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 So far U-Boot has not provided a way to handle creating such images in a
13 general way. Each SoC does what it needs to build an image, often packing or
14 concatenating images in the U-Boot build system.
16 Binman aims to provide a mechanism for building images, from simple
17 SPL + U-Boot combinations, to more complex arrangements with many parts.
23 Binman reads your board's device tree and finds a node which describes the
24 required image layout. It uses this to work out what to place where. The
25 output file normally contains the device tree, so it is in principle possible
26 to read an image and extract its constituent parts.
32 So far binman is pretty simple. It supports binary blobs, such as 'u-boot',
33 'spl' and 'fdt'. It supports empty entries (such as setting to 0xff). It can
34 place entries at a fixed location in the image, or fit them together with
35 suitable padding and alignment. It provides a way to process binaries before
36 they are included, by adding a Python plug-in. The device tree is available
37 to U-Boot at run-time so that the images can be interpreted.
39 Binman does not yet update the device tree with the final location of
40 everything when it is done. A simple C structure could be generated for
41 constrained environments like SPL (using dtoc) but this is also not
44 Binman can also support incorporating filesystems in the image if required.
45 For example x86 platforms may use CBFS in some cases.
47 Binman is intended for use with U-Boot but is designed to be general enough
48 to be useful in other image-packaging situations.
54 Packaging of firmware is quite a different task from building the various
55 parts. In many cases the various binaries which go into the image come from
56 separate build systems. For example, ARM Trusted Firmware is used on ARMv8
57 devices but is not built in the U-Boot tree. If a Linux kernel is included
58 in the firmware image, it is built elsewhere.
60 It is of course possible to add more and more build rules to the U-Boot
61 build system to cover these cases. It can shell out to other Makefiles and
62 build scripts. But it seems better to create a clear divide between building
63 software and packaging it.
65 At present this is handled by manual instructions, different for each board,
66 on how to create images that will boot. By turning these instructions into a
67 standard format, we can support making valid images for any board without
68 manual effort, lots of READMEs, etc.
71 - Each binary can have its own build system and tool chain without creating
72 any dependencies between them
73 - Avoids the need for a single-shot build: individual parts can be updated
74 and brought in as needed
75 - Provides for a standard image description available in the build and at
77 - SoC-specific image-signing tools can be accomodated
78 - Avoids cluttering the U-Boot build system with image-building code
79 - The image description is automatically available at run-time in U-Boot,
80 SPL. It can be made available to other software also
81 - The image description is easily readable (it's a text file in device-tree
82 format) and permits flexible packing of binaries
88 Binman uses the following terms:
90 - image - an output file containing a firmware image
91 - binary - an input binary that goes into the image
97 FIT is U-Boot's official image format. It supports multiple binaries with
98 load / execution addresses, compression. It also supports verification
99 through hashing and RSA signatures.
101 FIT was originally designed to support booting a Linux kernel (with an
102 optional ramdisk) and device tree chosen from various options in the FIT.
103 Now that U-Boot supports configuration via device tree, it is possible to
104 load U-Boot from a FIT, with the device tree chosen by SPL.
106 Binman considers FIT to be one of the binaries it can place in the image.
108 Where possible it is best to put as much as possible in the FIT, with binman
109 used to deal with cases not covered by FIT. Examples include initial
110 execution (since FIT itself does not have an executable header) and dealing
111 with device boundaries, such as the read-only/read-write separation in SPI
114 For U-Boot, binman should not be used to create ad-hoc images in place of
118 Relationship to mkimage
119 -----------------------
121 The mkimage tool provides a means to create a FIT. Traditionally it has
122 needed an image description file: a device tree, like binman, but in a
123 different format. More recently it has started to support a '-f auto' mode
124 which can generate that automatically.
126 More relevant to binman, mkimage also permits creation of many SoC-specific
127 image types. These can be listed by running 'mkimage -T list'. Examples
128 include 'rksd', the Rockchip SD/MMC boot format. The mkimage tool is often
129 called from the U-Boot build system for this reason.
