aboutsummaryrefslogtreecommitdiffstats
path: root/drivers/xen
diff options
context:
space:
mode:
authorJeremy Fitzhardinge <jeremy@xensource.com>2007-07-17 21:37:06 -0400
committerJeremy Fitzhardinge <jeremy@goop.org>2007-07-18 11:47:45 -0400
commit4bac07c993d03434ea902d3d4290d9e45944b66c (patch)
tree1930a1d8c23d3968d4644edda50791ff390bfe93 /drivers/xen
parentad9a86121f5a374b48ce2924f8a9d7e94a04db27 (diff)
xen: add the Xenbus sysfs and virtual device hotplug driver
This communicates with the machine control software via a registry residing in a controlling virtual machine. This allows dynamic creation, destruction and modification of virtual device configurations (network devices, block devices and CPUS, to name some examples). [ Greg, would you mind giving this a review? Thanks -J ] Signed-off-by: Ian Pratt <ian.pratt@xensource.com> Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk> Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com> Signed-off-by: Chris Wright <chrisw@sous-sol.org> Cc: Greg KH <greg@kroah.com>
Diffstat (limited to 'drivers/xen')
-rw-r--r--drivers/xen/Makefile1
-rw-r--r--drivers/xen/xenbus/Makefile7
-rw-r--r--drivers/xen/xenbus/xenbus_client.c569
-rw-r--r--drivers/xen/xenbus/xenbus_comms.c233
-rw-r--r--drivers/xen/xenbus/xenbus_comms.h46
-rw-r--r--drivers/xen/xenbus/xenbus_probe.c935
-rw-r--r--drivers/xen/xenbus/xenbus_probe.h74
-rw-r--r--drivers/xen/xenbus/xenbus_xs.c861
8 files changed, 2726 insertions, 0 deletions
diff --git a/drivers/xen/Makefile b/drivers/xen/Makefile
index eb42b521eef9..56592f0d6cef 100644
--- a/drivers/xen/Makefile
+++ b/drivers/xen/Makefile
@@ -1 +1,2 @@
1obj-y += grant-table.o 1obj-y += grant-table.o
2obj-y += xenbus/
diff --git a/drivers/xen/xenbus/Makefile b/drivers/xen/xenbus/Makefile
new file mode 100644
index 000000000000..5571f5b84223
--- /dev/null
+++ b/drivers/xen/xenbus/Makefile
@@ -0,0 +1,7 @@
1obj-y += xenbus.o
2
3xenbus-objs =
4xenbus-objs += xenbus_client.o
5xenbus-objs += xenbus_comms.o
6xenbus-objs += xenbus_xs.o
7xenbus-objs += xenbus_probe.o
diff --git a/drivers/xen/xenbus/xenbus_client.c b/drivers/xen/xenbus/xenbus_client.c
new file mode 100644
index 000000000000..9fd2f70ab46d
--- /dev/null
+++ b/drivers/xen/xenbus/xenbus_client.c
@@ -0,0 +1,569 @@
1/******************************************************************************
2 * Client-facing interface for the Xenbus driver. In other words, the
3 * interface between the Xenbus and the device-specific code, be it the
4 * frontend or the backend of that driver.
5 *
6 * Copyright (C) 2005 XenSource Ltd
7 *
8 * This program is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU General Public License version 2
10 * as published by the Free Software Foundation; or, when distributed
11 * separately from the Linux kernel or incorporated into other
12 * software packages, subject to the following license:
13 *
14 * Permission is hereby granted, free of charge, to any person obtaining a copy
15 * of this source file (the "Software"), to deal in the Software without
16 * restriction, including without limitation the rights to use, copy, modify,
17 * merge, publish, distribute, sublicense, and/or sell copies of the Software,
18 * and to permit persons to whom the Software is furnished to do so, subject to
19 * the following conditions:
20 *
21 * The above copyright notice and this permission notice shall be included in
22 * all copies or substantial portions of the Software.
23 *
24 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
25 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
26 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
27 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
28 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
29 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
30 * IN THE SOFTWARE.
31 */
32
33#include <linux/types.h>
34#include <linux/vmalloc.h>
35#include <asm/xen/hypervisor.h>
36#include <xen/interface/xen.h>
37#include <xen/interface/event_channel.h>
38#include <xen/events.h>
39#include <xen/grant_table.h>
40#include <xen/xenbus.h>
41
42const char *xenbus_strstate(enum xenbus_state state)
43{
44 static const char *const name[] = {
45 [ XenbusStateUnknown ] = "Unknown",
46 [ XenbusStateInitialising ] = "Initialising",
47 [ XenbusStateInitWait ] = "InitWait",
48 [ XenbusStateInitialised ] = "Initialised",
49 [ XenbusStateConnected ] = "Connected",
50 [ XenbusStateClosing ] = "Closing",
51 [ XenbusStateClosed ] = "Closed",
52 };
53 return (state < ARRAY_SIZE(name)) ? name[state] : "INVALID";
54}
55EXPORT_SYMBOL_GPL(xenbus_strstate);
56
57/**
58 * xenbus_watch_path - register a watch
59 * @dev: xenbus device
60 * @path: path to watch
61 * @watch: watch to register
62 * @callback: callback to register
63 *
64 * Register a @watch on the given path, using the given xenbus_watch structure
65 * for storage, and the given @callback function as the callback. Return 0 on
66 * success, or -errno on error. On success, the given @path will be saved as
67 * @watch->node, and remains the caller's to free. On error, @watch->node will
68 * be NULL, the device will switch to %XenbusStateClosing, and the error will
69 * be saved in the store.
70 */
71int xenbus_watch_path(struct xenbus_device *dev, const char *path,
72 struct xenbus_watch *watch,
73 void (*callback)(struct xenbus_watch *,
74 const char **, unsigned int))
75{
76 int err;
77
78 watch->node = path;
79 watch->callback = callback;
80
81 err = register_xenbus_watch(watch);
82
83 if (err) {
84 watch->node = NULL;
85 watch->callback = NULL;
86 xenbus_dev_fatal(dev, err, "adding watch on %s", path);
87 }
88
89 return err;
90}
91EXPORT_SYMBOL_GPL(xenbus_watch_path);
92
93
94/**
95 * xenbus_watch_pathfmt - register a watch on a sprintf-formatted path
96 * @dev: xenbus device
97 * @watch: watch to register
98 * @callback: callback to register
99 * @pathfmt: format of path to watch
100 *
101 * Register a watch on the given @path, using the given xenbus_watch
102 * structure for storage, and the given @callback function as the callback.
103 * Return 0 on success, or -errno on error. On success, the watched path
104 * (@path/@path2) will be saved as @watch->node, and becomes the caller's to
105 * kfree(). On error, watch->node will be NULL, so the caller has nothing to
106 * free, the device will switch to %XenbusStateClosing, and the error will be
107 * saved in the store.
108 */
109int xenbus_watch_pathfmt(struct xenbus_device *dev,
110 struct xenbus_watch *watch,
111 void (*callback)(struct xenbus_watch *,
112 const char **, unsigned int),
113 const char *pathfmt, ...)
114{
115 int err;
116 va_list ap;
117 char *path;
118
119 va_start(ap, pathfmt);
120 path = kvasprintf(GFP_KERNEL, pathfmt, ap);
121 va_end(ap);
122
123 if (!path) {
124 xenbus_dev_fatal(dev, -ENOMEM, "allocating path for watch");
125 return -ENOMEM;
126 }
127 err = xenbus_watch_path(dev, path, watch, callback);
128
129 if (err)
130 kfree(path);
131 return err;
132}
133EXPORT_SYMBOL_GPL(xenbus_watch_pathfmt);
134
135
136/**
137 * xenbus_switch_state
138 * @dev: xenbus device
139 * @xbt: transaction handle
140 * @state: new state
141 *
142 * Advertise in the store a change of the given driver to the given new_state.
143 * Return 0 on success, or -errno on error. On error, the device will switch
144 * to XenbusStateClosing, and the error will be saved in the store.
145 */
146int xenbus_switch_state(struct xenbus_device *dev, enum xenbus_state state)
147{
148 /* We check whether the state is currently set to the given value, and
149 if not, then the state is set. We don't want to unconditionally
150 write the given state, because we don't want to fire watches
151 unnecessarily. Furthermore, if the node has gone, we don't write
152 to it, as the device will be tearing down, and we don't want to
153 resurrect that directory.
154
155 Note that, because of this cached value of our state, this function
156 will not work inside a Xenstore transaction (something it was
157 trying to in the past) because dev->state would not get reset if
158 the transaction was aborted.
159
160 */
161
162 int current_state;
163 int err;
164
165 if (state == dev->state)
166 return 0;
167
168 err = xenbus_scanf(XBT_NIL, dev->nodename, "state", "%d",
169 &current_state);
170 if (err != 1)
171 return 0;
172
173 err = xenbus_printf(XBT_NIL, dev->nodename, "state", "%d", state);
174 if (err) {
175 if (state != XenbusStateClosing) /* Avoid looping */
176 xenbus_dev_fatal(dev, err, "writing new state");
177 return err;
178 }
179
180 dev->state = state;
181
182 return 0;
183}
184EXPORT_SYMBOL_GPL(xenbus_switch_state);
185
186int xenbus_frontend_closed(struct xenbus_device *dev)
187{
188 xenbus_switch_state(dev, XenbusStateClosed);
189 complete(&dev->down);
190 return 0;
191}
192EXPORT_SYMBOL_GPL(xenbus_frontend_closed);
193
194/**
195 * Return the path to the error node for the given device, or NULL on failure.
196 * If the value returned is non-NULL, then it is the caller's to kfree.
197 */
198static char *error_path(struct xenbus_device *dev)
199{
200 return kasprintf(GFP_KERNEL, "error/%s", dev->nodename);
201}
202
203
204static void xenbus_va_dev_error(struct xenbus_device *dev, int err,
205 const char *fmt, va_list ap)
206{
207 int ret;
208 unsigned int len;
209 char *printf_buffer = NULL;
210 char *path_buffer = NULL;
211
212#define PRINTF_BUFFER_SIZE 4096
213 printf_buffer = kmalloc(PRINTF_BUFFER_SIZE, GFP_KERNEL);
214 if (printf_buffer == NULL)
215 goto fail;
216
217 len = sprintf(printf_buffer, "%i ", -err);
218 ret = vsnprintf(printf_buffer+len, PRINTF_BUFFER_SIZE-len, fmt, ap);
219
220 BUG_ON(len + ret > PRINTF_BUFFER_SIZE-1);
221
222 dev_err(&dev->dev, "%s\n", printf_buffer);
223
224 path_buffer = error_path(dev);
225
226 if (path_buffer == NULL) {
227 dev_err(&dev->dev, "failed to write error node for %s (%s)\n",
228 dev->nodename, printf_buffer);
229 goto fail;
230 }
231
232 if (xenbus_write(XBT_NIL, path_buffer, "error", printf_buffer) != 0) {
233 dev_err(&dev->dev, "failed to write error node for %s (%s)\n",
234 dev->nodename, printf_buffer);
235 goto fail;
236 }
237
238fail:
239 kfree(printf_buffer);
240 kfree(path_buffer);
241}
242
243
244/**
245 * xenbus_dev_error
246 * @dev: xenbus device
247 * @err: error to report
248 * @fmt: error message format
249 *
250 * Report the given negative errno into the store, along with the given
251 * formatted message.
