summaryrefslogtreecommitdiffstats
path: root/source4/ldap_server/ldap_bind.c
blob: f37ef31c0adb15057e77e75bfca0a449072e5384 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
/* 
   Unix SMB/CIFS implementation.
   LDAP server
   Copyright (C) Stefan Metzmacher 2004
   
   This program is free software; you can redistribute it and/or modify
   it under the terms of the GNU General Public License as published by
   the Free Software Foundation; either version 3 of the License, or
   (at your option) any later version.
   
   This program is distributed in the hope that it will be useful,
   but WITHOUT ANY WARRANTY; without even the implied warranty of
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
   GNU General Public License for more details.
   
   You should have received a copy of the GNU General Public License
   along with this program.  If not, see <http://www.gnu.org/licenses/>.
*/

#include "includes.h"
#include "ldap_server/ldap_server.h"
#include "auth/auth.h"
#include "libcli/ldap/ldap.h"
#include "smbd/service.h"
#include "lib/ldb/include/ldb.h"
#include "lib/ldb/include/ldb_errors.h"
#include "dsdb/samdb/samdb.h"
#include "auth/gensec/gensec.h"
#include "param/param.h"

static NTSTATUS ldapsrv_BindSimple(struct ldapsrv_call *call)
{
	struct ldap_BindRequest *req = &call->request->r.BindRequest;
	struct ldapsrv_reply *reply;
	struct ldap_BindResponse *resp;

	int result;
	const char *errstr;
	const char *nt4_domain, *nt4_account;

	struct auth_session_info *session_info;

	NTSTATUS status;

	DEBUG(10, ("BindSimple dn: %s\n",req->dn));

	status = crack_auto_name_to_nt4_name(call, call->conn->connection->event.ctx, call->conn->lp_ctx, req->dn, &nt4_domain, &nt4_account);
	if (NT_STATUS_IS_OK(status)) {
		status = authenticate_username_pw(call,
						  call->conn->connection->event.ctx,
						  call->conn->connection->msg_ctx,
						  call->conn->lp_ctx,
						  nt4_domain, nt4_account, 
						  req->creds.password,
						  &session_info);
	}

	reply = ldapsrv_init_reply(call, LDAP_TAG_BindResponse);
	if (!reply) {
		return NT_STATUS_NO_MEMORY;
	}

	if (NT_STATUS_IS_OK(status)) {
		result = LDAP_SUCCESS;
		errstr = NULL;

		talloc_free(call->conn->session_info);
		call->conn->session_info = session_info;
		talloc_steal(call->conn, session_info);

		/* don't leak the old LDB */
		talloc_free(call->conn->ldb);

		status = ldapsrv_backend_Init(call->conn);		
		
		if (!NT_STATUS_IS_OK(status)) {
			result = LDAP_OPERATIONS_ERROR;
			errstr = talloc_asprintf(reply, "Simple Bind: Failed to advise ldb new credentials: %s", nt_errstr(status));
		}
	} else {
		status = auth_nt_status_squash(status);

		result = LDAP_INVALID_CREDENTIALS;
		errstr = talloc_asprintf(reply, "Simple Bind Failed: %s", nt_errstr(status));
	}

	resp = &reply->msg->r.BindResponse;
	resp->response.resultcode = result;
	resp->response.errormessage = errstr;
	resp->response.dn = NULL;
	resp->response.referral = NULL;
	resp->SASL.secblob = NULL;

	ldapsrv_queue_reply(call, reply);
	return NT_STATUS_OK;
}

struct ldapsrv_sasl_context {
	struct ldapsrv_connection *conn;
	struct socket_context *sasl_socket;
};

static void ldapsrv_set_sasl(void *private) 
{
	struct ldapsrv_sasl_context *ctx = talloc_get_type(private, struct ldapsrv_sasl_context);
	talloc_steal(ctx->conn->connection, ctx->sasl_socket);
	talloc_unlink(ctx->conn->connection, ctx->conn->connection->socket);

	ctx->conn->sockets.sasl = ctx->sasl_socket;
	ctx->conn->connection->socket = ctx->sasl_socket;
	packet_set_socket(ctx->conn->packet, ctx->conn->connection->socket);
}

static NTSTATUS ldapsrv_BindSASL(struct ldapsrv_call *call)
{
	struct ldap_BindRequest *req = &call->request->r.BindRequest;
	struct ldapsrv_reply *reply;
	struct ldap_BindResponse *resp;
	struct ldapsrv_connection *conn;
	int result = 0;
	const char *errstr=NULL;
	NTSTATUS status = NT_STATUS_OK;

	DEBUG(10, ("BindSASL dn: %s\n",req->dn));

	reply = ldapsrv_init_reply(call, LDAP_TAG_BindResponse);
	if (!reply) {
		return NT_STATUS_NO_MEMORY;
	}
	resp = &reply->msg->r.BindResponse;
	
	conn = call->conn;

