blob: b3bd36668db3a4cef12a7996dfb51f78180da129 [file] [log] [blame]
David Gibsonc125a182006-02-01 03:05:22 -08001 Booting the Linux/ppc kernel without Open Firmware
2 --------------------------------------------------
3
4
5(c) 2005 Benjamin Herrenschmidt <benh at kernel.crashing.org>,
6 IBM Corp.
7(c) 2005 Becky Bruce <becky.bruce at freescale.com>,
8 Freescale Semiconductor, FSL SOC and 32-bit additions
Vitaly Wool28f9ec32006-11-20 16:32:39 +03009(c) 2006 MontaVista Software, Inc.
10 Flash chip node definition
David Gibsonc125a182006-02-01 03:05:22 -080011
12 May 18, 2005: Rev 0.1 - Initial draft, no chapter III yet.
13
14 May 19, 2005: Rev 0.2 - Add chapter III and bits & pieces here or
15 clarifies the fact that a lot of things are
16 optional, the kernel only requires a very
17 small device tree, though it is encouraged
18 to provide an as complete one as possible.
19
20 May 24, 2005: Rev 0.3 - Precise that DT block has to be in RAM
21 - Misc fixes
22 - Define version 3 and new format version 16
23 for the DT block (version 16 needs kernel
24 patches, will be fwd separately).
25 String block now has a size, and full path
26 is replaced by unit name for more
27 compactness.
28 linux,phandle is made optional, only nodes
29 that are referenced by other nodes need it.
30 "name" property is now automatically
31 deduced from the unit name
32
33 June 1, 2005: Rev 0.4 - Correct confusion between OF_DT_END and
34 OF_DT_END_NODE in structure definition.
35 - Change version 16 format to always align
36 property data to 4 bytes. Since tokens are
37 already aligned, that means no specific
Matt LaPlante5d3f0832006-11-30 05:21:10 +010038 required alignment between property size
David Gibsonc125a182006-02-01 03:05:22 -080039 and property data. The old style variable
40 alignment would make it impossible to do
41 "simple" insertion of properties using
42 memove (thanks Milton for
43 noticing). Updated kernel patch as well
Matt LaPlante5d3f0832006-11-30 05:21:10 +010044 - Correct a few more alignment constraints
David Gibsonc125a182006-02-01 03:05:22 -080045 - Add a chapter about the device-tree
46 compiler and the textural representation of
47 the tree that can be "compiled" by dtc.
48
David Gibsonc125a182006-02-01 03:05:22 -080049 November 21, 2005: Rev 0.5
50 - Additions/generalizations for 32-bit
51 - Changed to reflect the new arch/powerpc
52 structure
53 - Added chapter VI
54
55
56 ToDo:
57 - Add some definitions of interrupt tree (simple/complex)
58 - Add some definitions for pci host bridges
59 - Add some common address format examples
60 - Add definitions for standard properties and "compatible"
61 names for cells that are not already defined by the existing
62 OF spec.
63 - Compare FSL SOC use of PCI to standard and make sure no new
64 node definition required.
65 - Add more information about node definitions for SOC devices
66 that currently have no standard, like the FSL CPM.
67
68
69I - Introduction
70================
71
72During the recent development of the Linux/ppc64 kernel, and more
73specifically, the addition of new platform types outside of the old
74IBM pSeries/iSeries pair, it was decided to enforce some strict rules
75regarding the kernel entry and bootloader <-> kernel interfaces, in
76order to avoid the degeneration that had become the ppc32 kernel entry
77point and the way a new platform should be added to the kernel. The
78legacy iSeries platform breaks those rules as it predates this scheme,
79but no new board support will be accepted in the main tree that
80doesn't follows them properly. In addition, since the advent of the
81arch/powerpc merged architecture for ppc32 and ppc64, new 32-bit
82platforms and 32-bit platforms which move into arch/powerpc will be
83required to use these rules as well.
84
85The main requirement that will be defined in more detail below is
86the presence of a device-tree whose format is defined after Open
87Firmware specification. However, in order to make life easier
88to embedded board vendors, the kernel doesn't require the device-tree
89to represent every device in the system and only requires some nodes
90and properties to be present. This will be described in detail in
91section III, but, for example, the kernel does not require you to
92create a node for every PCI device in the system. It is a requirement
93to have a node for PCI host bridges in order to provide interrupt
94routing informations and memory/IO ranges, among others. It is also
95recommended to define nodes for on chip devices and other busses that
96don't specifically fit in an existing OF specification. This creates a
97great flexibility in the way the kernel can then probe those and match
98drivers to device, without having to hard code all sorts of tables. It
99also makes it more flexible for board vendors to do minor hardware
100upgrades without significantly impacting the kernel code or cluttering
101it with special cases.
102
103
1041) Entry point for arch/powerpc
105-------------------------------
106
107 There is one and one single entry point to the kernel, at the start
108 of the kernel image. That entry point supports two calling
109 conventions:
110
111 a) Boot from Open Firmware. If your firmware is compatible
112 with Open Firmware (IEEE 1275) or provides an OF compatible
113 client interface API (support for "interpret" callback of
114 forth words isn't required), you can enter the kernel with:
115
116 r5 : OF callback pointer as defined by IEEE 1275
117 bindings to powerpc. Only the 32 bit client interface
118 is currently supported
119
120 r3, r4 : address & length of an initrd if any or 0
121
122 The MMU is either on or off; the kernel will run the
123 trampoline located in arch/powerpc/kernel/prom_init.c to
124 extract the device-tree and other information from open
125 firmware and build a flattened device-tree as described
126 in b). prom_init() will then re-enter the kernel using
127 the second method. This trampoline code runs in the
128 context of the firmware, which is supposed to handle all
129 exceptions during that time.
130
131 b) Direct entry with a flattened device-tree block. This entry
132 point is called by a) after the OF trampoline and can also be
133 called directly by a bootloader that does not support the Open
134 Firmware client interface. It is also used by "kexec" to
135 implement "hot" booting of a new kernel from a previous
136 running one. This method is what I will describe in more
137 details in this document, as method a) is simply standard Open
138 Firmware, and thus should be implemented according to the
139 various standard documents defining it and its binding to the
140 PowerPC platform. The entry point definition then becomes:
141
142 r3 : physical pointer to the device-tree block
143 (defined in chapter II) in RAM
144
145 r4 : physical pointer to the kernel itself. This is
146 used by the assembly code to properly disable the MMU
147 in case you are entering the kernel with MMU enabled
148 and a non-1:1 mapping.
149
Matt LaPlante2fe0ae72006-10-03 22:50:39 +0200150 r5 : NULL (as to differentiate with method a)
David Gibsonc125a182006-02-01 03:05:22 -0800151
152 Note about SMP entry: Either your firmware puts your other
153 CPUs in some sleep loop or spin loop in ROM where you can get
154 them out via a soft reset or some other means, in which case
155 you don't need to care, or you'll have to enter the kernel
156 with all CPUs. The way to do that with method b) will be
157 described in a later revision of this document.
158
159
1602) Board support
161----------------
162
16364-bit kernels:
164
165 Board supports (platforms) are not exclusive config options. An
166 arbitrary set of board supports can be built in a single kernel
167 image. The kernel will "know" what set of functions to use for a
168 given platform based on the content of the device-tree. Thus, you
169 should:
170
171 a) add your platform support as a _boolean_ option in
172 arch/powerpc/Kconfig, following the example of PPC_PSERIES,
173 PPC_PMAC and PPC_MAPLE. The later is probably a good
174 example of a board support to start from.
175
176 b) create your main platform file as
177 "arch/powerpc/platforms/myplatform/myboard_setup.c" and add it
178 to the Makefile under the condition of your CONFIG_
179 option. This file will define a structure of type "ppc_md"
180 containing the various callbacks that the generic code will
181 use to get to your platform specific code
182
183 c) Add a reference to your "ppc_md" structure in the
184 "machines" table in arch/powerpc/kernel/setup_64.c if you are
185 a 64-bit platform.
186
187 d) request and get assigned a platform number (see PLATFORM_*
188 constants in include/asm-powerpc/processor.h
189
19032-bit embedded kernels:
191
192 Currently, board support is essentially an exclusive config option.
193 The kernel is configured for a single platform. Part of the reason
194 for this is to keep kernels on embedded systems small and efficient;
195 part of this is due to the fact the code is already that way. In the
196 future, a kernel may support multiple platforms, but only if the
197 platforms feature the same core architectire. A single kernel build
198 cannot support both configurations with Book E and configurations
199 with classic Powerpc architectures.
200
201 32-bit embedded platforms that are moved into arch/powerpc using a
202 flattened device tree should adopt the merged tree practice of
203 setting ppc_md up dynamically, even though the kernel is currently
204 built with support for only a single platform at a time. This allows
205 unification of the setup code, and will make it easier to go to a
206 multiple-platform-support model in the future.
207
208NOTE: I believe the above will be true once Ben's done with the merge
209of the boot sequences.... someone speak up if this is wrong!
210
211 To add a 32-bit embedded platform support, follow the instructions
212 for 64-bit platforms above, with the exception that the Kconfig
213 option should be set up such that the kernel builds exclusively for
214 the platform selected. The processor type for the platform should
215 enable another config option to select the specific board
216 supported.
217
218NOTE: If ben doesn't merge the setup files, may need to change this to
219point to setup_32.c
220
221
222 I will describe later the boot process and various callbacks that
223 your platform should implement.
