aboutsummaryrefslogtreecommitdiff
path: root/hw/xen_nic.c
blob: 2364e823d45af92e8df0126644855d03933a2d4c (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
/*
 *  xen paravirt network card backend
 *
 *  (c) Gerd Hoffmann <kraxel@redhat.com>
 *
 *  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; under version 2 of the License.
 *
 *  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, write to the Free Software Foundation, Inc.,
 *  51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
 */

#include <stdio.h>
#include <stdlib.h>
#include <stdarg.h>
#include <string.h>
#include <unistd.h>
#include <signal.h>
#include <inttypes.h>
#include <fcntl.h>
#include <errno.h>
#include <pthread.h>
#include <sys/socket.h>
#include <sys/ioctl.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <sys/mman.h>
#include <sys/wait.h>
#include <linux/if.h>
#include <linux/if_tun.h>

#include <xs.h>
#include <xenctrl.h>
#include <xen/io/xenbus.h>
#include <xen/io/netif.h>

#include "hw.h"
#include "net.h"
#include "qemu-char.h"
#include "xen_backend.h"

/* ------------------------------------------------------------- */

struct XenNetDev {
    struct XenDevice      xendev;  /* must be first */
    char                  *mac;
    int                   tx_work;
    int                   tx_ring_ref;
    int                   rx_ring_ref;
    struct netif_tx_sring *txs;
    struct netif_rx_sring *rxs;
    netif_tx_back_ring_t  tx_ring;
    netif_rx_back_ring_t  rx_ring;
    VLANClientState       *vs;
};

/* ------------------------------------------------------------- */

static void net_tx_response(struct XenNetDev *netdev, netif_tx_request_t *txp, int8_t st)
{
    RING_IDX i = netdev->tx_ring.rsp_prod_pvt;
    netif_tx_response_t *resp;
    int notify;

    resp = RING_GET_RESPONSE(&netdev->tx_ring, i);
    resp->id     = txp->id;
    resp->status = st;

#if 0
    if (txp->flags & NETTXF_extra_info)
	RING_GET_RESPONSE(&netdev->tx_ring, ++i)->status = NETIF_RSP_NULL;
#endif

    netdev->tx_ring.rsp_prod_pvt = ++i;
    RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&netdev->tx_ring, notify);
    if (notify)
	xen_be_send_notify(&netdev->xendev);

    if (i == netdev->tx_ring.req_cons) {
	int more_to_do;
	RING_FINAL_CHECK_FOR_REQUESTS(&netdev->tx_ring, more_to_do);
	if (more_to_do)
	    netdev->tx_work++;
    }
}

static void net_tx_error(struct XenNetDev *netdev, netif_tx_request_t *txp, RING_IDX end)
{
#if 0
    /*
     * Hmm, why netback fails everything in the ring?
     * Should we do that even when not supporting SG and TSO?
     */
    RING_IDX cons = netdev->tx_ring.req_cons;

    do {
	make_tx_response(netif, txp, NETIF_RSP_ERROR);
	if (cons >= end)
	    break;
	txp = RING_GET_REQUEST(&netdev->tx_ring, cons++);
    } while (1);
    netdev->tx_ring.req_cons = cons;
    netif_schedule_work(netif);
    netif_put(netif);
#else
    net_tx_response(netdev, txp, NETIF_RSP_ERROR);
#endif
}

static void net_tx_packets(struct XenNetDev *netdev)
{
    netif_tx_request_t txreq;
    RING_IDX rc, rp;
    void *page;
    void *tmpbuf = NULL;

    for (;;) {
	rc = netdev->tx_ring.req_cons;
	rp = netdev->tx_ring.sring->req_prod;
	xen_rmb(); /* Ensure we see queued requests up to 'rp'. */

	while ((rc != rp)) {
	    if (RING_REQUEST_CONS_OVERFLOW(&netdev->tx_ring, rc))
		break;
	    memcpy(&txreq, RING_GET_REQUEST(&netdev->tx_ring, rc), sizeof(txreq));
	    netdev->tx_ring.req_cons = ++rc;

#if 1
	    /* should not happen in theory, we don't announce the *
	     * feature-{sg,gso,whatelse} flags in xenstore (yet?) */
	    if (txreq.flags & NETTXF_extra_info) {
		xen_be_printf(&netdev->xendev, 0, "FIXME: extra info flag\n");
		net_tx_error(netdev, &txreq, rc);
		continue;
	    }
	    if (txreq.flags & NETTXF_more_data) {
		xen_be_printf(&netdev->xendev, 0, "FIXME: more data flag\n");
		net_tx_error(netdev, &txreq, rc);
		continue;
	    }
#endif

	    if (txreq.size < 14) {
		xen_be_printf(&netdev->xendev, 0, "bad packet size: %d\n", txreq.size);
		net_tx_error(netdev, &txreq, rc);
		continue;
	    }

	    if ((txreq.offset + txreq.size) > XC_PAGE_SIZE) {
		xen_be_printf(&netdev->xendev, 0, "error: page crossing\n");
		net_tx_error(netdev, &txreq, rc);
		continue;
	    }

