1Remote Processor Framework
2
31. Introduction
4
5Modern SoCs typically have heterogeneous remote processor devices in asymmetric
6multiprocessing (AMP) configurations, which may be running different instances
7of operating system, whether it's Linux or any other flavor of real-time OS.
8
9OMAP4, for example, has dual Cortex-A9, dual Cortex-M3 and a C64x+ DSP.
10In a typical configuration, the dual cortex-A9 is running Linux in a SMP
11configuration, and each of the other three cores (two M3 cores and a DSP)
12is running its own instance of RTOS in an AMP configuration.
13
14The remoteproc framework allows different platforms/architectures to
15control (power on, load firmware, power off) those remote processors while
16abstracting the hardware differences, so the entire driver doesn't need to be
17duplicated. In addition, this framework also adds rpmsg virtio devices
18for remote processors that supports this kind of communication. This way,
19platform-specific remoteproc drivers only need to provide a few low-level
20handlers, and then all rpmsg drivers will then just work
21(for more information about the virtio-based rpmsg bus and its drivers,
22please read Documentation/rpmsg.txt).
23Registration of other types of virtio devices is now also possible. Firmwares
24just need to publish what kind of virtio devices do they support, and then
25remoteproc will add those devices. This makes it possible to reuse the
26existing virtio drivers with remote processor backends at a minimal development
27cost.
28
292. User API
30
31  int rproc_boot(struct rproc *rproc)
32    - Boot a remote processor (i.e. load its firmware, power it on, ...).
33      If the remote processor is already powered on, this function immediately
34      returns (successfully).
35      Returns 0 on success, and an appropriate error value otherwise.
36      Note: to use this function you should already have a valid rproc
37      handle. There are several ways to achieve that cleanly (devres, pdata,
38      the way remoteproc_rpmsg.c does this, or, if this becomes prevalent, we
39      might also consider using dev_archdata for this).
40
41  void rproc_shutdown(struct rproc *rproc)
42    - Power off a remote processor (previously booted with rproc_boot()).
43      In case @rproc is still being used by an additional user(s), then
44      this function will just decrement the power refcount and exit,
45      without really powering off the device.
46      Every call to rproc_boot() must (eventually) be accompanied by a call
47      to rproc_shutdown(). Calling rproc_shutdown() redundantly is a bug.
48      Notes:
49      - we're not decrementing the rproc's refcount, only the power refcount.
50        which means that the @rproc handle stays valid even after
51        rproc_shutdown() returns, and users can still use it with a subsequent
52        rproc_boot(), if needed.
53
543. Typical usage
55
56#include <linux/remoteproc.h>
57
58/* in case we were given a valid 'rproc' handle */
59int dummy_rproc_example(struct rproc *my_rproc)
60{
61	int ret;
62
63	/* let's power on and boot our remote processor */
64	ret = rproc_boot(my_rproc);
65	if (ret) {
66		/*
67		 * something went wrong. handle it and leave.
68		 */
69	}
70
71	/*
72	 * our remote processor is now powered on... give it some work
73	 */
74
75	/* let's shut it down now */
76	rproc_shutdown(my_rproc);
77}
78
794. API for implementors
80
81  struct rproc *rproc_alloc(struct device *dev, const char *name,
82				const struct rproc_ops *ops,
83				const char *firmware, int len)
84    - Allocate a new remote processor handle, but don't register
85      it yet. Required parameters are the underlying device, the
86      name of this remote processor, platform-specific ops handlers,
87      the name of the firmware to boot this rproc with, and the
88      length of private data needed by the allocating rproc driver (in bytes).
89
90      This function should be used by rproc implementations during
91      initialization of the remote processor.
92      After creating an rproc handle using this function, and when ready,
93      implementations should then call rproc_add() to complete
94      the registration of the remote processor.
95      On success, the new rproc is returned, and on failure, NULL.
96
97      Note: _never_ directly deallocate @rproc, even if it was not registered
98      yet. Instead, when you need to unroll rproc_alloc(), use rproc_put().