224
225
226II - The DT block format
227========================
228
229
230This chapter defines the actual format of the flattened device-tree
231passed to the kernel. The actual content of it and kernel requirements
232are described later. You can find example of code manipulating that
233format in various places, including arch/powerpc/kernel/prom_init.c
234which will generate a flattened device-tree from the Open Firmware
235representation, or the fs2dt utility which is part of the kexec tools
236which will generate one from a filesystem representation. It is
237expected that a bootloader like uboot provides a bit more support,
238that will be discussed later as well.
239
240Note: The block has to be in main memory. It has to be accessible in
241both real mode and virtual mode with no mapping other than main
242memory. If you are writing a simple flash bootloader, it should copy
243the block to RAM before passing it to the kernel.
244
245
2461) Header
247---------
248
249 The kernel is entered with r3 pointing to an area of memory that is
Matt LaPlanted6bc8ac2006-10-03 22:54:15 +0200250 roughly described in include/asm-powerpc/prom.h by the structure
David Gibsonc125a182006-02-01 03:05:22 -0800251 boot_param_header:
252
253struct boot_param_header {
254 u32 magic; /* magic word OF_DT_HEADER */
255 u32 totalsize; /* total size of DT block */
256 u32 off_dt_struct; /* offset to structure */
257 u32 off_dt_strings; /* offset to strings */
258 u32 off_mem_rsvmap; /* offset to memory reserve map
259*/
260 u32 version; /* format version */
261 u32 last_comp_version; /* last compatible version */
262
263 /* version 2 fields below */
264 u32 boot_cpuid_phys; /* Which physical CPU id we're
265 booting on */
266 /* version 3 fields below */
267 u32 size_dt_strings; /* size of the strings block */
268};
269
270 Along with the constants:
271
272/* Definitions used by the flattened device tree */
273#define OF_DT_HEADER 0xd00dfeed /* 4: version,
274 4: total size */
275#define OF_DT_BEGIN_NODE 0x1 /* Start node: full name
276*/
277#define OF_DT_END_NODE 0x2 /* End node */
278#define OF_DT_PROP 0x3 /* Property: name off,
279 size, content */
280#define OF_DT_END 0x9
281
282 All values in this header are in big endian format, the various
283 fields in this header are defined more precisely below. All
284 "offset" values are in bytes from the start of the header; that is
285 from the value of r3.
286
287 - magic
288
289 This is a magic value that "marks" the beginning of the
290 device-tree block header. It contains the value 0xd00dfeed and is
291 defined by the constant OF_DT_HEADER
292
293 - totalsize
294
295 This is the total size of the DT block including the header. The
296 "DT" block should enclose all data structures defined in this
297 chapter (who are pointed to by offsets in this header). That is,
298 the device-tree structure, strings, and the memory reserve map.
299
300 - off_dt_struct
301
302 This is an offset from the beginning of the header to the start
303 of the "structure" part the device tree. (see 2) device tree)
304
305 - off_dt_strings
306
307 This is an offset from the beginning of the header to the start
308 of the "strings" part of the device-tree
309
310 - off_mem_rsvmap
311
312 This is an offset from the beginning of the header to the start
313 of the reserved memory map. This map is a list of pairs of 64
314 bit integers. Each pair is a physical address and a size. The
315
316 list is terminated by an entry of size 0. This map provides the
317 kernel with a list of physical memory areas that are "reserved"
318 and thus not to be used for memory allocations, especially during
319 early initialization. The kernel needs to allocate memory during
320 boot for things like un-flattening the device-tree, allocating an
321 MMU hash table, etc... Those allocations must be done in such a
322 way to avoid overriding critical things like, on Open Firmware
323 capable machines, the RTAS instance, or on some pSeries, the TCE
324 tables used for the iommu. Typically, the reserve map should
325 contain _at least_ this DT block itself (header,total_size). If
326 you are passing an initrd to the kernel, you should reserve it as
327 well. You do not need to reserve the kernel image itself. The map
328 should be 64 bit aligned.
329
330 - version
331
332 This is the version of this structure. Version 1 stops
333 here. Version 2 adds an additional field boot_cpuid_phys.
334 Version 3 adds the size of the strings block, allowing the kernel
335 to reallocate it easily at boot and free up the unused flattened
336 structure after expansion. Version 16 introduces a new more
337 "compact" format for the tree itself that is however not backward
338 compatible. You should always generate a structure of the highest
339 version defined at the time of your implementation. Currently
Matt LaPlantefff92892006-10-03 22:47:42 +0200340 that is version 16, unless you explicitly aim at being backward
David Gibsonc125a182006-02-01 03:05:22 -0800341 compatible.
342
343 - last_comp_version
344
345 Last compatible version. This indicates down to what version of
346 the DT block you are backward compatible. For example, version 2
347 is backward compatible with version 1 (that is, a kernel build
348 for version 1 will be able to boot with a version 2 format). You
349 should put a 1 in this field if you generate a device tree of
350 version 1 to 3, or 0x10 if you generate a tree of version 0x10
351 using the new unit name format.
352
353 - boot_cpuid_phys
354
355 This field only exist on version 2 headers. It indicate which
356 physical CPU ID is calling the kernel entry point. This is used,
357 among others, by kexec. If you are on an SMP system, this value
358 should match the content of the "reg" property of the CPU node in
359 the device-tree corresponding to the CPU calling the kernel entry
360 point (see further chapters for more informations on the required
361 device-tree contents)
362
363
364 So the typical layout of a DT block (though the various parts don't
365 need to be in that order) looks like this (addresses go from top to
366 bottom):
367
368
369 ------------------------------
370 r3 -> | struct boot_param_header |
371 ------------------------------
372 | (alignment gap) (*) |
373 ------------------------------
374 | memory reserve map |
375 ------------------------------
376 | (alignment gap) |
377 ------------------------------
378 | |
379 | device-tree structure |
380 | |
381 ------------------------------
382 | (alignment gap) |
383 ------------------------------
384 | |
385 | device-tree strings |
386 | |
387 -----> ------------------------------
388 |
389 |
390 --- (r3 + totalsize)
391
392 (*) The alignment gaps are not necessarily present; their presence
393 and size are dependent on the various alignment requirements of
394 the individual data blocks.
395
396
3972) Device tree generalities
398---------------------------
399
400This device-tree itself is separated in two different blocks, a
401structure block and a strings block. Both need to be aligned to a 4
402byte boundary.
403
404First, let's quickly describe the device-tree concept before detailing
405the storage format. This chapter does _not_ describe the detail of the
406required types of nodes & properties for the kernel, this is done
407later in chapter III.
408
409The device-tree layout is strongly inherited from the definition of
410the Open Firmware IEEE 1275 device-tree. It's basically a tree of
411nodes, each node having two or more named properties. A property can
412have a value or not.
413
414It is a tree, so each node has one and only one parent except for the
415root node who has no parent.
416
417A node has 2 names. The actual node name is generally contained in a
418property of type "name" in the node property list whose value is a
419zero terminated string and is mandatory for version 1 to 3 of the
420format definition (as it is in Open Firmware). Version 0x10 makes it
421optional as it can generate it from the unit name defined below.
422
Matt LaPlante2fe0ae72006-10-03 22:50:39 +0200423There is also a "unit name" that is used to differentiate nodes with
David Gibsonc125a182006-02-01 03:05:22 -0800424the same name at the same level, it is usually made of the node
Matt LaPlante2fe0ae72006-10-03 22:50:39 +0200425names, the "@" sign, and a "unit address", which definition is
David Gibsonc125a182006-02-01 03:05:22 -0800426specific to the bus type the node sits on.
427
428The unit name doesn't exist as a property per-se but is included in
429the device-tree structure. It is typically used to represent "path" in
430the device-tree. More details about the actual format of these will be
431below.
432
433The kernel powerpc generic code does not make any formal use of the
434unit address (though some board support code may do) so the only real
435requirement here for the unit address is to ensure uniqueness of
436the node unit name at a given level of the tree. Nodes with no notion
437of address and no possible sibling of the same name (like /memory or
438/cpus) may omit the unit address in the context of this specification,
439or use the "@0" default unit address. The unit name is used to define
440a node "full path", which is the concatenation of all parent node
441unit names separated with "/".
442
443The root node doesn't have a defined name, and isn't required to have
444a name property either if you are using version 3 or earlier of the
445format. It also has no unit address (no @ symbol followed by a unit
446address). The root node unit name is thus an empty string. The full
447path to the root node is "/".
448
449Every node which actually represents an actual device (that is, a node
450which isn't only a virtual "container" for more nodes, like "/cpus"
451is) is also required to have a "device_type" property indicating the
452type of node .
453
454Finally, every node that can be referenced from a property in another
455node is required to have a "linux,phandle" property. Real open
456firmware implementations provide a unique "phandle" value for every
457node that the "prom_init()" trampoline code turns into
458"linux,phandle" properties. However, this is made optional if the
459flattened device tree is used directly. An example of a node
460referencing another node via "phandle" is when laying out the
461interrupt tree which will be described in a further version of this
462document.
463
464This "linux, phandle" property is a 32 bit value that uniquely
465identifies a node. You are free to use whatever values or system of
466values, internal pointers, or whatever to generate these, the only
467requirement is that every node for which you provide that property has
468a unique value for it.