	/* 
	 * TODO: a SASL bind with a different mechanism
	 *       should cancel an inprogress SASL bind.
	 *       (see RFC 4513)
	 */

	if (!conn->gensec) {
		conn->session_info = NULL;

		status = gensec_server_start(conn,
					     conn->connection->event.ctx,
					     conn->lp_ctx,
					     conn->connection->msg_ctx,
					     &conn->gensec);
		if (!NT_STATUS_IS_OK(status)) {
			DEBUG(1, ("Failed to start GENSEC server code: %s\n", nt_errstr(status)));
			result = LDAP_OPERATIONS_ERROR;
			errstr = talloc_asprintf(reply, "SASL: Failed to start authentication system: %s", 
						 nt_errstr(status));
		} else {
		
			gensec_set_target_service(conn->gensec, "ldap");
			
			gensec_set_credentials(conn->gensec, conn->server_credentials);
			
			gensec_want_feature(conn->gensec, GENSEC_FEATURE_SIGN);
			gensec_want_feature(conn->gensec, GENSEC_FEATURE_SEAL);
			gensec_want_feature(conn->gensec, GENSEC_FEATURE_ASYNC_REPLIES);
			
			status = gensec_start_mech_by_sasl_name(conn->gensec, req->creds.SASL.mechanism);
			
			if (!NT_STATUS_IS_OK(status)) {
				DEBUG(1, ("Failed to start GENSEC SASL[%s] server code: %s\n", 
					  req->creds.SASL.mechanism, nt_errstr(status)));
				result = LDAP_OPERATIONS_ERROR;
				errstr = talloc_asprintf(reply, "SASL:[%s]: Failed to start authentication backend: %s", 
							 req->creds.SASL.mechanism, nt_errstr(status));
			}
		}
	}

	if (NT_STATUS_IS_OK(status)) {
		DATA_BLOB input = data_blob(NULL, 0);
		DATA_BLOB output = data_blob(NULL, 0);

		if (req->creds.SASL.secblob) {
			input = *req->creds.SASL.secblob;
		}

		resp->SASL.secblob = talloc(reply, DATA_BLOB);
		NT_STATUS_HAVE_NO_MEMORY(resp->SASL.secblob);

		status = gensec_update(conn->gensec, reply,
				       input, &output);

		/* Windows 2000 mmc doesn't like secblob == NULL and reports a decoding error */
		resp->SASL.secblob = talloc(reply, DATA_BLOB);
		NT_STATUS_HAVE_NO_MEMORY(resp->SASL.secblob);
		*resp->SASL.secblob = output;
	} else {
		resp->SASL.secblob = NULL;
	}

	if (NT_STATUS_EQUAL(NT_STATUS_MORE_PROCESSING_REQUIRED, status)) {
		result = LDAP_SASL_BIND_IN_PROGRESS;
		errstr = NULL;
	} else if (NT_STATUS_IS_OK(status)) {
		struct auth_session_info *old_session_info=NULL;
		struct ldapsrv_sasl_context *ctx;

		result = LDAP_SUCCESS;
		errstr = NULL;

		ctx = talloc(call, struct ldapsrv_sasl_context); 

		if (!ctx) {
			status = NT_STATUS_NO_MEMORY;
		} else {
			ctx->conn = conn;
			status = gensec_socket_init(conn->gensec, 
						    conn->connection->socket,
						    conn->connection->event.ctx, 
						    stream_io_handler_callback,
						    conn->connection,
						    &ctx->sasl_socket);
		} 

		if (!ctx || !NT_STATUS_IS_OK(status)) {
			conn->session_info = old_session_info;
			result = LDAP_OPERATIONS_ERROR;
			errstr = talloc_asprintf(reply, 
						 "SASL:[%s]: Failed to setup SASL socket: %s", 
						 req->creds.SASL.mechanism, nt_errstr(status));
		} else {

			call->send_callback = ldapsrv_set_sasl;
			call->send_private = ctx;
		
			old_session_info = conn->session_info;
			conn->session_info = NULL;
			status = gensec_session_info(conn->gensec, &conn->session_info);
			if (!NT_STATUS_IS_OK(status)) {
				conn->session_info = old_session_info;
				result = LDAP_OPERATIONS_ERROR;
				errstr = talloc_asprintf(reply, 
							 "SASL:[%s]: Failed to get session info: %s", 
							 req->creds.SASL.mechanism, nt_errstr(status));
			} else {
				talloc_free(old_session_info);
				talloc_steal(conn, conn->session_info);
				
				/* don't leak the old LDB */
				talloc_free(conn->ldb);
				
				status = ldapsrv_backend_Init(conn);		
				
				if (!NT_STATUS_IS_OK(status)) {
					result = LDAP_OPERATIONS_ERROR;
					errstr = talloc_asprintf(reply, 
								 "SASL:[%s]: Failed to advise samdb of new credentials: %s", 
								 req->creds.SASL.mechanism, 
								 nt_errstr(status));
				}
			}
		}
	} else {
		status = auth_nt_status_squash(status);
		if (result == 0) {
			result = LDAP_INVALID_CREDENTIALS;
			errstr = talloc_asprintf(reply, "SASL:[%s]: %s", req->creds.SASL.mechanism, nt_errstr(status));
		}
	}

	resp->response.resultcode = result;
	resp->response.dn = NULL;
	resp->response.errormessage = errstr;
	resp->response.referral = NULL;

	ldapsrv_queue_reply(call, reply);
	return NT_STATUS_OK;
}

NTSTATUS ldapsrv_BindRequest(struct ldapsrv_call *call)
{
	struct ldap_BindRequest *req = &call->request->r.BindRequest;
	struct ldapsrv_reply *reply;
	struct ldap_BindResponse *resp;

