1 Booting the Linux/ppc kernel without Open Firmware
2 --------------------------------------------------
4 (c) 2005 Benjamin Herrenschmidt <benh at kernel.crashing.org>,
6 (c) 2005 Becky Bruce <becky.bruce at freescale.com>,
7 Freescale Semiconductor, FSL SOC and 32-bit additions
8 (c) 2006 MontaVista Software, Inc.
9 Flash chip node definition
15 1) Entry point for arch/powerpc
18 II - The DT block format
20 2) Device tree generalities
21 3) Device tree "structure" block
22 4) Device tree "strings" block
24 III - Required content of the device tree
25 1) Note about cells and address representation
26 2) Note about "compatible" properties
27 3) Note about "name" properties
28 4) Note about node and property names and character set
29 5) Required nodes and properties
33 d) the /memory node(s)
35 f) the /soc<SOCname> node
37 IV - "dtc", the device tree compiler
39 V - Recommendations for a bootloader
41 VI - System-on-a-chip devices and nodes
42 1) Defining child nodes of an SOC
43 2) Representing devices without a current OF specification
45 b) Gianfar-compatible ethernet nodes
47 d) Interrupt controllers
49 f) Freescale SOC USB controllers
50 g) Freescale SOC SEC Security Engines
51 h) Board Control and Status (BCSR)
52 i) Freescale QUICC Engine module (QE)
53 j) CFI or JEDEC memory-mapped NOR flash
54 k) Global Utilities Block
55 l) Freescale Communications Processor Module
56 m) Chipselect/Local Bus
57 n) 4xx/Axon EMAC ethernet nodes
59 p) Freescale Synchronous Serial Interface
60 q) USB EHCI controllers
62 VII - Marvell Discovery mv64[345]6x System Controller chips
63 1) The /system-controller node
64 2) Child nodes of /system-controller
65 a) Marvell Discovery MDIO bus
66 b) Marvell Discovery ethernet controller
67 c) Marvell Discovery PHY nodes
68 d) Marvell Discovery SDMA nodes
69 e) Marvell Discovery BRG nodes
70 f) Marvell Discovery CUNIT nodes
71 g) Marvell Discovery MPSCROUTING nodes
72 h) Marvell Discovery MPSCINTR nodes
73 i) Marvell Discovery MPSC nodes
74 j) Marvell Discovery Watch Dog Timer nodes
75 k) Marvell Discovery I2C nodes
76 l) Marvell Discovery PIC (Programmable Interrupt Controller) nodes
77 m) Marvell Discovery MPP (Multipurpose Pins) multiplexing nodes
78 n) Marvell Discovery GPP (General Purpose Pins) nodes
79 o) Marvell Discovery PCI host bridge node
80 p) Marvell Discovery CPU Error nodes
81 q) Marvell Discovery SRAM Controller nodes
82 r) Marvell Discovery PCI Error Handler nodes
83 s) Marvell Discovery Memory Controller nodes
85 VIII - Specifying interrupt information for devices
86 1) interrupts property
87 2) interrupt-parent property
88 3) OpenPIC Interrupt Controllers
89 4) ISA Interrupt Controllers
91 VIII - Specifying GPIO information for devices
93 2) gpio-controller nodes
95 Appendix A - Sample SOC node for MPC8540
101 May 18, 2005: Rev 0.1 - Initial draft, no chapter III yet.
103 May 19, 2005: Rev 0.2 - Add chapter III and bits & pieces here or
104 clarifies the fact that a lot of things are
105 optional, the kernel only requires a very
106 small device tree, though it is encouraged
107 to provide an as complete one as possible.
109 May 24, 2005: Rev 0.3 - Precise that DT block has to be in RAM
111 - Define version 3 and new format version 16
112 for the DT block (version 16 needs kernel
113 patches, will be fwd separately).
114 String block now has a size, and full path
115 is replaced by unit name for more
117 linux,phandle is made optional, only nodes
118 that are referenced by other nodes need it.
119 "name" property is now automatically
120 deduced from the unit name
122 June 1, 2005: Rev 0.4 - Correct confusion between OF_DT_END and
123 OF_DT_END_NODE in structure definition.
124 - Change version 16 format to always align
125 property data to 4 bytes. Since tokens are
126 already aligned, that means no specific
127 required alignment between property size
128 and property data. The old style variable
129 alignment would make it impossible to do
130 "simple" insertion of properties using
131 memmove (thanks Milton for
132 noticing). Updated kernel patch as well
133 - Correct a few more alignment constraints
134 - Add a chapter about the device-tree
135 compiler and the textural representation of
136 the tree that can be "compiled" by dtc.
138 November 21, 2005: Rev 0.5
139 - Additions/generalizations for 32-bit
140 - Changed to reflect the new arch/powerpc
146 - Add some definitions of interrupt tree (simple/complex)
147 - Add some definitions for PCI host bridges
148 - Add some common address format examples
149 - Add definitions for standard properties and "compatible"
150 names for cells that are not already defined by the existing
152 - Compare FSL SOC use of PCI to standard and make sure no new
153 node definition required.
154 - Add more information about node definitions for SOC devices
155 that currently have no standard, like the FSL CPM.
161 During the recent development of the Linux/ppc64 kernel, and more
162 specifically, the addition of new platform types outside of the old
163 IBM pSeries/iSeries pair, it was decided to enforce some strict rules
164 regarding the kernel entry and bootloader <-> kernel interfaces, in
165 order to avoid the degeneration that had become the ppc32 kernel entry
166 point and the way a new platform should be added to the kernel. The
167 legacy iSeries platform breaks those rules as it predates this scheme,
168 but no new board support will be accepted in the main tree that
169 doesn't follows them properly. In addition, since the advent of the
170 arch/powerpc merged architecture for ppc32 and ppc64, new 32-bit
171 platforms and 32-bit platforms which move into arch/powerpc will be
172 required to use these rules as well.
174 The main requirement that will be defined in more detail below is
175 the presence of a device-tree whose format is defined after Open
176 Firmware specification. However, in order to make life easier
177 to embedded board vendors, the kernel doesn't require the device-tree
178 to represent every device in the system and only requires some nodes
179 and properties to be present. This will be described in detail in
180 section III, but, for example, the kernel does not require you to
181 create a node for every PCI device in the system. It is a requirement
182 to have a node for PCI host bridges in order to provide interrupt
183 routing informations and memory/IO ranges, among others. It is also
184 recommended to define nodes for on chip devices and other busses that
185 don't specifically fit in an existing OF specification. This creates a
186 great flexibility in the way the kernel can then probe those and match
187 drivers to device, without having to hard code all sorts of tables. It
188 also makes it more flexible for board vendors to do minor hardware
189 upgrades without significantly impacting the kernel code or cluttering
190 it with special cases.
193 1) Entry point for arch/powerpc
194 -------------------------------
196 There is one and one single entry point to the kernel, at the start
197 of the kernel image. That entry point supports two calling
200 a) Boot from Open Firmware. If your firmware is compatible
201 with Open Firmware (IEEE 1275) or provides an OF compatible
202 client interface API (support for "interpret" callback of
203 forth words isn't required), you can enter the kernel with:
205 r5 : OF callback pointer as defined by IEEE 1275
206 bindings to powerpc. Only the 32-bit client interface
207 is currently supported
209 r3, r4 : address & length of an initrd if any or 0
211 The MMU is either on or off; the kernel will run the
212 trampoline located in arch/powerpc/kernel/prom_init.c to
213 extract the device-tree and other information from open
214 firmware and build a flattened device-tree as described
215 in b). prom_init() will then re-enter the kernel using
216 the second method. This trampoline code runs in the
217 context of the firmware, which is supposed to handle all
218 exceptions during that time.
220 b) Direct entry with a flattened device-tree block. This entry
221 point is called by a) after the OF trampoline and can also be
222 called directly by a bootloader that does not support the Open
223 Firmware client interface. It is also used by "kexec" to
224 implement "hot" booting of a new kernel from a previous
225 running one. This method is what I will describe in more
226 details in this document, as method a) is simply standard Open
227 Firmware, and thus should be implemented according to the
228 various standard documents defining it and its binding to the
229 PowerPC platform. The entry point definition then becomes:
231 r3 : physical pointer to the device-tree block
232 (defined in chapter II) in RAM
234 r4 : physical pointer to the kernel itself. This is
235 used by the assembly code to properly disable the MMU
236 in case you are entering the kernel with MMU enabled
237 and a non-1:1 mapping.
239 r5 : NULL (as to differentiate with method a)
241 Note about SMP entry: Either your firmware puts your other
242 CPUs in some sleep loop or spin loop in ROM where you can get
243 them out via a soft reset or some other means, in which case
244 you don't need to care, or you'll have to enter the kernel
245 with all CPUs. The way to do that with method b) will be
246 described in a later revision of this document.
254 Board supports (platforms) are not exclusive config options. An
255 arbitrary set of board supports can be built in a single kernel
256 image. The kernel will "know" what set of functions to use for a
257 given platform based on the content of the device-tree. Thus, you
260 a) add your platform support as a _boolean_ option in
261 arch/powerpc/Kconfig, following the example of PPC_PSERIES,
262 PPC_PMAC and PPC_MAPLE. The later is probably a good
263 example of a board support to start from.
265 b) create your main platform file as
266 "arch/powerpc/platforms/myplatform/myboard_setup.c" and add it
267 to the Makefile under the condition of your CONFIG_
268 option. This file will define a structure of type "ppc_md"
269 containing the various callbacks that the generic code will
270 use to get to your platform specific code
272 c) Add a reference to your "ppc_md" structure in the
273 "machines" table in arch/powerpc/kernel/setup_64.c if you are
276 d) request and get assigned a platform number (see PLATFORM_*
277 constants in include/asm-powerpc/processor.h
279 32-bit embedded kernels:
281 Currently, board support is essentially an exclusive config option.
282 The kernel is configured for a single platform. Part of the reason
283 for this is to keep kernels on embedded systems small and efficient;
284 part of this is due to the fact the code is already that way. In the
285 future, a kernel may support multiple platforms, but only if the
286 platforms feature the same core architecture. A single kernel build
287 cannot support both configurations with Book E and configurations
288 with classic Powerpc architectures.
290 32-bit embedded platforms that are moved into arch/powerpc using a
291 flattened device tree should adopt the merged tree practice of
292 setting ppc_md up dynamically, even though the kernel is currently
293 built with support for only a single platform at a time. This allows
294 unification of the setup code, and will make it easier to go to a
295 multiple-platform-support model in the future.
297 NOTE: I believe the above will be true once Ben's done with the merge
298 of the boot sequences.... someone speak up if this is wrong!
300 To add a 32-bit embedded platform support, follow the instructions
301 for 64-bit platforms above, with the exception that the Kconfig
302 option should be set up such that the kernel builds exclusively for
303 the platform selected. The processor type for the platform should
304 enable another config option to select the specific board
307 NOTE: If Ben doesn't merge the setup files, may need to change this to
311 I will describe later the boot process and various callbacks that
312 your platform should implement.
315 II - The DT block format
316 ========================
319 This chapter defines the actual format of the flattened device-tree
320 passed to the kernel. The actual content of it and kernel requirements
321 are described later. You can find example of code manipulating that
322 format in various places, including arch/powerpc/kernel/prom_init.c
323 which will generate a flattened device-tree from the Open Firmware
324 representation, or the fs2dt utility which is part of the kexec tools
325 which will generate one from a filesystem representation. It is
326 expected that a bootloader like uboot provides a bit more support,
327 that will be discussed later as well.
329 Note: The block has to be in main memory. It has to be accessible in
330 both real mode and virtual mode with no mapping other than main
331 memory. If you are writing a simple flash bootloader, it should copy
332 the block to RAM before passing it to the kernel.
338 The kernel is entered with r3 pointing to an area of memory that is
339 roughly described in include/asm-powerpc/prom.h by the structure
342 struct boot_param_header {
343 u32 magic; /* magic word OF_DT_HEADER */
344 u32 totalsize; /* total size of DT block */
345 u32 off_dt_struct; /* offset to structure */
346 u32 off_dt_strings; /* offset to strings */
347 u32 off_mem_rsvmap; /* offset to memory reserve map
349 u32 version; /* format version */
350 u32 last_comp_version; /* last compatible version */
352 /* version 2 fields below */
353 u32 boot_cpuid_phys; /* Which physical CPU id we're
355 /* version 3 fields below */
356 u32 size_dt_strings; /* size of the strings block */
358 /* version 17 fields below */
359 u32 size_dt_struct; /* size of the DT structure block */
362 Along with the constants:
364 /* Definitions used by the flattened device tree */
365 #define OF_DT_HEADER 0xd00dfeed /* 4: version,
367 #define OF_DT_BEGIN_NODE 0x1 /* Start node: full name
369 #define OF_DT_END_NODE 0x2 /* End node */
370 #define OF_DT_PROP 0x3 /* Property: name off,
372 #define OF_DT_END 0x9
374 All values in this header are in big endian format, the various
375 fields in this header are defined more precisely below. All
376 "offset" values are in bytes from the start of the header; that is
377 from the value of r3.
381 This is a magic value that "marks" the beginning of the
382 device-tree block header. It contains the value 0xd00dfeed and is
383 defined by the constant OF_DT_HEADER
387 This is the total size of the DT block including the header. The
388 "DT" block should enclose all data structures defined in this
389 chapter (who are pointed to by offsets in this header). That is,
390 the device-tree structure, strings, and the memory reserve map.
394 This is an offset from the beginning of the header to the start
395 of the "structure" part the device tree. (see 2) device tree)
399 This is an offset from the beginning of the header to the start
400 of the "strings" part of the device-tree
404 This is an offset from the beginning of the header to the start
405 of the reserved memory map. This map is a list of pairs of 64-
406 bit integers. Each pair is a physical address and a size. The
407 list is terminated by an entry of size 0. This map provides the
408 kernel with a list of physical memory areas that are "reserved"
409 and thus not to be used for memory allocations, especially during
410 early initialization. The kernel needs to allocate memory during
411 boot for things like un-flattening the device-tree, allocating an
412 MMU hash table, etc... Those allocations must be done in such a
413 way to avoid overriding critical things like, on Open Firmware
414 capable machines, the RTAS instance, or on some pSeries, the TCE
415 tables used for the iommu. Typically, the reserve map should
416 contain _at least_ this DT block itself (header,total_size). If
417 you are passing an initrd to the kernel, you should reserve it as
418 well. You do not need to reserve the kernel image itself. The map
419 should be 64-bit aligned.
423 This is the version of this structure. Version 1 stops
424 here. Version 2 adds an additional field boot_cpuid_phys.
425 Version 3 adds the size of the strings block, allowing the kernel
426 to reallocate it easily at boot and free up the unused flattened
427 structure after expansion. Version 16 introduces a new more
428 "compact" format for the tree itself that is however not backward
429 compatible. Version 17 adds an additional field, size_dt_struct,
430 allowing it to be reallocated or moved more easily (this is
431 particularly useful for bootloaders which need to make
432 adjustments to a device tree based on probed information). You
433 should always generate a structure of the highest version defined
434 at the time of your implementation. Currently that is version 17,
435 unless you explicitly aim at being backward compatible.
