?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
myri10ge: update version
[linux-2.6/cjktty.git] / drivers / net / myri10ge / myri10ge.c
blob1d2247554a35a0b5a6aff465207a2dc9a8243c21
1 /*************************************************************************
2 * myri10ge.c: Myricom Myri-10G Ethernet driver.
3 *
4 * Copyright (C) 2005 - 2011 Myricom, Inc.
5 * All rights reserved.
6 *
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
9 * are met:
10 * 1. Redistributions of source code must retain the above copyright
11 * notice, this list of conditions and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the distribution.
15 * 3. Neither the name of Myricom, Inc. nor the names of its contributors
16 * may be used to endorse or promote products derived from this software
17 * without specific prior written permission.
18 *
19 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
20 * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22 * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
23 * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
24 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
25 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
26 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
27 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
28 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
29 * POSSIBILITY OF SUCH DAMAGE.
30 *
31 *
32 * If the eeprom on your board is not recent enough, you will need to get a
33 * newer firmware image at:
34 * http://www.myri.com/scs/download-Myri10GE.html
35 *
36 * Contact Information:
37 * <help@myri.com>
38 * Myricom, Inc., 325N Santa Anita Avenue, Arcadia, CA 91006
39 *************************************************************************/
40
41 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
42
43 #include <linux/tcp.h>
44 #include <linux/netdevice.h>
45 #include <linux/skbuff.h>
46 #include <linux/string.h>
47 #include <linux/module.h>
48 #include <linux/pci.h>
49 #include <linux/dma-mapping.h>
50 #include <linux/etherdevice.h>
51 #include <linux/if_ether.h>
52 #include <linux/if_vlan.h>
53 #include <linux/inet_lro.h>
54 #include <linux/dca.h>
55 #include <linux/ip.h>
56 #include <linux/inet.h>
57 #include <linux/in.h>
58 #include <linux/ethtool.h>
59 #include <linux/firmware.h>
60 #include <linux/delay.h>
61 #include <linux/timer.h>
62 #include <linux/vmalloc.h>
63 #include <linux/crc32.h>
64 #include <linux/moduleparam.h>
65 #include <linux/io.h>
66 #include <linux/log2.h>
67 #include <linux/slab.h>
68 #include <linux/prefetch.h>
69 #include <net/checksum.h>
70 #include <net/ip.h>
71 #include <net/tcp.h>
72 #include <asm/byteorder.h>
73 #include <asm/io.h>
74 #include <asm/processor.h>
75 #ifdef CONFIG_MTRR
76 #include <asm/mtrr.h>
77 #endif
78
79 #include "myri10ge_mcp.h"
80 #include "myri10ge_mcp_gen_header.h"
81
82 #define MYRI10GE_VERSION_STR "1.5.3-1.534"
83
84 MODULE_DESCRIPTION("Myricom 10G driver (10GbE)");
85 MODULE_AUTHOR("Maintainer: help@myri.com");
86 MODULE_VERSION(MYRI10GE_VERSION_STR);
87 MODULE_LICENSE("Dual BSD/GPL");
88
89 #define MYRI10GE_MAX_ETHER_MTU 9014
90
91 #define MYRI10GE_ETH_STOPPED 0
92 #define MYRI10GE_ETH_STOPPING 1
93 #define MYRI10GE_ETH_STARTING 2
94 #define MYRI10GE_ETH_RUNNING 3
95 #define MYRI10GE_ETH_OPEN_FAILED 4
96
97 #define MYRI10GE_EEPROM_STRINGS_SIZE 256
98 #define MYRI10GE_MAX_SEND_DESC_TSO ((65536 / 2048) * 2)
99 #define MYRI10GE_MAX_LRO_DESCRIPTORS 8
100 #define MYRI10GE_LRO_MAX_PKTS 64
101
102 #define MYRI10GE_NO_CONFIRM_DATA htonl(0xffffffff)
103 #define MYRI10GE_NO_RESPONSE_RESULT 0xffffffff
104
105 #define MYRI10GE_ALLOC_ORDER 0
106 #define MYRI10GE_ALLOC_SIZE ((1 << MYRI10GE_ALLOC_ORDER) * PAGE_SIZE)
107 #define MYRI10GE_MAX_FRAGS_PER_FRAME (MYRI10GE_MAX_ETHER_MTU/MYRI10GE_ALLOC_SIZE + 1)
108
109 #define MYRI10GE_MAX_SLICES 32
110
111 struct myri10ge_rx_buffer_state {
112 struct page *page;
113 int page_offset;
114 DEFINE_DMA_UNMAP_ADDR(bus);
115 DEFINE_DMA_UNMAP_LEN(len);
116 };
117
118 struct myri10ge_tx_buffer_state {
119 struct sk_buff *skb;
120 int last;
121 DEFINE_DMA_UNMAP_ADDR(bus);
122 DEFINE_DMA_UNMAP_LEN(len);
123 };
124
125 struct myri10ge_cmd {
126 u32 data0;
127 u32 data1;
128 u32 data2;
129 };
130
131 struct myri10ge_rx_buf {
132 struct mcp_kreq_ether_recv __iomem *lanai; /* lanai ptr for recv ring */
133 struct mcp_kreq_ether_recv *shadow; /* host shadow of recv ring */
134 struct myri10ge_rx_buffer_state *info;
135 struct page *page;
136 dma_addr_t bus;
137 int page_offset;
138 int cnt;
139 int fill_cnt;
140 int alloc_fail;
141 int mask; /* number of rx slots -1 */
142 int watchdog_needed;
143 };
144
145 struct myri10ge_tx_buf {
146 struct mcp_kreq_ether_send __iomem *lanai; /* lanai ptr for sendq */
147 __be32 __iomem *send_go; /* "go" doorbell ptr */
148 __be32 __iomem *send_stop; /* "stop" doorbell ptr */
149 struct mcp_kreq_ether_send *req_list; /* host shadow of sendq */
150 char *req_bytes;
151 struct myri10ge_tx_buffer_state *info;
152 int mask; /* number of transmit slots -1 */
153 int req ____cacheline_aligned; /* transmit slots submitted */
154 int pkt_start; /* packets started */
155 int stop_queue;
156 int linearized;
157 int done ____cacheline_aligned; /* transmit slots completed */
158 int pkt_done; /* packets completed */
159 int wake_queue;
160 int queue_active;
161 };
162
163 struct myri10ge_rx_done {
164 struct mcp_slot *entry;
165 dma_addr_t bus;
166 int cnt;
167 int idx;
168 struct net_lro_mgr lro_mgr;
169 struct net_lro_desc lro_desc[MYRI10GE_MAX_LRO_DESCRIPTORS];
170 };
171
172 struct myri10ge_slice_netstats {
173 unsigned long rx_packets;
174 unsigned long tx_packets;
175 unsigned long rx_bytes;
176 unsigned long tx_bytes;
177 unsigned long rx_dropped;
178 unsigned long tx_dropped;
179 };
180
181 struct myri10ge_slice_state {
182 struct myri10ge_tx_buf tx; /* transmit ring */
183 struct myri10ge_rx_buf rx_small;
184 struct myri10ge_rx_buf rx_big;
185 struct myri10ge_rx_done rx_done;
186 struct net_device *dev;
187 struct napi_struct napi;
188 struct myri10ge_priv *mgp;
189 struct myri10ge_slice_netstats stats;
190 __be32 __iomem *irq_claim;
191 struct mcp_irq_data *fw_stats;
192 dma_addr_t fw_stats_bus;
193 int watchdog_tx_done;
194 int watchdog_tx_req;
195 int watchdog_rx_done;
196 int stuck;
197 #ifdef CONFIG_MYRI10GE_DCA
198 int cached_dca_tag;
199 int cpu;
200 __be32 __iomem *dca_tag;
201 #endif
202 char irq_desc[32];
203 };
204
205 struct myri10ge_priv {
206 struct myri10ge_slice_state *ss;
207 int tx_boundary; /* boundary transmits cannot cross */
208 int num_slices;
209 int running; /* running? */
210 int small_bytes;
211 int big_bytes;
212 int max_intr_slots;
213 struct net_device *dev;
214 u8 __iomem *sram;
215 int sram_size;
216 unsigned long board_span;
217 unsigned long iomem_base;
218 __be32 __iomem *irq_deassert;
219 char *mac_addr_string;
220 struct mcp_cmd_response *cmd;
221 dma_addr_t cmd_bus;
222 struct pci_dev *pdev;
223 int msi_enabled;
224 int msix_enabled;
225 struct msix_entry *msix_vectors;
226 #ifdef CONFIG_MYRI10GE_DCA
227 int dca_enabled;
228 int relaxed_order;
229 #endif
230 u32 link_state;
231 unsigned int rdma_tags_available;
232 int intr_coal_delay;
233 __be32 __iomem *intr_coal_delay_ptr;
234 int mtrr;
235 int wc_enabled;
236 int down_cnt;
237 wait_queue_head_t down_wq;
238 struct work_struct watchdog_work;
239 struct timer_list watchdog_timer;
240 int watchdog_resets;
241 int watchdog_pause;
242 int pause;
243 bool fw_name_allocated;
244 char *fw_name;
245 char eeprom_strings[MYRI10GE_EEPROM_STRINGS_SIZE];
246 char *product_code_string;
247 char fw_version[128];
248 int fw_ver_major;
249 int fw_ver_minor;
250 int fw_ver_tiny;
251 int adopted_rx_filter_bug;
252 u8 mac_addr[6]; /* eeprom mac address */
253 unsigned long serial_number;
254 int vendor_specific_offset;
255 int fw_multicast_support;
256 u32 features;
257 u32 max_tso6;
258 u32 read_dma;
259 u32 write_dma;
260 u32 read_write_dma;
261 u32 link_changes;
262 u32 msg_enable;
263 unsigned int board_number;
264 int rebooted;
265 };
266
267 static char *myri10ge_fw_unaligned = "myri10ge_ethp_z8e.dat";
268 static char *myri10ge_fw_aligned = "myri10ge_eth_z8e.dat";
269 static char *myri10ge_fw_rss_unaligned = "myri10ge_rss_ethp_z8e.dat";
270 static char *myri10ge_fw_rss_aligned = "myri10ge_rss_eth_z8e.dat";
271 MODULE_FIRMWARE("myri10ge_ethp_z8e.dat");
272 MODULE_FIRMWARE("myri10ge_eth_z8e.dat");
273 MODULE_FIRMWARE("myri10ge_rss_ethp_z8e.dat");
274 MODULE_FIRMWARE("myri10ge_rss_eth_z8e.dat");
275
276 /* Careful: must be accessed under kparam_block_sysfs_write */
277 static char *myri10ge_fw_name = NULL;
278 module_param(myri10ge_fw_name, charp, S_IRUGO | S_IWUSR);
279 MODULE_PARM_DESC(myri10ge_fw_name, "Firmware image name");
280
281 #define MYRI10GE_MAX_BOARDS 8
282 static char *myri10ge_fw_names[MYRI10GE_MAX_BOARDS] =
283 {[0 ... (MYRI10GE_MAX_BOARDS - 1)] = NULL };
284 module_param_array_named(myri10ge_fw_names, myri10ge_fw_names, charp, NULL,
285 0444);
286 MODULE_PARM_DESC(myri10ge_fw_name, "Firmware image names per board");
287
288 static int myri10ge_ecrc_enable = 1;
289 module_param(myri10ge_ecrc_enable, int, S_IRUGO);
290 MODULE_PARM_DESC(myri10ge_ecrc_enable, "Enable Extended CRC on PCI-E");
291
292 static int myri10ge_small_bytes = -1; /* -1 == auto */
293 module_param(myri10ge_small_bytes, int, S_IRUGO | S_IWUSR);
294 MODULE_PARM_DESC(myri10ge_small_bytes, "Threshold of small packets");
295
296 static int myri10ge_msi = 1; /* enable msi by default */
297 module_param(myri10ge_msi, int, S_IRUGO | S_IWUSR);
298 MODULE_PARM_DESC(myri10ge_msi, "Enable Message Signalled Interrupts");
299
300 static int myri10ge_intr_coal_delay = 75;
301 module_param(myri10ge_intr_coal_delay, int, S_IRUGO);
302 MODULE_PARM_DESC(myri10ge_intr_coal_delay, "Interrupt coalescing delay");
303
304 static int myri10ge_flow_control = 1;
305 module_param(myri10ge_flow_control, int, S_IRUGO);
306 MODULE_PARM_DESC(myri10ge_flow_control, "Pause parameter");
307
308 static int myri10ge_deassert_wait = 1;
309 module_param(myri10ge_deassert_wait, int, S_IRUGO | S_IWUSR);
310 MODULE_PARM_DESC(myri10ge_deassert_wait,
311 "Wait when deasserting legacy interrupts");
312
313 static int myri10ge_force_firmware = 0;
314 module_param(myri10ge_force_firmware, int, S_IRUGO);
315 MODULE_PARM_DESC(myri10ge_force_firmware,
316 "Force firmware to assume aligned completions");
317
318 static int myri10ge_initial_mtu = MYRI10GE_MAX_ETHER_MTU - ETH_HLEN;
319 module_param(myri10ge_initial_mtu, int, S_IRUGO);
320 MODULE_PARM_DESC(myri10ge_initial_mtu, "Initial MTU");
321
322 static int myri10ge_napi_weight = 64;
323 module_param(myri10ge_napi_weight, int, S_IRUGO);
324 MODULE_PARM_DESC(myri10ge_napi_weight, "Set NAPI weight");
325
326 static int myri10ge_watchdog_timeout = 1;
327 module_param(myri10ge_watchdog_timeout, int, S_IRUGO);
328 MODULE_PARM_DESC(myri10ge_watchdog_timeout, "Set watchdog timeout");
329
330 static int myri10ge_max_irq_loops = 1048576;
331 module_param(myri10ge_max_irq_loops, int, S_IRUGO);
332 MODULE_PARM_DESC(myri10ge_max_irq_loops,
333 "Set stuck legacy IRQ detection threshold");
334
335 #define MYRI10GE_MSG_DEFAULT NETIF_MSG_LINK
336
337 static int myri10ge_debug = -1; /* defaults above */
338 module_param(myri10ge_debug, int, 0);
339 MODULE_PARM_DESC(myri10ge_debug, "Debug level (0=none,...,16=all)");
340
341 static int myri10ge_lro_max_pkts = MYRI10GE_LRO_MAX_PKTS;
342 module_param(myri10ge_lro_max_pkts, int, S_IRUGO);
343 MODULE_PARM_DESC(myri10ge_lro_max_pkts,
344 "Number of LRO packets to be aggregated");
345
346 static int myri10ge_fill_thresh = 256;
347 module_param(myri10ge_fill_thresh, int, S_IRUGO | S_IWUSR);
348 MODULE_PARM_DESC(myri10ge_fill_thresh, "Number of empty rx slots allowed");
349
350 static int myri10ge_reset_recover = 1;
351
352 static int myri10ge_max_slices = 1;
353 module_param(myri10ge_max_slices, int, S_IRUGO);
354 MODULE_PARM_DESC(myri10ge_max_slices, "Max tx/rx queues");
355
356 static int myri10ge_rss_hash = MXGEFW_RSS_HASH_TYPE_SRC_DST_PORT;
357 module_param(myri10ge_rss_hash, int, S_IRUGO);
358 MODULE_PARM_DESC(myri10ge_rss_hash, "Type of RSS hashing to do");
359
360 static int myri10ge_dca = 1;
361 module_param(myri10ge_dca, int, S_IRUGO);
362 MODULE_PARM_DESC(myri10ge_dca, "Enable DCA if possible");
363
364 #define MYRI10GE_FW_OFFSET 1024*1024
365 #define MYRI10GE_HIGHPART_TO_U32(X) \
366 (sizeof (X) == 8) ? ((u32)((u64)(X) >> 32)) : (0)
367 #define MYRI10GE_LOWPART_TO_U32(X) ((u32)(X))
368
369 #define myri10ge_pio_copy(to,from,size) __iowrite64_copy(to,from,size/8)
370
371 static void myri10ge_set_multicast_list(struct net_device *dev);
372 static netdev_tx_t myri10ge_sw_tso(struct sk_buff *skb,
373 struct net_device *dev);
374
375 static inline void put_be32(__be32 val, __be32 __iomem * p)
376 {
377 __raw_writel((__force __u32) val, (__force void __iomem *)p);
378 }
379
380 static struct rtnl_link_stats64 *myri10ge_get_stats(struct net_device *dev,
381 struct rtnl_link_stats64 *stats);
382
383 static void set_fw_name(struct myri10ge_priv *mgp, char *name, bool allocated)
384 {
385 if (mgp->fw_name_allocated)
386 kfree(mgp->fw_name);
387 mgp->fw_name = name;
388 mgp->fw_name_allocated = allocated;
389 }
390
391 static int
392 myri10ge_send_cmd(struct myri10ge_priv *mgp, u32 cmd,
393 struct myri10ge_cmd *data, int atomic)
394 {
395 struct mcp_cmd *buf;
396 char buf_bytes[sizeof(*buf) + 8];
397 struct mcp_cmd_response *response = mgp->cmd;
398 char __iomem *cmd_addr = mgp->sram + MXGEFW_ETH_CMD;
399 u32 dma_low, dma_high, result, value;
400 int sleep_total = 0;
401
402 /* ensure buf is aligned to 8 bytes */
403 buf = (struct mcp_cmd *)ALIGN((unsigned long)buf_bytes, 8);
404
405 buf->data0 = htonl(data->data0);
406 buf->data1 = htonl(data->data1);
407 buf->data2 = htonl(data->data2);
408 buf->cmd = htonl(cmd);
409 dma_low = MYRI10GE_LOWPART_TO_U32(mgp->cmd_bus);
410 dma_high = MYRI10GE_HIGHPART_TO_U32(mgp->cmd_bus);
411
412 buf->response_addr.low = htonl(dma_low);
413 buf->response_addr.high = htonl(dma_high);
414 response->result = htonl(MYRI10GE_NO_RESPONSE_RESULT);
415 mb();
416 myri10ge_pio_copy(cmd_addr, buf, sizeof(*buf));
417
418 /* wait up to 15ms. Longest command is the DMA benchmark,
419 * which is capped at 5ms, but runs from a timeout handler
420 * that runs every 7.8ms. So a 15ms timeout leaves us with
421 * a 2.2ms margin
422 */
423 if (atomic) {
424 /* if atomic is set, do not sleep,
425 * and try to get the completion quickly
426 * (1ms will be enough for those commands) */
427 for (sleep_total = 0;
428 sleep_total < 1000 &&
429 response->result == htonl(MYRI10GE_NO_RESPONSE_RESULT);
430 sleep_total += 10) {
431 udelay(10);
432 mb();
433 }
434 } else {
435 /* use msleep for most command */
436 for (sleep_total = 0;
437 sleep_total < 15 &&
438 response->result == htonl(MYRI10GE_NO_RESPONSE_RESULT);
439 sleep_total++)
440 msleep(1);
441 }
442
443 result = ntohl(response->result);
444 value = ntohl(response->data);
445 if (result != MYRI10GE_NO_RESPONSE_RESULT) {
446 if (result == 0) {
447 data->data0 = value;
448 return 0;
449 } else if (result == MXGEFW_CMD_UNKNOWN) {
450 return -ENOSYS;
451 } else if (result == MXGEFW_CMD_ERROR_UNALIGNED) {
452 return -E2BIG;
453 } else if (result == MXGEFW_CMD_ERROR_RANGE &&
454 cmd == MXGEFW_CMD_ENABLE_RSS_QUEUES &&
455 (data->
456 data1 & MXGEFW_SLICE_ENABLE_MULTIPLE_TX_QUEUES) !=
457 0) {
458 return -ERANGE;
