summaryrefslogtreecommitdiffstats
path: root/drivers/infiniband/hw/ehca/ipz_pt_fn.c
blob: a090c679c39787f70a181d674bab1940c2f7fe87 (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
/*
 *  IBM eServer eHCA Infiniband device driver for Linux on POWER
 *
 *  internal queue handling
 *
 *  Authors: Waleri Fomin <fomin@de.ibm.com>
 *           Reinhard Ernst <rernst@de.ibm.com>
 *           Christoph Raisch <raisch@de.ibm.com>
 *
 *  Copyright (c) 2005 IBM Corporation
 *
 *  This source code is distributed under a dual license of GPL v2.0 and OpenIB
 *  BSD.
 *
 * OpenIB BSD License
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions are met:
 *
 * Redistributions of source code must retain the above copyright notice, this
 * list of conditions and the following disclaimer.
 *
 * Redistributions in binary form must reproduce the above copyright notice,
 * this list of conditions and the following disclaimer in the documentation
 * and/or other materials
 * provided with the distribution.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
 * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
 * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
 * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER
 * IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
 * POSSIBILITY OF SUCH DAMAGE.
 */

#include "ehca_tools.h"
#include "ipz_pt_fn.h"
#include "ehca_classes.h"

#define PAGES_PER_KPAGE (PAGE_SIZE >> EHCA_PAGESHIFT)

struct kmem_cache *small_qp_cache;

void *ipz_qpageit_get_inc(struct ipz_queue *queue)
{
	void *ret = ipz_qeit_get(queue);
	queue->current_q_offset += queue->pagesize;
	if (queue->current_q_offset > queue->queue_length) {
		queue->current_q_offset -= queue->pagesize;
		ret = NULL;
	}
	if (((u64)ret) % queue->pagesize) {
		ehca_gen_err("ERROR!! not at PAGE-Boundary");
		return NULL;
	}
	return ret;
}

void *ipz_qeit_eq_get_inc(struct ipz_queue *queue)
{
	void *ret = ipz_qeit_get(queue);
	u64 last_entry_in_q = queue->queue_length - queue->qe_size;

	queue->current_q_offset += queue->qe_size;
	if (queue->current_q_offset > last_entry_in_q) {
		queue->current_q_offset = 0;
		queue->toggle_state = (~queue->toggle_state) & 1;
	}

	return ret;
}

int ipz_queue_abs_to_offset(struct ipz_queue *queue, u64 addr, u64 *q_offset)
{
	int i;
	for (i = 0; i < queue->queue_length / queue->pagesize; i++) {
		u64 page = (u64)virt_to_abs(queue->queue_pages[i]);
		if (addr >= page && addr < page + queue->pagesize) {
			*q_offset = addr - page + i * queue->pagesize;
			return 0;
		}
	}
	return -EINVAL;
}

#if PAGE_SHIFT < EHCA_PAGESHIFT
#error Kernel pages must be at least as large than eHCA pages (4K) !
#endif

/*
 * allocate pages for queue:
 * outer loop allocates whole kernel pages (page aligned) and
 * inner loop divides a kernel page into smaller hca queue pages
 */
static int alloc_queue_pages(struct ipz_queue *queue, const u32 nr_of_pages)
{
	int k, f = 0;
	u8 *kpage;

	while (f < nr_of_pages) {
		kpage = (u8 *)get_zeroed_page(GFP_KERNEL);
		if (!kpage)
			goto out;

		for (k = 0; k < PAGES_PER_KPAGE && f < nr_of_pages; k++) {
			queue->queue_pages[f] = (struct ipz_page *)kpage;
			kpage += EHCA_PAGESIZE;
			f++;
		}
	}
	return 1;

out:
	for (f = 0; f < nr_of_pages && queue->queue_pages[f];
	     f += PAGES_PER_KPAGE)
		free_page((unsigned long)(queue->queue_pages)[f]);
	return 0;
}

static int alloc_small_queue_page(struct ipz_queue *queue, struct ehca_pd *pd)
{
	int order = ilog2(queue->pagesize) - 9;
	struct ipz_small_queue_page *page;
	unsigned long bit;

	mutex_lock(&pd->lock);

	if (!list_empty(&pd->free[order]))
		page = list_entry(pd->free[order].next,
				  struct ipz_small_queue_page, list);
	else {
		page = kmem_cache_zalloc(small_qp_cache, GFP_KERNEL);
		if (!page)
			goto out;