252 */
253void xenbus_dev_error(struct xenbus_device *dev, int err, const char *fmt, ...)
254{
255 va_list ap;
256
257 va_start(ap, fmt);
258 xenbus_va_dev_error(dev, err, fmt, ap);
259 va_end(ap);
260}
261EXPORT_SYMBOL_GPL(xenbus_dev_error);
262
263/**
264 * xenbus_dev_fatal
265 * @dev: xenbus device
266 * @err: error to report
267 * @fmt: error message format
268 *
269 * Equivalent to xenbus_dev_error(dev, err, fmt, args), followed by
270 * xenbus_switch_state(dev, NULL, XenbusStateClosing) to schedule an orderly
271 * closedown of this driver and its peer.
272 */
273
274void xenbus_dev_fatal(struct xenbus_device *dev, int err, const char *fmt, ...)
275{
276 va_list ap;
277
278 va_start(ap, fmt);
279 xenbus_va_dev_error(dev, err, fmt, ap);
280 va_end(ap);
281
282 xenbus_switch_state(dev, XenbusStateClosing);
283}
284EXPORT_SYMBOL_GPL(xenbus_dev_fatal);
285
286/**
287 * xenbus_grant_ring
288 * @dev: xenbus device
289 * @ring_mfn: mfn of ring to grant
290
291 * Grant access to the given @ring_mfn to the peer of the given device. Return
292 * 0 on success, or -errno on error. On error, the device will switch to
293 * XenbusStateClosing, and the error will be saved in the store.
294 */
295int xenbus_grant_ring(struct xenbus_device *dev, unsigned long ring_mfn)
296{
297 int err = gnttab_grant_foreign_access(dev->otherend_id, ring_mfn, 0);
298 if (err < 0)
299 xenbus_dev_fatal(dev, err, "granting access to ring page");
300 return err;
301}
302EXPORT_SYMBOL_GPL(xenbus_grant_ring);
303
304
305/**
306 * Allocate an event channel for the given xenbus_device, assigning the newly
307 * created local port to *port. Return 0 on success, or -errno on error. On
308 * error, the device will switch to XenbusStateClosing, and the error will be
309 * saved in the store.
310 */
311int xenbus_alloc_evtchn(struct xenbus_device *dev, int *port)
312{
313 struct evtchn_alloc_unbound alloc_unbound;
314 int err;
315
316 alloc_unbound.dom = DOMID_SELF;
317 alloc_unbound.remote_dom = dev->otherend_id;
318
319 err = HYPERVISOR_event_channel_op(EVTCHNOP_alloc_unbound,
320 &alloc_unbound);
321 if (err)
322 xenbus_dev_fatal(dev, err, "allocating event channel");
323 else
324 *port = alloc_unbound.port;
325
326 return err;
327}
328EXPORT_SYMBOL_GPL(xenbus_alloc_evtchn);
329
330
331/**
332 * Bind to an existing interdomain event channel in another domain. Returns 0
333 * on success and stores the local port in *port. On error, returns -errno,
334 * switches the device to XenbusStateClosing, and saves the error in XenStore.
335 */
336int xenbus_bind_evtchn(struct xenbus_device *dev, int remote_port, int *port)
337{
338 struct evtchn_bind_interdomain bind_interdomain;
339 int err;
340
341 bind_interdomain.remote_dom = dev->otherend_id;
342 bind_interdomain.remote_port = remote_port;
343
344 err = HYPERVISOR_event_channel_op(EVTCHNOP_bind_interdomain,
345 &bind_interdomain);
346 if (err)
347 xenbus_dev_fatal(dev, err,
348 "binding to event channel %d from domain %d",
349 remote_port, dev->otherend_id);
350 else
351 *port = bind_interdomain.local_port;
352
353 return err;
354}
355EXPORT_SYMBOL_GPL(xenbus_bind_evtchn);
356
357
358/**
359 * Free an existing event channel. Returns 0 on success or -errno on error.
360 */
361int xenbus_free_evtchn(struct xenbus_device *dev, int port)
362{
363 struct evtchn_close close;
364 int err;
365
366 close.port = port;
367
368 err = HYPERVISOR_event_channel_op(EVTCHNOP_close, &close);
369 if (err)
370 xenbus_dev_error(dev, err, "freeing event channel %d", port);
371
372 return err;
373}
374EXPORT_SYMBOL_GPL(xenbus_free_evtchn);
375
376
377/**
378 * xenbus_map_ring_valloc
379 * @dev: xenbus device
380 * @gnt_ref: grant reference
381 * @vaddr: pointer to address to be filled out by mapping
382 *
383 * Based on Rusty Russell's skeleton driver's map_page.
384 * Map a page of memory into this domain from another domain's grant table.
385 * xenbus_map_ring_valloc allocates a page of virtual address space, maps the
386 * page to that address, and sets *vaddr to that address.
387 * Returns 0 on success, and GNTST_* (see xen/include/interface/grant_table.h)
388 * or -ENOMEM on error. If an error is returned, device will switch to
389 * XenbusStateClosing and the error message will be saved in XenStore.
390 */
391int xenbus_map_ring_valloc(struct xenbus_device *dev, int gnt_ref, void **vaddr)
392{
393 struct gnttab_map_grant_ref op = {
394 .flags = GNTMAP_host_map,
395 .ref = gnt_ref,
396 .dom = dev->otherend_id,
397 };
398 struct vm_struct *area;
399
400 *vaddr = NULL;
401
402 area = alloc_vm_area(PAGE_SIZE);
403 if (!area)
404 return -ENOMEM;
405
406 op.host_addr = (unsigned long)area->addr;
407
408 if (HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, &op, 1))
409 BUG();
410
411 if (op.status != GNTST_okay) {
412 free_vm_area(area);
413 xenbus_dev_fatal(dev, op.status,
414 "mapping in shared page %d from domain %d",
415 gnt_ref, dev->otherend_id);
416 return op.status;
417 }
418
419 /* Stuff the handle in an unused field */
420 area->phys_addr = (unsigned long)op.handle;
421
422 *vaddr = area->addr;
423 return 0;
424}
425EXPORT_SYMBOL_GPL(xenbus_map_ring_valloc);
426
427
428/**
429 * xenbus_map_ring
430 * @dev: xenbus device
431 * @gnt_ref: grant reference
432 * @handle: pointer to grant handle to be filled
433 * @vaddr: address to be mapped to
434 *
435 * Map a page of memory into this domain from another domain's grant table.
436 * xenbus_map_ring does not allocate the virtual address space (you must do
437 * this yourself!). It only maps in the page to the specified address.
438 * Returns 0 on success, and GNTST_* (see xen/include/interface/grant_table.h)
439 * or -ENOMEM on error. If an error is returned, device will switch to
440 * XenbusStateClosing and the error message will be saved in XenStore.
441 */
442int xenbus_map_ring(struct xenbus_device *dev, int gnt_ref,
443 grant_handle_t *handle, void *vaddr)
444{
445 struct gnttab_map_grant_ref op = {
446 .host_addr = (unsigned long)vaddr,
447 .flags = GNTMAP_host_map,
448 .ref = gnt_ref,
449 .dom = dev->otherend_id,
450 };
451
452 if (HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, &op, 1))
453 BUG();
454
455 if (op.status != GNTST_okay) {
456 xenbus_dev_fatal(dev, op.status,
457 "mapping in shared page %d from domain %d",
458 gnt_ref, dev->otherend_id);
459 } else
460 *handle = op.handle;
461
462 return op.status;
463}
464EXPORT_SYMBOL_GPL(xenbus_map_ring);
465
466
467/**
468 * xenbus_unmap_ring_vfree
469 * @dev: xenbus device
470 * @vaddr: addr to unmap
471 *
472 * Based on Rusty Russell's skeleton driver's unmap_page.
473 * Unmap a page of memory in this domain that was imported from another domain.
474 * Use xenbus_unmap_ring_vfree if you mapped in your memory with
475 * xenbus_map_ring_valloc (it will free the virtual address space).
476 * Returns 0 on success and returns GNTST_* on error
477 * (see xen/include/interface/grant_table.h).
478 */
479int xenbus_unmap_ring_vfree(struct xenbus_device *dev, void *vaddr)
480{
481 struct vm_struct *area;
482 struct gnttab_unmap_grant_ref op = {
483 .host_addr = (unsigned long)vaddr,
484 };
485
486 /* It'd be nice if linux/vmalloc.h provided a find_vm_area(void *addr)
487 * method so that we don't have to muck with vmalloc internals here.
488 * We could force the user to hang on to their struct vm_struct from
489 * xenbus_map_ring_valloc, but these 6 lines considerably simplify
490 * this API.
491 */
492 read_lock(&vmlist_lock);
493 for (area = vmlist; area != NULL; area = area->next) {
494 if (area->addr == vaddr)
495 break;
496 }
497 read_unlock(&vmlist_lock);
498
499 if (!area) {
500 xenbus_dev_error(dev, -ENOENT,
501 "can't find mapped virtual address %p", vaddr);
502 return GNTST_bad_virt_addr;
503 }
504
505 op.handle = (grant_handle_t)area->phys_addr;
506
507 if (HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref, &op, 1))
508 BUG();
509
510 if (op.status == GNTST_okay)
511 free_vm_area(area);
512 else
513 xenbus_dev_error(dev, op.status,
514 "unmapping page at handle %d error %d",
515 (int16_t)area->phys_addr, op.status);
516
517 return op.status;
518}
519EXPORT_SYMBOL_GPL(xenbus_unmap_ring_vfree);
520
521
522/**
523 * xenbus_unmap_ring
524 * @dev: xenbus device
525 * @handle: grant handle
526 * @vaddr: addr to unmap
527 *
528 * Unmap a page of memory in this domain that was imported from another domain.
529 * Returns 0 on success and returns GNTST_* on error
530 * (see xen/include/interface/grant_table.h).
531 */
532int xenbus_unmap_ring(struct xenbus_device *dev,
533 grant_handle_t handle, void *vaddr)
534{
535 struct gnttab_unmap_grant_ref op = {
536 .host_addr = (unsigned long)vaddr,
537 .handle = handle,
538 };
539
540 if (HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref, &op, 1))
541 BUG();
542
543 if (op.status != GNTST_okay)
544 xenbus_dev_error(dev, op.status,
545 "unmapping page at handle %d error %d",
546 handle, op.status);
547
548 return op.status;
549}
550EXPORT_SYMBOL_GPL(xenbus_unmap_ring);
551
552
553/**
554 * xenbus_read_driver_state
555 * @path: path for driver
556 *
557 * Return the state of the driver rooted at the given store path, or
558 * XenbusStateUnknown if no state can be read.
559 */
560enum xenbus_state xenbus_read_driver_state(const char *path)
561{
562 enum xenbus_state result;
563 int err = xenbus_gather(XBT_NIL, path, "state", "%d", &result, NULL);
564 if (err)
565 result = XenbusStateUnknown;
566
567 return result;
568}
569EXPORT_SYMBOL_GPL(xenbus_read_driver_state);
diff --git a/drivers/xen/xenbus/xenbus_comms.c b/drivers/xen/xenbus/xenbus_comms.c
new file mode 100644
index 000000000000..6efbe3f29ca5
--- /dev/null
+++ b/drivers/xen/xenbus/xenbus_comms.c
@@ -0,0 +1,233 @@
1/******************************************************************************
2 * xenbus_comms.c
3 *
4 * Low level code to talks to Xen Store: ringbuffer and event channel.