459 } else {
460 dev_err(&mgp->pdev->dev,
461 "command %d failed, result = %d\n",
462 cmd, result);
463 return -ENXIO;
464 }
465 }
466
467 dev_err(&mgp->pdev->dev, "command %d timed out, result = %d\n",
468 cmd, result);
469 return -EAGAIN;
470 }
471
472 /*
473 * The eeprom strings on the lanaiX have the format
474 * SN=x\0
475 * MAC=x:x:x:x:x:x\0
476 * PT:ddd mmm xx xx:xx:xx xx\0
477 * PV:ddd mmm xx xx:xx:xx xx\0
478 */
479 static int myri10ge_read_mac_addr(struct myri10ge_priv *mgp)
480 {
481 char *ptr, *limit;
482 int i;
483
484 ptr = mgp->eeprom_strings;
485 limit = mgp->eeprom_strings + MYRI10GE_EEPROM_STRINGS_SIZE;
486
487 while (*ptr != '\0' && ptr < limit) {
488 if (memcmp(ptr, "MAC=", 4) == 0) {
489 ptr += 4;
490 mgp->mac_addr_string = ptr;
491 for (i = 0; i < 6; i++) {
492 if ((ptr + 2) > limit)
493 goto abort;
494 mgp->mac_addr[i] =
495 simple_strtoul(ptr, &ptr, 16);
496 ptr += 1;
497 }
498 }
499 if (memcmp(ptr, "PC=", 3) == 0) {
500 ptr += 3;
501 mgp->product_code_string = ptr;
502 }
503 if (memcmp((const void *)ptr, "SN=", 3) == 0) {
504 ptr += 3;
505 mgp->serial_number = simple_strtoul(ptr, &ptr, 10);
506 }
507 while (ptr < limit && *ptr++) ;
508 }
509
510 return 0;
511
512 abort:
513 dev_err(&mgp->pdev->dev, "failed to parse eeprom_strings\n");
514 return -ENXIO;
515 }
516
517 /*
518 * Enable or disable periodic RDMAs from the host to make certain
519 * chipsets resend dropped PCIe messages
520 */
521
522 static void myri10ge_dummy_rdma(struct myri10ge_priv *mgp, int enable)
523 {
524 char __iomem *submit;
525 __be32 buf[16] __attribute__ ((__aligned__(8)));
526 u32 dma_low, dma_high;
527 int i;
528
529 /* clear confirmation addr */
530 mgp->cmd->data = 0;
531 mb();
532
533 /* send a rdma command to the PCIe engine, and wait for the
534 * response in the confirmation address. The firmware should
535 * write a -1 there to indicate it is alive and well
536 */
537 dma_low = MYRI10GE_LOWPART_TO_U32(mgp->cmd_bus);
538 dma_high = MYRI10GE_HIGHPART_TO_U32(mgp->cmd_bus);
539
540 buf[0] = htonl(dma_high); /* confirm addr MSW */
541 buf[1] = htonl(dma_low); /* confirm addr LSW */
542 buf[2] = MYRI10GE_NO_CONFIRM_DATA; /* confirm data */
543 buf[3] = htonl(dma_high); /* dummy addr MSW */
544 buf[4] = htonl(dma_low); /* dummy addr LSW */
545 buf[5] = htonl(enable); /* enable? */
546
547 submit = mgp->sram + MXGEFW_BOOT_DUMMY_RDMA;
548
549 myri10ge_pio_copy(submit, &buf, sizeof(buf));
550 for (i = 0; mgp->cmd->data != MYRI10GE_NO_CONFIRM_DATA && i < 20; i++)
551 msleep(1);
552 if (mgp->cmd->data != MYRI10GE_NO_CONFIRM_DATA)
553 dev_err(&mgp->pdev->dev, "dummy rdma %s failed\n",
554 (enable ? "enable" : "disable"));
555 }
556
557 static int
558 myri10ge_validate_firmware(struct myri10ge_priv *mgp,
559 struct mcp_gen_header *hdr)
560 {
561 struct device *dev = &mgp->pdev->dev;
562
563 /* check firmware type */
564 if (ntohl(hdr->mcp_type) != MCP_TYPE_ETH) {
565 dev_err(dev, "Bad firmware type: 0x%x\n", ntohl(hdr->mcp_type));
566 return -EINVAL;
567 }
568
569 /* save firmware version for ethtool */
570 strncpy(mgp->fw_version, hdr->version, sizeof(mgp->fw_version));
571
572 sscanf(mgp->fw_version, "%d.%d.%d", &mgp->fw_ver_major,
573 &mgp->fw_ver_minor, &mgp->fw_ver_tiny);
574
575 if (!(mgp->fw_ver_major == MXGEFW_VERSION_MAJOR &&
576 mgp->fw_ver_minor == MXGEFW_VERSION_MINOR)) {
577 dev_err(dev, "Found firmware version %s\n", mgp->fw_version);
578 dev_err(dev, "Driver needs %d.%d\n", MXGEFW_VERSION_MAJOR,
579 MXGEFW_VERSION_MINOR);
580 return -EINVAL;
581 }
582 return 0;
583 }
584
585 static int myri10ge_load_hotplug_firmware(struct myri10ge_priv *mgp, u32 * size)
586 {
587 unsigned crc, reread_crc;
588 const struct firmware *fw;
589 struct device *dev = &mgp->pdev->dev;
590 unsigned char *fw_readback;
591 struct mcp_gen_header *hdr;
592 size_t hdr_offset;
593 int status;
594 unsigned i;
595
596 if ((status = request_firmware(&fw, mgp->fw_name, dev)) < 0) {
597 dev_err(dev, "Unable to load %s firmware image via hotplug\n",
598 mgp->fw_name);
599 status = -EINVAL;
600 goto abort_with_nothing;
601 }
602
603 /* check size */
604
605 if (fw->size >= mgp->sram_size - MYRI10GE_FW_OFFSET ||
606 fw->size < MCP_HEADER_PTR_OFFSET + 4) {
607 dev_err(dev, "Firmware size invalid:%d\n", (int)fw->size);
608 status = -EINVAL;
609 goto abort_with_fw;
610 }
611
612 /* check id */
613 hdr_offset = ntohl(*(__be32 *) (fw->data + MCP_HEADER_PTR_OFFSET));
614 if ((hdr_offset & 3) || hdr_offset + sizeof(*hdr) > fw->size) {
615 dev_err(dev, "Bad firmware file\n");
616 status = -EINVAL;
617 goto abort_with_fw;
618 }
619 hdr = (void *)(fw->data + hdr_offset);
620
621 status = myri10ge_validate_firmware(mgp, hdr);
622 if (status != 0)
623 goto abort_with_fw;
624
625 crc = crc32(~0, fw->data, fw->size);
626 for (i = 0; i < fw->size; i += 256) {
627 myri10ge_pio_copy(mgp->sram + MYRI10GE_FW_OFFSET + i,
628 fw->data + i,
629 min(256U, (unsigned)(fw->size - i)));
630 mb();
631 readb(mgp->sram);
632 }
633 fw_readback = vmalloc(fw->size);
634 if (!fw_readback) {
635 status = -ENOMEM;
636 goto abort_with_fw;
637 }
638 /* corruption checking is good for parity recovery and buggy chipset */
639 memcpy_fromio(fw_readback, mgp->sram + MYRI10GE_FW_OFFSET, fw->size);
640 reread_crc = crc32(~0, fw_readback, fw->size);
641 vfree(fw_readback);
642 if (crc != reread_crc) {
643 dev_err(dev, "CRC failed(fw-len=%u), got 0x%x (expect 0x%x)\n",
644 (unsigned)fw->size, reread_crc, crc);
645 status = -EIO;
646 goto abort_with_fw;
647 }
648 *size = (u32) fw->size;
649
650 abort_with_fw:
651 release_firmware(fw);
652
653 abort_with_nothing:
654 return status;
655 }
656
657 static int myri10ge_adopt_running_firmware(struct myri10ge_priv *mgp)
658 {
659 struct mcp_gen_header *hdr;
660 struct device *dev = &mgp->pdev->dev;
661 const size_t bytes = sizeof(struct mcp_gen_header);
662 size_t hdr_offset;
663 int status;
664
665 /* find running firmware header */
666 hdr_offset = swab32(readl(mgp->sram + MCP_HEADER_PTR_OFFSET));
667
668 if ((hdr_offset & 3) || hdr_offset + sizeof(*hdr) > mgp->sram_size) {
669 dev_err(dev, "Running firmware has bad header offset (%d)\n",
670 (int)hdr_offset);
671 return -EIO;
672 }
673
674 /* copy header of running firmware from SRAM to host memory to
675 * validate firmware */
676 hdr = kmalloc(bytes, GFP_KERNEL);
677 if (hdr == NULL) {
678 dev_err(dev, "could not malloc firmware hdr\n");
679 return -ENOMEM;
680 }
681 memcpy_fromio(hdr, mgp->sram + hdr_offset, bytes);
682 status = myri10ge_validate_firmware(mgp, hdr);
683 kfree(hdr);
684
685 /* check to see if adopted firmware has bug where adopting
686 * it will cause broadcasts to be filtered unless the NIC
687 * is kept in ALLMULTI mode */
688 if (mgp->fw_ver_major == 1 && mgp->fw_ver_minor == 4 &&
689 mgp->fw_ver_tiny >= 4 && mgp->fw_ver_tiny <= 11) {
690 mgp->adopted_rx_filter_bug = 1;
691 dev_warn(dev, "Adopting fw %d.%d.%d: "
692 "working around rx filter bug\n",
693 mgp->fw_ver_major, mgp->fw_ver_minor,
694 mgp->fw_ver_tiny);
695 }
696 return status;
697 }
698
699 static int myri10ge_get_firmware_capabilities(struct myri10ge_priv *mgp)
700 {
701 struct myri10ge_cmd cmd;
702 int status;
703
704 /* probe for IPv6 TSO support */
705 mgp->features = NETIF_F_SG | NETIF_F_HW_CSUM | NETIF_F_TSO;
706 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_MAX_TSO6_HDR_SIZE,
707 &cmd, 0);
708 if (status == 0) {
709 mgp->max_tso6 = cmd.data0;
710 mgp->features |= NETIF_F_TSO6;
711 }
712
713 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_RX_RING_SIZE, &cmd, 0);
714 if (status != 0) {
715 dev_err(&mgp->pdev->dev,
716 "failed MXGEFW_CMD_GET_RX_RING_SIZE\n");
717 return -ENXIO;
718 }
719
720 mgp->max_intr_slots = 2 * (cmd.data0 / sizeof(struct mcp_dma_addr));
721
722 return 0;
723 }
724
725 static int myri10ge_load_firmware(struct myri10ge_priv *mgp, int adopt)
726 {
727 char __iomem *submit;
728 __be32 buf[16] __attribute__ ((__aligned__(8)));
729 u32 dma_low, dma_high, size;
730 int status, i;
731
732 size = 0;
733 status = myri10ge_load_hotplug_firmware(mgp, &size);
734 if (status) {
735 if (!adopt)
736 return status;
737 dev_warn(&mgp->pdev->dev, "hotplug firmware loading failed\n");
738
739 /* Do not attempt to adopt firmware if there
740 * was a bad crc */
741 if (status == -EIO)
742 return status;
743
744 status = myri10ge_adopt_running_firmware(mgp);
745 if (status != 0) {
746 dev_err(&mgp->pdev->dev,
747 "failed to adopt running firmware\n");
748 return status;
749 }
750 dev_info(&mgp->pdev->dev,
751 "Successfully adopted running firmware\n");
752 if (mgp->tx_boundary == 4096) {
753 dev_warn(&mgp->pdev->dev,
754 "Using firmware currently running on NIC"
755 ". For optimal\n");
756 dev_warn(&mgp->pdev->dev,
757 "performance consider loading optimized "
758 "firmware\n");
759 dev_warn(&mgp->pdev->dev, "via hotplug\n");
760 }
761
762 set_fw_name(mgp, "adopted", false);
763 mgp->tx_boundary = 2048;
764 myri10ge_dummy_rdma(mgp, 1);
765 status = myri10ge_get_firmware_capabilities(mgp);
766 return status;
767 }
768
769 /* clear confirmation addr */
770 mgp->cmd->data = 0;
771 mb();
772
773 /* send a reload command to the bootstrap MCP, and wait for the
774 * response in the confirmation address. The firmware should
775 * write a -1 there to indicate it is alive and well
776 */
777 dma_low = MYRI10GE_LOWPART_TO_U32(mgp->cmd_bus);
778 dma_high = MYRI10GE_HIGHPART_TO_U32(mgp->cmd_bus);
779
780 buf[0] = htonl(dma_high); /* confirm addr MSW */
781 buf[1] = htonl(dma_low); /* confirm addr LSW */
782 buf[2] = MYRI10GE_NO_CONFIRM_DATA; /* confirm data */
783
784 /* FIX: All newest firmware should un-protect the bottom of
785 * the sram before handoff. However, the very first interfaces
786 * do not. Therefore the handoff copy must skip the first 8 bytes
787 */
788 buf[3] = htonl(MYRI10GE_FW_OFFSET + 8); /* where the code starts */
789 buf[4] = htonl(size - 8); /* length of code */
790 buf[5] = htonl(8); /* where to copy to */
791 buf[6] = htonl(0); /* where to jump to */
792
793 submit = mgp->sram + MXGEFW_BOOT_HANDOFF;
794
795 myri10ge_pio_copy(submit, &buf, sizeof(buf));
796 mb();
797 msleep(1);
798 mb();
799 i = 0;
800 while (mgp->cmd->data != MYRI10GE_NO_CONFIRM_DATA && i < 9) {
801 msleep(1 << i);
802 i++;
803 }
804 if (mgp->cmd->data != MYRI10GE_NO_CONFIRM_DATA) {
805 dev_err(&mgp->pdev->dev, "handoff failed\n");
806 return -ENXIO;
807 }
808 myri10ge_dummy_rdma(mgp, 1);
809 status = myri10ge_get_firmware_capabilities(mgp);
810
811 return status;
812 }
813
814 static int myri10ge_update_mac_address(struct myri10ge_priv *mgp, u8 * addr)
815 {
816 struct myri10ge_cmd cmd;
817 int status;
818
819 cmd.data0 = ((addr[0] << 24) | (addr[1] << 16)
820 | (addr[2] << 8) | addr[3]);
821
822 cmd.data1 = ((addr[4] << 8) | (addr[5]));
823
824 status = myri10ge_send_cmd(mgp, MXGEFW_SET_MAC_ADDRESS, &cmd, 0);
825 return status;
826 }
827
828 static int myri10ge_change_pause(struct myri10ge_priv *mgp, int pause)
829 {
830 struct myri10ge_cmd cmd;
831 int status, ctl;
832
833 ctl = pause ? MXGEFW_ENABLE_FLOW_CONTROL : MXGEFW_DISABLE_FLOW_CONTROL;
834 status = myri10ge_send_cmd(mgp, ctl, &cmd, 0);
835
836 if (status) {
837 netdev_err(mgp->dev, "Failed to set flow control mode\n");
838 return status;
839 }
840 mgp->pause = pause;
841 return 0;
842 }
843
844 static void
845 myri10ge_change_promisc(struct myri10ge_priv *mgp, int promisc, int atomic)
846 {
847 struct myri10ge_cmd cmd;
848 int status, ctl;
849
850 ctl = promisc ? MXGEFW_ENABLE_PROMISC : MXGEFW_DISABLE_PROMISC;
851 status = myri10ge_send_cmd(mgp, ctl, &cmd, atomic);
852 if (status)
853 netdev_err(mgp->dev, "Failed to set promisc mode\n");
854 }
855
856 static int myri10ge_dma_test(struct myri10ge_priv *mgp, int test_type)
857 {
858 struct myri10ge_cmd cmd;
859 int status;
860 u32 len;
861 struct page *dmatest_page;
862 dma_addr_t dmatest_bus;
863 char *test = " ";
864
865 dmatest_page = alloc_page(GFP_KERNEL);
866 if (!dmatest_page)
867 return -ENOMEM;
868 dmatest_bus = pci_map_page(mgp->pdev, dmatest_page, 0, PAGE_SIZE,
869 DMA_BIDIRECTIONAL);
870
871 /* Run a small DMA test.
872 * The magic multipliers to the length tell the firmware
873 * to do DMA read, write, or read+write tests. The
874 * results are returned in cmd.data0. The upper 16
875 * bits or the return is the number of transfers completed.
876 * The lower 16 bits is the time in 0.5us ticks that the
877 * transfers took to complete.
878 */
879
880 len = mgp->tx_boundary;
881
882 cmd.data0 = MYRI10GE_LOWPART_TO_U32(dmatest_bus);
883 cmd.data1 = MYRI10GE_HIGHPART_TO_U32(dmatest_bus);
884 cmd.data2 = len * 0x10000;
885 status = myri10ge_send_cmd(mgp, test_type, &cmd, 0);
886 if (status != 0) {
887 test = "read";
888 goto abort;
889 }
890 mgp->read_dma = ((cmd.data0 >> 16) * len * 2) / (cmd.data0 & 0xffff);
891 cmd.data0 = MYRI10GE_LOWPART_TO_U32(dmatest_bus);
892 cmd.data1 = MYRI10GE_HIGHPART_TO_U32(dmatest_bus);
893 cmd.data2 = len * 0x1;
894 status = myri10ge_send_cmd(mgp, test_type, &cmd, 0);
895 if (status != 0) {
896 test = "write";
897 goto abort;
898 }
899 mgp->write_dma = ((cmd.data0 >> 16) * len * 2) / (cmd.data0 & 0xffff);
900
901 cmd.data0 = MYRI10GE_LOWPART_TO_U32(dmatest_bus);
902 cmd.data1 = MYRI10GE_HIGHPART_TO_U32(dmatest_bus);
903 cmd.data2 = len * 0x10001;
904 status = myri10ge_send_cmd(mgp, test_type, &cmd, 0);
905 if (status != 0) {
906 test = "read/write";
907 goto abort;
908 }
909 mgp->read_write_dma = ((cmd.data0 >> 16) * len * 2 * 2) /
910 (cmd.data0 & 0xffff);
911
912 abort:
913 pci_unmap_page(mgp->pdev, dmatest_bus, PAGE_SIZE, DMA_BIDIRECTIONAL);
914 put_page(dmatest_page);
915
916 if (status != 0 && test_type != MXGEFW_CMD_UNALIGNED_TEST)
917 dev_warn(&mgp->pdev->dev, "DMA %s benchmark failed: %d\n",
918 test, status);
919
920 return status;
921 }
922
923 static int myri10ge_reset(struct myri10ge_priv *mgp)
924 {
925 struct myri10ge_cmd cmd;
926 struct myri10ge_slice_state *ss;
927 int i, status;
928 size_t bytes;
929 #ifdef CONFIG_MYRI10GE_DCA
930 unsigned long dca_tag_off;
931 #endif
932
933 /* try to send a reset command to the card to see if it
934 * is alive */
935 memset(&cmd, 0, sizeof(cmd));
936 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_RESET, &cmd, 0);
937 if (status != 0) {
938 dev_err(&mgp->pdev->dev, "failed reset\n");
939 return -ENXIO;
940 }
941
942 (void)myri10ge_dma_test(mgp, MXGEFW_DMA_TEST);
943 /*
944 * Use non-ndis mcp_slot (eg, 4 bytes total,
945 * no toeplitz hash value returned. Older firmware will
946 * not understand this command, but will use the correct
947 * sized mcp_slot, so we ignore error returns
948 */
949 cmd.data0 = MXGEFW_RSS_MCP_SLOT_TYPE_MIN;
950 (void)myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_RSS_MCP_SLOT_TYPE, &cmd, 0);
951
952 /* Now exchange information about interrupts */
953
954 bytes = mgp->max_intr_slots * sizeof(*mgp->ss[0].rx_done.entry);
955 cmd.data0 = (u32) bytes;
956 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_INTRQ_SIZE, &cmd, 0);
957
958 /*
959 * Even though we already know how many slices are supported
960 * via myri10ge_probe_slices() MXGEFW_CMD_GET_MAX_RSS_QUEUES
961 * has magic side effects, and must be called after a reset.
962 * It must be called prior to calling any RSS related cmds,
963 * including assigning an interrupt queue for anything but
964 * slice 0. It must also be called *after*
965 * MXGEFW_CMD_SET_INTRQ_SIZE, since the intrq size is used by
966 * the firmware to compute offsets.
