summaryrefslogtreecommitdiffstats
path: root/drivers/block/viodasd.c
blob: 709f809f79f1f9cc4b3e09cb9ce3e8b71b913f32 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
/* -*- linux-c -*-
 * viodasd.c
 *  Authors: Dave Boutcher <boutcher@us.ibm.com>
 *           Ryan Arnold <ryanarn@us.ibm.com>
 *           Colin Devilbiss <devilbis@us.ibm.com>
 *           Stephen Rothwell <sfr@au1.ibm.com>
 *
 * (C) Copyright 2000-2004 IBM Corporation
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License as
 * published by the Free Software Foundation; either version 2 of the
 * License, or (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
 *
 * This routine provides access to disk space (termed "DASD" in historical
 * IBM terms) owned and managed by an OS/400 partition running on the
 * same box as this Linux partition.
 *
 * All disk operations are performed by sending messages back and forth to
 * the OS/400 partition.
 */
#include <linux/major.h>
#include <linux/fs.h>
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/blkdev.h>
#include <linux/genhd.h>
#include <linux/hdreg.h>
#include <linux/errno.h>
#include <linux/init.h>
#include <linux/string.h>
#include <linux/dma-mapping.h>
#include <linux/completion.h>
#include <linux/device.h>
#include <linux/kernel.h>

#include <asm/uaccess.h>
#include <asm/vio.h>
#include <asm/iSeries/HvTypes.h>
#include <asm/iSeries/HvLpEvent.h>
#include <asm/iSeries/HvLpConfig.h>
#include <asm/iSeries/vio.h>

MODULE_DESCRIPTION("iSeries Virtual DASD");
MODULE_AUTHOR("Dave Boutcher");
MODULE_LICENSE("GPL");

/*
 * We only support 7 partitions per physical disk....so with minor
 * numbers 0-255 we get a maximum of 32 disks.
 */
#define VIOD_GENHD_NAME		"iseries/vd"
#define VIOD_GENHD_DEVFS_NAME	"iseries/disc"

#define VIOD_VERS		"1.64"

#define VIOD_KERN_WARNING	KERN_WARNING "viod: "
#define VIOD_KERN_INFO		KERN_INFO "viod: "

enum {
	PARTITION_SHIFT = 3,
	MAX_DISKNO = HVMAXARCHITECTEDVIRTUALDISKS,
	MAX_DISK_NAME = sizeof(((struct gendisk *)0)->disk_name)
};

static DEFINE_SPINLOCK(viodasd_spinlock);

#define VIOMAXREQ		16
#define VIOMAXBLOCKDMA		12

#define DEVICE_NO(cell)	((struct viodasd_device *)(cell) - &viodasd_devices[0])

struct open_data {
	u64	disk_size;
	u16	max_disk;
	u16	cylinders;
	u16	tracks;
	u16	sectors;
	u16	bytes_per_sector;
};

struct rw_data {
	u64	offset;
	struct {
		u32	token;
		u32	reserved;
		u64	len;
	} dma_info[VIOMAXBLOCKDMA];
};

struct vioblocklpevent {
	struct HvLpEvent	event;
	u32			reserved;
	u16			version;
	u16			sub_result;
	u16			disk;
	u16			flags;
	union {
		struct open_data	open_data;
		struct rw_data		rw_data;
		u64			changed;
	} u;
};

#define vioblockflags_ro   0x0001

enum vioblocksubtype {
	vioblockopen = 0x0001,
	vioblockclose = 0x0002,
	vioblockread = 0x0003,
	vioblockwrite = 0x0004,
	vioblockflush = 0x0005,
	vioblockcheck = 0x0007
};

struct viodasd_waitevent {
	struct completion	com;
	int			rc;
	u16			sub_result;
	int			max_disk;	/* open */
};

static const struct vio_error_entry viodasd_err_table[] = {
	{ 0x0201, EINVAL, "Invalid Range" },
	{ 0x0202, EINVAL, "Invalid Token" },
	{ 0x0203, EIO, "DMA Error" },
	{ 0x0204, EIO, "Use Error" },
	{ 0x0205, EIO, "Release Error" },
	{ 0x0206, EINVAL, "Invalid Disk" },
	{ 0x0207, EBUSY, "Cant Lock" },
	{ 0x0208, EIO, "Already Locked" },
	{ 0x0209, EIO, "Already Unlocked" },
	{ 0x020A, EIO, "Invalid Arg" },
	{ 0x020B, EIO, "Bad IFS File" },
	{ 0x020C, EROFS, "Read Only Device" },
	{ 0x02FF, EIO, "Internal Error" },
	{ 0x0000, 0, NULL },
};