469
470Here is an example of a simple device-tree. In this example, an "o"
471designates a node followed by the node unit name. Properties are
472presented with their name followed by their content. "content"
473represents an ASCII string (zero terminated) value, while <content>
474represents a 32 bit hexadecimal value. The various nodes in this
475example will be discussed in a later chapter. At this point, it is
476only meant to give you a idea of what a device-tree looks like. I have
477purposefully kept the "name" and "linux,phandle" properties which
478aren't necessary in order to give you a better idea of what the tree
479looks like in practice.
480
481 / o device-tree
482 |- name = "device-tree"
483 |- model = "MyBoardName"
484 |- compatible = "MyBoardFamilyName"
485 |- #address-cells = <2>
486 |- #size-cells = <2>
487 |- linux,phandle = <0>
488 |
489 o cpus
490 | | - name = "cpus"
491 | | - linux,phandle = <1>
492 | | - #address-cells = <1>
493 | | - #size-cells = <0>
494 | |
495 | o PowerPC,970@0
496 | |- name = "PowerPC,970"
497 | |- device_type = "cpu"
498 | |- reg = <0>
499 | |- clock-frequency = <5f5e1000>
500 | |- linux,boot-cpu
501 | |- linux,phandle = <2>
502 |
503 o memory@0
504 | |- name = "memory"
505 | |- device_type = "memory"
506 | |- reg = <00000000 00000000 00000000 20000000>
507 | |- linux,phandle = <3>
508 |
509 o chosen
510 |- name = "chosen"
511 |- bootargs = "root=/dev/sda2"
512 |- linux,platform = <00000600>
513 |- linux,phandle = <4>
514
515This tree is almost a minimal tree. It pretty much contains the
516minimal set of required nodes and properties to boot a linux kernel;
517that is, some basic model informations at the root, the CPUs, and the
518physical memory layout. It also includes misc information passed
519through /chosen, like in this example, the platform type (mandatory)
520and the kernel command line arguments (optional).
521
522The /cpus/PowerPC,970@0/linux,boot-cpu property is an example of a
523property without a value. All other properties have a value. The
524significance of the #address-cells and #size-cells properties will be
525explained in chapter IV which defines precisely the required nodes and
526properties and their content.
527
528
5293) Device tree "structure" block
530
531The structure of the device tree is a linearized tree structure. The
532"OF_DT_BEGIN_NODE" token starts a new node, and the "OF_DT_END_NODE"
533ends that node definition. Child nodes are simply defined before
534"OF_DT_END_NODE" (that is nodes within the node). A 'token' is a 32
535bit value. The tree has to be "finished" with a OF_DT_END token
536
537Here's the basic structure of a single node:
538
539 * token OF_DT_BEGIN_NODE (that is 0x00000001)
540 * for version 1 to 3, this is the node full path as a zero
541 terminated string, starting with "/". For version 16 and later,
542 this is the node unit name only (or an empty string for the
543 root node)
544 * [align gap to next 4 bytes boundary]
545 * for each property:
546 * token OF_DT_PROP (that is 0x00000003)
547 * 32 bit value of property value size in bytes (or 0 of no
548 * value)
549 * 32 bit value of offset in string block of property name
550 * property value data if any
551 * [align gap to next 4 bytes boundary]
552 * [child nodes if any]
553 * token OF_DT_END_NODE (that is 0x00000002)
554
Matt LaPlante53cb4722006-10-03 22:55:17 +0200555So the node content can be summarised as a start token, a full path,
556a list of properties, a list of child nodes, and an end token. Every
David Gibsonc125a182006-02-01 03:05:22 -0800557child node is a full node structure itself as defined above.
558
Matt LaPlante53cb4722006-10-03 22:55:17 +02005594) Device tree "strings" block
David Gibsonc125a182006-02-01 03:05:22 -0800560
561In order to save space, property names, which are generally redundant,
562are stored separately in the "strings" block. This block is simply the
563whole bunch of zero terminated strings for all property names
564concatenated together. The device-tree property definitions in the
565structure block will contain offset values from the beginning of the
566strings block.
567
568
569III - Required content of the device tree
570=========================================
571
572WARNING: All "linux,*" properties defined in this document apply only
573to a flattened device-tree. If your platform uses a real
574implementation of Open Firmware or an implementation compatible with
575the Open Firmware client interface, those properties will be created
576by the trampoline code in the kernel's prom_init() file. For example,
577that's where you'll have to add code to detect your board model and
Matt LaPlantea2ffd272006-10-03 22:49:15 +0200578set the platform number. However, when using the flattened device-tree
David Gibsonc125a182006-02-01 03:05:22 -0800579entry point, there is no prom_init() pass, and thus you have to
580provide those properties yourself.
581
582
5831) Note about cells and address representation
584----------------------------------------------
585
586The general rule is documented in the various Open Firmware
587documentations. If you chose to describe a bus with the device-tree
588and there exist an OF bus binding, then you should follow the
589specification. However, the kernel does not require every single
590device or bus to be described by the device tree.
591
592In general, the format of an address for a device is defined by the
593parent bus type, based on the #address-cells and #size-cells
594property. In the absence of such a property, the parent's parent
595values are used, etc... The kernel requires the root node to have
596those properties defining addresses format for devices directly mapped
597on the processor bus.
598
599Those 2 properties define 'cells' for representing an address and a
600size. A "cell" is a 32 bit number. For example, if both contain 2
601like the example tree given above, then an address and a size are both
602composed of 2 cells, and each is a 64 bit number (cells are
603concatenated and expected to be in big endian format). Another example
604is the way Apple firmware defines them, with 2 cells for an address
605and one cell for a size. Most 32-bit implementations should define
606#address-cells and #size-cells to 1, which represents a 32-bit value.
607Some 32-bit processors allow for physical addresses greater than 32
608bits; these processors should define #address-cells as 2.
609
610"reg" properties are always a tuple of the type "address size" where
611the number of cells of address and size is specified by the bus
612#address-cells and #size-cells. When a bus supports various address
613spaces and other flags relative to a given address allocation (like
614prefetchable, etc...) those flags are usually added to the top level
615bits of the physical address. For example, a PCI physical address is
616made of 3 cells, the bottom two containing the actual address itself
617while the top cell contains address space indication, flags, and pci
618bus & device numbers.
619
620For busses that support dynamic allocation, it's the accepted practice
621to then not provide the address in "reg" (keep it 0) though while
622providing a flag indicating the address is dynamically allocated, and
623then, to provide a separate "assigned-addresses" property that
624contains the fully allocated addresses. See the PCI OF bindings for
625details.
626
627In general, a simple bus with no address space bits and no dynamic
628allocation is preferred if it reflects your hardware, as the existing
629kernel address parsing functions will work out of the box. If you
630define a bus type with a more complex address format, including things
631like address space bits, you'll have to add a bus translator to the
632prom_parse.c file of the recent kernels for your bus type.
633
634The "reg" property only defines addresses and sizes (if #size-cells
Matt LaPlante992caac2006-10-03 22:52:05 +0200635is non-0) within a given bus. In order to translate addresses upward
David Gibsonc125a182006-02-01 03:05:22 -0800636(that is into parent bus addresses, and possibly into cpu physical
637addresses), all busses must contain a "ranges" property. If the
638"ranges" property is missing at a given level, it's assumed that
Matt LaPlante992caac2006-10-03 22:52:05 +0200639translation isn't possible. The format of the "ranges" property for a
David Gibsonc125a182006-02-01 03:05:22 -0800640bus is a list of:
641
642 bus address, parent bus address, size
643
644"bus address" is in the format of the bus this bus node is defining,
645that is, for a PCI bridge, it would be a PCI address. Thus, (bus
646address, size) defines a range of addresses for child devices. "parent
647bus address" is in the format of the parent bus of this bus. For
648example, for a PCI host controller, that would be a CPU address. For a
649PCI<->ISA bridge, that would be a PCI address. It defines the base
650address in the parent bus where the beginning of that range is mapped.
651
652For a new 64 bit powerpc board, I recommend either the 2/2 format or
653Apple's 2/1 format which is slightly more compact since sizes usually
654fit in a single 32 bit word. New 32 bit powerpc boards should use a
6551/1 format, unless the processor supports physical addresses greater
656than 32-bits, in which case a 2/1 format is recommended.
657
658
6592) Note about "compatible" properties
660-------------------------------------
661
662These properties are optional, but recommended in devices and the root
663node. The format of a "compatible" property is a list of concatenated
664zero terminated strings. They allow a device to express its
665compatibility with a family of similar devices, in some cases,
666allowing a single driver to match against several devices regardless
667of their actual names.
668
6693) Note about "name" properties
670-------------------------------
671
672While earlier users of Open Firmware like OldWorld macintoshes tended
673to use the actual device name for the "name" property, it's nowadays
674considered a good practice to use a name that is closer to the device
675class (often equal to device_type). For example, nowadays, ethernet
676controllers are named "ethernet", an additional "model" property
677defining precisely the chip type/model, and "compatible" property
678defining the family in case a single driver can driver more than one
679of these chips. However, the kernel doesn't generally put any
680restriction on the "name" property; it is simply considered good
681practice to follow the standard and its evolutions as closely as
682possible.
683
684Note also that the new format version 16 makes the "name" property
685optional. If it's absent for a node, then the node's unit name is then
686used to reconstruct the name. That is, the part of the unit name
687before the "@" sign is used (or the entire unit name if no "@" sign
688is present).