967 */
968
969 if (mgp->num_slices > 1) {
970
971 /* ask the maximum number of slices it supports */
972 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_MAX_RSS_QUEUES,
973 &cmd, 0);
974 if (status != 0) {
975 dev_err(&mgp->pdev->dev,
976 "failed to get number of slices\n");
977 }
978
979 /*
980 * MXGEFW_CMD_ENABLE_RSS_QUEUES must be called prior
981 * to setting up the interrupt queue DMA
982 */
983
984 cmd.data0 = mgp->num_slices;
985 cmd.data1 = MXGEFW_SLICE_INTR_MODE_ONE_PER_SLICE;
986 if (mgp->dev->real_num_tx_queues > 1)
987 cmd.data1 |= MXGEFW_SLICE_ENABLE_MULTIPLE_TX_QUEUES;
988 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_ENABLE_RSS_QUEUES,
989 &cmd, 0);
990
991 /* Firmware older than 1.4.32 only supports multiple
992 * RX queues, so if we get an error, first retry using a
993 * single TX queue before giving up */
994 if (status != 0 && mgp->dev->real_num_tx_queues > 1) {
995 netif_set_real_num_tx_queues(mgp->dev, 1);
996 cmd.data0 = mgp->num_slices;
997 cmd.data1 = MXGEFW_SLICE_INTR_MODE_ONE_PER_SLICE;
998 status = myri10ge_send_cmd(mgp,
999 MXGEFW_CMD_ENABLE_RSS_QUEUES,
1000 &cmd, 0);
1001 }
1002
1003 if (status != 0) {
1004 dev_err(&mgp->pdev->dev,
1005 "failed to set number of slices\n");
1006
1007 return status;
1008 }
1009 }
1010 for (i = 0; i < mgp->num_slices; i++) {
1011 ss = &mgp->ss[i];
1012 cmd.data0 = MYRI10GE_LOWPART_TO_U32(ss->rx_done.bus);
1013 cmd.data1 = MYRI10GE_HIGHPART_TO_U32(ss->rx_done.bus);
1014 cmd.data2 = i;
1015 status |= myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_INTRQ_DMA,
1016 &cmd, 0);
1017 }
1018
1019 status |=
1020 myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_IRQ_ACK_OFFSET, &cmd, 0);
1021 for (i = 0; i < mgp->num_slices; i++) {
1022 ss = &mgp->ss[i];
1023 ss->irq_claim =
1024 (__iomem __be32 *) (mgp->sram + cmd.data0 + 8 * i);
1025 }
1026 status |= myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_IRQ_DEASSERT_OFFSET,
1027 &cmd, 0);
1028 mgp->irq_deassert = (__iomem __be32 *) (mgp->sram + cmd.data0);
1029
1030 status |= myri10ge_send_cmd
1031 (mgp, MXGEFW_CMD_GET_INTR_COAL_DELAY_OFFSET, &cmd, 0);
1032 mgp->intr_coal_delay_ptr = (__iomem __be32 *) (mgp->sram + cmd.data0);
1033 if (status != 0) {
1034 dev_err(&mgp->pdev->dev, "failed set interrupt parameters\n");
1035 return status;
1036 }
1037 put_be32(htonl(mgp->intr_coal_delay), mgp->intr_coal_delay_ptr);
1038
1039 #ifdef CONFIG_MYRI10GE_DCA
1040 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_DCA_OFFSET, &cmd, 0);
1041 dca_tag_off = cmd.data0;
1042 for (i = 0; i < mgp->num_slices; i++) {
1043 ss = &mgp->ss[i];
1044 if (status == 0) {
1045 ss->dca_tag = (__iomem __be32 *)
1046 (mgp->sram + dca_tag_off + 4 * i);
1047 } else {
1048 ss->dca_tag = NULL;
1049 }
1050 }
1051 #endif /* CONFIG_MYRI10GE_DCA */
1052
1053 /* reset mcp/driver shared state back to 0 */
1054
1055 mgp->link_changes = 0;
1056 for (i = 0; i < mgp->num_slices; i++) {
1057 ss = &mgp->ss[i];
1058
1059 memset(ss->rx_done.entry, 0, bytes);
1060 ss->tx.req = 0;
1061 ss->tx.done = 0;
1062 ss->tx.pkt_start = 0;
1063 ss->tx.pkt_done = 0;
1064 ss->rx_big.cnt = 0;
1065 ss->rx_small.cnt = 0;
1066 ss->rx_done.idx = 0;
1067 ss->rx_done.cnt = 0;
1068 ss->tx.wake_queue = 0;
1069 ss->tx.stop_queue = 0;
1070 }
1071
1072 status = myri10ge_update_mac_address(mgp, mgp->dev->dev_addr);
1073 myri10ge_change_pause(mgp, mgp->pause);
1074 myri10ge_set_multicast_list(mgp->dev);
1075 return status;
1076 }
1077
1078 #ifdef CONFIG_MYRI10GE_DCA
1079 static int myri10ge_toggle_relaxed(struct pci_dev *pdev, int on)
1080 {
1081 int ret, cap, err;
1082 u16 ctl;
1083
1084 cap = pci_pcie_cap(pdev);
1085 if (!cap)
1086 return 0;
1087
1088 err = pci_read_config_word(pdev, cap + PCI_EXP_DEVCTL, &ctl);
1089 if (err)
1090 return 0;
1091
1092 ret = (ctl & PCI_EXP_DEVCTL_RELAX_EN) >> 4;
1093 if (ret != on) {
1094 ctl &= ~PCI_EXP_DEVCTL_RELAX_EN;
1095 ctl |= (on << 4);
1096 pci_write_config_word(pdev, cap + PCI_EXP_DEVCTL, ctl);
1097 }
1098 return ret;
1099 }
1100
1101 static void
1102 myri10ge_write_dca(struct myri10ge_slice_state *ss, int cpu, int tag)
1103 {
1104 ss->cached_dca_tag = tag;
1105 put_be32(htonl(tag), ss->dca_tag);
1106 }
1107
1108 static inline void myri10ge_update_dca(struct myri10ge_slice_state *ss)
1109 {
1110 int cpu = get_cpu();
1111 int tag;
1112
1113 if (cpu != ss->cpu) {
1114 tag = dca3_get_tag(&ss->mgp->pdev->dev, cpu);
1115 if (ss->cached_dca_tag != tag)
1116 myri10ge_write_dca(ss, cpu, tag);
1117 ss->cpu = cpu;
1118 }
1119 put_cpu();
1120 }
1121
1122 static void myri10ge_setup_dca(struct myri10ge_priv *mgp)
1123 {
1124 int err, i;
1125 struct pci_dev *pdev = mgp->pdev;
1126
1127 if (mgp->ss[0].dca_tag == NULL || mgp->dca_enabled)
1128 return;
1129 if (!myri10ge_dca) {
1130 dev_err(&pdev->dev, "dca disabled by administrator\n");
1131 return;
1132 }
1133 err = dca_add_requester(&pdev->dev);
1134 if (err) {
1135 if (err != -ENODEV)
1136 dev_err(&pdev->dev,
1137 "dca_add_requester() failed, err=%d\n", err);
1138 return;
1139 }
1140 mgp->relaxed_order = myri10ge_toggle_relaxed(pdev, 0);
1141 mgp->dca_enabled = 1;
1142 for (i = 0; i < mgp->num_slices; i++) {
1143 mgp->ss[i].cpu = -1;
1144 mgp->ss[i].cached_dca_tag = -1;
1145 myri10ge_update_dca(&mgp->ss[i]);
1146 }
1147 }
1148
1149 static void myri10ge_teardown_dca(struct myri10ge_priv *mgp)
1150 {
1151 struct pci_dev *pdev = mgp->pdev;
1152
1153 if (!mgp->dca_enabled)
1154 return;
1155 mgp->dca_enabled = 0;
1156 if (mgp->relaxed_order)
1157 myri10ge_toggle_relaxed(pdev, 1);
1158 dca_remove_requester(&pdev->dev);
1159 }
1160
1161 static int myri10ge_notify_dca_device(struct device *dev, void *data)
1162 {
1163 struct myri10ge_priv *mgp;
1164 unsigned long event;
1165
1166 mgp = dev_get_drvdata(dev);
1167 event = *(unsigned long *)data;
1168
1169 if (event == DCA_PROVIDER_ADD)
1170 myri10ge_setup_dca(mgp);
1171 else if (event == DCA_PROVIDER_REMOVE)
1172 myri10ge_teardown_dca(mgp);
1173 return 0;
1174 }
1175 #endif /* CONFIG_MYRI10GE_DCA */
1176
1177 static inline void
1178 myri10ge_submit_8rx(struct mcp_kreq_ether_recv __iomem * dst,
1179 struct mcp_kreq_ether_recv *src)
1180 {
1181 __be32 low;
1182
1183 low = src->addr_low;
1184 src->addr_low = htonl(DMA_BIT_MASK(32));
1185 myri10ge_pio_copy(dst, src, 4 * sizeof(*src));
1186 mb();
1187 myri10ge_pio_copy(dst + 4, src + 4, 4 * sizeof(*src));
1188 mb();
1189 src->addr_low = low;
1190 put_be32(low, &dst->addr_low);
1191 mb();
1192 }
1193
1194 static inline void myri10ge_vlan_ip_csum(struct sk_buff *skb, __wsum hw_csum)
1195 {
1196 struct vlan_hdr *vh = (struct vlan_hdr *)(skb->data);
1197
1198 if ((skb->protocol == htons(ETH_P_8021Q)) &&
1199 (vh->h_vlan_encapsulated_proto == htons(ETH_P_IP) ||
1200 vh->h_vlan_encapsulated_proto == htons(ETH_P_IPV6))) {
1201 skb->csum = hw_csum;
1202 skb->ip_summed = CHECKSUM_COMPLETE;
1203 }
1204 }
1205
1206 static inline void
1207 myri10ge_rx_skb_build(struct sk_buff *skb, u8 * va,
1208 struct skb_frag_struct *rx_frags, int len, int hlen)
1209 {
1210 struct skb_frag_struct *skb_frags;
1211
1212 skb->len = skb->data_len = len;
1213 skb->truesize = len + sizeof(struct sk_buff);
1214 /* attach the page(s) */
1215
1216 skb_frags = skb_shinfo(skb)->frags;
1217 while (len > 0) {
1218 memcpy(skb_frags, rx_frags, sizeof(*skb_frags));
1219 len -= rx_frags->size;
1220 skb_frags++;
1221 rx_frags++;
1222 skb_shinfo(skb)->nr_frags++;
1223 }
1224
1225 /* pskb_may_pull is not available in irq context, but
1226 * skb_pull() (for ether_pad and eth_type_trans()) requires
1227 * the beginning of the packet in skb_headlen(), move it
1228 * manually */
1229 skb_copy_to_linear_data(skb, va, hlen);
1230 skb_shinfo(skb)->frags[0].page_offset += hlen;
1231 skb_shinfo(skb)->frags[0].size -= hlen;
1232 skb->data_len -= hlen;
1233 skb->tail += hlen;
1234 skb_pull(skb, MXGEFW_PAD);
1235 }
1236
1237 static void
1238 myri10ge_alloc_rx_pages(struct myri10ge_priv *mgp, struct myri10ge_rx_buf *rx,
1239 int bytes, int watchdog)
1240 {
1241 struct page *page;
1242 int idx;
1243 #if MYRI10GE_ALLOC_SIZE > 4096
1244 int end_offset;
1245 #endif
1246
1247 if (unlikely(rx->watchdog_needed && !watchdog))
1248 return;
1249
1250 /* try to refill entire ring */
1251 while (rx->fill_cnt != (rx->cnt + rx->mask + 1)) {
1252 idx = rx->fill_cnt & rx->mask;
1253 if (rx->page_offset + bytes <= MYRI10GE_ALLOC_SIZE) {
1254 /* we can use part of previous page */
1255 get_page(rx->page);
1256 } else {
1257 /* we need a new page */
1258 page =
1259 alloc_pages(GFP_ATOMIC | __GFP_COMP,
1260 MYRI10GE_ALLOC_ORDER);
1261 if (unlikely(page == NULL)) {
1262 if (rx->fill_cnt - rx->cnt < 16)
1263 rx->watchdog_needed = 1;
1264 return;
1265 }
1266 rx->page = page;
1267 rx->page_offset = 0;
1268 rx->bus = pci_map_page(mgp->pdev, page, 0,
1269 MYRI10GE_ALLOC_SIZE,
1270 PCI_DMA_FROMDEVICE);
1271 }
1272 rx->info[idx].page = rx->page;
1273 rx->info[idx].page_offset = rx->page_offset;
1274 /* note that this is the address of the start of the
1275 * page */
1276 dma_unmap_addr_set(&rx->info[idx], bus, rx->bus);
1277 rx->shadow[idx].addr_low =
1278 htonl(MYRI10GE_LOWPART_TO_U32(rx->bus) + rx->page_offset);
1279 rx->shadow[idx].addr_high =
1280 htonl(MYRI10GE_HIGHPART_TO_U32(rx->bus));
1281
1282 /* start next packet on a cacheline boundary */
1283 rx->page_offset += SKB_DATA_ALIGN(bytes);
1284
1285 #if MYRI10GE_ALLOC_SIZE > 4096
1286 /* don't cross a 4KB boundary */
1287 end_offset = rx->page_offset + bytes - 1;
1288 if ((unsigned)(rx->page_offset ^ end_offset) > 4095)
1289 rx->page_offset = end_offset & ~4095;
1290 #endif
1291 rx->fill_cnt++;
1292
1293 /* copy 8 descriptors to the firmware at a time */
1294 if ((idx & 7) == 7) {
1295 myri10ge_submit_8rx(&rx->lanai[idx - 7],
1296 &rx->shadow[idx - 7]);
1297 }
1298 }
1299 }
1300
1301 static inline void
1302 myri10ge_unmap_rx_page(struct pci_dev *pdev,
1303 struct myri10ge_rx_buffer_state *info, int bytes)
1304 {
1305 /* unmap the recvd page if we're the only or last user of it */
1306 if (bytes >= MYRI10GE_ALLOC_SIZE / 2 ||
1307 (info->page_offset + 2 * bytes) > MYRI10GE_ALLOC_SIZE) {
1308 pci_unmap_page(pdev, (dma_unmap_addr(info, bus)
1309 & ~(MYRI10GE_ALLOC_SIZE - 1)),
1310 MYRI10GE_ALLOC_SIZE, PCI_DMA_FROMDEVICE);
1311 }
1312 }
1313
1314 #define MYRI10GE_HLEN 64 /* The number of bytes to copy from a
1315 * page into an skb */
1316
1317 static inline int
1318 myri10ge_rx_done(struct myri10ge_slice_state *ss, int len, __wsum csum,
1319 bool lro_enabled)
1320 {
1321 struct myri10ge_priv *mgp = ss->mgp;
1322 struct sk_buff *skb;
1323 struct skb_frag_struct rx_frags[MYRI10GE_MAX_FRAGS_PER_FRAME];
1324 struct myri10ge_rx_buf *rx;
1325 int i, idx, hlen, remainder, bytes;
1326 struct pci_dev *pdev = mgp->pdev;
1327 struct net_device *dev = mgp->dev;
1328 u8 *va;
1329
1330 if (len <= mgp->small_bytes) {
1331 rx = &ss->rx_small;
1332 bytes = mgp->small_bytes;
1333 } else {
1334 rx = &ss->rx_big;
1335 bytes = mgp->big_bytes;
1336 }
1337
1338 len += MXGEFW_PAD;
1339 idx = rx->cnt & rx->mask;
1340 va = page_address(rx->info[idx].page) + rx->info[idx].page_offset;
1341 prefetch(va);
1342 /* Fill skb_frag_struct(s) with data from our receive */
1343 for (i = 0, remainder = len; remainder > 0; i++) {
1344 myri10ge_unmap_rx_page(pdev, &rx->info[idx], bytes);
1345 rx_frags[i].page = rx->info[idx].page;
1346 rx_frags[i].page_offset = rx->info[idx].page_offset;
1347 if (remainder < MYRI10GE_ALLOC_SIZE)
1348 rx_frags[i].size = remainder;
1349 else
1350 rx_frags[i].size = MYRI10GE_ALLOC_SIZE;
1351 rx->cnt++;
1352 idx = rx->cnt & rx->mask;
1353 remainder -= MYRI10GE_ALLOC_SIZE;
1354 }
1355
1356 if (lro_enabled) {
1357 rx_frags[0].page_offset += MXGEFW_PAD;
1358 rx_frags[0].size -= MXGEFW_PAD;
1359 len -= MXGEFW_PAD;
1360 lro_receive_frags(&ss->rx_done.lro_mgr, rx_frags,
1361 /* opaque, will come back in get_frag_header */
1362 len, len,
1363 (void *)(__force unsigned long)csum, csum);
1364
1365 return 1;
1366 }
1367
1368 hlen = MYRI10GE_HLEN > len ? len : MYRI10GE_HLEN;
1369
1370 /* allocate an skb to attach the page(s) to. This is done
1371 * after trying LRO, so as to avoid skb allocation overheads */
1372
1373 skb = netdev_alloc_skb(dev, MYRI10GE_HLEN + 16);
1374 if (unlikely(skb == NULL)) {
1375 ss->stats.rx_dropped++;
1376 do {
1377 i--;
1378 put_page(rx_frags[i].page);
1379 } while (i != 0);
1380 return 0;
1381 }
1382
1383 /* Attach the pages to the skb, and trim off any padding */
1384 myri10ge_rx_skb_build(skb, va, rx_frags, len, hlen);
1385 if (skb_shinfo(skb)->frags[0].size <= 0) {
1386 put_page(skb_shinfo(skb)->frags[0].page);
1387 skb_shinfo(skb)->nr_frags = 0;
1388 }
1389 skb->protocol = eth_type_trans(skb, dev);
1390 skb_record_rx_queue(skb, ss - &mgp->ss[0]);
1391
1392 if (dev->features & NETIF_F_RXCSUM) {
1393 if ((skb->protocol == htons(ETH_P_IP)) ||
1394 (skb->protocol == htons(ETH_P_IPV6))) {
1395 skb->csum = csum;
1396 skb->ip_summed = CHECKSUM_COMPLETE;
1397 } else
1398 myri10ge_vlan_ip_csum(skb, csum);
1399 }
1400 netif_receive_skb(skb);
1401 return 1;
1402 }
1403
1404 static inline void
1405 myri10ge_tx_done(struct myri10ge_slice_state *ss, int mcp_index)
1406 {
1407 struct pci_dev *pdev = ss->mgp->pdev;
1408 struct myri10ge_tx_buf *tx = &ss->tx;
1409 struct netdev_queue *dev_queue;
1410 struct sk_buff *skb;
1411 int idx, len;
1412
1413 while (tx->pkt_done != mcp_index) {
1414 idx = tx->done & tx->mask;
1415 skb = tx->info[idx].skb;
1416
1417 /* Mark as free */
1418 tx->info[idx].skb = NULL;
1419 if (tx->info[idx].last) {
1420 tx->pkt_done++;
1421 tx->info[idx].last = 0;
1422 }
1423 tx->done++;
1424 len = dma_unmap_len(&tx->info[idx], len);
1425 dma_unmap_len_set(&tx->info[idx], len, 0);
1426 if (skb) {
1427 ss->stats.tx_bytes += skb->len;
1428 ss->stats.tx_packets++;
1429 dev_kfree_skb_irq(skb);
1430 if (len)
1431 pci_unmap_single(pdev,
1432 dma_unmap_addr(&tx->info[idx],
1433 bus), len,
1434 PCI_DMA_TODEVICE);
1435 } else {
1436 if (len)
1437 pci_unmap_page(pdev,
1438 dma_unmap_addr(&tx->info[idx],
1439 bus), len,
1440 PCI_DMA_TODEVICE);
1441 }
1442 }
1443
1444 dev_queue = netdev_get_tx_queue(ss->dev, ss - ss->mgp->ss);
1445 /*
1446 * Make a minimal effort to prevent the NIC from polling an
1447 * idle tx queue. If we can't get the lock we leave the queue
1448 * active. In this case, either a thread was about to start
1449 * using the queue anyway, or we lost a race and the NIC will
1450 * waste some of its resources polling an inactive queue for a
1451 * while.
1452 */
1453
1454 if ((ss->mgp->dev->real_num_tx_queues > 1) &&
1455 __netif_tx_trylock(dev_queue)) {
1456 if (tx->req == tx->done) {
1457 tx->queue_active = 0;
1458 put_be32(htonl(1), tx->send_stop);
1459 mb();
1460 mmiowb();
1461 }
1462 __netif_tx_unlock(dev_queue);
1463 }
1464
1465 /* start the queue if we've stopped it */
1466 if (netif_tx_queue_stopped(dev_queue) &&
1467 tx->req - tx->done < (tx->mask >> 1) &&
1468 ss->mgp->running == MYRI10GE_ETH_RUNNING) {
1469 tx->wake_queue++;
1470 netif_tx_wake_queue(dev_queue);
1471 }
1472 }
1473
1474 static inline int
1475 myri10ge_clean_rx_done(struct myri10ge_slice_state *ss, int budget)
1476 {
1477 struct myri10ge_rx_done *rx_done = &ss->rx_done;
1478 struct myri10ge_priv *mgp = ss->mgp;
1479 unsigned long rx_bytes = 0;
1480 unsigned long rx_packets = 0;
1481 unsigned long rx_ok;
1482 int idx = rx_done->idx;
1483 int cnt = rx_done->cnt;
1484 int work_done = 0;
1485 u16 length;
1486 __wsum checksum;
1487
1488 /*
1489 * Prevent compiler from generating more than one ->features memory
1490 * access to avoid theoretical race condition with functions that
1491 * change NETIF_F_LRO flag at runtime.
1492 */
1493 bool lro_enabled = ACCESS_ONCE(mgp->dev->features) & NETIF_F_LRO;
1494
1495 while (rx_done->entry[idx].length != 0 && work_done < budget) {
1496 length = ntohs(rx_done->entry[idx].length);
1497 rx_done->entry[idx].length = 0;
1498 checksum = csum_unfold(rx_done->entry[idx].checksum);
1499 rx_ok = myri10ge_rx_done(ss, length, checksum, lro_enabled);
1500 rx_packets += rx_ok;
1501 rx_bytes += rx_ok * (unsigned long)length;
1502 cnt++;
1503 idx = cnt & (mgp->max_intr_slots - 1);
1504 work_done++;
1505 }
1506 rx_done->idx = idx;
1507 rx_done->cnt = cnt;
1508 ss->stats.rx_packets += rx_packets;
1509 ss->stats.rx_bytes += rx_bytes;
1510
1511 if (lro_enabled)
1512 lro_flush_all(&rx_done->lro_mgr);
1513
1514 /* restock receive rings if needed */
1515 if (ss->rx_small.fill_cnt - ss->rx_small.cnt < myri10ge_fill_thresh)
1516 myri10ge_alloc_rx_pages(mgp, &ss->rx_small,
1517 mgp->small_bytes + MXGEFW_PAD, 0);
1518 if (ss->rx_big.fill_cnt - ss->rx_big.cnt < myri10ge_fill_thresh)
1519 myri10ge_alloc_rx_pages(mgp, &ss->rx_big, mgp->big_bytes, 0);
1520
1521 return work_done;
1522 }
1523
1524 static inline void myri10ge_check_statblock(struct myri10ge_priv *mgp)
1525 {
1526 struct mcp_irq_data *stats = mgp->ss[0].fw_stats;
1527
1528 if (unlikely(stats->stats_updated)) {
1529 unsigned link_up = ntohl(stats->link_up);
1530 if (mgp->link_state != link_up) {
1531 mgp->link_state = link_up;
1532
1533 if (mgp->link_state == MXGEFW_LINK_UP) {
1534 netif_info(mgp, link, mgp->dev, "link up\n");
1535 netif_carrier_on(mgp->dev);
1536 mgp->link_changes++;
1537 } else {
1538 netif_info(mgp, link, mgp->dev, "link %s\n",
1539 (link_up == MXGEFW_LINK_MYRINET ?
1540 "mismatch (Myrinet detected)" :
1541 "down"));
1542 netif_carrier_off(mgp->dev);
1543 mgp->link_changes++;
1544 }
1545 }
1546 if (mgp->rdma_tags_available !=
1547 ntohl(stats->rdma_tags_available)) {
1548 mgp->rdma_tags_available =
1549 ntohl(stats->rdma_tags_available);
1550 netdev_warn(mgp->dev, "RDMA timed out! %d tags left\n",
1551 mgp->rdma_tags_available);
1552 }
1553 mgp->down_cnt += stats->link_down;
1554 if (stats->link_down)
1555 wake_up(&mgp->down_wq);
1556 }
1557 }
1558
1559 static int myri10ge_poll(struct napi_struct *napi, int budget)
1560 {
1561 struct myri10ge_slice_state *ss =
1562 container_of(napi, struct myri10ge_slice_state, napi);
1563 int work_done;
1564
1565 #ifdef CONFIG_MYRI10GE_DCA
1566 if (ss->mgp->dca_enabled)
1567 myri10ge_update_dca(ss);
1568 #endif
1569
1570 /* process as many rx events as NAPI will allow */
1571 work_done = myri10ge_clean_rx_done(ss, budget);
1572
1573 if (work_done < budget) {
1574 napi_complete(napi);
1575 put_be32(htonl(3), ss->irq_claim);
1576 }
1577 return work_done;
1578 }
1579
1580 static irqreturn_t myri10ge_intr(int irq, void *arg)
1581 {
1582 struct myri10ge_slice_state *ss = arg;
1583 struct myri10ge_priv *mgp = ss->mgp;
1584 struct mcp_irq_data *stats = ss->fw_stats;
1585 struct myri10ge_tx_buf *tx = &ss->tx;
1586 u32 send_done_count;
1587 int i;
1588
1589 /* an interrupt on a non-zero receive-only slice is implicitly
1590 * valid since MSI-X irqs are not shared */
1591 if ((mgp->dev->real_num_tx_queues == 1) && (ss != mgp->ss)) {
1592 napi_schedule(&ss->napi);
1593 return IRQ_HANDLED;
1594 }
1595
1596 /* make sure it is our IRQ, and that the DMA has finished */
1597 if (unlikely(!stats->valid))
1598 return IRQ_NONE;
1599
1600 /* low bit indicates receives are present, so schedule
1601 * napi poll handler */
1602 if (stats->valid & 1)
1603 napi_schedule(&ss->napi);
1604
1605 if (!mgp->msi_enabled && !mgp->msix_enabled) {
1606 put_be32(0, mgp->irq_deassert);
1607 if (!myri10ge_deassert_wait)
1608 stats->valid = 0;
1609 mb();
1610 } else
1611 stats->valid = 0;
1612
1613 /* Wait for IRQ line to go low, if using INTx */
1614 i = 0;
1615 while (1) {
1616 i++;
1617 /* check for transmit completes and receives */
1618 send_done_count = ntohl(stats->send_done_count);
1619 if (send_done_count != tx->pkt_done)
1620 myri10ge_tx_done(ss, (int)send_done_count);
1621 if (unlikely(i > myri10ge_max_irq_loops)) {
1622 netdev_warn(mgp->dev, "irq stuck?\n");
1623 stats->valid = 0;
1624 schedule_work(&mgp->watchdog_work);
1625 }
1626 if (likely(stats->valid == 0))
1627 break;
1628 cpu_relax();
1629 barrier();
1630 }
1631
1632 /* Only slice 0 updates stats */
1633 if (ss == mgp->ss)
1634 myri10ge_check_statblock(mgp);
1635
1636 put_be32(htonl(3), ss->irq_claim + 1);
1637 return IRQ_HANDLED;
1638 }
1639
1640 static int
1641 myri10ge_get_settings(struct net_device *netdev, struct ethtool_cmd *cmd)
1642 {
1643 struct myri10ge_priv *mgp = netdev_priv(netdev);
1644 char *ptr;
1645 int i;
1646
1647 cmd->autoneg = AUTONEG_DISABLE;
1648 ethtool_cmd_speed_set(cmd, SPEED_10000);
1649 cmd->duplex = DUPLEX_FULL;
1650
1651 /*
1652 * parse the product code to deterimine the interface type
1653 * (CX4, XFP, Quad Ribbon Fiber) by looking at the character
1654 * after the 3rd dash in the driver's cached copy of the
1655 * EEPROM's product code string.
