1 /**************************************************************************
2
3 Copyright (c) 2007-2009, Chelsio Inc.
4 All rights reserved.
5
6 Redistribution and use in source and binary forms, with or without
7 modification, are permitted provided that the following conditions are met:
8
9 1. Redistributions of source code must retain the above copyright notice,
10 this list of conditions and the following disclaimer.
11
12 2. Neither the name of the Chelsio Corporation nor the names of its
13 contributors may be used to endorse or promote products derived from
14 this software without specific prior written permission.
15
16 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
17 AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18 IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19 ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
20 LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
21 CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
22 SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
23 INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
24 CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
25 ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
26 POSSIBILITY OF SUCH DAMAGE.
27
28 ***************************************************************************/
29
30 #include <sys/cdefs.h>
31 __FBSDID("$FreeBSD$");
32
33 #include "opt_inet.h"
34
35 #include <sys/param.h>
36 #include <sys/systm.h>
37 #include <sys/kernel.h>
38 #include <sys/bus.h>
39 #include <sys/module.h>
40 #include <sys/pciio.h>
41 #include <sys/conf.h>
42 #include <machine/bus.h>
43 #include <machine/resource.h>
44 #include <sys/bus_dma.h>
45 #include <sys/ktr.h>
46 #include <sys/rman.h>
47 #include <sys/ioccom.h>
48 #include <sys/mbuf.h>
49 #include <sys/linker.h>
50 #include <sys/firmware.h>
51 #include <sys/socket.h>
52 #include <sys/sockio.h>
53 #include <sys/smp.h>
54 #include <sys/sysctl.h>
55 #include <sys/syslog.h>
56 #include <sys/queue.h>
57 #include <sys/taskqueue.h>
58 #include <sys/proc.h>
59
60 #include <net/bpf.h>
61 #include <net/ethernet.h>
62 #include <net/if.h>
63 #include <net/if_var.h>
64 #include <net/if_arp.h>
65 #include <net/if_dl.h>
66 #include <net/if_media.h>
67 #include <net/if_types.h>
68 #include <net/if_vlan_var.h>
69
70 #include <netinet/in_systm.h>
71 #include <netinet/in.h>
72 #include <netinet/if_ether.h>
73 #include <netinet/ip.h>
74 #include <netinet/ip.h>
75 #include <netinet/tcp.h>
76 #include <netinet/udp.h>
77
78 #include <dev/pci/pcireg.h>
79 #include <dev/pci/pcivar.h>
80 #include <dev/pci/pci_private.h>
81
82 #include <cxgb_include.h>
83
84 #ifdef PRIV_SUPPORTED
85 #include <sys/priv.h>
86 #endif
87
88 static int cxgb_setup_interrupts(adapter_t *);
89 static void cxgb_teardown_interrupts(adapter_t *);
90 static void cxgb_init(void *);
91 static int cxgb_init_locked(struct port_info *);
92 static int cxgb_uninit_locked(struct port_info *);
93 static int cxgb_uninit_synchronized(struct port_info *);
94 static int cxgb_ioctl(struct ifnet *, unsigned long, caddr_t);
95 static int cxgb_media_change(struct ifnet *);
96 static int cxgb_ifm_type(int);
97 static void cxgb_build_medialist(struct port_info *);
98 static void cxgb_media_status(struct ifnet *, struct ifmediareq *);
99 static uint64_t cxgb_get_counter(struct ifnet *, ift_counter);
100 static int setup_sge_qsets(adapter_t *);
101 static void cxgb_async_intr(void *);
102 static void cxgb_tick_handler(void *, int);
103 static void cxgb_tick(void *);
104 static void link_check_callout(void *);
105 static void check_link_status(void *, int);
106 static void setup_rss(adapter_t *sc);
107 static int alloc_filters(struct adapter *);
108 static int setup_hw_filters(struct adapter *);
109 static int set_filter(struct adapter *, int, const struct filter_info *);
110 static inline void mk_set_tcb_field(struct cpl_set_tcb_field *, unsigned int,
111 unsigned int, u64, u64);
112 static inline void set_tcb_field_ulp(struct cpl_set_tcb_field *, unsigned int,
113 unsigned int, u64, u64);
114 #ifdef TCP_OFFLOAD
115 static int cpl_not_handled(struct sge_qset *, struct rsp_desc *, struct mbuf *);
116 #endif
117
118 /* Attachment glue for the PCI controller end of the device. Each port of
119 * the device is attached separately, as defined later.
120 */
121 static int cxgb_controller_probe(device_t);
122 static int cxgb_controller_attach(device_t);
123 static int cxgb_controller_detach(device_t);
124 static void cxgb_free(struct adapter *);
125 static __inline void reg_block_dump(struct adapter *ap, uint8_t *buf, unsigned int start,
126 unsigned int end);
127 static void cxgb_get_regs(adapter_t *sc, struct ch_ifconf_regs *regs, uint8_t *buf);
128 static int cxgb_get_regs_len(void);
129 static void touch_bars(device_t dev);
130 static void cxgb_update_mac_settings(struct port_info *p);
131 #ifdef TCP_OFFLOAD
132 static int toe_capability(struct port_info *, int);
133 #endif
134
135 static device_method_t cxgb_controller_methods[] = {
136 DEVMETHOD(device_probe, cxgb_controller_probe),
137 DEVMETHOD(device_attach, cxgb_controller_attach),
138 DEVMETHOD(device_detach, cxgb_controller_detach),
139
140 DEVMETHOD_END
141 };
142
143 static driver_t cxgb_controller_driver = {
144 "cxgbc",
145 cxgb_controller_methods,
146 sizeof(struct adapter)
147 };
148
149 static int cxgbc_mod_event(module_t, int, void *);
150 static devclass_t cxgb_controller_devclass;
151 DRIVER_MODULE(cxgbc, pci, cxgb_controller_driver, cxgb_controller_devclass,
152 cxgbc_mod_event, 0);
153 MODULE_VERSION(cxgbc, 1);
154 MODULE_DEPEND(cxgbc, firmware, 1, 1, 1);
155
156 /*
157 * Attachment glue for the ports. Attachment is done directly to the
158 * controller device.
159 */
160 static int cxgb_port_probe(device_t);
161 static int cxgb_port_attach(device_t);
162 static int cxgb_port_detach(device_t);
163
164 static device_method_t cxgb_port_methods[] = {
165 DEVMETHOD(device_probe, cxgb_port_probe),
166 DEVMETHOD(device_attach, cxgb_port_attach),
167 DEVMETHOD(device_detach, cxgb_port_detach),
168 { 0, 0 }
169 };
170
171 static driver_t cxgb_port_driver = {
172 "cxgb",
173 cxgb_port_methods,
174 0
175 };
176
177 static d_ioctl_t cxgb_extension_ioctl;
178 static d_open_t cxgb_extension_open;
179 static d_close_t cxgb_extension_close;
180
181 static struct cdevsw cxgb_cdevsw = {
182 .d_version = D_VERSION,
183 .d_flags = 0,
184 .d_open = cxgb_extension_open,
185 .d_close = cxgb_extension_close,
186 .d_ioctl = cxgb_extension_ioctl,
187 .d_name = "cxgb",
188 };
189
190 static devclass_t cxgb_port_devclass;
191 DRIVER_MODULE(cxgb, cxgbc, cxgb_port_driver, cxgb_port_devclass, 0, 0);
192 MODULE_VERSION(cxgb, 1);
193
194 static struct mtx t3_list_lock;
195 static SLIST_HEAD(, adapter) t3_list;
196 #ifdef TCP_OFFLOAD
197 static struct mtx t3_uld_list_lock;
198 static SLIST_HEAD(, uld_info) t3_uld_list;
199 #endif
200
201 /*
202 * The driver uses the best interrupt scheme available on a platform in the
203 * order MSI-X, MSI, legacy pin interrupts. This parameter determines which
204 * of these schemes the driver may consider as follows:
205 *
206 * msi = 2: choose from among all three options
207 * msi = 1 : only consider MSI and pin interrupts
208 * msi = 0: force pin interrupts
209 */
210 static int msi_allowed = 2;
211
212 SYSCTL_NODE(_hw, OID_AUTO, cxgb, CTLFLAG_RD, 0, "CXGB driver parameters");
213 SYSCTL_INT(_hw_cxgb, OID_AUTO, msi_allowed, CTLFLAG_RDTUN, &msi_allowed, 0,
214 "MSI-X, MSI, INTx selector");
215
216 /*
217 * The driver uses an auto-queue algorithm by default.
218 * To disable it and force a single queue-set per port, use multiq = 0
219 */
220 static int multiq = 1;
221 SYSCTL_INT(_hw_cxgb, OID_AUTO, multiq, CTLFLAG_RDTUN, &multiq, 0,
222 "use min(ncpus/ports, 8) queue-sets per port");
223
224 /*
225 * By default the driver will not update the firmware unless
226 * it was compiled against a newer version
227 *
228 */
229 static int force_fw_update = 0;
230 SYSCTL_INT(_hw_cxgb, OID_AUTO, force_fw_update, CTLFLAG_RDTUN, &force_fw_update, 0,
231 "update firmware even if up to date");
232
233 int cxgb_use_16k_clusters = -1;
234 SYSCTL_INT(_hw_cxgb, OID_AUTO, use_16k_clusters, CTLFLAG_RDTUN,
235 &cxgb_use_16k_clusters, 0, "use 16kB clusters for the jumbo queue ");
236
237 static int nfilters = -1;
238 SYSCTL_INT(_hw_cxgb, OID_AUTO, nfilters, CTLFLAG_RDTUN,
239 &nfilters, 0, "max number of entries in the filter table");
240
241 enum {
242 MAX_TXQ_ENTRIES = 16384,
243 MAX_CTRL_TXQ_ENTRIES = 1024,
244 MAX_RSPQ_ENTRIES = 16384,
245 MAX_RX_BUFFERS = 16384,
246 MAX_RX_JUMBO_BUFFERS = 16384,
247 MIN_TXQ_ENTRIES = 4,
248 MIN_CTRL_TXQ_ENTRIES = 4,
249 MIN_RSPQ_ENTRIES = 32,
250 MIN_FL_ENTRIES = 32,
251 MIN_FL_JUMBO_ENTRIES = 32
252 };
253
254 struct filter_info {
255 u32 sip;
256 u32 sip_mask;
257 u32 dip;
258 u16 sport;
259 u16 dport;
260 u32 vlan:12;
261 u32 vlan_prio:3;
262 u32 mac_hit:1;
263 u32 mac_idx:4;
264 u32 mac_vld:1;
265 u32 pkt_type:2;
266 u32 report_filter_id:1;
267 u32 pass:1;
268 u32 rss:1;
269 u32 qset:3;
270 u32 locked:1;
271 u32 valid:1;
272 };
273
274 enum { FILTER_NO_VLAN_PRI = 7 };
275
276 #define EEPROM_MAGIC 0x38E2F10C
277
278 #define PORT_MASK ((1 << MAX_NPORTS) - 1)
279
280 /* Table for probing the cards. The desc field isn't actually used */
281 struct cxgb_ident {
282 uint16_t vendor;
283 uint16_t device;
284 int index;
285 char *desc;
286 } cxgb_identifiers[] = {
287 {PCI_VENDOR_ID_CHELSIO, 0x0020, 0, "PE9000"},
288 {PCI_VENDOR_ID_CHELSIO, 0x0021, 1, "T302E"},
289 {PCI_VENDOR_ID_CHELSIO, 0x0022, 2, "T310E"},
290 {PCI_VENDOR_ID_CHELSIO, 0x0023, 3, "T320X"},
291 {PCI_VENDOR_ID_CHELSIO, 0x0024, 1, "T302X"},
292 {PCI_VENDOR_ID_CHELSIO, 0x0025, 3, "T320E"},
293 {PCI_VENDOR_ID_CHELSIO, 0x0026, 2, "T310X"},
294 {PCI_VENDOR_ID_CHELSIO, 0x0030, 2, "T3B10"},
295 {PCI_VENDOR_ID_CHELSIO, 0x0031, 3, "T3B20"},
296 {PCI_VENDOR_ID_CHELSIO, 0x0032, 1, "T3B02"},
297 {PCI_VENDOR_ID_CHELSIO, 0x0033, 4, "T3B04"},
298 {PCI_VENDOR_ID_CHELSIO, 0x0035, 6, "T3C10"},
299 {PCI_VENDOR_ID_CHELSIO, 0x0036, 3, "S320E-CR"},
300 {PCI_VENDOR_ID_CHELSIO, 0x0037, 7, "N320E-G2"},
301 {0, 0, 0, NULL}
302 };
303
304 static int set_eeprom(struct port_info *pi, const uint8_t *data, int len, int offset);
305
306
307 static __inline char
t3rev2char(struct adapter * adapter)308 t3rev2char(struct adapter *adapter)
309 {
310 char rev = 'z';
311
312 switch(adapter->params.rev) {
313 case T3_REV_A:
314 rev = 'a';
315 break;
316 case T3_REV_B:
317 case T3_REV_B2:
318 rev = 'b';
319 break;
320 case T3_REV_C:
321 rev = 'c';
322 break;
323 }
324 return rev;
325 }
326
327 static struct cxgb_ident *
cxgb_get_ident(device_t dev)328 cxgb_get_ident(device_t dev)
329 {
330 struct cxgb_ident *id;
331
332 for (id = cxgb_identifiers; id->desc != NULL; id++) {
333 if ((id->vendor == pci_get_vendor(dev)) &&
334 (id->device == pci_get_device(dev))) {
335 return (id);
336 }
337 }
338 return (NULL);
339 }
340
341 static const struct adapter_info *
cxgb_get_adapter_info(device_t dev)342 cxgb_get_adapter_info(device_t dev)
343 {
344 struct cxgb_ident *id;
345 const struct adapter_info *ai;
346
347 id = cxgb_get_ident(dev);
348 if (id == NULL)
349 return (NULL);
350
351 ai = t3_get_adapter_info(id->index);
352
353 return (ai);
354 }
355
356 static int
cxgb_controller_probe(device_t dev)357 cxgb_controller_probe(device_t dev)
358 {
359 const struct adapter_info *ai;
360 char *ports, buf[80];
361 int nports;
362
363 ai = cxgb_get_adapter_info(dev);
364 if (ai == NULL)
365 return (ENXIO);
366
367 nports = ai->nports0 + ai->nports1;
368 if (nports == 1)
369 ports = "port";
370 else
371 ports = "ports";
372
373 snprintf(buf, sizeof(buf), "%s, %d %s", ai->desc, nports, ports);
374 device_set_desc_copy(dev, buf);
375 return (BUS_PROBE_DEFAULT);
376 }
377
378 #define FW_FNAME "cxgb_t3fw"
379 #define TPEEPROM_NAME "cxgb_t3%c_tp_eeprom"
380 #define TPSRAM_NAME "cxgb_t3%c_protocol_sram"
381
382 static int
upgrade_fw(adapter_t * sc)383 upgrade_fw(adapter_t *sc)
384 {
385 const struct firmware *fw;
386 int status;
387 u32 vers;
388
389 if ((fw = firmware_get(FW_FNAME)) == NULL) {
390 device_printf(sc->dev, "Could not find firmware image %s\n", FW_FNAME);
391 return (ENOENT);
392 } else
393 device_printf(sc->dev, "installing firmware on card\n");
394 status = t3_load_fw(sc, (const uint8_t *)fw->data, fw->datasize);
395
396 if (status != 0) {
397 device_printf(sc->dev, "failed to install firmware: %d\n",
398 status);
399 } else {
400 t3_get_fw_version(sc, &vers);
401 snprintf(&sc->fw_version[0], sizeof(sc->fw_version), "%d.%d.%d",
402 G_FW_VERSION_MAJOR(vers), G_FW_VERSION_MINOR(vers),
403 G_FW_VERSION_MICRO(vers));
404 }
405
406 firmware_put(fw, FIRMWARE_UNLOAD);
407
408 return (status);
409 }
410
411 /*
412 * The cxgb_controller_attach function is responsible for the initial
413 * bringup of the device. Its responsibilities include:
414 *
415 * 1. Determine if the device supports MSI or MSI-X.
416 * 2. Allocate bus resources so that we can access the Base Address Register
417 * 3. Create and initialize mutexes for the controller and its control
418 * logic such as SGE and MDIO.
419 * 4. Call hardware specific setup routine for the adapter as a whole.
420 * 5. Allocate the BAR for doing MSI-X.
421 * 6. Setup the line interrupt iff MSI-X is not supported.
422 * 7. Create the driver's taskq.
423 * 8. Start one task queue service thread.
424 * 9. Check if the firmware and SRAM are up-to-date. They will be
425 * auto-updated later (before FULL_INIT_DONE), if required.
426 * 10. Create a child device for each MAC (port)
427 * 11. Initialize T3 private state.
428 * 12. Trigger the LED
429 * 13. Setup offload iff supported.
430 * 14. Reset/restart the tick callout.
431 * 15. Attach sysctls
432 *
433 * NOTE: Any modification or deviation from this list MUST be reflected in
434 * the above comment. Failure to do so will result in problems on various
435 * error conditions including link flapping.