99
100  void rproc_put(struct rproc *rproc)
101    - Free an rproc handle that was allocated by rproc_alloc.
102      This function essentially unrolls rproc_alloc(), by decrementing the
103      rproc's refcount. It doesn't directly free rproc; that would happen
104      only if there are no other references to rproc and its refcount now
105      dropped to zero.
106
107  int rproc_add(struct rproc *rproc)
108    - Register @rproc with the remoteproc framework, after it has been
109      allocated with rproc_alloc().
110      This is called by the platform-specific rproc implementation, whenever
111      a new remote processor device is probed.
112      Returns 0 on success and an appropriate error code otherwise.
113      Note: this function initiates an asynchronous firmware loading
114      context, which will look for virtio devices supported by the rproc's
115      firmware.
116      If found, those virtio devices will be created and added, so as a result
117      of registering this remote processor, additional virtio drivers might get
118      probed.
119
120  int rproc_del(struct rproc *rproc)
121    - Unroll rproc_add().
122      This function should be called when the platform specific rproc
123      implementation decides to remove the rproc device. it should
124      _only_ be called if a previous invocation of rproc_add()
125      has completed successfully.
126
127      After rproc_del() returns, @rproc is still valid, and its
128      last refcount should be decremented by calling rproc_put().
129
130      Returns 0 on success and -EINVAL if @rproc isn't valid.
131
132  void rproc_report_crash(struct rproc *rproc, enum rproc_crash_type type)
133    - Report a crash in a remoteproc
134      This function must be called every time a crash is detected by the
135      platform specific rproc implementation. This should not be called from a
136      non-remoteproc driver. This function can be called from atomic/interrupt
137      context.
138
1395. Implementation callbacks
140
141These callbacks should be provided by platform-specific remoteproc
142drivers:
143
144/**
145 * struct rproc_ops - platform-specific device handlers
146 * @start:	power on the device and boot it
147 * @stop:	power off the device
148 * @kick:	kick a virtqueue (virtqueue id given as a parameter)
149 */
150struct rproc_ops {
151	int (*start)(struct rproc *rproc);
152	int (*stop)(struct rproc *rproc);
153	void (*kick)(struct rproc *rproc, int vqid);
154};
155
156Every remoteproc implementation should at least provide the ->start and ->stop
157handlers. If rpmsg/virtio functionality is also desired, then the ->kick handler
158should be provided as well.
159
160The ->start() handler takes an rproc handle and should then power on the
161device and boot it (use rproc->priv to access platform-specific private data).
162The boot address, in case needed, can be found in rproc->bootaddr (remoteproc
163core puts there the ELF entry point).
164On success, 0 should be returned, and on failure, an appropriate error code.
165
166The ->stop() handler takes an rproc handle and powers the device down.
167On success, 0 is returned, and on failure, an appropriate error code.
168
169The ->kick() handler takes an rproc handle, and an index of a virtqueue
170where new message was placed in. Implementations should interrupt the remote
171processor and let it know it has pending messages. Notifying remote processors
172the exact virtqueue index to look in is optional: it is easy (and not
173too expensive) to go through the existing virtqueues and look for new buffers
174in the used rings.
175
1766. Binary Firmware Structure
177
178At this point remoteproc only supports ELF32 firmware binaries. However,
179it is quite expected that other platforms/devices which we'd want to
180support with this framework will be based on different binary formats.
181
182When those use cases show up, we will have to decouple the binary format
183from the framework core, so we can support several binary formats without
184duplicating common code.
185
186When the firmware is parsed, its various segments are loaded to memory
187according to the specified device address (might be a physical address
188if the remote processor is accessing memory directly).
189
190In addition to the standard ELF segments, most remote processors would
191also include a special section which we call "the resource table".
192
193The resource table contains system resources that the remote processor
194requires before it should be powered on, such as allocation of physically
195contiguous memory, or iommu mapping of certain on-chip peripherals.