1656 */
1657 ptr = mgp->product_code_string;
1658 if (ptr == NULL) {
1659 netdev_err(netdev, "Missing product code\n");
1660 return 0;
1661 }
1662 for (i = 0; i < 3; i++, ptr++) {
1663 ptr = strchr(ptr, '-');
1664 if (ptr == NULL) {
1665 netdev_err(netdev, "Invalid product code %s\n",
1666 mgp->product_code_string);
1667 return 0;
1668 }
1669 }
1670 if (*ptr == '2')
1671 ptr++;
1672 if (*ptr == 'R' || *ptr == 'Q' || *ptr == 'S') {
1673 /* We've found either an XFP, quad ribbon fiber, or SFP+ */
1674 cmd->port = PORT_FIBRE;
1675 cmd->supported |= SUPPORTED_FIBRE;
1676 cmd->advertising |= ADVERTISED_FIBRE;
1677 } else {
1678 cmd->port = PORT_OTHER;
1679 }
1680 if (*ptr == 'R' || *ptr == 'S')
1681 cmd->transceiver = XCVR_EXTERNAL;
1682 else
1683 cmd->transceiver = XCVR_INTERNAL;
1684
1685 return 0;
1686 }
1687
1688 static void
1689 myri10ge_get_drvinfo(struct net_device *netdev, struct ethtool_drvinfo *info)
1690 {
1691 struct myri10ge_priv *mgp = netdev_priv(netdev);
1692
1693 strlcpy(info->driver, "myri10ge", sizeof(info->driver));
1694 strlcpy(info->version, MYRI10GE_VERSION_STR, sizeof(info->version));
1695 strlcpy(info->fw_version, mgp->fw_version, sizeof(info->fw_version));
1696 strlcpy(info->bus_info, pci_name(mgp->pdev), sizeof(info->bus_info));
1697 }
1698
1699 static int
1700 myri10ge_get_coalesce(struct net_device *netdev, struct ethtool_coalesce *coal)
1701 {
1702 struct myri10ge_priv *mgp = netdev_priv(netdev);
1703
1704 coal->rx_coalesce_usecs = mgp->intr_coal_delay;
1705 return 0;
1706 }
1707
1708 static int
1709 myri10ge_set_coalesce(struct net_device *netdev, struct ethtool_coalesce *coal)
1710 {
1711 struct myri10ge_priv *mgp = netdev_priv(netdev);
1712
1713 mgp->intr_coal_delay = coal->rx_coalesce_usecs;
1714 put_be32(htonl(mgp->intr_coal_delay), mgp->intr_coal_delay_ptr);
1715 return 0;
1716 }
1717
1718 static void
1719 myri10ge_get_pauseparam(struct net_device *netdev,
1720 struct ethtool_pauseparam *pause)
1721 {
1722 struct myri10ge_priv *mgp = netdev_priv(netdev);
1723
1724 pause->autoneg = 0;
1725 pause->rx_pause = mgp->pause;
1726 pause->tx_pause = mgp->pause;
1727 }
1728
1729 static int
1730 myri10ge_set_pauseparam(struct net_device *netdev,
1731 struct ethtool_pauseparam *pause)
1732 {
1733 struct myri10ge_priv *mgp = netdev_priv(netdev);
1734
1735 if (pause->tx_pause != mgp->pause)
1736 return myri10ge_change_pause(mgp, pause->tx_pause);
1737 if (pause->rx_pause != mgp->pause)
1738 return myri10ge_change_pause(mgp, pause->rx_pause);
1739 if (pause->autoneg != 0)
1740 return -EINVAL;
1741 return 0;
1742 }
1743
1744 static void
1745 myri10ge_get_ringparam(struct net_device *netdev,
1746 struct ethtool_ringparam *ring)
1747 {
1748 struct myri10ge_priv *mgp = netdev_priv(netdev);
1749
1750 ring->rx_mini_max_pending = mgp->ss[0].rx_small.mask + 1;
1751 ring->rx_max_pending = mgp->ss[0].rx_big.mask + 1;
1752 ring->rx_jumbo_max_pending = 0;
1753 ring->tx_max_pending = mgp->ss[0].tx.mask + 1;
1754 ring->rx_mini_pending = ring->rx_mini_max_pending;
1755 ring->rx_pending = ring->rx_max_pending;
1756 ring->rx_jumbo_pending = ring->rx_jumbo_max_pending;
1757 ring->tx_pending = ring->tx_max_pending;
1758 }
1759
1760 static const char myri10ge_gstrings_main_stats[][ETH_GSTRING_LEN] = {
1761 "rx_packets", "tx_packets", "rx_bytes", "tx_bytes", "rx_errors",
1762 "tx_errors", "rx_dropped", "tx_dropped", "multicast", "collisions",
1763 "rx_length_errors", "rx_over_errors", "rx_crc_errors",
1764 "rx_frame_errors", "rx_fifo_errors", "rx_missed_errors",
1765 "tx_aborted_errors", "tx_carrier_errors", "tx_fifo_errors",
1766 "tx_heartbeat_errors", "tx_window_errors",
1767 /* device-specific stats */
1768 "tx_boundary", "WC", "irq", "MSI", "MSIX",
1769 "read_dma_bw_MBs", "write_dma_bw_MBs", "read_write_dma_bw_MBs",
1770 "serial_number", "watchdog_resets",
1771 #ifdef CONFIG_MYRI10GE_DCA
1772 "dca_capable_firmware", "dca_device_present",
1773 #endif
1774 "link_changes", "link_up", "dropped_link_overflow",
1775 "dropped_link_error_or_filtered",
1776 "dropped_pause", "dropped_bad_phy", "dropped_bad_crc32",
1777 "dropped_unicast_filtered", "dropped_multicast_filtered",
1778 "dropped_runt", "dropped_overrun", "dropped_no_small_buffer",
1779 "dropped_no_big_buffer"
1780 };
1781
1782 static const char myri10ge_gstrings_slice_stats[][ETH_GSTRING_LEN] = {
1783 "----------- slice ---------",
1784 "tx_pkt_start", "tx_pkt_done", "tx_req", "tx_done",
1785 "rx_small_cnt", "rx_big_cnt",
1786 "wake_queue", "stop_queue", "tx_linearized",
1787 "LRO aggregated", "LRO flushed", "LRO avg aggr", "LRO no_desc",
1788 };
1789
1790 #define MYRI10GE_NET_STATS_LEN 21
1791 #define MYRI10GE_MAIN_STATS_LEN ARRAY_SIZE(myri10ge_gstrings_main_stats)
1792 #define MYRI10GE_SLICE_STATS_LEN ARRAY_SIZE(myri10ge_gstrings_slice_stats)
1793
1794 static void
1795 myri10ge_get_strings(struct net_device *netdev, u32 stringset, u8 * data)
1796 {
1797 struct myri10ge_priv *mgp = netdev_priv(netdev);
1798 int i;
1799
1800 switch (stringset) {
1801 case ETH_SS_STATS:
1802 memcpy(data, *myri10ge_gstrings_main_stats,
1803 sizeof(myri10ge_gstrings_main_stats));
1804 data += sizeof(myri10ge_gstrings_main_stats);
1805 for (i = 0; i < mgp->num_slices; i++) {
1806 memcpy(data, *myri10ge_gstrings_slice_stats,
1807 sizeof(myri10ge_gstrings_slice_stats));
1808 data += sizeof(myri10ge_gstrings_slice_stats);
1809 }
1810 break;
1811 }
1812 }
1813
1814 static int myri10ge_get_sset_count(struct net_device *netdev, int sset)
1815 {
1816 struct myri10ge_priv *mgp = netdev_priv(netdev);
1817
1818 switch (sset) {
1819 case ETH_SS_STATS:
1820 return MYRI10GE_MAIN_STATS_LEN +
1821 mgp->num_slices * MYRI10GE_SLICE_STATS_LEN;
1822 default:
1823 return -EOPNOTSUPP;
1824 }
1825 }
1826
1827 static void
1828 myri10ge_get_ethtool_stats(struct net_device *netdev,
1829 struct ethtool_stats *stats, u64 * data)
1830 {
1831 struct myri10ge_priv *mgp = netdev_priv(netdev);
1832 struct myri10ge_slice_state *ss;
1833 struct rtnl_link_stats64 link_stats;
1834 int slice;
1835 int i;
1836
1837 /* force stats update */
1838 memset(&link_stats, 0, sizeof(link_stats));
1839 (void)myri10ge_get_stats(netdev, &link_stats);
1840 for (i = 0; i < MYRI10GE_NET_STATS_LEN; i++)
1841 data[i] = ((u64 *)&link_stats)[i];
1842
1843 data[i++] = (unsigned int)mgp->tx_boundary;
1844 data[i++] = (unsigned int)mgp->wc_enabled;
1845 data[i++] = (unsigned int)mgp->pdev->irq;
1846 data[i++] = (unsigned int)mgp->msi_enabled;
1847 data[i++] = (unsigned int)mgp->msix_enabled;
1848 data[i++] = (unsigned int)mgp->read_dma;
1849 data[i++] = (unsigned int)mgp->write_dma;
1850 data[i++] = (unsigned int)mgp->read_write_dma;
1851 data[i++] = (unsigned int)mgp->serial_number;
1852 data[i++] = (unsigned int)mgp->watchdog_resets;
1853 #ifdef CONFIG_MYRI10GE_DCA
1854 data[i++] = (unsigned int)(mgp->ss[0].dca_tag != NULL);
1855 data[i++] = (unsigned int)(mgp->dca_enabled);
1856 #endif
1857 data[i++] = (unsigned int)mgp->link_changes;
1858
1859 /* firmware stats are useful only in the first slice */
1860 ss = &mgp->ss[0];
1861 data[i++] = (unsigned int)ntohl(ss->fw_stats->link_up);
1862 data[i++] = (unsigned int)ntohl(ss->fw_stats->dropped_link_overflow);
1863 data[i++] =
1864 (unsigned int)ntohl(ss->fw_stats->dropped_link_error_or_filtered);
1865 data[i++] = (unsigned int)ntohl(ss->fw_stats->dropped_pause);
1866 data[i++] = (unsigned int)ntohl(ss->fw_stats->dropped_bad_phy);
1867 data[i++] = (unsigned int)ntohl(ss->fw_stats->dropped_bad_crc32);
1868 data[i++] = (unsigned int)ntohl(ss->fw_stats->dropped_unicast_filtered);
1869 data[i++] =
1870 (unsigned int)ntohl(ss->fw_stats->dropped_multicast_filtered);
1871 data[i++] = (unsigned int)ntohl(ss->fw_stats->dropped_runt);
1872 data[i++] = (unsigned int)ntohl(ss->fw_stats->dropped_overrun);
1873 data[i++] = (unsigned int)ntohl(ss->fw_stats->dropped_no_small_buffer);
1874 data[i++] = (unsigned int)ntohl(ss->fw_stats->dropped_no_big_buffer);
1875
1876 for (slice = 0; slice < mgp->num_slices; slice++) {
1877 ss = &mgp->ss[slice];
1878 data[i++] = slice;
1879 data[i++] = (unsigned int)ss->tx.pkt_start;
1880 data[i++] = (unsigned int)ss->tx.pkt_done;
1881 data[i++] = (unsigned int)ss->tx.req;
1882 data[i++] = (unsigned int)ss->tx.done;
1883 data[i++] = (unsigned int)ss->rx_small.cnt;
1884 data[i++] = (unsigned int)ss->rx_big.cnt;
1885 data[i++] = (unsigned int)ss->tx.wake_queue;
1886 data[i++] = (unsigned int)ss->tx.stop_queue;
1887 data[i++] = (unsigned int)ss->tx.linearized;
1888 data[i++] = ss->rx_done.lro_mgr.stats.aggregated;
1889 data[i++] = ss->rx_done.lro_mgr.stats.flushed;
1890 if (ss->rx_done.lro_mgr.stats.flushed)
1891 data[i++] = ss->rx_done.lro_mgr.stats.aggregated /
1892 ss->rx_done.lro_mgr.stats.flushed;
1893 else
1894 data[i++] = 0;
1895 data[i++] = ss->rx_done.lro_mgr.stats.no_desc;
1896 }
1897 }
1898
1899 static void myri10ge_set_msglevel(struct net_device *netdev, u32 value)
1900 {
1901 struct myri10ge_priv *mgp = netdev_priv(netdev);
1902 mgp->msg_enable = value;
1903 }
1904
1905 static u32 myri10ge_get_msglevel(struct net_device *netdev)
1906 {
1907 struct myri10ge_priv *mgp = netdev_priv(netdev);
1908 return mgp->msg_enable;
1909 }
1910
1911 /*
1912 * Use a low-level command to change the LED behavior. Rather than
1913 * blinking (which is the normal case), when identify is used, the
1914 * yellow LED turns solid.
1915 */
1916 static int myri10ge_led(struct myri10ge_priv *mgp, int on)
1917 {
1918 struct mcp_gen_header *hdr;
1919 struct device *dev = &mgp->pdev->dev;
1920 size_t hdr_off, pattern_off, hdr_len;
1921 u32 pattern = 0xfffffffe;
1922
1923 /* find running firmware header */
1924 hdr_off = swab32(readl(mgp->sram + MCP_HEADER_PTR_OFFSET));
1925 if ((hdr_off & 3) || hdr_off + sizeof(*hdr) > mgp->sram_size) {
1926 dev_err(dev, "Running firmware has bad header offset (%d)\n",
1927 (int)hdr_off);
1928 return -EIO;
1929 }
1930 hdr_len = swab32(readl(mgp->sram + hdr_off +
1931 offsetof(struct mcp_gen_header, header_length)));
1932 pattern_off = hdr_off + offsetof(struct mcp_gen_header, led_pattern);
1933 if (pattern_off >= (hdr_len + hdr_off)) {
1934 dev_info(dev, "Firmware does not support LED identification\n");
1935 return -EINVAL;
1936 }
1937 if (!on)
1938 pattern = swab32(readl(mgp->sram + pattern_off + 4));
1939 writel(htonl(pattern), mgp->sram + pattern_off);
1940 return 0;
1941 }
1942
1943 static int
1944 myri10ge_phys_id(struct net_device *netdev, enum ethtool_phys_id_state state)
1945 {
1946 struct myri10ge_priv *mgp = netdev_priv(netdev);
1947 int rc;
1948
1949 switch (state) {
1950 case ETHTOOL_ID_ACTIVE:
1951 rc = myri10ge_led(mgp, 1);
1952 break;
1953
1954 case ETHTOOL_ID_INACTIVE:
1955 rc = myri10ge_led(mgp, 0);
1956 break;
1957
1958 default:
1959 rc = -EINVAL;
1960 }
1961
1962 return rc;
1963 }
1964
1965 static const struct ethtool_ops myri10ge_ethtool_ops = {
1966 .get_settings = myri10ge_get_settings,
1967 .get_drvinfo = myri10ge_get_drvinfo,
1968 .get_coalesce = myri10ge_get_coalesce,
1969 .set_coalesce = myri10ge_set_coalesce,
1970 .get_pauseparam = myri10ge_get_pauseparam,
1971 .set_pauseparam = myri10ge_set_pauseparam,
1972 .get_ringparam = myri10ge_get_ringparam,
1973 .get_link = ethtool_op_get_link,
1974 .get_strings = myri10ge_get_strings,
1975 .get_sset_count = myri10ge_get_sset_count,
1976 .get_ethtool_stats = myri10ge_get_ethtool_stats,
1977 .set_msglevel = myri10ge_set_msglevel,
1978 .get_msglevel = myri10ge_get_msglevel,
1979 .set_phys_id = myri10ge_phys_id,
1980 };
1981
1982 static int myri10ge_allocate_rings(struct myri10ge_slice_state *ss)
1983 {
1984 struct myri10ge_priv *mgp = ss->mgp;
1985 struct myri10ge_cmd cmd;
1986 struct net_device *dev = mgp->dev;
1987 int tx_ring_size, rx_ring_size;
1988 int tx_ring_entries, rx_ring_entries;
1989 int i, slice, status;
1990 size_t bytes;
1991
1992 /* get ring sizes */
1993 slice = ss - mgp->ss;
1994 cmd.data0 = slice;
1995 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_SEND_RING_SIZE, &cmd, 0);
1996 tx_ring_size = cmd.data0;
1997 cmd.data0 = slice;
1998 status |= myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_RX_RING_SIZE, &cmd, 0);
1999 if (status != 0)
2000 return status;
2001 rx_ring_size = cmd.data0;
2002
2003 tx_ring_entries = tx_ring_size / sizeof(struct mcp_kreq_ether_send);
2004 rx_ring_entries = rx_ring_size / sizeof(struct mcp_dma_addr);
2005 ss->tx.mask = tx_ring_entries - 1;
2006 ss->rx_small.mask = ss->rx_big.mask = rx_ring_entries - 1;
2007
2008 status = -ENOMEM;
2009
2010 /* allocate the host shadow rings */
2011
2012 bytes = 8 + (MYRI10GE_MAX_SEND_DESC_TSO + 4)
2013 * sizeof(*ss->tx.req_list);
2014 ss->tx.req_bytes = kzalloc(bytes, GFP_KERNEL);
2015 if (ss->tx.req_bytes == NULL)
2016 goto abort_with_nothing;
2017
2018 /* ensure req_list entries are aligned to 8 bytes */
2019 ss->tx.req_list = (struct mcp_kreq_ether_send *)
2020 ALIGN((unsigned long)ss->tx.req_bytes, 8);
2021 ss->tx.queue_active = 0;
2022
2023 bytes = rx_ring_entries * sizeof(*ss->rx_small.shadow);
2024 ss->rx_small.shadow = kzalloc(bytes, GFP_KERNEL);
2025 if (ss->rx_small.shadow == NULL)
2026 goto abort_with_tx_req_bytes;
2027
2028 bytes = rx_ring_entries * sizeof(*ss->rx_big.shadow);
2029 ss->rx_big.shadow = kzalloc(bytes, GFP_KERNEL);
2030 if (ss->rx_big.shadow == NULL)
2031 goto abort_with_rx_small_shadow;
2032
2033 /* allocate the host info rings */
2034
2035 bytes = tx_ring_entries * sizeof(*ss->tx.info);
2036 ss->tx.info = kzalloc(bytes, GFP_KERNEL);
2037 if (ss->tx.info == NULL)
2038 goto abort_with_rx_big_shadow;
2039
2040 bytes = rx_ring_entries * sizeof(*ss->rx_small.info);
2041 ss->rx_small.info = kzalloc(bytes, GFP_KERNEL);
2042 if (ss->rx_small.info == NULL)
2043 goto abort_with_tx_info;
2044
2045 bytes = rx_ring_entries * sizeof(*ss->rx_big.info);
2046 ss->rx_big.info = kzalloc(bytes, GFP_KERNEL);
2047 if (ss->rx_big.info == NULL)
2048 goto abort_with_rx_small_info;
2049
2050 /* Fill the receive rings */
2051 ss->rx_big.cnt = 0;
2052 ss->rx_small.cnt = 0;
2053 ss->rx_big.fill_cnt = 0;
2054 ss->rx_small.fill_cnt = 0;
2055 ss->rx_small.page_offset = MYRI10GE_ALLOC_SIZE;
2056 ss->rx_big.page_offset = MYRI10GE_ALLOC_SIZE;
2057 ss->rx_small.watchdog_needed = 0;
2058 ss->rx_big.watchdog_needed = 0;
2059 if (mgp->small_bytes == 0) {
2060 ss->rx_small.fill_cnt = ss->rx_small.mask + 1;
2061 } else {
2062 myri10ge_alloc_rx_pages(mgp, &ss->rx_small,
2063 mgp->small_bytes + MXGEFW_PAD, 0);
2064 }
2065
2066 if (ss->rx_small.fill_cnt < ss->rx_small.mask + 1) {
2067 netdev_err(dev, "slice-%d: alloced only %d small bufs\n",
2068 slice, ss->rx_small.fill_cnt);
2069 goto abort_with_rx_small_ring;
2070 }
2071
2072 myri10ge_alloc_rx_pages(mgp, &ss->rx_big, mgp->big_bytes, 0);
2073 if (ss->rx_big.fill_cnt < ss->rx_big.mask + 1) {
2074 netdev_err(dev, "slice-%d: alloced only %d big bufs\n",
2075 slice, ss->rx_big.fill_cnt);
2076 goto abort_with_rx_big_ring;
2077 }
2078
2079 return 0;
2080
2081 abort_with_rx_big_ring:
2082 for (i = ss->rx_big.cnt; i < ss->rx_big.fill_cnt; i++) {
2083 int idx = i & ss->rx_big.mask;
2084 myri10ge_unmap_rx_page(mgp->pdev, &ss->rx_big.info[idx],
2085 mgp->big_bytes);
2086 put_page(ss->rx_big.info[idx].page);
2087 }
2088
2089 abort_with_rx_small_ring:
2090 if (mgp->small_bytes == 0)
2091 ss->rx_small.fill_cnt = ss->rx_small.cnt;
2092 for (i = ss->rx_small.cnt; i < ss->rx_small.fill_cnt; i++) {
2093 int idx = i & ss->rx_small.mask;
2094 myri10ge_unmap_rx_page(mgp->pdev, &ss->rx_small.info[idx],
2095 mgp->small_bytes + MXGEFW_PAD);
2096 put_page(ss->rx_small.info[idx].page);
2097 }
2098
2099 kfree(ss->rx_big.info);
2100
2101 abort_with_rx_small_info:
2102 kfree(ss->rx_small.info);
2103
2104 abort_with_tx_info:
2105 kfree(ss->tx.info);
2106
2107 abort_with_rx_big_shadow:
2108 kfree(ss->rx_big.shadow);
2109
2110 abort_with_rx_small_shadow:
2111 kfree(ss->rx_small.shadow);
2112
2113 abort_with_tx_req_bytes:
2114 kfree(ss->tx.req_bytes);
2115 ss->tx.req_bytes = NULL;
2116 ss->tx.req_list = NULL;
2117
2118 abort_with_nothing:
2119 return status;
2120 }
2121
2122 static void myri10ge_free_rings(struct myri10ge_slice_state *ss)
2123 {