436 */
437 static int
cxgb_controller_attach(device_t dev)438 cxgb_controller_attach(device_t dev)
439 {
440 device_t child;
441 const struct adapter_info *ai;
442 struct adapter *sc;
443 int i, error = 0;
444 uint32_t vers;
445 int port_qsets = 1;
446 int msi_needed, reg;
447 char buf[80];
448
449 sc = device_get_softc(dev);
450 sc->dev = dev;
451 sc->msi_count = 0;
452 ai = cxgb_get_adapter_info(dev);
453
454 snprintf(sc->lockbuf, ADAPTER_LOCK_NAME_LEN, "cxgb controller lock %d",
455 device_get_unit(dev));
456 ADAPTER_LOCK_INIT(sc, sc->lockbuf);
457
458 snprintf(sc->reglockbuf, ADAPTER_LOCK_NAME_LEN, "SGE reg lock %d",
459 device_get_unit(dev));
460 snprintf(sc->mdiolockbuf, ADAPTER_LOCK_NAME_LEN, "cxgb mdio lock %d",
461 device_get_unit(dev));
462 snprintf(sc->elmerlockbuf, ADAPTER_LOCK_NAME_LEN, "cxgb elmer lock %d",
463 device_get_unit(dev));
464
465 MTX_INIT(&sc->sge.reg_lock, sc->reglockbuf, NULL, MTX_SPIN);
466 MTX_INIT(&sc->mdio_lock, sc->mdiolockbuf, NULL, MTX_DEF);
467 MTX_INIT(&sc->elmer_lock, sc->elmerlockbuf, NULL, MTX_DEF);
468
469 mtx_lock(&t3_list_lock);
470 SLIST_INSERT_HEAD(&t3_list, sc, link);
471 mtx_unlock(&t3_list_lock);
472
473 /* find the PCIe link width and set max read request to 4KB*/
474 if (pci_find_cap(dev, PCIY_EXPRESS, ®) == 0) {
475 uint16_t lnk;
476
477 lnk = pci_read_config(dev, reg + PCIER_LINK_STA, 2);
478 sc->link_width = (lnk & PCIEM_LINK_STA_WIDTH) >> 4;
479 if (sc->link_width < 8 &&
480 (ai->caps & SUPPORTED_10000baseT_Full)) {
481 device_printf(sc->dev,
482 "PCIe x%d Link, expect reduced performance\n",
483 sc->link_width);
484 }
485
486 pci_set_max_read_req(dev, 4096);
487 }
488
489 touch_bars(dev);
490 pci_enable_busmaster(dev);
491 /*
492 * Allocate the registers and make them available to the driver.
493 * The registers that we care about for NIC mode are in BAR 0
494 */
495 sc->regs_rid = PCIR_BAR(0);
496 if ((sc->regs_res = bus_alloc_resource_any(dev, SYS_RES_MEMORY,
497 &sc->regs_rid, RF_ACTIVE)) == NULL) {
498 device_printf(dev, "Cannot allocate BAR region 0\n");
499 error = ENXIO;
500 goto out;
501 }
502
503 sc->bt = rman_get_bustag(sc->regs_res);
504 sc->bh = rman_get_bushandle(sc->regs_res);
505 sc->mmio_len = rman_get_size(sc->regs_res);
506
507 for (i = 0; i < MAX_NPORTS; i++)
508 sc->port[i].adapter = sc;
509
510 if (t3_prep_adapter(sc, ai, 1) < 0) {
511 printf("prep adapter failed\n");
512 error = ENODEV;
513 goto out;
514 }
515
516 sc->udbs_rid = PCIR_BAR(2);
517 sc->udbs_res = NULL;
518 if (is_offload(sc) &&
519 ((sc->udbs_res = bus_alloc_resource_any(dev, SYS_RES_MEMORY,
520 &sc->udbs_rid, RF_ACTIVE)) == NULL)) {
521 device_printf(dev, "Cannot allocate BAR region 1\n");
522 error = ENXIO;
523 goto out;
524 }
525
526 /* Allocate the BAR for doing MSI-X. If it succeeds, try to allocate
527 * enough messages for the queue sets. If that fails, try falling
528 * back to MSI. If that fails, then try falling back to the legacy
529 * interrupt pin model.
530 */
531 sc->msix_regs_rid = 0x20;
532 if ((msi_allowed >= 2) &&
533 (sc->msix_regs_res = bus_alloc_resource_any(dev, SYS_RES_MEMORY,
534 &sc->msix_regs_rid, RF_ACTIVE)) != NULL) {
535
536 if (multiq)
537 port_qsets = min(SGE_QSETS/sc->params.nports, mp_ncpus);
538 msi_needed = sc->msi_count = sc->params.nports * port_qsets + 1;
539
540 if (pci_msix_count(dev) == 0 ||
541 (error = pci_alloc_msix(dev, &sc->msi_count)) != 0 ||
542 sc->msi_count != msi_needed) {
543 device_printf(dev, "alloc msix failed - "
544 "msi_count=%d, msi_needed=%d, err=%d; "
545 "will try MSI\n", sc->msi_count,
546 msi_needed, error);
547 sc->msi_count = 0;
548 port_qsets = 1;
549 pci_release_msi(dev);
550 bus_release_resource(dev, SYS_RES_MEMORY,
551 sc->msix_regs_rid, sc->msix_regs_res);
552 sc->msix_regs_res = NULL;
553 } else {
554 sc->flags |= USING_MSIX;
555 sc->cxgb_intr = cxgb_async_intr;
556 device_printf(dev,
557 "using MSI-X interrupts (%u vectors)\n",
558 sc->msi_count);
559 }
560 }
561
562 if ((msi_allowed >= 1) && (sc->msi_count == 0)) {
563 sc->msi_count = 1;
564 if ((error = pci_alloc_msi(dev, &sc->msi_count)) != 0) {
565 device_printf(dev, "alloc msi failed - "
566 "err=%d; will try INTx\n", error);
567 sc->msi_count = 0;
568 port_qsets = 1;
569 pci_release_msi(dev);
570 } else {
571 sc->flags |= USING_MSI;
572 sc->cxgb_intr = t3_intr_msi;
573 device_printf(dev, "using MSI interrupts\n");
574 }
575 }
576 if (sc->msi_count == 0) {
577 device_printf(dev, "using line interrupts\n");
578 sc->cxgb_intr = t3b_intr;
579 }
580
581 /* Create a private taskqueue thread for handling driver events */
582 sc->tq = taskqueue_create("cxgb_taskq", M_NOWAIT,
583 taskqueue_thread_enqueue, &sc->tq);
584 if (sc->tq == NULL) {
585 device_printf(dev, "failed to allocate controller task queue\n");
586 goto out;
587 }
588
589 taskqueue_start_threads(&sc->tq, 1, PI_NET, "%s taskq",
590 device_get_nameunit(dev));
591 TASK_INIT(&sc->tick_task, 0, cxgb_tick_handler, sc);
592
593
594 /* Create a periodic callout for checking adapter status */
595 callout_init(&sc->cxgb_tick_ch, 1);
596
597 if (t3_check_fw_version(sc) < 0 || force_fw_update) {
598 /*
599 * Warn user that a firmware update will be attempted in init.
600 */
601 device_printf(dev, "firmware needs to be updated to version %d.%d.%d\n",
602 FW_VERSION_MAJOR, FW_VERSION_MINOR, FW_VERSION_MICRO);
603 sc->flags &= ~FW_UPTODATE;
604 } else {
605 sc->flags |= FW_UPTODATE;
606 }
607
608 if (t3_check_tpsram_version(sc) < 0) {
609 /*
610 * Warn user that a firmware update will be attempted in init.
611 */
612 device_printf(dev, "SRAM needs to be updated to version %c-%d.%d.%d\n",
613 t3rev2char(sc), TP_VERSION_MAJOR, TP_VERSION_MINOR, TP_VERSION_MICRO);
614 sc->flags &= ~TPS_UPTODATE;
615 } else {
616 sc->flags |= TPS_UPTODATE;
617 }
618
619 /*
620 * Create a child device for each MAC. The ethernet attachment
621 * will be done in these children.
622 */
623 for (i = 0; i < (sc)->params.nports; i++) {
624 struct port_info *pi;
625
626 if ((child = device_add_child(dev, "cxgb", -1)) == NULL) {
627 device_printf(dev, "failed to add child port\n");
628 error = EINVAL;
629 goto out;
630 }
631 pi = &sc->port[i];
632 pi->adapter = sc;
633 pi->nqsets = port_qsets;
634 pi->first_qset = i*port_qsets;
635 pi->port_id = i;
636 pi->tx_chan = i >= ai->nports0;
637 pi->txpkt_intf = pi->tx_chan ? 2 * (i - ai->nports0) + 1 : 2 * i;
638 sc->rxpkt_map[pi->txpkt_intf] = i;
639 sc->port[i].tx_chan = i >= ai->nports0;
640 sc->portdev[i] = child;
641 device_set_softc(child, pi);
642 }
643 if ((error = bus_generic_attach(dev)) != 0)
644 goto out;
645
646 /* initialize sge private state */
647 t3_sge_init_adapter(sc);
648
649 t3_led_ready(sc);
650
651 error = t3_get_fw_version(sc, &vers);
652 if (error)
653 goto out;
654
655 snprintf(&sc->fw_version[0], sizeof(sc->fw_version), "%d.%d.%d",
656 G_FW_VERSION_MAJOR(vers), G_FW_VERSION_MINOR(vers),
657 G_FW_VERSION_MICRO(vers));
658
659 snprintf(buf, sizeof(buf), "%s %sNIC\t E/C: %s S/N: %s",
660 ai->desc, is_offload(sc) ? "R" : "",
661 sc->params.vpd.ec, sc->params.vpd.sn);
662 device_set_desc_copy(dev, buf);
663
664 snprintf(&sc->port_types[0], sizeof(sc->port_types), "%x%x%x%x",
665 sc->params.vpd.port_type[0], sc->params.vpd.port_type[1],
666 sc->params.vpd.port_type[2], sc->params.vpd.port_type[3]);
667
668 device_printf(sc->dev, "Firmware Version %s\n", &sc->fw_version[0]);
669 callout_reset(&sc->cxgb_tick_ch, hz, cxgb_tick, sc);
670 t3_add_attach_sysctls(sc);
671
672 #ifdef TCP_OFFLOAD
673 for (i = 0; i < NUM_CPL_HANDLERS; i++)
674 sc->cpl_handler[i] = cpl_not_handled;
675 #endif
676
677 t3_intr_clear(sc);
678 error = cxgb_setup_interrupts(sc);
679 out:
680 if (error)
681 cxgb_free(sc);
682
683 return (error);
684 }
685
686 /*
687 * The cxgb_controller_detach routine is called with the device is
688 * unloaded from the system.
689 */
690
691 static int
cxgb_controller_detach(device_t dev)692 cxgb_controller_detach(device_t dev)
693 {
694 struct adapter *sc;
695
696 sc = device_get_softc(dev);
697
698 cxgb_free(sc);
699
700 return (0);
701 }
702
703 /*
704 * The cxgb_free() is called by the cxgb_controller_detach() routine
705 * to tear down the structures that were built up in
706 * cxgb_controller_attach(), and should be the final piece of work
707 * done when fully unloading the driver.
708 *
709 *
710 * 1. Shutting down the threads started by the cxgb_controller_attach()
711 * routine.
712 * 2. Stopping the lower level device and all callouts (cxgb_down_locked()).
713 * 3. Detaching all of the port devices created during the
714 * cxgb_controller_attach() routine.
715 * 4. Removing the device children created via cxgb_controller_attach().
716 * 5. Releasing PCI resources associated with the device.
717 * 6. Turning off the offload support, iff it was turned on.
718 * 7. Destroying the mutexes created in cxgb_controller_attach().
719 *
720 */
721 static void
cxgb_free(struct adapter * sc)722 cxgb_free(struct adapter *sc)
723 {
724 int i, nqsets = 0;
725
726 ADAPTER_LOCK(sc);
727 sc->flags |= CXGB_SHUTDOWN;
728 ADAPTER_UNLOCK(sc);
729
730 /*
731 * Make sure all child devices are gone.
732 */
733 bus_generic_detach(sc->dev);
734 for (i = 0; i < (sc)->params.nports; i++) {
735 if (sc->portdev[i] &&
736 device_delete_child(sc->dev, sc->portdev[i]) != 0)
737 device_printf(sc->dev, "failed to delete child port\n");
738 nqsets += sc->port[i].nqsets;
739 }
740
741 /*
742 * At this point, it is as if cxgb_port_detach has run on all ports, and
743 * cxgb_down has run on the adapter. All interrupts have been silenced,
744 * all open devices have been closed.
745 */
746 KASSERT(sc->open_device_map == 0, ("%s: device(s) still open (%x)",
747 __func__, sc->open_device_map));
748 for (i = 0; i < sc->params.nports; i++) {
749 KASSERT(sc->port[i].ifp == NULL, ("%s: port %i undead!",
750 __func__, i));
751 }
752
753 /*
754 * Finish off the adapter's callouts.
755 */
756 callout_drain(&sc->cxgb_tick_ch);
757 callout_drain(&sc->sge_timer_ch);
758
759 /*
760 * Release resources grabbed under FULL_INIT_DONE by cxgb_up. The
761 * sysctls are cleaned up by the kernel linker.
762 */
763 if (sc->flags & FULL_INIT_DONE) {
764 t3_free_sge_resources(sc, nqsets);
765 sc->flags &= ~FULL_INIT_DONE;
766 }
767
768 /*
769 * Release all interrupt resources.
770 */
771 cxgb_teardown_interrupts(sc);
772 if (sc->flags & (USING_MSI | USING_MSIX)) {
773 device_printf(sc->dev, "releasing msi message(s)\n");
774 pci_release_msi(sc->dev);
775 } else {
776 device_printf(sc->dev, "no msi message to release\n");
777 }
778
779 if (sc->msix_regs_res != NULL) {
780 bus_release_resource(sc->dev, SYS_RES_MEMORY, sc->msix_regs_rid,
781 sc->msix_regs_res);
782 }
783
784 /*
785 * Free the adapter's taskqueue.
786 */
787 if (sc->tq != NULL) {
788 taskqueue_free(sc->tq);
789 sc->tq = NULL;
790 }
791
792 free(sc->filters, M_DEVBUF);
793 t3_sge_free(sc);
794
795 if (sc->udbs_res != NULL)
796 bus_release_resource(sc->dev, SYS_RES_MEMORY, sc->udbs_rid,
797 sc->udbs_res);
798
799 if (sc->regs_res != NULL)
800 bus_release_resource(sc->dev, SYS_RES_MEMORY, sc->regs_rid,
801 sc->regs_res);
802
803 MTX_DESTROY(&sc->mdio_lock);
804 MTX_DESTROY(&sc->sge.reg_lock);
805 MTX_DESTROY(&sc->elmer_lock);
806 mtx_lock(&t3_list_lock);
807 SLIST_REMOVE(&t3_list, sc, adapter, link);
808 mtx_unlock(&t3_list_lock);
809 ADAPTER_LOCK_DEINIT(sc);
810 }
811
812 /**
813 * setup_sge_qsets - configure SGE Tx/Rx/response queues
814 * @sc: the controller softc
815 *
816 * Determines how many sets of SGE queues to use and initializes them.
817 * We support multiple queue sets per port if we have MSI-X, otherwise
818 * just one queue set per port.