131 Binman considers the output files created by mkimage to be binary blobs
132 which it can place in an image. Binman does not replace the mkimage tool or
133 this purpose. It would be possible in some situations to create a new entry
134 type for the images in mkimage, but this would not add functionality. It
135 seems better to use the mkimage tool to generate binaries and avoid blurring
136 the boundaries between building input files (mkimage) and packaging then
137 into a final image (binman).
140 Example use of binman in U-Boot
141 -------------------------------
143 Binman aims to replace some of the ad-hoc image creation in the U-Boot
146 Consider sunxi. It has the following steps:
148 1. It uses a custom mksunxiboot tool to build an SPL image called
149 sunxi-spl.bin. This should probably move into mkimage.
151 2. It uses mkimage to package U-Boot into a legacy image file (so that it can
152 hold the load and execution address) called u-boot.img.
154 3. It builds a final output image called u-boot-sunxi-with-spl.bin which
155 consists of sunxi-spl.bin, some padding and u-boot.img.
157 Binman is intended to replace the last step. The U-Boot build system builds
158 u-boot.bin and sunxi-spl.bin. Binman can then take over creation of
159 sunxi-spl.bin (by calling mksunxiboot, or hopefully one day mkimage). In any
160 case, it would then create the image from the component parts.
162 This simplifies the U-Boot Makefile somewhat, since various pieces of logic
163 can be replaced by a call to binman.
166 Example use of binman for x86
167 -----------------------------
169 In most cases x86 images have a lot of binary blobs, 'black-box' code
170 provided by Intel which must be run for the platform to work. Typically
171 these blobs are not relocatable and must be placed at fixed areas in the
174 Currently this is handled by ifdtool, which places microcode, FSP, MRC, VGA
175 BIOS, reference code and Intel ME binaries into a u-boot.rom file.
177 Binman is intended to replace all of this, with ifdtool left to handle only
178 the configuration of the Intel-format descriptor.
186 binman -b <board_name>
188 to build an image for a board. The board name is the same name used when
189 configuring U-Boot (e.g. for sandbox_defconfig the board name is 'sandbox').
190 Binman assumes that the input files for the build are in ../b/<board_name>.
192 Or you can specify this explicitly:
194 binman -I <build_path>
196 where <build_path> is the build directory containing the output of the U-Boot
199 (Future work will make this more configurable)
201 In either case, binman picks up the device tree file (u-boot.dtb) and looks
202 for its instructions in the 'binman' node.
204 Binman has a few other options which you can see by running 'binman -h'.
207 Enabling binman for a board
208 ---------------------------
210 At present binman is invoked from a rule in the main Makefile. Typically you
211 will have a rule like:
213 ifneq ($(CONFIG_ARCH_<something>),)
214 u-boot-<your_suffix>.bin: <input_file_1> <input_file_2> checkbinman FORCE
215 $(call if_changed,binman)
218 This assumes that u-boot-<your_suffix>.bin is a target, and is the final file
219 that you need to produce. You can make it a target by adding it to ALL-y
220 either in the main Makefile or in a config.mk file in your arch subdirectory.
222 Once binman is executed it will pick up its instructions from a device-tree
223 file, typically <soc>-u-boot.dtsi, where <soc> is your CONFIG_SYS_SOC value.
224 You can use other, more specific CONFIG options - see 'Automatic .dtsi
228 Image description format
229 ------------------------
231 The binman node is called 'binman'. An example image description is shown
235 filename = "u-boot-sunxi-with-spl.bin";
238 filename = "spl/sunxi-spl.bin";
241 offset = <CONFIG_SPL_PAD_TO>;
246 This requests binman to create an image file called u-boot-sunxi-with-spl.bin
247 consisting of a specially formatted SPL (spl/sunxi-spl.bin, built by the
248 normal U-Boot Makefile), some 0xff padding, and a U-Boot legacy image. The
249 padding comes from the fact that the second binary is placed at
250 CONFIG_SPL_PAD_TO. If that line were omitted then the U-Boot binary would
251 immediately follow the SPL binary.
253 The binman node describes an image. The sub-nodes describe entries in the
254 image. Each entry represents a region within the overall image. The name of
255 the entry (blob, u-boot) tells binman what to put there. For 'blob' we must
256 provide a filename. For 'u-boot', binman knows that this means 'u-boot.bin'.