	/* 
	 * TODO: we should fail the bind request
	 *       if there're any pending requests.
	 *
	 *       also a simple bind should cancel an
	 *       inprogress SASL bind.
	 *       (see RFC 4513)
	 */
	switch (req->mechanism) {
		case LDAP_AUTH_MECH_SIMPLE:
			return ldapsrv_BindSimple(call);
		case LDAP_AUTH_MECH_SASL:
			return ldapsrv_BindSASL(call);
	}

	reply = ldapsrv_init_reply(call, LDAP_TAG_BindResponse);
	if (!reply) {
		return NT_STATUS_NO_MEMORY;
	}

	resp = &reply->msg->r.BindResponse;
	resp->response.resultcode = 7;
	resp->response.dn = NULL;
	resp->response.errormessage = talloc_asprintf(reply, "Bad AuthenticationChoice [%d]", req->mechanism);
	resp->response.referral = NULL;
	resp->SASL.secblob = NULL;

	ldapsrv_queue_reply(call, reply);
	return NT_STATUS_OK;
}

NTSTATUS ldapsrv_UnbindRequest(struct ldapsrv_call *call)
{
	DEBUG(10, ("UnbindRequest\n"));
	return NT_STATUS_OK;
}
1363'>1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653
/*
 * This file is subject to the terms and conditions of the GNU General Public
 * License.  See the file "COPYING" in the main directory of this archive
 * for more details.
 *
 * Driver for SGI's IOC3 based Ethernet cards as found in the PCI card.
 *
 * Copyright (C) 1999, 2000, 2001, 2003 Ralf Baechle
 * Copyright (C) 1995, 1999, 2000, 2001 by Silicon Graphics, Inc.
 *
 * References:
 *  o IOC3 ASIC specification 4.51, 1996-04-18
 *  o IEEE 802.3 specification, 2000 edition
 *  o DP38840A Specification, National Semiconductor, March 1997
 *
 * To do:
 *
 *  o Handle allocation failures in ioc3_alloc_skb() more gracefully.
 *  o Handle allocation failures in ioc3_init_rings().
 *  o Use prefetching for large packets.  What is a good lower limit for
 *    prefetching?
 *  o We're probably allocating a bit too much memory.
 *  o Use hardware checksums.
 *  o Convert to using a IOC3 meta driver.
 *  o Which PHYs might possibly be attached to the IOC3 in real live,
 *    which workarounds are required for them?  Do we ever have Lucent's?
 *  o For the 2.5 branch kill the mii-tool ioctls.
 */

#define IOC3_NAME	"ioc3-eth"
#define IOC3_VERSION	"2.6.3-4"

#include <linux/init.h>
#include <linux/delay.h>
#include <linux/kernel.h>
#include <linux/mm.h>
#include <linux/errno.h>
#include <linux/module.h>
#include <linux/pci.h>
#include <linux/crc32.h>
#include <linux/mii.h>
#include <linux/in.h>
#include <linux/ip.h>
#include <linux/tcp.h>
#include <linux/udp.h>
#include <linux/dma-mapping.h>

#ifdef CONFIG_SERIAL_8250
#include <linux/serial_core.h>
#include <linux/serial_8250.h>
#endif

#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/ethtool.h>
#include <linux/skbuff.h>
#include <net/ip.h>

#include <asm/byteorder.h>
#include <asm/checksum.h>
#include <asm/io.h>
#include <asm/pgtable.h>
#include <asm/uaccess.h>
#include <asm/sn/types.h>
#include <asm/sn/sn0/addrs.h>
#include <asm/sn/sn0/hubni.h>
#include <asm/sn/sn0/hubio.h>
#include <asm/sn/klconfig.h>
#include <asm/sn/ioc3.h>
#include <asm/sn/sn0/ip27.h>
#include <asm/pci/bridge.h>

/*
 * 64 RX buffers.  This is tunable in the range of 16 <= x < 512.  The
 * value must be a power of two.
 */
#define RX_BUFFS 64

#define ETCSR_FD	((17<<ETCSR_IPGR2_SHIFT) | (11<<ETCSR_IPGR1_SHIFT) | 21)
#define ETCSR_HD	((21<<ETCSR_IPGR2_SHIFT) | (21<<ETCSR_IPGR1_SHIFT) | 21)

/* Private per NIC data of the driver.  */
struct ioc3_private {
	struct ioc3 *regs;
	unsigned long *rxr;		/* pointer to receiver ring */
	struct ioc3_etxd *txr;
	struct sk_buff *rx_skbs[512];
	struct sk_buff *tx_skbs[128];
	struct net_device_stats stats;
	int rx_ci;			/* RX consumer index */
	int rx_pi;			/* RX producer index */
	int tx_ci;			/* TX consumer index */
	int tx_pi;			/* TX producer index */
	int txqlen;
	u32 emcr, ehar_h, ehar_l;
	spinlock_t ioc3_lock;
	struct mii_if_info mii;
	struct pci_dev *pdev;

	/* Members used by autonegotiation  */
	struct timer_list ioc3_timer;
};

static inline struct net_device *priv_netdev(struct ioc3_private *dev)
{
	return (void *)dev - ((sizeof(struct net_device) + 31) & ~31);
}

static int ioc3_ioctl(struct net_device *dev, struct ifreq *rq, int cmd);
static void ioc3_set_multicast_list(struct net_device *dev);
static int ioc3_start_xmit(struct sk_buff *skb, struct net_device *dev);
static void ioc3_timeout(struct net_device *dev);
static inline unsigned int ioc3_hash(const unsigned char *addr);
static inline void ioc3_stop(struct ioc3_private *ip);
static void ioc3_init(struct net_device *dev);

static const char ioc3_str[] = "IOC3 Ethernet";
static const struct ethtool_ops ioc3_ethtool_ops;