5 *
6 * Copyright (C) 2005 Rusty Russell, IBM Corporation
7 *
8 * This program is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU General Public License version 2
10 * as published by the Free Software Foundation; or, when distributed
11 * separately from the Linux kernel or incorporated into other
12 * software packages, subject to the following license:
13 *
14 * Permission is hereby granted, free of charge, to any person obtaining a copy
15 * of this source file (the "Software"), to deal in the Software without
16 * restriction, including without limitation the rights to use, copy, modify,
17 * merge, publish, distribute, sublicense, and/or sell copies of the Software,
18 * and to permit persons to whom the Software is furnished to do so, subject to
19 * the following conditions:
20 *
21 * The above copyright notice and this permission notice shall be included in
22 * all copies or substantial portions of the Software.
23 *
24 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
25 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
26 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
27 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
28 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
29 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
30 * IN THE SOFTWARE.
31 */
32
33#include <linux/wait.h>
34#include <linux/interrupt.h>
35#include <linux/sched.h>
36#include <linux/err.h>
37#include <xen/xenbus.h>
38#include <asm/xen/hypervisor.h>
39#include <xen/events.h>
40#include <xen/page.h>
41#include "xenbus_comms.h"
42
43static int xenbus_irq;
44
45static DECLARE_WORK(probe_work, xenbus_probe);
46
47static DECLARE_WAIT_QUEUE_HEAD(xb_waitq);
48
49static irqreturn_t wake_waiting(int irq, void *unused)
50{
51 if (unlikely(xenstored_ready == 0)) {
52 xenstored_ready = 1;
53 schedule_work(&probe_work);
54 }
55
56 wake_up(&xb_waitq);
57 return IRQ_HANDLED;
58}
59
60static int check_indexes(XENSTORE_RING_IDX cons, XENSTORE_RING_IDX prod)
61{
62 return ((prod - cons) <= XENSTORE_RING_SIZE);
63}
64
65static void *get_output_chunk(XENSTORE_RING_IDX cons,
66 XENSTORE_RING_IDX prod,
67 char *buf, uint32_t *len)
68{
69 *len = XENSTORE_RING_SIZE - MASK_XENSTORE_IDX(prod);
70 if ((XENSTORE_RING_SIZE - (prod - cons)) < *len)
71 *len = XENSTORE_RING_SIZE - (prod - cons);
72 return buf + MASK_XENSTORE_IDX(prod);
73}
74
75static const void *get_input_chunk(XENSTORE_RING_IDX cons,
76 XENSTORE_RING_IDX prod,
77 const char *buf, uint32_t *len)
78{
79 *len = XENSTORE_RING_SIZE - MASK_XENSTORE_IDX(cons);
80 if ((prod - cons) < *len)
81 *len = prod - cons;
82 return buf + MASK_XENSTORE_IDX(cons);
83}
84
85/**
86 * xb_write - low level write
87 * @data: buffer to send
88 * @len: length of buffer
89 *
90 * Returns 0 on success, error otherwise.
91 */
92int xb_write(const void *data, unsigned len)
93{
94 struct xenstore_domain_interface *intf = xen_store_interface;
95 XENSTORE_RING_IDX cons, prod;
96 int rc;
97
98 while (len != 0) {
99 void *dst;
100 unsigned int avail;
101
102 rc = wait_event_interruptible(
103 xb_waitq,
104 (intf->req_prod - intf->req_cons) !=
105 XENSTORE_RING_SIZE);
106 if (rc < 0)
107 return rc;
108
109 /* Read indexes, then verify. */
110 cons = intf->req_cons;
111 prod = intf->req_prod;
112 if (!check_indexes(cons, prod)) {
113 intf->req_cons = intf->req_prod = 0;
114 return -EIO;
115 }
116
117 dst = get_output_chunk(cons, prod, intf->req, &avail);
118 if (avail == 0)
119 continue;
120 if (avail > len)
121 avail = len;
122
123 /* Must write data /after/ reading the consumer index. */
124 mb();
125
126 memcpy(dst, data, avail);
127 data += avail;
128 len -= avail;
129
130 /* Other side must not see new producer until data is there. */
131 wmb();
132 intf->req_prod += avail;
133
134 /* Implies mb(): other side will see the updated producer. */
135 notify_remote_via_evtchn(xen_store_evtchn);
136 }
137
138 return 0;
139}
140
141int xb_data_to_read(void)
142{
143 struct xenstore_domain_interface *intf = xen_store_interface;
144 return (intf->rsp_cons != intf->rsp_prod);
145}
146
147int xb_wait_for_data_to_read(void)
148{
149 return wait_event_interruptible(xb_waitq, xb_data_to_read());
150}
151
152int xb_read(void *data, unsigned len)
153{
154 struct xenstore_domain_interface *intf = xen_store_interface;
155 XENSTORE_RING_IDX cons, prod;
156 int rc;
157
158 while (len != 0) {
159 unsigned int avail;
160 const char *src;
161
162 rc = xb_wait_for_data_to_read();
163 if (rc < 0)
164 return rc;
165
166 /* Read indexes, then verify. */
167 cons = intf->rsp_cons;
168 prod = intf->rsp_prod;
169 if (!check_indexes(cons, prod)) {
170 intf->rsp_cons = intf->rsp_prod = 0;
171 return -EIO;
172 }
173
174 src = get_input_chunk(cons, prod, intf->rsp, &avail);
175 if (avail == 0)
176 continue;
177 if (avail > len)
178 avail = len;
179
180 /* Must read data /after/ reading the producer index. */
181 rmb();
182
183 memcpy(data, src, avail);
184 data += avail;
185 len -= avail;
186
187 /* Other side must not see free space until we've copied out */
188 mb();
189 intf->rsp_cons += avail;
190
191 pr_debug("Finished read of %i bytes (%i to go)\n", avail, len);
192
193 /* Implies mb(): other side will see the updated consumer. */
194 notify_remote_via_evtchn(xen_store_evtchn);
195 }
196
197 return 0;
198}
199
200/**
201 * xb_init_comms - Set up interrupt handler off store event channel.
202 */
203int xb_init_comms(void)
204{
205 struct xenstore_domain_interface *intf = xen_store_interface;
206 int err;
207
208 if (intf->req_prod != intf->req_cons)
209 printk(KERN_ERR "XENBUS request ring is not quiescent "
210 "(%08x:%08x)!\n", intf->req_cons, intf->req_prod);
211
212 if (intf->rsp_prod != intf->rsp_cons) {
213 printk(KERN_WARNING "XENBUS response ring is not quiescent "
214 "(%08x:%08x): fixing up\n",
215 intf->rsp_cons, intf->rsp_prod);
216 intf->rsp_cons = intf->rsp_prod;
217 }
218
219 if (xenbus_irq)
220 unbind_from_irqhandler(xenbus_irq, &xb_waitq);
221
222 err = bind_evtchn_to_irqhandler(
223 xen_store_evtchn, wake_waiting,
224 0, "xenbus", &xb_waitq);
225 if (err <= 0) {
226 printk(KERN_ERR "XENBUS request irq failed %i\n", err);
227 return err;
228 }
229
230 xenbus_irq = err;
231
232 return 0;
233}
diff --git a/drivers/xen/xenbus/xenbus_comms.h b/drivers/xen/xenbus/xenbus_comms.h
new file mode 100644
index 000000000000..c21db7513736
--- /dev/null
+++ b/drivers/xen/xenbus/xenbus_comms.h
@@ -0,0 +1,46 @@
1/*
2 * Private include for xenbus communications.
3 *
4 * Copyright (C) 2005 Rusty Russell, IBM Corporation
5 *
6 * This program is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU General Public License version 2
8 * as published by the Free Software Foundation; or, when distributed
9 * separately from the Linux kernel or incorporated into other
10 * software packages, subject to the following license:
11 *
12 * Permission is hereby granted, free of charge, to any person obtaining a copy
13 * of this source file (the "Software"), to deal in the Software without
14 * restriction, including without limitation the rights to use, copy, modify,
15 * merge, publish, distribute, sublicense, and/or sell copies of the Software,
16 * and to permit persons to whom the Software is furnished to do so, subject to
17 * the following conditions:
18 *
19 * The above copyright notice and this permission notice shall be included in
20 * all copies or substantial portions of the Software.
21 *
22 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
23 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
24 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
25 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
26 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
27 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
28 * IN THE SOFTWARE.
29 */
30
31#ifndef _XENBUS_COMMS_H
32#define _XENBUS_COMMS_H
33
34int xs_init(void);
35int xb_init_comms(void);
36
37/* Low level routines. */
38int xb_write(const void *data, unsigned len);
39int xb_read(void *data, unsigned len);
40int xb_data_to_read(void);
41int xb_wait_for_data_to_read(void);
42int xs_input_avail(void);
43extern struct xenstore_domain_interface *xen_store_interface;
44extern int xen_store_evtchn;
45
46#endif /* _XENBUS_COMMS_H */
diff --git a/drivers/xen/xenbus/xenbus_probe.c b/drivers/xen/xenbus/xenbus_probe.c
new file mode 100644
index 000000000000..0b769f7c4a48
--- /dev/null
+++ b/drivers/xen/xenbus/xenbus_probe.c
@@ -0,0 +1,935 @@
1/******************************************************************************
2 * Talks to Xen Store to figure out what devices we have.
3 *
4 * Copyright (C) 2005 Rusty Russell, IBM Corporation
5 * Copyright (C) 2005 Mike Wray, Hewlett-Packard
6 * Copyright (C) 2005, 2006 XenSource Ltd
7 *
8 * This program is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU General Public License version 2
10 * as published by the Free Software Foundation; or, when distributed
11 * separately from the Linux kernel or incorporated into other
12 * software packages, subject to the following license:
13 *
14 * Permission is hereby granted, free of charge, to any person obtaining a copy
15 * of this source file (the "Software"), to deal in the Software without
16 * restriction, including without limitation the rights to use, copy, modify,
17 * merge, publish, distribute, sublicense, and/or sell copies of the Software,
18 * and to permit persons to whom the Software is furnished to do so, subject to
19 * the following conditions:
20 *
21 * The above copyright notice and this permission notice shall be included in
22 * all copies or substantial portions of the Software.
23 *
24 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
25 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
26 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
27 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
28 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
29 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
30 * IN THE SOFTWARE.