		page->page = get_zeroed_page(GFP_KERNEL);
		if (!page->page) {
			kmem_cache_free(small_qp_cache, page);
			goto out;
		}

		list_add(&page->list, &pd->free[order]);
	}

	bit = find_first_zero_bit(page->bitmap, IPZ_SPAGE_PER_KPAGE >> order);
	__set_bit(bit, page->bitmap);
	page->fill++;

	if (page->fill == IPZ_SPAGE_PER_KPAGE >> order)
		list_move(&page->list, &pd->full[order]);

	mutex_unlock(&pd->lock);

	queue->queue_pages[0] = (void *)(page->page | (bit << (order + 9)));
	queue->small_page = page;
	return 1;

out:
	ehca_err(pd->ib_pd.device, "failed to allocate small queue page");
	return 0;
}

static void free_small_queue_page(struct ipz_queue *queue, struct ehca_pd *pd)
{
	int order = ilog2(queue->pagesize) - 9;
	struct ipz_small_queue_page *page = queue->small_page;
	unsigned long bit;
	int free_page = 0;

	bit = ((unsigned long)queue->queue_pages[0] & PAGE_MASK)
		>> (order + 9);

	mutex_lock(&pd->lock);

	__clear_bit(bit, page->bitmap);
	page->fill--;

	if (page->fill == 0) {
		list_del(&page->list);
		free_page = 1;
	}

	if (page->fill == (IPZ_SPAGE_PER_KPAGE >> order) - 1)
		/* the page was full until we freed the chunk */
		list_move_tail(&page->list, &pd->free[order]);

	mutex_unlock(&pd->lock);

	if (free_page) {
		free_page(page->page);
		kmem_cache_free(small_qp_cache, page);
	}
}

int ipz_queue_ctor(struct ehca_pd *pd, struct ipz_queue *queue,
		   const u32 nr_of_pages, const u32 pagesize,
		   const u32 qe_size, const u32 nr_of_sg,
		   int is_small)
{
	if (pagesize > PAGE_SIZE) {
		ehca_gen_err("FATAL ERROR: pagesize=%x "
			     "is greater than kernel page size", pagesize);
		return 0;
	}

	/* init queue fields */
	queue->queue_length = nr_of_pages * pagesize;
	queue->pagesize = pagesize;
	queue->qe_size = qe_size;
	queue->act_nr_of_sg = nr_of_sg;
	queue->current_q_offset = 0;
	queue->toggle_state = 1;
	queue->small_page = NULL;

	/* allocate queue page pointers */
	queue->queue_pages = vmalloc(nr_of_pages * sizeof(void *));
	if (!queue->queue_pages) {
		ehca_gen_err("Couldn't allocate queue page list");
		return 0;
	}
	memset(queue->queue_pages, 0, nr_of_pages * sizeof(void *));

	/* allocate actual queue pages */
	if (is_small) {
		if (!alloc_small_queue_page(queue, pd))
			goto ipz_queue_ctor_exit0;
	} else
		if (!alloc_queue_pages(queue, nr_of_pages))
			goto ipz_queue_ctor_exit0;

	return 1;

ipz_queue_ctor_exit0:
	ehca_gen_err("Couldn't alloc pages queue=%p "
		 "nr_of_pages=%x",  queue, nr_of_pages);
	vfree(queue->queue_pages);

	return 0;
}

int ipz_queue_dtor(struct ehca_pd *pd, struct ipz_queue *queue)
{
	int i, nr_pages;

	if (!queue || !queue->queue_pages) {
		ehca_gen_dbg("queue or queue_pages is NULL");
		return 0;
	}

	if (queue->small_page)
		free_small_queue_page(queue, pd);
	else {
		nr_pages = queue->queue_length / queue->pagesize;
		for (i = 0; i < nr_pages; i += PAGES_PER_KPAGE)
			free_page((unsigned long)queue->queue_pages[i]);
	}

	vfree(queue->queue_pages);

	return 1;
}

int ehca_init_small_qp_cache(void)
{
	small_qp_cache = kmem_cache_create("ehca_cache_small_qp",
					   sizeof(struct ipz_small_queue_page),
					   0, SLAB_HWCACHE_ALIGN, NULL);
	if (!small_qp_cache)
		return -ENOMEM;

	return 0;
}

void ehca_cleanup_small_qp_cache(void)
{
	kmem_cache_destroy(small_qp_cache);
}