) and permits flexible packing of binaries
92 Binman uses the following terms:
94 - image - an output file containing a firmware image
95 - binary - an input binary that goes into the image
101 FIT is U-Boot's official image format. It supports multiple binaries with
102 load / execution addresses, compression. It also supports verification
103 through hashing and RSA signatures.
105 FIT was originally designed to support booting a Linux kernel (with an
106 optional ramdisk) and device tree chosen from various options in the FIT.
107 Now that U-Boot supports configuration via device tree, it is possible to
108 load U-Boot from a FIT, with the device tree chosen by SPL.
110 Binman considers FIT to be one of the binaries it can place in the image.
112 Where possible it is best to put as much as possible in the FIT, with binman
113 used to deal with cases not covered by FIT. Examples include initial
114 execution (since FIT itself does not have an executable header) and dealing
115 with device boundaries, such as the read-only/read-write separation in SPI
118 For U-Boot, binman should not be used to create ad-hoc images in place of
121 Note that binman can itself create a FIT. This helps to move mkimage
122 invocations out of the Makefile and into binman image descriptions. It also
123 helps by removing the need for ad-hoc tools like `make_fit_atf.py`.
126 Relationship to mkimage
127 -----------------------
129 The mkimage tool provides a means to create a FIT. Traditionally it has
130 needed an image description file: a device tree, like binman, but in a
131 different format. More recently it has started to support a '-f auto' mode
132 which can generate that automatically.
134 More relevant to binman, mkimage also permits creation of many SoC-specific
135 image types. These can be listed by running 'mkimage -T list'. Examples
136 include 'rksd', the Rockchip SD/MMC boot format. The mkimage tool is often
137 called from the U-Boot build system for this reason.
139 Binman considers the output files created by mkimage to be binary blobs
140 which it can place in an image. Binman does not replace the mkimage tool or
141 this purpose. It would be possible in some situations to create a new entry
142 type for the images in mkimage, but this would not add functionality. It
143 seems better to use the mkimage tool to generate binaries and avoid blurring
144 the boundaries between building input files (mkimage) and packaging then
145 into a final image (binman).
147 Note that binman can itself invoke mkimage. This helps to move mkimage
148 invocations out of the Makefile and into binman image descriptions.
154 Example use of binman in U-Boot
155 -------------------------------
157 Binman aims to replace some of the ad-hoc image creation in the U-Boot
160 Consider sunxi. It has the following steps:
162 #. It uses a custom mksunxiboot tool to build an SPL image called
163 sunxi-spl.bin. This should probably move into mkimage.
165 #. It uses mkimage to package U-Boot into a legacy image file (so that it can
166 hold the load and execution address) called u-boot.img.
168 #. It builds a final output image called u-boot-sunxi-with-spl.bin which
169 consists of sunxi-spl.bin, some padding and u-boot.img.
171 Binman is intended to replace the last step. The U-Boot build system builds
172 u-boot.bin and sunxi-spl.bin. Binman can then take over creation of
173 sunxi-spl.bin by calling mksunxiboot or mkimage. In any case, it would then
174 create the image from the component parts.
176 This simplifies the U-Boot Makefile somewhat, since various pieces of logic
177 can be replaced by a call to binman.
180 Invoking binman within U-Boot
181 -----------------------------
183 Within U-Boot, binman is invoked by the build system, i.e. when you type 'make'
184 or use buildman to build U-Boot. There is no need to run binman independently
185 during development. Everything happens automatically and is set up for your
186 SoC or board so that binman produced the right things.
188 The general policy is that the Makefile builds all the binaries in INPUTS-y
189 (the 'inputs' rule), then binman is run to produce the final images (the 'all'
192 There should be only one invocation of binman in Makefile, the very last step
193 that pulls everything together. At present there are some arch-specific
194 invocations as well, but these should be dropped when those architectures are
195 converted to use binman properly.
197 As above, the term 'binary' is used for something in INPUTS-y and 'image' is
198 used for the things that binman creates. So the binaries are inputs to the
199 image(s) and it is the image that is actually loaded on the board.
201 Again, at present, there are a number of things created in Makefile which should
202 be done by binman (when we get around to it), like `u-boot-ivt.img`,
203 `lpc32xx-spl.img`, `u-boot-with-nand-spl.imx`, `u-boot-spl-padx4.sfp` and
204 `u-boot-mtk.bin`, just to pick on a few. When completed this will remove about
205 400 lines from `Makefile`.
207 Since binman is invoked only once, it must of course create all the images that
208 are needed, in that one invocation. It does this by working through the image
209 descriptions one by one, collecting the input binaries, processing them as
210 needed and producing the final images.
212 The same binaries may be used by multiple images. For example binman may be used
213 to produce an SD-card image and a SPI-flash image. In this case the binaries
214 going into the process are the same, but binman produces slightly different
217 For some SoCs, U-Boot is not the only project that produces the necessary
218 binaries. For example, ARM Trusted Firmware (ATF) is a project that produces
219 binaries which must be incorporate, such as `bl31.elf` or `bl31.bin`. For this
220 to work you must have built ATF before you build U-Boot and you must tell U-Boot
221 where to find the bl31 image, using the BL31 environment variable.
223 How do you know how to incorporate ATF? It is handled by the atf-bl31 entry type
224 (etype). An etype is an implementation of reading a binary into binman, in this
225 case the `bl31.bin` file. When you build U-Boot but do not set the BL31
226 environment variable, binman provides a help message, which comes from
227 `missing-blob-help`::
229 See the documentation for your board. You may need to build ARM Trusted
230 Firmware and build with BL31=/path/to/bl31.bin
232 The mechanism by which binman is advised of this is also in the Makefile. See
233 the `-a atf-bl31-path=${BL31}` piece in `cmd_binman`. This tells binman to
234 set the EntryArg `atf-bl31-path` to the value of the `BL31` environment
235 variable. Within binman, this EntryArg is picked up by the `Entry_atf_bl31`
236 etype. An EntryArg is simply an argument to the entry. The `atf-bl31-path`
237 name is documented in :ref:`etype_atf_bl31`.
239 Taking this a little further, when binman is used to create a FIT, it supports
240 using an ELF file, e.g. `bl31.elf` and splitting it into separate pieces (with
241 `fit,operation = "split-elf"`), each with its own load address.
244 Invoking binman outside U-Boot
245 ------------------------------
247 While binman is invoked from within the U-Boot build system, it is also possible
248 to invoke it separately. This is typically used in a production build system,
249 where signing is completed (with real keys) and any missing binaries are
252 For example, for build testing there is no need to provide a real signature,
253 nor is there any need to provide a real ATF BL31 binary (for example). These can
254 be added later by invoking binman again, providing all the required inputs
255 from the first time, plus any that were missing or placeholders.