689
6904) Note about node and property names and character set
691-------------------------------------------------------
692
Matt LaPlantea2ffd272006-10-03 22:49:15 +0200693While open firmware provides more flexible usage of 8859-1, this
David Gibsonc125a182006-02-01 03:05:22 -0800694specification enforces more strict rules. Nodes and properties should
695be comprised only of ASCII characters 'a' to 'z', '0' to
696'9', ',', '.', '_', '+', '#', '?', and '-'. Node names additionally
697allow uppercase characters 'A' to 'Z' (property names should be
698lowercase. The fact that vendors like Apple don't respect this rule is
699irrelevant here). Additionally, node and property names should always
700begin with a character in the range 'a' to 'z' (or 'A' to 'Z' for node
701names).
702
703The maximum number of characters for both nodes and property names
704is 31. In the case of node names, this is only the leftmost part of
705a unit name (the pure "name" property), it doesn't include the unit
706address which can extend beyond that limit.
707
708
7095) Required nodes and properties
710--------------------------------
711 These are all that are currently required. However, it is strongly
712 recommended that you expose PCI host bridges as documented in the
713 PCI binding to open firmware, and your interrupt tree as documented
714 in OF interrupt tree specification.
715
716 a) The root node
717
718 The root node requires some properties to be present:
719
720 - model : this is your board name/model
721 - #address-cells : address representation for "root" devices
722 - #size-cells: the size representation for "root" devices
Benjamin Herrenschmidte8222502006-03-28 23:15:54 +1100723 - device_type : This property shouldn't be necessary. However, if
724 you decide to create a device_type for your root node, make sure it
725 is _not_ "chrp" unless your platform is a pSeries or PAPR compliant
726 one for 64-bit, or a CHRP-type machine for 32-bit as this will
727 matched by the kernel this way.
David Gibsonc125a182006-02-01 03:05:22 -0800728
729 Additionally, some recommended properties are:
730
731 - compatible : the board "family" generally finds its way here,
732 for example, if you have 2 board models with a similar layout,
733 that typically get driven by the same platform code in the
734 kernel, you would use a different "model" property but put a
735 value in "compatible". The kernel doesn't directly use that
Matt LaPlante6c28f2c2006-10-03 22:46:31 +0200736 value (see /chosen/linux,platform for how the kernel chooses a
David Gibsonc125a182006-02-01 03:05:22 -0800737 platform type) but it is generally useful.
738
739 The root node is also generally where you add additional properties
740 specific to your board like the serial number if any, that sort of
Matt LaPlante6c28f2c2006-10-03 22:46:31 +0200741 thing. It is recommended that if you add any "custom" property whose
David Gibsonc125a182006-02-01 03:05:22 -0800742 name may clash with standard defined ones, you prefix them with your
743 vendor name and a comma.
744
745 b) The /cpus node
746
747 This node is the parent of all individual CPU nodes. It doesn't
748 have any specific requirements, though it's generally good practice
749 to have at least:
750
751 #address-cells = <00000001>
752 #size-cells = <00000000>
753
754 This defines that the "address" for a CPU is a single cell, and has
755 no meaningful size. This is not necessary but the kernel will assume
756 that format when reading the "reg" properties of a CPU node, see
757 below
758
759 c) The /cpus/* nodes
760
761 So under /cpus, you are supposed to create a node for every CPU on
762 the machine. There is no specific restriction on the name of the
763 CPU, though It's common practice to call it PowerPC,<name>. For
764 example, Apple uses PowerPC,G5 while IBM uses PowerPC,970FX.
765
766 Required properties:
767
768 - device_type : has to be "cpu"
769 - reg : This is the physical cpu number, it's a single 32 bit cell
770 and is also used as-is as the unit number for constructing the
771 unit name in the full path. For example, with 2 CPUs, you would
772 have the full path:
773 /cpus/PowerPC,970FX@0
774 /cpus/PowerPC,970FX@1
775 (unit addresses do not require leading zeroes)
776 - d-cache-line-size : one cell, L1 data cache line size in bytes
777 - i-cache-line-size : one cell, L1 instruction cache line size in
778 bytes
779 - d-cache-size : one cell, size of L1 data cache in bytes
780 - i-cache-size : one cell, size of L1 instruction cache in bytes
781 - linux, boot-cpu : Should be defined if this cpu is the boot cpu.
782
783 Recommended properties:
784
785 - timebase-frequency : a cell indicating the frequency of the
786 timebase in Hz. This is not directly used by the generic code,
787 but you are welcome to copy/paste the pSeries code for setting
788 the kernel timebase/decrementer calibration based on this
789 value.
790 - clock-frequency : a cell indicating the CPU core clock frequency
791 in Hz. A new property will be defined for 64 bit values, but if
792 your frequency is < 4Ghz, one cell is enough. Here as well as
793 for the above, the common code doesn't use that property, but
794 you are welcome to re-use the pSeries or Maple one. A future
795 kernel version might provide a common function for this.
796
797 You are welcome to add any property you find relevant to your board,
798 like some information about the mechanism used to soft-reset the
799 CPUs. For example, Apple puts the GPIO number for CPU soft reset
800 lines in there as a "soft-reset" property since they start secondary
801 CPUs by soft-resetting them.
802
803
804 d) the /memory node(s)
805
806 To define the physical memory layout of your board, you should
807 create one or more memory node(s). You can either create a single
808 node with all memory ranges in its reg property, or you can create
809 several nodes, as you wish. The unit address (@ part) used for the
810 full path is the address of the first range of memory defined by a
811 given node. If you use a single memory node, this will typically be
812 @0.
813
814 Required properties:
815
816 - device_type : has to be "memory"
817 - reg : This property contains all the physical memory ranges of
818 your board. It's a list of addresses/sizes concatenated
819 together, with the number of cells of each defined by the
820 #address-cells and #size-cells of the root node. For example,
Matt LaPlante6c28f2c2006-10-03 22:46:31 +0200821 with both of these properties being 2 like in the example given
David Gibsonc125a182006-02-01 03:05:22 -0800822 earlier, a 970 based machine with 6Gb of RAM could typically
823 have a "reg" property here that looks like:
824
825 00000000 00000000 00000000 80000000
826 00000001 00000000 00000001 00000000
827
828 That is a range starting at 0 of 0x80000000 bytes and a range
829 starting at 0x100000000 and of 0x100000000 bytes. You can see
830 that there is no memory covering the IO hole between 2Gb and
831 4Gb. Some vendors prefer splitting those ranges into smaller
832 segments, but the kernel doesn't care.
833
834 e) The /chosen node
835
836 This node is a bit "special". Normally, that's where open firmware
837 puts some variable environment information, like the arguments, or
838 phandle pointers to nodes like the main interrupt controller, or the
839 default input/output devices.
840
841 This specification makes a few of these mandatory, but also defines
842 some linux-specific properties that would be normally constructed by
843 the prom_init() trampoline when booting with an OF client interface,
844 but that you have to provide yourself when using the flattened format.
845
846 Required properties:
847
848 - linux,platform : This is your platform number as assigned by the
849 architecture maintainers
850
851 Recommended properties:
852
853 - bootargs : This zero-terminated string is passed as the kernel
854 command line
855 - linux,stdout-path : This is the full path to your standard
856 console device if any. Typically, if you have serial devices on
857 your board, you may want to put the full path to the one set as
858 the default console in the firmware here, for the kernel to pick
Matt LaPlante5d3f0832006-11-30 05:21:10 +0100859 it up as its own default console. If you look at the function
David Gibsonc125a182006-02-01 03:05:22 -0800860 set_preferred_console() in arch/ppc64/kernel/setup.c, you'll see
861 that the kernel tries to find out the default console and has
862 knowledge of various types like 8250 serial ports. You may want
863 to extend this function to add your own.
864 - interrupt-controller : This is one cell containing a phandle
865 value that matches the "linux,phandle" property of your main
866 interrupt controller node. May be used for interrupt routing.
867
868
869 Note that u-boot creates and fills in the chosen node for platforms
870 that use it.
871
872 f) the /soc<SOCname> node
873
874 This node is used to represent a system-on-a-chip (SOC) and must be
875 present if the processor is a SOC. The top-level soc node contains
876 information that is global to all devices on the SOC. The node name
877 should contain a unit address for the SOC, which is the base address
878 of the memory-mapped register set for the SOC. The name of an soc
879 node should start with "soc", and the remainder of the name should
880 represent the part number for the soc. For example, the MPC8540's
881 soc node would be called "soc8540".
882
883 Required properties:
884
885 - device_type : Should be "soc"
886 - ranges : Should be defined as specified in 1) to describe the
887 translation of SOC addresses for memory mapped SOC registers.
Becky Bruce7d4b95a2006-02-06 14:26:31 -0600888 - bus-frequency: Contains the bus frequency for the SOC node.
889 Typically, the value of this field is filled in by the boot
890 loader.
891
David Gibsonc125a182006-02-01 03:05:22 -0800892
893 Recommended properties:
894
895 - reg : This property defines the address and size of the
896 memory-mapped registers that are used for the SOC node itself.
897 It does not include the child device registers - these will be
898 defined inside each child node. The address specified in the
899 "reg" property should match the unit address of the SOC node.
900 - #address-cells : Address representation for "soc" devices. The
901 format of this field may vary depending on whether or not the
902 device registers are memory mapped. For memory mapped
903 registers, this field represents the number of cells needed to
904 represent the address of the registers. For SOCs that do not
905 use MMIO, a special address format should be defined that
906 contains enough cells to represent the required information.