439 Last compatible version. This indicates down to what version of
440 the DT block you are backward compatible. For example, version 2
441 is backward compatible with version 1 (that is, a kernel build
442 for version 1 will be able to boot with a version 2 format). You
443 should put a 1 in this field if you generate a device tree of
444 version 1 to 3, or 16 if you generate a tree of version 16 or 17
445 using the new unit name format.
449 This field only exist on version 2 headers. It indicate which
450 physical CPU ID is calling the kernel entry point. This is used,
451 among others, by kexec. If you are on an SMP system, this value
452 should match the content of the "reg" property of the CPU node in
453 the device-tree corresponding to the CPU calling the kernel entry
454 point (see further chapters for more informations on the required
455 device-tree contents)
459 This field only exists on version 3 and later headers. It
460 gives the size of the "strings" section of the device tree (which
461 starts at the offset given by off_dt_strings).
465 This field only exists on version 17 and later headers. It gives
466 the size of the "structure" section of the device tree (which
467 starts at the offset given by off_dt_struct).
469 So the typical layout of a DT block (though the various parts don't
470 need to be in that order) looks like this (addresses go from top to
474 ------------------------------
475 r3 -> | struct boot_param_header |
476 ------------------------------
477 | (alignment gap) (*) |
478 ------------------------------
479 | memory reserve map |
480 ------------------------------
482 ------------------------------
484 | device-tree structure |
486 ------------------------------
488 ------------------------------
490 | device-tree strings |
492 -----> ------------------------------
497 (*) The alignment gaps are not necessarily present; their presence
498 and size are dependent on the various alignment requirements of
499 the individual data blocks.
502 2) Device tree generalities
503 ---------------------------
505 This device-tree itself is separated in two different blocks, a
506 structure block and a strings block. Both need to be aligned to a 4
509 First, let's quickly describe the device-tree concept before detailing
510 the storage format. This chapter does _not_ describe the detail of the
511 required types of nodes & properties for the kernel, this is done
512 later in chapter III.
514 The device-tree layout is strongly inherited from the definition of
515 the Open Firmware IEEE 1275 device-tree. It's basically a tree of
516 nodes, each node having two or more named properties. A property can
519 It is a tree, so each node has one and only one parent except for the
520 root node who has no parent.
522 A node has 2 names. The actual node name is generally contained in a
523 property of type "name" in the node property list whose value is a
524 zero terminated string and is mandatory for version 1 to 3 of the
525 format definition (as it is in Open Firmware). Version 16 makes it
526 optional as it can generate it from the unit name defined below.
528 There is also a "unit name" that is used to differentiate nodes with
529 the same name at the same level, it is usually made of the node
530 names, the "@" sign, and a "unit address", which definition is
531 specific to the bus type the node sits on.
533 The unit name doesn't exist as a property per-se but is included in
534 the device-tree structure. It is typically used to represent "path" in
535 the device-tree. More details about the actual format of these will be
538 The kernel powerpc generic code does not make any formal use of the
539 unit address (though some board support code may do) so the only real
540 requirement here for the unit address is to ensure uniqueness of
541 the node unit name at a given level of the tree. Nodes with no notion
542 of address and no possible sibling of the same name (like /memory or
543 /cpus) may omit the unit address in the context of this specification,
544 or use the "@0" default unit address. The unit name is used to define
545 a node "full path", which is the concatenation of all parent node
546 unit names separated with "/".
548 The root node doesn't have a defined name, and isn't required to have
549 a name property either if you are using version 3 or earlier of the
550 format. It also has no unit address (no @ symbol followed by a unit
551 address). The root node unit name is thus an empty string. The full
552 path to the root node is "/".
554 Every node which actually represents an actual device (that is, a node
555 which isn't only a virtual "container" for more nodes, like "/cpus"
556 is) is also required to have a "device_type" property indicating the
559 Finally, every node that can be referenced from a property in another
560 node is required to have a "linux,phandle" property. Real open
561 firmware implementations provide a unique "phandle" value for every
562 node that the "prom_init()" trampoline code turns into
563 "linux,phandle" properties. However, this is made optional if the
564 flattened device tree is used directly. An example of a node
565 referencing another node via "phandle" is when laying out the
566 interrupt tree which will be described in a further version of this
569 This "linux, phandle" property is a 32-bit value that uniquely
570 identifies a node. You are free to use whatever values or system of
571 values, internal pointers, or whatever to generate these, the only
572 requirement is that every node for which you provide that property has
573 a unique value for it.
575 Here is an example of a simple device-tree. In this example, an "o"
576 designates a node followed by the node unit name. Properties are
577 presented with their name followed by their content. "content"
578 represents an ASCII string (zero terminated) value, while <content>
579 represents a 32-bit hexadecimal value. The various nodes in this
580 example will be discussed in a later chapter. At this point, it is
581 only meant to give you a idea of what a device-tree looks like. I have
582 purposefully kept the "name" and "linux,phandle" properties which
583 aren't necessary in order to give you a better idea of what the tree
584 looks like in practice.
587 |- name = "device-tree"
588 |- model = "MyBoardName"
589 |- compatible = "MyBoardFamilyName"
590 |- #address-cells = <2>
592 |- linux,phandle = <0>
596 | | - linux,phandle = <1>
597 | | - #address-cells = <1>
598 | | - #size-cells = <0>
601 | |- name = "PowerPC,970"
602 | |- device_type = "cpu"
604 | |- clock-frequency = <5f5e1000>
606 | |- linux,phandle = <2>
610 | |- device_type = "memory"
611 | |- reg = <00000000 00000000 00000000 20000000>
612 | |- linux,phandle = <3>
616 |- bootargs = "root=/dev/sda2"
617 |- linux,phandle = <4>
619 This tree is almost a minimal tree. It pretty much contains the
620 minimal set of required nodes and properties to boot a linux kernel;
621 that is, some basic model informations at the root, the CPUs, and the
622 physical memory layout. It also includes misc information passed
623 through /chosen, like in this example, the platform type (mandatory)
624 and the kernel command line arguments (optional).
626 The /cpus/PowerPC,970@0/64-bit property is an example of a
627 property without a value. All other properties have a value. The
628 significance of the #address-cells and #size-cells properties will be
629 explained in chapter IV which defines precisely the required nodes and
630 properties and their content.
633 3) Device tree "structure" block
635 The structure of the device tree is a linearized tree structure. The
636 "OF_DT_BEGIN_NODE" token starts a new node, and the "OF_DT_END_NODE"
637 ends that node definition. Child nodes are simply defined before
638 "OF_DT_END_NODE" (that is nodes within the node). A 'token' is a 32
639 bit value. The tree has to be "finished" with a OF_DT_END token
641 Here's the basic structure of a single node:
643 * token OF_DT_BEGIN_NODE (that is 0x00000001)
644 * for version 1 to 3, this is the node full path as a zero
645 terminated string, starting with "/". For version 16 and later,
646 this is the node unit name only (or an empty string for the
648 * [align gap to next 4 bytes boundary]
650 * token OF_DT_PROP (that is 0x00000003)
651 * 32-bit value of property value size in bytes (or 0 if no
653 * 32-bit value of offset in string block of property name
654 * property value data if any
655 * [align gap to next 4 bytes boundary]
656 * [child nodes if any]
657 * token OF_DT_END_NODE (that is 0x00000002)
659 So the node content can be summarized as a start token, a full path,
660 a list of properties, a list of child nodes, and an end token. Every
661 child node is a full node structure itself as defined above.
663 NOTE: The above definition requires that all property definitions for
664 a particular node MUST precede any subnode definitions for that node.
665 Although the structure would not be ambiguous if properties and
666 subnodes were intermingled, the kernel parser requires that the
667 properties come first (up until at least 2.6.22). Any tools
668 manipulating a flattened tree must take care to preserve this
671 4) Device tree "strings" block
673 In order to save space, property names, which are generally redundant,
674 are stored separately in the "strings" block. This block is simply the
675 whole bunch of zero terminated strings for all property names
676 concatenated together. The device-tree property definitions in the
677 structure block will contain offset values from the beginning of the
681 III - Required content of the device tree
682 =========================================
684 WARNING: All "linux,*" properties defined in this document apply only
685 to a flattened device-tree. If your platform uses a real
686 implementation of Open Firmware or an implementation compatible with
687 the Open Firmware client interface, those properties will be created
688 by the trampoline code in the kernel's prom_init() file. For example,
689 that's where you'll have to add code to detect your board model and
690 set the platform number. However, when using the flattened device-tree
691 entry point, there is no prom_init() pass, and thus you have to
692 provide those properties yourself.
695 1) Note about cells and address representation
696 ----------------------------------------------
698 The general rule is documented in the various Open Firmware
699 documentations. If you choose to describe a bus with the device-tree
700 and there exist an OF bus binding, then you should follow the
701 specification. However, the kernel does not require every single
702 device or bus to be described by the device tree.
704 In general, the format of an address for a device is defined by the
705 parent bus type, based on the #address-cells and #size-cells
706 properties. Note that the parent's parent definitions of #address-cells
707 and #size-cells are not inhereted so every node with children must specify
708 them. The kernel requires the root node to have those properties defining
709 addresses format for devices directly mapped on the processor bus.
711 Those 2 properties define 'cells' for representing an address and a
712 size. A "cell" is a 32-bit number. For example, if both contain 2
713 like the example tree given above, then an address and a size are both
714 composed of 2 cells, and each is a 64-bit number (cells are
715 concatenated and expected to be in big endian format). Another example
716 is the way Apple firmware defines them, with 2 cells for an address
717 and one cell for a size. Most 32-bit implementations should define
718 #address-cells and #size-cells to 1, which represents a 32-bit value.
719 Some 32-bit processors allow for physical addresses greater than 32
720 bits; these processors should define #address-cells as 2.
722 "reg" properties are always a tuple of the type "address size" where
723 the number of cells of address and size is specified by the bus
724 #address-cells and #size-cells. When a bus supports various address
725 spaces and other flags relative to a given address allocation (like
726 prefetchable, etc...) those flags are usually added to the top level
727 bits of the physical address. For example, a PCI physical address is
728 made of 3 cells, the bottom two containing the actual address itself
729 while the top cell contains address space indication, flags, and pci
730 bus & device numbers.
732 For busses that support dynamic allocation, it's the accepted practice
733 to then not provide the address in "reg" (keep it 0) though while
734 providing a flag indicating the address is dynamically allocated, and
735 then, to provide a separate "assigned-addresses" property that
736 contains the fully allocated addresses. See the PCI OF bindings for
739 In general, a simple bus with no address space bits and no dynamic
740 allocation is preferred if it reflects your hardware, as the existing
741 kernel address parsing functions will work out of the box. If you
742 define a bus type with a more complex address format, including things
743 like address space bits, you'll have to add a bus translator to the
744 prom_parse.c file of the recent kernels for your bus type.
746 The "reg" property only defines addresses and sizes (if #size-cells is
747 non-0) within a given bus. In order to translate addresses upward
748 (that is into parent bus addresses, and possibly into CPU physical
749 addresses), all busses must contain a "ranges" property. If the
750 "ranges" property is missing at a given level, it's assumed that
751 translation isn't possible, i.e., the registers are not visible on the
752 parent bus. The format of the "ranges" property for a bus is a list
755 bus address, parent bus address, size
757 "bus address" is in the format of the bus this bus node is defining,
758 that is, for a PCI bridge, it would be a PCI address. Thus, (bus
759 address, size) defines a range of addresses for child devices. "parent
760 bus address" is in the format of the parent bus of this bus. For
761 example, for a PCI host controller, that would be a CPU address. For a
762 PCI<->ISA bridge, that would be a PCI address. It defines the base
763 address in the parent bus where the beginning of that range is mapped.
765 For a new 64-bit powerpc board, I recommend either the 2/2 format or
766 Apple's 2/1 format which is slightly more compact since sizes usually
767 fit in a single 32-bit word. New 32-bit powerpc boards should use a
768 1/1 format, unless the processor supports physical addresses greater
769 than 32-bits, in which case a 2/1 format is recommended.
771 Alternatively, the "ranges" property may be empty, indicating that the
772 registers are visible on the parent bus using an identity mapping
773 translation. In other words, the parent bus address space is the same
774 as the child bus address space.
776 2) Note about "compatible" properties
777 -------------------------------------
779 These properties are optional, but recommended in devices and the root
780 node. The format of a "compatible" property is a list of concatenated
781 zero terminated strings. They allow a device to express its
782 compatibility with a family of similar devices, in some cases,
783 allowing a single driver to match against several devices regardless
784 of their actual names.
786 3) Note about "name" properties
787 -------------------------------
789 While earlier users of Open Firmware like OldWorld macintoshes tended
790 to use the actual device name for the "name" property, it's nowadays
791 considered a good practice to use a name that is closer to the device
792 class (often equal to device_type). For example, nowadays, ethernet
793 controllers are named "ethernet", an additional "model" property
794 defining precisely the chip type/model, and "compatible" property
795 defining the family in case a single driver can driver more than one
796 of these chips. However, the kernel doesn't generally put any
797 restriction on the "name" property; it is simply considered good
798 practice to follow the standard and its evolutions as closely as
801 Note also that the new format version 16 makes the "name" property
802 optional. If it's absent for a node, then the node's unit name is then
803 used to reconstruct the name. That is, the part of the unit name
804 before the "@" sign is used (or the entire unit name if no "@" sign
807 4) Note about node and property names and character set
808 -------------------------------------------------------
810 While open firmware provides more flexible usage of 8859-1, this
811 specification enforces more strict rules. Nodes and properties should
812 be comprised only of ASCII characters 'a' to 'z', '0' to
813 '9', ',', '.', '_', '+', '#', '?', and '-'. Node names additionally
814 allow uppercase characters 'A' to 'Z' (property names should be
815 lowercase. The fact that vendors like Apple don't respect this rule is
816 irrelevant here). Additionally, node and property names should always
817 begin with a character in the range 'a' to 'z' (or 'A' to 'Z' for node
820 The maximum number of characters for both nodes and property names
821 is 31. In the case of node names, this is only the leftmost part of
822 a unit name (the pure "name" property), it doesn't include the unit
823 address which can extend beyond that limit.
826 5) Required nodes and properties
827 --------------------------------
828 These are all that are currently required. However, it is strongly
829 recommended that you expose PCI host bridges as documented in the
830 PCI binding to open firmware, and your interrupt tree as documented
831 in OF interrupt tree specification.
835 The root node requires some properties to be present:
837 - model : this is your board name/model
838 - #address-cells : address representation for "root" devices
839 - #size-cells: the size representation for "root" devices
840 - device_type : This property shouldn't be necessary. However, if
841 you decide to create a device_type for your root node, make sure it
842 is _not_ "chrp" unless your platform is a pSeries or PAPR compliant
843 one for 64-bit, or a CHRP-type machine for 32-bit as this will
844 matched by the kernel this way.
846 Additionally, some recommended properties are:
848 - compatible : the board "family" generally finds its way here,
849 for example, if you have 2 board models with a similar layout,
850 that typically get driven by the same platform code in the
851 kernel, you would use a different "model" property but put a
852 value in "compatible". The kernel doesn't directly use that
853 value but it is generally useful.
