2 * Copyright (c) 2011 NetApp, Inc.
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
8 * 1. Redistributions of source code must retain the above copyright
9 * notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 * notice, this list of conditions and the following disclaimer in the
12 * documentation and/or other materials provided with the distribution.
14 * THIS SOFTWARE IS PROVIDED BY NETAPP, INC ``AS IS'' AND
15 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
16 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
17 * ARE DISCLAIMED. IN NO EVENT SHALL NETAPP, INC OR CONTRIBUTORS BE LIABLE
18 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
19 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
20 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
21 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
22 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
23 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29 #include <sys/cdefs.h>
30 __FBSDID("$FreeBSD$");
32 #include <sys/param.h>
33 #include <sys/linker_set.h>
36 #include <sys/ioctl.h>
55 #define VTBLK_RINGSZ 64
57 #define VTBLK_MAXSEGS 32
60 #define VTBLK_S_IOERR 1
61 #define VTBLK_S_UNSUPP 2
63 #define VTBLK_BLK_ID_BYTES 20
66 #define VTBLK_F_SEG_MAX (1 << 2) /* Maximum request segments */
67 #define VTBLK_F_BLK_SIZE (1 << 6) /* cfg block size valid */
72 #define VTBLK_S_HOSTCAPS \
75 VIRTIO_RING_F_INDIRECT_DESC ) /* indirect descriptors */
78 * Config space "registers"
81 uint64_t vbc_capacity;
82 uint32_t vbc_size_max;
87 uint32_t vbc_blk_size;
88 uint32_t vbc_sectors_max;
92 * Fixed-size block header
94 struct virtio_blk_hdr {
96 #define VBH_OP_WRITE 1
97 #define VBH_OP_FLUSH 4
98 #define VBH_OP_FLUSH_OUT 5
99 #define VBH_OP_IDENT 8
100 #define VBH_FLAG_BARRIER 0x80000000 /* OR'ed into vbh_type */
109 static int pci_vtblk_debug;
110 #define DPRINTF(params) if (pci_vtblk_debug) printf params
111 #define WPRINTF(params) printf params
116 struct pci_vtblk_softc {
117 struct virtio_softc vbsc_vs;
118 pthread_mutex_t vsc_mtx;
119 struct vqueue_info vbsc_vq;
121 struct vtblk_config vbsc_cfg;
122 char vbsc_ident[VTBLK_BLK_ID_BYTES];
125 static void pci_vtblk_reset(void *);
126 static void pci_vtblk_notify(void *, struct vqueue_info *);
127 static int pci_vtblk_cfgread(void *, int, int, uint32_t *);
128 static int pci_vtblk_cfgwrite(void *, int, int, uint32_t);
130 static struct virtio_consts vtblk_vi_consts = {
131 "vtblk", /* our name */
132 1, /* we support 1 virtqueue */
133 sizeof(struct vtblk_config), /* config reg size */
134 pci_vtblk_reset, /* reset */
135 pci_vtblk_notify, /* device-wide qnotify */
136 pci_vtblk_cfgread, /* read PCI config */
137 pci_vtblk_cfgwrite, /* write PCI config */
138 NULL, /* apply negotiated features */
139 VTBLK_S_HOSTCAPS, /* our capabilities */
143 pci_vtblk_reset(void *vsc)
145 struct pci_vtblk_softc *sc = vsc;
147 DPRINTF(("vtblk: device reset requested !\n"));
148 vi_reset_dev(&sc->vbsc_vs);
152 pci_vtblk_proc(struct pci_vtblk_softc *sc, struct vqueue_info *vq)
154 struct virtio_blk_hdr *vbh;
161 struct iovec iov[VTBLK_MAXSEGS + 2];
162 uint16_t flags[VTBLK_MAXSEGS + 2];
164 n = vq_getchain(vq, iov, VTBLK_MAXSEGS + 2, flags);
167 * The first descriptor will be the read-only fixed header,
168 * and the last is for status (hence +2 above and below).
169 * The remaining iov's are the actual data I/O vectors.
171 * XXX - note - this fails on crash dump, which does a
172 * VIRTIO_BLK_T_FLUSH with a zero transfer length
174 assert(n >= 2 && n <= VTBLK_MAXSEGS + 2);
176 assert((flags[0] & VRING_DESC_F_WRITE) == 0);
177 assert(iov[0].iov_len == sizeof(struct virtio_blk_hdr));
178 vbh = iov[0].iov_base;
180 status = iov[--n].iov_base;
181 assert(iov[n].iov_len == 1);
182 assert(flags[n] & VRING_DESC_F_WRITE);
186 * The guest should not be setting the BARRIER flag because
187 * we don't advertise the capability.