	    xen_be_printf(&netdev->xendev, 3, "tx packet ref %d, off %d, len %d, flags 0x%x%s%s%s%s\n",
			  txreq.gref, txreq.offset, txreq.size, txreq.flags,
			  (txreq.flags & NETTXF_csum_blank)     ? " csum_blank"     : "",
			  (txreq.flags & NETTXF_data_validated) ? " data_validated" : "",
			  (txreq.flags & NETTXF_more_data)      ? " more_data"      : "",
			  (txreq.flags & NETTXF_extra_info)     ? " extra_info"     : "");

	    page = xc_gnttab_map_grant_ref(netdev->xendev.gnttabdev,
					   netdev->xendev.dom,
					   txreq.gref, PROT_READ);
	    if (page == NULL) {
		xen_be_printf(&netdev->xendev, 0, "error: tx gref dereference failed (%d)\n",
                              txreq.gref);
		net_tx_error(netdev, &txreq, rc);
		continue;
	    }
	    if (txreq.flags & NETTXF_csum_blank) {
                /* have read-only mapping -> can't fill checksum in-place */
                if (!tmpbuf)
                    tmpbuf = malloc(PAGE_SIZE);
                memcpy(tmpbuf, page + txreq.offset, txreq.size);
		net_checksum_calculate(tmpbuf, txreq.size);
                qemu_send_packet(netdev->vs, tmpbuf, txreq.size);
            } else {
                qemu_send_packet(netdev->vs, page + txreq.offset, txreq.size);
            }
	    xc_gnttab_munmap(netdev->xendev.gnttabdev, page, 1);
	    net_tx_response(netdev, &txreq, NETIF_RSP_OKAY);
	}
	if (!netdev->tx_work)
	    break;
	netdev->tx_work = 0;
    }
    free(tmpbuf);
}

/* ------------------------------------------------------------- */

static void net_rx_response(struct XenNetDev *netdev,
			    netif_rx_request_t *req, int8_t st,
			    uint16_t offset, uint16_t size,
			    uint16_t flags)
{
    RING_IDX i = netdev->rx_ring.rsp_prod_pvt;
    netif_rx_response_t *resp;
    int notify;

    resp = RING_GET_RESPONSE(&netdev->rx_ring, i);
    resp->offset     = offset;
    resp->flags      = flags;
    resp->id         = req->id;
    resp->status     = (int16_t)size;
    if (st < 0)
	resp->status = (int16_t)st;

    xen_be_printf(&netdev->xendev, 3, "rx response: idx %d, status %d, flags 0x%x\n",
		  i, resp->status, resp->flags);

    netdev->rx_ring.rsp_prod_pvt = ++i;
    RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&netdev->rx_ring, notify);
    if (notify)
	xen_be_send_notify(&netdev->xendev);
}

#define NET_IP_ALIGN 2

static int net_rx_ok(void *opaque)
{
    struct XenNetDev *netdev = opaque;
    RING_IDX rc, rp;

    if (netdev->xendev.be_state != XenbusStateConnected)
	return 0;

    rc = netdev->rx_ring.req_cons;
    rp = netdev->rx_ring.sring->req_prod;
    xen_rmb();

    if (rc == rp || RING_REQUEST_CONS_OVERFLOW(&netdev->rx_ring, rc)) {
	xen_be_printf(&netdev->xendev, 2, "%s: no rx buffers (%d/%d)\n",
		      __FUNCTION__, rc, rp);
	return 0;
    }
    return 1;
}

static void net_rx_packet(void *opaque, const uint8_t *buf, int size)
{
    struct XenNetDev *netdev = opaque;
    netif_rx_request_t rxreq;
    RING_IDX rc, rp;
    void *page;

    if (netdev->xendev.be_state != XenbusStateConnected)
	return;

    rc = netdev->rx_ring.req_cons;
    rp = netdev->rx_ring.sring->req_prod;
    xen_rmb(); /* Ensure we see queued requests up to 'rp'. */

    if (rc == rp || RING_REQUEST_CONS_OVERFLOW(&netdev->rx_ring, rc)) {
	xen_be_printf(&netdev->xendev, 2, "no buffer, drop packet\n");
	return;
    }
    if (size > XC_PAGE_SIZE - NET_IP_ALIGN) {
	xen_be_printf(&netdev->xendev, 0, "packet too big (%d > %ld)",
		      size, XC_PAGE_SIZE - NET_IP_ALIGN);
	return;
    }

    memcpy(&rxreq, RING_GET_REQUEST(&netdev->rx_ring, rc), sizeof(rxreq));
    netdev->rx_ring.req_cons = ++rc;

    page = xc_gnttab_map_grant_ref(netdev->xendev.gnttabdev,
				   netdev->xendev.dom,
				   rxreq.gref, PROT_WRITE);
    if (page == NULL) {
	xen_be_printf(&netdev->xendev, 0, "error: rx gref dereference failed (%d)\n",
                      rxreq.gref);
	net_rx_response(netdev, &rxreq, NETIF_RSP_ERROR, 0, 0, 0);
	return;
    }
    memcpy(page + NET_IP_ALIGN, buf, size);
    xc_gnttab_munmap(netdev->xendev.gnttabdev, page, 1);
    net_rx_response(netdev, &rxreq, NETIF_RSP_OKAY, NET_IP_ALIGN, size, 0);
}