855 The root node is also generally where you add additional properties
856 specific to your board like the serial number if any, that sort of
857 thing. It is recommended that if you add any "custom" property whose
858 name may clash with standard defined ones, you prefix them with your
859 vendor name and a comma.
863 This node is the parent of all individual CPU nodes. It doesn't
864 have any specific requirements, though it's generally good practice
867 #address-cells = <00000001>
868 #size-cells = <00000000>
870 This defines that the "address" for a CPU is a single cell, and has
871 no meaningful size. This is not necessary but the kernel will assume
872 that format when reading the "reg" properties of a CPU node, see
877 So under /cpus, you are supposed to create a node for every CPU on
878 the machine. There is no specific restriction on the name of the
879 CPU, though It's common practice to call it PowerPC,<name>. For
880 example, Apple uses PowerPC,G5 while IBM uses PowerPC,970FX.
884 - device_type : has to be "cpu"
885 - reg : This is the physical CPU number, it's a single 32-bit cell
886 and is also used as-is as the unit number for constructing the
887 unit name in the full path. For example, with 2 CPUs, you would
889 /cpus/PowerPC,970FX@0
890 /cpus/PowerPC,970FX@1
891 (unit addresses do not require leading zeroes)
892 - d-cache-block-size : one cell, L1 data cache block size in bytes (*)
893 - i-cache-block-size : one cell, L1 instruction cache block size in
895 - d-cache-size : one cell, size of L1 data cache in bytes
896 - i-cache-size : one cell, size of L1 instruction cache in bytes
898 (*) The cache "block" size is the size on which the cache management
899 instructions operate. Historically, this document used the cache
900 "line" size here which is incorrect. The kernel will prefer the cache
901 block size and will fallback to cache line size for backward
904 Recommended properties:
906 - timebase-frequency : a cell indicating the frequency of the
907 timebase in Hz. This is not directly used by the generic code,
908 but you are welcome to copy/paste the pSeries code for setting
909 the kernel timebase/decrementer calibration based on this
911 - clock-frequency : a cell indicating the CPU core clock frequency
912 in Hz. A new property will be defined for 64-bit values, but if
913 your frequency is < 4Ghz, one cell is enough. Here as well as
914 for the above, the common code doesn't use that property, but
915 you are welcome to re-use the pSeries or Maple one. A future
916 kernel version might provide a common function for this.
917 - d-cache-line-size : one cell, L1 data cache line size in bytes
918 if different from the block size
919 - i-cache-line-size : one cell, L1 instruction cache line size in
920 bytes if different from the block size
922 You are welcome to add any property you find relevant to your board,
923 like some information about the mechanism used to soft-reset the
924 CPUs. For example, Apple puts the GPIO number for CPU soft reset
925 lines in there as a "soft-reset" property since they start secondary
926 CPUs by soft-resetting them.
929 d) the /memory node(s)
931 To define the physical memory layout of your board, you should
932 create one or more memory node(s). You can either create a single
933 node with all memory ranges in its reg property, or you can create
934 several nodes, as you wish. The unit address (@ part) used for the
935 full path is the address of the first range of memory defined by a
936 given node. If you use a single memory node, this will typically be
941 - device_type : has to be "memory"
942 - reg : This property contains all the physical memory ranges of
943 your board. It's a list of addresses/sizes concatenated
944 together, with the number of cells of each defined by the
945 #address-cells and #size-cells of the root node. For example,
946 with both of these properties being 2 like in the example given
947 earlier, a 970 based machine with 6Gb of RAM could typically
948 have a "reg" property here that looks like:
950 00000000 00000000 00000000 80000000
951 00000001 00000000 00000001 00000000
953 That is a range starting at 0 of 0x80000000 bytes and a range
954 starting at 0x100000000 and of 0x100000000 bytes. You can see
955 that there is no memory covering the IO hole between 2Gb and
956 4Gb. Some vendors prefer splitting those ranges into smaller
957 segments, but the kernel doesn't care.
961 This node is a bit "special". Normally, that's where open firmware
962 puts some variable environment information, like the arguments, or
963 the default input/output devices.
965 This specification makes a few of these mandatory, but also defines
966 some linux-specific properties that would be normally constructed by
967 the prom_init() trampoline when booting with an OF client interface,
968 but that you have to provide yourself when using the flattened format.
970 Recommended properties:
972 - bootargs : This zero-terminated string is passed as the kernel
974 - linux,stdout-path : This is the full path to your standard
975 console device if any. Typically, if you have serial devices on
976 your board, you may want to put the full path to the one set as
977 the default console in the firmware here, for the kernel to pick
978 it up as its own default console. If you look at the function
979 set_preferred_console() in arch/ppc64/kernel/setup.c, you'll see
980 that the kernel tries to find out the default console and has
981 knowledge of various types like 8250 serial ports. You may want
982 to extend this function to add your own.
984 Note that u-boot creates and fills in the chosen node for platforms
987 (Note: a practice that is now obsolete was to include a property
988 under /chosen called interrupt-controller which had a phandle value
989 that pointed to the main interrupt controller)
991 f) the /soc<SOCname> node
993 This node is used to represent a system-on-a-chip (SOC) and must be
994 present if the processor is a SOC. The top-level soc node contains
995 information that is global to all devices on the SOC. The node name
996 should contain a unit address for the SOC, which is the base address
997 of the memory-mapped register set for the SOC. The name of an soc
998 node should start with "soc", and the remainder of the name should
999 represent the part number for the soc. For example, the MPC8540's
1000 soc node would be called "soc8540".
1002 Required properties:
1004 - device_type : Should be "soc"
1005 - ranges : Should be defined as specified in 1) to describe the
1006 translation of SOC addresses for memory mapped SOC registers.
1007 - bus-frequency: Contains the bus frequency for the SOC node.
1008 Typically, the value of this field is filled in by the boot
1012 Recommended properties:
1014 - reg : This property defines the address and size of the
1015 memory-mapped registers that are used for the SOC node itself.
1016 It does not include the child device registers - these will be
1017 defined inside each child node. The address specified in the
1018 "reg" property should match the unit address of the SOC node.
1019 - #address-cells : Address representation for "soc" devices. The
1020 format of this field may vary depending on whether or not the
1021 device registers are memory mapped. For memory mapped
1022 registers, this field represents the number of cells needed to
1023 represent the address of the registers. For SOCs that do not
1024 use MMIO, a special address format should be defined that
1025 contains enough cells to represent the required information.
1026 See 1) above for more details on defining #address-cells.
1027 - #size-cells : Size representation for "soc" devices
1028 - #interrupt-cells : Defines the width of cells used to represent
1029 interrupts. Typically this value is <2>, which includes a
1030 32-bit number that represents the interrupt number, and a
1031 32-bit number that represents the interrupt sense and level.
1032 This field is only needed if the SOC contains an interrupt
1035 The SOC node may contain child nodes for each SOC device that the
1036 platform uses. Nodes should not be created for devices which exist
1037 on the SOC but are not used by a particular platform. See chapter VI
1038 for more information on how to specify devices that are part of a SOC.
1040 Example SOC node for the MPC8540:
1043 #address-cells = <1>;
1045 #interrupt-cells = <2>;
1046 device_type = "soc";
1047 ranges = <00000000 e0000000 00100000>
1048 reg = <e0000000 00003000>;
1049 bus-frequency = <0>;
1054 IV - "dtc", the device tree compiler
1055 ====================================
1058 dtc source code can be found at
1059 <http://ozlabs.org/~dgibson/dtc/dtc.tar.gz>
1061 WARNING: This version is still in early development stage; the
1062 resulting device-tree "blobs" have not yet been validated with the
1063 kernel. The current generated bloc lacks a useful reserve map (it will
1064 be fixed to generate an empty one, it's up to the bootloader to fill
1065 it up) among others. The error handling needs work, bugs are lurking,
1068 dtc basically takes a device-tree in a given format and outputs a
1069 device-tree in another format. The currently supported formats are:
1074 - "dtb": "blob" format, that is a flattened device-tree block
1076 header all in a binary blob.
1077 - "dts": "source" format. This is a text file containing a
1078 "source" for a device-tree. The format is defined later in this
1080 - "fs" format. This is a representation equivalent to the
1081 output of /proc/device-tree, that is nodes are directories and
1082 properties are files
1087 - "dtb": "blob" format
1088 - "dts": "source" format
1089 - "asm": assembly language file. This is a file that can be
1090 sourced by gas to generate a device-tree "blob". That file can
1091 then simply be added to your Makefile. Additionally, the
1092 assembly file exports some symbols that can be used.
1095 The syntax of the dtc tool is
1097 dtc [-I <input-format>] [-O <output-format>]
1098 [-o output-filename] [-V output_version] input_filename
1101 The "output_version" defines what version of the "blob" format will be
1102 generated. Supported versions are 1,2,3 and 16. The default is
1103 currently version 3 but that may change in the future to version 16.
1105 Additionally, dtc performs various sanity checks on the tree, like the
1106 uniqueness of linux, phandle properties, validity of strings, etc...
1108 The format of the .dts "source" file is "C" like, supports C and C++
1114 The above is the "device-tree" definition. It's the only statement
1115 supported currently at the toplevel.
1118 property1 = "string_value"; /* define a property containing a 0
1122 property2 = <1234abcd>; /* define a property containing a
1123 * numerical 32-bit value (hexadecimal)
1126 property3 = <12345678 12345678 deadbeef>;
1127 /* define a property containing 3
1128 * numerical 32-bit values (cells) in
1131 property4 = [0a 0b 0c 0d de ea ad be ef];
1132 /* define a property whose content is
1133 * an arbitrary array of bytes
1136 childnode@addresss { /* define a child node named "childnode"
1137 * whose unit name is "childnode at
1141 childprop = "hello\n"; /* define a property "childprop" of
1142 * childnode (in this case, a string)
1147 Nodes can contain other nodes etc... thus defining the hierarchical
1148 structure of the tree.
1150 Strings support common escape sequences from C: "\n", "\t", "\r",
1151 "\(octal value)", "\x(hex value)".
1153 It is also suggested that you pipe your source file through cpp (gcc
1154 preprocessor) so you can use #include's, #define for constants, etc...
1156 Finally, various options are planned but not yet implemented, like
1157 automatic generation of phandles, labels (exported to the asm file so
1158 you can point to a property content and change it easily from whatever
1159 you link the device-tree with), label or path instead of numeric value
1160 in some cells to "point" to a node (replaced by a phandle at compile
1161 time), export of reserve map address to the asm file, ability to
1162 specify reserve map content at compile time, etc...
1164 We may provide a .h include file with common definitions of that
1165 proves useful for some properties (like building PCI properties or
1166 interrupt maps) though it may be better to add a notion of struct
1167 definitions to the compiler...
1170 V - Recommendations for a bootloader
1171 ====================================
1174 Here are some various ideas/recommendations that have been proposed
1175 while all this has been defined and implemented.
1177 - The bootloader may want to be able to use the device-tree itself
1178 and may want to manipulate it (to add/edit some properties,
1179 like physical memory size or kernel arguments). At this point, 2
1180 choices can be made. Either the bootloader works directly on the
1181 flattened format, or the bootloader has its own internal tree
1182 representation with pointers (similar to the kernel one) and
1183 re-flattens the tree when booting the kernel. The former is a bit
1184 more difficult to edit/modify, the later requires probably a bit
1185 more code to handle the tree structure. Note that the structure
1186 format has been designed so it's relatively easy to "insert"
1187 properties or nodes or delete them by just memmoving things
1188 around. It contains no internal offsets or pointers for this
1191 - An example of code for iterating nodes & retrieving properties
1192 directly from the flattened tree format can be found in the kernel
1193 file arch/ppc64/kernel/prom.c, look at scan_flat_dt() function,
1194 its usage in early_init_devtree(), and the corresponding various
1195 early_init_dt_scan_*() callbacks. That code can be re-used in a
1196 GPL bootloader, and as the author of that code, I would be happy
1197 to discuss possible free licensing to any vendor who wishes to
1198 integrate all or part of this code into a non-GPL bootloader.
1202 VI - System-on-a-chip devices and nodes
1203 =======================================
1205 Many companies are now starting to develop system-on-a-chip
1206 processors, where the processor core (CPU) and many peripheral devices
1207 exist on a single piece of silicon. For these SOCs, an SOC node
1208 should be used that defines child nodes for the devices that make
1209 up the SOC. While platforms are not required to use this model in
1210 order to boot the kernel, it is highly encouraged that all SOC
1211 implementations define as complete a flat-device-tree as possible to
1212 describe the devices on the SOC. This will allow for the
1213 genericization of much of the kernel code.
1216 1) Defining child nodes of an SOC
1217 ---------------------------------
1219 Each device that is part of an SOC may have its own node entry inside
1220 the SOC node. For each device that is included in the SOC, the unit
1221 address property represents the address offset for this device's
1222 memory-mapped registers in the parent's address space. The parent's
1223 address space is defined by the "ranges" property in the top-level soc
1224 node. The "reg" property for each node that exists directly under the
1225 SOC node should contain the address mapping from the child address space
1226 to the parent SOC address space and the size of the device's
1227 memory-mapped register file.
1229 For many devices that may exist inside an SOC, there are predefined
1230 specifications for the format of the device tree node. All SOC child
1231 nodes should follow these specifications, except where noted in this
1234 See appendix A for an example partial SOC node definition for the
1238 2) Representing devices without a current OF specification
1239 ----------------------------------------------------------
1241 Currently, there are many devices on SOCs that do not have a standard
1242 representation pre-defined as part of the open firmware
1243 specifications, mainly because the boards that contain these SOCs are
1244 not currently booted using open firmware. This section contains
1245 descriptions for the SOC devices for which new nodes have been
1246 defined; this list will expand as more and more SOC-containing
1247 platforms are moved over to use the flattened-device-tree model.
1251 The MDIO is a bus to which the PHY devices are connected. For each
1252 device that exists on this bus, a child node should be created. See
1253 the definition of the PHY node below for an example of how to define
1256 Required properties:
1257 - reg : Offset and length of the register set for the device
1258 - compatible : Should define the compatible device type for the
1259 mdio. Currently, this is most likely to be "fsl,gianfar-mdio"
1265 compatible = "fsl,gianfar-mdio";
1273 b) Gianfar-compatible ethernet nodes
1275 Required properties:
1277 - device_type : Should be "network"
1278 - model : Model of the device. Can be "TSEC", "eTSEC", or "FEC"
1279 - compatible : Should be "gianfar"
1280 - reg : Offset and length of the register set for the device
1281 - mac-address : List of bytes representing the ethernet address of
1283 - interrupts : <a b> where a is the interrupt number and b is a
1284 field that represents an encoding of the sense and level
1285 information for the interrupt. This should be encoded based on
1286 the information in section 2) depending on the type of interrupt
1287 controller you have.
1288 - interrupt-parent : the phandle for the interrupt controller that
1289 services interrupts for this device.
1290 - phy-handle : The phandle for the PHY connected to this ethernet
1292 - fixed-link : <a b c d e> where a is emulated phy id - choose any,
1293 but unique to the all specified fixed-links, b is duplex - 0 half,
1294 1 full, c is link speed - d#10/d#100/d#1000, d is pause - 0 no
1295 pause, 1 pause, e is asym_pause - 0 no asym_pause, 1 asym_pause.