2124 struct myri10ge_priv *mgp = ss->mgp;
2125 struct sk_buff *skb;
2126 struct myri10ge_tx_buf *tx;
2127 int i, len, idx;
2128
2129 /* If not allocated, skip it */
2130 if (ss->tx.req_list == NULL)
2131 return;
2132
2133 for (i = ss->rx_big.cnt; i < ss->rx_big.fill_cnt; i++) {
2134 idx = i & ss->rx_big.mask;
2135 if (i == ss->rx_big.fill_cnt - 1)
2136 ss->rx_big.info[idx].page_offset = MYRI10GE_ALLOC_SIZE;
2137 myri10ge_unmap_rx_page(mgp->pdev, &ss->rx_big.info[idx],
2138 mgp->big_bytes);
2139 put_page(ss->rx_big.info[idx].page);
2140 }
2141
2142 if (mgp->small_bytes == 0)
2143 ss->rx_small.fill_cnt = ss->rx_small.cnt;
2144 for (i = ss->rx_small.cnt; i < ss->rx_small.fill_cnt; i++) {
2145 idx = i & ss->rx_small.mask;
2146 if (i == ss->rx_small.fill_cnt - 1)
2147 ss->rx_small.info[idx].page_offset =
2148 MYRI10GE_ALLOC_SIZE;
2149 myri10ge_unmap_rx_page(mgp->pdev, &ss->rx_small.info[idx],
2150 mgp->small_bytes + MXGEFW_PAD);
2151 put_page(ss->rx_small.info[idx].page);
2152 }
2153 tx = &ss->tx;
2154 while (tx->done != tx->req) {
2155 idx = tx->done & tx->mask;
2156 skb = tx->info[idx].skb;
2157
2158 /* Mark as free */
2159 tx->info[idx].skb = NULL;
2160 tx->done++;
2161 len = dma_unmap_len(&tx->info[idx], len);
2162 dma_unmap_len_set(&tx->info[idx], len, 0);
2163 if (skb) {
2164 ss->stats.tx_dropped++;
2165 dev_kfree_skb_any(skb);
2166 if (len)
2167 pci_unmap_single(mgp->pdev,
2168 dma_unmap_addr(&tx->info[idx],
2169 bus), len,
2170 PCI_DMA_TODEVICE);
2171 } else {
2172 if (len)
2173 pci_unmap_page(mgp->pdev,
2174 dma_unmap_addr(&tx->info[idx],
2175 bus), len,
2176 PCI_DMA_TODEVICE);
2177 }
2178 }
2179 kfree(ss->rx_big.info);
2180
2181 kfree(ss->rx_small.info);
2182
2183 kfree(ss->tx.info);
2184
2185 kfree(ss->rx_big.shadow);
2186
2187 kfree(ss->rx_small.shadow);
2188
2189 kfree(ss->tx.req_bytes);
2190 ss->tx.req_bytes = NULL;
2191 ss->tx.req_list = NULL;
2192 }
2193
2194 static int myri10ge_request_irq(struct myri10ge_priv *mgp)
2195 {
2196 struct pci_dev *pdev = mgp->pdev;
2197 struct myri10ge_slice_state *ss;
2198 struct net_device *netdev = mgp->dev;
2199 int i;
2200 int status;
2201
2202 mgp->msi_enabled = 0;
2203 mgp->msix_enabled = 0;
2204 status = 0;
2205 if (myri10ge_msi) {
2206 if (mgp->num_slices > 1) {
2207 status =
2208 pci_enable_msix(pdev, mgp->msix_vectors,
2209 mgp->num_slices);
2210 if (status == 0) {
2211 mgp->msix_enabled = 1;
2212 } else {
2213 dev_err(&pdev->dev,
2214 "Error %d setting up MSI-X\n", status);
2215 return status;
2216 }
2217 }
2218 if (mgp->msix_enabled == 0) {
2219 status = pci_enable_msi(pdev);
2220 if (status != 0) {
2221 dev_err(&pdev->dev,
2222 "Error %d setting up MSI; falling back to xPIC\n",
2223 status);
2224 } else {
2225 mgp->msi_enabled = 1;
2226 }
2227 }
2228 }
2229 if (mgp->msix_enabled) {
2230 for (i = 0; i < mgp->num_slices; i++) {
2231 ss = &mgp->ss[i];
2232 snprintf(ss->irq_desc, sizeof(ss->irq_desc),
2233 "%s:slice-%d", netdev->name, i);
2234 status = request_irq(mgp->msix_vectors[i].vector,
2235 myri10ge_intr, 0, ss->irq_desc,
2236 ss);
2237 if (status != 0) {
2238 dev_err(&pdev->dev,
2239 "slice %d failed to allocate IRQ\n", i);
2240 i--;
2241 while (i >= 0) {
2242 free_irq(mgp->msix_vectors[i].vector,
2243 &mgp->ss[i]);
2244 i--;
2245 }
2246 pci_disable_msix(pdev);
2247 return status;
2248 }
2249 }
2250 } else {
2251 status = request_irq(pdev->irq, myri10ge_intr, IRQF_SHARED,
2252 mgp->dev->name, &mgp->ss[0]);
2253 if (status != 0) {
2254 dev_err(&pdev->dev, "failed to allocate IRQ\n");
2255 if (mgp->msi_enabled)
2256 pci_disable_msi(pdev);
2257 }
2258 }
2259 return status;
2260 }
2261
2262 static void myri10ge_free_irq(struct myri10ge_priv *mgp)
2263 {
2264 struct pci_dev *pdev = mgp->pdev;
2265 int i;
2266
2267 if (mgp->msix_enabled) {
2268 for (i = 0; i < mgp->num_slices; i++)
2269 free_irq(mgp->msix_vectors[i].vector, &mgp->ss[i]);
2270 } else {
2271 free_irq(pdev->irq, &mgp->ss[0]);
2272 }
2273 if (mgp->msi_enabled)
2274 pci_disable_msi(pdev);
2275 if (mgp->msix_enabled)
2276 pci_disable_msix(pdev);
2277 }
2278
2279 static int
2280 myri10ge_get_frag_header(struct skb_frag_struct *frag, void **mac_hdr,
2281 void **ip_hdr, void **tcpudp_hdr,
2282 u64 * hdr_flags, void *priv)
2283 {
2284 struct ethhdr *eh;
2285 struct vlan_ethhdr *veh;
2286 struct iphdr *iph;
2287 u8 *va = page_address(frag->page) + frag->page_offset;
2288 unsigned long ll_hlen;
2289 /* passed opaque through lro_receive_frags() */
2290 __wsum csum = (__force __wsum) (unsigned long)priv;
2291
2292 /* find the mac header, aborting if not IPv4 */
2293
2294 eh = (struct ethhdr *)va;
2295 *mac_hdr = eh;
2296 ll_hlen = ETH_HLEN;
2297 if (eh->h_proto != htons(ETH_P_IP)) {
2298 if (eh->h_proto == htons(ETH_P_8021Q)) {
2299 veh = (struct vlan_ethhdr *)va;
2300 if (veh->h_vlan_encapsulated_proto != htons(ETH_P_IP))
2301 return -1;
2302
2303 ll_hlen += VLAN_HLEN;
2304
2305 /*
2306 * HW checksum starts ETH_HLEN bytes into
2307 * frame, so we must subtract off the VLAN
2308 * header's checksum before csum can be used
2309 */
2310 csum = csum_sub(csum, csum_partial(va + ETH_HLEN,
2311 VLAN_HLEN, 0));
2312 } else {
2313 return -1;
2314 }
2315 }
2316 *hdr_flags = LRO_IPV4;
2317
2318 iph = (struct iphdr *)(va + ll_hlen);
2319 *ip_hdr = iph;
2320 if (iph->protocol != IPPROTO_TCP)
2321 return -1;
2322 if (ip_is_fragment(iph))
2323 return -1;
2324 *hdr_flags |= LRO_TCP;
2325 *tcpudp_hdr = (u8 *) (*ip_hdr) + (iph->ihl << 2);
2326
2327 /* verify the IP checksum */
2328 if (unlikely(ip_fast_csum((u8 *) iph, iph->ihl)))
2329 return -1;
2330
2331 /* verify the checksum */
2332 if (unlikely(csum_tcpudp_magic(iph->saddr, iph->daddr,
2333 ntohs(iph->tot_len) - (iph->ihl << 2),
2334 IPPROTO_TCP, csum)))
2335 return -1;
2336
2337 return 0;
2338 }
2339
2340 static int myri10ge_get_txrx(struct myri10ge_priv *mgp, int slice)
2341 {
2342 struct myri10ge_cmd cmd;
2343 struct myri10ge_slice_state *ss;
2344 int status;
2345
2346 ss = &mgp->ss[slice];
2347 status = 0;
2348 if (slice == 0 || (mgp->dev->real_num_tx_queues > 1)) {
2349 cmd.data0 = slice;
2350 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_SEND_OFFSET,
2351 &cmd, 0);
2352 ss->tx.lanai = (struct mcp_kreq_ether_send __iomem *)
2353 (mgp->sram + cmd.data0);
2354 }
2355 cmd.data0 = slice;
2356 status |= myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_SMALL_RX_OFFSET,
2357 &cmd, 0);
2358 ss->rx_small.lanai = (struct mcp_kreq_ether_recv __iomem *)
2359 (mgp->sram + cmd.data0);
2360
2361 cmd.data0 = slice;
2362 status |= myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_BIG_RX_OFFSET, &cmd, 0);
2363 ss->rx_big.lanai = (struct mcp_kreq_ether_recv __iomem *)
2364 (mgp->sram + cmd.data0);
2365
2366 ss->tx.send_go = (__iomem __be32 *)
2367 (mgp->sram + MXGEFW_ETH_SEND_GO + 64 * slice);
2368 ss->tx.send_stop = (__iomem __be32 *)
2369 (mgp->sram + MXGEFW_ETH_SEND_STOP + 64 * slice);
2370 return status;
2371
2372 }
2373
2374 static int myri10ge_set_stats(struct myri10ge_priv *mgp, int slice)
2375 {
2376 struct myri10ge_cmd cmd;
2377 struct myri10ge_slice_state *ss;
2378 int status;
2379
2380 ss = &mgp->ss[slice];
2381 cmd.data0 = MYRI10GE_LOWPART_TO_U32(ss->fw_stats_bus);
2382 cmd.data1 = MYRI10GE_HIGHPART_TO_U32(ss->fw_stats_bus);
2383 cmd.data2 = sizeof(struct mcp_irq_data) | (slice << 16);
2384 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_STATS_DMA_V2, &cmd, 0);
2385 if (status == -ENOSYS) {
2386 dma_addr_t bus = ss->fw_stats_bus;
2387 if (slice != 0)
2388 return -EINVAL;
2389 bus += offsetof(struct mcp_irq_data, send_done_count);
2390 cmd.data0 = MYRI10GE_LOWPART_TO_U32(bus);
2391 cmd.data1 = MYRI10GE_HIGHPART_TO_U32(bus);
2392 status = myri10ge_send_cmd(mgp,
2393 MXGEFW_CMD_SET_STATS_DMA_OBSOLETE,
2394 &cmd, 0);
2395 /* Firmware cannot support multicast without STATS_DMA_V2 */
2396 mgp->fw_multicast_support = 0;
2397 } else {
2398 mgp->fw_multicast_support = 1;
2399 }
2400 return 0;
2401 }
2402
2403 static int myri10ge_open(struct net_device *dev)
2404 {
2405 struct myri10ge_slice_state *ss;
2406 struct myri10ge_priv *mgp = netdev_priv(dev);
2407 struct myri10ge_cmd cmd;
2408 int i, status, big_pow2, slice;
2409 u8 *itable;
2410 struct net_lro_mgr *lro_mgr;
2411
2412 if (mgp->running != MYRI10GE_ETH_STOPPED)
2413 return -EBUSY;
2414
2415 mgp->running = MYRI10GE_ETH_STARTING;
2416 status = myri10ge_reset(mgp);
2417 if (status != 0) {
2418 netdev_err(dev, "failed reset\n");
2419 goto abort_with_nothing;
2420 }
2421
2422 if (mgp->num_slices > 1) {
2423 cmd.data0 = mgp->num_slices;
2424 cmd.data1 = MXGEFW_SLICE_INTR_MODE_ONE_PER_SLICE;
2425 if (mgp->dev->real_num_tx_queues > 1)
2426 cmd.data1 |= MXGEFW_SLICE_ENABLE_MULTIPLE_TX_QUEUES;
2427 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_ENABLE_RSS_QUEUES,
2428 &cmd, 0);
2429 if (status != 0) {
2430 netdev_err(dev, "failed to set number of slices\n");
2431 goto abort_with_nothing;
2432 }
2433 /* setup the indirection table */
2434 cmd.data0 = mgp->num_slices;
2435 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_RSS_TABLE_SIZE,
2436 &cmd, 0);
2437
2438 status |= myri10ge_send_cmd(mgp,
2439 MXGEFW_CMD_GET_RSS_TABLE_OFFSET,
2440 &cmd, 0);
2441 if (status != 0) {
2442 netdev_err(dev, "failed to setup rss tables\n");
2443 goto abort_with_nothing;
2444 }
2445
2446 /* just enable an identity mapping */
2447 itable = mgp->sram + cmd.data0;
2448 for (i = 0; i < mgp->num_slices; i++)
2449 __raw_writeb(i, &itable[i]);
2450
2451 cmd.data0 = 1;
2452 cmd.data1 = myri10ge_rss_hash;
2453 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_RSS_ENABLE,
2454 &cmd, 0);
2455 if (status != 0) {
2456 netdev_err(dev, "failed to enable slices\n");
2457 goto abort_with_nothing;
2458 }
2459 }
2460
2461 status = myri10ge_request_irq(mgp);
2462 if (status != 0)
2463 goto abort_with_nothing;
2464
2465 /* decide what small buffer size to use. For good TCP rx
2466 * performance, it is important to not receive 1514 byte
2467 * frames into jumbo buffers, as it confuses the socket buffer
2468 * accounting code, leading to drops and erratic performance.
2469 */
2470
2471 if (dev->mtu <= ETH_DATA_LEN)
2472 /* enough for a TCP header */
2473 mgp->small_bytes = (128 > SMP_CACHE_BYTES)
2474 ? (128 - MXGEFW_PAD)
2475 : (SMP_CACHE_BYTES - MXGEFW_PAD);
2476 else
2477 /* enough for a vlan encapsulated ETH_DATA_LEN frame */
2478 mgp->small_bytes = VLAN_ETH_FRAME_LEN;
2479
2480 /* Override the small buffer size? */
2481 if (myri10ge_small_bytes >= 0)
2482 mgp->small_bytes = myri10ge_small_bytes;
2483
2484 /* Firmware needs the big buff size as a power of 2. Lie and
2485 * tell him the buffer is larger, because we only use 1
2486 * buffer/pkt, and the mtu will prevent overruns.
2487 */
2488 big_pow2 = dev->mtu + ETH_HLEN + VLAN_HLEN + MXGEFW_PAD;
2489 if (big_pow2 < MYRI10GE_ALLOC_SIZE / 2) {
2490 while (!is_power_of_2(big_pow2))
2491 big_pow2++;
2492 mgp->big_bytes = dev->mtu + ETH_HLEN + VLAN_HLEN + MXGEFW_PAD;
2493 } else {
2494 big_pow2 = MYRI10GE_ALLOC_SIZE;
2495 mgp->big_bytes = big_pow2;
2496 }
2497
2498 /* setup the per-slice data structures */
2499 for (slice = 0; slice < mgp->num_slices; slice++) {
2500 ss = &mgp->ss[slice];
2501
2502 status = myri10ge_get_txrx(mgp, slice);
2503 if (status != 0) {
2504 netdev_err(dev, "failed to get ring sizes or locations\n");
2505 goto abort_with_rings;
2506 }
2507 status = myri10ge_allocate_rings(ss);
2508 if (status != 0)
2509 goto abort_with_rings;
2510
2511 /* only firmware which supports multiple TX queues
2512 * supports setting up the tx stats on non-zero
2513 * slices */
2514 if (slice == 0 || mgp->dev->real_num_tx_queues > 1)
2515 status = myri10ge_set_stats(mgp, slice);
2516 if (status) {
2517 netdev_err(dev, "Couldn't set stats DMA\n");
2518 goto abort_with_rings;
2519 }
2520
2521 lro_mgr = &ss->rx_done.lro_mgr;
2522 lro_mgr->dev = dev;
2523 lro_mgr->features = LRO_F_NAPI;
2524 lro_mgr->ip_summed = CHECKSUM_COMPLETE;
2525 lro_mgr->ip_summed_aggr = CHECKSUM_UNNECESSARY;
2526 lro_mgr->max_desc = MYRI10GE_MAX_LRO_DESCRIPTORS;
2527 lro_mgr->lro_arr = ss->rx_done.lro_desc;
2528 lro_mgr->get_frag_header = myri10ge_get_frag_header;
2529 lro_mgr->max_aggr = myri10ge_lro_max_pkts;
2530 lro_mgr->frag_align_pad = 2;
2531 if (lro_mgr->max_aggr > MAX_SKB_FRAGS)
2532 lro_mgr->max_aggr = MAX_SKB_FRAGS;
2533
2534 /* must happen prior to any irq */
2535 napi_enable(&(ss)->napi);
2536 }
2537
2538 /* now give firmware buffers sizes, and MTU */
2539 cmd.data0 = dev->mtu + ETH_HLEN + VLAN_HLEN;
2540 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_MTU, &cmd, 0);
2541 cmd.data0 = mgp->small_bytes;
2542 status |=
2543 myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_SMALL_BUFFER_SIZE, &cmd, 0);
2544 cmd.data0 = big_pow2;
2545 status |=
2546 myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_BIG_BUFFER_SIZE, &cmd, 0);
2547 if (status) {
2548 netdev_err(dev, "Couldn't set buffer sizes\n");
2549 goto abort_with_rings;
2550 }
2551
2552 /*
2553 * Set Linux style TSO mode; this is needed only on newer
2554 * firmware versions. Older versions default to Linux
2555 * style TSO
2556 */
2557 cmd.data0 = 0;
2558 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_TSO_MODE, &cmd, 0);
2559 if (status && status != -ENOSYS) {
2560 netdev_err(dev, "Couldn't set TSO mode\n");
2561 goto abort_with_rings;
2562 }
2563
2564 mgp->link_state = ~0U;
2565 mgp->rdma_tags_available = 15;
2566
2567 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_ETHERNET_UP, &cmd, 0);
2568 if (status) {
2569 netdev_err(dev, "Couldn't bring up link\n");
2570 goto abort_with_rings;
2571 }
2572
2573 mgp->running = MYRI10GE_ETH_RUNNING;
2574 mgp->watchdog_timer.expires = jiffies + myri10ge_watchdog_timeout * HZ;
2575 add_timer(&mgp->watchdog_timer);
2576 netif_tx_wake_all_queues(dev);
2577
2578 return 0;
2579
2580 abort_with_rings:
2581 while (slice) {
2582 slice--;
2583 napi_disable(&mgp->ss[slice].napi);
2584 }
2585 for (i = 0; i < mgp->num_slices; i++)
2586 myri10ge_free_rings(&mgp->ss[i]);
2587
2588 myri10ge_free_irq(mgp);
2589
2590 abort_with_nothing:
2591 mgp->running = MYRI10GE_ETH_STOPPED;
2592 return -ENOMEM;
2593 }
2594
2595 static int myri10ge_close(struct net_device *dev)
2596 {
2597 struct myri10ge_priv *mgp = netdev_priv(dev);
2598 struct myri10ge_cmd cmd;
2599 int status, old_down_cnt;
2600 int i;
2601
2602 if (mgp->running != MYRI10GE_ETH_RUNNING)
2603 return 0;
2604
2605 if (mgp->ss[0].tx.req_bytes == NULL)
2606 return 0;
2607
2608 del_timer_sync(&mgp->watchdog_timer);
2609 mgp->running = MYRI10GE_ETH_STOPPING;
2610 for (i = 0; i < mgp->num_slices; i++) {
2611 napi_disable(&mgp->ss[i].napi);
2612 }
2613 netif_carrier_off(dev);
2614
2615 netif_tx_stop_all_queues(dev);
2616 if (mgp->rebooted == 0) {
2617 old_down_cnt = mgp->down_cnt;
2618 mb();
2619 status =
2620 myri10ge_send_cmd(mgp, MXGEFW_CMD_ETHERNET_DOWN, &cmd, 0);
2621 if (status)
2622 netdev_err(dev, "Couldn't bring down link\n");
2623
2624 wait_event_timeout(mgp->down_wq, old_down_cnt != mgp->down_cnt,
2625 HZ);
2626 if (old_down_cnt == mgp->down_cnt)
2627 netdev_err(dev, "never got down irq\n");
2628 }
2629 netif_tx_disable(dev);
2630 myri10ge_free_irq(mgp);
2631 for (i = 0; i < mgp->num_slices; i++)
2632 myri10ge_free_rings(&mgp->ss[i]);
2633
2634 mgp->running = MYRI10GE_ETH_STOPPED;
2635 return 0;
2636 }
2637
2638 /* copy an array of struct mcp_kreq_ether_send's to the mcp. Copy
2639 * backwards one at a time and handle ring wraps */
2640
2641 static inline void
2642 myri10ge_submit_req_backwards(struct myri10ge_tx_buf *tx,
2643 struct mcp_kreq_ether_send *src, int cnt)
2644 {
2645 int idx, starting_slot;
2646 starting_slot = tx->req;
2647 while (cnt > 1) {
2648 cnt--;
2649 idx = (starting_slot + cnt) & tx->mask;
2650 myri10ge_pio_copy(&tx->lanai[idx], &src[cnt], sizeof(*src));
2651 mb();
2652 }
2653 }
2654
2655 /*
2656 * copy an array of struct mcp_kreq_ether_send's to the mcp. Copy
2657 * at most 32 bytes at a time, so as to avoid involving the software
2658 * pio handler in the nic. We re-write the first segment's flags
2659 * to mark them valid only after writing the entire chain.