819 */
820 static int
setup_sge_qsets(adapter_t * sc)821 setup_sge_qsets(adapter_t *sc)
822 {
823 int i, j, err, irq_idx = 0, qset_idx = 0;
824 u_int ntxq = SGE_TXQ_PER_SET;
825
826 if ((err = t3_sge_alloc(sc)) != 0) {
827 device_printf(sc->dev, "t3_sge_alloc returned %d\n", err);
828 return (err);
829 }
830
831 if (sc->params.rev > 0 && !(sc->flags & USING_MSI))
832 irq_idx = -1;
833
834 for (i = 0; i < (sc)->params.nports; i++) {
835 struct port_info *pi = &sc->port[i];
836
837 for (j = 0; j < pi->nqsets; j++, qset_idx++) {
838 err = t3_sge_alloc_qset(sc, qset_idx, (sc)->params.nports,
839 (sc->flags & USING_MSIX) ? qset_idx + 1 : irq_idx,
840 &sc->params.sge.qset[qset_idx], ntxq, pi);
841 if (err) {
842 t3_free_sge_resources(sc, qset_idx);
843 device_printf(sc->dev,
844 "t3_sge_alloc_qset failed with %d\n", err);
845 return (err);
846 }
847 }
848 }
849
850 return (0);
851 }
852
853 static void
cxgb_teardown_interrupts(adapter_t * sc)854 cxgb_teardown_interrupts(adapter_t *sc)
855 {
856 int i;
857
858 for (i = 0; i < SGE_QSETS; i++) {
859 if (sc->msix_intr_tag[i] == NULL) {
860
861 /* Should have been setup fully or not at all */
862 KASSERT(sc->msix_irq_res[i] == NULL &&
863 sc->msix_irq_rid[i] == 0,
864 ("%s: half-done interrupt (%d).", __func__, i));
865
866 continue;
867 }
868
869 bus_teardown_intr(sc->dev, sc->msix_irq_res[i],
870 sc->msix_intr_tag[i]);
871 bus_release_resource(sc->dev, SYS_RES_IRQ, sc->msix_irq_rid[i],
872 sc->msix_irq_res[i]);
873
874 sc->msix_irq_res[i] = sc->msix_intr_tag[i] = NULL;
875 sc->msix_irq_rid[i] = 0;
876 }
877
878 if (sc->intr_tag) {
879 KASSERT(sc->irq_res != NULL,
880 ("%s: half-done interrupt.", __func__));
881
882 bus_teardown_intr(sc->dev, sc->irq_res, sc->intr_tag);
883 bus_release_resource(sc->dev, SYS_RES_IRQ, sc->irq_rid,
884 sc->irq_res);
885
886 sc->irq_res = sc->intr_tag = NULL;
887 sc->irq_rid = 0;
888 }
889 }
890
891 static int
cxgb_setup_interrupts(adapter_t * sc)892 cxgb_setup_interrupts(adapter_t *sc)
893 {
894 struct resource *res;
895 void *tag;
896 int i, rid, err, intr_flag = sc->flags & (USING_MSI | USING_MSIX);
897
898 sc->irq_rid = intr_flag ? 1 : 0;
899 sc->irq_res = bus_alloc_resource_any(sc->dev, SYS_RES_IRQ, &sc->irq_rid,
900 RF_SHAREABLE | RF_ACTIVE);
901 if (sc->irq_res == NULL) {
902 device_printf(sc->dev, "Cannot allocate interrupt (%x, %u)\n",
903 intr_flag, sc->irq_rid);
904 err = EINVAL;
905 sc->irq_rid = 0;
906 } else {
907 err = bus_setup_intr(sc->dev, sc->irq_res,
908 INTR_MPSAFE | INTR_TYPE_NET, NULL,
909 sc->cxgb_intr, sc, &sc->intr_tag);
910
911 if (err) {
912 device_printf(sc->dev,
913 "Cannot set up interrupt (%x, %u, %d)\n",
914 intr_flag, sc->irq_rid, err);
915 bus_release_resource(sc->dev, SYS_RES_IRQ, sc->irq_rid,
916 sc->irq_res);
917 sc->irq_res = sc->intr_tag = NULL;
918 sc->irq_rid = 0;
919 }
920 }
921
922 /* That's all for INTx or MSI */
923 if (!(intr_flag & USING_MSIX) || err)
924 return (err);
925
926 bus_describe_intr(sc->dev, sc->irq_res, sc->intr_tag, "err");
927 for (i = 0; i < sc->msi_count - 1; i++) {
928 rid = i + 2;
929 res = bus_alloc_resource_any(sc->dev, SYS_RES_IRQ, &rid,
930 RF_SHAREABLE | RF_ACTIVE);
931 if (res == NULL) {
932 device_printf(sc->dev, "Cannot allocate interrupt "
933 "for message %d\n", rid);
934 err = EINVAL;
935 break;
936 }
937
938 err = bus_setup_intr(sc->dev, res, INTR_MPSAFE | INTR_TYPE_NET,
939 NULL, t3_intr_msix, &sc->sge.qs[i], &tag);
940 if (err) {
941 device_printf(sc->dev, "Cannot set up interrupt "
942 "for message %d (%d)\n", rid, err);
943 bus_release_resource(sc->dev, SYS_RES_IRQ, rid, res);
944 break;
945 }
946
947 sc->msix_irq_rid[i] = rid;
948 sc->msix_irq_res[i] = res;
949 sc->msix_intr_tag[i] = tag;
950 bus_describe_intr(sc->dev, res, tag, "qs%d", i);
951 }
952
953 if (err)
954 cxgb_teardown_interrupts(sc);
955
956 return (err);
957 }
958
959
960 static int
cxgb_port_probe(device_t dev)961 cxgb_port_probe(device_t dev)
962 {
963 struct port_info *p;
964 char buf[80];
965 const char *desc;
966
967 p = device_get_softc(dev);
968 desc = p->phy.desc;
969 snprintf(buf, sizeof(buf), "Port %d %s", p->port_id, desc);
970 device_set_desc_copy(dev, buf);
971 return (0);
972 }
973
974
975 static int
cxgb_makedev(struct port_info * pi)976 cxgb_makedev(struct port_info *pi)
977 {
978
979 pi->port_cdev = make_dev(&cxgb_cdevsw, pi->ifp->if_dunit,
980 UID_ROOT, GID_WHEEL, 0600, "%s", if_name(pi->ifp));
981
982 if (pi->port_cdev == NULL)
983 return (ENOMEM);
984
985 pi->port_cdev->si_drv1 = (void *)pi;
986
987 return (0);
988 }
989
990 #define CXGB_CAP (IFCAP_VLAN_HWTAGGING | IFCAP_VLAN_MTU | IFCAP_HWCSUM | \
991 IFCAP_VLAN_HWCSUM | IFCAP_TSO | IFCAP_JUMBO_MTU | IFCAP_LRO | \
992 IFCAP_VLAN_HWTSO | IFCAP_LINKSTATE | IFCAP_HWCSUM_IPV6)
993 #define CXGB_CAP_ENABLE CXGB_CAP
994
995 static int
cxgb_port_attach(device_t dev)996 cxgb_port_attach(device_t dev)
997 {
998 struct port_info *p;
999 struct ifnet *ifp;
1000 int err;
1001 struct adapter *sc;
1002
1003 p = device_get_softc(dev);
1004 sc = p->adapter;
1005 snprintf(p->lockbuf, PORT_NAME_LEN, "cxgb port lock %d:%d",
1006 device_get_unit(device_get_parent(dev)), p->port_id);
1007 PORT_LOCK_INIT(p, p->lockbuf);
1008
1009 callout_init(&p->link_check_ch, 1);
1010 TASK_INIT(&p->link_check_task, 0, check_link_status, p);
1011
1012 /* Allocate an ifnet object and set it up */
1013 ifp = p->ifp = if_alloc(IFT_ETHER);
1014 if (ifp == NULL) {
1015 device_printf(dev, "Cannot allocate ifnet\n");
1016 return (ENOMEM);
1017 }
1018
1019 if_initname(ifp, device_get_name(dev), device_get_unit(dev));
1020 ifp->if_init = cxgb_init;
1021 ifp->if_softc = p;
1022 ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST;
1023 ifp->if_ioctl = cxgb_ioctl;
1024 ifp->if_transmit = cxgb_transmit;
1025 ifp->if_qflush = cxgb_qflush;
1026 ifp->if_get_counter = cxgb_get_counter;
1027
1028 ifp->if_capabilities = CXGB_CAP;
1029 #ifdef TCP_OFFLOAD
1030 if (is_offload(sc))
1031 ifp->if_capabilities |= IFCAP_TOE4;
1032 #endif
1033 ifp->if_capenable = CXGB_CAP_ENABLE;
1034 ifp->if_hwassist = CSUM_TCP | CSUM_UDP | CSUM_IP | CSUM_TSO |
1035 CSUM_UDP_IPV6 | CSUM_TCP_IPV6;
1036
1037 /*
1038 * Disable TSO on 4-port - it isn't supported by the firmware.
1039 */
1040 if (sc->params.nports > 2) {
1041 ifp->if_capabilities &= ~(IFCAP_TSO | IFCAP_VLAN_HWTSO);
1042 ifp->if_capenable &= ~(IFCAP_TSO | IFCAP_VLAN_HWTSO);
1043 ifp->if_hwassist &= ~CSUM_TSO;
1044 }
1045
1046 ether_ifattach(ifp, p->hw_addr);
1047
1048 #ifdef DEFAULT_JUMBO
1049 if (sc->params.nports <= 2)
1050 ifp->if_mtu = ETHERMTU_JUMBO;
1051 #endif
1052 if ((err = cxgb_makedev(p)) != 0) {
1053 printf("makedev failed %d\n", err);
1054 return (err);
1055 }
1056
1057 /* Create a list of media supported by this port */
1058 ifmedia_init(&p->media, IFM_IMASK, cxgb_media_change,
1059 cxgb_media_status);
1060 cxgb_build_medialist(p);
1061
1062 t3_sge_init_port(p);
1063
1064 return (err);
1065 }
1066
1067 /*
1068 * cxgb_port_detach() is called via the device_detach methods when
1069 * cxgb_free() calls the bus_generic_detach. It is responsible for
1070 * removing the device from the view of the kernel, i.e. from all
1071 * interfaces lists etc. This routine is only called when the driver is
1072 * being unloaded, not when the link goes down.
1073 */
1074 static int
cxgb_port_detach(device_t dev)1075 cxgb_port_detach(device_t dev)
1076 {
1077 struct port_info *p;
1078 struct adapter *sc;
1079 int i;
1080
1081 p = device_get_softc(dev);
1082 sc = p->adapter;
1083
1084 /* Tell cxgb_ioctl and if_init that the port is going away */
1085 ADAPTER_LOCK(sc);
1086 SET_DOOMED(p);
1087 wakeup(&sc->flags);
1088 while (IS_BUSY(sc))
1089 mtx_sleep(&sc->flags, &sc->lock, 0, "cxgbdtch", 0);
1090 SET_BUSY(sc);
1091 ADAPTER_UNLOCK(sc);
1092
1093 if (p->port_cdev != NULL)
1094 destroy_dev(p->port_cdev);
1095
1096 cxgb_uninit_synchronized(p);
1097 ether_ifdetach(p->ifp);
1098
1099 for (i = p->first_qset; i < p->first_qset + p->nqsets; i++) {
1100 struct sge_qset *qs = &sc->sge.qs[i];
1101 struct sge_txq *txq = &qs->txq[TXQ_ETH];
1102
1103 callout_drain(&txq->txq_watchdog);
1104 callout_drain(&txq->txq_timer);
1105 }
1106
1107 PORT_LOCK_DEINIT(p);
1108 if_free(p->ifp);
1109 p->ifp = NULL;
1110
1111 ADAPTER_LOCK(sc);
1112 CLR_BUSY(sc);
1113 wakeup_one(&sc->flags);
1114 ADAPTER_UNLOCK(sc);
1115 return (0);
1116 }
1117
1118 void
t3_fatal_err(struct adapter * sc)1119 t3_fatal_err(struct adapter *sc)
1120 {
1121 u_int fw_status[4];
1122
1123 if (sc->flags & FULL_INIT_DONE) {
1124 t3_sge_stop(sc);
1125 t3_write_reg(sc, A_XGM_TX_CTRL, 0);
1126 t3_write_reg(sc, A_XGM_RX_CTRL, 0);
1127 t3_write_reg(sc, XGM_REG(A_XGM_TX_CTRL, 1), 0);
1128 t3_write_reg(sc, XGM_REG(A_XGM_RX_CTRL, 1), 0);
1129 t3_intr_disable(sc);
1130 }
1131 device_printf(sc->dev,"encountered fatal error, operation suspended\n");
1132 if (!t3_cim_ctl_blk_read(sc, 0xa0, 4, fw_status))
1133 device_printf(sc->dev, "FW_ status: 0x%x, 0x%x, 0x%x, 0x%x\n",
1134 fw_status[0], fw_status[1], fw_status[2], fw_status[3]);
1135 }
1136
1137 int
t3_os_find_pci_capability(adapter_t * sc,int cap)1138 t3_os_find_pci_capability(adapter_t *sc, int cap)
1139 {
1140 device_t dev;
1141 struct pci_devinfo *dinfo;
1142 pcicfgregs *cfg;
1143 uint32_t status;
1144 uint8_t ptr;
1145
1146 dev = sc->dev;
1147 dinfo = device_get_ivars(dev);
1148 cfg = &dinfo->cfg;
1149
1150 status = pci_read_config(dev, PCIR_STATUS, 2);
1151 if (!(status & PCIM_STATUS_CAPPRESENT))
1152 return (0);
1153
1154 switch (cfg->hdrtype & PCIM_HDRTYPE) {
1155 case 0:
1156 case 1:
1157 ptr = PCIR_CAP_PTR;
1158 break;
1159 case 2:
1160 ptr = PCIR_CAP_PTR_2;
1161 break;
1162 default:
1163 return (0);
1164 break;
1165 }
1166 ptr = pci_read_config(dev, ptr, 1);
1167
1168 while (ptr != 0) {
1169 if (pci_read_config(dev, ptr + PCICAP_ID, 1) == cap)
1170 return (ptr);
1171 ptr = pci_read_config(dev, ptr + PCICAP_NEXTPTR, 1);
1172 }
1173
1174 return (0);
1175 }
1176
1177 int
t3_os_pci_save_state(struct adapter * sc)1178 t3_os_pci_save_state(struct adapter *sc)
1179 {
1180 device_t dev;
1181 struct pci_devinfo *dinfo;
1182
1183 dev = sc->dev;
1184 dinfo = device_get_ivars(dev);
1185
1186 pci_cfg_save(dev, dinfo, 0);
1187 return (0);
1188 }
1189
1190 int
t3_os_pci_restore_state(struct adapter * sc)1191 t3_os_pci_restore_state(struct adapter *sc)
1192 {
1193 device_t dev;
1194 struct pci_devinfo *dinfo;
1195
1196 dev = sc->dev;
1197 dinfo = device_get_ivars(dev);
1198
1199 pci_cfg_restore(dev, dinfo);
1200 return (0);
1201 }
1202
1203 /**
1204 * t3_os_link_changed - handle link status changes
1205 * @sc: the adapter associated with the link change
1206 * @port_id: the port index whose link status has changed
1207 * @link_status: the new status of the link
1208 * @speed: the new speed setting
1209 * @duplex: the new duplex setting
1210 * @fc: the new flow-control setting
1211 *
1212 * This is the OS-dependent handler for link status changes. The OS
1213 * neutral handler takes care of most of the processing for these events,
1214 * then calls this handler for any OS-specific processing.
1215 */
1216 void
t3_os_link_changed(adapter_t * adapter,int port_id,int link_status,int speed,int duplex,int fc,int mac_was_reset)1217 t3_os_link_changed(adapter_t *adapter, int port_id, int link_status, int speed,
1218 int duplex, int fc, int mac_was_reset)
1219 {
1220 struct port_info *pi = &adapter->port[port_id];
1221 struct ifnet *ifp = pi->ifp;
1222
1223 /* no race with detach, so ifp should always be good */
1224 KASSERT(ifp, ("%s: if detached.", __func__));
1225
1226 /* Reapply mac settings if they were lost due to a reset */
1227 if (mac_was_reset) {
1228 PORT_LOCK(pi);
1229 cxgb_update_mac_settings(pi);
1230 PORT_UNLOCK(pi);
1231 }
1232
1233 if (link_status) {
1234 ifp->if_baudrate = IF_Mbps(speed);
1235 if_link_state_change(ifp, LINK_STATE_UP);
1236 } else
1237 if_link_state_change(ifp, LINK_STATE_DOWN);
1238 }
1239
1240 /**
1241 * t3_os_phymod_changed - handle PHY module changes
1242 * @phy: the PHY reporting the module change
1243 * @mod_type: new module type
1244 *
1245 * This is the OS-dependent handler for PHY module changes. It is
1246 * invoked when a PHY module is removed or inserted for any OS-specific
1247 * processing.
1248 */
t3_os_phymod_changed(struct adapter * adap,int port_id)1249 void t3_os_phymod_changed(struct adapter *adap, int port_id)
1250 {
1251 static const char *mod_str[] = {
1252 NULL, "SR", "LR", "LRM", "TWINAX", "TWINAX-L", "unknown"
1253 };
1254 struct port_info *pi = &adap->port[port_id];
1255 int mod = pi->phy.modtype;
1256
1257 if (mod != pi->media.ifm_cur->ifm_data)
1258 cxgb_build_medialist(pi);
1259
1260 if (mod == phy_modtype_none)
1261 if_printf(pi->ifp, "PHY module unplugged\n");
1262 else {
1263 KASSERT(mod < ARRAY_SIZE(mod_str),
1264 ("invalid PHY module type %d", mod));
1265 if_printf(pi->ifp, "%s PHY module inserted\n", mod_str[mod]);
1266 }
1267 }
1268
1269 void
t3_os_set_hw_addr(adapter_t * adapter,int port_idx,u8 hw_addr[])1270 t3_os_set_hw_addr(adapter_t *adapter, int port_idx, u8 hw_addr[])
1271 {
1272
1273 /*
1274 * The ifnet might not be allocated before this gets called,
1275 * as this is called early on in attach by t3_prep_adapter
1276 * save the address off in the port structure
1277 */
1278 if (cxgb_debug)
1279 printf("set_hw_addr on idx %d addr %6D\n", port_idx, hw_addr, ":");
1280 bcopy(hw_addr, adapter->port[port_idx].hw_addr, ETHER_ADDR_LEN);
1281 }
1282
1283 /*
1284 * Programs the XGMAC based on the settings in the ifnet. These settings
1285 * include MTU, MAC address, mcast addresses, etc.
1286 */
1287 static void
cxgb_update_mac_settings(struct port_info * p)1288 cxgb_update_mac_settings(struct port_info *p)
1289 {
1290 struct ifnet *ifp = p->ifp;
1291 struct t3_rx_mode rm;
1292 struct cmac *mac = &p->mac;
1293 int mtu, hwtagging;
1294
1295 PORT_LOCK_ASSERT_OWNED(p);
1296
1297 bcopy(IF_LLADDR(ifp), p->hw_addr, ETHER_ADDR_LEN);
1298
1299 mtu = ifp->if_mtu;
1300 if (ifp->if_capenable & IFCAP_VLAN_MTU)
1301 mtu += ETHER_VLAN_ENCAP_LEN;
1302
1303 hwtagging = (ifp->if_capenable & IFCAP_VLAN_HWTAGGING) != 0;
1304
1305 t3_mac_set_mtu(mac, mtu);
1306 t3_set_vlan_accel(p->adapter, 1 << p->tx_chan, hwtagging);
1307 t3_mac_set_address(mac, 0, p->hw_addr);
1308 t3_init_rx_mode(&rm, p);
1309 t3_mac_set_rx_mode(mac, &rm);
1310 }
1311
1312
1313 static int
await_mgmt_replies(struct adapter * adap,unsigned long init_cnt,unsigned long n)1314 await_mgmt_replies(struct adapter *adap, unsigned long init_cnt,
1315 unsigned long n)
1316 {
1317 int attempts = 5;
1318
1319 while (adap->sge.qs[0].rspq.offload_pkts < init_cnt + n) {
1320 if (!--attempts)
1321 return (ETIMEDOUT);
1322 t3_os_sleep(10);
1323 }
1324 return 0;
1325 }
1326
1327 static int
init_tp_parity(struct adapter * adap)1328 init_tp_parity(struct adapter *adap)
1329 {
1330 int i;
1331 struct mbuf *m;
1332 struct cpl_set_tcb_field *greq;
1333 unsigned long cnt = adap->sge.qs[0].rspq.offload_pkts;
1334
1335 t3_tp_set_offload_mode(adap, 1);
1336
1337 for (i = 0; i < 16; i++) {
1338 struct cpl_smt_write_req *req;
1339
1340 m = m_gethdr(M_WAITOK, MT_DATA);
1341 req = mtod(m, struct cpl_smt_write_req *);
1342 m->m_len = m->m_pkthdr.len = sizeof(*req);
1343 memset(req, 0, sizeof(*req));
1344 req->wr.wrh_hi = htonl(V_WR_OP(FW_WROPCODE_FORWARD));
1345 OPCODE_TID(req) = htonl(MK_OPCODE_TID(CPL_SMT_WRITE_REQ, i));
1346 req->iff = i;
1347 t3_mgmt_tx(adap, m);
1348 }
1349
1350 for (i = 0; i < 2048; i++) {
1351 struct cpl_l2t_write_req *req;
1352
1353 m = m_gethdr(M_WAITOK, MT_DATA);
1354 req = mtod(m, struct cpl_l2t_write_req *);
1355 m->m_len = m->m_pkthdr.len = sizeof(*req);
1356 memset(req, 0, sizeof(*req));
1357 req->wr.wrh_hi = htonl(V_WR_OP(FW_WROPCODE_FORWARD));
1358 OPCODE_TID(req) = htonl(MK_OPCODE_TID(CPL_L2T_WRITE_REQ, i));
1359 req->params = htonl(V_L2T_W_IDX(i));
1360 t3_mgmt_tx(adap, m);
1361 }
1362
1363 for (i = 0; i < 2048; i++) {
1364 struct cpl_rte_write_req *req;
1365
1366 m = m_gethdr(M_WAITOK, MT_DATA);
1367 req = mtod(m, struct cpl_rte_write_req *);
1368 m->m_len = m->m_pkthdr.len = sizeof(*req);
1369 memset(req, 0, sizeof(*req));
1370 req->wr.wrh_hi = htonl(V_WR_OP(FW_WROPCODE_FORWARD));
1371 OPCODE_TID(req) = htonl(MK_OPCODE_TID(CPL_RTE_WRITE_REQ, i));
1372 req->l2t_idx = htonl(V_L2T_W_IDX(i));
1373 t3_mgmt_tx(adap, m);
1374 }
1375
1376 m = m_gethdr(M_WAITOK, MT_DATA);
1377 greq = mtod(m, struct cpl_set_tcb_field *);
1378 m->m_len = m->m_pkthdr.len = sizeof(*greq);
1379 memset(greq, 0, sizeof(*greq));
1380 greq->wr.wrh_hi = htonl(V_WR_OP(FW_WROPCODE_FORWARD));
1381 OPCODE_TID(greq) = htonl(MK_OPCODE_TID(CPL_SET_TCB_FIELD, 0));
1382 greq->mask = htobe64(1);
1383 t3_mgmt_tx(adap, m);
1384
1385 i = await_mgmt_replies(adap, cnt, 16 + 2048 + 2048 + 1);
1386 t3_tp_set_offload_mode(adap, 0);
1387 return (i);
1388 }
1389
1390 /**
1391 * setup_rss - configure Receive Side Steering (per-queue connection demux)
1392 * @adap: the adapter
1393 *
1394 * Sets up RSS to distribute packets to multiple receive queues. We
1395 * configure the RSS CPU lookup table to distribute to the number of HW
1396 * receive queues, and the response queue lookup table to narrow that
1397 * down to the response queues actually configured for each port.