196Remotecore will only power up the device after all the resource table's
197requirement are met.
198
199In addition to system resources, the resource table may also contain
200resource entries that publish the existence of supported features
201or configurations by the remote processor, such as trace buffers and
202supported virtio devices (and their configurations).
203
204The resource table begins with this header:
205
206/**
207 * struct resource_table - firmware resource table header
208 * @ver: version number
209 * @num: number of resource entries
210 * @reserved: reserved (must be zero)
211 * @offset: array of offsets pointing at the various resource entries
212 *
213 * The header of the resource table, as expressed by this structure,
214 * contains a version number (should we need to change this format in the
215 * future), the number of available resource entries, and their offsets
216 * in the table.
217 */
218struct resource_table {
219	u32 ver;
220	u32 num;
221	u32 reserved[2];
222	u32 offset[0];
223} __packed;
224
225Immediately following this header are the resource entries themselves,
226each of which begins with the following resource entry header:
227
228/**
229 * struct fw_rsc_hdr - firmware resource entry header
230 * @type: resource type
231 * @data: resource data
232 *
233 * Every resource entry begins with a 'struct fw_rsc_hdr' header providing
234 * its @type. The content of the entry itself will immediately follow
235 * this header, and it should be parsed according to the resource type.
236 */
237struct fw_rsc_hdr {
238	u32 type;
239	u8 data[0];
240} __packed;
241
242Some resources entries are mere announcements, where the host is informed
243of specific remoteproc configuration. Other entries require the host to
244do something (e.g. allocate a system resource). Sometimes a negotiation
245is expected, where the firmware requests a resource, and once allocated,
246the host should provide back its details (e.g. address of an allocated
247memory region).
248
249Here are the various resource types that are currently supported:
250
251/**
252 * enum fw_resource_type - types of resource entries
253 *
254 * @RSC_CARVEOUT:   request for allocation of a physically contiguous
255 *		    memory region.
256 * @RSC_DEVMEM:     request to iommu_map a memory-based peripheral.
257 * @RSC_TRACE:	    announces the availability of a trace buffer into which
258 *		    the remote processor will be writing logs.
259 * @RSC_VDEV:       declare support for a virtio device, and serve as its
260 *		    virtio header.
261 * @RSC_LAST:       just keep this one at the end
262 *
263 * Please note that these values are used as indices to the rproc_handle_rsc
264 * lookup table, so please keep them sane. Moreover, @RSC_LAST is used to
265 * check the validity of an index before the lookup table is accessed, so
266 * please update it as needed.
267 */
268enum fw_resource_type {
269	RSC_CARVEOUT	= 0,
270	RSC_DEVMEM	= 1,
271	RSC_TRACE	= 2,
272	RSC_VDEV	= 3,
273	RSC_LAST	= 4,
274};
275
276For more details regarding a specific resource type, please see its
277dedicated structure in include/linux/remoteproc.h.
278
279We also expect that platform-specific resource entries will show up
280at some point. When that happens, we could easily add a new RSC_PLATFORM
281type, and hand those resources to the platform-specific rproc driver to handle.
282
2837. Virtio and remoteproc
284
285The firmware should provide remoteproc information about virtio devices
286that it supports, and their configurations: a RSC_VDEV resource entry
287should specify the virtio device id (as in virtio_ids.h), virtio features,
288virtio config space, vrings information, etc.
289
290When a new remote processor is registered, the remoteproc framework
291will look for its resource table and will register the virtio devices
292it supports. A firmware may support any number of virtio devices, and
293of any type (a single remote processor can also easily support several
294rpmsg virtio devices this way, if desired).
295
296Of course, RSC_VDEV resource entries are only good enough for static
297allocation of virtio devices. Dynamic allocations will also be made possible
298using the rpmsg bus (similar to how we already do dynamic allocations of
299rpmsg channels; read more about it in rpmsg.txt).
300