summaryrefslogtreecommitdiff
path: root/drivers/staging/octeon/ethernet-rx.c
blob: a10fe3af9a9ce8e00d59c0620eac5acd91adc541 (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
/*
 * This file is based on code from OCTEON SDK by Cavium Networks.
 *
 * Copyright (c) 2003-2010 Cavium Networks
 *
 * This file is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License, Version 2, as
 * published by the Free Software Foundation.
 */

#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/cache.h>
#include <linux/cpumask.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/ip.h>
#include <linux/string.h>
#include <linux/prefetch.h>
#include <linux/ratelimit.h>
#include <linux/smp.h>
#include <linux/interrupt.h>
#include <net/dst.h>
#ifdef CONFIG_XFRM
#include <linux/xfrm.h>
#include <net/xfrm.h>
#endif /* CONFIG_XFRM */

#include <asm/octeon/octeon.h>

#include "ethernet-defines.h"
#include "ethernet-mem.h"
#include "ethernet-rx.h"
#include "octeon-ethernet.h"
#include "ethernet-util.h"

#include <asm/octeon/cvmx-helper.h>
#include <asm/octeon/cvmx-wqe.h>
#include <asm/octeon/cvmx-fau.h>
#include <asm/octeon/cvmx-pow.h>
#include <asm/octeon/cvmx-pip.h>
#include <asm/octeon/cvmx-scratch.h>

#include <asm/octeon/cvmx-gmxx-defs.h>

static struct napi_struct cvm_oct_napi;

/**
 * cvm_oct_do_interrupt - interrupt handler.
 * @cpl: Interrupt number. Unused
 * @dev_id: Cookie to identify the device. Unused
 *
 * The interrupt occurs whenever the POW has packets in our group.
 *
 */
static irqreturn_t cvm_oct_do_interrupt(int cpl, void *dev_id)
{
	/* Disable the IRQ and start napi_poll. */
	disable_irq_nosync(OCTEON_IRQ_WORKQ0 + pow_receive_group);
	napi_schedule(&cvm_oct_napi);

	return IRQ_HANDLED;
}

/**
 * cvm_oct_check_rcv_error - process receive errors
 * @work: Work queue entry pointing to the packet.
 *
 * Returns Non-zero if the packet can be dropped, zero otherwise.
 */
static inline int cvm_oct_check_rcv_error(cvmx_wqe_t *work)
{
	int port;

	if (octeon_has_feature(OCTEON_FEATURE_PKND))
		port = work->word0.pip.cn68xx.pknd;
	else
		port = work->word1.cn38xx.ipprt;

	if ((work->word2.snoip.err_code == 10) && (work->word1.len <= 64)) {
		/*
		 * Ignore length errors on min size packets. Some
		 * equipment incorrectly pads packets to 64+4FCS
		 * instead of 60+4FCS.  Note these packets still get
		 * counted as frame errors.
		 */
	} else if (work->word2.snoip.err_code == 5 ||
		   work->word2.snoip.err_code == 7) {
		/*
		 * We received a packet with either an alignment error
		 * or a FCS error. This may be signalling that we are
		 * running 10Mbps with GMXX_RXX_FRM_CTL[PRE_CHK]
		 * off. If this is the case we need to parse the
		 * packet to determine if we can remove a non spec
		 * preamble and generate a correct packet.
		 */
		int interface = cvmx_helper_get_interface_num(port);
		int index = cvmx_helper_get_interface_index_num(port);
		union cvmx_gmxx_rxx_frm_ctl gmxx_rxx_frm_ctl;

		gmxx_rxx_frm_ctl.u64 =
		    cvmx_read_csr(CVMX_GMXX_RXX_FRM_CTL(index, interface));
		if (gmxx_rxx_frm_ctl.s.pre_chk == 0) {
			u8 *ptr =
			    cvmx_phys_to_ptr(work->packet_ptr.s.addr);
			int i = 0;

			while (i < work->word1.len - 1) {
				if (*ptr != 0x55)
					break;
				ptr++;
				i++;
			}