1398 * We always configure the RSS mapping for two ports since the mapping
1399 * table has plenty of entries.
1400 */
1401 static void
setup_rss(adapter_t * adap)1402 setup_rss(adapter_t *adap)
1403 {
1404 int i;
1405 u_int nq[2];
1406 uint8_t cpus[SGE_QSETS + 1];
1407 uint16_t rspq_map[RSS_TABLE_SIZE];
1408
1409 for (i = 0; i < SGE_QSETS; ++i)
1410 cpus[i] = i;
1411 cpus[SGE_QSETS] = 0xff;
1412
1413 nq[0] = nq[1] = 0;
1414 for_each_port(adap, i) {
1415 const struct port_info *pi = adap2pinfo(adap, i);
1416
1417 nq[pi->tx_chan] += pi->nqsets;
1418 }
1419 for (i = 0; i < RSS_TABLE_SIZE / 2; ++i) {
1420 rspq_map[i] = nq[0] ? i % nq[0] : 0;
1421 rspq_map[i + RSS_TABLE_SIZE / 2] = nq[1] ? i % nq[1] + nq[0] : 0;
1422 }
1423
1424 /* Calculate the reverse RSS map table */
1425 for (i = 0; i < SGE_QSETS; ++i)
1426 adap->rrss_map[i] = 0xff;
1427 for (i = 0; i < RSS_TABLE_SIZE; ++i)
1428 if (adap->rrss_map[rspq_map[i]] == 0xff)
1429 adap->rrss_map[rspq_map[i]] = i;
1430
1431 t3_config_rss(adap, F_RQFEEDBACKENABLE | F_TNLLKPEN | F_TNLMAPEN |
1432 F_TNLPRTEN | F_TNL2TUPEN | F_TNL4TUPEN | F_OFDMAPEN |
1433 F_RRCPLMAPEN | V_RRCPLCPUSIZE(6) | F_HASHTOEPLITZ,
1434 cpus, rspq_map);
1435
1436 }
1437 static void
send_pktsched_cmd(struct adapter * adap,int sched,int qidx,int lo,int hi,int port)1438 send_pktsched_cmd(struct adapter *adap, int sched, int qidx, int lo,
1439 int hi, int port)
1440 {
1441 struct mbuf *m;
1442 struct mngt_pktsched_wr *req;
1443
1444 m = m_gethdr(M_NOWAIT, MT_DATA);
1445 if (m) {
1446 req = mtod(m, struct mngt_pktsched_wr *);
1447 req->wr.wrh_hi = htonl(V_WR_OP(FW_WROPCODE_MNGT));
1448 req->mngt_opcode = FW_MNGTOPCODE_PKTSCHED_SET;
1449 req->sched = sched;
1450 req->idx = qidx;
1451 req->min = lo;
1452 req->max = hi;
1453 req->binding = port;
1454 m->m_len = m->m_pkthdr.len = sizeof(*req);
1455 t3_mgmt_tx(adap, m);
1456 }
1457 }
1458
1459 static void
bind_qsets(adapter_t * sc)1460 bind_qsets(adapter_t *sc)
1461 {
1462 int i, j;
1463
1464 for (i = 0; i < (sc)->params.nports; ++i) {
1465 const struct port_info *pi = adap2pinfo(sc, i);
1466
1467 for (j = 0; j < pi->nqsets; ++j) {
1468 send_pktsched_cmd(sc, 1, pi->first_qset + j, -1,
1469 -1, pi->tx_chan);
1470
1471 }
1472 }
1473 }
1474
1475 static void
update_tpeeprom(struct adapter * adap)1476 update_tpeeprom(struct adapter *adap)
1477 {
1478 const struct firmware *tpeeprom;
1479
1480 uint32_t version;
1481 unsigned int major, minor;
1482 int ret, len;
1483 char rev, name[32];
1484
1485 t3_seeprom_read(adap, TP_SRAM_OFFSET, &version);
1486
1487 major = G_TP_VERSION_MAJOR(version);
1488 minor = G_TP_VERSION_MINOR(version);
1489 if (major == TP_VERSION_MAJOR && minor == TP_VERSION_MINOR)
1490 return;
1491
1492 rev = t3rev2char(adap);
1493 snprintf(name, sizeof(name), TPEEPROM_NAME, rev);
1494
1495 tpeeprom = firmware_get(name);
1496 if (tpeeprom == NULL) {
1497 device_printf(adap->dev,
1498 "could not load TP EEPROM: unable to load %s\n",
1499 name);
1500 return;
1501 }
1502
1503 len = tpeeprom->datasize - 4;
1504
1505 ret = t3_check_tpsram(adap, tpeeprom->data, tpeeprom->datasize);
1506 if (ret)
1507 goto release_tpeeprom;
1508
1509 if (len != TP_SRAM_LEN) {
1510 device_printf(adap->dev,
1511 "%s length is wrong len=%d expected=%d\n", name,
1512 len, TP_SRAM_LEN);
1513 return;
1514 }
1515
1516 ret = set_eeprom(&adap->port[0], tpeeprom->data, tpeeprom->datasize,
1517 TP_SRAM_OFFSET);
1518
1519 if (!ret) {
1520 device_printf(adap->dev,
1521 "Protocol SRAM image updated in EEPROM to %d.%d.%d\n",
1522 TP_VERSION_MAJOR, TP_VERSION_MINOR, TP_VERSION_MICRO);
1523 } else
1524 device_printf(adap->dev,
1525 "Protocol SRAM image update in EEPROM failed\n");
1526
1527 release_tpeeprom:
1528 firmware_put(tpeeprom, FIRMWARE_UNLOAD);
1529
1530 return;
1531 }
1532
1533 static int
update_tpsram(struct adapter * adap)1534 update_tpsram(struct adapter *adap)
1535 {
1536 const struct firmware *tpsram;
1537 int ret;
1538 char rev, name[32];
1539
1540 rev = t3rev2char(adap);
1541 snprintf(name, sizeof(name), TPSRAM_NAME, rev);
1542
1543 update_tpeeprom(adap);
1544
1545 tpsram = firmware_get(name);
1546 if (tpsram == NULL){
1547 device_printf(adap->dev, "could not load TP SRAM\n");
1548 return (EINVAL);
1549 } else
1550 device_printf(adap->dev, "updating TP SRAM\n");
1551
1552 ret = t3_check_tpsram(adap, tpsram->data, tpsram->datasize);
1553 if (ret)
1554 goto release_tpsram;
1555
1556 ret = t3_set_proto_sram(adap, tpsram->data);
1557 if (ret)
1558 device_printf(adap->dev, "loading protocol SRAM failed\n");
1559
1560 release_tpsram:
1561 firmware_put(tpsram, FIRMWARE_UNLOAD);
1562
1563 return ret;
1564 }
1565
1566 /**
1567 * cxgb_up - enable the adapter
1568 * @adap: adapter being enabled
1569 *
1570 * Called when the first port is enabled, this function performs the
1571 * actions necessary to make an adapter operational, such as completing
1572 * the initialization of HW modules, and enabling interrupts.
1573 */
1574 static int
cxgb_up(struct adapter * sc)1575 cxgb_up(struct adapter *sc)
1576 {
1577 int err = 0;
1578 unsigned int mxf = t3_mc5_size(&sc->mc5) - MC5_MIN_TIDS;
1579
1580 KASSERT(sc->open_device_map == 0, ("%s: device(s) already open (%x)",
1581 __func__, sc->open_device_map));
1582
1583 if ((sc->flags & FULL_INIT_DONE) == 0) {
1584
1585 ADAPTER_LOCK_ASSERT_NOTOWNED(sc);
1586
1587 if ((sc->flags & FW_UPTODATE) == 0)
1588 if ((err = upgrade_fw(sc)))
1589 goto out;
1590
1591 if ((sc->flags & TPS_UPTODATE) == 0)
1592 if ((err = update_tpsram(sc)))
1593 goto out;
1594
1595 if (is_offload(sc) && nfilters != 0) {
1596 sc->params.mc5.nservers = 0;
1597
1598 if (nfilters < 0)
1599 sc->params.mc5.nfilters = mxf;
1600 else
1601 sc->params.mc5.nfilters = min(nfilters, mxf);
1602 }
1603
1604 err = t3_init_hw(sc, 0);
1605 if (err)
1606 goto out;
1607
1608 t3_set_reg_field(sc, A_TP_PARA_REG5, 0, F_RXDDPOFFINIT);
1609 t3_write_reg(sc, A_ULPRX_TDDP_PSZ, V_HPZ0(PAGE_SHIFT - 12));
1610
1611 err = setup_sge_qsets(sc);
1612 if (err)
1613 goto out;
1614
1615 alloc_filters(sc);
1616 setup_rss(sc);
1617
1618 t3_add_configured_sysctls(sc);
1619 sc->flags |= FULL_INIT_DONE;
1620 }
1621
1622 t3_intr_clear(sc);
1623 t3_sge_start(sc);
1624 t3_intr_enable(sc);
1625
1626 if (sc->params.rev >= T3_REV_C && !(sc->flags & TP_PARITY_INIT) &&
1627 is_offload(sc) && init_tp_parity(sc) == 0)
1628 sc->flags |= TP_PARITY_INIT;
1629
1630 if (sc->flags & TP_PARITY_INIT) {
1631 t3_write_reg(sc, A_TP_INT_CAUSE, F_CMCACHEPERR | F_ARPLUTPERR);
1632 t3_write_reg(sc, A_TP_INT_ENABLE, 0x7fbfffff);
1633 }
1634
1635 if (!(sc->flags & QUEUES_BOUND)) {
1636 bind_qsets(sc);
1637 setup_hw_filters(sc);
1638 sc->flags |= QUEUES_BOUND;
1639 }
1640
1641 t3_sge_reset_adapter(sc);
1642 out:
1643 return (err);
1644 }
1645
1646 /*
1647 * Called when the last open device is closed. Does NOT undo all of cxgb_up's
1648 * work. Specifically, the resources grabbed under FULL_INIT_DONE are released
1649 * during controller_detach, not here.
1650 */
1651 static void
cxgb_down(struct adapter * sc)1652 cxgb_down(struct adapter *sc)
1653 {
1654 t3_sge_stop(sc);
1655 t3_intr_disable(sc);
1656 }
1657
1658 /*
1659 * if_init for cxgb ports.
1660 */
1661 static void
cxgb_init(void * arg)1662 cxgb_init(void *arg)
1663 {
1664 struct port_info *p = arg;
1665 struct adapter *sc = p->adapter;
1666
1667 ADAPTER_LOCK(sc);
1668 cxgb_init_locked(p); /* releases adapter lock */
1669 ADAPTER_LOCK_ASSERT_NOTOWNED(sc);
1670 }
1671
1672 static int
cxgb_init_locked(struct port_info * p)1673 cxgb_init_locked(struct port_info *p)
1674 {
1675 struct adapter *sc = p->adapter;
1676 struct ifnet *ifp = p->ifp;
1677 struct cmac *mac = &p->mac;
1678 int i, rc = 0, may_sleep = 0, gave_up_lock = 0;
1679
1680 ADAPTER_LOCK_ASSERT_OWNED(sc);
1681
1682 while (!IS_DOOMED(p) && IS_BUSY(sc)) {
1683 gave_up_lock = 1;
1684 if (mtx_sleep(&sc->flags, &sc->lock, PCATCH, "cxgbinit", 0)) {
1685 rc = EINTR;
1686 goto done;
1687 }
1688 }
1689 if (IS_DOOMED(p)) {
1690 rc = ENXIO;
1691 goto done;
1692 }
1693 KASSERT(!IS_BUSY(sc), ("%s: controller busy.", __func__));
1694
1695 /*
1696 * The code that runs during one-time adapter initialization can sleep
1697 * so it's important not to hold any locks across it.
1698 */
1699 may_sleep = sc->flags & FULL_INIT_DONE ? 0 : 1;
1700
1701 if (may_sleep) {
1702 SET_BUSY(sc);
1703 gave_up_lock = 1;
1704 ADAPTER_UNLOCK(sc);
1705 }
1706
1707 if (sc->open_device_map == 0 && ((rc = cxgb_up(sc)) != 0))
1708 goto done;
1709
1710 PORT_LOCK(p);
1711 if (isset(&sc->open_device_map, p->port_id) &&
1712 (ifp->if_drv_flags & IFF_DRV_RUNNING)) {
1713 PORT_UNLOCK(p);
1714 goto done;
1715 }
1716 t3_port_intr_enable(sc, p->port_id);
1717 if (!mac->multiport)
1718 t3_mac_init(mac);
1719 cxgb_update_mac_settings(p);
1720 t3_link_start(&p->phy, mac, &p->link_config);
1721 t3_mac_enable(mac, MAC_DIRECTION_RX | MAC_DIRECTION_TX);
1722 ifp->if_drv_flags |= IFF_DRV_RUNNING;
1723 ifp->if_drv_flags &= ~IFF_DRV_OACTIVE;
1724 PORT_UNLOCK(p);
1725
1726 for (i = p->first_qset; i < p->first_qset + p->nqsets; i++) {
1727 struct sge_qset *qs = &sc->sge.qs[i];
1728 struct sge_txq *txq = &qs->txq[TXQ_ETH];
1729
1730 callout_reset_on(&txq->txq_watchdog, hz, cxgb_tx_watchdog, qs,
1731 txq->txq_watchdog.c_cpu);
1732 }
1733
1734 /* all ok */
1735 setbit(&sc->open_device_map, p->port_id);
1736 callout_reset(&p->link_check_ch,
1737 p->phy.caps & SUPPORTED_LINK_IRQ ? hz * 3 : hz / 4,
1738 link_check_callout, p);
1739
1740 done:
1741 if (may_sleep) {
1742 ADAPTER_LOCK(sc);
1743 KASSERT(IS_BUSY(sc), ("%s: controller not busy.", __func__));
1744 CLR_BUSY(sc);
1745 }
1746 if (gave_up_lock)
1747 wakeup_one(&sc->flags);
1748 ADAPTER_UNLOCK(sc);
1749 return (rc);
1750 }
1751
1752 static int
cxgb_uninit_locked(struct port_info * p)1753 cxgb_uninit_locked(struct port_info *p)
1754 {
1755 struct adapter *sc = p->adapter;
1756 int rc;
1757
1758 ADAPTER_LOCK_ASSERT_OWNED(sc);
1759
1760 while (!IS_DOOMED(p) && IS_BUSY(sc)) {
1761 if (mtx_sleep(&sc->flags, &sc->lock, PCATCH, "cxgbunin", 0)) {
1762 rc = EINTR;
1763 goto done;
1764 }
1765 }
1766 if (IS_DOOMED(p)) {
1767 rc = ENXIO;
1768 goto done;
1769 }
1770 KASSERT(!IS_BUSY(sc), ("%s: controller busy.", __func__));
1771 SET_BUSY(sc);
1772 ADAPTER_UNLOCK(sc);
1773
1774 rc = cxgb_uninit_synchronized(p);
1775
1776 ADAPTER_LOCK(sc);
1777 KASSERT(IS_BUSY(sc), ("%s: controller not busy.", __func__));
1778 CLR_BUSY(sc);
1779 wakeup_one(&sc->flags);
1780 done:
1781 ADAPTER_UNLOCK(sc);
1782 return (rc);
1783 }
1784
1785 /*
1786 * Called on "ifconfig down", and from port_detach
1787 */
1788 static int
cxgb_uninit_synchronized(struct port_info * pi)1789 cxgb_uninit_synchronized(struct port_info *pi)
1790 {
1791 struct adapter *sc = pi->adapter;
1792 struct ifnet *ifp = pi->ifp;
1793
1794 /*
1795 * taskqueue_drain may cause a deadlock if the adapter lock is held.
1796 */
1797 ADAPTER_LOCK_ASSERT_NOTOWNED(sc);
1798
1799 /*
1800 * Clear this port's bit from the open device map, and then drain all
1801 * the tasks that can access/manipulate this port's port_info or ifp.
1802 * We disable this port's interrupts here and so the slow/ext
1803 * interrupt tasks won't be enqueued. The tick task will continue to
1804 * be enqueued every second but the runs after this drain will not see
1805 * this port in the open device map.
1806 *
1807 * A well behaved task must take open_device_map into account and ignore
1808 * ports that are not open.