907 See 1) above for more details on defining #address-cells.
908 - #size-cells : Size representation for "soc" devices
909 - #interrupt-cells : Defines the width of cells used to represent
910 interrupts. Typically this value is <2>, which includes a
911 32-bit number that represents the interrupt number, and a
912 32-bit number that represents the interrupt sense and level.
913 This field is only needed if the SOC contains an interrupt
914 controller.
915
916 The SOC node may contain child nodes for each SOC device that the
917 platform uses. Nodes should not be created for devices which exist
918 on the SOC but are not used by a particular platform. See chapter VI
919 for more information on how to specify devices that are part of an
920SOC.
921
922 Example SOC node for the MPC8540:
923
924 soc8540@e0000000 {
925 #address-cells = <1>;
926 #size-cells = <1>;
927 #interrupt-cells = <2>;
928 device_type = "soc";
929 ranges = <00000000 e0000000 00100000>
930 reg = <e0000000 00003000>;
Becky Bruce7d4b95a2006-02-06 14:26:31 -0600931 bus-frequency = <0>;
David Gibsonc125a182006-02-01 03:05:22 -0800932 }
933
934
935
936IV - "dtc", the device tree compiler
937====================================
938
939
940dtc source code can be found at
941<http://ozlabs.org/~dgibson/dtc/dtc.tar.gz>
942
943WARNING: This version is still in early development stage; the
944resulting device-tree "blobs" have not yet been validated with the
945kernel. The current generated bloc lacks a useful reserve map (it will
946be fixed to generate an empty one, it's up to the bootloader to fill
947it up) among others. The error handling needs work, bugs are lurking,
948etc...
949
950dtc basically takes a device-tree in a given format and outputs a
951device-tree in another format. The currently supported formats are:
952
953 Input formats:
954 -------------
955
956 - "dtb": "blob" format, that is a flattened device-tree block
957 with
958 header all in a binary blob.
959 - "dts": "source" format. This is a text file containing a
960 "source" for a device-tree. The format is defined later in this
961 chapter.
962 - "fs" format. This is a representation equivalent to the
963 output of /proc/device-tree, that is nodes are directories and
964 properties are files
965
966 Output formats:
967 ---------------
968
969 - "dtb": "blob" format
970 - "dts": "source" format
971 - "asm": assembly language file. This is a file that can be
972 sourced by gas to generate a device-tree "blob". That file can
973 then simply be added to your Makefile. Additionally, the
Matt LaPlante6c28f2c2006-10-03 22:46:31 +0200974 assembly file exports some symbols that can be used.
David Gibsonc125a182006-02-01 03:05:22 -0800975
976
977The syntax of the dtc tool is
978
979 dtc [-I <input-format>] [-O <output-format>]
980 [-o output-filename] [-V output_version] input_filename
981
982
983The "output_version" defines what versio of the "blob" format will be
984generated. Supported versions are 1,2,3 and 16. The default is
985currently version 3 but that may change in the future to version 16.
986
987Additionally, dtc performs various sanity checks on the tree, like the
Matt LaPlante6c28f2c2006-10-03 22:46:31 +0200988uniqueness of linux, phandle properties, validity of strings, etc...
David Gibsonc125a182006-02-01 03:05:22 -0800989
990The format of the .dts "source" file is "C" like, supports C and C++
Matt LaPlante6c28f2c2006-10-03 22:46:31 +0200991style comments.
David Gibsonc125a182006-02-01 03:05:22 -0800992
993/ {
994}
995
996The above is the "device-tree" definition. It's the only statement
997supported currently at the toplevel.
998
999/ {
1000 property1 = "string_value"; /* define a property containing a 0
1001 * terminated string
1002 */
1003
1004 property2 = <1234abcd>; /* define a property containing a
1005 * numerical 32 bits value (hexadecimal)
1006 */
1007
1008 property3 = <12345678 12345678 deadbeef>;
1009 /* define a property containing 3
1010 * numerical 32 bits values (cells) in
1011 * hexadecimal
1012 */
1013 property4 = [0a 0b 0c 0d de ea ad be ef];
1014 /* define a property whose content is
1015 * an arbitrary array of bytes
1016 */
1017
1018 childnode@addresss { /* define a child node named "childnode"
1019 * whose unit name is "childnode at
1020 * address"
1021 */
1022
1023 childprop = "hello\n"; /* define a property "childprop" of
1024 * childnode (in this case, a string)
1025 */
1026 };
1027};
1028
1029Nodes can contain other nodes etc... thus defining the hierarchical
1030structure of the tree.
1031
1032Strings support common escape sequences from C: "\n", "\t", "\r",
1033"\(octal value)", "\x(hex value)".
1034
1035It is also suggested that you pipe your source file through cpp (gcc
1036preprocessor) so you can use #include's, #define for constants, etc...
1037
1038Finally, various options are planned but not yet implemented, like
1039automatic generation of phandles, labels (exported to the asm file so
1040you can point to a property content and change it easily from whatever
1041you link the device-tree with), label or path instead of numeric value
1042in some cells to "point" to a node (replaced by a phandle at compile
1043time), export of reserve map address to the asm file, ability to
1044specify reserve map content at compile time, etc...
1045
1046We may provide a .h include file with common definitions of that
1047proves useful for some properties (like building PCI properties or
1048interrupt maps) though it may be better to add a notion of struct
1049definitions to the compiler...
1050
1051
1052V - Recommendations for a bootloader
1053====================================
1054
1055
1056Here are some various ideas/recommendations that have been proposed
1057while all this has been defined and implemented.
1058
1059 - The bootloader may want to be able to use the device-tree itself
1060 and may want to manipulate it (to add/edit some properties,
1061 like physical memory size or kernel arguments). At this point, 2
1062 choices can be made. Either the bootloader works directly on the
1063 flattened format, or the bootloader has its own internal tree
1064 representation with pointers (similar to the kernel one) and
1065 re-flattens the tree when booting the kernel. The former is a bit
1066 more difficult to edit/modify, the later requires probably a bit
1067 more code to handle the tree structure. Note that the structure
1068 format has been designed so it's relatively easy to "insert"
1069 properties or nodes or delete them by just memmoving things
1070 around. It contains no internal offsets or pointers for this
1071 purpose.
1072
Matt LaPlanted6bc8ac2006-10-03 22:54:15 +02001073 - An example of code for iterating nodes & retrieving properties
David Gibsonc125a182006-02-01 03:05:22 -08001074 directly from the flattened tree format can be found in the kernel
1075 file arch/ppc64/kernel/prom.c, look at scan_flat_dt() function,
Matt LaPlanted6bc8ac2006-10-03 22:54:15 +02001076 its usage in early_init_devtree(), and the corresponding various
David Gibsonc125a182006-02-01 03:05:22 -08001077 early_init_dt_scan_*() callbacks. That code can be re-used in a
1078 GPL bootloader, and as the author of that code, I would be happy
Matt LaPlanted6bc8ac2006-10-03 22:54:15 +02001079 to discuss possible free licencing to any vendor who wishes to
David Gibsonc125a182006-02-01 03:05:22 -08001080 integrate all or part of this code into a non-GPL bootloader.
1081
1082
1083
1084VI - System-on-a-chip devices and nodes
1085=======================================
1086
1087Many companies are now starting to develop system-on-a-chip
1088processors, where the processor core (cpu) and many peripheral devices
1089exist on a single piece of silicon. For these SOCs, an SOC node
1090should be used that defines child nodes for the devices that make
1091up the SOC. While platforms are not required to use this model in
1092order to boot the kernel, it is highly encouraged that all SOC
1093implementations define as complete a flat-device-tree as possible to
1094describe the devices on the SOC. This will allow for the
1095genericization of much of the kernel code.
1096
1097
10981) Defining child nodes of an SOC
1099---------------------------------
1100
1101Each device that is part of an SOC may have its own node entry inside
1102the SOC node. For each device that is included in the SOC, the unit
1103address property represents the address offset for this device's
1104memory-mapped registers in the parent's address space. The parent's
1105address space is defined by the "ranges" property in the top-level soc
1106node. The "reg" property for each node that exists directly under the
1107SOC node should contain the address mapping from the child address space
1108to the parent SOC address space and the size of the device's
1109memory-mapped register file.
1110
1111For many devices that may exist inside an SOC, there are predefined
1112specifications for the format of the device tree node. All SOC child
1113nodes should follow these specifications, except where noted in this
1114document.
1115
1116See appendix A for an example partial SOC node definition for the
1117MPC8540.
1118
1119
11202) Specifying interrupt information for SOC devices
1121---------------------------------------------------
1122
1123Each device that is part of an SOC and which generates interrupts
1124should have the following properties:
1125
1126 - interrupt-parent : contains the phandle of the interrupt
1127 controller which handles interrupts for this device
1128 - interrupts : a list of tuples representing the interrupt
Matt LaPlante5d3f0832006-11-30 05:21:10 +01001129 number and the interrupt sense and level for each interrupt
David Gibsonc125a182006-02-01 03:05:22 -08001130 for this device.
1131
1132This information is used by the kernel to build the interrupt table
1133for the interrupt controllers in the system.