31 */
32
33#define DPRINTK(fmt, args...) \
34 pr_debug("xenbus_probe (%s:%d) " fmt ".\n", \
35 __func__, __LINE__, ##args)
36
37#include <linux/kernel.h>
38#include <linux/err.h>
39#include <linux/string.h>
40#include <linux/ctype.h>
41#include <linux/fcntl.h>
42#include <linux/mm.h>
43#include <linux/notifier.h>
44#include <linux/kthread.h>
45#include <linux/mutex.h>
46#include <linux/io.h>
47
48#include <asm/page.h>
49#include <asm/pgtable.h>
50#include <asm/xen/hypervisor.h>
51#include <xen/xenbus.h>
52#include <xen/events.h>
53#include <xen/page.h>
54
55#include "xenbus_comms.h"
56#include "xenbus_probe.h"
57
58int xen_store_evtchn;
59struct xenstore_domain_interface *xen_store_interface;
60static unsigned long xen_store_mfn;
61
62static BLOCKING_NOTIFIER_HEAD(xenstore_chain);
63
64static void wait_for_devices(struct xenbus_driver *xendrv);
65
66static int xenbus_probe_frontend(const char *type, const char *name);
67
68static void xenbus_dev_shutdown(struct device *_dev);
69
70/* If something in array of ids matches this device, return it. */
71static const struct xenbus_device_id *
72match_device(const struct xenbus_device_id *arr, struct xenbus_device *dev)
73{
74 for (; *arr->devicetype != '\0'; arr++) {
75 if (!strcmp(arr->devicetype, dev->devicetype))
76 return arr;
77 }
78 return NULL;
79}
80
81int xenbus_match(struct device *_dev, struct device_driver *_drv)
82{
83 struct xenbus_driver *drv = to_xenbus_driver(_drv);
84
85 if (!drv->ids)
86 return 0;
87
88 return match_device(drv->ids, to_xenbus_device(_dev)) != NULL;
89}
90
91/* device/<type>/<id> => <type>-<id> */
92static int frontend_bus_id(char bus_id[BUS_ID_SIZE], const char *nodename)
93{
94 nodename = strchr(nodename, '/');
95 if (!nodename || strlen(nodename + 1) >= BUS_ID_SIZE) {
96 printk(KERN_WARNING "XENBUS: bad frontend %s\n", nodename);
97 return -EINVAL;
98 }
99
100 strlcpy(bus_id, nodename + 1, BUS_ID_SIZE);
101 if (!strchr(bus_id, '/')) {
102 printk(KERN_WARNING "XENBUS: bus_id %s no slash\n", bus_id);
103 return -EINVAL;
104 }
105 *strchr(bus_id, '/') = '-';
106 return 0;
107}
108
109
110static void free_otherend_details(struct xenbus_device *dev)
111{
112 kfree(dev->otherend);
113 dev->otherend = NULL;
114}
115
116
117static void free_otherend_watch(struct xenbus_device *dev)
118{
119 if (dev->otherend_watch.node) {
120 unregister_xenbus_watch(&dev->otherend_watch);
121 kfree(dev->otherend_watch.node);
122 dev->otherend_watch.node = NULL;
123 }
124}
125
126
127int read_otherend_details(struct xenbus_device *xendev,
128 char *id_node, char *path_node)
129{
130 int err = xenbus_gather(XBT_NIL, xendev->nodename,
131 id_node, "%i", &xendev->otherend_id,
132 path_node, NULL, &xendev->otherend,
133 NULL);
134 if (err) {
135 xenbus_dev_fatal(xendev, err,
136 "reading other end details from %s",
137 xendev->nodename);
138 return err;
139 }
140 if (strlen(xendev->otherend) == 0 ||
141 !xenbus_exists(XBT_NIL, xendev->otherend, "")) {
142 xenbus_dev_fatal(xendev, -ENOENT,
143 "unable to read other end from %s. "
144 "missing or inaccessible.",
145 xendev->nodename);
146 free_otherend_details(xendev);
147 return -ENOENT;
148 }
149
150 return 0;
151}
152
153
154static int read_backend_details(struct xenbus_device *xendev)
155{
156 return read_otherend_details(xendev, "backend-id", "backend");
157}
158
159
160/* Bus type for frontend drivers. */
161static struct xen_bus_type xenbus_frontend = {
162 .root = "device",
163 .levels = 2, /* device/type/<id> */
164 .get_bus_id = frontend_bus_id,
165 .probe = xenbus_probe_frontend,
166 .bus = {
167 .name = "xen",
168 .match = xenbus_match,
169 .probe = xenbus_dev_probe,
170 .remove = xenbus_dev_remove,
171 .shutdown = xenbus_dev_shutdown,
172 },
173};
174
175static void otherend_changed(struct xenbus_watch *watch,
176 const char **vec, unsigned int len)
177{
178 struct xenbus_device *dev =
179 container_of(watch, struct xenbus_device, otherend_watch);
180 struct xenbus_driver *drv = to_xenbus_driver(dev->dev.driver);
181 enum xenbus_state state;
182
183 /* Protect us against watches firing on old details when the otherend
184 details change, say immediately after a resume. */
185 if (!dev->otherend ||
186 strncmp(dev->otherend, vec[XS_WATCH_PATH],
187 strlen(dev->otherend))) {
188 dev_dbg(&dev->dev, "Ignoring watch at %s", vec[XS_WATCH_PATH]);
189 return;
190 }
191
192 state = xenbus_read_driver_state(dev->otherend);
193
194 dev_dbg(&dev->dev, "state is %d, (%s), %s, %s",
195 state, xenbus_strstate(state), dev->otherend_watch.node,
196 vec[XS_WATCH_PATH]);
197
198 /*
199 * Ignore xenbus transitions during shutdown. This prevents us doing
200 * work that can fail e.g., when the rootfs is gone.
201 */
202 if (system_state > SYSTEM_RUNNING) {
203 struct xen_bus_type *bus = bus;
204 bus = container_of(dev->dev.bus, struct xen_bus_type, bus);
205 /* If we're frontend, drive the state machine to Closed. */
206 /* This should cause the backend to release our resources. */
207 if ((bus == &xenbus_frontend) && (state == XenbusStateClosing))
208 xenbus_frontend_closed(dev);
209 return;
210 }
211
212 if (drv->otherend_changed)
213 drv->otherend_changed(dev, state);
214}
215
216
217static int talk_to_otherend(struct xenbus_device *dev)
218{
219 struct xenbus_driver *drv = to_xenbus_driver(dev->dev.driver);
220
221 free_otherend_watch(dev);
222 free_otherend_details(dev);
223
224 return drv->read_otherend_details(dev);
225}
226
227
228static int watch_otherend(struct xenbus_device *dev)
229{
230 return xenbus_watch_pathfmt(dev, &dev->otherend_watch, otherend_changed,
231 "%s/%s", dev->otherend, "state");
232}
233
234
235int xenbus_dev_probe(struct device *_dev)
236{
237 struct xenbus_device *dev = to_xenbus_device(_dev);
238 struct xenbus_driver *drv = to_xenbus_driver(_dev->driver);
239 const struct xenbus_device_id *id;
240 int err;
241
242 DPRINTK("%s", dev->nodename);
243
244 if (!drv->probe) {
245 err = -ENODEV;
246 goto fail;
247 }
248
249 id = match_device(drv->ids, dev);
250 if (!id) {
251 err = -ENODEV;
252 goto fail;
253 }
254
255 err = talk_to_otherend(dev);
256 if (err) {
257 dev_warn(&dev->dev, "talk_to_otherend on %s failed.\n",
258 dev->nodename);
259 return err;
260 }
261
262 err = drv->probe(dev, id);
263 if (err)
264 goto fail;
265
266 err = watch_otherend(dev);
267 if (err) {
268 dev_warn(&dev->dev, "watch_otherend on %s failed.\n",
269 dev->nodename);
270 return err;
271 }
272
273 return 0;
274fail:
275 xenbus_dev_error(dev, err, "xenbus_dev_probe on %s", dev->nodename);
276 xenbus_switch_state(dev, XenbusStateClosed);
277 return -ENODEV;
278}
279
280int xenbus_dev_remove(struct device *_dev)
281{
282 struct xenbus_device *dev = to_xenbus_device(_dev);
283 struct xenbus_driver *drv = to_xenbus_driver(_dev->driver);
284
285 DPRINTK("%s", dev->nodename);
286
287 free_otherend_watch(dev);
288 free_otherend_details(dev);
289
290 if (drv->remove)
291 drv->remove(dev);
292
293 xenbus_switch_state(dev, XenbusStateClosed);
294 return 0;
295}
296
297static void xenbus_dev_shutdown(struct device *_dev)
298{
299 struct xenbus_device *dev = to_xenbus_device(_dev);
300 unsigned long timeout = 5*HZ;
301
302 DPRINTK("%s", dev->nodename);
303
304 get_device(&dev->dev);
305 if (dev->state != XenbusStateConnected) {
306 printk(KERN_INFO "%s: %s: %s != Connected, skipping\n", __func__,
307 dev->nodename, xenbus_strstate(dev->state));
308 goto out;
309 }
310 xenbus_switch_state(dev, XenbusStateClosing);
311 timeout = wait_for_completion_timeout(&dev->down, timeout);
312 if (!timeout)
313 printk(KERN_INFO "%s: %s timeout closing device\n",
314 __func__, dev->nodename);
315 out:
316 put_device(&dev->dev);
317}
318
319int xenbus_register_driver_common(struct xenbus_driver *drv,
320 struct xen_bus_type *bus,
321 struct module *owner,
322 const char *mod_name)
323{
324 drv->driver.name = drv->name;
325 drv->driver.bus = &bus->bus;
326 drv->driver.owner = owner;
327 drv->driver.mod_name = mod_name;
328
329 return driver_register(&drv->driver);
330}
331
332int __xenbus_register_frontend(struct xenbus_driver *drv,
333 struct module *owner, const char *mod_name)
334{
335 int ret;
336
337 drv->read_otherend_details = read_backend_details;
338
339 ret = xenbus_register_driver_common(drv, &xenbus_frontend,
340 owner, mod_name);
341 if (ret)
342 return ret;
343
344 /* If this driver is loaded as a module wait for devices to attach. */
345 wait_for_devices(drv);
346
347 return 0;
348}
349EXPORT_SYMBOL_GPL(__xenbus_register_frontend);
350
351void xenbus_unregister_driver(struct xenbus_driver *drv)
352{
353 driver_unregister(&drv->driver);
354}
355EXPORT_SYMBOL_GPL(xenbus_unregister_driver);
356
357struct xb_find_info
358{
359 struct xenbus_device *dev;
360 const char *nodename;
361};
362
363static int cmp_dev(struct device *dev, void *data)
364{
365 struct xenbus_device *xendev = to_xenbus_device(dev);
366 struct xb_find_info *info = data;
367
368 if (!strcmp(xendev->nodename, info->nodename)) {
369 info->dev = xendev;
370 get_device(dev);
371 return 1;
372 }
373 return 0;
374}
375
376struct xenbus_device *xenbus_device_find(const char *nodename,
377 struct bus_type *bus)
378{
379 struct xb_find_info info = { .dev = NULL, .nodename = nodename };
380
381 bus_for_each_dev(bus, NULL, &info, cmp_dev);
382 return info.dev;
383}
384
385static int cleanup_dev(struct device *dev, void *data)
386{
387 struct xenbus_device *xendev = to_xenbus_device(dev);
388 struct xb_find_info *info = data;
389 int len = strlen(info->nodename);
390
391 DPRINTK("%s", info->nodename);
392
393 /* Match the info->nodename path, or any subdirectory of that path. */
394 if (strncmp(xendev->nodename, info->nodename, len))
395 return 0;
396
397 /* If the node name is longer, ensure it really is a subdirectory. */
398 if ((strlen(xendev->nodename) > len) && (xendev->nodename[len] != '/'))