257 So in practice binman is often used twice:
259 - once within the U-Boot build system, for development and testing
260 - again outside U-Boot to assembly and final production images
262 While the same input binaries are used in each case, you will of course you will
263 need to create your own binman command line, similar to that in `cmd_binman` in
264 the Makefile. You may find the -I and --toolpath options useful. The
265 device tree file is provided to binman in binary form, so there is no need to
266 have access to the original `.dts` sources.
269 Assembling the image description
270 --------------------------------
272 Since binman uses the device tree for its image description, you can use the
273 same files that describe your board's hardware to describe how the image is
274 assembled. Typically the images description is in a common file used by all
275 boards with a particular SoC (e.g. `imx8mp-u-boot.dtsi`).
277 Where a particular boards needs to make changes, it can override properties in
278 the SoC file, just as it would for any other device tree property. It can also
279 add a image that is specific to the board.
281 Another way to control the image description to make use of CONFIG options in
282 the description. For example, if the start offset of a particular entry varies
283 by board, you can add a Kconfig for that and reference it in the description::
289 offset = <CONFIG_SPL_PAD_TO>;
293 The SoC can provide a default value but boards can override that as needed and
294 binman will take care of it.
296 It is even possible to control which entries appear in the image, by using the
299 #ifdef CONFIG_HAVE_MRC
301 offset = <CONFIG_X86_MRC_ADDR>;
305 Only boards which enable `HAVE_MRC` will include this entry.
307 Obviously a similar approach can be used to control which images are produced,
308 with a Kconfig option to enable a SPI image, for example. However there is
309 generally no harm in producing an image that is not used. If a board uses MMC
310 but not SPI, but the SoC supports booting from both, then both images can be
311 produced, with only on or other being used by particular boards. This can help
312 reduce the need for having multiple defconfig targets for a board where the
313 only difference is the boot media, enabling / disabling secure boot, etc.
315 Of course you can use the device tree itself to pass any board-specific
316 information that is needed by U-Boot at runtime (see binman_syms_ for how to
317 make binman insert these values directly into executables like SPL).
319 There is one more way this can be done: with individual .dtsi files for each
320 image supported by the SoC. Then the board `.dts` file can include the ones it
321 wants. This is not recommended, since it is likely to be difficult to maintain
322 and harder to understand the relationship between the different boards.
325 Producing images for multiple boards
326 ------------------------------------
328 When invoked within U-Boot, binman only builds a single set of images, for
329 the chosen board. This is set by the `CONFIG_DEFAULT_DEVICE_TREE` option.
331 However, U-Boot generally builds all the device tree files associated with an
332 SoC. These are written to the (e.g. for ARM) `arch/arm/dts` directory. Each of
333 these contains the full binman description for that board. Often the best
334 approach is to build a single image that includes all these device tree binaries
335 and allow SPL to select the correct one on boot.
337 However, it is also possible to build separate images for each board, simply by
338 invoking binman multiple times, once for each device tree file, using a
339 different output directory. This will produce one set of images for each board.
342 Example use of binman for x86
343 -----------------------------
345 In most cases x86 images have a lot of binary blobs, 'black-box' code
346 provided by Intel which must be run for the platform to work. Typically
347 these blobs are not relocatable and must be placed at fixed areas in the
350 Currently this is handled by ifdtool, which places microcode, FSP, MRC, VGA
351 BIOS, reference code and Intel ME binaries into a u-boot.rom file.
353 Binman is intended to replace all of this, with ifdtool left to handle only
354 the configuration of the Intel-format descriptor.
360 First install prerequisites, e.g:
364 sudo apt-get install python-pyelftools python3-pyelftools lzma-alone \
367 You can run binman directly if you put it on your PATH. But if you want to
368 install into your `~/.local` Python directory, use:
372 pip install tools/patman tools/dtoc tools/binman
374 Note that binman makes use of libraries from patman and dtoc, which is why these
375 need to be installed. Also you need `libfdt` and `pylibfdt` which can be
380 git clone git://git.kernel.org/pub/scm/utils/dtc/dtc.git
383 make NO_PYTHON=1 install
385 This installs the `libfdt.so` library into `~/lib` so you can use
386 `LD_LIBRARY_PATH=~/lib` when running binman. If you want to install it in the
387 system-library directory, replace the last line with:
391 make NO_PYTHON=1 PREFIX=/ install
400 make NO_PYTHON=1 PREFIX=/ install
401 binman build -b <board_name>
403 to build an image for a board. The board name is the same name used when
404 configuring U-Boot (e.g. for sandbox_defconfig the board name is 'sandbox').
405 Binman assumes that the input files for the build are in ../b/<board_name>.
407 Or you can specify this explicitly:
411 make NO_PYTHON=1 PREFIX=/ install
412 binman build -I <build_path>
414 where <build_path> is the build directory containing the output of the U-Boot
417 (Future work will make this more configurable)
419 In either case, binman picks up the device tree file (u-boot.dtb) and looks
420 for its instructions in the 'binman' node.
422 Binman has a few other options which you can see by running 'binman -h'.
425 Enabling binman for a board
426 ---------------------------
428 At present binman is invoked from a rule in the main Makefile. You should be
429 able to enable CONFIG_BINMAN to enable this rule.
431 The output file is typically named image.bin and is located in the output
432 directory. If input files are needed to you add these to INPUTS-y either in the
433 main Makefile or in a config.mk file in your arch subdirectory.
435 Once binman is executed it will pick up its instructions from a device-tree
436 file, typically <soc>-u-boot.dtsi, where <soc> is your CONFIG_SYS_SOC value.
437 You can use other, more specific CONFIG options - see 'Automatic .dtsi
442 Access to binman entry offsets at run time (symbols)
443 ----------------------------------------------------
445 Binman assembles images and determines where each entry is placed in the image.
446 This information may be useful to U-Boot at run time. For example, in SPL it
447 is useful to be able to find the location of U-Boot so that it can be executed
448 when SPL is finished.
450 Binman allows you to declare symbols in the SPL image which are filled in
451 with their correct values during the build. For example:
455 binman_sym_declare(ulong, u_boot_any, image_pos);
457 declares a ulong value which will be assigned to the image-pos of any U-Boot
458 image (u-boot.bin, u-boot.img, u-boot-nodtb.bin) that is present in the image.
459 You can access this value with something like:
463 ulong u_boot_offset = binman_sym(ulong, u_boot_any, image_pos);
465 Thus u_boot_offset will be set to the image-pos of U-Boot in memory, assuming
466 that the whole image has been loaded, or is available in flash. You can then
467 jump to that address to start U-Boot.