1134
1135Sense and level information should be encoded as follows:
1136
1137 Devices connected to openPIC-compatible controllers should encode
1138 sense and polarity as follows:
1139
Benjamin Herrenschmidt3efbdd12006-08-30 08:58:00 +10001140 0 = low to high edge sensitive type enabled
David Gibsonc125a182006-02-01 03:05:22 -08001141 1 = active low level sensitive type enabled
Benjamin Herrenschmidt3efbdd12006-08-30 08:58:00 +10001142 2 = active high level sensitive type enabled
1143 3 = high to low edge sensitive type enabled
David Gibsonc125a182006-02-01 03:05:22 -08001144
1145 ISA PIC interrupt controllers should adhere to the ISA PIC
1146 encodings listed below:
1147
1148 0 = active low level sensitive type enabled
1149 1 = active high level sensitive type enabled
1150 2 = high to low edge sensitive type enabled
1151 3 = low to high edge sensitive type enabled
1152
1153
1154
11553) Representing devices without a current OF specification
1156----------------------------------------------------------
1157
1158Currently, there are many devices on SOCs that do not have a standard
1159representation pre-defined as part of the open firmware
1160specifications, mainly because the boards that contain these SOCs are
1161not currently booted using open firmware. This section contains
1162descriptions for the SOC devices for which new nodes have been
1163defined; this list will expand as more and more SOC-containing
1164platforms are moved over to use the flattened-device-tree model.
1165
1166 a) MDIO IO device
1167
1168 The MDIO is a bus to which the PHY devices are connected. For each
1169 device that exists on this bus, a child node should be created. See
1170 the definition of the PHY node below for an example of how to define
1171 a PHY.
1172
1173 Required properties:
1174 - reg : Offset and length of the register set for the device
1175 - device_type : Should be "mdio"
1176 - compatible : Should define the compatible device type for the
1177 mdio. Currently, this is most likely to be "gianfar"
1178
1179 Example:
1180
1181 mdio@24520 {
1182 reg = <24520 20>;
Becky Bruce7d4b95a2006-02-06 14:26:31 -06001183 device_type = "mdio";
1184 compatible = "gianfar";
David Gibsonc125a182006-02-01 03:05:22 -08001185
1186 ethernet-phy@0 {
1187 ......
1188 };
1189 };
1190
1191
1192 b) Gianfar-compatible ethernet nodes
1193
1194 Required properties:
1195
1196 - device_type : Should be "network"
1197 - model : Model of the device. Can be "TSEC", "eTSEC", or "FEC"
1198 - compatible : Should be "gianfar"
1199 - reg : Offset and length of the register set for the device
Jon Loeligerf5831652006-08-17 08:42:35 -05001200 - mac-address : List of bytes representing the ethernet address of
David Gibsonc125a182006-02-01 03:05:22 -08001201 this controller
1202 - interrupts : <a b> where a is the interrupt number and b is a
1203 field that represents an encoding of the sense and level
1204 information for the interrupt. This should be encoded based on
1205 the information in section 2) depending on the type of interrupt
1206 controller you have.
1207 - interrupt-parent : the phandle for the interrupt controller that
1208 services interrupts for this device.
1209 - phy-handle : The phandle for the PHY connected to this ethernet
1210 controller.
1211
1212 Example:
1213
1214 ethernet@24000 {
1215 #size-cells = <0>;
1216 device_type = "network";
1217 model = "TSEC";
1218 compatible = "gianfar";
1219 reg = <24000 1000>;
Jon Loeligerf5831652006-08-17 08:42:35 -05001220 mac-address = [ 00 E0 0C 00 73 00 ];
David Gibsonc125a182006-02-01 03:05:22 -08001221 interrupts = <d 3 e 3 12 3>;
1222 interrupt-parent = <40000>;
1223 phy-handle = <2452000>
1224 };
1225
1226
1227
1228 c) PHY nodes
1229
1230 Required properties:
1231
1232 - device_type : Should be "ethernet-phy"
1233 - interrupts : <a b> where a is the interrupt number and b is a
1234 field that represents an encoding of the sense and level
1235 information for the interrupt. This should be encoded based on
1236 the information in section 2) depending on the type of interrupt
1237 controller you have.
1238 - interrupt-parent : the phandle for the interrupt controller that
1239 services interrupts for this device.
1240 - reg : The ID number for the phy, usually a small integer
1241 - linux,phandle : phandle for this node; likely referenced by an
1242 ethernet controller node.
1243
1244
1245 Example:
1246
1247 ethernet-phy@0 {
1248 linux,phandle = <2452000>
1249 interrupt-parent = <40000>;
1250 interrupts = <35 1>;
1251 reg = <0>;
1252 device_type = "ethernet-phy";
1253 };
1254
1255
1256 d) Interrupt controllers
1257
1258 Some SOC devices contain interrupt controllers that are different
1259 from the standard Open PIC specification. The SOC device nodes for
1260 these types of controllers should be specified just like a standard
1261 OpenPIC controller. Sense and level information should be encoded
1262 as specified in section 2) of this chapter for each device that
1263 specifies an interrupt.
1264
1265 Example :
1266
1267 pic@40000 {
1268 linux,phandle = <40000>;
1269 clock-frequency = <0>;
1270 interrupt-controller;
1271 #address-cells = <0>;
1272 reg = <40000 40000>;
1273 built-in;
1274 compatible = "chrp,open-pic";
1275 device_type = "open-pic";
1276 big-endian;
1277 };
1278
1279
1280 e) I2C
1281
1282 Required properties :
1283
1284 - device_type : Should be "i2c"
1285 - reg : Offset and length of the register set for the device
1286
1287 Recommended properties :
1288
1289 - compatible : Should be "fsl-i2c" for parts compatible with
1290 Freescale I2C specifications.
1291 - interrupts : <a b> where a is the interrupt number and b is a
1292 field that represents an encoding of the sense and level
1293 information for the interrupt. This should be encoded based on
1294 the information in section 2) depending on the type of interrupt
1295 controller you have.
1296 - interrupt-parent : the phandle for the interrupt controller that
1297 services interrupts for this device.
1298 - dfsrr : boolean; if defined, indicates that this I2C device has
1299 a digital filter sampling rate register
1300 - fsl5200-clocking : boolean; if defined, indicated that this device
1301 uses the FSL 5200 clocking mechanism.
1302
1303 Example :
1304
1305 i2c@3000 {
1306 interrupt-parent = <40000>;
1307 interrupts = <1b 3>;
1308 reg = <3000 18>;
1309 device_type = "i2c";
1310 compatible = "fsl-i2c";
1311 dfsrr;
1312 };
1313
1314
Becky Brucead71f122006-02-07 13:44:08 -06001315 f) Freescale SOC USB controllers
1316
1317 The device node for a USB controller that is part of a Freescale
1318 SOC is as described in the document "Open Firmware Recommended
1319 Practice : Universal Serial Bus" with the following modifications
1320 and additions :
1321
1322 Required properties :
1323 - compatible : Should be "fsl-usb2-mph" for multi port host usb
1324 controllers, or "fsl-usb2-dr" for dual role usb controllers
1325 - phy_type : For multi port host usb controllers, should be one of
1326 "ulpi", or "serial". For dual role usb controllers, should be
1327 one of "ulpi", "utmi", "utmi_wide", or "serial".
1328 - reg : Offset and length of the register set for the device
1329 - port0 : boolean; if defined, indicates port0 is connected for
1330 fsl-usb2-mph compatible controllers. Either this property or
1331 "port1" (or both) must be defined for "fsl-usb2-mph" compatible
1332 controllers.
1333 - port1 : boolean; if defined, indicates port1 is connected for
1334 fsl-usb2-mph compatible controllers. Either this property or
1335 "port0" (or both) must be defined for "fsl-usb2-mph" compatible
1336 controllers.
1337
1338 Recommended properties :
1339 - interrupts : <a b> where a is the interrupt number and b is a
1340 field that represents an encoding of the sense and level
1341 information for the interrupt. This should be encoded based on
1342 the information in section 2) depending on the type of interrupt
1343 controller you have.
1344 - interrupt-parent : the phandle for the interrupt controller that
1345 services interrupts for this device.
1346
1347 Example multi port host usb controller device node :
1348 usb@22000 {
1349 device_type = "usb";
1350 compatible = "fsl-usb2-mph";
1351 reg = <22000 1000>;
1352 #address-cells = <1>;
1353 #size-cells = <0>;
1354 interrupt-parent = <700>;
1355 interrupts = <27 1>;
1356 phy_type = "ulpi";
1357 port0;
1358 port1;
1359 };
1360
1361 Example dual role usb controller device node :
1362 usb@23000 {
1363 device_type = "usb";
1364 compatible = "fsl-usb2-dr";
1365 reg = <23000 1000>;
1366 #address-cells = <1>;
1367 #size-cells = <0>;
1368 interrupt-parent = <700>;
1369 interrupts = <26 1>;
1370 phy = "ulpi";
1371 };
1372
1373
Kim Phillipsb88a0b12006-03-22 14:39:03 -06001374 g) Freescale SOC SEC Security Engines
1375
1376 Required properties:
1377
1378 - device_type : Should be "crypto"
1379 - model : Model of the device. Should be "SEC1" or "SEC2"
1380 - compatible : Should be "talitos"
1381 - reg : Offset and length of the register set for the device
1382 - interrupts : <a b> where a is the interrupt number and b is a
1383 field that represents an encoding of the sense and level
1384 information for the interrupt. This should be encoded based on
1385 the information in section 2) depending on the type of interrupt
1386 controller you have.