1297 Recommended properties:
1299 - phy-connection-type : a string naming the controller/PHY interface type,
1300 i.e., "mii" (default), "rmii", "gmii", "rgmii", "rgmii-id", "sgmii",
1301 "tbi", or "rtbi". This property is only really needed if the connection
1302 is of type "rgmii-id", as all other connection types are detected by
1310 device_type = "network";
1312 compatible = "gianfar";
1314 mac-address = [ 00 E0 0C 00 73 00 ];
1315 interrupts = <d 3 e 3 12 3>;
1316 interrupt-parent = <40000>;
1317 phy-handle = <2452000>
1324 Required properties:
1326 - device_type : Should be "ethernet-phy"
1327 - interrupts : <a b> where a is the interrupt number and b is a
1328 field that represents an encoding of the sense and level
1329 information for the interrupt. This should be encoded based on
1330 the information in section 2) depending on the type of interrupt
1331 controller you have.
1332 - interrupt-parent : the phandle for the interrupt controller that
1333 services interrupts for this device.
1334 - reg : The ID number for the phy, usually a small integer
1335 - linux,phandle : phandle for this node; likely referenced by an
1336 ethernet controller node.
1342 linux,phandle = <2452000>
1343 interrupt-parent = <40000>;
1344 interrupts = <35 1>;
1346 device_type = "ethernet-phy";
1350 d) Interrupt controllers
1352 Some SOC devices contain interrupt controllers that are different
1353 from the standard Open PIC specification. The SOC device nodes for
1354 these types of controllers should be specified just like a standard
1355 OpenPIC controller. Sense and level information should be encoded
1356 as specified in section 2) of this chapter for each device that
1357 specifies an interrupt.
1362 linux,phandle = <40000>;
1363 clock-frequency = <0>;
1364 interrupt-controller;
1365 #address-cells = <0>;
1366 reg = <40000 40000>;
1368 compatible = "chrp,open-pic";
1369 device_type = "open-pic";
1376 Required properties :
1378 - device_type : Should be "i2c"
1379 - reg : Offset and length of the register set for the device
1381 Recommended properties :
1383 - compatible : Should be "fsl-i2c" for parts compatible with
1384 Freescale I2C specifications.
1385 - interrupts : <a b> where a is the interrupt number and b is a
1386 field that represents an encoding of the sense and level
1387 information for the interrupt. This should be encoded based on
1388 the information in section 2) depending on the type of interrupt
1389 controller you have.
1390 - interrupt-parent : the phandle for the interrupt controller that
1391 services interrupts for this device.
1392 - dfsrr : boolean; if defined, indicates that this I2C device has
1393 a digital filter sampling rate register
1394 - fsl5200-clocking : boolean; if defined, indicated that this device
1395 uses the FSL 5200 clocking mechanism.
1400 interrupt-parent = <40000>;
1401 interrupts = <1b 3>;
1403 device_type = "i2c";
1404 compatible = "fsl-i2c";
1409 f) Freescale SOC USB controllers
1411 The device node for a USB controller that is part of a Freescale
1412 SOC is as described in the document "Open Firmware Recommended
1413 Practice : Universal Serial Bus" with the following modifications
1416 Required properties :
1417 - compatible : Should be "fsl-usb2-mph" for multi port host USB
1418 controllers, or "fsl-usb2-dr" for dual role USB controllers
1419 - phy_type : For multi port host USB controllers, should be one of
1420 "ulpi", or "serial". For dual role USB controllers, should be
1421 one of "ulpi", "utmi", "utmi_wide", or "serial".
1422 - reg : Offset and length of the register set for the device
1423 - port0 : boolean; if defined, indicates port0 is connected for
1424 fsl-usb2-mph compatible controllers. Either this property or
1425 "port1" (or both) must be defined for "fsl-usb2-mph" compatible
1427 - port1 : boolean; if defined, indicates port1 is connected for
1428 fsl-usb2-mph compatible controllers. Either this property or
1429 "port0" (or both) must be defined for "fsl-usb2-mph" compatible
1431 - dr_mode : indicates the working mode for "fsl-usb2-dr" compatible
1432 controllers. Can be "host", "peripheral", or "otg". Default to
1433 "host" if not defined for backward compatibility.
1435 Recommended properties :
1436 - interrupts : <a b> where a is the interrupt number and b is a
1437 field that represents an encoding of the sense and level
1438 information for the interrupt. This should be encoded based on
1439 the information in section 2) depending on the type of interrupt
1440 controller you have.
1441 - interrupt-parent : the phandle for the interrupt controller that
1442 services interrupts for this device.
1444 Example multi port host USB controller device node :
1446 compatible = "fsl-usb2-mph";
1448 #address-cells = <1>;
1450 interrupt-parent = <700>;
1451 interrupts = <27 1>;
1457 Example dual role USB controller device node :
1459 compatible = "fsl-usb2-dr";
1461 #address-cells = <1>;
1463 interrupt-parent = <700>;
1464 interrupts = <26 1>;
1470 g) Freescale SOC SEC Security Engines
1472 Required properties:
1474 - device_type : Should be "crypto"
1475 - model : Model of the device. Should be "SEC1" or "SEC2"
1476 - compatible : Should be "talitos"
1477 - reg : Offset and length of the register set for the device
1478 - interrupts : <a b> where a is the interrupt number and b is a
1479 field that represents an encoding of the sense and level
1480 information for the interrupt. This should be encoded based on
1481 the information in section 2) depending on the type of interrupt
1482 controller you have.
1483 - interrupt-parent : the phandle for the interrupt controller that
1484 services interrupts for this device.
1485 - num-channels : An integer representing the number of channels
1487 - channel-fifo-len : An integer representing the number of
1488 descriptor pointers each channel fetch fifo can hold.
1489 - exec-units-mask : The bitmask representing what execution units
1490 (EUs) are available. It's a single 32-bit cell. EU information
1491 should be encoded following the SEC's Descriptor Header Dword
1492 EU_SEL0 field documentation, i.e. as follows:
1494 bit 0 = reserved - should be 0
1495 bit 1 = set if SEC has the ARC4 EU (AFEU)
1496 bit 2 = set if SEC has the DES/3DES EU (DEU)
1497 bit 3 = set if SEC has the message digest EU (MDEU)
1498 bit 4 = set if SEC has the random number generator EU (RNG)
1499 bit 5 = set if SEC has the public key EU (PKEU)
1500 bit 6 = set if SEC has the AES EU (AESU)
1501 bit 7 = set if SEC has the Kasumi EU (KEU)
1503 bits 8 through 31 are reserved for future SEC EUs.
1505 - descriptor-types-mask : The bitmask representing what descriptors
1506 are available. It's a single 32-bit cell. Descriptor type
1507 information should be encoded following the SEC's Descriptor
1508 Header Dword DESC_TYPE field documentation, i.e. as follows:
1510 bit 0 = set if SEC supports the aesu_ctr_nonsnoop desc. type
1511 bit 1 = set if SEC supports the ipsec_esp descriptor type
1512 bit 2 = set if SEC supports the common_nonsnoop desc. type
1513 bit 3 = set if SEC supports the 802.11i AES ccmp desc. type
1514 bit 4 = set if SEC supports the hmac_snoop_no_afeu desc. type
1515 bit 5 = set if SEC supports the srtp descriptor type
1516 bit 6 = set if SEC supports the non_hmac_snoop_no_afeu desc.type
1517 bit 7 = set if SEC supports the pkeu_assemble descriptor type
1518 bit 8 = set if SEC supports the aesu_key_expand_output desc.type
1519 bit 9 = set if SEC supports the pkeu_ptmul descriptor type
1520 bit 10 = set if SEC supports the common_nonsnoop_afeu desc. type
1521 bit 11 = set if SEC supports the pkeu_ptadd_dbl descriptor type
1523 ..and so on and so forth.
1529 device_type = "crypto";
1531 compatible = "talitos";
1532 reg = <30000 10000>;
1533 interrupts = <1d 3>;
1534 interrupt-parent = <40000>;
1536 channel-fifo-len = <18>;
1537 exec-units-mask = <000000fe>;
1538 descriptor-types-mask = <012b0ebf>;
1541 h) Board Control and Status (BCSR)
1543 Required properties:
1545 - device_type : Should be "board-control"
1546 - reg : Offset and length of the register set for the device
1551 device_type = "board-control";
1552 reg = <f8000000 8000>;
1555 i) Freescale QUICC Engine module (QE)
1556 This represents qe module that is installed on PowerQUICC II Pro.
1558 NOTE: This is an interim binding; it should be updated to fit
1559 in with the CPM binding later in this document.
1561 Basically, it is a bus of devices, that could act more or less
1562 as a complete entity (UCC, USB etc ). All of them should be siblings on
1563 the "root" qe node, using the common properties from there.
1564 The description below applies to the qe of MPC8360 and
1565 more nodes and properties would be extended in the future.
1569 Required properties:
1570 - compatible : should be "fsl,qe";
1571 - model : precise model of the QE, Can be "QE", "CPM", or "CPM2"
1572 - reg : offset and length of the device registers.
1573 - bus-frequency : the clock frequency for QUICC Engine.
1575 Recommended properties
1576 - brg-frequency : the internal clock source frequency for baud-rate
1581 #address-cells = <1>;
1583 #interrupt-cells = <2>;
1584 compatible = "fsl,qe";
1585 ranges = <0 e0100000 00100000>;
1586 reg = <e0100000 480>;
1587 brg-frequency = <0>;
1588 bus-frequency = <179A7B00>;
1592 ii) SPI (Serial Peripheral Interface)
1594 Required properties:
1595 - cell-index : SPI controller index.
1596 - compatible : should be "fsl,spi".
1597 - mode : the SPI operation mode, it can be "cpu" or "cpu-qe".
1598 - reg : Offset and length of the register set for the device
1599 - interrupts : <a b> where a is the interrupt number and b is a
1600 field that represents an encoding of the sense and level
1601 information for the interrupt. This should be encoded based on
1602 the information in section 2) depending on the type of interrupt
1603 controller you have.
1604 - interrupt-parent : the phandle for the interrupt controller that
1605 services interrupts for this device.
1610 compatible = "fsl,spi";
1612 interrupts = <82 0>;
1613 interrupt-parent = <700>;
1618 iii) USB (Universal Serial Bus Controller)
1620 Required properties:
1621 - compatible : could be "qe_udc" or "fhci-hcd".
1622 - mode : the could be "host" or "slave".
1623 - reg : Offset and length of the register set for the device
1624 - interrupts : <a b> where a is the interrupt number and b is a
1625 field that represents an encoding of the sense and level
1626 information for the interrupt. This should be encoded based on
1627 the information in section 2) depending on the type of interrupt
1628 controller you have.
1629 - interrupt-parent : the phandle for the interrupt controller that
1630 services interrupts for this device.
1634 compatible = "qe_udc";
1636 interrupts = <8b 0>;
1637 interrupt-parent = <700>;
1642 iv) UCC (Unified Communications Controllers)
1644 Required properties:
1645 - device_type : should be "network", "hldc", "uart", "transparent"
1646 "bisync", "atm", or "serial".
1647 - compatible : could be "ucc_geth" or "fsl_atm" and so on.
1648 - cell-index : the ucc number(1-8), corresponding to UCCx in UM.
1649 - reg : Offset and length of the register set for the device
1650 - interrupts : <a b> where a is the interrupt number and b is a
1651 field that represents an encoding of the sense and level
1652 information for the interrupt. This should be encoded based on
1653 the information in section 2) depending on the type of interrupt
1654 controller you have.
1655 - interrupt-parent : the phandle for the interrupt controller that
1656 services interrupts for this device.
1657 - pio-handle : The phandle for the Parallel I/O port configuration.
1658 - port-number : for UART drivers, the port number to use, between 0 and 3.
1659 This usually corresponds to the /dev/ttyQE device, e.g. <0> = /dev/ttyQE0.
1660 The port number is added to the minor number of the device. Unlike the
1661 CPM UART driver, the port-number is required for the QE UART driver.
1662 - soft-uart : for UART drivers, if specified this means the QE UART device
1663 driver should use "Soft-UART" mode, which is needed on some SOCs that have
1664 broken UART hardware. Soft-UART is provided via a microcode upload.
1665 - rx-clock-name: the UCC receive clock source
1666 "none": clock source is disabled
1667 "brg1" through "brg16": clock source is BRG1-BRG16, respectively
1668 "clk1" through "clk24": clock source is CLK1-CLK24, respectively
1669 - tx-clock-name: the UCC transmit clock source
1670 "none": clock source is disabled
1671 "brg1" through "brg16": clock source is BRG1-BRG16, respectively
1672 "clk1" through "clk24": clock source is CLK1-CLK24, respectively
1673 The following two properties are deprecated. rx-clock has been replaced
1674 with rx-clock-name, and tx-clock has been replaced with tx-clock-name.
1675 Drivers that currently use the deprecated properties should continue to
1676 do so, in order to support older device trees, but they should be updated
1677 to check for the new properties first.
1678 - rx-clock : represents the UCC receive clock source.
1679 0x00 : clock source is disabled;
1680 0x1~0x10 : clock source is BRG1~BRG16 respectively;
1681 0x11~0x28: clock source is QE_CLK1~QE_CLK24 respectively.
1682 - tx-clock: represents the UCC transmit clock source;
1683 0x00 : clock source is disabled;
1684 0x1~0x10 : clock source is BRG1~BRG16 respectively;
1685 0x11~0x28: clock source is QE_CLK1~QE_CLK24 respectively.
1687 Required properties for network device_type:
1688 - mac-address : list of bytes representing the ethernet address.
1689 - phy-handle : The phandle for the PHY connected to this controller.
1691 Recommended properties:
1692 - phy-connection-type : a string naming the controller/PHY interface type,
1693 i.e., "mii" (default), "rmii", "gmii", "rgmii", "rgmii-id" (Internal
1694 Delay), "rgmii-txid" (delay on TX only), "rgmii-rxid" (delay on RX only),
1699 device_type = "network";
1700 compatible = "ucc_geth";
1703 interrupts = <a0 0>;
1704 interrupt-parent = <700>;
1705 mac-address = [ 00 04 9f 00 23 23 ];
1708 phy-handle = <212000>;
1709 phy-connection-type = "gmii";
1710 pio-handle = <140001>;
1714 v) Parallel I/O Ports
1716 This node configures Parallel I/O ports for CPUs with QE support.
1717 The node should reside in the "soc" node of the tree. For each
1718 device that using parallel I/O ports, a child node should be created.
1719 See the definition of the Pin configuration nodes below for more
1722 Required properties:
1723 - device_type : should be "par_io".
1724 - reg : offset to the register set and its length.
1725 - num-ports : number of Parallel I/O ports
1730 #address-cells = <1>;
1732 device_type = "par_io";
1739 vi) Pin configuration nodes
1741 Required properties:
1742 - linux,phandle : phandle of this node; likely referenced by a QE
1744 - pio-map : array of pin configurations. Each pin is defined by 6
1745 integers. The six numbers are respectively: port, pin, dir,
1746 open_drain, assignment, has_irq.
1747 - port : port number of the pin; 0-6 represent port A-G in UM.
1748 - pin : pin number in the port.