1809 */
1810 clrbit(&sc->open_device_map, pi->port_id);
1811 t3_port_intr_disable(sc, pi->port_id);
1812 taskqueue_drain(sc->tq, &sc->slow_intr_task);
1813 taskqueue_drain(sc->tq, &sc->tick_task);
1814
1815 callout_drain(&pi->link_check_ch);
1816 taskqueue_drain(sc->tq, &pi->link_check_task);
1817
1818 PORT_LOCK(pi);
1819 ifp->if_drv_flags &= ~(IFF_DRV_RUNNING | IFF_DRV_OACTIVE);
1820
1821 /* disable pause frames */
1822 t3_set_reg_field(sc, A_XGM_TX_CFG + pi->mac.offset, F_TXPAUSEEN, 0);
1823
1824 /* Reset RX FIFO HWM */
1825 t3_set_reg_field(sc, A_XGM_RXFIFO_CFG + pi->mac.offset,
1826 V_RXFIFOPAUSEHWM(M_RXFIFOPAUSEHWM), 0);
1827
1828 DELAY(100 * 1000);
1829
1830 /* Wait for TXFIFO empty */
1831 t3_wait_op_done(sc, A_XGM_TXFIFO_CFG + pi->mac.offset,
1832 F_TXFIFO_EMPTY, 1, 20, 5);
1833
1834 DELAY(100 * 1000);
1835 t3_mac_disable(&pi->mac, MAC_DIRECTION_RX);
1836
1837 pi->phy.ops->power_down(&pi->phy, 1);
1838
1839 PORT_UNLOCK(pi);
1840
1841 pi->link_config.link_ok = 0;
1842 t3_os_link_changed(sc, pi->port_id, 0, 0, 0, 0, 0);
1843
1844 if (sc->open_device_map == 0)
1845 cxgb_down(pi->adapter);
1846
1847 return (0);
1848 }
1849
1850 /*
1851 * Mark lro enabled or disabled in all qsets for this port
1852 */
1853 static int
cxgb_set_lro(struct port_info * p,int enabled)1854 cxgb_set_lro(struct port_info *p, int enabled)
1855 {
1856 int i;
1857 struct adapter *adp = p->adapter;
1858 struct sge_qset *q;
1859
1860 for (i = 0; i < p->nqsets; i++) {
1861 q = &adp->sge.qs[p->first_qset + i];
1862 q->lro.enabled = (enabled != 0);
1863 }
1864 return (0);
1865 }
1866
1867 static int
cxgb_ioctl(struct ifnet * ifp,unsigned long command,caddr_t data)1868 cxgb_ioctl(struct ifnet *ifp, unsigned long command, caddr_t data)
1869 {
1870 struct port_info *p = ifp->if_softc;
1871 struct adapter *sc = p->adapter;
1872 struct ifreq *ifr = (struct ifreq *)data;
1873 int flags, error = 0, mtu;
1874 uint32_t mask;
1875
1876 switch (command) {
1877 case SIOCSIFMTU:
1878 ADAPTER_LOCK(sc);
1879 error = IS_DOOMED(p) ? ENXIO : (IS_BUSY(sc) ? EBUSY : 0);
1880 if (error) {
1881 fail:
1882 ADAPTER_UNLOCK(sc);
1883 return (error);
1884 }
1885
1886 mtu = ifr->ifr_mtu;
1887 if ((mtu < ETHERMIN) || (mtu > ETHERMTU_JUMBO)) {
1888 error = EINVAL;
1889 } else {
1890 ifp->if_mtu = mtu;
1891 PORT_LOCK(p);
1892 cxgb_update_mac_settings(p);
1893 PORT_UNLOCK(p);
1894 }
1895 ADAPTER_UNLOCK(sc);
1896 break;
1897 case SIOCSIFFLAGS:
1898 ADAPTER_LOCK(sc);
1899 if (IS_DOOMED(p)) {
1900 error = ENXIO;
1901 goto fail;
1902 }
1903 if (ifp->if_flags & IFF_UP) {
1904 if (ifp->if_drv_flags & IFF_DRV_RUNNING) {
1905 flags = p->if_flags;
1906 if (((ifp->if_flags ^ flags) & IFF_PROMISC) ||
1907 ((ifp->if_flags ^ flags) & IFF_ALLMULTI)) {
1908 if (IS_BUSY(sc)) {
1909 error = EBUSY;
1910 goto fail;
1911 }
1912 PORT_LOCK(p);
1913 cxgb_update_mac_settings(p);
1914 PORT_UNLOCK(p);
1915 }
1916 ADAPTER_UNLOCK(sc);
1917 } else
1918 error = cxgb_init_locked(p);
1919 p->if_flags = ifp->if_flags;
1920 } else if (ifp->if_drv_flags & IFF_DRV_RUNNING)
1921 error = cxgb_uninit_locked(p);
1922 else
1923 ADAPTER_UNLOCK(sc);
1924
1925 ADAPTER_LOCK_ASSERT_NOTOWNED(sc);
1926 break;
1927 case SIOCADDMULTI:
1928 case SIOCDELMULTI:
1929 ADAPTER_LOCK(sc);
1930 error = IS_DOOMED(p) ? ENXIO : (IS_BUSY(sc) ? EBUSY : 0);
1931 if (error)
1932 goto fail;
1933
1934 if (ifp->if_drv_flags & IFF_DRV_RUNNING) {
1935 PORT_LOCK(p);
1936 cxgb_update_mac_settings(p);
1937 PORT_UNLOCK(p);
1938 }
1939 ADAPTER_UNLOCK(sc);
1940
1941 break;
1942 case SIOCSIFCAP:
1943 ADAPTER_LOCK(sc);
1944 error = IS_DOOMED(p) ? ENXIO : (IS_BUSY(sc) ? EBUSY : 0);
1945 if (error)
1946 goto fail;
1947
1948 mask = ifr->ifr_reqcap ^ ifp->if_capenable;
1949 if (mask & IFCAP_TXCSUM) {
1950 ifp->if_capenable ^= IFCAP_TXCSUM;
1951 ifp->if_hwassist ^= (CSUM_TCP | CSUM_UDP | CSUM_IP);
1952
1953 if (IFCAP_TSO4 & ifp->if_capenable &&
1954 !(IFCAP_TXCSUM & ifp->if_capenable)) {
1955 mask &= ~IFCAP_TSO4;
1956 ifp->if_capenable &= ~IFCAP_TSO4;
1957 if_printf(ifp,
1958 "tso4 disabled due to -txcsum.\n");
1959 }
1960 }
1961 if (mask & IFCAP_TXCSUM_IPV6) {
1962 ifp->if_capenable ^= IFCAP_TXCSUM_IPV6;
1963 ifp->if_hwassist ^= (CSUM_UDP_IPV6 | CSUM_TCP_IPV6);
1964
1965 if (IFCAP_TSO6 & ifp->if_capenable &&
1966 !(IFCAP_TXCSUM_IPV6 & ifp->if_capenable)) {
1967 mask &= ~IFCAP_TSO6;
1968 ifp->if_capenable &= ~IFCAP_TSO6;
1969 if_printf(ifp,
1970 "tso6 disabled due to -txcsum6.\n");
1971 }
1972 }
1973 if (mask & IFCAP_RXCSUM)
1974 ifp->if_capenable ^= IFCAP_RXCSUM;
1975 if (mask & IFCAP_RXCSUM_IPV6)
1976 ifp->if_capenable ^= IFCAP_RXCSUM_IPV6;
1977
1978 /*
1979 * Note that we leave CSUM_TSO alone (it is always set). The
1980 * kernel takes both IFCAP_TSOx and CSUM_TSO into account before
1981 * sending a TSO request our way, so it's sufficient to toggle
1982 * IFCAP_TSOx only.
1983 */
1984 if (mask & IFCAP_TSO4) {
1985 if (!(IFCAP_TSO4 & ifp->if_capenable) &&
1986 !(IFCAP_TXCSUM & ifp->if_capenable)) {
1987 if_printf(ifp, "enable txcsum first.\n");
1988 error = EAGAIN;
1989 goto fail;
1990 }
1991 ifp->if_capenable ^= IFCAP_TSO4;
1992 }
1993 if (mask & IFCAP_TSO6) {
1994 if (!(IFCAP_TSO6 & ifp->if_capenable) &&
1995 !(IFCAP_TXCSUM_IPV6 & ifp->if_capenable)) {
1996 if_printf(ifp, "enable txcsum6 first.\n");
1997 error = EAGAIN;
1998 goto fail;
1999 }
2000 ifp->if_capenable ^= IFCAP_TSO6;
2001 }
2002 if (mask & IFCAP_LRO) {
2003 ifp->if_capenable ^= IFCAP_LRO;
2004
2005 /* Safe to do this even if cxgb_up not called yet */
2006 cxgb_set_lro(p, ifp->if_capenable & IFCAP_LRO);
2007 }
2008 #ifdef TCP_OFFLOAD
2009 if (mask & IFCAP_TOE4) {
2010 int enable = (ifp->if_capenable ^ mask) & IFCAP_TOE4;
2011
2012 error = toe_capability(p, enable);
2013 if (error == 0)
2014 ifp->if_capenable ^= mask;
2015 }
2016 #endif
2017 if (mask & IFCAP_VLAN_HWTAGGING) {
2018 ifp->if_capenable ^= IFCAP_VLAN_HWTAGGING;
2019 if (ifp->if_drv_flags & IFF_DRV_RUNNING) {
2020 PORT_LOCK(p);
2021 cxgb_update_mac_settings(p);
2022 PORT_UNLOCK(p);
2023 }
2024 }
2025 if (mask & IFCAP_VLAN_MTU) {
2026 ifp->if_capenable ^= IFCAP_VLAN_MTU;
2027 if (ifp->if_drv_flags & IFF_DRV_RUNNING) {
2028 PORT_LOCK(p);
2029 cxgb_update_mac_settings(p);
2030 PORT_UNLOCK(p);
2031 }
2032 }
2033 if (mask & IFCAP_VLAN_HWTSO)
2034 ifp->if_capenable ^= IFCAP_VLAN_HWTSO;
2035 if (mask & IFCAP_VLAN_HWCSUM)
2036 ifp->if_capenable ^= IFCAP_VLAN_HWCSUM;
2037
2038 #ifdef VLAN_CAPABILITIES
2039 VLAN_CAPABILITIES(ifp);
2040 #endif
2041 ADAPTER_UNLOCK(sc);
2042 break;
2043 case SIOCSIFMEDIA:
2044 case SIOCGIFMEDIA:
2045 error = ifmedia_ioctl(ifp, ifr, &p->media, command);
2046 break;
2047 default:
2048 error = ether_ioctl(ifp, command, data);
2049 }
2050
2051 return (error);
2052 }
2053
2054 static int
cxgb_media_change(struct ifnet * ifp)2055 cxgb_media_change(struct ifnet *ifp)
2056 {
2057 return (EOPNOTSUPP);
2058 }
2059
2060 /*
2061 * Translates phy->modtype to the correct Ethernet media subtype.
2062 */
2063 static int
cxgb_ifm_type(int mod)2064 cxgb_ifm_type(int mod)
2065 {
2066 switch (mod) {
2067 case phy_modtype_sr:
2068 return (IFM_10G_SR);
2069 case phy_modtype_lr:
2070 return (IFM_10G_LR);
2071 case phy_modtype_lrm:
2072 return (IFM_10G_LRM);
2073 case phy_modtype_twinax:
2074 return (IFM_10G_TWINAX);
2075 case phy_modtype_twinax_long:
2076 return (IFM_10G_TWINAX_LONG);
2077 case phy_modtype_none:
2078 return (IFM_NONE);
2079 case phy_modtype_unknown:
2080 return (IFM_UNKNOWN);
2081 }
2082
2083 KASSERT(0, ("%s: modtype %d unknown", __func__, mod));
2084 return (IFM_UNKNOWN);
2085 }
2086
2087 /*
2088 * Rebuilds the ifmedia list for this port, and sets the current media.
2089 */
2090 static void
cxgb_build_medialist(struct port_info * p)2091 cxgb_build_medialist(struct port_info *p)
2092 {
2093 struct cphy *phy = &p->phy;
2094 struct ifmedia *media = &p->media;
2095 int mod = phy->modtype;
2096 int m = IFM_ETHER | IFM_FDX;
2097
2098 PORT_LOCK(p);
2099
2100 ifmedia_removeall(media);
2101 if (phy->caps & SUPPORTED_TP && phy->caps & SUPPORTED_Autoneg) {
2102 /* Copper (RJ45) */
2103
2104 if (phy->caps & SUPPORTED_10000baseT_Full)
2105 ifmedia_add(media, m | IFM_10G_T, mod, NULL);
2106
2107 if (phy->caps & SUPPORTED_1000baseT_Full)
2108 ifmedia_add(media, m | IFM_1000_T, mod, NULL);
2109
2110 if (phy->caps & SUPPORTED_100baseT_Full)
2111 ifmedia_add(media, m | IFM_100_TX, mod, NULL);
2112
2113 if (phy->caps & SUPPORTED_10baseT_Full)
2114 ifmedia_add(media, m | IFM_10_T, mod, NULL);
2115
2116 ifmedia_add(media, IFM_ETHER | IFM_AUTO, mod, NULL);
2117 ifmedia_set(media, IFM_ETHER | IFM_AUTO);
2118
2119 } else if (phy->caps & SUPPORTED_TP) {
2120 /* Copper (CX4) */
2121
2122 KASSERT(phy->caps & SUPPORTED_10000baseT_Full,
2123 ("%s: unexpected cap 0x%x", __func__, phy->caps));
2124
2125 ifmedia_add(media, m | IFM_10G_CX4, mod, NULL);
2126 ifmedia_set(media, m | IFM_10G_CX4);
2127
2128 } else if (phy->caps & SUPPORTED_FIBRE &&
2129 phy->caps & SUPPORTED_10000baseT_Full) {
2130 /* 10G optical (but includes SFP+ twinax) */
2131
2132 m |= cxgb_ifm_type(mod);
2133 if (IFM_SUBTYPE(m) == IFM_NONE)
2134 m &= ~IFM_FDX;
2135
2136 ifmedia_add(media, m, mod, NULL);
2137 ifmedia_set(media, m);
2138
2139 } else if (phy->caps & SUPPORTED_FIBRE &&
2140 phy->caps & SUPPORTED_1000baseT_Full) {
2141 /* 1G optical */
2142
2143 /* XXX: Lie and claim to be SX, could actually be any 1G-X */
2144 ifmedia_add(media, m | IFM_1000_SX, mod, NULL);
2145 ifmedia_set(media, m | IFM_1000_SX);
2146
2147 } else {
2148 KASSERT(0, ("%s: don't know how to handle 0x%x.", __func__,
2149 phy->caps));
2150 }
2151
2152 PORT_UNLOCK(p);
2153 }
2154
2155 static void
cxgb_media_status(struct ifnet * ifp,struct ifmediareq * ifmr)2156 cxgb_media_status(struct ifnet *ifp, struct ifmediareq *ifmr)
2157 {
2158 struct port_info *p = ifp->if_softc;
2159 struct ifmedia_entry *cur = p->media.ifm_cur;
2160 int speed = p->link_config.speed;
2161
2162 if (cur->ifm_data != p->phy.modtype) {
2163 cxgb_build_medialist(p);
2164 cur = p->media.ifm_cur;
2165 }
2166
2167 ifmr->ifm_status = IFM_AVALID;
2168 if (!p->link_config.link_ok)
2169 return;
2170
2171 ifmr->ifm_status |= IFM_ACTIVE;
2172
2173 /*
2174 * active and current will differ iff current media is autoselect. That
2175 * can happen only for copper RJ45.
2176 */
2177 if (IFM_SUBTYPE(cur->ifm_media) != IFM_AUTO)
2178 return;
2179 KASSERT(p->phy.caps & SUPPORTED_TP && p->phy.caps & SUPPORTED_Autoneg,
2180 ("%s: unexpected PHY caps 0x%x", __func__, p->phy.caps));
2181
2182 ifmr->ifm_active = IFM_ETHER | IFM_FDX;
2183 if (speed == SPEED_10000)
2184 ifmr->ifm_active |= IFM_10G_T;
2185 else if (speed == SPEED_1000)
2186 ifmr->ifm_active |= IFM_1000_T;
2187 else if (speed == SPEED_100)
2188 ifmr->ifm_active |= IFM_100_TX;
2189 else if (speed == SPEED_10)
2190 ifmr->ifm_active |= IFM_10_T;
2191 else
2192 KASSERT(0, ("%s: link up but speed unknown (%u)", __func__,
2193 speed));
2194 }
2195
2196 static uint64_t
cxgb_get_counter(struct ifnet * ifp,ift_counter c)2197 cxgb_get_counter(struct ifnet *ifp, ift_counter c)
2198 {
2199 struct port_info *pi = ifp->if_softc;
2200 struct adapter *sc = pi->adapter;
2201 struct cmac *mac = &pi->mac;
2202 struct mac_stats *mstats = &mac->stats;
2203
2204 cxgb_refresh_stats(pi);
2205
2206 switch (c) {
2207 case IFCOUNTER_IPACKETS:
2208 return (mstats->rx_frames);
2209
2210 case IFCOUNTER_IERRORS:
2211 return (mstats->rx_jabber + mstats->rx_data_errs +
2212 mstats->rx_sequence_errs + mstats->rx_runt +
2213 mstats->rx_too_long + mstats->rx_mac_internal_errs +
2214 mstats->rx_short + mstats->rx_fcs_errs);
2215
2216 case IFCOUNTER_OPACKETS:
2217 return (mstats->tx_frames);
2218
2219 case IFCOUNTER_OERRORS:
2220 return (mstats->tx_excess_collisions + mstats->tx_underrun +
2221 mstats->tx_len_errs + mstats->tx_mac_internal_errs +
2222 mstats->tx_excess_deferral + mstats->tx_fcs_errs);
2223
2224 case IFCOUNTER_COLLISIONS:
2225 return (mstats->tx_total_collisions);
2226
2227 case IFCOUNTER_IBYTES:
2228 return (mstats->rx_octets);
2229
2230 case IFCOUNTER_OBYTES:
2231 return (mstats->tx_octets);
2232
2233 case IFCOUNTER_IMCASTS:
2234 return (mstats->rx_mcast_frames);
2235
2236 case IFCOUNTER_OMCASTS:
2237 return (mstats->tx_mcast_frames);
2238
2239 case IFCOUNTER_IQDROPS:
2240 return (mstats->rx_cong_drops);
2241
2242 case IFCOUNTER_OQDROPS: {
2243 int i;
2244 uint64_t drops;
2245
2246 drops = 0;
2247 if (sc->flags & FULL_INIT_DONE) {
2248 for (i = pi->first_qset; i < pi->first_qset + pi->nqsets; i++)
2249 drops += sc->sge.qs[i].txq[TXQ_ETH].txq_mr->br_drops;
2250 }
2251
2252 return (drops);
2253
2254 }
2255
2256 default:
2257 return (if_get_counter_default(ifp, c));
2258 }
2259 }
2260
2261 static void
cxgb_async_intr(void * data)2262 cxgb_async_intr(void *data)
2263 {
2264 adapter_t *sc = data;
2265
2266 t3_write_reg(sc, A_PL_INT_ENABLE0, 0);
2267 (void) t3_read_reg(sc, A_PL_INT_ENABLE0);
2268 taskqueue_enqueue(sc->tq, &sc->slow_intr_task);
2269 }
2270
2271 static void
link_check_callout(void * arg)2272 link_check_callout(void *arg)
2273 {
2274 struct port_info *pi = arg;
2275 struct adapter *sc = pi->adapter;
2276
2277 if (!isset(&sc->open_device_map, pi->port_id))
2278 return;
2279
2280 taskqueue_enqueue(sc->tq, &pi->link_check_task);
2281 }
2282
2283 static void
check_link_status(void * arg,int pending)2284 check_link_status(void *arg, int pending)
2285 {
2286 struct port_info *pi = arg;
2287 struct adapter *sc = pi->adapter;
2288
2289 if (!isset(&sc->open_device_map, pi->port_id))
2290 return;
2291
2292 t3_link_changed(sc, pi->port_id);
2293
2294 if (pi->link_fault || !(pi->phy.caps & SUPPORTED_LINK_IRQ) ||
2295 pi->link_config.link_ok == 0)
2296 callout_reset(&pi->link_check_ch, hz, link_check_callout, pi);
2297 }
2298
2299 void
t3_os_link_intr(struct port_info * pi)2300 t3_os_link_intr(struct port_info *pi)
2301 {
2302 /*
2303 * Schedule a link check in the near future. If the link is flapping
2304 * rapidly we'll keep resetting the callout and delaying the check until
2305 * things stabilize a bit.