/* We use this to acquire receive skb's that we can DMA directly into. */

#define IOC3_CACHELINE	128UL

static inline unsigned long aligned_rx_skb_addr(unsigned long addr)
{
	return (~addr + 1) & (IOC3_CACHELINE - 1UL);
}

static inline struct sk_buff * ioc3_alloc_skb(unsigned long length,
	unsigned int gfp_mask)
{
	struct sk_buff *skb;

	skb = alloc_skb(length + IOC3_CACHELINE - 1, gfp_mask);
	if (likely(skb)) {
		int offset = aligned_rx_skb_addr((unsigned long) skb->data);
		if (offset)
			skb_reserve(skb, offset);
	}

	return skb;
}

static inline unsigned long ioc3_map(void *ptr, unsigned long vdev)
{
#ifdef CONFIG_SGI_IP27
	vdev <<= 57;   /* Shift to PCI64_ATTR_VIRTUAL */

	return vdev | (0xaUL << PCI64_ATTR_TARG_SHFT) | PCI64_ATTR_PREF |
	       ((unsigned long)ptr & TO_PHYS_MASK);
#else
	return virt_to_bus(ptr);
#endif
}

/* BEWARE: The IOC3 documentation documents the size of rx buffers as
   1644 while it's actually 1664.  This one was nasty to track down ...  */
#define RX_OFFSET		10
#define RX_BUF_ALLOC_SIZE	(1664 + RX_OFFSET + IOC3_CACHELINE)

/* DMA barrier to separate cached and uncached accesses.  */
#define BARRIER()							\
	__asm__("sync" ::: "memory")


#define IOC3_SIZE 0x100000

/*
 * IOC3 is a big endian device
 *
 * Unorthodox but makes the users of these macros more readable - the pointer
 * to the IOC3's memory mapped registers is expected as struct ioc3 * ioc3
 * in the environment.
 */
#define ioc3_r_mcr()		be32_to_cpu(ioc3->mcr)
#define ioc3_w_mcr(v)		do { ioc3->mcr = cpu_to_be32(v); } while (0)
#define ioc3_w_gpcr_s(v)	do { ioc3->gpcr_s = cpu_to_be32(v); } while (0)
#define ioc3_r_emcr()		be32_to_cpu(ioc3->emcr)
#define ioc3_w_emcr(v)		do { ioc3->emcr = cpu_to_be32(v); } while (0)
#define ioc3_r_eisr()		be32_to_cpu(ioc3->eisr)
#define ioc3_w_eisr(v)		do { ioc3->eisr = cpu_to_be32(v); } while (0)
#define ioc3_r_eier()		be32_to_cpu(ioc3->eier)
#define ioc3_w_eier(v)		do { ioc3->eier = cpu_to_be32(v); } while (0)
#define ioc3_r_ercsr()		be32_to_cpu(ioc3->ercsr)
#define ioc3_w_ercsr(v)		do { ioc3->ercsr = cpu_to_be32(v); } while (0)
#define ioc3_r_erbr_h()		be32_to_cpu(ioc3->erbr_h)
#define ioc3_w_erbr_h(v)	do { ioc3->erbr_h = cpu_to_be32(v); } while (0)
#define ioc3_r_erbr_l()		be32_to_cpu(ioc3->erbr_l)
#define ioc3_w_erbr_l(v)	do { ioc3->erbr_l = cpu_to_be32(v); } while (0)
#define ioc3_r_erbar()		be32_to_cpu(ioc3->erbar)
#define ioc3_w_erbar(v)		do { ioc3->erbar = cpu_to_be32(v); } while (0)
#define ioc3_r_ercir()		be32_to_cpu(ioc3->ercir)
#define ioc3_w_ercir(v)		do { ioc3->ercir = cpu_to_be32(v); } while (0)
#define ioc3_r_erpir()		be32_to_cpu(ioc3->erpir)
#define ioc3_w_erpir(v)		do { ioc3->erpir = cpu_to_be32(v); } while (0)
#define ioc3_r_ertr()		be32_to_cpu(ioc3->ertr)
#define ioc3_w_ertr(v)		do { ioc3->ertr = cpu_to_be32(v); } while (0)
#define ioc3_r_etcsr()		be32_to_cpu(ioc3->etcsr)
#define ioc3_w_etcsr(v)		do { ioc3->etcsr = cpu_to_be32(v); } while (0)
#define ioc3_r_ersr()		be32_to_cpu(ioc3->ersr)
#define ioc3_w_ersr(v)		do { ioc3->ersr = cpu_to_be32(v); } while (0)
#define ioc3_r_etcdc()		be32_to_cpu(ioc3->etcdc)
#define ioc3_w_etcdc(v)		do { ioc3->etcdc = cpu_to_be32(v); } while (0)
#define ioc3_r_ebir()		be32_to_cpu(ioc3->ebir)
#define ioc3_w_ebir(v)		do { ioc3->ebir = cpu_to_be32(v); } while (0)
#define ioc3_r_etbr_h()		be32_to_cpu(ioc3->etbr_h)
#define ioc3_w_etbr_h(v)	do { ioc3->etbr_h = cpu_to_be32(v); } while (0)
#define ioc3_r_etbr_l()		be32_to_cpu(ioc3->etbr_l)
#define ioc3_w_etbr_l(v)	do { ioc3->etbr_l = cpu_to_be32(v); } while (0)
#define ioc3_r_etcir()		be32_to_cpu(ioc3->etcir)
#define ioc3_w_etcir(v)		do { ioc3->etcir = cpu_to_be32(v); } while (0)
#define ioc3_r_etpir()		be32_to_cpu(ioc3->etpir)
#define ioc3_w_etpir(v)		do { ioc3->etpir = cpu_to_be32(v); } while (0)
#define ioc3_r_emar_h()		be32_to_cpu(ioc3->emar_h)
#define ioc3_w_emar_h(v)	do { ioc3->emar_h = cpu_to_be32(v); } while (0)
#define ioc3_r_emar_l()		be32_to_cpu(ioc3->emar_l)
#define ioc3_w_emar_l(v)	do { ioc3->emar_l = cpu_to_be32(v); } while (0)
#define ioc3_r_ehar_h()		be32_to_cpu(ioc3->ehar_h)
#define ioc3_w_ehar_h(v)	do { ioc3->ehar_h = cpu_to_be32(v); } while (0)
#define ioc3_r_ehar_l()		be32_to_cpu(ioc3->ehar_l)
#define ioc3_w_ehar_l(v)	do { ioc3->ehar_l = cpu_to_be32(v); } while (0)
#define ioc3_r_micr()		be32_to_cpu(ioc3->micr)
#define ioc3_w_micr(v)		do { ioc3->micr = cpu_to_be32(v); } while (0)
#define ioc3_r_midr_r()		be32_to_cpu(ioc3->midr_r)
#define ioc3_w_midr_r(v)	do { ioc3->midr_r = cpu_to_be32(v); } while (0)
#define ioc3_r_midr_w()		be32_to_cpu(ioc3->midr_w)
#define ioc3_w_midr_w(v)	do { ioc3->midr_w = cpu_to_be32(v); } while (0)