1387 - interrupt-parent : the phandle for the interrupt controller that
1388 services interrupts for this device.
1389 - num-channels : An integer representing the number of channels
1390 available.
1391 - channel-fifo-len : An integer representing the number of
1392 descriptor pointers each channel fetch fifo can hold.
1393 - exec-units-mask : The bitmask representing what execution units
1394 (EUs) are available. It's a single 32 bit cell. EU information
1395 should be encoded following the SEC's Descriptor Header Dword
1396 EU_SEL0 field documentation, i.e. as follows:
1397
1398 bit 0 = reserved - should be 0
1399 bit 1 = set if SEC has the ARC4 EU (AFEU)
1400 bit 2 = set if SEC has the DES/3DES EU (DEU)
1401 bit 3 = set if SEC has the message digest EU (MDEU)
1402 bit 4 = set if SEC has the random number generator EU (RNG)
1403 bit 5 = set if SEC has the public key EU (PKEU)
1404 bit 6 = set if SEC has the AES EU (AESU)
1405 bit 7 = set if SEC has the Kasumi EU (KEU)
1406
1407 bits 8 through 31 are reserved for future SEC EUs.
1408
1409 - descriptor-types-mask : The bitmask representing what descriptors
1410 are available. It's a single 32 bit cell. Descriptor type
1411 information should be encoded following the SEC's Descriptor
1412 Header Dword DESC_TYPE field documentation, i.e. as follows:
1413
1414 bit 0 = set if SEC supports the aesu_ctr_nonsnoop desc. type
1415 bit 1 = set if SEC supports the ipsec_esp descriptor type
1416 bit 2 = set if SEC supports the common_nonsnoop desc. type
1417 bit 3 = set if SEC supports the 802.11i AES ccmp desc. type
1418 bit 4 = set if SEC supports the hmac_snoop_no_afeu desc. type
1419 bit 5 = set if SEC supports the srtp descriptor type
1420 bit 6 = set if SEC supports the non_hmac_snoop_no_afeu desc.type
1421 bit 7 = set if SEC supports the pkeu_assemble descriptor type
1422 bit 8 = set if SEC supports the aesu_key_expand_output desc.type
1423 bit 9 = set if SEC supports the pkeu_ptmul descriptor type
1424 bit 10 = set if SEC supports the common_nonsnoop_afeu desc. type
1425 bit 11 = set if SEC supports the pkeu_ptadd_dbl descriptor type
1426
1427 ..and so on and so forth.
1428
1429 Example:
1430
1431 /* MPC8548E */
1432 crypto@30000 {
1433 device_type = "crypto";
1434 model = "SEC2";
1435 compatible = "talitos";
1436 reg = <30000 10000>;
1437 interrupts = <1d 3>;
1438 interrupt-parent = <40000>;
1439 num-channels = <4>;
Kim Phillipscbdb54d2006-07-03 15:10:14 -05001440 channel-fifo-len = <18>;
Kim Phillipsb88a0b12006-03-22 14:39:03 -06001441 exec-units-mask = <000000fe>;
Kim Phillipscbdb54d2006-07-03 15:10:14 -05001442 descriptor-types-mask = <012b0ebf>;
Kim Phillipsb88a0b12006-03-22 14:39:03 -06001443 };
1444
Li Yang9a1ab882006-10-02 20:08:59 -05001445 h) Board Control and Status (BCSR)
1446
1447 Required properties:
1448
1449 - device_type : Should be "board-control"
1450 - reg : Offset and length of the register set for the device
1451
1452 Example:
1453
1454 bcsr@f8000000 {
1455 device_type = "board-control";
1456 reg = <f8000000 8000>;
1457 };
1458
1459 i) Freescale QUICC Engine module (QE)
1460 This represents qe module that is installed on PowerQUICC II Pro.
1461 Hopefully it will merge backward compatibility with CPM/CPM2.
1462 Basically, it is a bus of devices, that could act more or less
1463 as a complete entity (UCC, USB etc ). All of them should be siblings on
1464 the "root" qe node, using the common properties from there.
1465 The description below applies to the the qe of MPC8360 and
1466 more nodes and properties would be extended in the future.
1467
1468 i) Root QE device
1469
1470 Required properties:
1471 - device_type : should be "qe";
1472 - model : precise model of the QE, Can be "QE", "CPM", or "CPM2"
1473 - reg : offset and length of the device registers.
1474 - bus-frequency : the clock frequency for QUICC Engine.
1475
1476 Recommended properties
1477 - brg-frequency : the internal clock source frequency for baud-rate
1478 generators in Hz.
1479
1480 Example:
1481 qe@e0100000 {
1482 #address-cells = <1>;
1483 #size-cells = <1>;
1484 #interrupt-cells = <2>;
1485 device_type = "qe";
1486 model = "QE";
1487 ranges = <0 e0100000 00100000>;
1488 reg = <e0100000 480>;
1489 brg-frequency = <0>;
1490 bus-frequency = <179A7B00>;
1491 }
1492
1493
1494 ii) SPI (Serial Peripheral Interface)
1495
1496 Required properties:
1497 - device_type : should be "spi".
1498 - compatible : should be "fsl_spi".
1499 - mode : the spi operation mode, it can be "cpu" or "qe".
1500 - reg : Offset and length of the register set for the device
1501 - interrupts : <a b> where a is the interrupt number and b is a
1502 field that represents an encoding of the sense and level
1503 information for the interrupt. This should be encoded based on
1504 the information in section 2) depending on the type of interrupt
1505 controller you have.
1506 - interrupt-parent : the phandle for the interrupt controller that
1507 services interrupts for this device.
1508
1509 Example:
1510 spi@4c0 {
1511 device_type = "spi";
1512 compatible = "fsl_spi";
1513 reg = <4c0 40>;
1514 interrupts = <82 0>;
1515 interrupt-parent = <700>;
1516 mode = "cpu";
1517 };
1518
1519
1520 iii) USB (Universal Serial Bus Controller)
1521
1522 Required properties:
1523 - device_type : should be "usb".
1524 - compatible : could be "qe_udc" or "fhci-hcd".
1525 - mode : the could be "host" or "slave".
1526 - reg : Offset and length of the register set for the device
1527 - interrupts : <a b> where a is the interrupt number and b is a
1528 field that represents an encoding of the sense and level
1529 information for the interrupt. This should be encoded based on
1530 the information in section 2) depending on the type of interrupt
1531 controller you have.
1532 - interrupt-parent : the phandle for the interrupt controller that
1533 services interrupts for this device.
1534
1535 Example(slave):
1536 usb@6c0 {
1537 device_type = "usb";
1538 compatible = "qe_udc";
1539 reg = <6c0 40>;
1540 interrupts = <8b 0>;
1541 interrupt-parent = <700>;
1542 mode = "slave";
1543 };
1544
1545
1546 iv) UCC (Unified Communications Controllers)
1547
1548 Required properties:
1549 - device_type : should be "network", "hldc", "uart", "transparent"
1550 "bisync" or "atm".
1551 - compatible : could be "ucc_geth" or "fsl_atm" and so on.
1552 - model : should be "UCC".
1553 - device-id : the ucc number(1-8), corresponding to UCCx in UM.
1554 - reg : Offset and length of the register set for the device
1555 - interrupts : <a b> where a is the interrupt number and b is a
1556 field that represents an encoding of the sense and level
1557 information for the interrupt. This should be encoded based on
1558 the information in section 2) depending on the type of interrupt
1559 controller you have.
1560 - interrupt-parent : the phandle for the interrupt controller that
1561 services interrupts for this device.
1562 - pio-handle : The phandle for the Parallel I/O port configuration.
1563 - rx-clock : represents the UCC receive clock source.
1564 0x00 : clock source is disabled;
1565 0x1~0x10 : clock source is BRG1~BRG16 respectively;
1566 0x11~0x28: clock source is QE_CLK1~QE_CLK24 respectively.
1567 - tx-clock: represents the UCC transmit clock source;
1568 0x00 : clock source is disabled;
1569 0x1~0x10 : clock source is BRG1~BRG16 respectively;
1570 0x11~0x28: clock source is QE_CLK1~QE_CLK24 respectively.
1571
1572 Required properties for network device_type:
1573 - mac-address : list of bytes representing the ethernet address.
1574 - phy-handle : The phandle for the PHY connected to this controller.
1575
1576 Example:
1577 ucc@2000 {
1578 device_type = "network";
1579 compatible = "ucc_geth";
1580 model = "UCC";
1581 device-id = <1>;
1582 reg = <2000 200>;
1583 interrupts = <a0 0>;
1584 interrupt-parent = <700>;
1585 mac-address = [ 00 04 9f 00 23 23 ];
1586 rx-clock = "none";
1587 tx-clock = "clk9";
1588 phy-handle = <212000>;
1589 pio-handle = <140001>;
1590 };
1591
1592
1593 v) Parallel I/O Ports
1594
1595 This node configures Parallel I/O ports for CPUs with QE support.
1596 The node should reside in the "soc" node of the tree. For each
1597 device that using parallel I/O ports, a child node should be created.
1598 See the definition of the Pin configuration nodes below for more
1599 information.
1600
1601 Required properties:
1602 - device_type : should be "par_io".
1603 - reg : offset to the register set and its length.