258 Entries are normally placed into the image sequentially, one after the other.
259 The image size is the total size of all entries. As you can see, you can
260 specify the start offset of an entry using the 'offset' property.
262 Note that due to a device tree requirement, all entries must have a unique
263 name. If you want to put the same binary in the image multiple times, you can
264 use any unique name, with the 'type' property providing the type.
266 The attributes supported for entries are described below.
269 This sets the offset of an entry within the image or section containing
270 it. The first byte of the image is normally at offset 0. If 'offset' is
271 not provided, binman sets it to the end of the previous region, or the
272 start of the image's entry area (normally 0) if there is no previous
276 This sets the alignment of the entry. The entry offset is adjusted
277 so that the entry starts on an aligned boundary within the image. For
278 example 'align = <16>' means that the entry will start on a 16-byte
279 boundary. Alignment shold be a power of 2. If 'align' is not
280 provided, no alignment is performed.
283 This sets the size of the entry. The contents will be padded out to
284 this size. If this is not provided, it will be set to the size of the
288 Padding before the contents of the entry. Normally this is 0, meaning
289 that the contents start at the beginning of the entry. This can be
290 offset the entry contents a little. Defaults to 0.
293 Padding after the contents of the entry. Normally this is 0, meaning
294 that the entry ends at the last byte of content (unless adjusted by
295 other properties). This allows room to be created in the image for
296 this entry to expand later. Defaults to 0.
299 This sets the alignment of the entry size. For example, to ensure
300 that the size of an entry is a multiple of 64 bytes, set this to 64.
301 If 'align-size' is not provided, no alignment is performed.
304 This sets the alignment of the end of an entry. Some entries require
305 that they end on an alignment boundary, regardless of where they
306 start. This does not move the start of the entry, so the contents of
307 the entry will still start at the beginning. But there may be padding
308 at the end. If 'align-end' is not provided, no alignment is performed.
311 For 'blob' types this provides the filename containing the binary to
312 put into the entry. If binman knows about the entry type (like
313 u-boot-bin), then there is no need to specify this.
316 Sets the type of an entry. This defaults to the entry name, but it is
317 possible to use any name, and then add (for example) 'type = "u-boot"'
321 Indicates that the offset of this entry should not be set by placing
322 it immediately after the entry before. Instead, is set by another
323 entry which knows where this entry should go. When this boolean
324 property is present, binman will give an error if another entry does
325 not set the offset (with the GetOffsets() method).
328 This cannot be set on entry (or at least it is ignored if it is), but
329 with the -u option, binman will set it to the absolute image position
330 for each entry. This makes it easy to find out exactly where the entry
331 ended up in the image, regardless of parent sections, etc.
334 Expand the size of this entry to fit available space. This space is only
335 limited by the size of the image/section and the position of the next
338 The attributes supported for images and sections are described below. Several
339 are similar to those for entries.
342 Sets the image size in bytes, for example 'size = <0x100000>' for a
346 This is similar to 'offset' in entries, setting the offset of a section
347 within the image or section containing it. The first byte of the section
348 is normally at offset 0. If 'offset' is not provided, binman sets it to
349 the end of the previous region, or the start of the image's entry area
350 (normally 0) if there is no previous region.
353 This sets the alignment of the image size. For example, to ensure
354 that the image ends on a 512-byte boundary, use 'align-size = <512>'.
355 If 'align-size' is not provided, no alignment is performed.
358 This sets the padding before the image entries. The first entry will
359 be positioned after the padding. This defaults to 0.
362 This sets the padding after the image entries. The padding will be
363 placed after the last entry. This defaults to 0.
366 This specifies the pad byte to use when padding in the image. It
367 defaults to 0. To use 0xff, you would add 'pad-byte = <0xff>'.
370 This specifies the image filename. It defaults to 'image.bin'.
373 This causes binman to reorder the entries as needed to make sure they
374 are in increasing positional order. This can be used when your entry
375 order may not match the positional order. A common situation is where
376 the 'offset' properties are set by CONFIG options, so their ordering is
379 This is a boolean property so needs no value. To enable it, add a
380 line 'sort-by-offset;' to your description.