static inline u32 mcr_pack(u32 pulse, u32 sample)
{
	return (pulse << 10) | (sample << 2);
}

static int nic_wait(struct ioc3 *ioc3)
{
	u32 mcr;

        do {
                mcr = ioc3_r_mcr();
        } while (!(mcr & 2));

        return mcr & 1;
}

static int nic_reset(struct ioc3 *ioc3)
{
        int presence;

	ioc3_w_mcr(mcr_pack(500, 65));
	presence = nic_wait(ioc3);

	ioc3_w_mcr(mcr_pack(0, 500));
	nic_wait(ioc3);

        return presence;
}

static inline int nic_read_bit(struct ioc3 *ioc3)
{
	int result;

	ioc3_w_mcr(mcr_pack(6, 13));
	result = nic_wait(ioc3);
	ioc3_w_mcr(mcr_pack(0, 100));
	nic_wait(ioc3);

	return result;
}

static inline void nic_write_bit(struct ioc3 *ioc3, int bit)
{
	if (bit)
		ioc3_w_mcr(mcr_pack(6, 110));
	else
		ioc3_w_mcr(mcr_pack(80, 30));

	nic_wait(ioc3);
}

/*
 * Read a byte from an iButton device
 */
static u32 nic_read_byte(struct ioc3 *ioc3)
{
	u32 result = 0;
	int i;

	for (i = 0; i < 8; i++)
		result = (result >> 1) | (nic_read_bit(ioc3) << 7);

	return result;
}

/*
 * Write a byte to an iButton device
 */
static void nic_write_byte(struct ioc3 *ioc3, int byte)
{
	int i, bit;

	for (i = 8; i; i--) {
		bit = byte & 1;
		byte >>= 1;

		nic_write_bit(ioc3, bit);
	}
}

static u64 nic_find(struct ioc3 *ioc3, int *last)
{
	int a, b, index, disc;
	u64 address = 0;

	nic_reset(ioc3);
	/* Search ROM.  */
	nic_write_byte(ioc3, 0xf0);

	/* Algorithm from ``Book of iButton Standards''.  */
	for (index = 0, disc = 0; index < 64; index++) {
		a = nic_read_bit(ioc3);
		b = nic_read_bit(ioc3);

		if (a && b) {
			printk("NIC search failed (not fatal).\n");
			*last = 0;
			return 0;
		}

		if (!a && !b) {
			if (index == *last) {
				address |= 1UL << index;
			} else if (index > *last) {
				address &= ~(1UL << index);
				disc = index;
			} else if ((address & (1UL << index)) == 0)
				disc = index;
			nic_write_bit(ioc3, address & (1UL << index));
			continue;
		} else {
			if (a)
				address |= 1UL << index;
			else
				address &= ~(1UL << index);
			nic_write_bit(ioc3, a);
			continue;
		}
	}

	*last = disc;

	return address;
}

static int nic_init(struct ioc3 *ioc3)
{
	const char *type;
	u8 crc;
	u8 serial[6];
	int save = 0, i;

	type = "unknown";

	while (1) {
		u64 reg;
		reg = nic_find(ioc3, &save);

		switch (reg & 0xff) {
		case 0x91:
			type = "DS1981U";
			break;
		default:
			if (save == 0) {
				/* Let the caller try again.  */
				return -1;
			}
			continue;
		}

		nic_reset(ioc3);

		/* Match ROM.  */
		nic_write_byte(ioc3, 0x55);
		for (i = 0; i < 8; i++)
			nic_write_byte(ioc3, (reg >> (i << 3)) & 0xff);

		reg >>= 8; /* Shift out type.  */
		for (i = 0; i < 6; i++) {
			serial[i] = reg & 0xff;
			reg >>= 8;
		}
		crc = reg & 0xff;
		break;
	}

	printk("Found %s NIC", type);
	if (type != "unknown") {
		printk (" registration number %02x:%02x:%02x:%02x:%02x:%02x,"
			" CRC %02x", serial[0], serial[1], serial[2],
			serial[3], serial[4], serial[5], crc);
	}
	printk(".\n");