			if (*ptr == 0xd5) {
				/* Port received 0xd5 preamble */
				work->packet_ptr.s.addr += i + 1;
				work->word1.len -= i + 5;
			} else if ((*ptr & 0xf) == 0xd) {
				/* Port received 0xd preamble */
				work->packet_ptr.s.addr += i;
				work->word1.len -= i + 4;
				for (i = 0; i < work->word1.len; i++) {
					*ptr =
					    ((*ptr & 0xf0) >> 4) |
					    ((*(ptr + 1) & 0xf) << 4);
					ptr++;
				}
			} else {
				printk_ratelimited("Port %d unknown preamble, packet dropped\n",
						   port);
				cvm_oct_free_work(work);
				return 1;
			}
		}
	} else {
		printk_ratelimited("Port %d receive error code %d, packet dropped\n",
				   port, work->word2.snoip.err_code);
		cvm_oct_free_work(work);
		return 1;
	}

	return 0;
}

/**
 * cvm_oct_napi_poll - the NAPI poll function.
 * @napi: The NAPI instance, or null if called from cvm_oct_poll_controller
 * @budget: Maximum number of packets to receive.
 *
 * Returns the number of packets processed.
 */
static int cvm_oct_napi_poll(struct napi_struct *napi, int budget)
{
	const int	coreid = cvmx_get_core_num();
	u64	old_group_mask;
	u64	old_scratch;
	int		rx_count = 0;
	int		did_work_request = 0;
	int		packet_not_copied;

	/* Prefetch cvm_oct_device since we know we need it soon */
	prefetch(cvm_oct_device);

	if (USE_ASYNC_IOBDMA) {
		/* Save scratch in case userspace is using it */
		CVMX_SYNCIOBDMA;
		old_scratch = cvmx_scratch_read64(CVMX_SCR_SCRATCH);
	}

	/* Only allow work for our group (and preserve priorities) */
	if (OCTEON_IS_MODEL(OCTEON_CN68XX)) {
		old_group_mask = cvmx_read_csr(CVMX_SSO_PPX_GRP_MSK(coreid));
		cvmx_write_csr(CVMX_SSO_PPX_GRP_MSK(coreid),
			       1ull << pow_receive_group);
		cvmx_read_csr(CVMX_SSO_PPX_GRP_MSK(coreid)); /* Flush */
	} else {
		old_group_mask = cvmx_read_csr(CVMX_POW_PP_GRP_MSKX(coreid));
		cvmx_write_csr(CVMX_POW_PP_GRP_MSKX(coreid),
			       (old_group_mask & ~0xFFFFull) |
			       1 << pow_receive_group);
	}

	if (USE_ASYNC_IOBDMA) {
		cvmx_pow_work_request_async(CVMX_SCR_SCRATCH, CVMX_POW_NO_WAIT);
		did_work_request = 1;
	}

	while (rx_count < budget) {
		struct sk_buff *skb = NULL;
		struct sk_buff **pskb = NULL;
		int skb_in_hw;
		cvmx_wqe_t *work;
		int port;

		if (USE_ASYNC_IOBDMA && did_work_request)
			work = cvmx_pow_work_response_async(CVMX_SCR_SCRATCH);
		else
			work = cvmx_pow_work_request_sync(CVMX_POW_NO_WAIT);

		prefetch(work);
		did_work_request = 0;
		if (!work) {
			if (OCTEON_IS_MODEL(OCTEON_CN68XX)) {
				cvmx_write_csr(CVMX_SSO_WQ_IQ_DIS,
					       1ull << pow_receive_group);
				cvmx_write_csr(CVMX_SSO_WQ_INT,
					       1ull << pow_receive_group);
			} else {
				union cvmx_pow_wq_int wq_int;

				wq_int.u64 = 0;
				wq_int.s.iq_dis = 1 << pow_receive_group;
				wq_int.s.wq_int = 1 << pow_receive_group;
				cvmx_write_csr(CVMX_POW_WQ_INT, wq_int.u64);
			}
			break;
		}
		pskb = (struct sk_buff **)
			(cvm_oct_get_buffer_ptr(work->packet_ptr) -
			sizeof(void *));
		prefetch(pskb);

		if (USE_ASYNC_IOBDMA && rx_count < (budget - 1)) {
			cvmx_pow_work_request_async_nocheck(CVMX_SCR_SCRATCH,
							    CVMX_POW_NO_WAIT);
			did_work_request = 1;
		}
		rx_count++;

		skb_in_hw = work->word2.s.bufs == 1;
		if (likely(skb_in_hw)) {
			skb = *pskb;
			prefetch(&skb->head);
			prefetch(&skb->len);
		}

		if (octeon_has_feature(OCTEON_FEATURE_PKND))
			port = work->word0.pip.cn68xx.pknd;
		else
			port = work->word1.cn38xx.ipprt;

		prefetch(cvm_oct_device[port]);