/*
 * Figure out the biggest I/O request (in sectors) we can accept
 */
#define VIODASD_MAXSECTORS (4096 / 512 * VIOMAXBLOCKDMA)

/*
 * Number of disk I/O requests we've sent to OS/400
 */
static int num_req_outstanding;

/*
 * This is our internal structure for keeping track of disk devices
 */
struct viodasd_device {
	u16		cylinders;
	u16		tracks;
	u16		sectors;
	u16		bytes_per_sector;
	u64		size;
	int		read_only;
	spinlock_t	q_lock;
	struct gendisk	*disk;
	struct device	*dev;
} viodasd_devices[MAX_DISKNO];

/*
 * External open entry point.
 */
static int viodasd_open(struct inode *ino, struct file *fil)
{
	struct viodasd_device *d = ino->i_bdev->bd_disk->private_data;
	HvLpEvent_Rc hvrc;
	struct viodasd_waitevent we;
	u16 flags = 0;

	if (d->read_only) {
		if ((fil != NULL) && (fil->f_mode & FMODE_WRITE))
			return -EROFS;
		flags = vioblockflags_ro;
	}

	init_completion(&we.com);

	/* Send the open event to OS/400 */
	hvrc = HvCallEvent_signalLpEventFast(viopath_hostLp,
			HvLpEvent_Type_VirtualIo,
			viomajorsubtype_blockio | vioblockopen,
			HvLpEvent_AckInd_DoAck, HvLpEvent_AckType_ImmediateAck,
			viopath_sourceinst(viopath_hostLp),
			viopath_targetinst(viopath_hostLp),
			(u64)(unsigned long)&we, VIOVERSION << 16,
			((u64)DEVICE_NO(d) << 48) | ((u64)flags << 32),
			0, 0, 0);
	if (hvrc != 0) {
		printk(VIOD_KERN_WARNING "HV open failed %d\n", (int)hvrc);
		return -EIO;
	}

	wait_for_completion(&we.com);

	/* Check the return code */
	if (we.rc != 0) {
		const struct vio_error_entry *err =
			vio_lookup_rc(viodasd_err_table, we.sub_result);

		printk(VIOD_KERN_WARNING
				"bad rc opening disk: %d:0x%04x (%s)\n",
				(int)we.rc, we.sub_result, err->msg);
		return -EIO;
	}

	return 0;
}

/*
 * External release entry point.
 */
static int viodasd_release(struct inode *ino, struct file *fil)
{
	struct viodasd_device *d = ino->i_bdev->bd_disk->private_data;
	HvLpEvent_Rc hvrc;

	/* Send the event to OS/400.  We DON'T expect a response */
	hvrc = HvCallEvent_signalLpEventFast(viopath_hostLp,
			HvLpEvent_Type_VirtualIo,
			viomajorsubtype_blockio | vioblockclose,
			HvLpEvent_AckInd_NoAck, HvLpEvent_AckType_ImmediateAck,
			viopath_sourceinst(viopath_hostLp),
			viopath_targetinst(viopath_hostLp),
			0, VIOVERSION << 16,
			((u64)DEVICE_NO(d) << 48) /* | ((u64)flags << 32) */,
			0, 0, 0);
	if (hvrc != 0)
		printk(VIOD_KERN_WARNING "HV close call failed %d\n",
				(int)hvrc);
	return 0;
}