383 Normally only a single image is generated. To create more than one
384 image, put this property in the binman node. For example, this will
385 create image1.bin containing u-boot.bin, and image2.bin containing
386 both spl/u-boot-spl.bin and u-boot.bin:
404 For x86 machines the ROM offsets start just before 4GB and extend
405 up so that the image finished at the 4GB boundary. This boolean
406 option can be enabled to support this. The image size must be
407 provided so that binman knows when the image should start. For an
408 8MB ROM, the offset of the first entry would be 0xfff80000 with
409 this option, instead of 0 without this option.
412 This property specifies the entry offset of the first entry.
414 For PowerPC mpc85xx based CPU, CONFIG_SYS_TEXT_BASE is the entry
415 offset of the first entry. It can be 0xeff40000 or 0xfff40000 for
416 nor flash boot, 0x201000 for sd boot etc.
418 'end-at-4gb' property is not applicable where CONFIG_SYS_TEXT_BASE +
421 Examples of the above options can be found in the tests. See the
422 tools/binman/test directory.
424 It is possible to have the same binary appear multiple times in the image,
425 either by using a unit number suffix (u-boot@0, u-boot@1) or by using a
426 different name for each and specifying the type with the 'type' attribute.
429 Sections and hierachical images
430 -------------------------------
432 Sometimes it is convenient to split an image into several pieces, each of which
433 contains its own set of binaries. An example is a flash device where part of
434 the image is read-only and part is read-write. We can set up sections for each
435 of these, and place binaries in them independently. The image is still produced
436 as a single output file.
438 This feature provides a way of creating hierarchical images. For example here
439 is an example image with two copies of U-Boot. One is read-only (ro), intended
440 to be written only in the factory. Another is read-write (rw), so that it can be
441 upgraded in the field. The sizes are fixed so that the ro/rw boundary is known
442 and can be programmed:
460 This image could be placed into a SPI flash chip, with the protection boundary
463 A few special properties are provided for sections:
466 Indicates that this section is read-only. This has no impact on binman's
467 operation, but his property can be read at run time.
470 This string is prepended to all the names of the binaries in the
471 section. In the example above, the 'u-boot' binaries which actually be
472 renamed to 'ro-u-boot' and 'rw-u-boot'. This can be useful to
473 distinguish binaries with otherwise identical names.
479 For details on the various entry types supported by binman and how to use them,
480 see README.entries. This is generated from the source code using:
482 binman -E >tools/binman/README.entries
488 It is possible to ask binman to hash the contents of an entry and write that
489 value back to the device-tree node. For example:
499 Here, a new 'value' property will be written to the 'hash' node containing
500 the hash of the 'u-boot' entry. Only SHA256 is supported at present. Whole
501 sections can be hased if desired, by adding the 'hash' node to the section.
503 The has value can be chcked at runtime by hashing the data actually read and
504 comparing this has to the value in the device tree.
507 Order of image creation
508 -----------------------
510 Image creation proceeds in the following order, for each entry in the image.
512 1. AddMissingProperties() - binman can add calculated values to the device
513 tree as part of its processing, for example the offset and size of each
514 entry. This method adds any properties associated with this, expanding the
515 device tree as needed. These properties can have placeholder values which are
516 set later by SetCalculatedProperties(). By that stage the size of sections
517 cannot be changed (since it would cause the images to need to be repacked),
518 but the correct values can be inserted.
520 2. ProcessFdt() - process the device tree information as required by the
521 particular entry. This may involve adding or deleting properties. If the
522 processing is complete, this method should return True. If the processing
523 cannot complete because it needs the ProcessFdt() method of another entry to
524 run first, this method should return False, in which case it will be called
527 3. GetEntryContents() - the contents of each entry are obtained, normally by
528 reading from a file. This calls the Entry.ObtainContents() to read the
529 contents. The default version of Entry.ObtainContents() calls
530 Entry.GetDefaultFilename() and then reads that file. So a common mechanism
531 to select a file to read is to override that function in the subclass. The
532 functions must return True when they have read the contents. Binman will
533 retry calling the functions a few times if False is returned, allowing
534 dependencies between the contents of different entries.