399 return 0;
400
401 info->dev = xendev;
402 get_device(dev);
403 return 1;
404}
405
406static void xenbus_cleanup_devices(const char *path, struct bus_type *bus)
407{
408 struct xb_find_info info = { .nodename = path };
409
410 do {
411 info.dev = NULL;
412 bus_for_each_dev(bus, NULL, &info, cleanup_dev);
413 if (info.dev) {
414 device_unregister(&info.dev->dev);
415 put_device(&info.dev->dev);
416 }
417 } while (info.dev);
418}
419
420static void xenbus_dev_release(struct device *dev)
421{
422 if (dev)
423 kfree(to_xenbus_device(dev));
424}
425
426static ssize_t xendev_show_nodename(struct device *dev,
427 struct device_attribute *attr, char *buf)
428{
429 return sprintf(buf, "%s\n", to_xenbus_device(dev)->nodename);
430}
431DEVICE_ATTR(nodename, S_IRUSR | S_IRGRP | S_IROTH, xendev_show_nodename, NULL);
432
433static ssize_t xendev_show_devtype(struct device *dev,
434 struct device_attribute *attr, char *buf)
435{
436 return sprintf(buf, "%s\n", to_xenbus_device(dev)->devicetype);
437}
438DEVICE_ATTR(devtype, S_IRUSR | S_IRGRP | S_IROTH, xendev_show_devtype, NULL);
439
440
441int xenbus_probe_node(struct xen_bus_type *bus,
442 const char *type,
443 const char *nodename)
444{
445 int err;
446 struct xenbus_device *xendev;
447 size_t stringlen;
448 char *tmpstring;
449
450 enum xenbus_state state = xenbus_read_driver_state(nodename);
451
452 if (state != XenbusStateInitialising) {
453 /* Device is not new, so ignore it. This can happen if a
454 device is going away after switching to Closed. */
455 return 0;
456 }
457
458 stringlen = strlen(nodename) + 1 + strlen(type) + 1;
459 xendev = kzalloc(sizeof(*xendev) + stringlen, GFP_KERNEL);
460 if (!xendev)
461 return -ENOMEM;
462
463 xendev->state = XenbusStateInitialising;
464
465 /* Copy the strings into the extra space. */
466
467 tmpstring = (char *)(xendev + 1);
468 strcpy(tmpstring, nodename);
469 xendev->nodename = tmpstring;
470
471 tmpstring += strlen(tmpstring) + 1;
472 strcpy(tmpstring, type);
473 xendev->devicetype = tmpstring;
474 init_completion(&xendev->down);
475
476 xendev->dev.bus = &bus->bus;
477 xendev->dev.release = xenbus_dev_release;
478
479 err = bus->get_bus_id(xendev->dev.bus_id, xendev->nodename);
480 if (err)
481 goto fail;
482
483 /* Register with generic device framework. */
484 err = device_register(&xendev->dev);
485 if (err)
486 goto fail;
487
488 err = device_create_file(&xendev->dev, &dev_attr_nodename);
489 if (err)
490 goto fail_unregister;
491
492 err = device_create_file(&xendev->dev, &dev_attr_devtype);
493 if (err)
494 goto fail_remove_file;
495
496 return 0;
497fail_remove_file:
498 device_remove_file(&xendev->dev, &dev_attr_nodename);
499fail_unregister:
500 device_unregister(&xendev->dev);
501fail:
502 kfree(xendev);
503 return err;
504}
505
506/* device/<typename>/<name> */
507static int xenbus_probe_frontend(const char *type, const char *name)
508{
509 char *nodename;
510 int err;
511
512 nodename = kasprintf(GFP_KERNEL, "%s/%s/%s",
513 xenbus_frontend.root, type, name);
514 if (!nodename)
515 return -ENOMEM;
516
517 DPRINTK("%s", nodename);
518
519 err = xenbus_probe_node(&xenbus_frontend, type, nodename);
520 kfree(nodename);
521 return err;
522}
523
524static int xenbus_probe_device_type(struct xen_bus_type *bus, const char *type)
525{
526 int err = 0;
527 char **dir;
528 unsigned int dir_n = 0;
529 int i;
530
531 dir = xenbus_directory(XBT_NIL, bus->root, type, &dir_n);
532 if (IS_ERR(dir))
533 return PTR_ERR(dir);
534
535 for (i = 0; i < dir_n; i++) {
536 err = bus->probe(type, dir[i]);
537 if (err)
538 break;
539 }
540 kfree(dir);
541 return err;
542}
543
544int xenbus_probe_devices(struct xen_bus_type *bus)
545{
546 int err = 0;
547 char **dir;
548 unsigned int i, dir_n;
549
550 dir = xenbus_directory(XBT_NIL, bus->root, "", &dir_n);
551 if (IS_ERR(dir))
552 return PTR_ERR(dir);
553
554 for (i = 0; i < dir_n; i++) {
555 err = xenbus_probe_device_type(bus, dir[i]);
556 if (err)
557 break;
558 }
559 kfree(dir);
560 return err;
561}
562
563static unsigned int char_count(const char *str, char c)
564{
565 unsigned int i, ret = 0;
566
567 for (i = 0; str[i]; i++)
568 if (str[i] == c)
569 ret++;
570 return ret;
571}
572
573static int strsep_len(const char *str, char c, unsigned int len)
574{
575 unsigned int i;
576
577 for (i = 0; str[i]; i++)
578 if (str[i] == c) {
579 if (len == 0)
580 return i;
581 len--;
582 }
583 return (len == 0) ? i : -ERANGE;
584}
585
586void xenbus_dev_changed(const char *node, struct xen_bus_type *bus)
587{
588 int exists, rootlen;
589 struct xenbus_device *dev;
590 char type[BUS_ID_SIZE];
591 const char *p, *root;
592
593 if (char_count(node, '/') < 2)
594 return;
595
596 exists = xenbus_exists(XBT_NIL, node, "");
597 if (!exists) {
598 xenbus_cleanup_devices(node, &bus->bus);
599 return;
600 }
601
602 /* backend/<type>/... or device/<type>/... */
603 p = strchr(node, '/') + 1;
604 snprintf(type, BUS_ID_SIZE, "%.*s", (int)strcspn(p, "/"), p);
605 type[BUS_ID_SIZE-1] = '\0';
606
607 rootlen = strsep_len(node, '/', bus->levels);
608 if (rootlen < 0)
609 return;
610 root = kasprintf(GFP_KERNEL, "%.*s", rootlen, node);
611 if (!root)
612 return;
613
614 dev = xenbus_device_find(root, &bus->bus);
615 if (!dev)
616 xenbus_probe_node(bus, type, root);
617 else
618 put_device(&dev->dev);
619
620 kfree(root);
621}
622
623static void frontend_changed(struct xenbus_watch *watch,
624 const char **vec, unsigned int len)
625{
626 DPRINTK("");
627
628 xenbus_dev_changed(vec[XS_WATCH_PATH], &xenbus_frontend);
629}
630
631/* We watch for devices appearing and vanishing. */
632static struct xenbus_watch fe_watch = {
633 .node = "device",
634 .callback = frontend_changed,
635};
636
637static int suspend_dev(struct device *dev, void *data)
638{
639 int err = 0;
640 struct xenbus_driver *drv;
641 struct xenbus_device *xdev;
642
643 DPRINTK("");
644
645 if (dev->driver == NULL)
646 return 0;
647 drv = to_xenbus_driver(dev->driver);
648 xdev = container_of(dev, struct xenbus_device, dev);
649 if (drv->suspend)
650 err = drv->suspend(xdev);
651 if (err)
652 printk(KERN_WARNING
653 "xenbus: suspend %s failed: %i\n", dev->bus_id, err);
654 return 0;
655}
656
657static int suspend_cancel_dev(struct device *dev, void *data)
658{
659 int err = 0;
660 struct xenbus_driver *drv;
661 struct xenbus_device *xdev;
662
663 DPRINTK("");
664
665 if (dev->driver == NULL)
666 return 0;
667 drv = to_xenbus_driver(dev->driver);
668 xdev = container_of(dev, struct xenbus_device, dev);
669 if (drv->suspend_cancel)
670 err = drv->suspend_cancel(xdev);
671 if (err)
672 printk(KERN_WARNING
673 "xenbus: suspend_cancel %s failed: %i\n",
674 dev->bus_id, err);
675 return 0;
676}
677
678static int resume_dev(struct device *dev, void *data)
679{
680 int err;
681 struct xenbus_driver *drv;
682 struct xenbus_device *xdev;
683
684 DPRINTK("");
685
686 if (dev->driver == NULL)
687 return 0;
688
689 drv = to_xenbus_driver(dev->driver);
690 xdev = container_of(dev, struct xenbus_device, dev);
691
692 err = talk_to_otherend(xdev);
693 if (err) {
694 printk(KERN_WARNING
695 "xenbus: resume (talk_to_otherend) %s failed: %i\n",
696 dev->bus_id, err);
697 return err;
698 }
699
700 xdev->state = XenbusStateInitialising;
701
702 if (drv->resume) {
703 err = drv->resume(xdev);
704 if (err) {
705 printk(KERN_WARNING
706 "xenbus: resume %s failed: %i\n",
707 dev->bus_id, err);
708 return err;
709 }
710 }
711
712 err = watch_otherend(xdev);
713 if (err) {
714 printk(KERN_WARNING
715 "xenbus_probe: resume (watch_otherend) %s failed: "
716 "%d.\n", dev->bus_id, err);
717 return err;
718 }
719
720 return 0;
721}
722
723void xenbus_suspend(void)
724{
725 DPRINTK("");
726
727 bus_for_each_dev(&xenbus_frontend.bus, NULL, NULL, suspend_dev);
728 xenbus_backend_suspend(suspend_dev);
729 xs_suspend();
730}
731EXPORT_SYMBOL_GPL(xenbus_suspend);
732
733void xenbus_resume(void)
734{
735 xb_init_comms();
736 xs_resume();
737 bus_for_each_dev(&xenbus_frontend.bus, NULL, NULL, resume_dev);
738 xenbus_backend_resume(resume_dev);
739}
740EXPORT_SYMBOL_GPL(xenbus_resume);
741
742void xenbus_suspend_cancel(void)
743{
744 xs_suspend_cancel();
745 bus_for_each_dev(&xenbus_frontend.bus, NULL, NULL, suspend_cancel_dev);
746 xenbus_backend_resume(suspend_cancel_dev);
747}
748EXPORT_SYMBOL_GPL(xenbus_suspend_cancel);
749
750/* A flag to determine if xenstored is 'ready' (i.e. has started) */
751int xenstored_ready = 0;
752
753
754int register_xenstore_notifier(struct notifier_block *nb)
755{
756 int ret = 0;
757
758 if (xenstored_ready > 0)
759 ret = nb->notifier_call(nb, 0, NULL);
760 else
761 blocking_notifier_chain_register(&xenstore_chain, nb);
762
763 return ret;
764}
765EXPORT_SYMBOL_GPL(register_xenstore_notifier);
766
767void unregister_xenstore_notifier(struct notifier_block *nb)
768{
769 blocking_notifier_chain_unregister(&xenstore_chain, nb);
770}
771EXPORT_SYMBOL_GPL(unregister_xenstore_notifier);
772
773void xenbus_probe(struct work_struct *unused)
774{
775 BUG_ON((xenstored_ready <= 0));
776
777 /* Enumerate devices in xenstore and watch for changes. */
778 xenbus_probe_devices(&xenbus_frontend);
779 register_xenbus_watch(&fe_watch);
780 xenbus_backend_probe_and_watch();
781
782 /* Notify others that xenstore is up */
783 blocking_notifier_call_chain(&xenstore_chain, 0, NULL);
784}
785
786static int __init xenbus_probe_init(void)
787{
788 int err = 0;
789
790 DPRINTK("");
791
792 err = -ENODEV;
793 if (!is_running_on_xen())
794 goto out_error;
795
796 /* Register ourselves with the kernel bus subsystem */
797 err = bus_register(&xenbus_frontend.bus);
798 if (err)
799 goto out_error;
800
801 err = xenbus_backend_bus_register();
802 if (err)
803 goto out_unreg_front;
804
805 /*
806 * Domain0 doesn't have a store_evtchn or store_mfn yet.