2306 */
2307 callout_reset(&pi->link_check_ch, hz / 4, link_check_callout, pi);
2308 }
2309
2310 static void
check_t3b2_mac(struct adapter * sc)2311 check_t3b2_mac(struct adapter *sc)
2312 {
2313 int i;
2314
2315 if (sc->flags & CXGB_SHUTDOWN)
2316 return;
2317
2318 for_each_port(sc, i) {
2319 struct port_info *p = &sc->port[i];
2320 int status;
2321 #ifdef INVARIANTS
2322 struct ifnet *ifp = p->ifp;
2323 #endif
2324
2325 if (!isset(&sc->open_device_map, p->port_id) || p->link_fault ||
2326 !p->link_config.link_ok)
2327 continue;
2328
2329 KASSERT(ifp->if_drv_flags & IFF_DRV_RUNNING,
2330 ("%s: state mismatch (drv_flags %x, device_map %x)",
2331 __func__, ifp->if_drv_flags, sc->open_device_map));
2332
2333 PORT_LOCK(p);
2334 status = t3b2_mac_watchdog_task(&p->mac);
2335 if (status == 1)
2336 p->mac.stats.num_toggled++;
2337 else if (status == 2) {
2338 struct cmac *mac = &p->mac;
2339
2340 cxgb_update_mac_settings(p);
2341 t3_link_start(&p->phy, mac, &p->link_config);
2342 t3_mac_enable(mac, MAC_DIRECTION_RX | MAC_DIRECTION_TX);
2343 t3_port_intr_enable(sc, p->port_id);
2344 p->mac.stats.num_resets++;
2345 }
2346 PORT_UNLOCK(p);
2347 }
2348 }
2349
2350 static void
cxgb_tick(void * arg)2351 cxgb_tick(void *arg)
2352 {
2353 adapter_t *sc = (adapter_t *)arg;
2354
2355 if (sc->flags & CXGB_SHUTDOWN)
2356 return;
2357
2358 taskqueue_enqueue(sc->tq, &sc->tick_task);
2359 callout_reset(&sc->cxgb_tick_ch, hz, cxgb_tick, sc);
2360 }
2361
2362 void
cxgb_refresh_stats(struct port_info * pi)2363 cxgb_refresh_stats(struct port_info *pi)
2364 {
2365 struct timeval tv;
2366 const struct timeval interval = {0, 250000}; /* 250ms */
2367
2368 getmicrotime(&tv);
2369 timevalsub(&tv, &interval);
2370 if (timevalcmp(&tv, &pi->last_refreshed, <))
2371 return;
2372
2373 PORT_LOCK(pi);
2374 t3_mac_update_stats(&pi->mac);
2375 PORT_UNLOCK(pi);
2376 getmicrotime(&pi->last_refreshed);
2377 }
2378
2379 static void
cxgb_tick_handler(void * arg,int count)2380 cxgb_tick_handler(void *arg, int count)
2381 {
2382 adapter_t *sc = (adapter_t *)arg;
2383 const struct adapter_params *p = &sc->params;
2384 int i;
2385 uint32_t cause, reset;
2386
2387 if (sc->flags & CXGB_SHUTDOWN || !(sc->flags & FULL_INIT_DONE))
2388 return;
2389
2390 if (p->rev == T3_REV_B2 && p->nports < 4 && sc->open_device_map)
2391 check_t3b2_mac(sc);
2392
2393 cause = t3_read_reg(sc, A_SG_INT_CAUSE) & (F_RSPQSTARVE | F_FLEMPTY);
2394 if (cause) {
2395 struct sge_qset *qs = &sc->sge.qs[0];
2396 uint32_t mask, v;
2397
2398 v = t3_read_reg(sc, A_SG_RSPQ_FL_STATUS) & ~0xff00;
2399
2400 mask = 1;
2401 for (i = 0; i < SGE_QSETS; i++) {
2402 if (v & mask)
2403 qs[i].rspq.starved++;
2404 mask <<= 1;
2405 }
2406
2407 mask <<= SGE_QSETS; /* skip RSPQXDISABLED */
2408
2409 for (i = 0; i < SGE_QSETS * 2; i++) {
2410 if (v & mask) {
2411 qs[i / 2].fl[i % 2].empty++;
2412 }
2413 mask <<= 1;
2414 }
2415
2416 /* clear */
2417 t3_write_reg(sc, A_SG_RSPQ_FL_STATUS, v);
2418 t3_write_reg(sc, A_SG_INT_CAUSE, cause);
2419 }
2420
2421 for (i = 0; i < sc->params.nports; i++) {
2422 struct port_info *pi = &sc->port[i];
2423 struct cmac *mac = &pi->mac;
2424
2425 if (!isset(&sc->open_device_map, pi->port_id))
2426 continue;
2427
2428 cxgb_refresh_stats(pi);
2429
2430 if (mac->multiport)
2431 continue;
2432
2433 /* Count rx fifo overflows, once per second */
2434 cause = t3_read_reg(sc, A_XGM_INT_CAUSE + mac->offset);
2435 reset = 0;
2436 if (cause & F_RXFIFO_OVERFLOW) {
2437 mac->stats.rx_fifo_ovfl++;
2438 reset |= F_RXFIFO_OVERFLOW;
2439 }
2440 t3_write_reg(sc, A_XGM_INT_CAUSE + mac->offset, reset);
2441 }
2442 }
2443
2444 static void
touch_bars(device_t dev)2445 touch_bars(device_t dev)
2446 {
2447 /*
2448 * Don't enable yet
2449 */
2450 #if !defined(__LP64__) && 0
2451 u32 v;
2452
2453 pci_read_config_dword(pdev, PCI_BASE_ADDRESS_1, &v);
2454 pci_write_config_dword(pdev, PCI_BASE_ADDRESS_1, v);
2455 pci_read_config_dword(pdev, PCI_BASE_ADDRESS_3, &v);
2456 pci_write_config_dword(pdev, PCI_BASE_ADDRESS_3, v);
2457 pci_read_config_dword(pdev, PCI_BASE_ADDRESS_5, &v);
2458 pci_write_config_dword(pdev, PCI_BASE_ADDRESS_5, v);
2459 #endif
2460 }
2461
2462 static int
set_eeprom(struct port_info * pi,const uint8_t * data,int len,int offset)2463 set_eeprom(struct port_info *pi, const uint8_t *data, int len, int offset)
2464 {
2465 uint8_t *buf;
2466 int err = 0;
2467 u32 aligned_offset, aligned_len, *p;
2468 struct adapter *adapter = pi->adapter;
2469
2470
2471 aligned_offset = offset & ~3;
2472 aligned_len = (len + (offset & 3) + 3) & ~3;
2473
2474 if (aligned_offset != offset || aligned_len != len) {
2475 buf = malloc(aligned_len, M_DEVBUF, M_WAITOK|M_ZERO);
2476 if (!buf)
2477 return (ENOMEM);
2478 err = t3_seeprom_read(adapter, aligned_offset, (u32 *)buf);
2479 if (!err && aligned_len > 4)
2480 err = t3_seeprom_read(adapter,
2481 aligned_offset + aligned_len - 4,
2482 (u32 *)&buf[aligned_len - 4]);
2483 if (err)
2484 goto out;
2485 memcpy(buf + (offset & 3), data, len);
2486 } else
2487 buf = (uint8_t *)(uintptr_t)data;
2488
2489 err = t3_seeprom_wp(adapter, 0);
2490 if (err)
2491 goto out;
2492
2493 for (p = (u32 *)buf; !err && aligned_len; aligned_len -= 4, p++) {
2494 err = t3_seeprom_write(adapter, aligned_offset, *p);
2495 aligned_offset += 4;
2496 }
2497
2498 if (!err)
2499 err = t3_seeprom_wp(adapter, 1);
2500 out:
2501 if (buf != data)
2502 free(buf, M_DEVBUF);
2503 return err;
2504 }
2505
2506
2507 static int
in_range(int val,int lo,int hi)2508 in_range(int val, int lo, int hi)
2509 {
2510 return val < 0 || (val <= hi && val >= lo);
2511 }
2512
2513 static int
cxgb_extension_open(struct cdev * dev,int flags,int fmp,struct thread * td)2514 cxgb_extension_open(struct cdev *dev, int flags, int fmp, struct thread *td)
2515 {
2516 return (0);
2517 }
2518
2519 static int
cxgb_extension_close(struct cdev * dev,int flags,int fmt,struct thread * td)2520 cxgb_extension_close(struct cdev *dev, int flags, int fmt, struct thread *td)
2521 {
2522 return (0);
2523 }
2524
2525 static int
cxgb_extension_ioctl(struct cdev * dev,unsigned long cmd,caddr_t data,int fflag,struct thread * td)2526 cxgb_extension_ioctl(struct cdev *dev, unsigned long cmd, caddr_t data,
2527 int fflag, struct thread *td)
2528 {
2529 int mmd, error = 0;
2530 struct port_info *pi = dev->si_drv1;
2531 adapter_t *sc = pi->adapter;
2532
2533 #ifdef PRIV_SUPPORTED
2534 if (priv_check(td, PRIV_DRIVER)) {
2535 if (cxgb_debug)
2536 printf("user does not have access to privileged ioctls\n");
2537 return (EPERM);
2538 }
2539 #else
2540 if (suser(td)) {
2541 if (cxgb_debug)
2542 printf("user does not have access to privileged ioctls\n");
2543 return (EPERM);
2544 }
2545 #endif
2546
2547 switch (cmd) {
2548 case CHELSIO_GET_MIIREG: {
2549 uint32_t val;
2550 struct cphy *phy = &pi->phy;
2551 struct ch_mii_data *mid = (struct ch_mii_data *)data;
2552
2553 if (!phy->mdio_read)
2554 return (EOPNOTSUPP);
2555 if (is_10G(sc)) {
2556 mmd = mid->phy_id >> 8;
2557 if (!mmd)
2558 mmd = MDIO_DEV_PCS;
2559 else if (mmd > MDIO_DEV_VEND2)
2560 return (EINVAL);
2561
2562 error = phy->mdio_read(sc, mid->phy_id & 0x1f, mmd,
2563 mid->reg_num, &val);
2564 } else
2565 error = phy->mdio_read(sc, mid->phy_id & 0x1f, 0,
2566 mid->reg_num & 0x1f, &val);
2567 if (error == 0)
2568 mid->val_out = val;
2569 break;
2570 }
2571 case CHELSIO_SET_MIIREG: {
2572 struct cphy *phy = &pi->phy;
2573 struct ch_mii_data *mid = (struct ch_mii_data *)data;
2574
2575 if (!phy->mdio_write)
2576 return (EOPNOTSUPP);
2577 if (is_10G(sc)) {
2578 mmd = mid->phy_id >> 8;
2579 if (!mmd)
2580 mmd = MDIO_DEV_PCS;
2581 else if (mmd > MDIO_DEV_VEND2)
2582 return (EINVAL);
2583
2584 error = phy->mdio_write(sc, mid->phy_id & 0x1f,
2585 mmd, mid->reg_num, mid->val_in);
2586 } else
2587 error = phy->mdio_write(sc, mid->phy_id & 0x1f, 0,
2588 mid->reg_num & 0x1f,
2589 mid->val_in);
2590 break;
2591 }
2592 case CHELSIO_SETREG: {
2593 struct ch_reg *edata = (struct ch_reg *)data;
2594 if ((edata->addr & 0x3) != 0 || edata->addr >= sc->mmio_len)
2595 return (EFAULT);
2596 t3_write_reg(sc, edata->addr, edata->val);
2597 break;
2598 }
2599 case CHELSIO_GETREG: {
2600 struct ch_reg *edata = (struct ch_reg *)data;
2601 if ((edata->addr & 0x3) != 0 || edata->addr >= sc->mmio_len)
2602 return (EFAULT);
2603 edata->val = t3_read_reg(sc, edata->addr);
2604 break;
2605 }
2606 case CHELSIO_GET_SGE_CONTEXT: {
2607 struct ch_cntxt *ecntxt = (struct ch_cntxt *)data;
2608 mtx_lock_spin(&sc->sge.reg_lock);
2609 switch (ecntxt->cntxt_type) {
2610 case CNTXT_TYPE_EGRESS:
2611 error = -t3_sge_read_ecntxt(sc, ecntxt->cntxt_id,
2612 ecntxt->data);
2613 break;
2614 case CNTXT_TYPE_FL:
2615 error = -t3_sge_read_fl(sc, ecntxt->cntxt_id,
2616 ecntxt->data);
2617 break;
2618 case CNTXT_TYPE_RSP:
2619 error = -t3_sge_read_rspq(sc, ecntxt->cntxt_id,
2620 ecntxt->data);
2621 break;
2622 case CNTXT_TYPE_CQ:
2623 error = -t3_sge_read_cq(sc, ecntxt->cntxt_id,
2624 ecntxt->data);
2625 break;
2626 default:
2627 error = EINVAL;
2628 break;
2629 }
2630 mtx_unlock_spin(&sc->sge.reg_lock);
2631 break;
2632 }
2633 case CHELSIO_GET_SGE_DESC: {
2634 struct ch_desc *edesc = (struct ch_desc *)data;
2635 int ret;
2636 if (edesc->queue_num >= SGE_QSETS * 6)
2637 return (EINVAL);
2638 ret = t3_get_desc(&sc->sge.qs[edesc->queue_num / 6],
2639 edesc->queue_num % 6, edesc->idx, edesc->data);
2640 if (ret < 0)
2641 return (EINVAL);
2642 edesc->size = ret;
2643 break;
2644 }
2645 case CHELSIO_GET_QSET_PARAMS: {
2646 struct qset_params *q;
2647 struct ch_qset_params *t = (struct ch_qset_params *)data;
2648 int q1 = pi->first_qset;
2649 int nqsets = pi->nqsets;
2650 int i;
2651
2652 if (t->qset_idx >= nqsets)
2653 return EINVAL;
2654
2655 i = q1 + t->qset_idx;
2656 q = &sc->params.sge.qset[i];
2657 t->rspq_size = q->rspq_size;
2658 t->txq_size[0] = q->txq_size[0];
2659 t->txq_size[1] = q->txq_size[1];
2660 t->txq_size[2] = q->txq_size[2];
2661 t->fl_size[0] = q->fl_size;
2662 t->fl_size[1] = q->jumbo_size;
2663 t->polling = q->polling;
2664 t->lro = q->lro;
2665 t->intr_lat = q->coalesce_usecs;
2666 t->cong_thres = q->cong_thres;
2667 t->qnum = i;
2668
2669 if ((sc->flags & FULL_INIT_DONE) == 0)
2670 t->vector = 0;
2671 else if (sc->flags & USING_MSIX)
2672 t->vector = rman_get_start(sc->msix_irq_res[i]);
2673 else
2674 t->vector = rman_get_start(sc->irq_res);
2675
2676 break;
2677 }
2678 case CHELSIO_GET_QSET_NUM: {
2679 struct ch_reg *edata = (struct ch_reg *)data;
2680 edata->val = pi->nqsets;
2681 break;
2682 }
2683 case CHELSIO_LOAD_FW: {
2684 uint8_t *fw_data;
2685 uint32_t vers;
2686 struct ch_mem_range *t = (struct ch_mem_range *)data;
2687
2688 /*
2689 * You're allowed to load a firmware only before FULL_INIT_DONE
2690 *
2691 * FW_UPTODATE is also set so the rest of the initialization
2692 * will not overwrite what was loaded here. This gives you the
2693 * flexibility to load any firmware (and maybe shoot yourself in
2694 * the foot).