1749 - dir : direction of the pin, should encode as follows:
1751 0 = The pin is disabled
1752 1 = The pin is an output
1753 2 = The pin is an input
1756 - open_drain : indicates the pin is normal or wired-OR:
1758 0 = The pin is actively driven as an output
1759 1 = The pin is an open-drain driver. As an output, the pin is
1760 driven active-low, otherwise it is three-stated.
1762 - assignment : function number of the pin according to the Pin Assignment
1763 tables in User Manual. Each pin can have up to 4 possible functions in
1764 QE and two options for CPM.
1765 - has_irq : indicates if the pin is used as source of external
1770 linux,phandle = <140001>;
1772 /* port pin dir open_drain assignment has_irq */
1773 0 3 1 0 1 0 /* TxD0 */
1774 0 4 1 0 1 0 /* TxD1 */
1775 0 5 1 0 1 0 /* TxD2 */
1776 0 6 1 0 1 0 /* TxD3 */
1777 1 6 1 0 3 0 /* TxD4 */
1778 1 7 1 0 1 0 /* TxD5 */
1779 1 9 1 0 2 0 /* TxD6 */
1780 1 a 1 0 2 0 /* TxD7 */
1781 0 9 2 0 1 0 /* RxD0 */
1782 0 a 2 0 1 0 /* RxD1 */
1783 0 b 2 0 1 0 /* RxD2 */
1784 0 c 2 0 1 0 /* RxD3 */
1785 0 d 2 0 1 0 /* RxD4 */
1786 1 1 2 0 2 0 /* RxD5 */
1787 1 0 2 0 2 0 /* RxD6 */
1788 1 4 2 0 2 0 /* RxD7 */
1789 0 7 1 0 1 0 /* TX_EN */
1790 0 8 1 0 1 0 /* TX_ER */
1791 0 f 2 0 1 0 /* RX_DV */
1792 0 10 2 0 1 0 /* RX_ER */
1793 0 0 2 0 1 0 /* RX_CLK */
1794 2 9 1 0 3 0 /* GTX_CLK - CLK10 */
1795 2 8 2 0 1 0>; /* GTX125 - CLK9 */
1798 vii) Multi-User RAM (MURAM)
1800 Required properties:
1801 - compatible : should be "fsl,qe-muram", "fsl,cpm-muram".
1802 - mode : the could be "host" or "slave".
1803 - ranges : Should be defined as specified in 1) to describe the
1804 translation of MURAM addresses.
1805 - data-only : sub-node which defines the address area under MURAM
1806 bus that can be allocated as data/parameter
1811 compatible = "fsl,qe-muram", "fsl,cpm-muram";
1812 ranges = <0 00010000 0000c000>;
1815 compatible = "fsl,qe-muram-data",
1816 "fsl,cpm-muram-data";
1821 viii) Uploaded QE firmware
1823 If a new firwmare has been uploaded to the QE (usually by the
1824 boot loader), then a 'firmware' child node should be added to the QE
1825 node. This node provides information on the uploaded firmware that
1826 device drivers may need.
1828 Required properties:
1829 - id: The string name of the firmware. This is taken from the 'id'
1830 member of the qe_firmware structure of the uploaded firmware.
1831 Device drivers can search this string to determine if the
1832 firmware they want is already present.
1833 - extended-modes: The Extended Modes bitfield, taken from the
1834 firmware binary. It is a 64-bit number represented
1835 as an array of two 32-bit numbers.
1836 - virtual-traps: The virtual traps, taken from the firmware binary.
1837 It is an array of 8 32-bit numbers.
1843 extended-modes = <0 0>;
1844 virtual-traps = <0 0 0 0 0 0 0 0>;
1847 j) CFI or JEDEC memory-mapped NOR flash
1849 Flash chips (Memory Technology Devices) are often used for solid state
1850 file systems on embedded devices.
1852 - compatible : should contain the specific model of flash chip(s)
1853 used, if known, followed by either "cfi-flash" or "jedec-flash"
1854 - reg : Address range of the flash chip
1855 - bank-width : Width (in bytes) of the flash bank. Equal to the
1856 device width times the number of interleaved chips.
1857 - device-width : (optional) Width of a single flash chip. If
1858 omitted, assumed to be equal to 'bank-width'.
1859 - #address-cells, #size-cells : Must be present if the flash has
1860 sub-nodes representing partitions (see below). In this case
1861 both #address-cells and #size-cells must be equal to 1.
1863 For JEDEC compatible devices, the following additional properties
1866 - vendor-id : Contains the flash chip's vendor id (1 byte).
1867 - device-id : Contains the flash chip's device id (1 byte).
1869 In addition to the information on the flash bank itself, the
1870 device tree may optionally contain additional information
1871 describing partitions of the flash address space. This can be
1872 used on platforms which have strong conventions about which
1873 portions of the flash are used for what purposes, but which don't
1874 use an on-flash partition table such as RedBoot.
1876 Each partition is represented as a sub-node of the flash device.
1877 Each node's name represents the name of the corresponding
1878 partition of the flash device.
1881 - reg : The partition's offset and size within the flash bank.
1882 - label : (optional) The label / name for this flash partition.
1883 If omitted, the label is taken from the node name (excluding
1885 - read-only : (optional) This parameter, if present, is a hint to
1886 Linux that this flash partition should only be mounted
1887 read-only. This is usually used for flash partitions
1888 containing early-boot firmware images or data which should not
1894 compatible = "amd,am29lv128ml", "cfi-flash";
1895 reg = <ff000000 01000000>;
1898 #address-cells = <1>;
1906 reg = <f80000 80000>;
1911 k) Global Utilities Block
1913 The global utilities block controls power management, I/O device
1914 enabling, power-on-reset configuration monitoring, general-purpose
1915 I/O signal configuration, alternate function selection for multiplexed
1916 signals, and clock control.
1918 Required properties:
1920 - compatible : Should define the compatible device type for
1922 - reg : Offset and length of the register set for the device.
1924 Recommended properties:
1926 - fsl,has-rstcr : Indicates that the global utilities register set
1927 contains a functioning "reset control register" (i.e. the board
1928 is wired to reset upon setting the HRESET_REQ bit in this register).
1932 global-utilities@e0000 { /* global utilities block */
1933 compatible = "fsl,mpc8548-guts";
1938 l) Freescale Communications Processor Module
1940 NOTE: This is an interim binding, and will likely change slightly,
1941 as more devices are supported. The QE bindings especially are
1947 - compatible : "fsl,cpm1", "fsl,cpm2", or "fsl,qe".
1948 - reg : A 48-byte region beginning with CPCR.
1952 #address-cells = <1>;
1954 #interrupt-cells = <2>;
1955 compatible = "fsl,mpc8272-cpm", "fsl,cpm2";
1959 ii) Properties common to mulitple CPM/QE devices
1961 - fsl,cpm-command : This value is ORed with the opcode and command flag
1962 to specify the device on which a CPM command operates.
1964 - fsl,cpm-brg : Indicates which baud rate generator the device
1965 is associated with. If absent, an unused BRG
1966 should be dynamically allocated. If zero, the
1967 device uses an external clock rather than a BRG.
1969 - reg : Unless otherwise specified, the first resource represents the
1970 scc/fcc/ucc registers, and the second represents the device's
1971 parameter RAM region (if it has one).
1975 Currently defined compatibles:
1985 device_type = "serial";
1986 compatible = "fsl,mpc8272-scc-uart",
1987 "fsl,cpm2-scc-uart";
1988 reg = <11a00 20 8000 100>;
1989 interrupts = <28 8>;
1990 interrupt-parent = <&PIC>;
1992 fsl,cpm-command = <00800000>;
1997 Currently defined compatibles:
2001 - fsl,cpm2-fcc-enet (third resource is GFEMR)
2007 device_type = "network";
2008 compatible = "fsl,mpc8272-fcc-enet",
2009 "fsl,cpm2-fcc-enet";
2010 reg = <11300 20 8400 100 11390 1>;
2011 local-mac-address = [ 00 00 00 00 00 00 ];
2012 interrupts = <20 8>;
2013 interrupt-parent = <&PIC>;
2014 phy-handle = <&PHY0>;
2015 fsl,cpm-command = <12000300>;
2020 Currently defined compatibles:
2021 fsl,pq1-fec-mdio (reg is same as first resource of FEC device)
2022 fsl,cpm2-mdio-bitbang (reg is port C registers)
2024 Properties for fsl,cpm2-mdio-bitbang:
2025 fsl,mdio-pin : pin of port C controlling mdio data
2026 fsl,mdc-pin : pin of port C controlling mdio clock
2031 device_type = "mdio";
2032 compatible = "fsl,mpc8272ads-mdio-bitbang",
2033 "fsl,mpc8272-mdio-bitbang",
2034 "fsl,cpm2-mdio-bitbang";
2036 #address-cells = <1>;
2038 fsl,mdio-pin = <12>;
2042 v) Baud Rate Generators
2044 Currently defined compatibles:
2050 - reg : There may be an arbitrary number of reg resources; BRG
2051 numbers are assigned to these in order.
2052 - clock-frequency : Specifies the base frequency driving
2058 compatible = "fsl,mpc8272-brg",
2061 reg = <119f0 10 115f0 10>;
2062 clock-frequency = <d#25000000>;
2065 vi) Interrupt Controllers
2067 Currently defined compatibles:
2069 - only one interrupt cell
2072 - second interrupt cell is level/sense:
2078 interrupt-controller@10c00 {
2079 #interrupt-cells = <2>;
2080 interrupt-controller;
2082 compatible = "mpc8272-pic", "fsl,cpm2-pic";
2085 vii) USB (Universal Serial Bus Controller)
2088 - compatible : "fsl,cpm1-usb", "fsl,cpm2-usb", "fsl,qe-usb"
2092 #address-cells = <1>;
2094 compatible = "fsl,cpm2-usb";
2095 reg = <11b60 18 8b00 100>;
2097 interrupt-parent = <&PIC>;
2098 fsl,cpm-command = <2e600000>;
2101 viii) Multi-User RAM (MURAM)
2103 The multi-user/dual-ported RAM is expressed as a bus under the CPM node.
2105 Ranges must be set up subject to the following restrictions:
2107 - Children's reg nodes must be offsets from the start of all muram, even
2108 if the user-data area does not begin at zero.
2109 - If multiple range entries are used, the difference between the parent
2110 address and the child address must be the same in all, so that a single
2111 mapping can cover them all while maintaining the ability to determine
2112 CPM-side offsets with pointer subtraction. It is recommended that
2113 multiple range entries not be used.
2114 - A child address of zero must be translatable, even if no reg resources
2117 A child "data" node must exist, compatible with "fsl,cpm-muram-data", to
2118 indicate the portion of muram that is usable by the OS for arbitrary
2119 purposes. The data node may have an arbitrary number of reg resources,
2120 all of which contribute to the allocatable muram pool.
2122 Example, based on mpc8272:
2125 #address-cells = <1>;
2127 ranges = <0 0 10000>;
2130 compatible = "fsl,cpm-muram-data";
2131 reg = <0 2000 9800 800>;
2135 m) Chipselect/Local Bus
2138 - name : Should be localbus
2139 - #address-cells : Should be either two or three. The first cell is the
2140 chipselect number, and the remaining cells are the
2141 offset into the chipselect.
2142 - #size-cells : Either one or two, depending on how large each chipselect
2144 - ranges : Each range corresponds to a single chipselect, and cover
2145 the entire access window as configured.
2149 compatible = "fsl,mpc8272-localbus",
2151 #address-cells = <2>;
2153 reg = <f0010100 40>;
2155 ranges = <0 0 fe000000 02000000
2156 1 0 f4500000 00008000>;
2159 compatible = "jedec-flash";
2160 reg = <0 0 2000000>;
2167 compatible = "fsl,mpc8272ads-bcsr";
2172 n) 4xx/Axon EMAC ethernet nodes
2174 The EMAC ethernet controller in IBM and AMCC 4xx chips, and also
2175 the Axon bridge. To operate this needs to interact with a ths
2176 special McMAL DMA controller, and sometimes an RGMII or ZMII
2177 interface. In addition to the nodes and properties described
2178 below, the node for the OPB bus on which the EMAC sits must have a
2179 correct clock-frequency property.
2181 i) The EMAC node itself
2183 Required properties:
2184 - device_type : "network"
2186 - compatible : compatible list, contains 2 entries, first is
2187 "ibm,emac-CHIP" where CHIP is the host ASIC (440gx,
2188 405gp, Axon) and second is either "ibm,emac" or
2189 "ibm,emac4". For Axon, thus, we have: "ibm,emac-axon",
2191 - interrupts : <interrupt mapping for EMAC IRQ and WOL IRQ>
2192 - interrupt-parent : optional, if needed for interrupt mapping
2193 - reg : <registers mapping>
2194 - local-mac-address : 6 bytes, MAC address
2195 - mal-device : phandle of the associated McMAL node
2196 - mal-tx-channel : 1 cell, index of the tx channel on McMAL associated
2198 - mal-rx-channel : 1 cell, index of the rx channel on McMAL associated
2200 - cell-index : 1 cell, hardware index of the EMAC cell on a given
2201 ASIC (typically 0x0 and 0x1 for EMAC0 and EMAC1 on
2203 - max-frame-size : 1 cell, maximum frame size supported in bytes
2204 - rx-fifo-size : 1 cell, Rx fifo size in bytes for 10 and 100 Mb/sec
2207 - tx-fifo-size : 1 cell, Tx fifo size in bytes for 10 and 100 Mb/sec
2210 - fifo-entry-size : 1 cell, size of a fifo entry (used to calculate
2212 For Axon, 0x00000010
2213 - mal-burst-size : 1 cell, MAL burst size (used to calculate thresholds)
2215 For Axon, 0x00000100 (I think ...)
2216 - phy-mode : string, mode of operations of the PHY interface.
2217 Supported values are: "mii", "rmii", "smii", "rgmii",
2218 "tbi", "gmii", rtbi", "sgmii".
2219 For Axon on CAB, it is "rgmii"
2220 - mdio-device : 1 cell, required iff using shared MDIO registers
2221 (440EP). phandle of the EMAC to use to drive the
2222 MDIO lines for the PHY used by this EMAC.
2223 - zmii-device : 1 cell, required iff connected to a ZMII. phandle of
2224 the ZMII device node
2225 - zmii-channel : 1 cell, required iff connected to a ZMII. Which ZMII
2226 channel or 0xffffffff if ZMII is only used for MDIO.
2227 - rgmii-device : 1 cell, required iff connected to an RGMII. phandle
2228 of the RGMII device node.
2229 For Axon: phandle of plb5/plb4/opb/rgmii
2230 - rgmii-channel : 1 cell, required iff connected to an RGMII. Which
2231 RGMII channel is used by this EMAC.
2232 Fox Axon: present, whatever value is appropriate for each
2233 EMAC, that is the content of the current (bogus) "phy-port"
2236 Optional properties:
2237 - phy-address : 1 cell, optional, MDIO address of the PHY. If absent,
2238 a search is performed.
2239 - phy-map : 1 cell, optional, bitmap of addresses to probe the PHY
2240 for, used if phy-address is absent. bit 0x00000001 is
2242 For Axon it can be absent, thouugh my current driver
2243 doesn't handle phy-address yet so for now, keep
2245 - rx-fifo-size-gige : 1 cell, Rx fifo size in bytes for 1000 Mb/sec
2246 operations (if absent the value is the same as
2247 rx-fifo-size). For Axon, either absent or 2048.