807 */
808 if (is_initial_xendomain()) {
809 /* dom0 not yet supported */
810 } else {
811 xenstored_ready = 1;
812 xen_store_evtchn = xen_start_info->store_evtchn;
813 xen_store_mfn = xen_start_info->store_mfn;
814 }
815 xen_store_interface = mfn_to_virt(xen_store_mfn);
816
817 /* Initialize the interface to xenstore. */
818 err = xs_init();
819 if (err) {
820 printk(KERN_WARNING
821 "XENBUS: Error initializing xenstore comms: %i\n", err);
822 goto out_unreg_back;
823 }
824
825 if (!is_initial_xendomain())
826 xenbus_probe(NULL);
827
828 return 0;
829
830 out_unreg_back:
831 xenbus_backend_bus_unregister();
832
833 out_unreg_front:
834 bus_unregister(&xenbus_frontend.bus);
835
836 out_error:
837 return err;
838}
839
840postcore_initcall(xenbus_probe_init);
841
842MODULE_LICENSE("GPL");
843
844static int is_disconnected_device(struct device *dev, void *data)
845{
846 struct xenbus_device *xendev = to_xenbus_device(dev);
847 struct device_driver *drv = data;
848
849 /*
850 * A device with no driver will never connect. We care only about
851 * devices which should currently be in the process of connecting.
852 */
853 if (!dev->driver)
854 return 0;
855
856 /* Is this search limited to a particular driver? */
857 if (drv && (dev->driver != drv))
858 return 0;
859
860 return (xendev->state != XenbusStateConnected);
861}
862
863static int exists_disconnected_device(struct device_driver *drv)
864{
865 return bus_for_each_dev(&xenbus_frontend.bus, NULL, drv,
866 is_disconnected_device);
867}
868
869static int print_device_status(struct device *dev, void *data)
870{
871 struct xenbus_device *xendev = to_xenbus_device(dev);
872 struct device_driver *drv = data;
873
874 /* Is this operation limited to a particular driver? */
875 if (drv && (dev->driver != drv))
876 return 0;
877
878 if (!dev->driver) {
879 /* Information only: is this too noisy? */
880 printk(KERN_INFO "XENBUS: Device with no driver: %s\n",
881 xendev->nodename);
882 } else if (xendev->state != XenbusStateConnected) {
883 printk(KERN_WARNING "XENBUS: Timeout connecting "
884 "to device: %s (state %d)\n",
885 xendev->nodename, xendev->state);
886 }
887
888 return 0;
889}
890
891/* We only wait for device setup after most initcalls have run. */
892static int ready_to_wait_for_devices;
893
894/*
895 * On a 10 second timeout, wait for all devices currently configured. We need
896 * to do this to guarantee that the filesystems and / or network devices
897 * needed for boot are available, before we can allow the boot to proceed.
898 *
899 * This needs to be on a late_initcall, to happen after the frontend device
900 * drivers have been initialised, but before the root fs is mounted.
901 *
902 * A possible improvement here would be to have the tools add a per-device
903 * flag to the store entry, indicating whether it is needed at boot time.
904 * This would allow people who knew what they were doing to accelerate their
905 * boot slightly, but of course needs tools or manual intervention to set up
906 * those flags correctly.
907 */
908static void wait_for_devices(struct xenbus_driver *xendrv)
909{
910 unsigned long timeout = jiffies + 10*HZ;
911 struct device_driver *drv = xendrv ? &xendrv->driver : NULL;
912
913 if (!ready_to_wait_for_devices || !is_running_on_xen())
914 return;
915
916 while (exists_disconnected_device(drv)) {
917 if (time_after(jiffies, timeout))
918 break;
919 schedule_timeout_interruptible(HZ/10);
920 }
921
922 bus_for_each_dev(&xenbus_frontend.bus, NULL, drv,
923 print_device_status);
924}
925
926#ifndef MODULE
927static int __init boot_wait_for_devices(void)
928{
929 ready_to_wait_for_devices = 1;
930 wait_for_devices(NULL);
931 return 0;
932}
933
934late_initcall(boot_wait_for_devices);
935#endif
diff --git a/drivers/xen/xenbus/xenbus_probe.h b/drivers/xen/xenbus/xenbus_probe.h
new file mode 100644
index 000000000000..e09b19415a40
--- /dev/null
+++ b/drivers/xen/xenbus/xenbus_probe.h
@@ -0,0 +1,74 @@
1/******************************************************************************
2 * xenbus_probe.h
3 *
4 * Talks to Xen Store to figure out what devices we have.
5 *
6 * Copyright (C) 2005 Rusty Russell, IBM Corporation
7 * Copyright (C) 2005 XenSource Ltd.
8 *
9 * This program is free software; you can redistribute it and/or
10 * modify it under the terms of the GNU General Public License version 2
11 * as published by the Free Software Foundation; or, when distributed
12 * separately from the Linux kernel or incorporated into other
13 * software packages, subject to the following license:
14 *
15 * Permission is hereby granted, free of charge, to any person obtaining a copy
16 * of this source file (the "Software"), to deal in the Software without
17 * restriction, including without limitation the rights to use, copy, modify,
18 * merge, publish, distribute, sublicense, and/or sell copies of the Software,
19 * and to permit persons to whom the Software is furnished to do so, subject to
20 * the following conditions:
21 *
22 * The above copyright notice and this permission notice shall be included in
23 * all copies or substantial portions of the Software.
24 *
25 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
26 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
27 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
28 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
29 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
30 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
31 * IN THE SOFTWARE.
32 */
33
34#ifndef _XENBUS_PROBE_H
35#define _XENBUS_PROBE_H
36
37#ifdef CONFIG_XEN_BACKEND
38extern void xenbus_backend_suspend(int (*fn)(struct device *, void *));
39extern void xenbus_backend_resume(int (*fn)(struct device *, void *));
40extern void xenbus_backend_probe_and_watch(void);
41extern int xenbus_backend_bus_register(void);
42extern void xenbus_backend_bus_unregister(void);
43#else
44static inline void xenbus_backend_suspend(int (*fn)(struct device *, void *)) {}
45static inline void xenbus_backend_resume(int (*fn)(struct device *, void *)) {}
46static inline void xenbus_backend_probe_and_watch(void) {}
47static inline int xenbus_backend_bus_register(void) { return 0; }
48static inline void xenbus_backend_bus_unregister(void) {}
49#endif
50
51struct xen_bus_type
52{
53 char *root;
54 unsigned int levels;
55 int (*get_bus_id)(char bus_id[BUS_ID_SIZE], const char *nodename);
56 int (*probe)(const char *type, const char *dir);
57 struct bus_type bus;
58};
59
60extern int xenbus_match(struct device *_dev, struct device_driver *_drv);
61extern int xenbus_dev_probe(struct device *_dev);
62extern int xenbus_dev_remove(struct device *_dev);
63extern int xenbus_register_driver_common(struct xenbus_driver *drv,
64 struct xen_bus_type *bus,
65 struct module *owner,
66 const char *mod_name);
67extern int xenbus_probe_node(struct xen_bus_type *bus,
68 const char *type,
69 const char *nodename);
70extern int xenbus_probe_devices(struct xen_bus_type *bus);
71
72extern void xenbus_dev_changed(const char *node, struct xen_bus_type *bus);
73
74#endif
diff --git a/drivers/xen/xenbus/xenbus_xs.c b/drivers/xen/xenbus/xenbus_xs.c
new file mode 100644
index 000000000000..9e943fbce81b
--- /dev/null
+++ b/drivers/xen/xenbus/xenbus_xs.c
@@ -0,0 +1,861 @@
1/******************************************************************************
2 * xenbus_xs.c
3 *
4 * This is the kernel equivalent of the "xs" library. We don't need everything
5 * and we use xenbus_comms for communication.
6 *
7 * Copyright (C) 2005 Rusty Russell, IBM Corporation
8 *
9 * This program is free software; you can redistribute it and/or
10 * modify it under the terms of the GNU General Public License version 2
11 * as published by the Free Software Foundation; or, when distributed
12 * separately from the Linux kernel or incorporated into other
13 * software packages, subject to the following license:
14 *
15 * Permission is hereby granted, free of charge, to any person obtaining a copy
16 * of this source file (the "Software"), to deal in the Software without
17 * restriction, including without limitation the rights to use, copy, modify,
18 * merge, publish, distribute, sublicense, and/or sell copies of the Software,
19 * and to permit persons to whom the Software is furnished to do so, subject to
20 * the following conditions:
21 *
22 * The above copyright notice and this permission notice shall be included in
23 * all copies or substantial portions of the Software.
24 *
25 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
26 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
27 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
28 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
29 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
30 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
31 * IN THE SOFTWARE.
32 */
33
34#include <linux/unistd.h>
35#include <linux/errno.h>
36#include <linux/types.h>
37#include <linux/uio.h>
38#include <linux/kernel.h>
39#include <linux/string.h>
40#include <linux/err.h>
41#include <linux/slab.h>
42#include <linux/fcntl.h>
43#include <linux/kthread.h>
44#include <linux/rwsem.h>
45#include <linux/module.h>
46#include <linux/mutex.h>
47#include <xen/xenbus.h>
48#include "xenbus_comms.h"
49
50struct xs_stored_msg {
51 struct list_head list;
52
53 struct xsd_sockmsg hdr;
54
55 union {
56 /* Queued replies. */
57 struct {
58 char *body;
59 } reply;
60
61 /* Queued watch events. */
62 struct {
63 struct xenbus_watch *handle;
64 char **vec;
65 unsigned int vec_size;
66 } watch;
67 } u;
68};
69
70struct xs_handle {
71 /* A list of replies. Currently only one will ever be outstanding. */
72 struct list_head reply_list;
73 spinlock_t reply_lock;
74 wait_queue_head_t reply_waitq;
75
76 /*
77 * Mutex ordering: transaction_mutex -> watch_mutex -> request_mutex.
78 * response_mutex is never taken simultaneously with the other three.
79 */
80
81 /* One request at a time. */
82 struct mutex request_mutex;
83
84 /* Protect xenbus reader thread against save/restore. */
85 struct mutex response_mutex;
86
87 /* Protect transactions against save/restore. */
88 struct rw_semaphore transaction_mutex;
89
90 /* Protect watch (de)register against save/restore. */
91 struct rw_semaphore watch_mutex;
92};
93
94static struct xs_handle xs_state;
95
96/* List of registered watches, and a lock to protect it. */
97static LIST_HEAD(watches);
98static DEFINE_SPINLOCK(watches_lock);
99
100/* List of pending watch callback events, and a lock to protect it. */
101static LIST_HEAD(watch_events);
102static DEFINE_SPINLOCK(watch_events_lock);
103
104/*
105 * Details of the xenwatch callback kernel thread. The thread waits on the
106 * watch_events_waitq for work to do (queued on watch_events list). When it
107 * wakes up it acquires the xenwatch_mutex before reading the list and
108 * carrying out work.