		/* Immediately throw away all packets with receive errors */
		if (unlikely(work->word2.snoip.rcv_error)) {
			if (cvm_oct_check_rcv_error(work))
				continue;
		}

		/*
		 * We can only use the zero copy path if skbuffs are
		 * in the FPA pool and the packet fits in a single
		 * buffer.
		 */
		if (likely(skb_in_hw)) {
			skb->data = skb->head + work->packet_ptr.s.addr -
				cvmx_ptr_to_phys(skb->head);
			prefetch(skb->data);
			skb->len = work->word1.len;
			skb_set_tail_pointer(skb, skb->len);
			packet_not_copied = 1;
		} else {
			/*
			 * We have to copy the packet. First allocate
			 * an skbuff for it.
			 */
			skb = dev_alloc_skb(work->word1.len);
			if (!skb) {
				cvm_oct_free_work(work);
				continue;
			}

			/*
			 * Check if we've received a packet that was
			 * entirely stored in the work entry.
			 */
			if (unlikely(work->word2.s.bufs == 0)) {
				u8 *ptr = work->packet_data;

				if (likely(!work->word2.s.not_IP)) {
					/*
					 * The beginning of the packet
					 * moves for IP packets.
					 */
					if (work->word2.s.is_v6)
						ptr += 2;
					else
						ptr += 6;
				}
				memcpy(skb_put(skb, work->word1.len), ptr,
				       work->word1.len);
				/* No packet buffers to free */
			} else {
				int segments = work->word2.s.bufs;
				union cvmx_buf_ptr segment_ptr =
				    work->packet_ptr;
				int len = work->word1.len;

				while (segments--) {
					union cvmx_buf_ptr next_ptr =
					    *(union cvmx_buf_ptr *)
					      cvmx_phys_to_ptr(
					      segment_ptr.s.addr - 8);

			/*
			 * Octeon Errata PKI-100: The segment size is
			 * wrong. Until it is fixed, calculate the
			 * segment size based on the packet pool
			 * buffer size. When it is fixed, the
			 * following line should be replaced with this
			 * one: int segment_size =
			 * segment_ptr.s.size;
			 */
					int segment_size =
					    CVMX_FPA_PACKET_POOL_SIZE -
					    (segment_ptr.s.addr -
					     (((segment_ptr.s.addr >> 7) -
					       segment_ptr.s.back) << 7));
					/*
					 * Don't copy more than what
					 * is left in the packet.
					 */
					if (segment_size > len)
						segment_size = len;
					/* Copy the data into the packet */
					memcpy(skb_put(skb, segment_size),
					       cvmx_phys_to_ptr(
					       segment_ptr.s.addr),
					       segment_size);
					len -= segment_size;
					segment_ptr = next_ptr;
				}
			}
			packet_not_copied = 0;
		}
		if (likely((port < TOTAL_NUMBER_OF_PORTS) &&
			   cvm_oct_device[port])) {
			struct net_device *dev = cvm_oct_device[port];
			struct octeon_ethernet *priv = netdev_priv(dev);

			/*
			 * Only accept packets for devices that are
			 * currently up.
			 */
			if (likely(dev->flags & IFF_UP)) {
				skb->protocol = eth_type_trans(skb, dev);
				skb->dev = dev;

				if (unlikely(work->word2.s.not_IP ||
					     work->word2.s.IP_exc ||
					     work->word2.s.L4_error ||
					     !work->word2.s.tcp_or_udp))
					skb->ip_summed = CHECKSUM_NONE;
				else
					skb->ip_summed = CHECKSUM_UNNECESSARY;