	return 0;
}

/*
 * Read the NIC (Number-In-a-Can) device used to store the MAC address on
 * SN0 / SN00 nodeboards and PCI cards.
 */
static void ioc3_get_eaddr_nic(struct ioc3_private *ip)
{
	struct ioc3 *ioc3 = ip->regs;
	u8 nic[14];
	int tries = 2; /* There may be some problem with the battery?  */
	int i;

	ioc3_w_gpcr_s(1 << 21);

	while (tries--) {
		if (!nic_init(ioc3))
			break;
		udelay(500);
	}

	if (tries < 0) {
		printk("Failed to read MAC address\n");
		return;
	}

	/* Read Memory.  */
	nic_write_byte(ioc3, 0xf0);
	nic_write_byte(ioc3, 0x00);
	nic_write_byte(ioc3, 0x00);

	for (i = 13; i >= 0; i--)
		nic[i] = nic_read_byte(ioc3);

	for (i = 2; i < 8; i++)
		priv_netdev(ip)->dev_addr[i - 2] = nic[i];
}

/*
 * Ok, this is hosed by design.  It's necessary to know what machine the
 * NIC is in in order to know how to read the NIC address.  We also have
 * to know if it's a PCI card or a NIC in on the node board ...
 */
static void ioc3_get_eaddr(struct ioc3_private *ip)
{
	int i;


	ioc3_get_eaddr_nic(ip);

	printk("Ethernet address is ");
	for (i = 0; i < 6; i++) {
		printk("%02x", priv_netdev(ip)->dev_addr[i]);
		if (i < 5)
			printk(":");
	}
	printk(".\n");
}

static void __ioc3_set_mac_address(struct net_device *dev)
{
	struct ioc3_private *ip = netdev_priv(dev);
	struct ioc3 *ioc3 = ip->regs;

	ioc3_w_emar_h((dev->dev_addr[5] <<  8) | dev->dev_addr[4]);
	ioc3_w_emar_l((dev->dev_addr[3] << 24) | (dev->dev_addr[2] << 16) |
	              (dev->dev_addr[1] <<  8) | dev->dev_addr[0]);
}

static int ioc3_set_mac_address(struct net_device *dev, void *addr)
{
	struct ioc3_private *ip = netdev_priv(dev);
	struct sockaddr *sa = addr;

	memcpy(dev->dev_addr, sa->sa_data, dev->addr_len);

	spin_lock_irq(&ip->ioc3_lock);
	__ioc3_set_mac_address(dev);
	spin_unlock_irq(&ip->ioc3_lock);

	return 0;
}

/*
 * Caller must hold the ioc3_lock ever for MII readers.  This is also
 * used to protect the transmitter side but it's low contention.
 */
static int ioc3_mdio_read(struct net_device *dev, int phy, int reg)
{
	struct ioc3_private *ip = netdev_priv(dev);
	struct ioc3 *ioc3 = ip->regs;

	while (ioc3_r_micr() & MICR_BUSY);
	ioc3_w_micr((phy << MICR_PHYADDR_SHIFT) | reg | MICR_READTRIG);
	while (ioc3_r_micr() & MICR_BUSY);

	return ioc3_r_midr_r() & MIDR_DATA_MASK;
}

static void ioc3_mdio_write(struct net_device *dev, int phy, int reg, int data)
{
	struct ioc3_private *ip = netdev_priv(dev);
	struct ioc3 *ioc3 = ip->regs;

	while (ioc3_r_micr() & MICR_BUSY);
	ioc3_w_midr_w(data);
	ioc3_w_micr((phy << MICR_PHYADDR_SHIFT) | reg);
	while (ioc3_r_micr() & MICR_BUSY);
}

static int ioc3_mii_init(struct ioc3_private *ip);

static struct net_device_stats *ioc3_get_stats(struct net_device *dev)
{
	struct ioc3_private *ip = netdev_priv(dev);
	struct ioc3 *ioc3 = ip->regs;

	ip->stats.collisions += (ioc3_r_etcdc() & ETCDC_COLLCNT_MASK);
	return &ip->stats;
}

#ifdef CONFIG_SGI_IOC3_ETH_HW_RX_CSUM

static void ioc3_tcpudp_checksum(struct sk_buff *skb, uint32_t hwsum, int len)
{
	struct ethhdr *eh = eth_hdr(skb);
	uint32_t csum, ehsum;
	unsigned int proto;
	struct iphdr *ih;
	uint16_t *ew;
	unsigned char *cp;

	/*
	 * Did hardware handle the checksum at all?  The cases we can handle
	 * are:
	 *
	 * - TCP and UDP checksums of IPv4 only.
	 * - IPv6 would be doable but we keep that for later ...
	 * - Only unfragmented packets.  Did somebody already tell you
	 *   fragmentation is evil?
	 * - don't care about packet size.  Worst case when processing a
	 *   malformed packet we'll try to access the packet at ip header +
	 *   64 bytes which is still inside the skb.  Even in the unlikely
	 *   case where the checksum is right the higher layers will still
	 *   drop the packet as appropriate.
	 */
	if (eh->h_proto != ntohs(ETH_P_IP))
		return;

	ih = (struct iphdr *) ((char *)eh + ETH_HLEN);
	if (ih->frag_off & htons(IP_MF | IP_OFFSET))
		return;

	proto = ih->protocol;
	if (proto != IPPROTO_TCP && proto != IPPROTO_UDP)
		return;

	/* Same as tx - compute csum of pseudo header  */
	csum = hwsum +
	       (ih->tot_len - (ih->ihl << 2)) +
	       htons((uint16_t)ih->protocol) +
	       (ih->saddr >> 16) + (ih->saddr & 0xffff) +
	       (ih->daddr >> 16) + (ih->daddr & 0xffff);