2248 - tx-fifo-size-gige : 1 cell, Tx fifo size in bytes for 1000 Mb/sec
2249 operations (if absent the value is the same as
2250 tx-fifo-size). For Axon, either absent or 2048.
2251 - tah-device : 1 cell, optional. If connected to a TAH engine for
2252 offload, phandle of the TAH device node.
2253 - tah-channel : 1 cell, optional. If appropriate, channel used on the
2258 EMAC0: ethernet@40000800 {
2259 device_type = "network";
2260 compatible = "ibm,emac-440gp", "ibm,emac";
2261 interrupt-parent = <&UIC1>;
2262 interrupts = <1c 4 1d 4>;
2263 reg = <40000800 70>;
2264 local-mac-address = [00 04 AC E3 1B 1E];
2265 mal-device = <&MAL0>;
2266 mal-tx-channel = <0 1>;
2267 mal-rx-channel = <0>;
2269 max-frame-size = <5dc>;
2270 rx-fifo-size = <1000>;
2271 tx-fifo-size = <800>;
2273 phy-map = <00000001>;
2274 zmii-device = <&ZMII0>;
2280 Required properties:
2281 - device_type : "dma-controller"
2282 - compatible : compatible list, containing 2 entries, first is
2283 "ibm,mcmal-CHIP" where CHIP is the host ASIC (like
2284 emac) and the second is either "ibm,mcmal" or
2286 For Axon, "ibm,mcmal-axon","ibm,mcmal2"
2287 - interrupts : <interrupt mapping for the MAL interrupts sources:
2288 5 sources: tx_eob, rx_eob, serr, txde, rxde>.
2289 For Axon: This is _different_ from the current
2290 firmware. We use the "delayed" interrupts for txeob
2291 and rxeob. Thus we end up with mapping those 5 MPIC
2292 interrupts, all level positive sensitive: 10, 11, 32,
2294 - dcr-reg : < DCR registers range >
2295 - dcr-parent : if needed for dcr-reg
2296 - num-tx-chans : 1 cell, number of Tx channels
2297 - num-rx-chans : 1 cell, number of Rx channels
2301 Required properties:
2302 - compatible : compatible list, containing 2 entries, first is
2303 "ibm,zmii-CHIP" where CHIP is the host ASIC (like
2304 EMAC) and the second is "ibm,zmii".
2305 For Axon, there is no ZMII node.
2306 - reg : <registers mapping>
2310 Required properties:
2311 - compatible : compatible list, containing 2 entries, first is
2312 "ibm,rgmii-CHIP" where CHIP is the host ASIC (like
2313 EMAC) and the second is "ibm,rgmii".
2314 For Axon, "ibm,rgmii-axon","ibm,rgmii"
2315 - reg : <registers mapping>
2316 - revision : as provided by the RGMII new version register if
2318 For Axon: 0x0000012a
2322 The Xilinx EDK toolchain ships with a set of IP cores (devices) for use
2323 in Xilinx Spartan and Virtex FPGAs. The devices cover the whole range
2324 of standard device types (network, serial, etc.) and miscellanious
2325 devices (gpio, LCD, spi, etc). Also, since these devices are
2326 implemented within the fpga fabric every instance of the device can be
2327 synthesised with different options that change the behaviour.
2329 Each IP-core has a set of parameters which the FPGA designer can use to
2330 control how the core is synthesized. Historically, the EDK tool would
2331 extract the device parameters relevant to device drivers and copy them
2332 into an 'xparameters.h' in the form of #define symbols. This tells the
2333 device drivers how the IP cores are configured, but it requres the kernel
2334 to be recompiled every time the FPGA bitstream is resynthesized.
2336 The new approach is to export the parameters into the device tree and
2337 generate a new device tree each time the FPGA bitstream changes. The
2338 parameters which used to be exported as #defines will now become
2339 properties of the device node. In general, device nodes for IP-cores
2340 will take the following form:
2342 (name): (generic-name)@(base-address) {
2343 compatible = "xlnx,(ip-core-name)-(HW_VER)"
2344 [, (list of compatible devices), ...];
2345 reg = <(baseaddr) (size)>;
2346 interrupt-parent = <&interrupt-controller-phandle>;
2347 interrupts = < ... >;
2348 xlnx,(parameter1) = "(string-value)";
2349 xlnx,(parameter2) = <(int-value)>;
2352 (generic-name): an open firmware-style name that describes the
2353 generic class of device. Preferably, this is one word, such
2354 as 'serial' or 'ethernet'.
2355 (ip-core-name): the name of the ip block (given after the BEGIN
2356 directive in system.mhs). Should be in lowercase
2357 and all underscores '_' converted to dashes '-'.
2358 (name): is derived from the "PARAMETER INSTANCE" value.
2359 (parameter#): C_* parameters from system.mhs. The C_ prefix is
2360 dropped from the parameter name, the name is converted
2361 to lowercase and all underscore '_' characters are
2362 converted to dashes '-'.
2363 (baseaddr): the baseaddr parameter value (often named C_BASEADDR).
2364 (HW_VER): from the HW_VER parameter.
2365 (size): the address range size (often C_HIGHADDR - C_BASEADDR + 1).
2367 Typically, the compatible list will include the exact IP core version
2368 followed by an older IP core version which implements the same
2369 interface or any other device with the same interface.
2371 'reg', 'interrupt-parent' and 'interrupts' are all optional properties.
2373 For example, the following block from system.mhs:
2376 PARAMETER INSTANCE = opb_uartlite_0
2377 PARAMETER HW_VER = 1.00.b
2378 PARAMETER C_BAUDRATE = 115200
2379 PARAMETER C_DATA_BITS = 8
2380 PARAMETER C_ODD_PARITY = 0
2381 PARAMETER C_USE_PARITY = 0
2382 PARAMETER C_CLK_FREQ = 50000000
2383 PARAMETER C_BASEADDR = 0xEC100000
2384 PARAMETER C_HIGHADDR = 0xEC10FFFF
2385 BUS_INTERFACE SOPB = opb_7
2386 PORT OPB_Clk = CLK_50MHz
2387 PORT Interrupt = opb_uartlite_0_Interrupt
2388 PORT RX = opb_uartlite_0_RX
2389 PORT TX = opb_uartlite_0_TX
2390 PORT OPB_Rst = sys_bus_reset_0
2393 becomes the following device tree node:
2395 opb_uartlite_0: serial@ec100000 {
2396 device_type = "serial";
2397 compatible = "xlnx,opb-uartlite-1.00.b";
2398 reg = <ec100000 10000>;
2399 interrupt-parent = <&opb_intc_0>;
2400 interrupts = <1 0>; // got this from the opb_intc parameters
2401 current-speed = <d#115200>; // standard serial device prop
2402 clock-frequency = <d#50000000>; // standard serial device prop
2403 xlnx,data-bits = <8>;
2404 xlnx,odd-parity = <0>;
2405 xlnx,use-parity = <0>;
2408 Some IP cores actually implement 2 or more logical devices. In
2409 this case, the device should still describe the whole IP core with
2410 a single node and add a child node for each logical device. The
2411 ranges property can be used to translate from parent IP-core to the
2412 registers of each device. In addition, the parent node should be
2413 compatible with the bus type 'xlnx,compound', and should contain
2414 #address-cells and #size-cells, as with any other bus. (Note: this
2415 makes the assumption that both logical devices have the same bus
2416 binding. If this is not true, then separate nodes should be used
2417 for each logical device). The 'cell-index' property can be used to
2418 enumerate logical devices within an IP core. For example, the
2419 following is the system.mhs entry for the dual ps2 controller found
2420 on the ml403 reference design.
2422 BEGIN opb_ps2_dual_ref
2423 PARAMETER INSTANCE = opb_ps2_dual_ref_0
2424 PARAMETER HW_VER = 1.00.a
2425 PARAMETER C_BASEADDR = 0xA9000000
2426 PARAMETER C_HIGHADDR = 0xA9001FFF
2427 BUS_INTERFACE SOPB = opb_v20_0
2428 PORT Sys_Intr1 = ps2_1_intr
2429 PORT Sys_Intr2 = ps2_2_intr
2430 PORT Clkin1 = ps2_clk_rx_1
2431 PORT Clkin2 = ps2_clk_rx_2
2432 PORT Clkpd1 = ps2_clk_tx_1
2433 PORT Clkpd2 = ps2_clk_tx_2
2434 PORT Rx1 = ps2_d_rx_1
2435 PORT Rx2 = ps2_d_rx_2
2436 PORT Txpd1 = ps2_d_tx_1
2437 PORT Txpd2 = ps2_d_tx_2
2440 It would result in the following device tree nodes:
2442 opb_ps2_dual_ref_0: opb-ps2-dual-ref@a9000000 {
2443 #address-cells = <1>;
2445 compatible = "xlnx,compound";
2446 ranges = <0 a9000000 2000>;
2447 // If this device had extra parameters, then they would
2450 compatible = "xlnx,opb-ps2-dual-ref-1.00.a";
2452 interrupt-parent = <&opb_intc_0>;
2457 compatible = "xlnx,opb-ps2-dual-ref-1.00.a";
2459 interrupt-parent = <&opb_intc_0>;
2465 Also, the system.mhs file defines bus attachments from the processor
2466 to the devices. The device tree structure should reflect the bus
2467 attachments. Again an example; this system.mhs fragment:
2469 BEGIN ppc405_virtex4
2470 PARAMETER INSTANCE = ppc405_0
2471 PARAMETER HW_VER = 1.01.a
2472 BUS_INTERFACE DPLB = plb_v34_0
2473 BUS_INTERFACE IPLB = plb_v34_0
2477 PARAMETER INSTANCE = opb_intc_0
2478 PARAMETER HW_VER = 1.00.c
2479 PARAMETER C_BASEADDR = 0xD1000FC0
2480 PARAMETER C_HIGHADDR = 0xD1000FDF
2481 BUS_INTERFACE SOPB = opb_v20_0
2485 PARAMETER INSTANCE = opb_uart16550_0
2486 PARAMETER HW_VER = 1.00.d
2487 PARAMETER C_BASEADDR = 0xa0000000
2488 PARAMETER C_HIGHADDR = 0xa0001FFF
2489 BUS_INTERFACE SOPB = opb_v20_0
2493 PARAMETER INSTANCE = plb_v34_0
2494 PARAMETER HW_VER = 1.02.a
2497 BEGIN plb_bram_if_cntlr
2498 PARAMETER INSTANCE = plb_bram_if_cntlr_0
2499 PARAMETER HW_VER = 1.00.b
2500 PARAMETER C_BASEADDR = 0xFFFF0000
2501 PARAMETER C_HIGHADDR = 0xFFFFFFFF
2502 BUS_INTERFACE SPLB = plb_v34_0
2505 BEGIN plb2opb_bridge
2506 PARAMETER INSTANCE = plb2opb_bridge_0
2507 PARAMETER HW_VER = 1.01.a
2508 PARAMETER C_RNG0_BASEADDR = 0x20000000
2509 PARAMETER C_RNG0_HIGHADDR = 0x3FFFFFFF
2510 PARAMETER C_RNG1_BASEADDR = 0x60000000
2511 PARAMETER C_RNG1_HIGHADDR = 0x7FFFFFFF
2512 PARAMETER C_RNG2_BASEADDR = 0x80000000
2513 PARAMETER C_RNG2_HIGHADDR = 0xBFFFFFFF
2514 PARAMETER C_RNG3_BASEADDR = 0xC0000000
2515 PARAMETER C_RNG3_HIGHADDR = 0xDFFFFFFF
2516 BUS_INTERFACE SPLB = plb_v34_0
2517 BUS_INTERFACE MOPB = opb_v20_0
2520 Gives this device tree (some properties removed for clarity):
2523 #address-cells = <1>;
2525 compatible = "xlnx,plb-v34-1.02.a";
2526 device_type = "ibm,plb";
2527 ranges; // 1:1 translation
2529 plb_bram_if_cntrl_0: bram@ffff0000 {
2530 reg = <ffff0000 10000>;
2534 #address-cells = <1>;
2536 ranges = <20000000 20000000 20000000
2537 60000000 60000000 20000000
2538 80000000 80000000 40000000
2539 c0000000 c0000000 20000000>;
2541 opb_uart16550_0: serial@a0000000 {
2542 reg = <a00000000 2000>;
2545 opb_intc_0: interrupt-controller@d1000fc0 {
2546 reg = <d1000fc0 20>;
2551 That covers the general approach to binding xilinx IP cores into the
2552 device tree. The following are bindings for specific devices:
2554 i) Xilinx ML300 Framebuffer
2556 Simple framebuffer device from the ML300 reference design (also on the
2557 ML403 reference design as well as others).
2559 Optional properties:
2560 - resolution = <xres yres> : pixel resolution of framebuffer. Some
2561 implementations use a different resolution.
2562 Default is <d#640 d#480>
2563 - virt-resolution = <xvirt yvirt> : Size of framebuffer in memory.
2564 Default is <d#1024 d#480>.
2565 - rotate-display (empty) : rotate display 180 degrees.
2567 ii) Xilinx SystemACE
2569 The Xilinx SystemACE device is used to program FPGAs from an FPGA
2570 bitstream stored on a CF card. It can also be used as a generic CF
2573 Optional properties:
2574 - 8-bit (empty) : Set this property for SystemACE in 8 bit mode
2576 iii) Xilinx EMAC and Xilinx TEMAC
2578 Xilinx Ethernet devices. In addition to general xilinx properties
2579 listed above, nodes for these devices should include a phy-handle
2580 property, and may include other common network device properties
2581 like local-mac-address.
2585 Xilinx uartlite devices are simple fixed speed serial ports.
2588 - current-speed : Baud rate of uartlite
2592 Xilinx hwicap devices provide access to the configuration logic
2593 of the FPGA through the Internal Configuration Access Port
2594 (ICAP). The ICAP enables partial reconfiguration of the FPGA,
2595 readback of the configuration information, and some control over
2596 'warm boots' of the FPGA fabric.
2598 Required properties:
2599 - xlnx,family : The family of the FPGA, necessary since the
2600 capabilities of the underlying ICAP hardware
2601 differ between different families. May be
2602 'virtex2p', 'virtex4', or 'virtex5'.
2604 vi) Xilinx Uart 16550
2606 Xilinx UART 16550 devices are very similar to the NS16550 but with
2607 different register spacing and an offset from the base address.
2610 - clock-frequency : Frequency of the clock input
2611 - reg-offset : A value of 3 is required
2612 - reg-shift : A value of 2 is required
2615 p) Freescale Synchronous Serial Interface
2617 The SSI is a serial device that communicates with audio codecs. It can
2618 be programmed in AC97, I2S, left-justified, or right-justified modes.
2620 Required properties:
2621 - compatible : compatible list, containing "fsl,ssi"
2622 - cell-index : the SSI, <0> = SSI1, <1> = SSI2, and so on
2623 - reg : offset and length of the register set for the device
2624 - interrupts : <a b> where a is the interrupt number and b is a
2625 field that represents an encoding of the sense and
2626 level information for the interrupt. This should be
2627 encoded based on the information in section 2)
2628 depending on the type of interrupt controller you
2630 - interrupt-parent : the phandle for the interrupt controller that
2631 services interrupts for this device.