109 */
110static pid_t xenwatch_pid;
111static DEFINE_MUTEX(xenwatch_mutex);
112static DECLARE_WAIT_QUEUE_HEAD(watch_events_waitq);
113
114static int get_error(const char *errorstring)
115{
116 unsigned int i;
117
118 for (i = 0; strcmp(errorstring, xsd_errors[i].errstring) != 0; i++) {
119 if (i == ARRAY_SIZE(xsd_errors) - 1) {
120 printk(KERN_WARNING
121 "XENBUS xen store gave: unknown error %s",
122 errorstring);
123 return EINVAL;
124 }
125 }
126 return xsd_errors[i].errnum;
127}
128
129static void *read_reply(enum xsd_sockmsg_type *type, unsigned int *len)
130{
131 struct xs_stored_msg *msg;
132 char *body;
133
134 spin_lock(&xs_state.reply_lock);
135
136 while (list_empty(&xs_state.reply_list)) {
137 spin_unlock(&xs_state.reply_lock);
138 /* XXX FIXME: Avoid synchronous wait for response here. */
139 wait_event(xs_state.reply_waitq,
140 !list_empty(&xs_state.reply_list));
141 spin_lock(&xs_state.reply_lock);
142 }
143
144 msg = list_entry(xs_state.reply_list.next,
145 struct xs_stored_msg, list);
146 list_del(&msg->list);
147
148 spin_unlock(&xs_state.reply_lock);
149
150 *type = msg->hdr.type;
151 if (len)
152 *len = msg->hdr.len;
153 body = msg->u.reply.body;
154
155 kfree(msg);
156
157 return body;
158}
159
160void *xenbus_dev_request_and_reply(struct xsd_sockmsg *msg)
161{
162 void *ret;
163 struct xsd_sockmsg req_msg = *msg;
164 int err;
165
166 if (req_msg.type == XS_TRANSACTION_START)
167 down_read(&xs_state.transaction_mutex);
168
169 mutex_lock(&xs_state.request_mutex);
170
171 err = xb_write(msg, sizeof(*msg) + msg->len);
172 if (err) {
173 msg->type = XS_ERROR;
174 ret = ERR_PTR(err);
175 } else
176 ret = read_reply(&msg->type, &msg->len);
177
178 mutex_unlock(&xs_state.request_mutex);
179
180 if ((msg->type == XS_TRANSACTION_END) ||
181 ((req_msg.type == XS_TRANSACTION_START) &&
182 (msg->type == XS_ERROR)))
183 up_read(&xs_state.transaction_mutex);
184
185 return ret;
186}
187
188/* Send message to xs, get kmalloc'ed reply. ERR_PTR() on error. */
189static void *xs_talkv(struct xenbus_transaction t,
190 enum xsd_sockmsg_type type,
191 const struct kvec *iovec,
192 unsigned int num_vecs,
193 unsigned int *len)
194{
195 struct xsd_sockmsg msg;
196 void *ret = NULL;
197 unsigned int i;
198 int err;
199
200 msg.tx_id = t.id;
201 msg.req_id = 0;
202 msg.type = type;
203 msg.len = 0;
204 for (i = 0; i < num_vecs; i++)
205 msg.len += iovec[i].iov_len;
206
207 mutex_lock(&xs_state.request_mutex);
208
209 err = xb_write(&msg, sizeof(msg));
210 if (err) {
211 mutex_unlock(&xs_state.request_mutex);
212 return ERR_PTR(err);
213 }
214
215 for (i = 0; i < num_vecs; i++) {
216 err = xb_write(iovec[i].iov_base, iovec[i].iov_len);
217 if (err) {
218 mutex_unlock(&xs_state.request_mutex);
219 return ERR_PTR(err);
220 }
221 }
222
223 ret = read_reply(&msg.type, len);
224
225 mutex_unlock(&xs_state.request_mutex);
226
227 if (IS_ERR(ret))
228 return ret;
229
230 if (msg.type == XS_ERROR) {
231 err = get_error(ret);
232 kfree(ret);
233 return ERR_PTR(-err);
234 }
235
236 if (msg.type != type) {
237 if (printk_ratelimit())
238 printk(KERN_WARNING
239 "XENBUS unexpected type [%d], expected [%d]\n",
240 msg.type, type);
241 kfree(ret);
242 return ERR_PTR(-EINVAL);
243 }
244 return ret;
245}
246
247/* Simplified version of xs_talkv: single message. */
248static void *xs_single(struct xenbus_transaction t,
249 enum xsd_sockmsg_type type,
250 const char *string,
251 unsigned int *len)
252{
253 struct kvec iovec;
254
255 iovec.iov_base = (void *)string;
256 iovec.iov_len = strlen(string) + 1;
257 return xs_talkv(t, type, &iovec, 1, len);
258}
259
260/* Many commands only need an ack, don't care what it says. */
261static int xs_error(char *reply)
262{
263 if (IS_ERR(reply))
264 return PTR_ERR(reply);
265 kfree(reply);
266 return 0;
267}
268
269static unsigned int count_strings(const char *strings, unsigned int len)
270{
271 unsigned int num;
272 const char *p;
273
274 for (p = strings, num = 0; p < strings + len; p += strlen(p) + 1)
275 num++;
276
277 return num;
278}
279
280/* Return the path to dir with /name appended. Buffer must be kfree()'ed. */
281static char *join(const char *dir, const char *name)
282{
283 char *buffer;
284
285 if (strlen(name) == 0)
286 buffer = kasprintf(GFP_KERNEL, "%s", dir);
287 else
288 buffer = kasprintf(GFP_KERNEL, "%s/%s", dir, name);
289 return (!buffer) ? ERR_PTR(-ENOMEM) : buffer;
290}
291
292static char **split(char *strings, unsigned int len, unsigned int *num)
293{
294 char *p, **ret;
295
296 /* Count the strings. */
297 *num = count_strings(strings, len);
298
299 /* Transfer to one big alloc for easy freeing. */
300 ret = kmalloc(*num * sizeof(char *) + len, GFP_KERNEL);
301 if (!ret) {
302 kfree(strings);
303 return ERR_PTR(-ENOMEM);
304 }
305 memcpy(&ret[*num], strings, len);
306 kfree(strings);
307
308 strings = (char *)&ret[*num];
309 for (p = strings, *num = 0; p < strings + len; p += strlen(p) + 1)
310 ret[(*num)++] = p;
311
312 return ret;
313}
314
315char **xenbus_directory(struct xenbus_transaction t,
316 const char *dir, const char *node, unsigned int *num)
317{
318 char *strings, *path;
319 unsigned int len;
320
321 path = join(dir, node);
322 if (IS_ERR(path))
323 return (char **)path;
324
325 strings = xs_single(t, XS_DIRECTORY, path, &len);
326 kfree(path);
327 if (IS_ERR(strings))
328 return (char **)strings;
329
330 return split(strings, len, num);
331}
332EXPORT_SYMBOL_GPL(xenbus_directory);
333
334/* Check if a path exists. Return 1 if it does. */
335int xenbus_exists(struct xenbus_transaction t,
336 const char *dir, const char *node)
337{
338 char **d;
339 int dir_n;
340
341 d = xenbus_directory(t, dir, node, &dir_n);
342 if (IS_ERR(d))
343 return 0;
344 kfree(d);
345 return 1;
346}
347EXPORT_SYMBOL_GPL(xenbus_exists);
348
349/* Get the value of a single file.
350 * Returns a kmalloced value: call free() on it after use.
351 * len indicates length in bytes.
352 */
353void *xenbus_read(struct xenbus_transaction t,
354 const char *dir, const char *node, unsigned int *len)
355{
356 char *path;
357 void *ret;
358
359 path = join(dir, node);
360 if (IS_ERR(path))
361 return (void *)path;
362
363 ret = xs_single(t, XS_READ, path, len);
364 kfree(path);
365 return ret;
366}
367EXPORT_SYMBOL_GPL(xenbus_read);
368
369/* Write the value of a single file.
370 * Returns -err on failure.
371 */
372int xenbus_write(struct xenbus_transaction t,
373 const char *dir, const char *node, const char *string)
374{
375 const char *path;
376 struct kvec iovec[2];
377 int ret;
378
379 path = join(dir, node);
380 if (IS_ERR(path))
381 return PTR_ERR(path);
382
383 iovec[0].iov_base = (void *)path;
384 iovec[0].iov_len = strlen(path) + 1;
385 iovec[1].iov_base = (void *)string;
386 iovec[1].iov_len = strlen(string);
387
388 ret = xs_error(xs_talkv(t, XS_WRITE, iovec, ARRAY_SIZE(iovec), NULL));
389 kfree(path);
390 return ret;
391}
392EXPORT_SYMBOL_GPL(xenbus_write);
393
394/* Create a new directory. */
395int xenbus_mkdir(struct xenbus_transaction t,
396 const char *dir, const char *node)
397{
398 char *path;
399 int ret;
400
401 path = join(dir, node);
402 if (IS_ERR(path))
403 return PTR_ERR(path);
404
405 ret = xs_error(xs_single(t, XS_MKDIR, path, NULL));
406 kfree(path);
407 return ret;
408}
409EXPORT_SYMBOL_GPL(xenbus_mkdir);
410
411/* Destroy a file or directory (directories must be empty). */
412int xenbus_rm(struct xenbus_transaction t, const char *dir, const char *node)
413{
414 char *path;
415 int ret;
416
417 path = join(dir, node);
418 if (IS_ERR(path))
419 return PTR_ERR(path);
420
421 ret = xs_error(xs_single(t, XS_RM, path, NULL));
422 kfree(path);
423 return ret;
424}
425EXPORT_SYMBOL_GPL(xenbus_rm);
426
427/* Start a transaction: changes by others will not be seen during this
428 * transaction, and changes will not be visible to others until end.
429 */
430int xenbus_transaction_start(struct xenbus_transaction *t)
431{
432 char *id_str;
433
434 down_read(&xs_state.transaction_mutex);
435
436 id_str = xs_single(XBT_NIL, XS_TRANSACTION_START, "", NULL);
437 if (IS_ERR(id_str)) {
438 up_read(&xs_state.transaction_mutex);
439 return PTR_ERR(id_str);
440 }
441
442 t->id = simple_strtoul(id_str, NULL, 0);
443 kfree(id_str);
444 return 0;
445}
446EXPORT_SYMBOL_GPL(xenbus_transaction_start);
447
448/* End a transaction.
449 * If abandon is true, transaction is discarded instead of committed.
450 */
451int xenbus_transaction_end(struct xenbus_transaction t, int abort)
452{
453 char abortstr[2];
454 int err;
455
456 if (abort)
457 strcpy(abortstr, "F");
458 else
459 strcpy(abortstr, "T");
460
461 err = xs_error(xs_single(t, XS_TRANSACTION_END, abortstr, NULL));
462
463 up_read(&xs_state.transaction_mutex);
464
465 return err;
466}
467EXPORT_SYMBOL_GPL(xenbus_transaction_end);
468
469/* Single read and scanf: returns -errno or num scanned. */
470int xenbus_scanf(struct xenbus_transaction t,
471 const char *dir, const char *node, const char *fmt, ...)
472{
473 va_list ap;
474 int ret;
475 char *val;
476
477 val = xenbus_read(t, dir, node, NULL);
478 if (IS_ERR(val))
479 return PTR_ERR(val);
480
481 va_start(ap, fmt);
482 ret = vsscanf(val, fmt, ap);
483 va_end(ap);
484 kfree(val);
485 /* Distinctive errno. */
486 if (ret == 0)
487 return -ERANGE;
488 return ret;
489}
490EXPORT_SYMBOL_GPL(xenbus_scanf);
491
492/* Single printf and write: returns -errno or 0. */
493int xenbus_printf(struct xenbus_transaction t,
494 const char *dir, const char *node, const char *fmt, ...)