469 At present this feature is only supported in SPL and TPL. In principle it is
470 possible to fill in such symbols in U-Boot proper, as well, but a future C
471 library is planned for this instead, to read from the device tree.
473 As well as image-pos, it is possible to read the size of an entry and its
474 offset (which is the start position of the entry within its parent).
476 A small technical note: Binman automatically adds the base address of the image
477 (i.e. __image_copy_start) to the value of the image-pos symbol, so that when the
478 image is loaded to its linked address, the value will be correct and actually
479 point into the image.
481 For example, say SPL is at the start of the image and linked to start at address
482 80108000. If U-Boot's image-pos is 0x8000 then binman will write an image-pos
483 for U-Boot of 80110000 into the SPL binary, since it assumes the image is loaded
484 to 80108000, with SPL at 80108000 and U-Boot at 80110000.
486 For x86 devices (with the end-at-4gb property) this base address is not added
487 since it is assumed that images are XIP and the offsets already include the
491 Access to binman entry offsets at run time (fdt)
492 ------------------------------------------------
494 Binman can update the U-Boot FDT to include the final position and size of
495 each entry in the images it processes. The option to enable this is -u and it
496 causes binman to make sure that the 'offset', 'image-pos' and 'size' properties
497 are set correctly for every entry. Since it is not necessary to specify these in
498 the image definition, binman calculates the final values and writes these to
499 the device tree. These can be used by U-Boot at run-time to find the location
502 Alternatively, an FDT map entry can be used to add a special FDT containing
503 just the information about the image. This is preceded by a magic string so can
504 be located anywhere in the image. An image header (typically at the start or end
505 of the image) can be used to point to the FDT map. See fdtmap and image-header
506 entries for more information.
512 The -m option causes binman to output a .map file for each image that it
513 generates. This shows the offset and size of each entry. For example::
516 00000000 00000028 main-section
517 00000000 00000010 section@0
518 00000000 00000004 u-boot
519 00000010 00000010 section@1
520 00000000 00000004 u-boot
522 This shows a hierarchical image with two sections, each with a single entry. The
523 offsets of the sections are absolute hex byte offsets within the image. The
524 offsets of the entries are relative to their respective sections. The size of
525 each entry is also shown, in bytes (hex). The indentation shows the entries
526 nested inside their sections.
529 Passing command-line arguments to entries
530 -----------------------------------------
532 Sometimes it is useful to pass binman the value of an entry property from the
533 command line. For example some entries need access to files and it is not
534 always convenient to put these filenames in the image definition (device tree).
536 The -a option supports this::
542 <prop> is the property to set
543 <value> is the value to set it to
545 Not all properties can be provided this way. Only some entries support it,
546 typically for filenames.
549 Image description format
550 ========================
552 The binman node is called 'binman'. An example image description is shown
556 filename = "u-boot-sunxi-with-spl.bin";
559 filename = "spl/sunxi-spl.bin";
562 offset = <CONFIG_SPL_PAD_TO>;
567 This requests binman to create an image file called u-boot-sunxi-with-spl.bin
568 consisting of a specially formatted SPL (spl/sunxi-spl.bin, built by the
569 normal U-Boot Makefile), some 0xff padding, and a U-Boot legacy image. The
570 padding comes from the fact that the second binary is placed at
571 CONFIG_SPL_PAD_TO. If that line were omitted then the U-Boot binary would
572 immediately follow the SPL binary.
574 The binman node describes an image. The sub-nodes describe entries in the
575 image. Each entry represents a region within the overall image. The name of
576 the entry (blob, u-boot) tells binman what to put there. For 'blob' we must
577 provide a filename. For 'u-boot', binman knows that this means 'u-boot.bin'.
579 Entries are normally placed into the image sequentially, one after the other.
580 The image size is the total size of all entries. As you can see, you can
581 specify the start offset of an entry using the 'offset' property.
583 Note that due to a device tree requirement, all entries must have a unique
584 name. If you want to put the same binary in the image multiple times, you can
585 use any unique name, with the 'type' property providing the type.
587 The attributes supported for entries are described below.
590 This sets the offset of an entry within the image or section containing
591 it. The first byte of the image is normally at offset 0. If 'offset' is
592 not provided, binman sets it to the end of the previous region, or the
593 start of the image's entry area (normally 0) if there is no previous
597 This sets the alignment of the entry. The entry offset is adjusted
598 so that the entry starts on an aligned boundary within the containing
599 section or image. For example 'align = <16>' means that the entry will
600 start on a 16-byte boundary. This may mean that padding is added before
601 the entry. The padding is part of the containing section but is not
602 included in the entry, meaning that an empty space may be created before
603 the entry starts. Alignment should be a power of 2. If 'align' is not
604 provided, no alignment is performed.
607 This sets the size of the entry. The contents will be padded out to
608 this size. If this is not provided, it will be set to the size of the
612 Padding before the contents of the entry. Normally this is 0, meaning
613 that the contents start at the beginning of the entry. This can be used
614 to offset the entry contents a little. While this does not affect the
615 contents of the entry within binman itself (the padding is performed
616 only when its parent section is assembled), the end result will be that
617 the entry starts with the padding bytes, so may grow. Defaults to 0.
620 Padding after the contents of the entry. Normally this is 0, meaning
621 that the entry ends at the last byte of content (unless adjusted by
622 other properties). This allows room to be created in the image for
623 this entry to expand later. While this does not affect the contents of
624 the entry within binman itself (the padding is performed only when its
625 parent section is assembled), the end result will be that the entry ends
626 with the padding bytes, so may grow. Defaults to 0.
629 This sets the alignment of the entry size. For example, to ensure
630 that the size of an entry is a multiple of 64 bytes, set this to 64.
631 While this does not affect the contents of the entry within binman
632 itself (the padding is performed only when its parent section is
633 assembled), the end result is that the entry ends with the padding
634 bytes, so may grow. If 'align-size' is not provided, no alignment is
638 This sets the alignment of the end of an entry with respect to the
639 containing section. Some entries require that they end on an alignment
640 boundary, regardless of where they start. This does not move the start
641 of the entry, so the contents of the entry will still start at the
642 beginning. But there may be padding at the end. While this does not
643 affect the contents of the entry within binman itself (the padding is
644 performed only when its parent section is assembled), the end result
645 is that the entry ends with the padding bytes, so may grow.
646 If 'align-end' is not provided, no alignment is performed.
649 For 'blob' types this provides the filename containing the binary to
650 put into the entry. If binman knows about the entry type (like
651 u-boot-bin), then there is no need to specify this.
654 Sets the type of an entry. This defaults to the entry name, but it is
655 possible to use any name, and then add (for example) 'type = "u-boot"'
659 Indicates that the offset of this entry should not be set by placing
660 it immediately after the entry before. Instead, is set by another
661 entry which knows where this entry should go. When this boolean
662 property is present, binman will give an error if another entry does
663 not set the offset (with the GetOffsets() method).