189 type = vbh->vbh_type & ~VBH_FLAG_BARRIER;
190 writeop = (type == VBH_OP_WRITE);
192 offset = vbh->vbh_sector * DEV_BSIZE;
195 for (i = 1; i < n; i++) {
197 * - write op implies read-only descriptor,
198 * - read/ident op implies write-only descriptor,
199 * therefore test the inverse of the descriptor bit
202 assert(((flags[i] & VRING_DESC_F_WRITE) == 0) == writeop);
203 iolen += iov[i].iov_len;
206 DPRINTF(("virtio-block: %s op, %d bytes, %d segs, offset %ld\n\r",
207 writeop ? "write" : "read/ident", iolen, i - 1, offset));
211 err = pwritev(sc->vbsc_fd, iov + 1, i - 1, offset);
214 err = preadv(sc->vbsc_fd, iov + 1, i - 1, offset);
217 /* Assume a single buffer */
218 strlcpy(iov[1].iov_base, sc->vbsc_ident,
219 MIN(iov[1].iov_len, sizeof(sc->vbsc_ident)));
223 case VBH_OP_FLUSH_OUT:
224 err = fsync(sc->vbsc_fd);
231 /* convert errno into a virtio block error return */
234 *status = VTBLK_S_UNSUPP;
236 *status = VTBLK_S_IOERR;
238 *status = VTBLK_S_OK;
241 * Return the descriptor back to the host.
242 * We wrote 1 byte (our status) to host.
248 pci_vtblk_notify(void *vsc, struct vqueue_info *vq)
250 struct pci_vtblk_softc *sc = vsc;
253 while (vq_has_descs(vq))
254 pci_vtblk_proc(sc, vq);
255 vq_endchains(vq, 1); /* Generate interrupt if appropriate. */
259 pci_vtblk_init(struct vmctx *ctx, struct pci_devinst *pi, char *opts)
264 struct pci_vtblk_softc *sc;
270 printf("virtio-block: backing device required\n");
275 * The supplied backing file has to exist
277 fd = open(opts, O_RDWR);
279 perror("Could not open backing file");
283 if (fstat(fd, &sbuf) < 0) {
284 perror("Could not stat backing file");
290 * Deal with raw devices
294 if (S_ISCHR(sbuf.st_mode)) {
295 if (ioctl(fd, DIOCGMEDIASIZE, &size) < 0 ||
296 ioctl(fd, DIOCGSECTORSIZE, §sz)) {
297 perror("Could not fetch dev blk/sector size");
305 sc = calloc(1, sizeof(struct pci_vtblk_softc));
307 /* record fd of storage device/file */
310 pthread_mutex_init(&sc->vsc_mtx, NULL);
312 /* init virtio softc and virtqueues */
313 vi_softc_linkup(&sc->vbsc_vs, &vtblk_vi_consts, sc, pi, &sc->vbsc_vq);
314 sc->vbsc_vs.vs_mtx = &sc->vsc_mtx;
316 sc->vbsc_vq.vq_qsize = VTBLK_RINGSZ;
317 /* sc->vbsc_vq.vq_notify = we have no per-queue notify */
320 * Create an identifier for the backing file. Use parts of the
321 * md5 sum of the filename
324 MD5Update(&mdctx, opts, strlen(opts));
325 MD5Final(digest, &mdctx);
326 sprintf(sc->vbsc_ident, "BHYVE-%02X%02X-%02X%02X-%02X%02X",
327 digest[0], digest[1], digest[2], digest[3], digest[4], digest[5]);
329 /* setup virtio block config space */
330 sc->vbsc_cfg.vbc_capacity = size / DEV_BSIZE; /* 512-byte units */
331 sc->vbsc_cfg.vbc_seg_max = VTBLK_MAXSEGS;
332 sc->vbsc_cfg.vbc_blk_size = sectsz;
333 sc->vbsc_cfg.vbc_size_max = 0; /* not negotiated */
334 sc->vbsc_cfg.vbc_geom_c = 0; /* no geometry */
335 sc->vbsc_cfg.vbc_geom_h = 0;
336 sc->vbsc_cfg.vbc_geom_s = 0;
337 sc->vbsc_cfg.vbc_sectors_max = 0;
340 * Should we move some of this into virtio.c? Could
341 * have the device, class, and subdev_0 as fields in
342 * the virtio constants structure.
344 pci_set_cfgdata16(pi, PCIR_DEVICE, VIRTIO_DEV_BLOCK);
345 pci_set_cfgdata16(pi, PCIR_VENDOR, VIRTIO_VENDOR);
346 pci_set_cfgdata8(pi, PCIR_CLASS, PCIC_STORAGE);
347 pci_set_cfgdata16(pi, PCIR_SUBDEV_0, VIRTIO_TYPE_BLOCK);
349 pci_lintr_request(pi);
351 if (vi_intr_init(&sc->vbsc_vs, 1, fbsdrun_virtio_msix()))
353 vi_set_io_bar(&sc->vbsc_vs, 0);
358 pci_vtblk_cfgwrite(void *vsc, int offset, int size, uint32_t value)
361 DPRINTF(("vtblk: write to readonly reg %d\n\r", offset));
366 pci_vtblk_cfgread(void *vsc, int offset, int size, uint32_t *retval)
368 struct pci_vtblk_softc *sc = vsc;
371 /* our caller has already verified offset and size */
372 ptr = (uint8_t *)&sc->vbsc_cfg + offset;
373 memcpy(retval, ptr, size);
377 struct pci_devemu pci_de_vblk = {
378 .pe_emu = "virtio-blk",
379 .pe_init = pci_vtblk_init,
380 .pe_barwrite = vi_pci_write,
381 .pe_barread = vi_pci_read
383 PCI_EMUL_SET(pci_de_vblk);