	/* Sum up ethernet dest addr, src addr and protocol  */
	ew = (uint16_t *) eh;
	ehsum = ew[0] + ew[1] + ew[2] + ew[3] + ew[4] + ew[5] + ew[6];

	ehsum = (ehsum & 0xffff) + (ehsum >> 16);
	ehsum = (ehsum & 0xffff) + (ehsum >> 16);

	csum += 0xffff ^ ehsum;

	/* In the next step we also subtract the 1's complement
	   checksum of the trailing ethernet CRC.  */
	cp = (char *)eh + len;	/* points at trailing CRC */
	if (len & 1) {
		csum += 0xffff ^ (uint16_t) ((cp[1] << 8) | cp[0]);
		csum += 0xffff ^ (uint16_t) ((cp[3] << 8) | cp[2]);
	} else {
		csum += 0xffff ^ (uint16_t) ((cp[0] << 8) | cp[1]);
		csum += 0xffff ^ (uint16_t) ((cp[2] << 8) | cp[3]);
	}

	csum = (csum & 0xffff) + (csum >> 16);
	csum = (csum & 0xffff) + (csum >> 16);

	if (csum == 0xffff)
		skb->ip_summed = CHECKSUM_UNNECESSARY;
}
#endif /* CONFIG_SGI_IOC3_ETH_HW_RX_CSUM */

static inline void ioc3_rx(struct ioc3_private *ip)
{
	struct sk_buff *skb, *new_skb;
	struct ioc3 *ioc3 = ip->regs;
	int rx_entry, n_entry, len;
	struct ioc3_erxbuf *rxb;
	unsigned long *rxr;
	u32 w0, err;

	rxr = (unsigned long *) ip->rxr;		/* Ring base */
	rx_entry = ip->rx_ci;				/* RX consume index */
	n_entry = ip->rx_pi;

	skb = ip->rx_skbs[rx_entry];
	rxb = (struct ioc3_erxbuf *) (skb->data - RX_OFFSET);
	w0 = be32_to_cpu(rxb->w0);

	while (w0 & ERXBUF_V) {
		err = be32_to_cpu(rxb->err);		/* It's valid ...  */
		if (err & ERXBUF_GOODPKT) {
			len = ((w0 >> ERXBUF_BYTECNT_SHIFT) & 0x7ff) - 4;
			skb_trim(skb, len);
			skb->protocol = eth_type_trans(skb, priv_netdev(ip));

			new_skb = ioc3_alloc_skb(RX_BUF_ALLOC_SIZE, GFP_ATOMIC);
			if (!new_skb) {
				/* Ouch, drop packet and just recycle packet
				   to keep the ring filled.  */
				ip->stats.rx_dropped++;
				new_skb = skb;
				goto next;
			}

#ifdef CONFIG_SGI_IOC3_ETH_HW_RX_CSUM
			ioc3_tcpudp_checksum(skb, w0 & ERXBUF_IPCKSUM_MASK,len);
#endif

			netif_rx(skb);

			ip->rx_skbs[rx_entry] = NULL;	/* Poison  */

			new_skb->dev = priv_netdev(ip);

			/* Because we reserve afterwards. */
			skb_put(new_skb, (1664 + RX_OFFSET));
			rxb = (struct ioc3_erxbuf *) new_skb->data;
			skb_reserve(new_skb, RX_OFFSET);

			priv_netdev(ip)->last_rx = jiffies;
			ip->stats.rx_packets++;		/* Statistics */
			ip->stats.rx_bytes += len;
		} else {
 			/* The frame is invalid and the skb never
                           reached the network layer so we can just
                           recycle it.  */
 			new_skb = skb;
 			ip->stats.rx_errors++;
		}
		if (err & ERXBUF_CRCERR)	/* Statistics */
			ip->stats.rx_crc_errors++;
		if (err & ERXBUF_FRAMERR)
			ip->stats.rx_frame_errors++;
next:
		ip->rx_skbs[n_entry] = new_skb;
		rxr[n_entry] = cpu_to_be64(ioc3_map(rxb, 1));
		rxb->w0 = 0;				/* Clear valid flag */
		n_entry = (n_entry + 1) & 511;		/* Update erpir */

		/* Now go on to the next ring entry.  */
		rx_entry = (rx_entry + 1) & 511;
		skb = ip->rx_skbs[rx_entry];
		rxb = (struct ioc3_erxbuf *) (skb->data - RX_OFFSET);
		w0 = be32_to_cpu(rxb->w0);
	}
	ioc3_w_erpir((n_entry << 3) | ERPIR_ARM);
	ip->rx_pi = n_entry;
	ip->rx_ci = rx_entry;
}

static inline void ioc3_tx(struct ioc3_private *ip)
{
	unsigned long packets, bytes;
	struct ioc3 *ioc3 = ip->regs;
	int tx_entry, o_entry;
	struct sk_buff *skb;
	u32 etcir;

	spin_lock(&ip->ioc3_lock);
	etcir = ioc3_r_etcir();

	tx_entry = (etcir >> 7) & 127;
	o_entry = ip->tx_ci;
	packets = 0;
	bytes = 0;

	while (o_entry != tx_entry) {
		packets++;
		skb = ip->tx_skbs[o_entry];
		bytes += skb->len;
		dev_kfree_skb_irq(skb);
		ip->tx_skbs[o_entry] = NULL;

		o_entry = (o_entry + 1) & 127;		/* Next */

		etcir = ioc3_r_etcir();			/* More pkts sent?  */
		tx_entry = (etcir >> 7) & 127;
	}