				/* Increment RX stats for virtual ports */
				if (port >= CVMX_PIP_NUM_INPUT_PORTS) {
					priv->stats.rx_packets++;
					priv->stats.rx_bytes += skb->len;
				}
				netif_receive_skb(skb);
			} else {
				/*
				 * Drop any packet received for a device that
				 * isn't up.
				 */
				priv->stats.rx_dropped++;
				dev_kfree_skb_irq(skb);
			}
		} else {
			/*
			 * Drop any packet received for a device that
			 * doesn't exist.
			 */
			printk_ratelimited("Port %d not controlled by Linux, packet dropped\n",
					   port);
			dev_kfree_skb_irq(skb);
		}
		/*
		 * Check to see if the skbuff and work share the same
		 * packet buffer.
		 */
		if (likely(packet_not_copied)) {
			/*
			 * This buffer needs to be replaced, increment
			 * the number of buffers we need to free by
			 * one.
			 */
			cvmx_fau_atomic_add32(FAU_NUM_PACKET_BUFFERS_TO_FREE,
					      1);

			cvmx_fpa_free(work, CVMX_FPA_WQE_POOL, 1);
		} else {
			cvm_oct_free_work(work);
		}
	}
	/* Restore the original POW group mask */
	if (OCTEON_IS_MODEL(OCTEON_CN68XX)) {
		cvmx_write_csr(CVMX_SSO_PPX_GRP_MSK(coreid), old_group_mask);
		cvmx_read_csr(CVMX_SSO_PPX_GRP_MSK(coreid)); /* Flush */
	} else {
		cvmx_write_csr(CVMX_POW_PP_GRP_MSKX(coreid), old_group_mask);
	}

	if (USE_ASYNC_IOBDMA) {
		/* Restore the scratch area */
		cvmx_scratch_write64(CVMX_SCR_SCRATCH, old_scratch);
	}
	cvm_oct_rx_refill_pool(0);

	if (rx_count < budget && napi) {
		/* No more work */
		napi_complete(napi);
		enable_irq(OCTEON_IRQ_WORKQ0 + pow_receive_group);
	}
	return rx_count;
}

#ifdef CONFIG_NET_POLL_CONTROLLER
/**
 * cvm_oct_poll_controller - poll for receive packets
 * device.
 *
 * @dev:    Device to poll. Unused
 */
void cvm_oct_poll_controller(struct net_device *dev)
{
	cvm_oct_napi_poll(NULL, 16);
}
#endif

void cvm_oct_rx_initialize(void)
{
	int i;
	struct net_device *dev_for_napi = NULL;

	for (i = 0; i < TOTAL_NUMBER_OF_PORTS; i++) {
		if (cvm_oct_device[i]) {
			dev_for_napi = cvm_oct_device[i];
			break;
		}
	}

	if (!dev_for_napi)
		panic("No net_devices were allocated.");

	netif_napi_add(dev_for_napi, &cvm_oct_napi, cvm_oct_napi_poll,
		       rx_napi_weight);
	napi_enable(&cvm_oct_napi);

	/* Register an IRQ handler to receive POW interrupts */
	i = request_irq(OCTEON_IRQ_WORKQ0 + pow_receive_group,
			cvm_oct_do_interrupt, 0, "Ethernet", cvm_oct_device);

	if (i)
		panic("Could not acquire Ethernet IRQ %d\n",
		      OCTEON_IRQ_WORKQ0 + pow_receive_group);

	disable_irq_nosync(OCTEON_IRQ_WORKQ0 + pow_receive_group);

	/* Enable POW interrupt when our port has at least one packet */
	if (OCTEON_IS_MODEL(OCTEON_CN68XX)) {
		union cvmx_sso_wq_int_thrx int_thr;
		union cvmx_pow_wq_int_pc int_pc;

		int_thr.u64 = 0;
		int_thr.s.tc_en = 1;
		int_thr.s.tc_thr = 1;
		cvmx_write_csr(CVMX_SSO_WQ_INT_THRX(pow_receive_group),
			       int_thr.u64);

		int_pc.u64 = 0;
		int_pc.s.pc_thr = 5;
		cvmx_write_csr(CVMX_SSO_WQ_INT_PC, int_pc.u64);
	} else {
		union cvmx_pow_wq_int_thrx int_thr;
		union cvmx_pow_wq_int_pc int_pc;

		int_thr.u64 = 0;
		int_thr.s.tc_en = 1;
		int_thr.s.tc_thr = 1;
		cvmx_write_csr(CVMX_POW_WQ_INT_THRX(pow_receive_group),
			       int_thr.u64);

		int_pc.u64 = 0;
		int_pc.s.pc_thr = 5;
		cvmx_write_csr(CVMX_POW_WQ_INT_PC, int_pc.u64);
	}

	/* Schedule NAPI now. This will indirectly enable the interrupt. */
	napi_schedule(&cvm_oct_napi);
}

void cvm_oct_rx_shutdown(void)
{
	netif_napi_del(&cvm_oct_napi);
}