/* ------------------------------------------------------------- */

static int net_init(struct XenDevice *xendev)
{
    struct XenNetDev *netdev = container_of(xendev, struct XenNetDev, xendev);
    VLANState *vlan;

    /* read xenstore entries */
    if (netdev->mac == NULL)
	netdev->mac = xenstore_read_be_str(&netdev->xendev, "mac");

    /* do we have all we need? */
    if (netdev->mac == NULL)
	return -1;

    vlan = qemu_find_vlan(netdev->xendev.dev);
    netdev->vs = qemu_new_vlan_client(vlan, "xen", NULL,
                                      net_rx_packet, net_rx_ok, NULL,
                                      netdev);
    snprintf(netdev->vs->info_str, sizeof(netdev->vs->info_str),
             "nic: xenbus vif macaddr=%s", netdev->mac);

    /* fill info */
    xenstore_write_be_int(&netdev->xendev, "feature-rx-copy", 1);
    xenstore_write_be_int(&netdev->xendev, "feature-rx-flip", 0);

    return 0;
}

static int net_connect(struct XenDevice *xendev)
{
    struct XenNetDev *netdev = container_of(xendev, struct XenNetDev, xendev);
    int rx_copy;

    if (xenstore_read_fe_int(&netdev->xendev, "tx-ring-ref",
				   &netdev->tx_ring_ref) == -1)
	return -1;
    if (xenstore_read_fe_int(&netdev->xendev, "rx-ring-ref",
				   &netdev->rx_ring_ref) == -1)
	return 1;
    if (xenstore_read_fe_int(&netdev->xendev, "event-channel",
				   &netdev->xendev.remote_port) == -1)
	return -1;

    if (xenstore_read_fe_int(&netdev->xendev, "request-rx-copy", &rx_copy) == -1)
	rx_copy = 0;
    if (rx_copy == 0) {
	xen_be_printf(&netdev->xendev, 0, "frontend doesn't support rx-copy.\n");
	return -1;
    }

    netdev->txs = xc_gnttab_map_grant_ref(netdev->xendev.gnttabdev,
					  netdev->xendev.dom,
					  netdev->tx_ring_ref,
					  PROT_READ | PROT_WRITE);
    netdev->rxs = xc_gnttab_map_grant_ref(netdev->xendev.gnttabdev,
					  netdev->xendev.dom,
					  netdev->rx_ring_ref,
					  PROT_READ | PROT_WRITE);
    if (!netdev->txs || !netdev->rxs)
	return -1;
    BACK_RING_INIT(&netdev->tx_ring, netdev->txs, XC_PAGE_SIZE);
    BACK_RING_INIT(&netdev->rx_ring, netdev->rxs, XC_PAGE_SIZE);

    xen_be_bind_evtchn(&netdev->xendev);

    xen_be_printf(&netdev->xendev, 1, "ok: tx-ring-ref %d, rx-ring-ref %d, "
		  "remote port %d, local port %d\n",
		  netdev->tx_ring_ref, netdev->rx_ring_ref,
		  netdev->xendev.remote_port, netdev->xendev.local_port);
    return 0;
}

static void net_disconnect(struct XenDevice *xendev)
{
    struct XenNetDev *netdev = container_of(xendev, struct XenNetDev, xendev);

    xen_be_unbind_evtchn(&netdev->xendev);

    if (netdev->txs) {
	xc_gnttab_munmap(netdev->xendev.gnttabdev, netdev->txs, 1);
	netdev->txs = NULL;
    }
    if (netdev->rxs) {
	xc_gnttab_munmap(netdev->xendev.gnttabdev, netdev->rxs, 1);
	netdev->rxs = NULL;
    }
    if (netdev->vs) {
        qemu_del_vlan_client(netdev->vs);
        netdev->vs = NULL;
    }
}

static void net_event(struct XenDevice *xendev)
{
    struct XenNetDev *netdev = container_of(xendev, struct XenNetDev, xendev);
    net_tx_packets(netdev);
}

static int net_free(struct XenDevice *xendev)
{
    struct XenNetDev *netdev = container_of(xendev, struct XenNetDev, xendev);

    qemu_free(netdev->mac);
    return 0;
}

/* ------------------------------------------------------------- */

struct XenDevOps xen_netdev_ops = {
    .size       = sizeof(struct XenNetDev),
    .flags      = DEVOPS_FLAG_NEED_GNTDEV,
    .init       = net_init,
    .connect    = net_connect,
    .event      = net_event,
    .disconnect = net_disconnect,
    .free       = net_free,
};