536 4. GetEntryOffsets() - calls Entry.GetOffsets() for each entry. This can
537 return a dict containing entries that need updating. The key should be the
538 entry name and the value is a tuple (offset, size). This allows an entry to
539 provide the offset and size for other entries. The default implementation
540 of GetEntryOffsets() returns {}.
542 5. PackEntries() - calls Entry.Pack() which figures out the offset and
543 size of an entry. The 'current' image offset is passed in, and the function
544 returns the offset immediately after the entry being packed. The default
545 implementation of Pack() is usually sufficient.
547 6. CheckSize() - checks that the contents of all the entries fits within
548 the image size. If the image does not have a defined size, the size is set
549 large enough to hold all the entries.
551 7. CheckEntries() - checks that the entries do not overlap, nor extend
554 8. SetCalculatedProperties() - update any calculated properties in the device
555 tree. This sets the correct 'offset' and 'size' vaues, for example.
557 9. ProcessEntryContents() - this calls Entry.ProcessContents() on each entry.
558 The default implementatoin does nothing. This can be overriden to adjust the
559 contents of an entry in some way. For example, it would be possible to create
560 an entry containing a hash of the contents of some other entries. At this
561 stage the offset and size of entries should not be adjusted.
563 10. WriteSymbols() - write the value of symbols into the U-Boot SPL binary.
564 See 'Access to binman entry offsets at run time' below for a description of
565 what happens in this stage.
567 11. BuildImage() - builds the image and writes it to a file. This is the final
571 Automatic .dtsi inclusion
572 -------------------------
574 It is sometimes inconvenient to add a 'binman' node to the .dts file for each
575 board. This can be done by using #include to bring in a common file. Another
576 approach supported by the U-Boot build system is to automatically include
577 a common header. You can then put the binman node (and anything else that is
578 specific to U-Boot, such as u-boot,dm-pre-reloc properies) in that header
581 Binman will search for the following files in arch/<arch>/dts:
583 <dts>-u-boot.dtsi where <dts> is the base name of the .dts file
584 <CONFIG_SYS_SOC>-u-boot.dtsi
585 <CONFIG_SYS_CPU>-u-boot.dtsi
586 <CONFIG_SYS_VENDOR>-u-boot.dtsi
589 U-Boot will only use the first one that it finds. If you need to include a
590 more general file you can do that from the more specific file using #include.
591 If you are having trouble figuring out what is going on, you can uncomment
592 the 'warning' line in scripts/Makefile.lib to see what it has found:
594 # Uncomment for debugging
595 # This shows all the files that were considered and the one that we chose.
596 # u_boot_dtsi_options_debug = $(u_boot_dtsi_options_raw)
599 Access to binman entry offsets at run time (symbols)
600 ----------------------------------------------------
602 Binman assembles images and determines where each entry is placed in the image.
603 This information may be useful to U-Boot at run time. For example, in SPL it
604 is useful to be able to find the location of U-Boot so that it can be executed
605 when SPL is finished.
607 Binman allows you to declare symbols in the SPL image which are filled in
608 with their correct values during the build. For example:
610 binman_sym_declare(ulong, u_boot_any, offset);
612 declares a ulong value which will be assigned to the offset of any U-Boot
613 image (u-boot.bin, u-boot.img, u-boot-nodtb.bin) that is present in the image.
614 You can access this value with something like:
616 ulong u_boot_offset = binman_sym(ulong, u_boot_any, offset);
618 Thus u_boot_offset will be set to the offset of U-Boot in memory, assuming that
619 the whole image has been loaded, or is available in flash. You can then jump to
620 that address to start U-Boot.
622 At present this feature is only supported in SPL. In principle it is possible
623 to fill in such symbols in U-Boot proper, as well.
626 Access to binman entry offsets at run time (fdt)
627 ------------------------------------------------
629 Binman can update the U-Boot FDT to include the final position and size of
630 each entry in the images it processes. The option to enable this is -u and it
631 causes binman to make sure that the 'offset', 'image-pos' and 'size' properties
632 are set correctly for every entry. Since it is not necessary to specify these in
633 the image definition, binman calculates the final values and writes these to
634 the device tree. These can be used by U-Boot at run-time to find the location
641 Binman support compression for 'blob' entries (those of type 'blob' and
642 derivatives). To enable this for an entry, add a 'compression' property:
645 filename = "datafile";
649 The entry will then contain the compressed data, using the 'lz4' compression
650 algorithm. Currently this is the only one that is supported.