2660 */
2661
2662 static inline void
2663 myri10ge_submit_req(struct myri10ge_tx_buf *tx, struct mcp_kreq_ether_send *src,
2664 int cnt)
2665 {
2666 int idx, i;
2667 struct mcp_kreq_ether_send __iomem *dstp, *dst;
2668 struct mcp_kreq_ether_send *srcp;
2669 u8 last_flags;
2670
2671 idx = tx->req & tx->mask;
2672
2673 last_flags = src->flags;
2674 src->flags = 0;
2675 mb();
2676 dst = dstp = &tx->lanai[idx];
2677 srcp = src;
2678
2679 if ((idx + cnt) < tx->mask) {
2680 for (i = 0; i < (cnt - 1); i += 2) {
2681 myri10ge_pio_copy(dstp, srcp, 2 * sizeof(*src));
2682 mb(); /* force write every 32 bytes */
2683 srcp += 2;
2684 dstp += 2;
2685 }
2686 } else {
2687 /* submit all but the first request, and ensure
2688 * that it is submitted below */
2689 myri10ge_submit_req_backwards(tx, src, cnt);
2690 i = 0;
2691 }
2692 if (i < cnt) {
2693 /* submit the first request */
2694 myri10ge_pio_copy(dstp, srcp, sizeof(*src));
2695 mb(); /* barrier before setting valid flag */
2696 }
2697
2698 /* re-write the last 32-bits with the valid flags */
2699 src->flags = last_flags;
2700 put_be32(*((__be32 *) src + 3), (__be32 __iomem *) dst + 3);
2701 tx->req += cnt;
2702 mb();
2703 }
2704
2705 /*
2706 * Transmit a packet. We need to split the packet so that a single
2707 * segment does not cross myri10ge->tx_boundary, so this makes segment
2708 * counting tricky. So rather than try to count segments up front, we
2709 * just give up if there are too few segments to hold a reasonably
2710 * fragmented packet currently available. If we run
2711 * out of segments while preparing a packet for DMA, we just linearize
2712 * it and try again.
2713 */
2714
2715 static netdev_tx_t myri10ge_xmit(struct sk_buff *skb,
2716 struct net_device *dev)
2717 {
2718 struct myri10ge_priv *mgp = netdev_priv(dev);
2719 struct myri10ge_slice_state *ss;
2720 struct mcp_kreq_ether_send *req;
2721 struct myri10ge_tx_buf *tx;
2722 struct skb_frag_struct *frag;
2723 struct netdev_queue *netdev_queue;
2724 dma_addr_t bus;
2725 u32 low;
2726 __be32 high_swapped;
2727 unsigned int len;
2728 int idx, last_idx, avail, frag_cnt, frag_idx, count, mss, max_segments;
2729 u16 pseudo_hdr_offset, cksum_offset, queue;
2730 int cum_len, seglen, boundary, rdma_count;
2731 u8 flags, odd_flag;
2732
2733 queue = skb_get_queue_mapping(skb);
2734 ss = &mgp->ss[queue];
2735 netdev_queue = netdev_get_tx_queue(mgp->dev, queue);
2736 tx = &ss->tx;
2737
2738 again:
2739 req = tx->req_list;
2740 avail = tx->mask - 1 - (tx->req - tx->done);
2741
2742 mss = 0;
2743 max_segments = MXGEFW_MAX_SEND_DESC;
2744
2745 if (skb_is_gso(skb)) {
2746 mss = skb_shinfo(skb)->gso_size;
2747 max_segments = MYRI10GE_MAX_SEND_DESC_TSO;
2748 }
2749
2750 if ((unlikely(avail < max_segments))) {
2751 /* we are out of transmit resources */
2752 tx->stop_queue++;
2753 netif_tx_stop_queue(netdev_queue);
2754 return NETDEV_TX_BUSY;
2755 }
2756
2757 /* Setup checksum offloading, if needed */
2758 cksum_offset = 0;
2759 pseudo_hdr_offset = 0;
2760 odd_flag = 0;
2761 flags = (MXGEFW_FLAGS_NO_TSO | MXGEFW_FLAGS_FIRST);
2762 if (likely(skb->ip_summed == CHECKSUM_PARTIAL)) {
2763 cksum_offset = skb_checksum_start_offset(skb);
2764 pseudo_hdr_offset = cksum_offset + skb->csum_offset;
2765 /* If the headers are excessively large, then we must
2766 * fall back to a software checksum */
2767 if (unlikely(!mss && (cksum_offset > 255 ||
2768 pseudo_hdr_offset > 127))) {
2769 if (skb_checksum_help(skb))
2770 goto drop;
2771 cksum_offset = 0;
2772 pseudo_hdr_offset = 0;
2773 } else {
2774 odd_flag = MXGEFW_FLAGS_ALIGN_ODD;
2775 flags |= MXGEFW_FLAGS_CKSUM;
2776 }
2777 }
2778
2779 cum_len = 0;
2780
2781 if (mss) { /* TSO */
2782 /* this removes any CKSUM flag from before */
2783 flags = (MXGEFW_FLAGS_TSO_HDR | MXGEFW_FLAGS_FIRST);
2784
2785 /* negative cum_len signifies to the
2786 * send loop that we are still in the
2787 * header portion of the TSO packet.
2788 * TSO header can be at most 1KB long */
2789 cum_len = -(skb_transport_offset(skb) + tcp_hdrlen(skb));
2790
2791 /* for IPv6 TSO, the checksum offset stores the
2792 * TCP header length, to save the firmware from
2793 * the need to parse the headers */
2794 if (skb_is_gso_v6(skb)) {
2795 cksum_offset = tcp_hdrlen(skb);
2796 /* Can only handle headers <= max_tso6 long */
2797 if (unlikely(-cum_len > mgp->max_tso6))
2798 return myri10ge_sw_tso(skb, dev);
2799 }
2800 /* for TSO, pseudo_hdr_offset holds mss.
2801 * The firmware figures out where to put
2802 * the checksum by parsing the header. */
2803 pseudo_hdr_offset = mss;
2804 } else
2805 /* Mark small packets, and pad out tiny packets */
2806 if (skb->len <= MXGEFW_SEND_SMALL_SIZE) {
2807 flags |= MXGEFW_FLAGS_SMALL;
2808
2809 /* pad frames to at least ETH_ZLEN bytes */
2810 if (unlikely(skb->len < ETH_ZLEN)) {
2811 if (skb_padto(skb, ETH_ZLEN)) {
2812 /* The packet is gone, so we must
2813 * return 0 */
2814 ss->stats.tx_dropped += 1;
2815 return NETDEV_TX_OK;
2816 }
2817 /* adjust the len to account for the zero pad
2818 * so that the nic can know how long it is */
2819 skb->len = ETH_ZLEN;
2820 }
2821 }
2822
2823 /* map the skb for DMA */
2824 len = skb_headlen(skb);
2825 idx = tx->req & tx->mask;
2826 tx->info[idx].skb = skb;
2827 bus = pci_map_single(mgp->pdev, skb->data, len, PCI_DMA_TODEVICE);
2828 dma_unmap_addr_set(&tx->info[idx], bus, bus);
2829 dma_unmap_len_set(&tx->info[idx], len, len);
2830
2831 frag_cnt = skb_shinfo(skb)->nr_frags;
2832 frag_idx = 0;
2833 count = 0;
2834 rdma_count = 0;
2835
2836 /* "rdma_count" is the number of RDMAs belonging to the
2837 * current packet BEFORE the current send request. For
2838 * non-TSO packets, this is equal to "count".
2839 * For TSO packets, rdma_count needs to be reset
2840 * to 0 after a segment cut.
2841 *
2842 * The rdma_count field of the send request is
2843 * the number of RDMAs of the packet starting at
2844 * that request. For TSO send requests with one ore more cuts
2845 * in the middle, this is the number of RDMAs starting
2846 * after the last cut in the request. All previous
2847 * segments before the last cut implicitly have 1 RDMA.
2848 *
2849 * Since the number of RDMAs is not known beforehand,
2850 * it must be filled-in retroactively - after each
2851 * segmentation cut or at the end of the entire packet.
2852 */
2853
2854 while (1) {
2855 /* Break the SKB or Fragment up into pieces which
2856 * do not cross mgp->tx_boundary */
2857 low = MYRI10GE_LOWPART_TO_U32(bus);
2858 high_swapped = htonl(MYRI10GE_HIGHPART_TO_U32(bus));
2859 while (len) {
2860 u8 flags_next;
2861 int cum_len_next;
2862
2863 if (unlikely(count == max_segments))
2864 goto abort_linearize;
2865
2866 boundary =
2867 (low + mgp->tx_boundary) & ~(mgp->tx_boundary - 1);
2868 seglen = boundary - low;
2869 if (seglen > len)
2870 seglen = len;
2871 flags_next = flags & ~MXGEFW_FLAGS_FIRST;
2872 cum_len_next = cum_len + seglen;
2873 if (mss) { /* TSO */
2874 (req - rdma_count)->rdma_count = rdma_count + 1;
2875
2876 if (likely(cum_len >= 0)) { /* payload */
2877 int next_is_first, chop;
2878
2879 chop = (cum_len_next > mss);
2880 cum_len_next = cum_len_next % mss;
2881 next_is_first = (cum_len_next == 0);
2882 flags |= chop * MXGEFW_FLAGS_TSO_CHOP;
2883 flags_next |= next_is_first *
2884 MXGEFW_FLAGS_FIRST;
2885 rdma_count |= -(chop | next_is_first);
2886 rdma_count += chop & !next_is_first;
2887 } else if (likely(cum_len_next >= 0)) { /* header ends */
2888 int small;
2889
2890 rdma_count = -1;
2891 cum_len_next = 0;
2892 seglen = -cum_len;
2893 small = (mss <= MXGEFW_SEND_SMALL_SIZE);
2894 flags_next = MXGEFW_FLAGS_TSO_PLD |
2895 MXGEFW_FLAGS_FIRST |
2896 (small * MXGEFW_FLAGS_SMALL);
2897 }
2898 }
2899 req->addr_high = high_swapped;
2900 req->addr_low = htonl(low);
2901 req->pseudo_hdr_offset = htons(pseudo_hdr_offset);
2902 req->pad = 0; /* complete solid 16-byte block; does this matter? */
2903 req->rdma_count = 1;
2904 req->length = htons(seglen);
2905 req->cksum_offset = cksum_offset;
2906 req->flags = flags | ((cum_len & 1) * odd_flag);
2907
2908 low += seglen;
2909 len -= seglen;
2910 cum_len = cum_len_next;
2911 flags = flags_next;
2912 req++;
2913 count++;
2914 rdma_count++;
2915 if (cksum_offset != 0 && !(mss && skb_is_gso_v6(skb))) {
2916 if (unlikely(cksum_offset > seglen))
2917 cksum_offset -= seglen;
2918 else
2919 cksum_offset = 0;
2920 }
2921 }
2922 if (frag_idx == frag_cnt)
2923 break;
2924
2925 /* map next fragment for DMA */
2926 idx = (count + tx->req) & tx->mask;
2927 frag = &skb_shinfo(skb)->frags[frag_idx];
2928 frag_idx++;
2929 len = frag->size;
2930 bus = pci_map_page(mgp->pdev, frag->page, frag->page_offset,
2931 len, PCI_DMA_TODEVICE);
2932 dma_unmap_addr_set(&tx->info[idx], bus, bus);
2933 dma_unmap_len_set(&tx->info[idx], len, len);
2934 }
2935
2936 (req - rdma_count)->rdma_count = rdma_count;
2937 if (mss)
2938 do {
2939 req--;
2940 req->flags |= MXGEFW_FLAGS_TSO_LAST;
2941 } while (!(req->flags & (MXGEFW_FLAGS_TSO_CHOP |
2942 MXGEFW_FLAGS_FIRST)));
2943 idx = ((count - 1) + tx->req) & tx->mask;
2944 tx->info[idx].last = 1;
2945 myri10ge_submit_req(tx, tx->req_list, count);
2946 /* if using multiple tx queues, make sure NIC polls the
2947 * current slice */
2948 if ((mgp->dev->real_num_tx_queues > 1) && tx->queue_active == 0) {
2949 tx->queue_active = 1;
2950 put_be32(htonl(1), tx->send_go);
2951 mb();
2952 mmiowb();
2953 }
2954 tx->pkt_start++;
2955 if ((avail - count) < MXGEFW_MAX_SEND_DESC) {
2956 tx->stop_queue++;
2957 netif_tx_stop_queue(netdev_queue);
2958 }
2959 return NETDEV_TX_OK;
2960
2961 abort_linearize:
2962 /* Free any DMA resources we've alloced and clear out the skb
2963 * slot so as to not trip up assertions, and to avoid a
2964 * double-free if linearizing fails */
2965
2966 last_idx = (idx + 1) & tx->mask;
2967 idx = tx->req & tx->mask;
2968 tx->info[idx].skb = NULL;
2969 do {
2970 len = dma_unmap_len(&tx->info[idx], len);
2971 if (len) {
2972 if (tx->info[idx].skb != NULL)
2973 pci_unmap_single(mgp->pdev,
2974 dma_unmap_addr(&tx->info[idx],
2975 bus), len,
2976 PCI_DMA_TODEVICE);
2977 else
2978 pci_unmap_page(mgp->pdev,
2979 dma_unmap_addr(&tx->info[idx],
2980 bus), len,
2981 PCI_DMA_TODEVICE);
2982 dma_unmap_len_set(&tx->info[idx], len, 0);
2983 tx->info[idx].skb = NULL;
2984 }
2985 idx = (idx + 1) & tx->mask;
2986 } while (idx != last_idx);
2987 if (skb_is_gso(skb)) {
2988 netdev_err(mgp->dev, "TSO but wanted to linearize?!?!?\n");
2989 goto drop;
2990 }
2991
2992 if (skb_linearize(skb))
2993 goto drop;
2994
2995 tx->linearized++;
2996 goto again;
2997
2998 drop:
2999 dev_kfree_skb_any(skb);
3000 ss->stats.tx_dropped += 1;
3001 return NETDEV_TX_OK;
3002
3003 }
3004
3005 static netdev_tx_t myri10ge_sw_tso(struct sk_buff *skb,
3006 struct net_device *dev)
3007 {
3008 struct sk_buff *segs, *curr;
3009 struct myri10ge_priv *mgp = netdev_priv(dev);
3010 struct myri10ge_slice_state *ss;
3011 netdev_tx_t status;
3012
3013 segs = skb_gso_segment(skb, dev->features & ~NETIF_F_TSO6);
3014 if (IS_ERR(segs))
3015 goto drop;
3016
3017 while (segs) {
3018 curr = segs;
3019 segs = segs->next;
3020 curr->next = NULL;
3021 status = myri10ge_xmit(curr, dev);
3022 if (status != 0) {
3023 dev_kfree_skb_any(curr);
3024 if (segs != NULL) {
3025 curr = segs;
3026 segs = segs->next;
3027 curr->next = NULL;
3028 dev_kfree_skb_any(segs);
3029 }
3030 goto drop;
3031 }
3032 }
3033 dev_kfree_skb_any(skb);
3034 return NETDEV_TX_OK;
3035
3036 drop:
3037 ss = &mgp->ss[skb_get_queue_mapping(skb)];
3038 dev_kfree_skb_any(skb);
3039 ss->stats.tx_dropped += 1;
3040 return NETDEV_TX_OK;
3041 }
3042
3043 static struct rtnl_link_stats64 *myri10ge_get_stats(struct net_device *dev,
3044 struct rtnl_link_stats64 *stats)
3045 {
3046 const struct myri10ge_priv *mgp = netdev_priv(dev);
3047 const struct myri10ge_slice_netstats *slice_stats;
3048 int i;
3049
3050 for (i = 0; i < mgp->num_slices; i++) {
3051 slice_stats = &mgp->ss[i].stats;
3052 stats->rx_packets += slice_stats->rx_packets;
3053 stats->tx_packets += slice_stats->tx_packets;
3054 stats->rx_bytes += slice_stats->rx_bytes;
3055 stats->tx_bytes += slice_stats->tx_bytes;
3056 stats->rx_dropped += slice_stats->rx_dropped;
3057 stats->tx_dropped += slice_stats->tx_dropped;
3058 }
3059 return stats;
3060 }
3061
3062 static void myri10ge_set_multicast_list(struct net_device *dev)
3063 {
3064 struct myri10ge_priv *mgp = netdev_priv(dev);
3065 struct myri10ge_cmd cmd;
3066 struct netdev_hw_addr *ha;
3067 __be32 data[2] = { 0, 0 };
3068 int err;
3069
3070 /* can be called from atomic contexts,
3071 * pass 1 to force atomicity in myri10ge_send_cmd() */
3072 myri10ge_change_promisc(mgp, dev->flags & IFF_PROMISC, 1);
3073
3074 /* This firmware is known to not support multicast */
3075 if (!mgp->fw_multicast_support)
3076 return;
3077
3078 /* Disable multicast filtering */
3079
3080 err = myri10ge_send_cmd(mgp, MXGEFW_ENABLE_ALLMULTI, &cmd, 1);
3081 if (err != 0) {
3082 netdev_err(dev, "Failed MXGEFW_ENABLE_ALLMULTI, error status: %d\n",
3083 err);
3084 goto abort;
3085 }
3086
3087 if ((dev->flags & IFF_ALLMULTI) || mgp->adopted_rx_filter_bug) {
3088 /* request to disable multicast filtering, so quit here */
3089 return;
3090 }
3091
3092 /* Flush the filters */
3093
3094 err = myri10ge_send_cmd(mgp, MXGEFW_LEAVE_ALL_MULTICAST_GROUPS,
3095 &cmd, 1);
3096 if (err != 0) {
3097 netdev_err(dev, "Failed MXGEFW_LEAVE_ALL_MULTICAST_GROUPS, error status: %d\n",
3098 err);
3099 goto abort;
3100 }
3101
3102 /* Walk the multicast list, and add each address */
3103 netdev_for_each_mc_addr(ha, dev) {
3104 memcpy(data, &ha->addr, 6);
3105 cmd.data0 = ntohl(data[0]);
3106 cmd.data1 = ntohl(data[1]);
3107 err = myri10ge_send_cmd(mgp, MXGEFW_JOIN_MULTICAST_GROUP,
3108 &cmd, 1);
3109
3110 if (err != 0) {
3111 netdev_err(dev, "Failed MXGEFW_JOIN_MULTICAST_GROUP, error status:%d %pM\n",
3112 err, ha->addr);
3113 goto abort;
3114 }
3115 }
3116 /* Enable multicast filtering */
3117 err = myri10ge_send_cmd(mgp, MXGEFW_DISABLE_ALLMULTI, &cmd, 1);
3118 if (err != 0) {
3119 netdev_err(dev, "Failed MXGEFW_DISABLE_ALLMULTI, error status: %d\n",
3120 err);
3121 goto abort;
3122 }
3123
3124 return;
3125
3126 abort:
3127 return;
3128 }
3129
3130 static int myri10ge_set_mac_address(struct net_device *dev, void *addr)
3131 {
3132 struct sockaddr *sa = addr;
3133 struct myri10ge_priv *mgp = netdev_priv(dev);
3134 int status;
3135
3136 if (!is_valid_ether_addr(sa->sa_data))
3137 return -EADDRNOTAVAIL;
3138
3139 status = myri10ge_update_mac_address(mgp, sa->sa_data);
3140 if (status != 0) {
3141 netdev_err(dev, "changing mac address failed with %d\n",
3142 status);
3143 return status;
3144 }
3145
3146 /* change the dev structure */
3147 memcpy(dev->dev_addr, sa->sa_data, 6);
3148 return 0;
3149 }
3150
3151 static u32 myri10ge_fix_features(struct net_device *dev, u32 features)
3152 {
3153 if (!(features & NETIF_F_RXCSUM))
3154 features &= ~NETIF_F_LRO;
3155
3156 return features;
3157 }
3158
3159 static int myri10ge_change_mtu(struct net_device *dev, int new_mtu)
3160 {
3161 struct myri10ge_priv *mgp = netdev_priv(dev);
3162 int error = 0;
3163
3164 if ((new_mtu < 68) || (ETH_HLEN + new_mtu > MYRI10GE_MAX_ETHER_MTU)) {
3165 netdev_err(dev, "new mtu (%d) is not valid\n", new_mtu);
3166 return -EINVAL;
3167 }
3168 netdev_info(dev, "changing mtu from %d to %d\n", dev->mtu, new_mtu);
3169 if (mgp->running) {
3170 /* if we change the mtu on an active device, we must
3171 * reset the device so the firmware sees the change */
3172 myri10ge_close(dev);
3173 dev->mtu = new_mtu;
3174 myri10ge_open(dev);
3175 } else
3176 dev->mtu = new_mtu;
3177
3178 return error;
3179 }
3180
3181 /*
3182 * Enable ECRC to align PCI-E Completion packets on an 8-byte boundary.