666 This cannot be set on entry (or at least it is ignored if it is), but
667 with the -u option, binman will set it to the absolute image position
668 for each entry. This makes it easy to find out exactly where the entry
669 ended up in the image, regardless of parent sections, etc.
672 Extend the size of this entry to fit available space. This space is only
673 limited by the size of the image/section and the position of the next
677 Sets the compression algortihm to use (for blobs only). See the entry
678 documentation for details.
681 Sets the tag of the message to show if this entry is missing. This is
682 used for external blobs. When they are missing it is helpful to show
683 information about what needs to be fixed. See missing-blob-help for the
684 message for each tag.
687 By default binman substitutes entries with expanded versions if available,
688 so that a `u-boot` entry type turns into `u-boot-expanded`, for example. The
689 `--no-expanded` command-line option disables this globally. The
690 `no-expanded` property disables this just for a single entry. Put the
691 `no-expanded` boolean property in the node to select this behaviour.
693 The attributes supported for images and sections are described below. Several
694 are similar to those for entries.
697 Sets the image size in bytes, for example 'size = <0x100000>' for a
701 This is similar to 'offset' in entries, setting the offset of a section
702 within the image or section containing it. The first byte of the section
703 is normally at offset 0. If 'offset' is not provided, binman sets it to
704 the end of the previous region, or the start of the image's entry area
705 (normally 0) if there is no previous region.
708 This sets the alignment of the image size. For example, to ensure
709 that the image ends on a 512-byte boundary, use 'align-size = <512>'.
710 If 'align-size' is not provided, no alignment is performed.
713 This sets the padding before the image entries. The first entry will
714 be positioned after the padding. This defaults to 0.
717 This sets the padding after the image entries. The padding will be
718 placed after the last entry. This defaults to 0.
721 This specifies the pad byte to use when padding in the image. It
722 defaults to 0. To use 0xff, you would add 'pad-byte = <0xff>'.
725 This specifies the image filename. It defaults to 'image.bin'.
728 This causes binman to reorder the entries as needed to make sure they
729 are in increasing positional order. This can be used when your entry
730 order may not match the positional order. A common situation is where
731 the 'offset' properties are set by CONFIG options, so their ordering is
734 This is a boolean property so needs no value. To enable it, add a
735 line 'sort-by-offset;' to your description.
738 Normally only a single image is generated. To create more than one
739 image, put this property in the binman node. For example, this will
740 create image1.bin containing u-boot.bin, and image2.bin containing
741 both spl/u-boot-spl.bin and u-boot.bin::
759 For x86 machines the ROM offsets start just before 4GB and extend
760 up so that the image finished at the 4GB boundary. This boolean
761 option can be enabled to support this. The image size must be
762 provided so that binman knows when the image should start. For an
763 8MB ROM, the offset of the first entry would be 0xfff80000 with
764 this option, instead of 0 without this option.
767 This property specifies the entry offset of the first entry.
769 For PowerPC mpc85xx based CPU, CONFIG_SYS_TEXT_BASE is the entry
770 offset of the first entry. It can be 0xeff40000 or 0xfff40000 for
771 nor flash boot, 0x201000 for sd boot etc.
773 'end-at-4gb' property is not applicable where CONFIG_SYS_TEXT_BASE +
777 Specifies the default alignment for entries in this section, if they do
778 not specify an alignment. Note that this only applies to top-level entries
779 in the section (direct subentries), not any subentries of those entries.
780 This means that each section must specify its own default alignment, if
783 Examples of the above options can be found in the tests. See the
784 tools/binman/test directory.
786 It is possible to have the same binary appear multiple times in the image,
787 either by using a unit number suffix (u-boot@0, u-boot@1) or by using a
788 different name for each and specifying the type with the 'type' attribute.
791 Sections and hierachical images
792 -------------------------------
794 Sometimes it is convenient to split an image into several pieces, each of which
795 contains its own set of binaries. An example is a flash device where part of
796 the image is read-only and part is read-write. We can set up sections for each
797 of these, and place binaries in them independently. The image is still produced
798 as a single output file.
800 This feature provides a way of creating hierarchical images. For example here
801 is an example image with two copies of U-Boot. One is read-only (ro), intended
802 to be written only in the factory. Another is read-write (rw), so that it can be
803 upgraded in the field. The sizes are fixed so that the ro/rw boundary is known
804 and can be programmed::
822 This image could be placed into a SPI flash chip, with the protection boundary
825 A few special properties are provided for sections:
828 Indicates that this section is read-only. This has no impact on binman's
829 operation, but his property can be read at run time.
832 This string is prepended to all the names of the binaries in the
833 section. In the example above, the 'u-boot' binaries which actually be
834 renamed to 'ro-u-boot' and 'rw-u-boot'. This can be useful to
835 distinguish binaries with otherwise identical names.
841 Image nodes act like sections but also have a few extra properties:
844 Output filename for the image. This defaults to image.bin (or in the
845 case of multiple images <nodename>.bin where <nodename> is the name of
849 Create an image that can be repacked. With this option it is possible
850 to change anything in the image after it is created, including updating
851 the position and size of image components. By default this is not
852 permitted since it is not possibly to know whether this might violate a
853 constraint in the image description. For example, if a section has to
854 increase in size to hold a larger binary, that might cause the section
855 to fall out of its allow region (e.g. read-only portion of flash).
857 Adding this property causes the original offset and size values in the
858 image description to be stored in the FDT and fdtmap.
864 It is possible to ask binman to hash the contents of an entry and write that
865 value back to the device-tree node. For example::
875 Here, a new 'value' property will be written to the 'hash' node containing
876 the hash of the 'u-boot' entry. Only SHA256 is supported at present. Whole
877 sections can be hased if desired, by adding the 'hash' node to the section.
879 The has value can be chcked at runtime by hashing the data actually read and
880 comparing this has to the value in the device tree.
886 Binman automatically replaces 'u-boot' with an expanded version of that, i.e.
887 'u-boot-expanded'. This means that when you write::
895 type = "u-boot-expanded';
898 which in turn expands to::
910 U-Boot's various phase binaries actually comprise two or three pieces.
911 For example, u-boot.bin has the executable followed by a devicetree.
913 With binman we want to be able to update that devicetree with full image
914 information so that it is accessible to the executable. This is tricky
915 if it is not clear where the devicetree starts.
917 The above feature ensures that the devicetree is clearly separated from the
918 U-Boot executable and can be updated separately by binman as needed. It can be
919 disabled with the --no-expanded flag if required.