/* External ioctl entry point.
 */
static int viodasd_ioctl(struct inode *ino, struct file *fil,
			 unsigned int cmd, unsigned long arg)
{
	unsigned char sectors;
	unsigned char heads;
	unsigned short cylinders;
	struct hd_geometry *geo;
	struct gendisk *gendisk;
	struct viodasd_device *d;

	switch (cmd) {
	case HDIO_GETGEO:
		geo = (struct hd_geometry *)arg;
		if (geo == NULL)
			return -EINVAL;
		if (!access_ok(VERIFY_WRITE, geo, sizeof(*geo)))
			return -EFAULT;
		gendisk = ino->i_bdev->bd_disk;
		d = gendisk->private_data;
		sectors = d->sectors;
		if (sectors == 0)
			sectors = 32;
		heads = d->tracks;
		if (heads == 0)
			heads = 64;
		cylinders = d->cylinders;
		if (cylinders == 0)
			cylinders = get_capacity(gendisk) / (sectors * heads);
		if (__put_user(sectors, &geo->sectors) ||
		    __put_user(heads, &geo->heads) ||
		    __put_user(cylinders, &geo->cylinders) ||
		    __put_user(get_start_sect(ino->i_bdev), &geo->start))
			return -EFAULT;
		return 0;
	}

	return -EINVAL;
}

/*
 * Our file operations table
 */
static struct block_device_operations viodasd_fops = {
	.owner = THIS_MODULE,
	.open = viodasd_open,
	.release = viodasd_release,
	.ioctl = viodasd_ioctl,
};

/*
 * End a request
 */
static void viodasd_end_request(struct request *req, int uptodate,
		int num_sectors)
{
	if (end_that_request_first(req, uptodate, num_sectors))
		return;
	add_disk_randomness(req->rq_disk);
	end_that_request_last(req);
}

/*
 * Send an actual I/O request to OS/400
 */
static int send_request(struct request *req)
{
	u64 start;
	int direction;
	int nsg;
	u16 viocmd;
	HvLpEvent_Rc hvrc;
	struct vioblocklpevent *bevent;
	struct scatterlist sg[VIOMAXBLOCKDMA];
	int sgindex;
	int statindex;
	struct viodasd_device *d;
	unsigned long flags;

	start = (u64)req->sector << 9;

	if (rq_data_dir(req) == READ) {
		direction = DMA_FROM_DEVICE;
		viocmd = viomajorsubtype_blockio | vioblockread;
		statindex = 0;
	} else {
		direction = DMA_TO_DEVICE;
		viocmd = viomajorsubtype_blockio | vioblockwrite;
		statindex = 1;
	}

        d = req->rq_disk->private_data;

	/* Now build the scatter-gather list */
	nsg = blk_rq_map_sg(req->q, req, sg);
	nsg = dma_map_sg(d->dev, sg, nsg, direction);

	spin_lock_irqsave(&viodasd_spinlock, flags);
	num_req_outstanding++;

	/* This optimization handles a single DMA block */
	if (nsg == 1)
		hvrc = HvCallEvent_signalLpEventFast(viopath_hostLp,
				HvLpEvent_Type_VirtualIo, viocmd,
				HvLpEvent_AckInd_DoAck,
				HvLpEvent_AckType_ImmediateAck,
				viopath_sourceinst(viopath_hostLp),
				viopath_targetinst(viopath_hostLp),
				(u64)(unsigned long)req, VIOVERSION << 16,
				((u64)DEVICE_NO(d) << 48), start,
				((u64)sg_dma_address(&sg[0])) << 32,
				sg_dma_len(&sg[0]));
	else {
		bevent = (struct vioblocklpevent *)
			vio_get_event_buffer(viomajorsubtype_blockio);
		if (bevent == NULL) {
			printk(VIOD_KERN_WARNING
			       "error allocating disk event buffer\n");
			goto error_ret;
		}

		/*
		 * Now build up the actual request.  Note that we store
		 * the pointer to the request in the correlation
		 * token so we can match the response up later
		 */
		memset(bevent, 0, sizeof(struct vioblocklpevent));
		bevent->event.xFlags.xValid = 1;
		bevent->event.xFlags.xFunction = HvLpEvent_Function_Int;
		bevent->event.xFlags.xAckInd = HvLpEvent_AckInd_DoAck;
		bevent->event.xFlags.xAckType = HvLpEvent_AckType_ImmediateAck;
		bevent->event.xType = HvLpEvent_Type_VirtualIo;
		bevent->event.xSubtype = viocmd;
		bevent->event.xSourceLp = HvLpConfig_getLpIndex();
		bevent->event.xTargetLp = viopath_hostLp;
		bevent->event.xSizeMinus1 =
			offsetof(struct vioblocklpevent, u.rw_data.dma_info) +
			(sizeof(bevent->u.rw_data.dma_info[0]) * nsg) - 1;
		bevent->event.xSourceInstanceId =
			viopath_sourceinst(viopath_hostLp);
		bevent->event.xTargetInstanceId =
			viopath_targetinst(viopath_hostLp);
		bevent->event.xCorrelationToken = (u64)req;
		bevent->version = VIOVERSION;
		bevent->disk = DEVICE_NO(d);
		bevent->u.rw_data.offset = start;