2695 */
2696
2697 ADAPTER_LOCK(sc);
2698 if (sc->open_device_map || sc->flags & FULL_INIT_DONE) {
2699 ADAPTER_UNLOCK(sc);
2700 return (EBUSY);
2701 }
2702
2703 fw_data = malloc(t->len, M_DEVBUF, M_NOWAIT);
2704 if (!fw_data)
2705 error = ENOMEM;
2706 else
2707 error = copyin(t->buf, fw_data, t->len);
2708
2709 if (!error)
2710 error = -t3_load_fw(sc, fw_data, t->len);
2711
2712 if (t3_get_fw_version(sc, &vers) == 0) {
2713 snprintf(&sc->fw_version[0], sizeof(sc->fw_version),
2714 "%d.%d.%d", G_FW_VERSION_MAJOR(vers),
2715 G_FW_VERSION_MINOR(vers), G_FW_VERSION_MICRO(vers));
2716 }
2717
2718 if (!error)
2719 sc->flags |= FW_UPTODATE;
2720
2721 free(fw_data, M_DEVBUF);
2722 ADAPTER_UNLOCK(sc);
2723 break;
2724 }
2725 case CHELSIO_LOAD_BOOT: {
2726 uint8_t *boot_data;
2727 struct ch_mem_range *t = (struct ch_mem_range *)data;
2728
2729 boot_data = malloc(t->len, M_DEVBUF, M_NOWAIT);
2730 if (!boot_data)
2731 return ENOMEM;
2732
2733 error = copyin(t->buf, boot_data, t->len);
2734 if (!error)
2735 error = -t3_load_boot(sc, boot_data, t->len);
2736
2737 free(boot_data, M_DEVBUF);
2738 break;
2739 }
2740 case CHELSIO_GET_PM: {
2741 struct ch_pm *m = (struct ch_pm *)data;
2742 struct tp_params *p = &sc->params.tp;
2743
2744 if (!is_offload(sc))
2745 return (EOPNOTSUPP);
2746
2747 m->tx_pg_sz = p->tx_pg_size;
2748 m->tx_num_pg = p->tx_num_pgs;
2749 m->rx_pg_sz = p->rx_pg_size;
2750 m->rx_num_pg = p->rx_num_pgs;
2751 m->pm_total = p->pmtx_size + p->chan_rx_size * p->nchan;
2752
2753 break;
2754 }
2755 case CHELSIO_SET_PM: {
2756 struct ch_pm *m = (struct ch_pm *)data;
2757 struct tp_params *p = &sc->params.tp;
2758
2759 if (!is_offload(sc))
2760 return (EOPNOTSUPP);
2761 if (sc->flags & FULL_INIT_DONE)
2762 return (EBUSY);
2763
2764 if (!m->rx_pg_sz || (m->rx_pg_sz & (m->rx_pg_sz - 1)) ||
2765 !m->tx_pg_sz || (m->tx_pg_sz & (m->tx_pg_sz - 1)))
2766 return (EINVAL); /* not power of 2 */
2767 if (!(m->rx_pg_sz & 0x14000))
2768 return (EINVAL); /* not 16KB or 64KB */
2769 if (!(m->tx_pg_sz & 0x1554000))
2770 return (EINVAL);
2771 if (m->tx_num_pg == -1)
2772 m->tx_num_pg = p->tx_num_pgs;
2773 if (m->rx_num_pg == -1)
2774 m->rx_num_pg = p->rx_num_pgs;
2775 if (m->tx_num_pg % 24 || m->rx_num_pg % 24)
2776 return (EINVAL);
2777 if (m->rx_num_pg * m->rx_pg_sz > p->chan_rx_size ||
2778 m->tx_num_pg * m->tx_pg_sz > p->chan_tx_size)
2779 return (EINVAL);
2780
2781 p->rx_pg_size = m->rx_pg_sz;
2782 p->tx_pg_size = m->tx_pg_sz;
2783 p->rx_num_pgs = m->rx_num_pg;
2784 p->tx_num_pgs = m->tx_num_pg;
2785 break;
2786 }
2787 case CHELSIO_SETMTUTAB: {
2788 struct ch_mtus *m = (struct ch_mtus *)data;
2789 int i;
2790
2791 if (!is_offload(sc))
2792 return (EOPNOTSUPP);
2793 if (offload_running(sc))
2794 return (EBUSY);
2795 if (m->nmtus != NMTUS)
2796 return (EINVAL);
2797 if (m->mtus[0] < 81) /* accommodate SACK */
2798 return (EINVAL);
2799
2800 /*
2801 * MTUs must be in ascending order
2802 */
2803 for (i = 1; i < NMTUS; ++i)
2804 if (m->mtus[i] < m->mtus[i - 1])
2805 return (EINVAL);
2806
2807 memcpy(sc->params.mtus, m->mtus, sizeof(sc->params.mtus));
2808 break;
2809 }
2810 case CHELSIO_GETMTUTAB: {
2811 struct ch_mtus *m = (struct ch_mtus *)data;
2812
2813 if (!is_offload(sc))
2814 return (EOPNOTSUPP);
2815
2816 memcpy(m->mtus, sc->params.mtus, sizeof(m->mtus));
2817 m->nmtus = NMTUS;
2818 break;
2819 }
2820 case CHELSIO_GET_MEM: {
2821 struct ch_mem_range *t = (struct ch_mem_range *)data;
2822 struct mc7 *mem;
2823 uint8_t *useraddr;
2824 u64 buf[32];
2825
2826 /*
2827 * Use these to avoid modifying len/addr in the return
2828 * struct
2829 */
2830 uint32_t len = t->len, addr = t->addr;
2831
2832 if (!is_offload(sc))
2833 return (EOPNOTSUPP);
2834 if (!(sc->flags & FULL_INIT_DONE))
2835 return (EIO); /* need the memory controllers */
2836 if ((addr & 0x7) || (len & 0x7))
2837 return (EINVAL);
2838 if (t->mem_id == MEM_CM)
2839 mem = &sc->cm;
2840 else if (t->mem_id == MEM_PMRX)
2841 mem = &sc->pmrx;
2842 else if (t->mem_id == MEM_PMTX)
2843 mem = &sc->pmtx;
2844 else
2845 return (EINVAL);
2846
2847 /*
2848 * Version scheme:
2849 * bits 0..9: chip version
2850 * bits 10..15: chip revision
2851 */
2852 t->version = 3 | (sc->params.rev << 10);
2853
2854 /*
2855 * Read 256 bytes at a time as len can be large and we don't
2856 * want to use huge intermediate buffers.
2857 */
2858 useraddr = (uint8_t *)t->buf;
2859 while (len) {
2860 unsigned int chunk = min(len, sizeof(buf));
2861
2862 error = t3_mc7_bd_read(mem, addr / 8, chunk / 8, buf);
2863 if (error)
2864 return (-error);
2865 if (copyout(buf, useraddr, chunk))
2866 return (EFAULT);
2867 useraddr += chunk;
2868 addr += chunk;
2869 len -= chunk;
2870 }
2871 break;
2872 }
2873 case CHELSIO_READ_TCAM_WORD: {
2874 struct ch_tcam_word *t = (struct ch_tcam_word *)data;
2875
2876 if (!is_offload(sc))
2877 return (EOPNOTSUPP);
2878 if (!(sc->flags & FULL_INIT_DONE))
2879 return (EIO); /* need MC5 */
2880 return -t3_read_mc5_range(&sc->mc5, t->addr, 1, t->buf);
2881 break;
2882 }
2883 case CHELSIO_SET_TRACE_FILTER: {
2884 struct ch_trace *t = (struct ch_trace *)data;
2885 const struct trace_params *tp;
2886
2887 tp = (const struct trace_params *)&t->sip;
2888 if (t->config_tx)
2889 t3_config_trace_filter(sc, tp, 0, t->invert_match,
2890 t->trace_tx);
2891 if (t->config_rx)
2892 t3_config_trace_filter(sc, tp, 1, t->invert_match,
2893 t->trace_rx);
2894 break;
2895 }
2896 case CHELSIO_SET_PKTSCHED: {
2897 struct ch_pktsched_params *p = (struct ch_pktsched_params *)data;
2898 if (sc->open_device_map == 0)
2899 return (EAGAIN);
2900 send_pktsched_cmd(sc, p->sched, p->idx, p->min, p->max,
2901 p->binding);
2902 break;
2903 }
2904 case CHELSIO_IFCONF_GETREGS: {
2905 struct ch_ifconf_regs *regs = (struct ch_ifconf_regs *)data;
2906 int reglen = cxgb_get_regs_len();
2907 uint8_t *buf = malloc(reglen, M_DEVBUF, M_NOWAIT);
2908 if (buf == NULL) {
2909 return (ENOMEM);
2910 }
2911 if (regs->len > reglen)
2912 regs->len = reglen;
2913 else if (regs->len < reglen)
2914 error = ENOBUFS;
2915
2916 if (!error) {
2917 cxgb_get_regs(sc, regs, buf);
2918 error = copyout(buf, regs->data, reglen);
2919 }
2920 free(buf, M_DEVBUF);
2921
2922 break;
2923 }
2924 case CHELSIO_SET_HW_SCHED: {
2925 struct ch_hw_sched *t = (struct ch_hw_sched *)data;
2926 unsigned int ticks_per_usec = core_ticks_per_usec(sc);
2927
2928 if ((sc->flags & FULL_INIT_DONE) == 0)
2929 return (EAGAIN); /* need TP to be initialized */
2930 if (t->sched >= NTX_SCHED || !in_range(t->mode, 0, 1) ||
2931 !in_range(t->channel, 0, 1) ||
2932 !in_range(t->kbps, 0, 10000000) ||
2933 !in_range(t->class_ipg, 0, 10000 * 65535 / ticks_per_usec) ||
2934 !in_range(t->flow_ipg, 0,
2935 dack_ticks_to_usec(sc, 0x7ff)))
2936 return (EINVAL);
2937
2938 if (t->kbps >= 0) {
2939 error = t3_config_sched(sc, t->kbps, t->sched);
2940 if (error < 0)
2941 return (-error);
2942 }
2943 if (t->class_ipg >= 0)
2944 t3_set_sched_ipg(sc, t->sched, t->class_ipg);
2945 if (t->flow_ipg >= 0) {
2946 t->flow_ipg *= 1000; /* us -> ns */
2947 t3_set_pace_tbl(sc, &t->flow_ipg, t->sched, 1);
2948 }
2949 if (t->mode >= 0) {
2950 int bit = 1 << (S_TX_MOD_TIMER_MODE + t->sched);
2951
2952 t3_set_reg_field(sc, A_TP_TX_MOD_QUEUE_REQ_MAP,
2953 bit, t->mode ? bit : 0);
2954 }
2955 if (t->channel >= 0)
2956 t3_set_reg_field(sc, A_TP_TX_MOD_QUEUE_REQ_MAP,
2957 1 << t->sched, t->channel << t->sched);
2958 break;
2959 }
2960 case CHELSIO_GET_EEPROM: {
2961 int i;
2962 struct ch_eeprom *e = (struct ch_eeprom *)data;
2963 uint8_t *buf;
2964
2965 if (e->offset & 3 || e->offset >= EEPROMSIZE ||
2966 e->len > EEPROMSIZE || e->offset + e->len > EEPROMSIZE) {
2967 return (EINVAL);
2968 }
2969
2970 buf = malloc(EEPROMSIZE, M_DEVBUF, M_NOWAIT);
2971 if (buf == NULL) {
2972 return (ENOMEM);
2973 }
2974 e->magic = EEPROM_MAGIC;
2975 for (i = e->offset & ~3; !error && i < e->offset + e->len; i += 4)
2976 error = -t3_seeprom_read(sc, i, (uint32_t *)&buf[i]);
2977
2978 if (!error)
2979 error = copyout(buf + e->offset, e->data, e->len);
2980
2981 free(buf, M_DEVBUF);
2982 break;
2983 }
2984 case CHELSIO_CLEAR_STATS: {
2985 if (!(sc->flags & FULL_INIT_DONE))
2986 return EAGAIN;
2987
2988 PORT_LOCK(pi);
2989 t3_mac_update_stats(&pi->mac);
2990 memset(&pi->mac.stats, 0, sizeof(pi->mac.stats));
2991 PORT_UNLOCK(pi);
2992 break;
2993 }
2994 case CHELSIO_GET_UP_LA: {
2995 struct ch_up_la *la = (struct ch_up_la *)data;
2996 uint8_t *buf = malloc(LA_BUFSIZE, M_DEVBUF, M_NOWAIT);
2997 if (buf == NULL) {
2998 return (ENOMEM);
2999 }
3000 if (la->bufsize < LA_BUFSIZE)
3001 error = ENOBUFS;
3002
3003 if (!error)
3004 error = -t3_get_up_la(sc, &la->stopped, &la->idx,
3005 &la->bufsize, buf);
3006 if (!error)
3007 error = copyout(buf, la->data, la->bufsize);
3008
3009 free(buf, M_DEVBUF);
3010 break;
3011 }
3012 case CHELSIO_GET_UP_IOQS: {
3013 struct ch_up_ioqs *ioqs = (struct ch_up_ioqs *)data;
3014 uint8_t *buf = malloc(IOQS_BUFSIZE, M_DEVBUF, M_NOWAIT);
3015 uint32_t *v;
3016
3017 if (buf == NULL) {
3018 return (ENOMEM);
3019 }
3020 if (ioqs->bufsize < IOQS_BUFSIZE)
3021 error = ENOBUFS;
3022
3023 if (!error)
3024 error = -t3_get_up_ioqs(sc, &ioqs->bufsize, buf);
3025
3026 if (!error) {
3027 v = (uint32_t *)buf;
3028
3029 ioqs->ioq_rx_enable = *v++;
3030 ioqs->ioq_tx_enable = *v++;
3031 ioqs->ioq_rx_status = *v++;
3032 ioqs->ioq_tx_status = *v++;
3033
3034 error = copyout(v, ioqs->data, ioqs->bufsize);
3035 }
3036
3037 free(buf, M_DEVBUF);
3038 break;
3039 }
3040 case CHELSIO_SET_FILTER: {
3041 struct ch_filter *f = (struct ch_filter *)data;
3042 struct filter_info *p;
3043 unsigned int nfilters = sc->params.mc5.nfilters;
3044
3045 if (!is_offload(sc))
3046 return (EOPNOTSUPP); /* No TCAM */
3047 if (!(sc->flags & FULL_INIT_DONE))
3048 return (EAGAIN); /* mc5 not setup yet */
3049 if (nfilters == 0)
3050 return (EBUSY); /* TOE will use TCAM */
3051
3052 /* sanity checks */
3053 if (f->filter_id >= nfilters ||
3054 (f->val.dip && f->mask.dip != 0xffffffff) ||
3055 (f->val.sport && f->mask.sport != 0xffff) ||
3056 (f->val.dport && f->mask.dport != 0xffff) ||
3057 (f->val.vlan && f->mask.vlan != 0xfff) ||
3058 (f->val.vlan_prio &&
3059 f->mask.vlan_prio != FILTER_NO_VLAN_PRI) ||
3060 (f->mac_addr_idx != 0xffff && f->mac_addr_idx > 15) ||
3061 f->qset >= SGE_QSETS ||
3062 sc->rrss_map[f->qset] >= RSS_TABLE_SIZE)
3063 return (EINVAL);
3064
3065 /* Was allocated with M_WAITOK */
3066 KASSERT(sc->filters, ("filter table NULL\n"));
3067
3068 p = &sc->filters[f->filter_id];
3069 if (p->locked)
3070 return (EPERM);
3071
3072 bzero(p, sizeof(*p));
3073 p->sip = f->val.sip;
3074 p->sip_mask = f->mask.sip;
3075 p->dip = f->val.dip;
3076 p->sport = f->val.sport;
3077 p->dport = f->val.dport;
3078 p->vlan = f->mask.vlan ? f->val.vlan : 0xfff;
3079 p->vlan_prio = f->mask.vlan_prio ? (f->val.vlan_prio & 6) :
3080 FILTER_NO_VLAN_PRI;
3081 p->mac_hit = f->mac_hit;
3082 p->mac_vld = f->mac_addr_idx != 0xffff;
3083 p->mac_idx = f->mac_addr_idx;
3084 p->pkt_type = f->proto;
3085 p->report_filter_id = f->want_filter_id;
3086 p->pass = f->pass;
3087 p->rss = f->rss;
3088 p->qset = f->qset;
3089
3090 error = set_filter(sc, f->filter_id, p);
3091 if (error == 0)
3092 p->valid = 1;
3093 break;
3094 }
3095 case CHELSIO_DEL_FILTER: {
3096 struct ch_filter *f = (struct ch_filter *)data;
3097 struct filter_info *p;
3098 unsigned int nfilters = sc->params.mc5.nfilters;
3099
3100 if (!is_offload(sc))
3101 return (EOPNOTSUPP);
3102 if (!(sc->flags & FULL_INIT_DONE))
3103 return (EAGAIN);
3104 if (nfilters == 0 || sc->filters == NULL)
3105 return (EINVAL);
3106 if (f->filter_id >= nfilters)
3107 return (EINVAL);
3108
3109 p = &sc->filters[f->filter_id];
3110 if (p->locked)
3111 return (EPERM);
3112 if (!p->valid)
3113 return (EFAULT); /* Read "Bad address" as "Bad index" */
3114
3115 bzero(p, sizeof(*p));
3116 p->sip = p->sip_mask = 0xffffffff;
3117 p->vlan = 0xfff;
3118 p->vlan_prio = FILTER_NO_VLAN_PRI;
3119 p->pkt_type = 1;
3120 error = set_filter(sc, f->filter_id, p);
3121 break;
3122 }
3123 case CHELSIO_GET_FILTER: {
3124 struct ch_filter *f = (struct ch_filter *)data;
3125 struct filter_info *p;
3126 unsigned int i, nfilters = sc->params.mc5.nfilters;
3127
3128 if (!is_offload(sc))
3129 return (EOPNOTSUPP);
3130 if (!(sc->flags & FULL_INIT_DONE))
3131 return (EAGAIN);
3132 if (nfilters == 0 || sc->filters == NULL)
3133 return (EINVAL);
3134
3135 i = f->filter_id == 0xffffffff ? 0 : f->filter_id + 1;
3136 for (; i < nfilters; i++) {
3137 p = &sc->filters[i];
3138 if (!p->valid)
3139 continue;
3140
3141 bzero(f, sizeof(*f));
3142
3143 f->filter_id = i;
3144 f->val.sip = p->sip;
3145 f->mask.sip = p->sip_mask;
3146 f->val.dip = p->dip;
3147 f->mask.dip = p->dip ? 0xffffffff : 0;
3148 f->val.sport = p->sport;
3149 f->mask.sport = p->sport ? 0xffff : 0;
3150 f->val.dport = p->dport;
3151 f->mask.dport = p->dport ? 0xffff : 0;
3152 f->val.vlan = p->vlan == 0xfff ? 0 : p->vlan;
3153 f->mask.vlan = p->vlan == 0xfff ? 0 : 0xfff;
3154 f->val.vlan_prio = p->vlan_prio == FILTER_NO_VLAN_PRI ?