921 The same applies for u-boot-spl and u-boot-tpl. In those cases, the expansion
922 includes the BSS padding, so for example::
931 type = "u-boot-expanded';
934 which in turn expands to::
949 Of course we should not expand SPL if it has no devicetree. Also if the BSS
950 padding is not needed (because BSS is in RAM as with CONFIG_SPL_SEPARATE_BSS),
951 the 'u-boot-spl-bss-pad' subnode should not be created. The use of the expaned
952 entry type is controlled by the UseExpanded() method. In the SPL case it checks
953 the 'spl-dtb' entry arg, which is 'y' or '1' if SPL has a devicetree.
955 For the BSS case, a 'spl-bss-pad' entry arg controls whether it is present. All
956 entry args are provided by the U-Boot Makefile.
962 Binman support compression for 'blob' entries (those of type 'blob' and
963 derivatives). To enable this for an entry, add a 'compress' property::
966 filename = "datafile";
970 The entry will then contain the compressed data, using the 'lz4' compression
971 algorithm. Currently this is the only one that is supported. The uncompressed
972 size is written to the node in an 'uncomp-size' property, if -u is used.
974 Compression is also supported for sections. In that case the entire section is
975 compressed in one block, including all its contents. This means that accessing
976 an entry from the section required decompressing the entire section. Also, the
977 size of a section indicates the space that it consumes in its parent section
978 (and typically the image). With compression, the section may contain more data,
979 and the uncomp-size property indicates that, as above. The contents of the
980 section is compressed first, before any padding is added. This ensures that the
981 padding itself is not compressed, which would be a waste of time.
984 Automatic .dtsi inclusion
985 -------------------------
987 It is sometimes inconvenient to add a 'binman' node to the .dts file for each
988 board. This can be done by using #include to bring in a common file. Another
989 approach supported by the U-Boot build system is to automatically include
990 a common header. You can then put the binman node (and anything else that is
991 specific to U-Boot, such as u-boot,dm-pre-reloc properies) in that header
994 Binman will search for the following files in arch/<arch>/dts::
996 <dts>-u-boot.dtsi where <dts> is the base name of the .dts file
997 <CONFIG_SYS_SOC>-u-boot.dtsi
998 <CONFIG_SYS_CPU>-u-boot.dtsi
999 <CONFIG_SYS_VENDOR>-u-boot.dtsi
1002 U-Boot will only use the first one that it finds. If you need to include a
1003 more general file you can do that from the more specific file using #include.
1004 If you are having trouble figuring out what is going on, you can use
1005 `DEVICE_TREE_DEBUG=1` with your build::
1007 make DEVICE_TREE_DEBUG=1
1008 scripts/Makefile.lib:334: Automatic .dtsi inclusion: options:
1009 arch/arm/dts/juno-r2-u-boot.dtsi arch/arm/dts/-u-boot.dtsi
1010 arch/arm/dts/armv8-u-boot.dtsi arch/arm/dts/armltd-u-boot.dtsi
1011 arch/arm/dts/u-boot.dtsi ... found: "arch/arm/dts/juno-r2-u-boot.dtsi"
1014 Updating an ELF file
1015 ====================
1017 For the EFI app, where U-Boot is loaded from UEFI and runs as an app, there is
1018 no way to update the devicetree after U-Boot is built. Normally this works by
1019 creating a new u-boot.dtb.out with he updated devicetree, which is automatically
1020 built into the output image. With ELF this is not possible since the ELF is
1021 not part of an image, just a stand-along file. We must create an updated ELF
1022 file with the new devicetree.
1024 This is handled by the --update-fdt-in-elf option. It takes four arguments,
1027 infile - filename of input ELF file, e.g. 'u-boot's
1028 outfile - filename of output ELF file, e.g. 'u-boot.out'
1029 begin_sym - symbol at the start of the embedded devicetree, e.g.
1031 end_sym - symbol at the start of the embedded devicetree, e.g.
1034 When this flag is used, U-Boot does all the normal packaging, but as an
1035 additional step, it creates a new ELF file with the new devicetree embedded in
1038 If logging is enabled you will see a message like this::
1040 Updating file 'u-boot' with data length 0x400a (16394) between symbols
1041 '__dtb_dt_begin' and '__dtb_dt_end'
1043 There must be enough space for the updated devicetree. If not, an error like
1044 the following is produced::
1046 ValueError: Not enough space in 'u-boot' for data length 0x400a (16394);
1047 size is 0x1744 (5956)
1053 For details on the various entry types supported by binman and how to use them,
1054 see entries.rst which is generated from the source code using:
1056 binman entry-docs >tools/binman/entries.rst
1070 It is possible to list the entries in an existing firmware image created by
1071 binman, provided that there is an 'fdtmap' entry in the image. For example::
1073 $ binman ls -i image.bin
1074 Name Image-pos Size Entry-type Offset Uncomp-size
1075 ----------------------------------------------------------------------
1076 main-section c00 section 0
1078 section 5fc section 4
1080 u-boot 138 4 u-boot 38
1081 u-boot-dtb 180 108 u-boot-dtb 80 3b5
1082 u-boot-dtb 500 1ff u-boot-dtb 400 3b5
1083 fdtmap 6fc 381 fdtmap 6fc
1084 image-header bf8 8 image-header bf8
1086 This shows the hierarchy of the image, the position, size and type of each
1087 entry, the offset of each entry within its parent and the uncompressed size if
1088 the entry is compressed.
1090 It is also possible to list just some files in an image, e.g.::
1092 $ binman ls -i image.bin section/cbfs
1093 Name Image-pos Size Entry-type Offset Uncomp-size
1094 --------------------------------------------------------------------
1096 u-boot 138 4 u-boot 38
1097 u-boot-dtb 180 108 u-boot-dtb 80 3b5
1101 $ binman ls -i image.bin "*cb*" "*head*"
1102 Name Image-pos Size Entry-type Offset Uncomp-size
1103 ----------------------------------------------------------------------
1105 u-boot 138 4 u-boot 38
1106 u-boot-dtb 180 108 u-boot-dtb 80 3b5
1107 image-header bf8 8 image-header bf8
1109 If an older version of binman is used to list images created by a newer one, it
1110 is possible that it will contain entry types that are not supported. These still
1111 show with the correct type, but binman just sees them as blobs (plain binary
1112 data). Any special features of that etype are not supported by the old binman.