495{
496 va_list ap;
497 int ret;
498#define PRINTF_BUFFER_SIZE 4096
499 char *printf_buffer;
500
501 printf_buffer = kmalloc(PRINTF_BUFFER_SIZE, GFP_KERNEL);
502 if (printf_buffer == NULL)
503 return -ENOMEM;
504
505 va_start(ap, fmt);
506 ret = vsnprintf(printf_buffer, PRINTF_BUFFER_SIZE, fmt, ap);
507 va_end(ap);
508
509 BUG_ON(ret > PRINTF_BUFFER_SIZE-1);
510 ret = xenbus_write(t, dir, node, printf_buffer);
511
512 kfree(printf_buffer);
513
514 return ret;
515}
516EXPORT_SYMBOL_GPL(xenbus_printf);
517
518/* Takes tuples of names, scanf-style args, and void **, NULL terminated. */
519int xenbus_gather(struct xenbus_transaction t, const char *dir, ...)
520{
521 va_list ap;
522 const char *name;
523 int ret = 0;
524
525 va_start(ap, dir);
526 while (ret == 0 && (name = va_arg(ap, char *)) != NULL) {
527 const char *fmt = va_arg(ap, char *);
528 void *result = va_arg(ap, void *);
529 char *p;
530
531 p = xenbus_read(t, dir, name, NULL);
532 if (IS_ERR(p)) {
533 ret = PTR_ERR(p);
534 break;
535 }
536 if (fmt) {
537 if (sscanf(p, fmt, result) == 0)
538 ret = -EINVAL;
539 kfree(p);
540 } else
541 *(char **)result = p;
542 }
543 va_end(ap);
544 return ret;
545}
546EXPORT_SYMBOL_GPL(xenbus_gather);
547
548static int xs_watch(const char *path, const char *token)
549{
550 struct kvec iov[2];
551
552 iov[0].iov_base = (void *)path;
553 iov[0].iov_len = strlen(path) + 1;
554 iov[1].iov_base = (void *)token;
555 iov[1].iov_len = strlen(token) + 1;
556
557 return xs_error(xs_talkv(XBT_NIL, XS_WATCH, iov,
558 ARRAY_SIZE(iov), NULL));
559}
560
561static int xs_unwatch(const char *path, const char *token)
562{
563 struct kvec iov[2];
564
565 iov[0].iov_base = (char *)path;
566 iov[0].iov_len = strlen(path) + 1;
567 iov[1].iov_base = (char *)token;
568 iov[1].iov_len = strlen(token) + 1;
569
570 return xs_error(xs_talkv(XBT_NIL, XS_UNWATCH, iov,
571 ARRAY_SIZE(iov), NULL));
572}
573
574static struct xenbus_watch *find_watch(const char *token)
575{
576 struct xenbus_watch *i, *cmp;
577
578 cmp = (void *)simple_strtoul(token, NULL, 16);
579
580 list_for_each_entry(i, &watches, list)
581 if (i == cmp)
582 return i;
583
584 return NULL;
585}
586
587/* Register callback to watch this node. */
588int register_xenbus_watch(struct xenbus_watch *watch)
589{
590 /* Pointer in ascii is the token. */
591 char token[sizeof(watch) * 2 + 1];
592 int err;
593
594 sprintf(token, "%lX", (long)watch);
595
596 down_read(&xs_state.watch_mutex);
597
598 spin_lock(&watches_lock);
599 BUG_ON(find_watch(token));
600 list_add(&watch->list, &watches);
601 spin_unlock(&watches_lock);
602
603 err = xs_watch(watch->node, token);
604
605 /* Ignore errors due to multiple registration. */
606 if ((err != 0) && (err != -EEXIST)) {
607 spin_lock(&watches_lock);
608 list_del(&watch->list);
609 spin_unlock(&watches_lock);
610 }
611
612 up_read(&xs_state.watch_mutex);
613
614 return err;
615}
616EXPORT_SYMBOL_GPL(register_xenbus_watch);
617
618void unregister_xenbus_watch(struct xenbus_watch *watch)
619{
620 struct xs_stored_msg *msg, *tmp;
621 char token[sizeof(watch) * 2 + 1];
622 int err;
623
624 sprintf(token, "%lX", (long)watch);
625
626 down_read(&xs_state.watch_mutex);
627
628 spin_lock(&watches_lock);
629 BUG_ON(!find_watch(token));
630 list_del(&watch->list);
631 spin_unlock(&watches_lock);
632
633 err = xs_unwatch(watch->node, token);
634 if (err)
635 printk(KERN_WARNING
636 "XENBUS Failed to release watch %s: %i\n",
637 watch->node, err);
638
639 up_read(&xs_state.watch_mutex);
640
641 /* Make sure there are no callbacks running currently (unless
642 its us) */
643 if (current->pid != xenwatch_pid)
644 mutex_lock(&xenwatch_mutex);
645
646 /* Cancel pending watch events. */
647 spin_lock(&watch_events_lock);
648 list_for_each_entry_safe(msg, tmp, &watch_events, list) {
649 if (msg->u.watch.handle != watch)
650 continue;
651 list_del(&msg->list);
652 kfree(msg->u.watch.vec);
653 kfree(msg);
654 }
655 spin_unlock(&watch_events_lock);
656
657 if (current->pid != xenwatch_pid)
658 mutex_unlock(&xenwatch_mutex);
659}
660EXPORT_SYMBOL_GPL(unregister_xenbus_watch);
661
662void xs_suspend(void)
663{
664 down_write(&xs_state.transaction_mutex);
665 down_write(&xs_state.watch_mutex);
666 mutex_lock(&xs_state.request_mutex);
667 mutex_lock(&xs_state.response_mutex);
668}
669
670void xs_resume(void)
671{
672 struct xenbus_watch *watch;
673 char token[sizeof(watch) * 2 + 1];
674
675 mutex_unlock(&xs_state.response_mutex);
676 mutex_unlock(&xs_state.request_mutex);
677 up_write(&xs_state.transaction_mutex);
678
679 /* No need for watches_lock: the watch_mutex is sufficient. */
680 list_for_each_entry(watch, &watches, list) {
681 sprintf(token, "%lX", (long)watch);
682 xs_watch(watch->node, token);
683 }
684
685 up_write(&xs_state.watch_mutex);
686}
687
688void xs_suspend_cancel(void)
689{
690 mutex_unlock(&xs_state.response_mutex);
691 mutex_unlock(&xs_state.request_mutex);
692 up_write(&xs_state.watch_mutex);
693 up_write(&xs_state.transaction_mutex);
694}
695
696static int xenwatch_thread(void *unused)
697{
698 struct list_head *ent;
699 struct xs_stored_msg *msg;
700
701 for (;;) {
702 wait_event_interruptible(watch_events_waitq,
703 !list_empty(&watch_events));
704
705 if (kthread_should_stop())
706 break;
707
708 mutex_lock(&xenwatch_mutex);
709
710 spin_lock(&watch_events_lock);
711 ent = watch_events.next;
712 if (ent != &watch_events)
713 list_del(ent);
714 spin_unlock(&watch_events_lock);
715
716 if (ent != &watch_events) {
717 msg = list_entry(ent, struct xs_stored_msg, list);
718 msg->u.watch.handle->callback(
719 msg->u.watch.handle,
720 (const char **)msg->u.watch.vec,
721 msg->u.watch.vec_size);
722 kfree(msg->u.watch.vec);
723 kfree(msg);
724 }
725
726 mutex_unlock(&xenwatch_mutex);
727 }
728
729 return 0;
730}
731
732static int process_msg(void)
733{
734 struct xs_stored_msg *msg;
735 char *body;
736 int err;
737
738 /*
739 * We must disallow save/restore while reading a xenstore message.
740 * A partial read across s/r leaves us out of sync with xenstored.
741 */
742 for (;;) {
743 err = xb_wait_for_data_to_read();
744 if (err)
745 return err;
746 mutex_lock(&xs_state.response_mutex);
747 if (xb_data_to_read())
748 break;
749 /* We raced with save/restore: pending data 'disappeared'. */
750 mutex_unlock(&xs_state.response_mutex);
751 }
752
753
754 msg = kmalloc(sizeof(*msg), GFP_KERNEL);
755 if (msg == NULL) {
756 err = -ENOMEM;
757 goto out;
758 }
759
760 err = xb_read(&msg->hdr, sizeof(msg->hdr));
761 if (err) {
762 kfree(msg);
763 goto out;
764 }
765
766 body = kmalloc(msg->hdr.len + 1, GFP_KERNEL);
767 if (body == NULL) {
768 kfree(msg);
769 err = -ENOMEM;
770 goto out;
771 }
772
773 err = xb_read(body, msg->hdr.len);
774 if (err) {
775 kfree(body);
776 kfree(msg);
777 goto out;
778 }
779 body[msg->hdr.len] = '\0';
780
781 if (msg->hdr.type == XS_WATCH_EVENT) {
782 msg->u.watch.vec = split(body, msg->hdr.len,
783 &msg->u.watch.vec_size);
784 if (IS_ERR(msg->u.watch.vec)) {
785 kfree(msg);
786 err = PTR_ERR(msg->u.watch.vec);
787 goto out;
788 }
789
790 spin_lock(&watches_lock);
791 msg->u.watch.handle = find_watch(
792 msg->u.watch.vec[XS_WATCH_TOKEN]);
793 if (msg->u.watch.handle != NULL) {
794 spin_lock(&watch_events_lock);
795 list_add_tail(&msg->list, &watch_events);
796 wake_up(&watch_events_waitq);
797 spin_unlock(&watch_events_lock);
798 } else {
799 kfree(msg->u.watch.vec);
800 kfree(msg);
801 }
802 spin_unlock(&watches_lock);
803 } else {
804 msg->u.reply.body = body;
805 spin_lock(&xs_state.reply_lock);
806 list_add_tail(&msg->list, &xs_state.reply_list);
807 spin_unlock(&xs_state.reply_lock);
808 wake_up(&xs_state.reply_waitq);
809 }
810
811 out:
812 mutex_unlock(&xs_state.response_mutex);
813 return err;
814}
815
816static int xenbus_thread(void *unused)
817{
818 int err;
819
820 for (;;) {
821 err = process_msg();
822 if (err)
823 printk(KERN_WARNING "XENBUS error %d while reading "
824 "message\n", err);
825 if (kthread_should_stop())
826 break;
827 }
828
829 return 0;
830}
831
832int xs_init(void)
833{
834 int err;
835 struct task_struct *task;
836
837 INIT_LIST_HEAD(&xs_state.reply_list);
838 spin_lock_init(&xs_state.reply_lock);
839 init_waitqueue_head(&xs_state.reply_waitq);
840
841 mutex_init(&xs_state.request_mutex);
842 mutex_init(&xs_state.response_mutex);
843 init_rwsem(&xs_state.transaction_mutex);
844 init_rwsem(&xs_state.watch_mutex);
845
846 /* Initialize the shared memory rings to talk to xenstored */
847 err = xb_init_comms();
848 if (err)
849 return err;
850
851 task = kthread_run(xenwatch_thread, NULL, "xenwatch");
852 if (IS_ERR(task))
853 return PTR_ERR(task);
854 xenwatch_pid = task->pid;
855
856 task = kthread_run(xenbus_thread, NULL, "xenbus");
857 if (IS_ERR(task))
858 return PTR_ERR(task);
859
860 return 0;
861}