		/*
		 * Copy just the dma information from the sg list
		 * into the request
		 */
		for (sgindex = 0; sgindex < nsg; sgindex++) {
			bevent->u.rw_data.dma_info[sgindex].token =
				sg_dma_address(&sg[sgindex]);
			bevent->u.rw_data.dma_info[sgindex].len =
				sg_dma_len(&sg[sgindex]);
		}

		/* Send the request */
		hvrc = HvCallEvent_signalLpEvent(&bevent->event);
		vio_free_event_buffer(viomajorsubtype_blockio, bevent);
	}

	if (hvrc != HvLpEvent_Rc_Good) {
		printk(VIOD_KERN_WARNING
		       "error sending disk event to OS/400 (rc %d)\n",
		       (int)hvrc);
		goto error_ret;
	}
	spin_unlock_irqrestore(&viodasd_spinlock, flags);
	return 0;

error_ret:
	num_req_outstanding--;
	spin_unlock_irqrestore(&viodasd_spinlock, flags);
	dma_unmap_sg(d->dev, sg, nsg, direction);
	return -1;
}

/*
 * This is the external request processing routine
 */
static void do_viodasd_request(request_queue_t *q)
{
	struct request *req;

	/*
	 * If we already have the maximum number of requests
	 * outstanding to OS/400 just bail out. We'll come
	 * back later.
	 */
	while (num_req_outstanding < VIOMAXREQ) {
		req = elv_next_request(q);
		if (req == NULL)
			return;
		/* dequeue the current request from the queue */
		blkdev_dequeue_request(req);
		/* check that request contains a valid command */
		if (!blk_fs_request(req)) {
			viodasd_end_request(req, 0, req->hard_nr_sectors);
			continue;
		}
		/* Try sending the request */
		if (send_request(req) != 0)
			viodasd_end_request(req, 0, req->hard_nr_sectors);
	}
}

/*
 * Probe a single disk and fill in the viodasd_device structure
 * for it.
 */
static void probe_disk(struct viodasd_device *d)
{
	HvLpEvent_Rc hvrc;
	struct viodasd_waitevent we;
	int dev_no = DEVICE_NO(d);
	struct gendisk *g;
	struct request_queue *q;
	u16 flags = 0;

retry:
	init_completion(&we.com);

	/* Send the open event to OS/400 */
	hvrc = HvCallEvent_signalLpEventFast(viopath_hostLp,
			HvLpEvent_Type_VirtualIo,
			viomajorsubtype_blockio | vioblockopen,
			HvLpEvent_AckInd_DoAck, HvLpEvent_AckType_ImmediateAck,
			viopath_sourceinst(viopath_hostLp),
			viopath_targetinst(viopath_hostLp),
			(u64)(unsigned long)&we, VIOVERSION << 16,
			((u64)dev_no << 48) | ((u64)flags<< 32),
			0, 0, 0);
	if (hvrc != 0) {
		printk(VIOD_KERN_WARNING "bad rc on HV open %d\n", (int)hvrc);
		return;
	}

	wait_for_completion(&we.com);

	if (we.rc != 0) {
		if (flags != 0)
			return;
		/* try again with read only flag set */
		flags = vioblockflags_ro;
		goto retry;
	}
	if (we.max_disk > (MAX_DISKNO - 1)) {
		static int warned;

		if (warned == 0) {
			warned++;
			printk(VIOD_KERN_INFO
				"Only examining the first %d "
				"of %d disks connected\n",
				MAX_DISKNO, we.max_disk + 1);
		}
	}