3183 * Only do it if the bridge is a root port since we don't want to disturb
3184 * any other device, except if forced with myri10ge_ecrc_enable > 1.
3185 */
3186
3187 static void myri10ge_enable_ecrc(struct myri10ge_priv *mgp)
3188 {
3189 struct pci_dev *bridge = mgp->pdev->bus->self;
3190 struct device *dev = &mgp->pdev->dev;
3191 int cap;
3192 unsigned err_cap;
3193 u16 val;
3194 u8 ext_type;
3195 int ret;
3196
3197 if (!myri10ge_ecrc_enable || !bridge)
3198 return;
3199
3200 /* check that the bridge is a root port */
3201 cap = pci_pcie_cap(bridge);
3202 pci_read_config_word(bridge, cap + PCI_CAP_FLAGS, &val);
3203 ext_type = (val & PCI_EXP_FLAGS_TYPE) >> 4;
3204 if (ext_type != PCI_EXP_TYPE_ROOT_PORT) {
3205 if (myri10ge_ecrc_enable > 1) {
3206 struct pci_dev *prev_bridge, *old_bridge = bridge;
3207
3208 /* Walk the hierarchy up to the root port
3209 * where ECRC has to be enabled */
3210 do {
3211 prev_bridge = bridge;
3212 bridge = bridge->bus->self;
3213 if (!bridge || prev_bridge == bridge) {
3214 dev_err(dev,
3215 "Failed to find root port"
3216 " to force ECRC\n");
3217 return;
3218 }
3219 cap = pci_pcie_cap(bridge);
3220 pci_read_config_word(bridge,
3221 cap + PCI_CAP_FLAGS, &val);
3222 ext_type = (val & PCI_EXP_FLAGS_TYPE) >> 4;
3223 } while (ext_type != PCI_EXP_TYPE_ROOT_PORT);
3224
3225 dev_info(dev,
3226 "Forcing ECRC on non-root port %s"
3227 " (enabling on root port %s)\n",
3228 pci_name(old_bridge), pci_name(bridge));
3229 } else {
3230 dev_err(dev,
3231 "Not enabling ECRC on non-root port %s\n",
3232 pci_name(bridge));
3233 return;
3234 }
3235 }
3236
3237 cap = pci_find_ext_capability(bridge, PCI_EXT_CAP_ID_ERR);
3238 if (!cap)
3239 return;
3240
3241 ret = pci_read_config_dword(bridge, cap + PCI_ERR_CAP, &err_cap);
3242 if (ret) {
3243 dev_err(dev, "failed reading ext-conf-space of %s\n",
3244 pci_name(bridge));
3245 dev_err(dev, "\t pci=nommconf in use? "
3246 "or buggy/incomplete/absent ACPI MCFG attr?\n");
3247 return;
3248 }
3249 if (!(err_cap & PCI_ERR_CAP_ECRC_GENC))
3250 return;
3251
3252 err_cap |= PCI_ERR_CAP_ECRC_GENE;
3253 pci_write_config_dword(bridge, cap + PCI_ERR_CAP, err_cap);
3254 dev_info(dev, "Enabled ECRC on upstream bridge %s\n", pci_name(bridge));
3255 }
3256
3257 /*
3258 * The Lanai Z8E PCI-E interface achieves higher Read-DMA throughput
3259 * when the PCI-E Completion packets are aligned on an 8-byte
3260 * boundary. Some PCI-E chip sets always align Completion packets; on
3261 * the ones that do not, the alignment can be enforced by enabling
3262 * ECRC generation (if supported).
3263 *
3264 * When PCI-E Completion packets are not aligned, it is actually more
3265 * efficient to limit Read-DMA transactions to 2KB, rather than 4KB.
3266 *
3267 * If the driver can neither enable ECRC nor verify that it has
3268 * already been enabled, then it must use a firmware image which works
3269 * around unaligned completion packets (myri10ge_rss_ethp_z8e.dat), and it
3270 * should also ensure that it never gives the device a Read-DMA which is
3271 * larger than 2KB by setting the tx_boundary to 2KB. If ECRC is
3272 * enabled, then the driver should use the aligned (myri10ge_rss_eth_z8e.dat)
3273 * firmware image, and set tx_boundary to 4KB.
3274 */
3275
3276 static void myri10ge_firmware_probe(struct myri10ge_priv *mgp)
3277 {
3278 struct pci_dev *pdev = mgp->pdev;
3279 struct device *dev = &pdev->dev;
3280 int status;
3281
3282 mgp->tx_boundary = 4096;
3283 /*
3284 * Verify the max read request size was set to 4KB
3285 * before trying the test with 4KB.
3286 */
3287 status = pcie_get_readrq(pdev);
3288 if (status < 0) {
3289 dev_err(dev, "Couldn't read max read req size: %d\n", status);
3290 goto abort;
3291 }
3292 if (status != 4096) {
3293 dev_warn(dev, "Max Read Request size != 4096 (%d)\n", status);
3294 mgp->tx_boundary = 2048;
3295 }
3296 /*
3297 * load the optimized firmware (which assumes aligned PCIe
3298 * completions) in order to see if it works on this host.
3299 */
3300 set_fw_name(mgp, myri10ge_fw_aligned, false);
3301 status = myri10ge_load_firmware(mgp, 1);
3302 if (status != 0) {
3303 goto abort;
3304 }
3305
3306 /*
3307 * Enable ECRC if possible
3308 */
3309 myri10ge_enable_ecrc(mgp);
3310
3311 /*
3312 * Run a DMA test which watches for unaligned completions and
3313 * aborts on the first one seen.
3314 */
3315
3316 status = myri10ge_dma_test(mgp, MXGEFW_CMD_UNALIGNED_TEST);
3317 if (status == 0)
3318 return; /* keep the aligned firmware */
3319
3320 if (status != -E2BIG)
3321 dev_warn(dev, "DMA test failed: %d\n", status);
3322 if (status == -ENOSYS)
3323 dev_warn(dev, "Falling back to ethp! "
3324 "Please install up to date fw\n");
3325 abort:
3326 /* fall back to using the unaligned firmware */
3327 mgp->tx_boundary = 2048;
3328 set_fw_name(mgp, myri10ge_fw_unaligned, false);
3329 }
3330
3331 static void myri10ge_select_firmware(struct myri10ge_priv *mgp)
3332 {
3333 int overridden = 0;
3334
3335 if (myri10ge_force_firmware == 0) {
3336 int link_width, exp_cap;
3337 u16 lnk;
3338
3339 exp_cap = pci_pcie_cap(mgp->pdev);
3340 pci_read_config_word(mgp->pdev, exp_cap + PCI_EXP_LNKSTA, &lnk);
3341 link_width = (lnk >> 4) & 0x3f;
3342
3343 /* Check to see if Link is less than 8 or if the
3344 * upstream bridge is known to provide aligned
3345 * completions */
3346 if (link_width < 8) {
3347 dev_info(&mgp->pdev->dev, "PCIE x%d Link\n",
3348 link_width);
3349 mgp->tx_boundary = 4096;
3350 set_fw_name(mgp, myri10ge_fw_aligned, false);
3351 } else {
3352 myri10ge_firmware_probe(mgp);
3353 }
3354 } else {
3355 if (myri10ge_force_firmware == 1) {
3356 dev_info(&mgp->pdev->dev,
3357 "Assuming aligned completions (forced)\n");
3358 mgp->tx_boundary = 4096;
3359 set_fw_name(mgp, myri10ge_fw_aligned, false);
3360 } else {
3361 dev_info(&mgp->pdev->dev,
3362 "Assuming unaligned completions (forced)\n");
3363 mgp->tx_boundary = 2048;
3364 set_fw_name(mgp, myri10ge_fw_unaligned, false);
3365 }
3366 }
3367
3368 kparam_block_sysfs_write(myri10ge_fw_name);
3369 if (myri10ge_fw_name != NULL) {
3370 char *fw_name = kstrdup(myri10ge_fw_name, GFP_KERNEL);
3371 if (fw_name) {
3372 overridden = 1;
3373 set_fw_name(mgp, fw_name, true);
3374 }
3375 }
3376 kparam_unblock_sysfs_write(myri10ge_fw_name);
3377
3378 if (mgp->board_number < MYRI10GE_MAX_BOARDS &&
3379 myri10ge_fw_names[mgp->board_number] != NULL &&
3380 strlen(myri10ge_fw_names[mgp->board_number])) {
3381 set_fw_name(mgp, myri10ge_fw_names[mgp->board_number], false);
3382 overridden = 1;
3383 }
3384 if (overridden)
3385 dev_info(&mgp->pdev->dev, "overriding firmware to %s\n",
3386 mgp->fw_name);
3387 }
3388
3389 static void myri10ge_mask_surprise_down(struct pci_dev *pdev)
3390 {
3391 struct pci_dev *bridge = pdev->bus->self;
3392 int cap;
3393 u32 mask;
3394
3395 if (bridge == NULL)
3396 return;
3397
3398 cap = pci_find_ext_capability(bridge, PCI_EXT_CAP_ID_ERR);
3399 if (cap) {
3400 /* a sram parity error can cause a surprise link
3401 * down; since we expect and can recover from sram
3402 * parity errors, mask surprise link down events */
3403 pci_read_config_dword(bridge, cap + PCI_ERR_UNCOR_MASK, &mask);
3404 mask |= 0x20;
3405 pci_write_config_dword(bridge, cap + PCI_ERR_UNCOR_MASK, mask);
3406 }
3407 }
3408
3409 #ifdef CONFIG_PM
3410 static int myri10ge_suspend(struct pci_dev *pdev, pm_message_t state)
3411 {
3412 struct myri10ge_priv *mgp;
3413 struct net_device *netdev;
3414
3415 mgp = pci_get_drvdata(pdev);
3416 if (mgp == NULL)
3417 return -EINVAL;
3418 netdev = mgp->dev;
3419
3420 netif_device_detach(netdev);
3421 if (netif_running(netdev)) {
3422 netdev_info(netdev, "closing\n");
3423 rtnl_lock();
3424 myri10ge_close(netdev);
3425 rtnl_unlock();
3426 }
3427 myri10ge_dummy_rdma(mgp, 0);
3428 pci_save_state(pdev);
3429 pci_disable_device(pdev);
3430
3431 return pci_set_power_state(pdev, pci_choose_state(pdev, state));
3432 }
3433
3434 static int myri10ge_resume(struct pci_dev *pdev)
3435 {
3436 struct myri10ge_priv *mgp;
3437 struct net_device *netdev;
3438 int status;
3439 u16 vendor;
3440
3441 mgp = pci_get_drvdata(pdev);
3442 if (mgp == NULL)
3443 return -EINVAL;
3444 netdev = mgp->dev;
3445 pci_set_power_state(pdev, 0); /* zeros conf space as a side effect */
3446 msleep(5); /* give card time to respond */
3447 pci_read_config_word(mgp->pdev, PCI_VENDOR_ID, &vendor);
3448 if (vendor == 0xffff) {
3449 netdev_err(mgp->dev, "device disappeared!\n");
3450 return -EIO;
3451 }
3452
3453 pci_restore_state(pdev);
3454
3455 status = pci_enable_device(pdev);
3456 if (status) {
3457 dev_err(&pdev->dev, "failed to enable device\n");
3458 return status;
3459 }
3460
3461 pci_set_master(pdev);
3462
3463 myri10ge_reset(mgp);
3464 myri10ge_dummy_rdma(mgp, 1);
3465
3466 /* Save configuration space to be restored if the
3467 * nic resets due to a parity error */
3468 pci_save_state(pdev);
3469
3470 if (netif_running(netdev)) {
3471 rtnl_lock();
3472 status = myri10ge_open(netdev);
3473 rtnl_unlock();
3474 if (status != 0)
3475 goto abort_with_enabled;
3476
3477 }
3478 netif_device_attach(netdev);
3479
3480 return 0;
3481
3482 abort_with_enabled:
3483 pci_disable_device(pdev);
3484 return -EIO;
3485
3486 }
3487 #endif /* CONFIG_PM */
3488
3489 static u32 myri10ge_read_reboot(struct myri10ge_priv *mgp)
3490 {
3491 struct pci_dev *pdev = mgp->pdev;
3492 int vs = mgp->vendor_specific_offset;
3493 u32 reboot;
3494
3495 /*enter read32 mode */
3496 pci_write_config_byte(pdev, vs + 0x10, 0x3);
3497
3498 /*read REBOOT_STATUS (0xfffffff0) */
3499 pci_write_config_dword(pdev, vs + 0x18, 0xfffffff0);
3500 pci_read_config_dword(pdev, vs + 0x14, &reboot);
3501 return reboot;
3502 }
3503
3504 static void
3505 myri10ge_check_slice(struct myri10ge_slice_state *ss, int *reset_needed,
3506 int *busy_slice_cnt, u32 rx_pause_cnt)
3507 {
3508 struct myri10ge_priv *mgp = ss->mgp;
3509 int slice = ss - mgp->ss;
3510
3511 if (ss->tx.req != ss->tx.done &&
3512 ss->tx.done == ss->watchdog_tx_done &&
3513 ss->watchdog_tx_req != ss->watchdog_tx_done) {
3514 /* nic seems like it might be stuck.. */
3515 if (rx_pause_cnt != mgp->watchdog_pause) {
3516 if (net_ratelimit())
3517 netdev_warn(mgp->dev, "slice %d: TX paused, "
3518 "check link partner\n", slice);
3519 } else {
3520 netdev_warn(mgp->dev,
3521 "slice %d: TX stuck %d %d %d %d %d %d\n",
3522 slice, ss->tx.queue_active, ss->tx.req,
3523 ss->tx.done, ss->tx.pkt_start,
3524 ss->tx.pkt_done,
3525 (int)ntohl(mgp->ss[slice].fw_stats->
3526 send_done_count));
3527 *reset_needed = 1;
3528 ss->stuck = 1;
3529 }
3530 }
3531 if (ss->watchdog_tx_done != ss->tx.done ||
3532 ss->watchdog_rx_done != ss->rx_done.cnt) {
3533 *busy_slice_cnt += 1;
3534 }
3535 ss->watchdog_tx_done = ss->tx.done;
3536 ss->watchdog_tx_req = ss->tx.req;
3537 ss->watchdog_rx_done = ss->rx_done.cnt;
3538 }
3539
3540 /*
3541 * This watchdog is used to check whether the board has suffered
3542 * from a parity error and needs to be recovered.
3543 */
3544 static void myri10ge_watchdog(struct work_struct *work)
3545 {
3546 struct myri10ge_priv *mgp =
3547 container_of(work, struct myri10ge_priv, watchdog_work);
3548 struct myri10ge_slice_state *ss;
3549 u32 reboot, rx_pause_cnt;
3550 int status, rebooted;
3551 int i;
3552 int reset_needed = 0;
3553 int busy_slice_cnt = 0;
3554 u16 cmd, vendor;
3555
3556 mgp->watchdog_resets++;
3557 pci_read_config_word(mgp->pdev, PCI_COMMAND, &cmd);
3558 rebooted = 0;
3559 if ((cmd & PCI_COMMAND_MASTER) == 0) {
3560 /* Bus master DMA disabled? Check to see
3561 * if the card rebooted due to a parity error
3562 * For now, just report it */
3563 reboot = myri10ge_read_reboot(mgp);
3564 netdev_err(mgp->dev, "NIC rebooted (0x%x),%s resetting\n",
3565 reboot, myri10ge_reset_recover ? "" : " not");
3566 if (myri10ge_reset_recover == 0)
3567 return;
3568 rtnl_lock();
3569 mgp->rebooted = 1;
3570 rebooted = 1;
3571 myri10ge_close(mgp->dev);
3572 myri10ge_reset_recover--;
3573 mgp->rebooted = 0;
3574 /*
3575 * A rebooted nic will come back with config space as
3576 * it was after power was applied to PCIe bus.
3577 * Attempt to restore config space which was saved
3578 * when the driver was loaded, or the last time the
3579 * nic was resumed from power saving mode.
3580 */
3581 pci_restore_state(mgp->pdev);
3582
3583 /* save state again for accounting reasons */
3584 pci_save_state(mgp->pdev);
3585
3586 } else {
3587 /* if we get back -1's from our slot, perhaps somebody
3588 * powered off our card. Don't try to reset it in
3589 * this case */
3590 if (cmd == 0xffff) {
3591 pci_read_config_word(mgp->pdev, PCI_VENDOR_ID, &vendor);
3592 if (vendor == 0xffff) {
3593 netdev_err(mgp->dev, "device disappeared!\n");
3594 return;
3595 }
3596 }
3597 /* Perhaps it is a software error. See if stuck slice
3598 * has recovered, reset if not */
3599 rx_pause_cnt = ntohl(mgp->ss[0].fw_stats->dropped_pause);
3600 for (i = 0; i < mgp->num_slices; i++) {
3601 ss = mgp->ss;
3602 if (ss->stuck) {
3603 myri10ge_check_slice(ss, &reset_needed,
3604 &busy_slice_cnt,
3605 rx_pause_cnt);
3606 ss->stuck = 0;
3607 }
3608 }
3609 if (!reset_needed) {
3610 netdev_dbg(mgp->dev, "not resetting\n");
3611 return;
3612 }
3613
3614 netdev_err(mgp->dev, "device timeout, resetting\n");
3615 }
3616
3617 if (!rebooted) {
3618 rtnl_lock();
3619 myri10ge_close(mgp->dev);
3620 }
3621 status = myri10ge_load_firmware(mgp, 1);
3622 if (status != 0)
3623 netdev_err(mgp->dev, "failed to load firmware\n");
3624 else
3625 myri10ge_open(mgp->dev);
3626 rtnl_unlock();
3627 }
3628
3629 /*
3630 * We use our own timer routine rather than relying upon
3631 * netdev->tx_timeout because we have a very large hardware transmit
3632 * queue. Due to the large queue, the netdev->tx_timeout function
3633 * cannot detect a NIC with a parity error in a timely fashion if the
3634 * NIC is lightly loaded.
3635 */
3636 static void myri10ge_watchdog_timer(unsigned long arg)
3637 {
3638 struct myri10ge_priv *mgp;
3639 struct myri10ge_slice_state *ss;
3640 int i, reset_needed, busy_slice_cnt;
3641 u32 rx_pause_cnt;
3642 u16 cmd;
3643
3644 mgp = (struct myri10ge_priv *)arg;
3645
3646 rx_pause_cnt = ntohl(mgp->ss[0].fw_stats->dropped_pause);
3647 busy_slice_cnt = 0;
3648 for (i = 0, reset_needed = 0;
3649 i < mgp->num_slices && reset_needed == 0; ++i) {
3650
3651 ss = &mgp->ss[i];
3652 if (ss->rx_small.watchdog_needed) {
3653 myri10ge_alloc_rx_pages(mgp, &ss->rx_small,
3654 mgp->small_bytes + MXGEFW_PAD,
3655 1);
3656 if (ss->rx_small.fill_cnt - ss->rx_small.cnt >=
3657 myri10ge_fill_thresh)
3658 ss->rx_small.watchdog_needed = 0;
3659 }
3660 if (ss->rx_big.watchdog_needed) {
3661 myri10ge_alloc_rx_pages(mgp, &ss->rx_big,
3662 mgp->big_bytes, 1);
3663 if (ss->rx_big.fill_cnt - ss->rx_big.cnt >=
3664 myri10ge_fill_thresh)
3665 ss->rx_big.watchdog_needed = 0;
3666 }
3667 myri10ge_check_slice(ss, &reset_needed, &busy_slice_cnt,
3668 rx_pause_cnt);
3669 }
3670 /* if we've sent or received no traffic, poll the NIC to
3671 * ensure it is still there. Otherwise, we risk not noticing
3672 * an error in a timely fashion */
3673 if (busy_slice_cnt == 0) {
3674 pci_read_config_word(mgp->pdev, PCI_COMMAND, &cmd);
3675 if ((cmd & PCI_COMMAND_MASTER) == 0) {
3676 reset_needed = 1;
3677 }
3678 }
3679 mgp->watchdog_pause = rx_pause_cnt;
3680
3681 if (reset_needed) {
3682 schedule_work(&mgp->watchdog_work);
3683 } else {
3684 /* rearm timer */
3685 mod_timer(&mgp->watchdog_timer,
3686 jiffies + myri10ge_watchdog_timeout * HZ);
3687 }
3688 }
3689
3690 static void myri10ge_free_slices(struct myri10ge_priv *mgp)
3691 {
3692 struct myri10ge_slice_state *ss;
3693 struct pci_dev *pdev = mgp->pdev;
3694 size_t bytes;
3695 int i;
3696
3697 if (mgp->ss == NULL)
3698 return;
3699
3700 for (i = 0; i < mgp->num_slices; i++) {
3701 ss = &mgp->ss[i];
3702 if (ss->rx_done.entry != NULL) {
3703 bytes = mgp->max_intr_slots *
3704 sizeof(*ss->rx_done.entry);
3705 dma_free_coherent(&pdev->dev, bytes,
3706 ss->rx_done.entry, ss->rx_done.bus);
3707 ss->rx_done.entry = NULL;
3708 }
3709 if (ss->fw_stats != NULL) {
3710 bytes = sizeof(*ss->fw_stats);
3711 dma_free_coherent(&pdev->dev, bytes,
3712 ss->fw_stats, ss->fw_stats_bus);
3713 ss->fw_stats = NULL;
3714 }
3715 netif_napi_del(&ss->napi);
3716 }
3717 kfree(mgp->ss);
3718 mgp->ss = NULL;
3719 }
3720
3721 static int myri10ge_alloc_slices(struct myri10ge_priv *mgp)
3722 {
3723 struct myri10ge_slice_state *ss;
3724 struct pci_dev *pdev = mgp->pdev;
3725 size_t bytes;
3726 int i;
3727
3728 bytes = sizeof(*mgp->ss) * mgp->num_slices;
3729 mgp->ss = kzalloc(bytes, GFP_KERNEL);
3730 if (mgp->ss == NULL) {
3731 return -ENOMEM;
3732 }
3733
3734 for (i = 0; i < mgp->num_slices; i++) {
3735 ss = &mgp->ss[i];
3736 bytes = mgp->max_intr_slots * sizeof(*ss->rx_done.entry);
3737 ss->rx_done.entry = dma_alloc_coherent(&pdev->dev, bytes,
3738 &ss->rx_done.bus,
3739 GFP_KERNEL);
3740 if (ss->rx_done.entry == NULL)
3741 goto abort;
3742 memset(ss->rx_done.entry, 0, bytes);
3743 bytes = sizeof(*ss->fw_stats);
3744 ss->fw_stats = dma_alloc_coherent(&pdev->dev, bytes,
3745 &ss->fw_stats_bus,
3746 GFP_KERNEL);
3747 if (ss->fw_stats == NULL)
3748 goto abort;
3749 ss->mgp = mgp;
3750 ss->dev = mgp->dev;
3751 netif_napi_add(ss->dev, &ss->napi, myri10ge_poll,
3752 myri10ge_napi_weight);
3753 }
3754 return 0;
3755 abort:
3756 myri10ge_free_slices(mgp);
3757 return -ENOMEM;
3758 }
3759
3760 /*
3761 * This function determines the number of slices supported.