1115 Extracting files from images
1116 ----------------------------
1118 You can extract files from an existing firmware image created by binman,
1119 provided that there is an 'fdtmap' entry in the image. For example::
1121 $ binman extract -i image.bin section/cbfs/u-boot
1123 which will write the uncompressed contents of that entry to the file 'u-boot' in
1124 the current directory. You can also extract to a particular file, in this case
1127 $ binman extract -i image.bin section/cbfs/u-boot -f u-boot.bin
1129 It is possible to extract all files into a destination directory, which will
1130 put files in subdirectories matching the entry hierarchy::
1132 $ binman extract -i image.bin -O outdir
1134 or just a selection::
1136 $ binman extract -i image.bin "*u-boot*" -O outdir
1138 Some entry types have alternative formats, for example fdtmap which allows
1139 extracted just the devicetree binary without the fdtmap header::
1141 $ binman extract -i /tmp/b/odroid-c4/image.bin -f out.dtb -F fdt fdtmap
1144 // magic: 0xd00dfeed
1145 // totalsize: 0x8ab (2219)
1146 // off_dt_struct: 0x38
1147 // off_dt_strings: 0x82c
1148 // off_mem_rsvmap: 0x28
1150 // last_comp_version: 2
1151 // boot_cpuid_phys: 0x0
1152 // size_dt_strings: 0x7f
1153 // size_dt_struct: 0x7f4
1156 image-node = "binman";
1157 image-pos = <0x00000000>;
1158 size = <0x0011162b>;
1161 Use `-F list` to see what alternative formats are available::
1163 $ binman extract -i /tmp/b/odroid-c4/image.bin -F list
1164 Flag (-F) Entry type Description
1165 fdt fdtmap Extract the devicetree blob from the fdtmap
1168 Replacing files in an image
1169 ---------------------------
1171 You can replace files in an existing firmware image created by binman, provided
1172 that there is an 'fdtmap' entry in the image. For example::
1174 $ binman replace -i image.bin section/cbfs/u-boot
1176 which will write the contents of the file 'u-boot' from the current directory
1177 to the that entry, compressing if necessary. If the entry size changes, you must
1178 add the 'allow-repack' property to the original image before generating it (see
1179 above), otherwise you will get an error.
1181 You can also use a particular file, in this case u-boot.bin::
1183 $ binman replace -i image.bin section/cbfs/u-boot -f u-boot.bin
1185 It is possible to replace all files from a source directory which uses the same
1186 hierarchy as the entries::
1188 $ binman replace -i image.bin -I indir
1190 Files that are missing will generate a warning.
1192 You can also replace just a selection of entries::
1194 $ binman replace -i image.bin "*u-boot*" -I indir
1200 Binman normally operates silently unless there is an error, in which case it
1201 just displays the error. The -D/--debug option can be used to create a full
1202 backtrace when errors occur. You can use BINMAN_DEBUG=1 when building to select
1205 Internally binman logs some output while it is running. This can be displayed
1206 by increasing the -v/--verbosity from the default of 1:
1210 2: notices (important messages)
1211 3: info about major operations
1212 4: detailed information about each operation
1213 5: debug (all output)
1215 You can use BINMAN_VERBOSE=5 (for example) when building to select this.
1221 `Bintool` is the name binman gives to a binary tool which it uses to create and
1222 manipulate binaries that binman cannot handle itself. Bintools are often
1223 necessary since Binman only supports a subset of the available file formats
1226 Many SoC vendors invent ways to load code into their SoC using new file formats,
1227 sometimes changing the format with successive SoC generations. Sometimes the
1228 tool is available as Open Source. Sometimes it is a pre-compiled binary that
1229 must be downloaded from the vendor's website. Sometimes it is available in
1230 source form but difficult or slow to build.
1232 Even for images that use bintools, binman still assembles the image from its
1233 image description. It may handle parts of the image natively and part with
1236 Binman relies on these tools so provides various features to manage them:
1238 - Determining whether the tool is currently installed
1239 - Downloading or building the tool
1240 - Determining the version of the tool that is installed
1241 - Deciding which tools are needed to build an image
1243 The Bintool class is an interface to the tool, a thin level of abstration, using
1244 Python functions to run the tool for each purpose (e.g. creating a new
1245 structure, adding a file to an existing structure) rather than just lists of
1248 As with external blobs, bintools (which are like 'external' tools) can be
1249 missing. When building an image requires a bintool and it is not installed,
1250 binman detects this and reports the problem, but continues to build an image.
1251 This is useful in CI systems which want to check that everything is correct but
1252 don't have access to the bintools.
1254 To make this work, all calls to bintools (e.g. with Bintool.run_cmd()) must cope
1255 with the tool being missing, i.e. when None is returned, by:
1257 - Calling self.record_missing_bintool()
1258 - Setting up some fake contents so binman can continue
1260 Of course the image will not work, but binman reports which bintools are needed
1261 and also provide a way to fetch them.
1263 To see the available bintools, use::
1267 To fetch tools which are missing, use::
1269 binman tool --fetch missing
1271 You can also use `--fetch all` to fetch all tools or `--fetch <tool>` to fetch
1272 a particular tool. Some tools are built from source code, in which case you will
1273 need to have at least the `build-essential` and `git` packages installed.
1275 Bintool Documentation
1276 =====================
1278 To provide details on the various bintools supported by binman, bintools.rst is
1279 generated from the source code using:
1281 binman bintool-docs >tools/binman/bintools.rst
1292 Order of image creation
1293 -----------------------
1295 Image creation proceeds in the following order, for each entry in the image.
1297 1. AddMissingProperties() - binman can add calculated values to the device
1298 tree as part of its processing, for example the offset and size of each
1299 entry. This method adds any properties associated with this, expanding the
1300 device tree as needed. These properties can have placeholder values which are
1301 set later by SetCalculatedProperties(). By that stage the size of sections
1302 cannot be changed (since it would cause the images to need to be repacked),
1303 but the correct values can be inserted.
1305 2. ProcessFdt() - process the device tree information as required by the
1306 particular entry. This may involve adding or deleting properties. If the
1307 processing is complete, this method should return True. If the processing
1308 cannot complete because it needs the ProcessFdt() method of another entry to
1309 run first, this method should return False, in which case it will be called
1312 3. GetEntryContents() - the contents of each entry are obtained, normally by
1313 reading from a file. This calls the Entry.ObtainContents() to read the
1314 contents. The default version of Entry.ObtainContents() calls
1315 Entry.GetDefaultFilename() and then reads that file. So a common mechanism
1316 to select a file to read is to override that function in the subclass. The
1317 functions must return True when they have read the contents. Binman will
1318 retry calling the functions a few times if False is returned, allowing
1319 dependencies between the contents of different entries.
1321 4. GetEntryOffsets() - calls Entry.GetOffsets() for each entry. This can
1322 return a dict containing entries that need updating. The key should be the
1323 entry name and the value is a tuple (offset, size). This allows an entry to
1324 provide the offset and size for other entries. The default implementation
1325 of GetEntryOffsets() returns {}.