	/* Send the close event to OS/400.  We DON'T expect a response */
	hvrc = HvCallEvent_signalLpEventFast(viopath_hostLp,
			HvLpEvent_Type_VirtualIo,
			viomajorsubtype_blockio | vioblockclose,
			HvLpEvent_AckInd_NoAck, HvLpEvent_AckType_ImmediateAck,
			viopath_sourceinst(viopath_hostLp),
			viopath_targetinst(viopath_hostLp),
			0, VIOVERSION << 16,
			((u64)dev_no << 48) | ((u64)flags << 32),
			0, 0, 0);
	if (hvrc != 0) {
		printk(VIOD_KERN_WARNING
		       "bad rc sending event to OS/400 %d\n", (int)hvrc);
		return;
	}
	/* create the request queue for the disk */
	spin_lock_init(&d->q_lock);
	q = blk_init_queue(do_viodasd_request, &d->q_lock);
	if (q == NULL) {
		printk(VIOD_KERN_WARNING "cannot allocate queue for disk %d\n",
				dev_no);
		return;
	}
	g = alloc_disk(1 << PARTITION_SHIFT);
	if (g == NULL) {
		printk(VIOD_KERN_WARNING
				"cannot allocate disk structure for disk %d\n",
				dev_no);
		blk_cleanup_queue(q);
		return;
	}

	d->disk = g;
	blk_queue_max_hw_segments(q, VIOMAXBLOCKDMA);
	blk_queue_max_phys_segments(q, VIOMAXBLOCKDMA);
	blk_queue_max_sectors(q, VIODASD_MAXSECTORS);
	g->major = VIODASD_MAJOR;
	g->first_minor = dev_no << PARTITION_SHIFT;
	if (dev_no >= 26)
		snprintf(g->disk_name, sizeof(g->disk_name),
				VIOD_GENHD_NAME "%c%c",
				'a' + (dev_no / 26) - 1, 'a' + (dev_no % 26));
	else
		snprintf(g->disk_name, sizeof(g->disk_name),
				VIOD_GENHD_NAME "%c", 'a' + (dev_no % 26));
	snprintf(g->devfs_name, sizeof(g->devfs_name),
			"%s%d", VIOD_GENHD_DEVFS_NAME, dev_no);
	g->fops = &viodasd_fops;
	g->queue = q;
	g->private_data = d;
	g->driverfs_dev = d->dev;
	set_capacity(g, d->size >> 9);

	printk(VIOD_KERN_INFO "disk %d: %lu sectors (%lu MB) "
			"CHS=%d/%d/%d sector size %d%s\n",
			dev_no, (unsigned long)(d->size >> 9),
			(unsigned long)(d->size >> 20),
			(int)d->cylinders, (int)d->tracks,
			(int)d->sectors, (int)d->bytes_per_sector,
			d->read_only ? " (RO)" : "");

	/* register us in the global list */
	add_disk(g);
}

/* returns the total number of scatterlist elements converted */
static int block_event_to_scatterlist(const struct vioblocklpevent *bevent,
		struct scatterlist *sg, int *total_len)
{
	int i, numsg;
	const struct rw_data *rw_data = &bevent->u.rw_data;
	static const int offset =
		offsetof(struct vioblocklpevent, u.rw_data.dma_info);
	static const int element_size = sizeof(rw_data->dma_info[0]);

	numsg = ((bevent->event.xSizeMinus1 + 1) - offset) / element_size;
	if (numsg > VIOMAXBLOCKDMA)
		numsg = VIOMAXBLOCKDMA;

	*total_len = 0;
	memset(sg, 0, sizeof(sg[0]) * VIOMAXBLOCKDMA);

	for (i = 0; (i < numsg) && (rw_data->dma_info[i].len > 0); ++i) {
		sg_dma_address(&sg[i]) = rw_data->dma_info[i].token;
		sg_dma_len(&sg[i]) = rw_data->dma_info[i].len;
		*total_len += rw_data->dma_info[i].len;
	}
	return i;
}

/*
 * Restart all queues, starting with the one _after_ the disk given,
 * thus reducing the chance of starvation of higher numbered disks.
 */
static void viodasd_restart_all_queues_starting_from(int first_index)
{
	int i;

	for (i = first_index + 1; i < MAX_DISKNO; ++i)
		if (viodasd_devices[i].disk)
			blk_run_queue(viodasd_devices[i].disk->queue);
	for (i = 0; i <= first_index; ++i)
		if (viodasd_devices[i].disk)
			blk_run_queue(viodasd_devices[i].disk->queue);
}