1604 - num-ports : number of Parallel I/O ports
1605
1606 Example:
1607 par_io@1400 {
1608 reg = <1400 100>;
1609 #address-cells = <1>;
1610 #size-cells = <0>;
1611 device_type = "par_io";
1612 num-ports = <7>;
1613 ucc_pin@01 {
1614 ......
1615 };
1616
1617
1618 vi) Pin configuration nodes
1619
1620 Required properties:
1621 - linux,phandle : phandle of this node; likely referenced by a QE
1622 device.
1623 - pio-map : array of pin configurations. Each pin is defined by 6
1624 integers. The six numbers are respectively: port, pin, dir,
1625 open_drain, assignment, has_irq.
1626 - port : port number of the pin; 0-6 represent port A-G in UM.
1627 - pin : pin number in the port.
1628 - dir : direction of the pin, should encode as follows:
1629
1630 0 = The pin is disabled
1631 1 = The pin is an output
1632 2 = The pin is an input
1633 3 = The pin is I/O
1634
1635 - open_drain : indicates the pin is normal or wired-OR:
1636
1637 0 = The pin is actively driven as an output
1638 1 = The pin is an open-drain driver. As an output, the pin is
1639 driven active-low, otherwise it is three-stated.
1640
1641 - assignment : function number of the pin according to the Pin Assignment
1642 tables in User Manual. Each pin can have up to 4 possible functions in
1643 QE and two options for CPM.
1644 - has_irq : indicates if the pin is used as source of exteral
1645 interrupts.
1646
1647 Example:
1648 ucc_pin@01 {
1649 linux,phandle = <140001>;
1650 pio-map = <
1651 /* port pin dir open_drain assignment has_irq */
1652 0 3 1 0 1 0 /* TxD0 */
1653 0 4 1 0 1 0 /* TxD1 */
1654 0 5 1 0 1 0 /* TxD2 */
1655 0 6 1 0 1 0 /* TxD3 */
1656 1 6 1 0 3 0 /* TxD4 */
1657 1 7 1 0 1 0 /* TxD5 */
1658 1 9 1 0 2 0 /* TxD6 */
1659 1 a 1 0 2 0 /* TxD7 */
1660 0 9 2 0 1 0 /* RxD0 */
1661 0 a 2 0 1 0 /* RxD1 */
1662 0 b 2 0 1 0 /* RxD2 */
1663 0 c 2 0 1 0 /* RxD3 */
1664 0 d 2 0 1 0 /* RxD4 */
1665 1 1 2 0 2 0 /* RxD5 */
1666 1 0 2 0 2 0 /* RxD6 */
1667 1 4 2 0 2 0 /* RxD7 */
1668 0 7 1 0 1 0 /* TX_EN */
1669 0 8 1 0 1 0 /* TX_ER */
1670 0 f 2 0 1 0 /* RX_DV */
1671 0 10 2 0 1 0 /* RX_ER */
1672 0 0 2 0 1 0 /* RX_CLK */
1673 2 9 1 0 3 0 /* GTX_CLK - CLK10 */
1674 2 8 2 0 1 0>; /* GTX125 - CLK9 */
1675 };
1676
1677 vii) Multi-User RAM (MURAM)
1678
1679 Required properties:
1680 - device_type : should be "muram".
1681 - mode : the could be "host" or "slave".
1682 - ranges : Should be defined as specified in 1) to describe the
1683 translation of MURAM addresses.
1684 - data-only : sub-node which defines the address area under MURAM
1685 bus that can be allocated as data/parameter
1686
1687 Example:
1688
1689 muram@10000 {
1690 device_type = "muram";
1691 ranges = <0 00010000 0000c000>;
1692
1693 data-only@0{
1694 reg = <0 c000>;
1695 };
1696 };
Kim Phillipsb88a0b12006-03-22 14:39:03 -06001697
Vitaly Wool28f9ec32006-11-20 16:32:39 +03001698 g) Flash chip nodes
1699
1700 Flash chips (Memory Technology Devices) are often used for solid state
1701 file systems on embedded devices.
1702
1703 Required properties:
1704
1705 - device_type : has to be "rom"
1706 - compatible : Should specify what this ROM device is compatible with
1707 (i.e. "onenand"). Currently, this is most likely to be "direct-mapped"
1708 (which corresponds to the MTD physmap mapping driver).
1709 - regs : Offset and length of the register set (or memory mapping) for
1710 the device.
1711
1712 Recommended properties :
1713
1714 - bank-width : Width of the flash data bus in bytes. Required
1715 for the NOR flashes (compatible == "direct-mapped" and others) ONLY.
1716 - partitions : Several pairs of 32-bit values where the first value is
1717 partition's offset from the start of the device and the second one is
1718 partition size in bytes with LSB used to signify a read only
1719 partititon (so, the parition size should always be an even number).
1720 - partition-names : The list of concatenated zero terminated strings
1721 representing the partition names.
1722
1723 Example:
1724
1725 flash@ff000000 {
1726 device_type = "rom";
1727 compatible = "direct-mapped";
1728 regs = <ff000000 01000000>;
1729 bank-width = <4>;
1730 partitions = <00000000 00f80000
1731 00f80000 00080001>;
1732 partition-names = "fs\0firmware";
1733 };
1734
David Gibsonc125a182006-02-01 03:05:22 -08001735 More devices will be defined as this spec matures.
1736
1737
1738Appendix A - Sample SOC node for MPC8540
1739========================================
1740
1741Note that the #address-cells and #size-cells for the SoC node
1742in this example have been explicitly listed; these are likely
1743not necessary as they are usually the same as the root node.
1744
1745 soc8540@e0000000 {
1746 #address-cells = <1>;
1747 #size-cells = <1>;
1748 #interrupt-cells = <2>;
1749 device_type = "soc";
1750 ranges = <00000000 e0000000 00100000>
1751 reg = <e0000000 00003000>;
Becky Bruce7d4b95a2006-02-06 14:26:31 -06001752 bus-frequency = <0>;
David Gibsonc125a182006-02-01 03:05:22 -08001753
1754 mdio@24520 {
1755 reg = <24520 20>;
1756 device_type = "mdio";
1757 compatible = "gianfar";
1758
1759 ethernet-phy@0 {
1760 linux,phandle = <2452000>
1761 interrupt-parent = <40000>;
1762 interrupts = <35 1>;
1763 reg = <0>;
1764 device_type = "ethernet-phy";
1765 };
1766
1767 ethernet-phy@1 {
1768 linux,phandle = <2452001>
1769 interrupt-parent = <40000>;
1770 interrupts = <35 1>;
1771 reg = <1>;
1772 device_type = "ethernet-phy";
1773 };
1774
1775 ethernet-phy@3 {
1776 linux,phandle = <2452002>
1777 interrupt-parent = <40000>;
1778 interrupts = <35 1>;
1779 reg = <3>;
1780 device_type = "ethernet-phy";
1781 };
1782
1783 };
1784
1785 ethernet@24000 {
1786 #size-cells = <0>;
1787 device_type = "network";
1788 model = "TSEC";
1789 compatible = "gianfar";
1790 reg = <24000 1000>;
Jon Loeligerf5831652006-08-17 08:42:35 -05001791 mac-address = [ 00 E0 0C 00 73 00 ];
David Gibsonc125a182006-02-01 03:05:22 -08001792 interrupts = <d 3 e 3 12 3>;
1793 interrupt-parent = <40000>;
1794 phy-handle = <2452000>;
1795 };
1796
1797 ethernet@25000 {
1798 #address-cells = <1>;
1799 #size-cells = <0>;
1800 device_type = "network";
1801 model = "TSEC";
1802 compatible = "gianfar";
1803 reg = <25000 1000>;
Jon Loeligerf5831652006-08-17 08:42:35 -05001804 mac-address = [ 00 E0 0C 00 73 01 ];
David Gibsonc125a182006-02-01 03:05:22 -08001805 interrupts = <13 3 14 3 18 3>;
1806 interrupt-parent = <40000>;
1807 phy-handle = <2452001>;
1808 };
1809
1810 ethernet@26000 {
1811 #address-cells = <1>;
1812 #size-cells = <0>;
1813 device_type = "network";
1814 model = "FEC";
1815 compatible = "gianfar";
1816 reg = <26000 1000>;
Jon Loeligerf5831652006-08-17 08:42:35 -05001817 mac-address = [ 00 E0 0C 00 73 02 ];
David Gibsonc125a182006-02-01 03:05:22 -08001818 interrupts = <19 3>;
1819 interrupt-parent = <40000>;
1820 phy-handle = <2452002>;
1821 };
1822
1823 serial@4500 {
1824 device_type = "serial";
1825 compatible = "ns16550";
1826 reg = <4500 100>;
1827 clock-frequency = <0>;
1828 interrupts = <1a 3>;
1829 interrupt-parent = <40000>;
1830 };
1831
1832 pic@40000 {
1833 linux,phandle = <40000>;
1834 clock-frequency = <0>;
1835 interrupt-controller;
1836 #address-cells = <0>;
1837 reg = <40000 40000>;
1838 built-in;
1839 compatible = "chrp,open-pic";
1840 device_type = "open-pic";
1841 big-endian;
1842 };
1843
1844 i2c@3000 {
1845 interrupt-parent = <40000>;
1846 interrupts = <1b 3>;
1847 reg = <3000 18>;
1848 device_type = "i2c";
1849 compatible = "fsl-i2c";
1850 dfsrr;
1851 };
1852
1853 };