1327 5. PackEntries() - calls Entry.Pack() which figures out the offset and
1328 size of an entry. The 'current' image offset is passed in, and the function
1329 returns the offset immediately after the entry being packed. The default
1330 implementation of Pack() is usually sufficient.
1332 Note: for sections, this also checks that the entries do not overlap, nor extend
1333 outside the section. If the section does not have a defined size, the size is
1334 set large enough to hold all the entries.
1336 6. SetImagePos() - sets the image position of every entry. This is the absolute
1337 position 'image-pos', as opposed to 'offset' which is relative to the containing
1338 section. This must be done after all offsets are known, which is why it is quite
1339 late in the ordering.
1341 7. SetCalculatedProperties() - update any calculated properties in the device
1342 tree. This sets the correct 'offset' and 'size' vaues, for example.
1344 8. ProcessEntryContents() - this calls Entry.ProcessContents() on each entry.
1345 The default implementatoin does nothing. This can be overriden to adjust the
1346 contents of an entry in some way. For example, it would be possible to create
1347 an entry containing a hash of the contents of some other entries. At this
1348 stage the offset and size of entries should not be adjusted unless absolutely
1349 necessary, since it requires a repack (going back to PackEntries()).
1351 9. ResetForPack() - if the ProcessEntryContents() step failed, in that an entry
1352 has changed its size, then there is no alternative but to go back to step 5 and
1353 try again, repacking the entries with the updated size. ResetForPack() removes
1354 the fixed offset/size values added by binman, so that the packing can start from
1357 10. WriteSymbols() - write the value of symbols into the U-Boot SPL binary.
1358 See 'Access to binman entry offsets at run time' below for a description of
1359 what happens in this stage.
1361 11. BuildImage() - builds the image and writes it to a file
1363 12. WriteMap() - writes a text file containing a map of the image. This is the
1370 Binman can make use of external command-line tools to handle processing of
1371 entry contents or to generate entry contents. These tools are executed using
1372 the 'tools' module's Run() method. The tools generally must exist on the PATH,
1373 but the --toolpath option can be used to specify additional search paths to
1374 use. This option can be specified multiple times to add more than one path.
1376 For some compile tools binman will use the versions specified by commonly-used
1377 environment variables like CC and HOSTCC for the C compiler, based on whether
1378 the tool's output will be used for the target or for the host machine. If those
1379 aren't given, it will also try to derive target-specific versions from the
1380 CROSS_COMPILE environment variable during a cross-compilation.
1382 If the tool is not available in the path you can use BINMAN_TOOLPATHS to specify
1383 a space-separated list of paths to search, e.g.::
1385 BINMAN_TOOLPATHS="/tools/g12a /tools/tegra" binman ...
1391 Binary blobs, even if the source code is available, complicate building
1392 firmware. The instructions can involve multiple steps and the binaries may be
1393 hard to build or obtain. Binman at least provides a unified description of how
1394 to build the final image, no matter what steps are needed to get there.
1396 Binman also provides a `blob-ext` entry type that pulls in a binary blob from an
1397 external file. If the file is missing, binman can optionally complete the build
1398 and just report a warning. Use the `-M/--allow-missing` option to enble this.
1399 This is useful in CI systems which want to check that everything is correct but
1400 don't have access to the blobs.
1402 If the blobs are in a different directory, you can specify this with the `-I`
1405 For U-Boot, you can use set the BINMAN_INDIRS environment variable to provide a
1406 space-separated list of directories to search for binary blobs::
1408 BINMAN_INDIRS="odroid-c4/fip/g12a \
1409 odroid-c4/build/board/hardkernel/odroidc4/firmware \
1410 odroid-c4/build/scp_task" binman ...
1415 Binman is a critical tool and is designed to be very testable. Entry
1416 implementations target 100% test coverage. Run 'binman test -T' to check this.
1418 To enable Python test coverage on Debian-type distributions (e.g. Ubuntu)::
1420 $ sudo apt-get install python-coverage python3-coverage python-pytest
1426 This section provides some guidance for some of the less obvious error messages
1430 Expected __bss_size symbol
1431 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1435 binman: Node '/binman/u-boot-spl-ddr/u-boot-spl/u-boot-spl-bss-pad':
1436 Expected __bss_size symbol in spl/u-boot-spl
1438 This indicates that binman needs the `__bss_size` symbol to be defined in the
1439 SPL binary, where `spl/u-boot-spl` is the ELF file containing the symbols. The
1440 symbol tells binman the size of the BSS region, in bytes. It needs this to be
1441 able to pad the image so that the following entries do not overlap the BSS,
1442 which would cause them to be overwritte by variable access in SPL.
1444 This symbols is normally defined in the linker script, immediately after
1445 _bss_start and __bss_end are defined, like this::
1447 __bss_size = __bss_end - __bss_start;
1449 You may need to add it to your linker script if you get this error.
1455 Binman tries to run tests concurrently. This means that the tests make use of
1456 all available CPUs to run.
1460 $ sudo apt-get install python-subunit python3-subunit
1462 Use '-P 1' to disable this. It is automatically disabled when code coverage is
1463 being used (-T) since they are incompatible.
1469 Sometimes when debugging tests it is useful to keep the input and output
1470 directories so they can be examined later. Use -X or --test-preserve-dirs for
1474 Running tests on non-x86 architectures
1475 --------------------------------------
1477 Binman's tests have been written under the assumption that they'll be run on a
1478 x86-like host and there hasn't been an attempt to make them portable yet.
1479 However, it's possible to run the tests by cross-compiling to x86.
1481 To install an x86 cross-compiler on Debian-type distributions (e.g. Ubuntu)::
1483 $ sudo apt-get install gcc-x86-64-linux-gnu
1485 Then, you can run the tests under cross-compilation::
1487 $ CROSS_COMPILE=x86_64-linux-gnu- binman test -T
1489 You can also use gcc-i686-linux-gnu similar to the above.
1492 Writing new entries and debugging
1493 ---------------------------------
1495 The behaviour of entries is defined by the Entry class. All other entries are
1496 a subclass of this. An important subclass is Entry_blob which takes binary
1497 data from a file and places it in the entry. In fact most entry types are
1498 subclasses of Entry_blob.
1500 Each entry type is a separate file in the tools/binman/etype directory. Each
1501 file contains a class called Entry_<type> where <type> is the entry type.
1502 New entry types can be supported by adding new files in that directory.
1503 These will automatically be detected by binman when needed.
1505 Entry properties are documented in entry.py. The entry subclasses are free
1506 to change the values of properties to support special behaviour. For example,
1507 when Entry_blob loads a file, it sets content_size to the size of the file.