/*
 * For read and write requests, decrement the number of outstanding requests,
 * Free the DMA buffers we allocated.
 */
static int viodasd_handle_read_write(struct vioblocklpevent *bevent)
{
	int num_sg, num_sect, pci_direction, total_len;
	struct request *req;
	struct scatterlist sg[VIOMAXBLOCKDMA];
	struct HvLpEvent *event = &bevent->event;
	unsigned long irq_flags;
	struct viodasd_device *d;
	int error;
	spinlock_t *qlock;

	num_sg = block_event_to_scatterlist(bevent, sg, &total_len);
	num_sect = total_len >> 9;
	if (event->xSubtype == (viomajorsubtype_blockio | vioblockread))
		pci_direction = DMA_FROM_DEVICE;
	else
		pci_direction = DMA_TO_DEVICE;
	req = (struct request *)bevent->event.xCorrelationToken;
	d = req->rq_disk->private_data;

	dma_unmap_sg(d->dev, sg, num_sg, pci_direction);

	/*
	 * Since this is running in interrupt mode, we need to make sure
	 * we're not stepping on any global I/O operations
	 */
	spin_lock_irqsave(&viodasd_spinlock, irq_flags);
	num_req_outstanding--;
	spin_unlock_irqrestore(&viodasd_spinlock, irq_flags);

	error = event->xRc != HvLpEvent_Rc_Good;
	if (error) {
		const struct vio_error_entry *err;
		err = vio_lookup_rc(viodasd_err_table, bevent->sub_result);
		printk(VIOD_KERN_WARNING "read/write error %d:0x%04x (%s)\n",
				event->xRc, bevent->sub_result, err->msg);
		num_sect = req->hard_nr_sectors;
	}
	qlock = req->q->queue_lock;
	spin_lock_irqsave(qlock, irq_flags);
	viodasd_end_request(req, !error, num_sect);
	spin_unlock_irqrestore(qlock, irq_flags);

	/* Finally, try to get more requests off of this device's queue */
	viodasd_restart_all_queues_starting_from(DEVICE_NO(d));

	return 0;
}

/* This routine handles incoming block LP events */
static void handle_block_event(struct HvLpEvent *event)
{
	struct vioblocklpevent *bevent = (struct vioblocklpevent *)event;
	struct viodasd_waitevent *pwe;

	if (event == NULL)
		/* Notification that a partition went away! */
		return;
	/* First, we should NEVER get an int here...only acks */
	if (event->xFlags.xFunction == HvLpEvent_Function_Int) {
		printk(VIOD_KERN_WARNING
		       "Yikes! got an int in viodasd event handler!\n");
		if (event->xFlags.xAckInd == HvLpEvent_AckInd_DoAck) {
			event->xRc = HvLpEvent_Rc_InvalidSubtype;
			HvCallEvent_ackLpEvent(event);
		}
	}

	switch (event->xSubtype & VIOMINOR_SUBTYPE_MASK) {
	case vioblockopen:
		/*
		 * Handle a response to an open request.  We get all the
		 * disk information in the response, so update it.  The
		 * correlation token contains a pointer to a waitevent
		 * structure that has a completion in it.  update the
		 * return code in the waitevent structure and post the
		 * completion to wake up the guy who sent the request
		 */
		pwe = (struct viodasd_waitevent *)event->xCorrelationToken;
		pwe->rc = event->xRc;
		pwe->sub_result = bevent->sub_result;
		if (event->xRc == HvLpEvent_Rc_Good) {
			const struct open_data *data = &bevent->u.open_data;
			struct viodasd_device *device =
				&viodasd_devices[bevent->disk];
			device->read_only =
				bevent->flags & vioblockflags_ro;
			device->size = data->disk_size;
			device->cylinders = data->cylinders;
			device->tracks = data->tracks;
			device->sectors = data->sectors;
			device->bytes_per_sector = data->bytes_per_sector;
			pwe->max_disk = data->max_disk;
		}
		complete(&pwe->com);
		break;
	case vioblockclose:
		break;
	case vioblockread:
	case vioblockwrite:
		viodasd_handle_read_write(bevent);
		break;

	default:
		printk(VIOD_KERN_WARNING "invalid subtype!");
		if (event->xFlags.xAckInd == HvLpEvent_AckInd_DoAck) {
			event->xRc = HvLpEvent_Rc_InvalidSubtype;
			HvCallEvent_ackLpEvent(event);
		}
	}
}