3762 * The number slices is the minimum of the number of CPUS,
3763 * the number of MSI-X irqs supported, the number of slices
3764 * supported by the firmware
3765 */
3766 static void myri10ge_probe_slices(struct myri10ge_priv *mgp)
3767 {
3768 struct myri10ge_cmd cmd;
3769 struct pci_dev *pdev = mgp->pdev;
3770 char *old_fw;
3771 bool old_allocated;
3772 int i, status, ncpus, msix_cap;
3773
3774 mgp->num_slices = 1;
3775 msix_cap = pci_find_capability(pdev, PCI_CAP_ID_MSIX);
3776 ncpus = num_online_cpus();
3777
3778 if (myri10ge_max_slices == 1 || msix_cap == 0 ||
3779 (myri10ge_max_slices == -1 && ncpus < 2))
3780 return;
3781
3782 /* try to load the slice aware rss firmware */
3783 old_fw = mgp->fw_name;
3784 old_allocated = mgp->fw_name_allocated;
3785 /* don't free old_fw if we override it. */
3786 mgp->fw_name_allocated = false;
3787
3788 if (myri10ge_fw_name != NULL) {
3789 dev_info(&mgp->pdev->dev, "overriding rss firmware to %s\n",
3790 myri10ge_fw_name);
3791 set_fw_name(mgp, myri10ge_fw_name, false);
3792 } else if (old_fw == myri10ge_fw_aligned)
3793 set_fw_name(mgp, myri10ge_fw_rss_aligned, false);
3794 else
3795 set_fw_name(mgp, myri10ge_fw_rss_unaligned, false);
3796 status = myri10ge_load_firmware(mgp, 0);
3797 if (status != 0) {
3798 dev_info(&pdev->dev, "Rss firmware not found\n");
3799 if (old_allocated)
3800 kfree(old_fw);
3801 return;
3802 }
3803
3804 /* hit the board with a reset to ensure it is alive */
3805 memset(&cmd, 0, sizeof(cmd));
3806 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_RESET, &cmd, 0);
3807 if (status != 0) {
3808 dev_err(&mgp->pdev->dev, "failed reset\n");
3809 goto abort_with_fw;
3810 }
3811
3812 mgp->max_intr_slots = cmd.data0 / sizeof(struct mcp_slot);
3813
3814 /* tell it the size of the interrupt queues */
3815 cmd.data0 = mgp->max_intr_slots * sizeof(struct mcp_slot);
3816 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_SET_INTRQ_SIZE, &cmd, 0);
3817 if (status != 0) {
3818 dev_err(&mgp->pdev->dev, "failed MXGEFW_CMD_SET_INTRQ_SIZE\n");
3819 goto abort_with_fw;
3820 }
3821
3822 /* ask the maximum number of slices it supports */
3823 status = myri10ge_send_cmd(mgp, MXGEFW_CMD_GET_MAX_RSS_QUEUES, &cmd, 0);
3824 if (status != 0)
3825 goto abort_with_fw;
3826 else
3827 mgp->num_slices = cmd.data0;
3828
3829 /* Only allow multiple slices if MSI-X is usable */
3830 if (!myri10ge_msi) {
3831 goto abort_with_fw;
3832 }
3833
3834 /* if the admin did not specify a limit to how many
3835 * slices we should use, cap it automatically to the
3836 * number of CPUs currently online */
3837 if (myri10ge_max_slices == -1)
3838 myri10ge_max_slices = ncpus;
3839
3840 if (mgp->num_slices > myri10ge_max_slices)
3841 mgp->num_slices = myri10ge_max_slices;
3842
3843 /* Now try to allocate as many MSI-X vectors as we have
3844 * slices. We give up on MSI-X if we can only get a single
3845 * vector. */
3846
3847 mgp->msix_vectors = kcalloc(mgp->num_slices, sizeof(*mgp->msix_vectors),
3848 GFP_KERNEL);
3849 if (mgp->msix_vectors == NULL)
3850 goto disable_msix;
3851 for (i = 0; i < mgp->num_slices; i++) {
3852 mgp->msix_vectors[i].entry = i;
3853 }
3854
3855 while (mgp->num_slices > 1) {
3856 /* make sure it is a power of two */
3857 while (!is_power_of_2(mgp->num_slices))
3858 mgp->num_slices--;
3859 if (mgp->num_slices == 1)
3860 goto disable_msix;
3861 status = pci_enable_msix(pdev, mgp->msix_vectors,
3862 mgp->num_slices);
3863 if (status == 0) {
3864 pci_disable_msix(pdev);
3865 if (old_allocated)
3866 kfree(old_fw);
3867 return;
3868 }
3869 if (status > 0)
3870 mgp->num_slices = status;
3871 else
3872 goto disable_msix;
3873 }
3874
3875 disable_msix:
3876 if (mgp->msix_vectors != NULL) {
3877 kfree(mgp->msix_vectors);
3878 mgp->msix_vectors = NULL;
3879 }
3880
3881 abort_with_fw:
3882 mgp->num_slices = 1;
3883 set_fw_name(mgp, old_fw, old_allocated);
3884 myri10ge_load_firmware(mgp, 0);
3885 }
3886
3887 static const struct net_device_ops myri10ge_netdev_ops = {
3888 .ndo_open = myri10ge_open,
3889 .ndo_stop = myri10ge_close,
3890 .ndo_start_xmit = myri10ge_xmit,
3891 .ndo_get_stats64 = myri10ge_get_stats,
3892 .ndo_validate_addr = eth_validate_addr,
3893 .ndo_change_mtu = myri10ge_change_mtu,
3894 .ndo_fix_features = myri10ge_fix_features,
3895 .ndo_set_multicast_list = myri10ge_set_multicast_list,
3896 .ndo_set_mac_address = myri10ge_set_mac_address,
3897 };
3898
3899 static int myri10ge_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
3900 {
3901 struct net_device *netdev;
3902 struct myri10ge_priv *mgp;
3903 struct device *dev = &pdev->dev;
3904 int i;
3905 int status = -ENXIO;
3906 int dac_enabled;
3907 unsigned hdr_offset, ss_offset;
3908 static int board_number;
3909
3910 netdev = alloc_etherdev_mq(sizeof(*mgp), MYRI10GE_MAX_SLICES);
3911 if (netdev == NULL) {
3912 dev_err(dev, "Could not allocate ethernet device\n");
3913 return -ENOMEM;
3914 }
3915
3916 SET_NETDEV_DEV(netdev, &pdev->dev);
3917
3918 mgp = netdev_priv(netdev);
3919 mgp->dev = netdev;
3920 mgp->pdev = pdev;
3921 mgp->pause = myri10ge_flow_control;
3922 mgp->intr_coal_delay = myri10ge_intr_coal_delay;
3923 mgp->msg_enable = netif_msg_init(myri10ge_debug, MYRI10GE_MSG_DEFAULT);
3924 mgp->board_number = board_number;
3925 init_waitqueue_head(&mgp->down_wq);
3926
3927 if (pci_enable_device(pdev)) {
3928 dev_err(&pdev->dev, "pci_enable_device call failed\n");
3929 status = -ENODEV;
3930 goto abort_with_netdev;
3931 }
3932
3933 /* Find the vendor-specific cap so we can check
3934 * the reboot register later on */
3935 mgp->vendor_specific_offset
3936 = pci_find_capability(pdev, PCI_CAP_ID_VNDR);
3937
3938 /* Set our max read request to 4KB */
3939 status = pcie_set_readrq(pdev, 4096);
3940 if (status != 0) {
3941 dev_err(&pdev->dev, "Error %d writing PCI_EXP_DEVCTL\n",
3942 status);
3943 goto abort_with_enabled;
3944 }
3945
3946 myri10ge_mask_surprise_down(pdev);
3947 pci_set_master(pdev);
3948 dac_enabled = 1;
3949 status = pci_set_dma_mask(pdev, DMA_BIT_MASK(64));
3950 if (status != 0) {
3951 dac_enabled = 0;
3952 dev_err(&pdev->dev,
3953 "64-bit pci address mask was refused, "
3954 "trying 32-bit\n");
3955 status = pci_set_dma_mask(pdev, DMA_BIT_MASK(32));
3956 }
3957 if (status != 0) {
3958 dev_err(&pdev->dev, "Error %d setting DMA mask\n", status);
3959 goto abort_with_enabled;
3960 }
3961 (void)pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64));
3962 mgp->cmd = dma_alloc_coherent(&pdev->dev, sizeof(*mgp->cmd),
3963 &mgp->cmd_bus, GFP_KERNEL);
3964 if (mgp->cmd == NULL)
3965 goto abort_with_enabled;
3966
3967 mgp->board_span = pci_resource_len(pdev, 0);
3968 mgp->iomem_base = pci_resource_start(pdev, 0);
3969 mgp->mtrr = -1;
3970 mgp->wc_enabled = 0;
3971 #ifdef CONFIG_MTRR
3972 mgp->mtrr = mtrr_add(mgp->iomem_base, mgp->board_span,
3973 MTRR_TYPE_WRCOMB, 1);
3974 if (mgp->mtrr >= 0)
3975 mgp->wc_enabled = 1;
3976 #endif
3977 mgp->sram = ioremap_wc(mgp->iomem_base, mgp->board_span);
3978 if (mgp->sram == NULL) {
3979 dev_err(&pdev->dev, "ioremap failed for %ld bytes at 0x%lx\n",
3980 mgp->board_span, mgp->iomem_base);
3981 status = -ENXIO;
3982 goto abort_with_mtrr;
3983 }
3984 hdr_offset =
3985 ntohl(__raw_readl(mgp->sram + MCP_HEADER_PTR_OFFSET)) & 0xffffc;
3986 ss_offset = hdr_offset + offsetof(struct mcp_gen_header, string_specs);
3987 mgp->sram_size = ntohl(__raw_readl(mgp->sram + ss_offset));
3988 if (mgp->sram_size > mgp->board_span ||
3989 mgp->sram_size <= MYRI10GE_FW_OFFSET) {
3990 dev_err(&pdev->dev,
3991 "invalid sram_size %dB or board span %ldB\n",
3992 mgp->sram_size, mgp->board_span);
3993 goto abort_with_ioremap;
3994 }
3995 memcpy_fromio(mgp->eeprom_strings,
3996 mgp->sram + mgp->sram_size, MYRI10GE_EEPROM_STRINGS_SIZE);
3997 memset(mgp->eeprom_strings + MYRI10GE_EEPROM_STRINGS_SIZE - 2, 0, 2);
3998 status = myri10ge_read_mac_addr(mgp);
3999 if (status)
4000 goto abort_with_ioremap;
4001
4002 for (i = 0; i < ETH_ALEN; i++)
4003 netdev->dev_addr[i] = mgp->mac_addr[i];
4004
4005 myri10ge_select_firmware(mgp);
4006
4007 status = myri10ge_load_firmware(mgp, 1);
4008 if (status != 0) {
4009 dev_err(&pdev->dev, "failed to load firmware\n");
4010 goto abort_with_ioremap;
4011 }
4012 myri10ge_probe_slices(mgp);
4013 status = myri10ge_alloc_slices(mgp);
4014 if (status != 0) {
4015 dev_err(&pdev->dev, "failed to alloc slice state\n");
4016 goto abort_with_firmware;
4017 }
4018 netif_set_real_num_tx_queues(netdev, mgp->num_slices);
4019 netif_set_real_num_rx_queues(netdev, mgp->num_slices);
4020 status = myri10ge_reset(mgp);
4021 if (status != 0) {
4022 dev_err(&pdev->dev, "failed reset\n");
4023 goto abort_with_slices;
4024 }
4025 #ifdef CONFIG_MYRI10GE_DCA
4026 myri10ge_setup_dca(mgp);
4027 #endif
4028 pci_set_drvdata(pdev, mgp);
4029 if ((myri10ge_initial_mtu + ETH_HLEN) > MYRI10GE_MAX_ETHER_MTU)
4030 myri10ge_initial_mtu = MYRI10GE_MAX_ETHER_MTU - ETH_HLEN;
4031 if ((myri10ge_initial_mtu + ETH_HLEN) < 68)
4032 myri10ge_initial_mtu = 68;
4033
4034 netdev->netdev_ops = &myri10ge_netdev_ops;
4035 netdev->mtu = myri10ge_initial_mtu;
4036 netdev->base_addr = mgp->iomem_base;
4037 netdev->hw_features = mgp->features | NETIF_F_LRO | NETIF_F_RXCSUM;
4038 netdev->features = netdev->hw_features;
4039
4040 if (dac_enabled)
4041 netdev->features |= NETIF_F_HIGHDMA;
4042
4043 netdev->vlan_features |= mgp->features;
4044 if (mgp->fw_ver_tiny < 37)
4045 netdev->vlan_features &= ~NETIF_F_TSO6;
4046 if (mgp->fw_ver_tiny < 32)
4047 netdev->vlan_features &= ~NETIF_F_TSO;
4048
4049 /* make sure we can get an irq, and that MSI can be
4050 * setup (if available). Also ensure netdev->irq
4051 * is set to correct value if MSI is enabled */
4052 status = myri10ge_request_irq(mgp);
4053 if (status != 0)
4054 goto abort_with_firmware;
4055 netdev->irq = pdev->irq;
4056 myri10ge_free_irq(mgp);
4057
4058 /* Save configuration space to be restored if the
4059 * nic resets due to a parity error */
4060 pci_save_state(pdev);
4061
4062 /* Setup the watchdog timer */
4063 setup_timer(&mgp->watchdog_timer, myri10ge_watchdog_timer,
4064 (unsigned long)mgp);
4065
4066 SET_ETHTOOL_OPS(netdev, &myri10ge_ethtool_ops);
4067 INIT_WORK(&mgp->watchdog_work, myri10ge_watchdog);
4068 status = register_netdev(netdev);
4069 if (status != 0) {
4070 dev_err(&pdev->dev, "register_netdev failed: %d\n", status);
4071 goto abort_with_state;
4072 }
4073 if (mgp->msix_enabled)
4074 dev_info(dev, "%d MSI-X IRQs, tx bndry %d, fw %s, WC %s\n",
4075 mgp->num_slices, mgp->tx_boundary, mgp->fw_name,
4076 (mgp->wc_enabled ? "Enabled" : "Disabled"));
4077 else
4078 dev_info(dev, "%s IRQ %d, tx bndry %d, fw %s, WC %s\n",
4079 mgp->msi_enabled ? "MSI" : "xPIC",
4080 netdev->irq, mgp->tx_boundary, mgp->fw_name,
4081 (mgp->wc_enabled ? "Enabled" : "Disabled"));
4082
4083 board_number++;
4084 return 0;
4085
4086 abort_with_state:
4087 pci_restore_state(pdev);
4088
4089 abort_with_slices:
4090 myri10ge_free_slices(mgp);
4091
4092 abort_with_firmware:
4093 myri10ge_dummy_rdma(mgp, 0);
4094
4095 abort_with_ioremap:
4096 if (mgp->mac_addr_string != NULL)
4097 dev_err(&pdev->dev,
4098 "myri10ge_probe() failed: MAC=%s, SN=%ld\n",
4099 mgp->mac_addr_string, mgp->serial_number);
4100 iounmap(mgp->sram);
4101
4102 abort_with_mtrr:
4103 #ifdef CONFIG_MTRR
4104 if (mgp->mtrr >= 0)
4105 mtrr_del(mgp->mtrr, mgp->iomem_base, mgp->board_span);
4106 #endif
4107 dma_free_coherent(&pdev->dev, sizeof(*mgp->cmd),
4108 mgp->cmd, mgp->cmd_bus);
4109
4110 abort_with_enabled:
4111 pci_disable_device(pdev);
4112
4113 abort_with_netdev:
4114 set_fw_name(mgp, NULL, false);
4115 free_netdev(netdev);
4116 return status;
4117 }
4118
4119 /*
4120 * myri10ge_remove
4121 *
4122 * Does what is necessary to shutdown one Myrinet device. Called
4123 * once for each Myrinet card by the kernel when a module is
4124 * unloaded.
4125 */
4126 static void myri10ge_remove(struct pci_dev *pdev)
4127 {
4128 struct myri10ge_priv *mgp;
4129 struct net_device *netdev;
4130
4131 mgp = pci_get_drvdata(pdev);
4132 if (mgp == NULL)
4133 return;
4134
4135 cancel_work_sync(&mgp->watchdog_work);
4136 netdev = mgp->dev;
4137 unregister_netdev(netdev);
4138
4139 #ifdef CONFIG_MYRI10GE_DCA
4140 myri10ge_teardown_dca(mgp);
4141 #endif
4142 myri10ge_dummy_rdma(mgp, 0);
4143
4144 /* avoid a memory leak */
4145 pci_restore_state(pdev);
4146
4147 iounmap(mgp->sram);
4148
4149 #ifdef CONFIG_MTRR
4150 if (mgp->mtrr >= 0)
4151 mtrr_del(mgp->mtrr, mgp->iomem_base, mgp->board_span);
4152 #endif
4153 myri10ge_free_slices(mgp);
4154 if (mgp->msix_vectors != NULL)
4155 kfree(mgp->msix_vectors);
4156 dma_free_coherent(&pdev->dev, sizeof(*mgp->cmd),
4157 mgp->cmd, mgp->cmd_bus);
4158
4159 set_fw_name(mgp, NULL, false);
4160 free_netdev(netdev);
4161 pci_disable_device(pdev);
4162 pci_set_drvdata(pdev, NULL);
4163 }
4164
4165 #define PCI_DEVICE_ID_MYRICOM_MYRI10GE_Z8E 0x0008
4166 #define PCI_DEVICE_ID_MYRICOM_MYRI10GE_Z8E_9 0x0009
4167
4168 static DEFINE_PCI_DEVICE_TABLE(myri10ge_pci_tbl) = {
4169 {PCI_DEVICE(PCI_VENDOR_ID_MYRICOM, PCI_DEVICE_ID_MYRICOM_MYRI10GE_Z8E)},
4170 {PCI_DEVICE
4171 (PCI_VENDOR_ID_MYRICOM, PCI_DEVICE_ID_MYRICOM_MYRI10GE_Z8E_9)},
4172 {0},
4173 };
4174
4175 MODULE_DEVICE_TABLE(pci, myri10ge_pci_tbl);
4176
4177 static struct pci_driver myri10ge_driver = {
4178 .name = "myri10ge",
4179 .probe = myri10ge_probe,
4180 .remove = myri10ge_remove,
4181 .id_table = myri10ge_pci_tbl,
4182 #ifdef CONFIG_PM
4183 .suspend = myri10ge_suspend,
4184 .resume = myri10ge_resume,
4185 #endif
4186 };
4187
4188 #ifdef CONFIG_MYRI10GE_DCA
4189 static int
4190 myri10ge_notify_dca(struct notifier_block *nb, unsigned long event, void *p)
4191 {
4192 int err = driver_for_each_device(&myri10ge_driver.driver,
4193 NULL, &event,
4194 myri10ge_notify_dca_device);
4195
4196 if (err)
4197 return NOTIFY_BAD;
4198 return NOTIFY_DONE;
4199 }
4200
4201 static struct notifier_block myri10ge_dca_notifier = {
4202 .notifier_call = myri10ge_notify_dca,
4203 .next = NULL,
4204 .priority = 0,
4205 };
4206 #endif /* CONFIG_MYRI10GE_DCA */
4207
4208 static __init int myri10ge_init_module(void)
4209 {
4210 pr_info("Version %s\n", MYRI10GE_VERSION_STR);
4211
4212 if (myri10ge_rss_hash > MXGEFW_RSS_HASH_TYPE_MAX) {
4213 pr_err("Illegal rssh hash type %d, defaulting to source port\n",
4214 myri10ge_rss_hash);
4215 myri10ge_rss_hash = MXGEFW_RSS_HASH_TYPE_SRC_PORT;
4216 }
4217 #ifdef CONFIG_MYRI10GE_DCA
4218 dca_register_notify(&myri10ge_dca_notifier);
4219 #endif
4220 if (myri10ge_max_slices > MYRI10GE_MAX_SLICES)
4221 myri10ge_max_slices = MYRI10GE_MAX_SLICES;
4222
4223 return pci_register_driver(&myri10ge_driver);
4224 }
4225
4226 module_init(myri10ge_init_module);
4227
4228 static __exit void myri10ge_cleanup_module(void)
4229 {
4230 #ifdef CONFIG_MYRI10GE_DCA
4231 dca_unregister_notify(&myri10ge_dca_notifier);
4232 #endif
4233 pci_unregister_driver(&myri10ge_driver);
4234 }
4235
4236 module_exit(myri10ge_cleanup_module);
CJK support for tty framebuffer vt