1508 Entry classes can adjust other entries. For example, an entry that knows
1509 where other entries should be positioned can set up those entries' offsets
1510 so they don't need to be set in the binman decription. It can also adjust
1513 Most of the time such essoteric behaviour is not needed, but it can be
1514 essential for complex images.
1516 If you need to specify a particular device-tree compiler to use, you can define
1517 the DTC environment variable. This can be useful when the system dtc is too
1520 To enable a full backtrace and other debugging features in binman, pass
1521 BINMAN_DEBUG=1 to your build::
1523 make qemu-x86_defconfig
1526 To enable verbose logging from binman, base BINMAN_VERBOSE to your build, which
1527 adds a -v<level> option to the call to binman::
1529 make qemu-x86_defconfig
1530 make BINMAN_VERBOSE=5
1533 Building sections in parallel
1534 -----------------------------
1536 By default binman uses multiprocessing to speed up compilation of large images.
1537 This works at a section level, with one thread for each entry in the section.
1538 This can speed things up if the entries are large and use compression.
1540 This feature can be disabled with the '-T' flag, which defaults to a suitable
1541 value for your machine. This depends on the Python version, e.g on v3.8 it uses
1542 12 threads on an 8-core machine. See ConcurrentFutures_ for more details.
1544 The special value -T0 selects single-threaded mode, useful for debugging during
1545 development, since dealing with exceptions and problems in threads is more
1546 difficult. This avoids any use of ThreadPoolExecutor.
1549 Collecting data for an entry type
1550 ---------------------------------
1552 Some entry types deal with data obtained from others. For example,
1553 `Entry_mkimage` calls the `mkimage` tool with data from its subnodes::
1556 args = "-n test -T script";
1565 This shows mkimage being passed a file consisting of SPL and U-Boot proper. It
1566 is created by calling `Entry.collect_contents_to_file()`. Note that in this
1567 case, the data is passed to mkimage for processing but does not appear
1568 separately in the image. It may not appear at all, depending on what mkimage
1569 does. The contents of the `mkimage` entry are entirely dependent on the
1570 processing done by the entry, with the provided subnodes (`u-boot-spl` and
1571 `u-boot`) simply providing the input data for that processing.
1573 Note that `Entry.collect_contents_to_file()` simply concatenates the data from
1574 the different entries together, with no control over alignment, etc. Another
1575 approach is to subclass `Entry_section` so that those features become available,
1576 such as `size` and `pad-byte`. Then the contents of the entry can be obtained by
1577 calling `super().BuildSectionData()` in the entry's BuildSectionData()
1578 implementation to get the input data, then write it to a file and process it
1581 There are other ways to obtain data also, depending on the situation. If the
1582 entry type is simply signing data which exists elsewhere in the image, then
1583 you can use `Entry_collection` as a base class. It lets you use a property
1584 called `content` which lists the entries containing data to be processed. This
1585 is used by `Entry_vblock`, for example::
1591 content = <&u_boot &dtb>;
1592 keyblock = "firmware.keyblock";
1593 signprivate = "firmware_data_key.vbprivk";
1595 kernelkey = "kernel_subkey.vbpubk";
1596 preamble-flags = <1>;
1602 which shows an image containing `u-boot` and `u-boot-dtb`, with the `vblock`
1603 image collecting their contents to produce input for its signing process,
1604 without affecting those entries, which still appear in the final image
1607 Another example is where an entry type needs several independent pieces of input
1608 to function. For example, `Entry_fip` allows a number of different binary blobs
1609 to be placed in their own individual places in a custom data structure in the
1610 output image. To make that work you can add subnodes for each of them and call
1611 `Entry.Create()` on each subnode, as `Entry_fip` does. Then the data for each
1612 blob can come from any suitable place, such as an `Entry_u_boot` or an
1613 `Entry_blob` or anything else::
1616 fip-hdr-flags = /bits/ 64 <0x123>;
1618 fip-flags = /bits/ 64 <0x123456789abcdef>;
1619 filename = "bl31.bin";
1623 fip-uuid = [fc 65 13 92 4a 5b 11 ec
1624 94 35 ff 2d 1c fc 79 9c];
1628 The `soc-fw` node is a `blob-ext` (i.e. it reads in a named binary file) whereas
1629 `u-boot` is a normal entry type. This works because `Entry_fip` selects the
1630 `blob-ext` entry type if the node name (here `soc-fw`) is recognised as being
1633 When adding new entry types you are encouraged to use subnodes to provide the
1634 data for processing, unless the `content` approach is more suitable. Consider
1635 whether the input entries are contained within (or consumed by) the entry, vs
1636 just being 'referenced' by the entry. In the latter case, the `content` approach
1637 makes more sense. Ad-hoc properties and other methods of obtaining data are
1638 discouraged, since it adds to confusion for users.
1643 Binman takes a lot of inspiration from a Chrome OS tool called
1644 'cros_bundle_firmware', which I wrote some years ago. That tool was based on
1645 a reasonably simple and sound design but has expanded greatly over the
1646 years. In particular its handling of x86 images is convoluted.
1648 Quite a few lessons have been learned which are hopefully applied here.
1654 On the face of it, a tool to create firmware images should be fairly simple:
1655 just find all the input binaries and place them at the right place in the
1656 image. The difficulty comes from the wide variety of input types (simple
1657 flat binaries containing code, packaged data with various headers), packing
1658 requirments (alignment, spacing, device boundaries) and other required
1659 features such as hierarchical images.
1661 The design challenge is to make it easy to create simple images, while
1662 allowing the more complex cases to be supported. For example, for most
1663 images we don't much care exactly where each binary ends up, so we should
1664 not have to specify that unnecessarily.
1666 New entry types should aim to provide simple usage where possible. If new
1667 core features are needed, they can be added in the Entry base class.
1675 - Use of-platdata to make the information available to code that is unable
1676 to use device tree (such as a very small SPL image). For now, limited info is
1677 available via linker symbols
1678 - Allow easy building of images by specifying just the board name
1679 - Support building an image for a board (-b) more completely, with a
1680 configurable build directory
1681 - Detect invalid properties in nodes
1682 - Sort the fdtmap by offset
1683 - Output temporary files to a different directory
1684 - Rationalise the fdt, fdt_util and pylibfdt modules which currently have some
1685 overlapping and confusing functionality
1686 - Update the fdt library to use a better format for Prop.value (the current one
1687 is useful for dtoc but not much else)
1688 - Figure out how to make Fdt support changing the node order, so that
1689 Node.AddSubnode() can support adding a node before another, existing node.
1690 Perhaps it should completely regenerate the flat tree?
1693 Simon Glass <sjg@chromium.org>
1696 .. _ConcurrentFutures: https://docs.python.org/3/library/concurrent.futures.html#concurrent.futures.ThreadPoolExecutor