/*
 * Get the driver to reprobe for more disks.
 */
static ssize_t probe_disks(struct device_driver *drv, const char *buf,
		size_t count)
{
	struct viodasd_device *d;

	for (d = viodasd_devices; d < &viodasd_devices[MAX_DISKNO]; d++) {
		if (d->disk == NULL)
			probe_disk(d);
	}
	return count;
}
static DRIVER_ATTR(probe, S_IWUSR, NULL, probe_disks);

static int viodasd_probe(struct vio_dev *vdev, const struct vio_device_id *id)
{
	struct viodasd_device *d = &viodasd_devices[vdev->unit_address];

	d->dev = &vdev->dev;
	probe_disk(d);
	if (d->disk == NULL)
		return -ENODEV;
	return 0;
}

static int viodasd_remove(struct vio_dev *vdev)
{
	struct viodasd_device *d;

	d = &viodasd_devices[vdev->unit_address];
	if (d->disk) {
		del_gendisk(d->disk);
		blk_cleanup_queue(d->disk->queue);
		put_disk(d->disk);
		d->disk = NULL;
	}
	d->dev = NULL;
	return 0;
}

/**
 * viodasd_device_table: Used by vio.c to match devices that we
 * support.
 */
static struct vio_device_id viodasd_device_table[] __devinitdata = {
	{ "viodasd", "" },
	{ "", "" }
};
MODULE_DEVICE_TABLE(vio, viodasd_device_table);

static struct vio_driver viodasd_driver = {
	.id_table = viodasd_device_table,
	.probe = viodasd_probe,
	.remove = viodasd_remove,
	.driver = {
		.name = "viodasd",
		.owner = THIS_MODULE,
	}
};

/*
 * Initialize the whole device driver.  Handle module and non-module
 * versions
 */
static int __init viodasd_init(void)
{
	int rc;

	/* Try to open to our host lp */
	if (viopath_hostLp == HvLpIndexInvalid)
		vio_set_hostlp();

	if (viopath_hostLp == HvLpIndexInvalid) {
		printk(VIOD_KERN_WARNING "invalid hosting partition\n");
		return -EIO;
	}

	printk(VIOD_KERN_INFO "vers " VIOD_VERS ", hosting partition %d\n",
			viopath_hostLp);

        /* register the block device */
	if (register_blkdev(VIODASD_MAJOR, VIOD_GENHD_NAME)) {
		printk(VIOD_KERN_WARNING
				"Unable to get major number %d for %s\n",
				VIODASD_MAJOR, VIOD_GENHD_NAME);
		return -EIO;
	}
	/* Actually open the path to the hosting partition */
	if (viopath_open(viopath_hostLp, viomajorsubtype_blockio,
				VIOMAXREQ + 2)) {
		printk(VIOD_KERN_WARNING
		       "error opening path to host partition %d\n",
		       viopath_hostLp);
		unregister_blkdev(VIODASD_MAJOR, VIOD_GENHD_NAME);
		return -EIO;
	}

	/* Initialize our request handler */
	vio_setHandler(viomajorsubtype_blockio, handle_block_event);

	rc = vio_register_driver(&viodasd_driver);
	if (rc == 0)
		driver_create_file(&viodasd_driver.driver, &driver_attr_probe);
	return rc;
}
module_init(viodasd_init);

void viodasd_exit(void)
{
	driver_remove_file(&viodasd_driver.driver, &driver_attr_probe);
	vio_unregister_driver(&viodasd_driver);
	vio_clearHandler(viomajorsubtype_blockio);
	unregister_blkdev(VIODASD_MAJOR, VIOD_GENHD_NAME);
	viopath_close(viopath_hostLp, viomajorsubtype_blockio, VIOMAXREQ + 2);
}

module_exit(viodasd_exit);