657 The -m option causes binman to output a .map file for each image that it
658 generates. This shows the offset and size of each entry. For example:
661 00000000 00000028 main-section
662 00000000 00000010 section@0
663 00000000 00000004 u-boot
664 00000010 00000010 section@1
665 00000000 00000004 u-boot
667 This shows a hierarchical image with two sections, each with a single entry. The
668 offsets of the sections are absolute hex byte offsets within the image. The
669 offsets of the entries are relative to their respective sections. The size of
670 each entry is also shown, in bytes (hex). The indentation shows the entries
671 nested inside their sections.
674 Passing command-line arguments to entries
675 -----------------------------------------
677 Sometimes it is useful to pass binman the value of an entry property from the
678 command line. For example some entries need access to files and it is not
679 always convenient to put these filenames in the image definition (device tree).
681 The-a option supports this:
687 <prop> is the property to set
688 <value> is the value to set it to
690 Not all properties can be provided this way. Only some entries support it,
691 typically for filenames.
697 Binman is a critical tool and is designed to be very testable. Entry
698 implementations target 100% test coverage. Run 'binman -T' to check this.
700 To enable Python test coverage on Debian-type distributions (e.g. Ubuntu):
702 $ sudo apt-get install python-coverage python-pytest
705 Advanced Features / Technical docs
706 ----------------------------------
708 The behaviour of entries is defined by the Entry class. All other entries are
709 a subclass of this. An important subclass is Entry_blob which takes binary
710 data from a file and places it in the entry. In fact most entry types are
711 subclasses of Entry_blob.
713 Each entry type is a separate file in the tools/binman/etype directory. Each
714 file contains a class called Entry_<type> where <type> is the entry type.
715 New entry types can be supported by adding new files in that directory.
716 These will automatically be detected by binman when needed.
718 Entry properties are documented in entry.py. The entry subclasses are free
719 to change the values of properties to support special behaviour. For example,
720 when Entry_blob loads a file, it sets content_size to the size of the file.
721 Entry classes can adjust other entries. For example, an entry that knows
722 where other entries should be positioned can set up those entries' offsets
723 so they don't need to be set in the binman decription. It can also adjust
726 Most of the time such essoteric behaviour is not needed, but it can be
727 essential for complex images.
729 If you need to specify a particular device-tree compiler to use, you can define
730 the DTC environment variable. This can be useful when the system dtc is too
733 To enable a full backtrace and other debugging features in binman, pass
734 BINMAN_DEBUG=1 to your build:
736 make sandbox_defconfig
743 Binman takes a lot of inspiration from a Chrome OS tool called
744 'cros_bundle_firmware', which I wrote some years ago. That tool was based on
745 a reasonably simple and sound design but has expanded greatly over the
746 years. In particular its handling of x86 images is convoluted.
748 Quite a few lessons have been learned which are hopefully applied here.
754 On the face of it, a tool to create firmware images should be fairly simple:
755 just find all the input binaries and place them at the right place in the
756 image. The difficulty comes from the wide variety of input types (simple
757 flat binaries containing code, packaged data with various headers), packing
758 requirments (alignment, spacing, device boundaries) and other required
759 features such as hierarchical images.
761 The design challenge is to make it easy to create simple images, while
762 allowing the more complex cases to be supported. For example, for most
763 images we don't much care exactly where each binary ends up, so we should
764 not have to specify that unnecessarily.
766 New entry types should aim to provide simple usage where possible. If new
767 core features are needed, they can be added in the Entry base class.
774 - Use of-platdata to make the information available to code that is unable
775 to use device tree (such as a very small SPL image)
776 - Allow easy building of images by specifying just the board name
777 - Produce a full Python binding for libfdt (for upstream). This is nearing
778 completion but some work remains
779 - Add an option to decode an image into the constituent binaries
780 - Support building an image for a board (-b) more completely, with a
781 configurable build directory
782 - Consider making binman work with buildman, although if it is used in the
783 Makefile, this will be automatic
786 Simon Glass <sjg@chromium.org>