2632 - fsl,mode : the operating mode for the SSI interface
2633 "i2s-slave" - I2S mode, SSI is clock slave
2634 "i2s-master" - I2S mode, SSI is clock master
2635 "lj-slave" - left-justified mode, SSI is clock slave
2636 "lj-master" - l.j. mode, SSI is clock master
2637 "rj-slave" - right-justified mode, SSI is clock slave
2638 "rj-master" - r.j., SSI is clock master
2639 "ac97-slave" - AC97 mode, SSI is clock slave
2640 "ac97-master" - AC97 mode, SSI is clock master
2642 Optional properties:
2643 - codec-handle : phandle to a 'codec' node that defines an audio
2644 codec connected to this SSI. This node is typically
2645 a child of an I2C or other control node.
2647 Child 'codec' node required properties:
2648 - compatible : compatible list, contains the name of the codec
2650 Child 'codec' node optional properties:
2651 - clock-frequency : The frequency of the input clock, which typically
2652 comes from an on-board dedicated oscillator.
2654 * Freescale 83xx DMA Controller
2656 Freescale PowerPC 83xx have on chip general purpose DMA controllers.
2658 Required properties:
2660 - compatible : compatible list, contains 2 entries, first is
2661 "fsl,CHIP-dma", where CHIP is the processor
2662 (mpc8349, mpc8360, etc.) and the second is
2664 - reg : <registers mapping for DMA general status reg>
2665 - ranges : Should be defined as specified in 1) to describe the
2666 DMA controller channels.
2667 - cell-index : controller index. 0 for controller @ 0x8100
2668 - interrupts : <interrupt mapping for DMA IRQ>
2669 - interrupt-parent : optional, if needed for interrupt mapping
2672 - DMA channel nodes:
2673 - compatible : compatible list, contains 2 entries, first is
2674 "fsl,CHIP-dma-channel", where CHIP is the processor
2675 (mpc8349, mpc8350, etc.) and the second is
2676 "fsl,elo-dma-channel"
2677 - reg : <registers mapping for channel>
2678 - cell-index : dma channel index starts at 0.
2680 Optional properties:
2681 - interrupts : <interrupt mapping for DMA channel IRQ>
2682 (on 83xx this is expected to be identical to
2683 the interrupts property of the parent node)
2684 - interrupt-parent : optional, if needed for interrupt mapping
2688 #address-cells = <1>;
2690 compatible = "fsl,mpc8349-dma", "fsl,elo-dma";
2692 ranges = <0 8100 1a4>;
2693 interrupt-parent = <&ipic>;
2694 interrupts = <47 8>;
2697 compatible = "fsl,mpc8349-dma-channel", "fsl,elo-dma-channel";
2702 compatible = "fsl,mpc8349-dma-channel", "fsl,elo-dma-channel";
2707 compatible = "fsl,mpc8349-dma-channel", "fsl,elo-dma-channel";
2712 compatible = "fsl,mpc8349-dma-channel", "fsl,elo-dma-channel";
2718 * Freescale 85xx/86xx DMA Controller
2720 Freescale PowerPC 85xx/86xx have on chip general purpose DMA controllers.
2722 Required properties:
2724 - compatible : compatible list, contains 2 entries, first is
2725 "fsl,CHIP-dma", where CHIP is the processor
2726 (mpc8540, mpc8540, etc.) and the second is
2728 - reg : <registers mapping for DMA general status reg>
2729 - cell-index : controller index. 0 for controller @ 0x21000,
2730 1 for controller @ 0xc000
2731 - ranges : Should be defined as specified in 1) to describe the
2732 DMA controller channels.
2734 - DMA channel nodes:
2735 - compatible : compatible list, contains 2 entries, first is
2736 "fsl,CHIP-dma-channel", where CHIP is the processor
2737 (mpc8540, mpc8560, etc.) and the second is
2738 "fsl,eloplus-dma-channel"
2739 - cell-index : dma channel index starts at 0.
2740 - reg : <registers mapping for channel>
2741 - interrupts : <interrupt mapping for DMA channel IRQ>
2742 - interrupt-parent : optional, if needed for interrupt mapping
2746 #address-cells = <1>;
2748 compatible = "fsl,mpc8540-dma", "fsl,eloplus-dma";
2750 ranges = <0 21100 200>;
2753 compatible = "fsl,mpc8540-dma-channel", "fsl,eloplus-dma-channel";
2756 interrupt-parent = <&mpic>;
2757 interrupts = <14 2>;
2760 compatible = "fsl,mpc8540-dma-channel", "fsl,eloplus-dma-channel";
2763 interrupt-parent = <&mpic>;
2764 interrupts = <15 2>;
2767 compatible = "fsl,mpc8540-dma-channel", "fsl,eloplus-dma-channel";
2770 interrupt-parent = <&mpic>;
2771 interrupts = <16 2>;
2774 compatible = "fsl,mpc8540-dma-channel", "fsl,eloplus-dma-channel";
2777 interrupt-parent = <&mpic>;
2778 interrupts = <17 2>;
2782 * Freescale 8xxx/3.0 Gb/s SATA nodes
2784 SATA nodes are defined to describe on-chip Serial ATA controllers.
2785 Each SATA port should have its own node.
2787 Required properties:
2788 - compatible : compatible list, contains 2 entries, first is
2789 "fsl,CHIP-sata", where CHIP is the processor
2790 (mpc8315, mpc8379, etc.) and the second is
2792 - interrupts : <interrupt mapping for SATA IRQ>
2793 - cell-index : controller index.
2794 1 for controller @ 0x18000
2795 2 for controller @ 0x19000
2796 3 for controller @ 0x1a000
2797 4 for controller @ 0x1b000
2799 Optional properties:
2800 - interrupt-parent : optional, if needed for interrupt mapping
2801 - reg : <registers mapping>
2806 compatible = "fsl,mpc8379-sata", "fsl,pq-sata";
2807 reg = <0x18000 0x1000>;
2809 interrupts = <2c 8>;
2810 interrupt-parent = < &ipic >;
2813 q) USB EHCI controllers
2815 Required properties:
2816 - compatible : should be "usb-ehci".
2817 - reg : should contain at least address and length of the standard EHCI
2818 register set for the device. Optional platform-dependent registers
2819 (debug-port or other) can be also specified here, but only after
2820 definition of standard EHCI registers.
2821 - interrupts : one EHCI interrupt should be described here.
2822 If device registers are implemented in big endian mode, the device
2823 node should have "big-endian-regs" property.
2824 If controller implementation operates with big endian descriptors,
2825 "big-endian-desc" property should be specified.
2826 If both big endian registers and descriptors are used by the controller
2827 implementation, "big-endian" property can be specified instead of having
2828 both "big-endian-regs" and "big-endian-desc".
2830 Example (Sequoia 440EPx):
2832 compatible = "ibm,usb-ehci-440epx", "usb-ehci";
2833 interrupt-parent = <&UIC0>;
2834 interrupts = <1a 4>;
2835 reg = <0 e0000300 90 0 e0000390 70>;
2839 r) Freescale Display Interface Unit
2841 The Freescale DIU is a LCD controller, with proper hardware, it can also
2844 Required properties:
2845 - compatible : should be "fsl-diu".
2846 - reg : should contain at least address and length of the DIU register
2848 - Interrupts : one DIU interrupt should be describe here.
2850 Example (MPC8610HPCD)
2852 compatible = "fsl,diu";
2853 reg = <0x2c000 100>;
2854 interrupts = <72 2>;
2855 interrupt-parent = <&mpic>;
2858 s) Freescale on board FPGA
2860 This is the memory-mapped registers for on board FPGA.
2862 Required properities:
2863 - compatible : should be "fsl,fpga-pixis".
2864 - reg : should contain the address and the lenght of the FPPGA register
2867 Example (MPC8610HPCD)
2868 board-control@e8000000 {
2869 compatible = "fsl,fpga-pixis";
2870 reg = <0xe8000000 32>;
2873 VII - Marvell Discovery mv64[345]6x System Controller chips
2874 ===========================================================
2876 The Marvell mv64[345]60 series of system controller chips contain
2877 many of the peripherals needed to implement a complete computer
2878 system. In this section, we define device tree nodes to describe
2879 the system controller chip itself and each of the peripherals
2880 which it contains. Compatible string values for each node are
2881 prefixed with the string "marvell,", for Marvell Technology Group Ltd.
2883 1) The /system-controller node
2885 This node is used to represent the system-controller and must be
2886 present when the system uses a system contller chip. The top-level
2887 system-controller node contains information that is global to all
2888 devices within the system controller chip. The node name begins
2889 with "system-controller" followed by the unit address, which is
2890 the base address of the memory-mapped register set for the system
2893 Required properties:
2895 - ranges : Describes the translation of system controller addresses
2896 for memory mapped registers.
2897 - clock-frequency: Contains the main clock frequency for the system
2899 - reg : This property defines the address and size of the
2900 memory-mapped registers contained within the system controller
2901 chip. The address specified in the "reg" property should match
2902 the unit address of the system-controller node.
2903 - #address-cells : Address representation for system controller
2904 devices. This field represents the number of cells needed to
2905 represent the address of the memory-mapped registers of devices
2906 within the system controller chip.
2907 - #size-cells : Size representation for for the memory-mapped
2908 registers within the system controller chip.
2909 - #interrupt-cells : Defines the width of cells used to represent
2912 Optional properties:
2914 - model : The specific model of the system controller chip. Such
2915 as, "mv64360", "mv64460", or "mv64560".
2916 - compatible : A string identifying the compatibility identifiers
2917 of the system controller chip.
2919 The system-controller node contains child nodes for each system
2920 controller device that the platform uses. Nodes should not be created
2921 for devices which exist on the system controller chip but are not used
2923 Example Marvell Discovery mv64360 system-controller node:
2925 system-controller@f1000000 { /* Marvell Discovery mv64360 */
2926 #address-cells = <1>;
2928 model = "mv64360"; /* Default */
2929 compatible = "marvell,mv64360";
2930 clock-frequency = <133333333>;
2931 reg = <0xf1000000 0x10000>;
2932 virtual-reg = <0xf1000000>;
2933 ranges = <0x88000000 0x88000000 0x1000000 /* PCI 0 I/O Space */
2934 0x80000000 0x80000000 0x8000000 /* PCI 0 MEM Space */
2935 0xa0000000 0xa0000000 0x4000000 /* User FLASH */
2936 0x00000000 0xf1000000 0x0010000 /* Bridge's regs */
2937 0xf2000000 0xf2000000 0x0040000>;/* Integrated SRAM */
2939 [ child node definitions... ]
2942 2) Child nodes of /system-controller
2944 a) Marvell Discovery MDIO bus
2946 The MDIO is a bus to which the PHY devices are connected. For each
2947 device that exists on this bus, a child node should be created. See
2948 the definition of the PHY node below for an example of how to define
2951 Required properties:
2952 - #address-cells : Should be <1>
2953 - #size-cells : Should be <0>
2954 - device_type : Should be "mdio"
2955 - compatible : Should be "marvell,mv64360-mdio"
2960 #address-cells = <1>;
2962 device_type = "mdio";
2963 compatible = "marvell,mv64360-mdio";
2971 b) Marvell Discovery ethernet controller
2973 The Discover ethernet controller is described with two levels
2974 of nodes. The first level describes an ethernet silicon block
2975 and the second level describes up to 3 ethernet nodes within
2976 that block. The reason for the multiple levels is that the
2977 registers for the node are interleaved within a single set
2978 of registers. The "ethernet-block" level describes the
2979 shared register set, and the "ethernet" nodes describe ethernet
2980 port-specific properties.
2984 Required properties:
2985 - #address-cells : <1>
2987 - compatible : "marvell,mv64360-eth-block"
2988 - reg : Offset and length of the register set for this block
2990 Example Discovery Ethernet block node:
2991 ethernet-block@2000 {
2992 #address-cells = <1>;
2994 compatible = "marvell,mv64360-eth-block";
2995 reg = <0x2000 0x2000>;
3003 Required properties:
3004 - device_type : Should be "network".
3005 - compatible : Should be "marvell,mv64360-eth".
3006 - reg : Should be <0>, <1>, or <2>, according to which registers
3007 within the silicon block the device uses.
3008 - interrupts : <a> where a is the interrupt number for the port.
3009 - interrupt-parent : the phandle for the interrupt controller
3010 that services interrupts for this device.
3011 - phy : the phandle for the PHY connected to this ethernet
3013 - local-mac-address : 6 bytes, MAC address
3015 Example Discovery Ethernet port node:
3017 device_type = "network";
3018 compatible = "marvell,mv64360-eth";
3021 interrupt-parent = <&PIC>;
3023 local-mac-address = [ 00 00 00 00 00 00 ];
3028 c) Marvell Discovery PHY nodes
3030 Required properties:
3031 - device_type : Should be "ethernet-phy"
3032 - interrupts : <a> where a is the interrupt number for this phy.
3033 - interrupt-parent : the phandle for the interrupt controller that
3034 services interrupts for this device.
3035 - reg : The ID number for the phy, usually a small integer
3037 Example Discovery PHY node:
3039 device_type = "ethernet-phy";
3040 compatible = "broadcom,bcm5421";
3041 interrupts = <76>; /* GPP 12 */
3042 interrupt-parent = <&PIC>;
3047 d) Marvell Discovery SDMA nodes
3049 Represent DMA hardware associated with the MPSC (multiprotocol
3050 serial controllers).
3052 Required properties:
3053 - compatible : "marvell,mv64360-sdma"
3054 - reg : Offset and length of the register set for this device
3055 - interrupts : <a> where a is the interrupt number for the DMA
3057 - interrupt-parent : the phandle for the interrupt controller
3058 that services interrupts for this device.
3060 Example Discovery SDMA node:
3062 compatible = "marvell,mv64360-sdma";
3063 reg = <0x4000 0xc18>;
3064 virtual-reg = <0xf1004000>;
3066 interrupt-parent = <&PIC>;
3070 e) Marvell Discovery BRG nodes
3072 Represent baud rate generator hardware associated with the MPSC
3073 (multiprotocol serial controllers).
3075 Required properties:
3076 - compatible : "marvell,mv64360-brg"
3077 - reg : Offset and length of the register set for this device
3078 - clock-src : A value from 0 to 15 which selects the clock
3079 source for the baud rate generator. This value corresponds
3080 to the CLKS value in the BRGx configuration register. See
3081 the mv64x60 User's Manual.
3082 - clock-frequence : The frequency (in Hz) of the baud rate
3083 generator's input clock.
3084 - current-speed : The current speed setting (presumably by
3085 firmware) of the baud rate generator.