3155 0 : p->vlan_prio;
3156 f->mask.vlan_prio = p->vlan_prio == FILTER_NO_VLAN_PRI ?
3157 0 : FILTER_NO_VLAN_PRI;
3158 f->mac_hit = p->mac_hit;
3159 f->mac_addr_idx = p->mac_vld ? p->mac_idx : 0xffff;
3160 f->proto = p->pkt_type;
3161 f->want_filter_id = p->report_filter_id;
3162 f->pass = p->pass;
3163 f->rss = p->rss;
3164 f->qset = p->qset;
3165
3166 break;
3167 }
3168
3169 if (i == nfilters)
3170 f->filter_id = 0xffffffff;
3171 break;
3172 }
3173 default:
3174 return (EOPNOTSUPP);
3175 break;
3176 }
3177
3178 return (error);
3179 }
3180
3181 static __inline void
reg_block_dump(struct adapter * ap,uint8_t * buf,unsigned int start,unsigned int end)3182 reg_block_dump(struct adapter *ap, uint8_t *buf, unsigned int start,
3183 unsigned int end)
3184 {
3185 uint32_t *p = (uint32_t *)(buf + start);
3186
3187 for ( ; start <= end; start += sizeof(uint32_t))
3188 *p++ = t3_read_reg(ap, start);
3189 }
3190
3191 #define T3_REGMAP_SIZE (3 * 1024)
3192 static int
cxgb_get_regs_len(void)3193 cxgb_get_regs_len(void)
3194 {
3195 return T3_REGMAP_SIZE;
3196 }
3197
3198 static void
cxgb_get_regs(adapter_t * sc,struct ch_ifconf_regs * regs,uint8_t * buf)3199 cxgb_get_regs(adapter_t *sc, struct ch_ifconf_regs *regs, uint8_t *buf)
3200 {
3201
3202 /*
3203 * Version scheme:
3204 * bits 0..9: chip version
3205 * bits 10..15: chip revision
3206 * bit 31: set for PCIe cards
3207 */
3208 regs->version = 3 | (sc->params.rev << 10) | (is_pcie(sc) << 31);
3209
3210 /*
3211 * We skip the MAC statistics registers because they are clear-on-read.
3212 * Also reading multi-register stats would need to synchronize with the
3213 * periodic mac stats accumulation. Hard to justify the complexity.
3214 */
3215 memset(buf, 0, cxgb_get_regs_len());
3216 reg_block_dump(sc, buf, 0, A_SG_RSPQ_CREDIT_RETURN);
3217 reg_block_dump(sc, buf, A_SG_HI_DRB_HI_THRSH, A_ULPRX_PBL_ULIMIT);
3218 reg_block_dump(sc, buf, A_ULPTX_CONFIG, A_MPS_INT_CAUSE);
3219 reg_block_dump(sc, buf, A_CPL_SWITCH_CNTRL, A_CPL_MAP_TBL_DATA);
3220 reg_block_dump(sc, buf, A_SMB_GLOBAL_TIME_CFG, A_XGM_SERDES_STAT3);
3221 reg_block_dump(sc, buf, A_XGM_SERDES_STATUS0,
3222 XGM_REG(A_XGM_SERDES_STAT3, 1));
3223 reg_block_dump(sc, buf, XGM_REG(A_XGM_SERDES_STATUS0, 1),
3224 XGM_REG(A_XGM_RX_SPI4_SOP_EOP_CNT, 1));
3225 }
3226
3227 static int
alloc_filters(struct adapter * sc)3228 alloc_filters(struct adapter *sc)
3229 {
3230 struct filter_info *p;
3231 unsigned int nfilters = sc->params.mc5.nfilters;
3232
3233 if (nfilters == 0)
3234 return (0);
3235
3236 p = malloc(sizeof(*p) * nfilters, M_DEVBUF, M_WAITOK | M_ZERO);
3237 sc->filters = p;
3238
3239 p = &sc->filters[nfilters - 1];
3240 p->vlan = 0xfff;
3241 p->vlan_prio = FILTER_NO_VLAN_PRI;
3242 p->pass = p->rss = p->valid = p->locked = 1;
3243
3244 return (0);
3245 }
3246
3247 static int
setup_hw_filters(struct adapter * sc)3248 setup_hw_filters(struct adapter *sc)
3249 {
3250 int i, rc;
3251 unsigned int nfilters = sc->params.mc5.nfilters;
3252
3253 if (!sc->filters)
3254 return (0);
3255
3256 t3_enable_filters(sc);
3257
3258 for (i = rc = 0; i < nfilters && !rc; i++) {
3259 if (sc->filters[i].locked)
3260 rc = set_filter(sc, i, &sc->filters[i]);
3261 }
3262
3263 return (rc);
3264 }
3265
3266 static int
set_filter(struct adapter * sc,int id,const struct filter_info * f)3267 set_filter(struct adapter *sc, int id, const struct filter_info *f)
3268 {
3269 int len;
3270 struct mbuf *m;
3271 struct ulp_txpkt *txpkt;
3272 struct work_request_hdr *wr;
3273 struct cpl_pass_open_req *oreq;
3274 struct cpl_set_tcb_field *sreq;
3275
3276 len = sizeof(*wr) + sizeof(*oreq) + 2 * sizeof(*sreq);
3277 KASSERT(len <= MHLEN, ("filter request too big for an mbuf"));
3278
3279 id += t3_mc5_size(&sc->mc5) - sc->params.mc5.nroutes -
3280 sc->params.mc5.nfilters;
3281
3282 m = m_gethdr(M_WAITOK, MT_DATA);
3283 m->m_len = m->m_pkthdr.len = len;
3284 bzero(mtod(m, char *), len);
3285
3286 wr = mtod(m, struct work_request_hdr *);
3287 wr->wrh_hi = htonl(V_WR_OP(FW_WROPCODE_BYPASS) | F_WR_ATOMIC);
3288
3289 oreq = (struct cpl_pass_open_req *)(wr + 1);
3290 txpkt = (struct ulp_txpkt *)oreq;
3291 txpkt->cmd_dest = htonl(V_ULPTX_CMD(ULP_TXPKT));
3292 txpkt->len = htonl(V_ULPTX_NFLITS(sizeof(*oreq) / 8));
3293 OPCODE_TID(oreq) = htonl(MK_OPCODE_TID(CPL_PASS_OPEN_REQ, id));
3294 oreq->local_port = htons(f->dport);
3295 oreq->peer_port = htons(f->sport);
3296 oreq->local_ip = htonl(f->dip);
3297 oreq->peer_ip = htonl(f->sip);
3298 oreq->peer_netmask = htonl(f->sip_mask);
3299 oreq->opt0h = 0;
3300 oreq->opt0l = htonl(F_NO_OFFLOAD);
3301 oreq->opt1 = htonl(V_MAC_MATCH_VALID(f->mac_vld) |
3302 V_CONN_POLICY(CPL_CONN_POLICY_FILTER) |
3303 V_VLAN_PRI(f->vlan_prio >> 1) |
3304 V_VLAN_PRI_VALID(f->vlan_prio != FILTER_NO_VLAN_PRI) |
3305 V_PKT_TYPE(f->pkt_type) | V_OPT1_VLAN(f->vlan) |
3306 V_MAC_MATCH(f->mac_idx | (f->mac_hit << 4)));
3307
3308 sreq = (struct cpl_set_tcb_field *)(oreq + 1);
3309 set_tcb_field_ulp(sreq, id, 1, 0x1800808000ULL,
3310 (f->report_filter_id << 15) | (1 << 23) |
3311 ((u64)f->pass << 35) | ((u64)!f->rss << 36));
3312 set_tcb_field_ulp(sreq + 1, id, 0, 0xffffffff, (2 << 19) | 1);
3313 t3_mgmt_tx(sc, m);
3314
3315 if (f->pass && !f->rss) {
3316 len = sizeof(*sreq);
3317 m = m_gethdr(M_WAITOK, MT_DATA);
3318 m->m_len = m->m_pkthdr.len = len;
3319 bzero(mtod(m, char *), len);
3320 sreq = mtod(m, struct cpl_set_tcb_field *);
3321 sreq->wr.wrh_hi = htonl(V_WR_OP(FW_WROPCODE_FORWARD));
3322 mk_set_tcb_field(sreq, id, 25, 0x3f80000,
3323 (u64)sc->rrss_map[f->qset] << 19);
3324 t3_mgmt_tx(sc, m);
3325 }
3326 return 0;
3327 }
3328
3329 static inline void
mk_set_tcb_field(struct cpl_set_tcb_field * req,unsigned int tid,unsigned int word,u64 mask,u64 val)3330 mk_set_tcb_field(struct cpl_set_tcb_field *req, unsigned int tid,
3331 unsigned int word, u64 mask, u64 val)
3332 {
3333 OPCODE_TID(req) = htonl(MK_OPCODE_TID(CPL_SET_TCB_FIELD, tid));
3334 req->reply = V_NO_REPLY(1);
3335 req->cpu_idx = 0;
3336 req->word = htons(word);
3337 req->mask = htobe64(mask);
3338 req->val = htobe64(val);
3339 }
3340
3341 static inline void
set_tcb_field_ulp(struct cpl_set_tcb_field * req,unsigned int tid,unsigned int word,u64 mask,u64 val)3342 set_tcb_field_ulp(struct cpl_set_tcb_field *req, unsigned int tid,
3343 unsigned int word, u64 mask, u64 val)
3344 {
3345 struct ulp_txpkt *txpkt = (struct ulp_txpkt *)req;
3346
3347 txpkt->cmd_dest = htonl(V_ULPTX_CMD(ULP_TXPKT));
3348 txpkt->len = htonl(V_ULPTX_NFLITS(sizeof(*req) / 8));
3349 mk_set_tcb_field(req, tid, word, mask, val);
3350 }
3351
3352 void
t3_iterate(void (* func)(struct adapter *,void *),void * arg)3353 t3_iterate(void (*func)(struct adapter *, void *), void *arg)
3354 {
3355 struct adapter *sc;
3356
3357 mtx_lock(&t3_list_lock);
3358 SLIST_FOREACH(sc, &t3_list, link) {
3359 /*
3360 * func should not make any assumptions about what state sc is
3361 * in - the only guarantee is that sc->sc_lock is a valid lock.
3362 */
3363 func(sc, arg);
3364 }
3365 mtx_unlock(&t3_list_lock);
3366 }
3367
3368 #ifdef TCP_OFFLOAD
3369 static int
toe_capability(struct port_info * pi,int enable)3370 toe_capability(struct port_info *pi, int enable)
3371 {
3372 int rc;
3373 struct adapter *sc = pi->adapter;
3374
3375 ADAPTER_LOCK_ASSERT_OWNED(sc);
3376
3377 if (!is_offload(sc))
3378 return (ENODEV);
3379
3380 if (enable) {
3381 if (!(sc->flags & FULL_INIT_DONE)) {
3382 log(LOG_WARNING,
3383 "You must enable a cxgb interface first\n");
3384 return (EAGAIN);
3385 }
3386
3387 if (isset(&sc->offload_map, pi->port_id))
3388 return (0);
3389
3390 if (!(sc->flags & TOM_INIT_DONE)) {
3391 rc = t3_activate_uld(sc, ULD_TOM);
3392 if (rc == EAGAIN) {
3393 log(LOG_WARNING,
3394 "You must kldload t3_tom.ko before trying "
3395 "to enable TOE on a cxgb interface.\n");
3396 }
3397 if (rc != 0)
3398 return (rc);
3399 KASSERT(sc->tom_softc != NULL,
3400 ("%s: TOM activated but softc NULL", __func__));
3401 KASSERT(sc->flags & TOM_INIT_DONE,
3402 ("%s: TOM activated but flag not set", __func__));
3403 }
3404
3405 setbit(&sc->offload_map, pi->port_id);
3406
3407 /*
3408 * XXX: Temporary code to allow iWARP to be enabled when TOE is
3409 * enabled on any port. Need to figure out how to enable,
3410 * disable, load, and unload iWARP cleanly.
3411 */
3412 if (!isset(&sc->offload_map, MAX_NPORTS) &&
3413 t3_activate_uld(sc, ULD_IWARP) == 0)
3414 setbit(&sc->offload_map, MAX_NPORTS);
3415 } else {
3416 if (!isset(&sc->offload_map, pi->port_id))
3417 return (0);
3418
3419 KASSERT(sc->flags & TOM_INIT_DONE,
3420 ("%s: TOM never initialized?", __func__));
3421 clrbit(&sc->offload_map, pi->port_id);
3422 }
3423
3424 return (0);
3425 }
3426
3427 /*
3428 * Add an upper layer driver to the global list.
3429 */
3430 int
t3_register_uld(struct uld_info * ui)3431 t3_register_uld(struct uld_info *ui)
3432 {
3433 int rc = 0;
3434 struct uld_info *u;
3435
3436 mtx_lock(&t3_uld_list_lock);
3437 SLIST_FOREACH(u, &t3_uld_list, link) {
3438 if (u->uld_id == ui->uld_id) {
3439 rc = EEXIST;
3440 goto done;
3441 }
3442 }
3443
3444 SLIST_INSERT_HEAD(&t3_uld_list, ui, link);
3445 ui->refcount = 0;
3446 done:
3447 mtx_unlock(&t3_uld_list_lock);
3448 return (rc);
3449 }
3450
3451 int
t3_unregister_uld(struct uld_info * ui)3452 t3_unregister_uld(struct uld_info *ui)
3453 {
3454 int rc = EINVAL;
3455 struct uld_info *u;
3456
3457 mtx_lock(&t3_uld_list_lock);
3458
3459 SLIST_FOREACH(u, &t3_uld_list, link) {
3460 if (u == ui) {
3461 if (ui->refcount > 0) {
3462 rc = EBUSY;
3463 goto done;
3464 }
3465
3466 SLIST_REMOVE(&t3_uld_list, ui, uld_info, link);
3467 rc = 0;
3468 goto done;
3469 }
3470 }
3471 done:
3472 mtx_unlock(&t3_uld_list_lock);
3473 return (rc);
3474 }
3475
3476 int
t3_activate_uld(struct adapter * sc,int id)3477 t3_activate_uld(struct adapter *sc, int id)
3478 {
3479 int rc = EAGAIN;
3480 struct uld_info *ui;
3481
3482 mtx_lock(&t3_uld_list_lock);
3483
3484 SLIST_FOREACH(ui, &t3_uld_list, link) {
3485 if (ui->uld_id == id) {
3486 rc = ui->activate(sc);
3487 if (rc == 0)
3488 ui->refcount++;
3489 goto done;
3490 }
3491 }
3492 done:
3493 mtx_unlock(&t3_uld_list_lock);
3494
3495 return (rc);
3496 }
3497
3498 int
t3_deactivate_uld(struct adapter * sc,int id)3499 t3_deactivate_uld(struct adapter *sc, int id)
3500 {
3501 int rc = EINVAL;
3502 struct uld_info *ui;
3503
3504 mtx_lock(&t3_uld_list_lock);
3505
3506 SLIST_FOREACH(ui, &t3_uld_list, link) {
3507 if (ui->uld_id == id) {
3508 rc = ui->deactivate(sc);
3509 if (rc == 0)
3510 ui->refcount--;
3511 goto done;
3512 }
3513 }
3514 done:
3515 mtx_unlock(&t3_uld_list_lock);
3516
3517 return (rc);
3518 }
3519
3520 static int
cpl_not_handled(struct sge_qset * qs __unused,struct rsp_desc * r __unused,struct mbuf * m)3521 cpl_not_handled(struct sge_qset *qs __unused, struct rsp_desc *r __unused,
3522 struct mbuf *m)
3523 {
3524 m_freem(m);
3525 return (EDOOFUS);
3526 }
3527
3528 int
t3_register_cpl_handler(struct adapter * sc,int opcode,cpl_handler_t h)3529 t3_register_cpl_handler(struct adapter *sc, int opcode, cpl_handler_t h)
3530 {
3531 uintptr_t *loc, new;
3532
3533 if (opcode >= NUM_CPL_HANDLERS)
3534 return (EINVAL);
3535
3536 new = h ? (uintptr_t)h : (uintptr_t)cpl_not_handled;
3537 loc = (uintptr_t *) &sc->cpl_handler[opcode];
3538 atomic_store_rel_ptr(loc, new);
3539
3540 return (0);
3541 }
3542 #endif
3543
3544 static int
cxgbc_mod_event(module_t mod,int cmd,void * arg)3545 cxgbc_mod_event(module_t mod, int cmd, void *arg)
3546 {
3547 int rc = 0;
3548
3549 switch (cmd) {
3550 case MOD_LOAD:
3551 mtx_init(&t3_list_lock, "T3 adapters", 0, MTX_DEF);
3552 SLIST_INIT(&t3_list);
3553 #ifdef TCP_OFFLOAD
3554 mtx_init(&t3_uld_list_lock, "T3 ULDs", 0, MTX_DEF);
3555 SLIST_INIT(&t3_uld_list);
3556 #endif
3557 break;
3558
3559 case MOD_UNLOAD:
3560 #ifdef TCP_OFFLOAD
3561 mtx_lock(&t3_uld_list_lock);
3562 if (!SLIST_EMPTY(&t3_uld_list)) {
3563 rc = EBUSY;
3564 mtx_unlock(&t3_uld_list_lock);
3565 break;
3566 }
3567 mtx_unlock(&t3_uld_list_lock);
3568 mtx_destroy(&t3_uld_list_lock);
3569 #endif
3570 mtx_lock(&t3_list_lock);
3571 if (!SLIST_EMPTY(&t3_list)) {
3572 rc = EBUSY;
3573 mtx_unlock(&t3_list_lock);
3574 break;
3575 }
3576 mtx_unlock(&t3_list_lock);
3577 mtx_destroy(&t3_list_lock);
3578 break;
3579 }
3580
3581 return (rc);
3582 }
3583