	ip->stats.tx_packets += packets;
	ip->stats.tx_bytes += bytes;
	ip->txqlen -= packets;

	if (ip->txqlen < 128)
		netif_wake_queue(priv_netdev(ip));

	ip->tx_ci = o_entry;
	spin_unlock(&ip->ioc3_lock);
}

/*
 * Deal with fatal IOC3 errors.  This condition might be caused by a hard or
 * software problems, so we should try to recover
 * more gracefully if this ever happens.  In theory we might be flooded
 * with such error interrupts if something really goes wrong, so we might
 * also consider to take the interface down.
 */
static void ioc3_error(struct ioc3_private *ip, u32 eisr)
{
	struct net_device *dev = priv_netdev(ip);
	unsigned char *iface = dev->name;

	spin_lock(&ip->ioc3_lock);

	if (eisr & EISR_RXOFLO)
		printk(KERN_ERR "%s: RX overflow.\n", iface);
	if (eisr & EISR_RXBUFOFLO)
		printk(KERN_ERR "%s: RX buffer overflow.\n", iface);
	if (eisr & EISR_RXMEMERR)
		printk(KERN_ERR "%s: RX PCI error.\n", iface);
	if (eisr & EISR_RXPARERR)
		printk(KERN_ERR "%s: RX SSRAM parity error.\n", iface);
	if (eisr & EISR_TXBUFUFLO)
		printk(KERN_ERR "%s: TX buffer underflow.\n", iface);
	if (eisr & EISR_TXMEMERR)
		printk(KERN_ERR "%s: TX PCI error.\n", iface);

	ioc3_stop(ip);
	ioc3_init(dev);
	ioc3_mii_init(ip);

	netif_wake_queue(dev);

	spin_unlock(&ip->ioc3_lock);
}

/* The interrupt handler does all of the Rx thread work and cleans up
   after the Tx thread.  */
static irqreturn_t ioc3_interrupt(int irq, void *_dev)
{
	struct net_device *dev = (struct net_device *)_dev;
	struct ioc3_private *ip = netdev_priv(dev);
	struct ioc3 *ioc3 = ip->regs;
	const u32 enabled = EISR_RXTIMERINT | EISR_RXOFLO | EISR_RXBUFOFLO |
	                    EISR_RXMEMERR | EISR_RXPARERR | EISR_TXBUFUFLO |
	                    EISR_TXEXPLICIT | EISR_TXMEMERR;
	u32 eisr;

	eisr = ioc3_r_eisr() & enabled;

	ioc3_w_eisr(eisr);
	(void) ioc3_r_eisr();				/* Flush */

	if (eisr & (EISR_RXOFLO | EISR_RXBUFOFLO | EISR_RXMEMERR |
	            EISR_RXPARERR | EISR_TXBUFUFLO | EISR_TXMEMERR))
		ioc3_error(ip, eisr);
	if (eisr & EISR_RXTIMERINT)
		ioc3_rx(ip);
	if (eisr & EISR_TXEXPLICIT)
		ioc3_tx(ip);

	return IRQ_HANDLED;
}

static inline void ioc3_setup_duplex(struct ioc3_private *ip)
{
	struct ioc3 *ioc3 = ip->regs;

	if (ip->mii.full_duplex) {
		ioc3_w_etcsr(ETCSR_FD);
		ip->emcr |= EMCR_DUPLEX;
	} else {
		ioc3_w_etcsr(ETCSR_HD);
		ip->emcr &= ~EMCR_DUPLEX;
	}
	ioc3_w_emcr(ip->emcr);
}

static void ioc3_timer(unsigned long data)
{
	struct ioc3_private *ip = (struct ioc3_private *) data;

	/* Print the link status if it has changed */
	mii_check_media(&ip->mii, 1, 0);
	ioc3_setup_duplex(ip);

	ip->ioc3_timer.expires = jiffies + ((12 * HZ)/10); /* 1.2s */
	add_timer(&ip->ioc3_timer);
}

/*
 * Try to find a PHY.  There is no apparent relation between the MII addresses
 * in the SGI documentation and what we find in reality, so we simply probe
 * for the PHY.  It seems IOC3 PHYs usually live on address 31.  One of my
 * onboard IOC3s has the special oddity that probing doesn't seem to find it
 * yet the interface seems to work fine, so if probing fails we for now will
 * simply default to PHY 31 instead of bailing out.
 */
static int ioc3_mii_init(struct ioc3_private *ip)
{
	struct net_device *dev = priv_netdev(ip);
	int i, found = 0, res = 0;
	int ioc3_phy_workaround = 1;
	u16 word;

	for (i = 0; i < 32; i++) {
		word = ioc3_mdio_read(dev, i, MII_PHYSID1);

		if (word != 0xffff && word != 0x0000) {
			found = 1;
			break;			/* Found a PHY		*/
		}
	}

	if (!found) {
		if (ioc3_phy_workaround)
			i = 31;
		else {
			ip->mii.phy_id = -1;
			res = -ENODEV;
			goto out;
		}
	}

	ip->mii.phy_id = i;
	ip->ioc3_timer.expires = jiffies + (12 * HZ)/10;  /* 1.2 sec. */
	ip->ioc3_timer.data = (unsigned long) ip;
	ip->ioc3_timer.function = &ioc3_timer;
	add_timer(&ip->ioc3_timer);

out:
	return res;
}

static inline void ioc3_clean_rx_ring(struct ioc3_private *ip)
{
	struct sk_buff *skb;
	int i;

	for (i = ip->rx_ci; i & 15; i++) {