3087 Example Discovery BRG node:
3089 compatible = "marvell,mv64360-brg";
3092 clock-frequency = <133333333>;
3093 current-speed = <9600>;
3097 f) Marvell Discovery CUNIT nodes
3099 Represent the Serial Communications Unit device hardware.
3101 Required properties:
3102 - reg : Offset and length of the register set for this device
3104 Example Discovery CUNIT node:
3106 reg = <0xf200 0x200>;
3110 g) Marvell Discovery MPSCROUTING nodes
3112 Represent the Discovery's MPSC routing hardware
3114 Required properties:
3115 - reg : Offset and length of the register set for this device
3117 Example Discovery CUNIT node:
3123 h) Marvell Discovery MPSCINTR nodes
3125 Represent the Discovery's MPSC DMA interrupt hardware registers
3126 (SDMA cause and mask registers).
3128 Required properties:
3129 - reg : Offset and length of the register set for this device
3131 Example Discovery MPSCINTR node:
3133 reg = <0xb800 0x100>;
3137 i) Marvell Discovery MPSC nodes
3139 Represent the Discovery's MPSC (Multiprotocol Serial Controller)
3142 Required properties:
3143 - device_type : "serial"
3144 - compatible : "marvell,mv64360-mpsc"
3145 - reg : Offset and length of the register set for this device
3146 - sdma : the phandle for the SDMA node used by this port
3147 - brg : the phandle for the BRG node used by this port
3148 - cunit : the phandle for the CUNIT node used by this port
3149 - mpscrouting : the phandle for the MPSCROUTING node used by this port
3150 - mpscintr : the phandle for the MPSCINTR node used by this port
3151 - cell-index : the hardware index of this cell in the MPSC core
3152 - max_idle : value needed for MPSC CHR3 (Maximum Frame Length)
3154 - interrupts : <a> where a is the interrupt number for the MPSC.
3155 - interrupt-parent : the phandle for the interrupt controller
3156 that services interrupts for this device.
3158 Example Discovery MPSCINTR node:
3160 device_type = "serial";
3161 compatible = "marvell,mv64360-mpsc";
3162 reg = <0x8000 0x38>;
3163 virtual-reg = <0xf1008000>;
3167 mpscrouting = <&MPSCROUTING>;
3168 mpscintr = <&MPSCINTR>;
3172 interrupt-parent = <&PIC>;
3176 j) Marvell Discovery Watch Dog Timer nodes
3178 Represent the Discovery's watchdog timer hardware
3180 Required properties:
3181 - compatible : "marvell,mv64360-wdt"
3182 - reg : Offset and length of the register set for this device
3184 Example Discovery Watch Dog Timer node:
3186 compatible = "marvell,mv64360-wdt";
3191 k) Marvell Discovery I2C nodes
3193 Represent the Discovery's I2C hardware
3195 Required properties:
3196 - device_type : "i2c"
3197 - compatible : "marvell,mv64360-i2c"
3198 - reg : Offset and length of the register set for this device
3199 - interrupts : <a> where a is the interrupt number for the I2C.
3200 - interrupt-parent : the phandle for the interrupt controller
3201 that services interrupts for this device.
3203 Example Discovery I2C node:
3204 compatible = "marvell,mv64360-i2c";
3205 reg = <0xc000 0x20>;
3206 virtual-reg = <0xf100c000>;
3208 interrupt-parent = <&PIC>;
3212 l) Marvell Discovery PIC (Programmable Interrupt Controller) nodes
3214 Represent the Discovery's PIC hardware
3216 Required properties:
3217 - #interrupt-cells : <1>
3218 - #address-cells : <0>
3219 - compatible : "marvell,mv64360-pic"
3220 - reg : Offset and length of the register set for this device
3221 - interrupt-controller
3223 Example Discovery PIC node:
3225 #interrupt-cells = <1>;
3226 #address-cells = <0>;
3227 compatible = "marvell,mv64360-pic";
3229 interrupt-controller;
3233 m) Marvell Discovery MPP (Multipurpose Pins) multiplexing nodes
3235 Represent the Discovery's MPP hardware
3237 Required properties:
3238 - compatible : "marvell,mv64360-mpp"
3239 - reg : Offset and length of the register set for this device
3241 Example Discovery MPP node:
3243 compatible = "marvell,mv64360-mpp";
3244 reg = <0xf000 0x10>;
3248 n) Marvell Discovery GPP (General Purpose Pins) nodes
3250 Represent the Discovery's GPP hardware
3252 Required properties:
3253 - compatible : "marvell,mv64360-gpp"
3254 - reg : Offset and length of the register set for this device
3256 Example Discovery GPP node:
3258 compatible = "marvell,mv64360-gpp";
3259 reg = <0xf100 0x20>;
3263 o) Marvell Discovery PCI host bridge node
3265 Represents the Discovery's PCI host bridge device. The properties
3266 for this node conform to Rev 2.1 of the PCI Bus Binding to IEEE
3267 1275-1994. A typical value for the compatible property is
3268 "marvell,mv64360-pci".
3270 Example Discovery PCI host bridge node
3272 #address-cells = <3>;
3274 #interrupt-cells = <1>;
3275 device_type = "pci";
3276 compatible = "marvell,mv64360-pci";
3278 ranges = <0x01000000 0x0 0x0
3279 0x88000000 0x0 0x01000000
3280 0x02000000 0x0 0x80000000
3281 0x80000000 0x0 0x08000000>;
3282 bus-range = <0 255>;
3283 clock-frequency = <66000000>;
3284 interrupt-parent = <&PIC>;
3285 interrupt-map-mask = <0xf800 0x0 0x0 0x7>;
3288 0x5000 0 0 1 &PIC 80
3289 0x5000 0 0 2 &PIC 81
3290 0x5000 0 0 3 &PIC 91
3291 0x5000 0 0 4 &PIC 93
3294 0x5800 0 0 1 &PIC 91
3295 0x5800 0 0 2 &PIC 93
3296 0x5800 0 0 3 &PIC 80
3297 0x5800 0 0 4 &PIC 81
3300 0x6000 0 0 1 &PIC 91
3301 0x6000 0 0 2 &PIC 93
3302 0x6000 0 0 3 &PIC 80
3303 0x6000 0 0 4 &PIC 81
3306 0x6800 0 0 1 &PIC 93
3307 0x6800 0 0 2 &PIC 80
3308 0x6800 0 0 3 &PIC 81
3309 0x6800 0 0 4 &PIC 91
3314 p) Marvell Discovery CPU Error nodes
3316 Represent the Discovery's CPU error handler device.
3318 Required properties:
3319 - compatible : "marvell,mv64360-cpu-error"
3320 - reg : Offset and length of the register set for this device
3321 - interrupts : the interrupt number for this device
3322 - interrupt-parent : the phandle for the interrupt controller
3323 that services interrupts for this device.
3325 Example Discovery CPU Error node:
3327 compatible = "marvell,mv64360-cpu-error";
3328 reg = <0x70 0x10 0x128 0x28>;
3330 interrupt-parent = <&PIC>;
3334 q) Marvell Discovery SRAM Controller nodes
3336 Represent the Discovery's SRAM controller device.
3338 Required properties:
3339 - compatible : "marvell,mv64360-sram-ctrl"
3340 - reg : Offset and length of the register set for this device
3341 - interrupts : the interrupt number for this device
3342 - interrupt-parent : the phandle for the interrupt controller
3343 that services interrupts for this device.
3345 Example Discovery SRAM Controller node:
3347 compatible = "marvell,mv64360-sram-ctrl";
3350 interrupt-parent = <&PIC>;
3354 r) Marvell Discovery PCI Error Handler nodes
3356 Represent the Discovery's PCI error handler device.
3358 Required properties:
3359 - compatible : "marvell,mv64360-pci-error"
3360 - reg : Offset and length of the register set for this device
3361 - interrupts : the interrupt number for this device
3362 - interrupt-parent : the phandle for the interrupt controller
3363 that services interrupts for this device.
3365 Example Discovery PCI Error Handler node:
3367 compatible = "marvell,mv64360-pci-error";
3368 reg = <0x1d40 0x40 0xc28 0x4>;
3370 interrupt-parent = <&PIC>;
3374 s) Marvell Discovery Memory Controller nodes
3376 Represent the Discovery's memory controller device.
3378 Required properties:
3379 - compatible : "marvell,mv64360-mem-ctrl"
3380 - reg : Offset and length of the register set for this device
3381 - interrupts : the interrupt number for this device
3382 - interrupt-parent : the phandle for the interrupt controller
3383 that services interrupts for this device.
3385 Example Discovery Memory Controller node:
3387 compatible = "marvell,mv64360-mem-ctrl";
3388 reg = <0x1400 0x60>;
3390 interrupt-parent = <&PIC>;
3394 VIII - Specifying interrupt information for devices
3395 ===================================================
3397 The device tree represents the busses and devices of a hardware
3398 system in a form similar to the physical bus topology of the
3401 In addition, a logical 'interrupt tree' exists which represents the
3402 hierarchy and routing of interrupts in the hardware.
3404 The interrupt tree model is fully described in the
3405 document "Open Firmware Recommended Practice: Interrupt
3406 Mapping Version 0.9". The document is available at:
3407 <http://playground.sun.com/1275/practice>.
3409 1) interrupts property
3410 ----------------------
3412 Devices that generate interrupts to a single interrupt controller
3413 should use the conventional OF representation described in the
3414 OF interrupt mapping documentation.
3416 Each device which generates interrupts must have an 'interrupt'
3417 property. The interrupt property value is an arbitrary number of
3418 of 'interrupt specifier' values which describe the interrupt or
3419 interrupts for the device.
3421 The encoding of an interrupt specifier is determined by the
3422 interrupt domain in which the device is located in the
3423 interrupt tree. The root of an interrupt domain specifies in
3424 its #interrupt-cells property the number of 32-bit cells
3425 required to encode an interrupt specifier. See the OF interrupt
3426 mapping documentation for a detailed description of domains.
3428 For example, the binding for the OpenPIC interrupt controller
3429 specifies an #interrupt-cells value of 2 to encode the interrupt
3430 number and level/sense information. All interrupt children in an
3431 OpenPIC interrupt domain use 2 cells per interrupt in their interrupts
3434 The PCI bus binding specifies a #interrupt-cell value of 1 to encode
3435 which interrupt pin (INTA,INTB,INTC,INTD) is used.
3437 2) interrupt-parent property
3438 ----------------------------
3440 The interrupt-parent property is specified to define an explicit
3441 link between a device node and its interrupt parent in
3442 the interrupt tree. The value of interrupt-parent is the
3443 phandle of the parent node.
3445 If the interrupt-parent property is not defined for a node, it's
3446 interrupt parent is assumed to be an ancestor in the node's
3447 _device tree_ hierarchy.
3449 3) OpenPIC Interrupt Controllers
3450 --------------------------------
3452 OpenPIC interrupt controllers require 2 cells to encode
3453 interrupt information. The first cell defines the interrupt
3454 number. The second cell defines the sense and level
3457 Sense and level information should be encoded as follows:
3459 0 = low to high edge sensitive type enabled
3460 1 = active low level sensitive type enabled
3461 2 = active high level sensitive type enabled
3462 3 = high to low edge sensitive type enabled
3464 4) ISA Interrupt Controllers
3465 ----------------------------
3467 ISA PIC interrupt controllers require 2 cells to encode
3468 interrupt information. The first cell defines the interrupt
3469 number. The second cell defines the sense and level
3472 ISA PIC interrupt controllers should adhere to the ISA PIC
3473 encodings listed below:
3475 0 = active low level sensitive type enabled
3476 1 = active high level sensitive type enabled
3477 2 = high to low edge sensitive type enabled
3478 3 = low to high edge sensitive type enabled
3480 VIII - Specifying GPIO information for devices
3481 ==============================================
3486 Nodes that makes use of GPIOs should define them using `gpios' property,
3487 format of which is: <&gpio-controller1-phandle gpio1-specifier
3488 &gpio-controller2-phandle gpio2-specifier
3489 0 /* holes are permitted, means no GPIO 3 */
3490 &gpio-controller4-phandle gpio4-specifier
3493 Note that gpio-specifier length is controller dependent.
3495 gpio-specifier may encode: bank, pin position inside the bank,
3496 whether pin is open-drain and whether pin is logically inverted.
3498 Example of the node using GPIOs:
3501 gpios = <&qe_pio_e 18 0>;
3504 In this example gpio-specifier is "18 0" and encodes GPIO pin number,
3505 and empty GPIO flags as accepted by the "qe_pio_e" gpio-controller.
3507 2) gpio-controller nodes
3508 ------------------------
3510 Every GPIO controller node must have #gpio-cells property defined,
3511 this information will be used to translate gpio-specifiers.
3513 Example of two SOC GPIO banks defined as gpio-controller nodes:
3515 qe_pio_a: gpio-controller@1400 {
3517 compatible = "fsl,qe-pario-bank-a", "fsl,qe-pario-bank";
3518 reg = <0x1400 0x18>;
3522 qe_pio_e: gpio-controller@1460 {
3524 compatible = "fsl,qe-pario-bank-e", "fsl,qe-pario-bank";
3525 reg = <0x1460 0x18>;
3529 Appendix A - Sample SOC node for MPC8540
3530 ========================================
3532 Note that the #address-cells and #size-cells for the SoC node
3533 in this example have been explicitly listed; these are likely
3534 not necessary as they are usually the same as the root node.
3537 #address-cells = <1>;
3539 #interrupt-cells = <2>;
3540 device_type = "soc";
3541 ranges = <00000000 e0000000 00100000>
3542 reg = <e0000000 00003000>;
3543 bus-frequency = <0>;
3547 device_type = "mdio";
3548 compatible = "gianfar";
3551 linux,phandle = <2452000>
3552 interrupt-parent = <40000>;
3553 interrupts = <35 1>;
3555 device_type = "ethernet-phy";
3559 linux,phandle = <2452001>
3560 interrupt-parent = <40000>;
3561 interrupts = <35 1>;
3563 device_type = "ethernet-phy";
3567 linux,phandle = <2452002>
3568 interrupt-parent = <40000>;
3569 interrupts = <35 1>;
3571 device_type = "ethernet-phy";
3578 device_type = "network";
3580 compatible = "gianfar";
3582 mac-address = [ 00 E0 0C 00 73 00 ];
3583 interrupts = <d 3 e 3 12 3>;
3584 interrupt-parent = <40000>;
3585 phy-handle = <2452000>;
3589 #address-cells = <1>;
3591 device_type = "network";
3593 compatible = "gianfar";
3595 mac-address = [ 00 E0 0C 00 73 01 ];
3596 interrupts = <13 3 14 3 18 3>;
3597 interrupt-parent = <40000>;
3598 phy-handle = <2452001>;
3602 #address-cells = <1>;
3604 device_type = "network";
3606 compatible = "gianfar";
3608 mac-address = [ 00 E0 0C 00 73 02 ];
3609 interrupts = <19 3>;
3610 interrupt-parent = <40000>;
3611 phy-handle = <2452002>;
3615 device_type = "serial";
3616 compatible = "ns16550";
3618 clock-frequency = <0>;
3619 interrupts = <1a 3>;
3620 interrupt-parent = <40000>;
3624 linux,phandle = <40000>;
3625 clock-frequency = <0>;
3626 interrupt-controller;
3627 #address-cells = <0>;
3628 reg = <40000 40000>;
3630 compatible = "chrp,open-pic";
3631 device_type = "open-pic";
3636 interrupt-parent = <40000>;
3637 interrupts = <1b 3>;
3639 device_type = "i2c";
3640 compatible = "fsl-i2c";