Spdk/lib/bdev/virtio/bdev_virtio.c
Dariusz Stojaczyk 661feb15a8 bdev_virtio: defer bdev registration until the scan finishes
bdev_virtio I/O poller used to interfere with scan poller.
e.g. GPT read could be picked up by scan poller.

Even if such request is instantly rejected, this would cause
both GPT and scan to fail. Instead, defer bdev_register
until after the virtio_hw scan has finished.

I/O poller logic will still have to be greatly refactored to
add proper multiqueue/multibdev support. However, this patch
lifts scan processing responsibility from any I/O pollers.

Change-Id: I201de8aa0dc1db71ed836fd5e74d55604950f271
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/378064
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
2017-09-14 13:48:09 -04:00

542 lines
13 KiB
C

/*-
* BSD LICENSE
*
* Copyright (c) Intel Corporation.
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
*
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in
* the documentation and/or other materials provided with the
* distribution.
* * Neither the name of Intel Corporation nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
* A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
* OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
* LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#include "spdk/stdinc.h"
#include "spdk/bdev.h"
#include "spdk/conf.h"
#include "spdk/env.h"
#include "spdk/io_channel.h"
#include "spdk/string.h"
#include "spdk/endian.h"
#include "spdk/stdinc.h"
#include "spdk/util.h"
#include "spdk_internal/bdev.h"
#include "spdk_internal/log.h"
#include <getopt.h>
#include <sys/param.h>
#include <linux/virtio_scsi.h>
#include <virtio_dev.h>
#include <virtio_user/virtio_user_dev.h>
#include "spdk/scsi_spec.h"
#define BDEV_VIRTIO_MAX_TARGET 64
#define BDEV_VIRTIO_SCAN_PAYLOAD_SIZE 256
static int bdev_virtio_initialize(void);
static void bdev_virtio_finish(void);
struct virtio_scsi_io_ctx {
struct virtio_req vreq;
struct virtio_scsi_cmd_req req;
struct virtio_scsi_cmd_resp resp;
};
struct virtio_scsi_scan_buf {
struct virtio_scsi_io_ctx io_ctx;
struct iovec iov;
uint8_t payload[BDEV_VIRTIO_SCAN_PAYLOAD_SIZE];
};
struct virtio_scsi_scan_base {
struct virtio_hw *hw;
struct spdk_bdev_poller *scan_poller;
unsigned refcount;
/* Disks to be registered after the scan finishes */
TAILQ_HEAD(, virtio_scsi_disk) found_disks;
struct virtio_scsi_scan_buf buf[BDEV_VIRTIO_MAX_TARGET];
};
struct virtio_scsi_disk {
struct spdk_bdev bdev;
struct virtio_hw *hw;
uint64_t num_blocks;
uint32_t block_size;
TAILQ_ENTRY(virtio_scsi_disk) link;
};
struct bdev_virtio_io_channel {
struct virtio_hw *hw;
struct spdk_bdev_poller *poller;
};
static int
bdev_virtio_get_ctx_size(void)
{
return sizeof(struct virtio_scsi_io_ctx);
}
SPDK_BDEV_MODULE_REGISTER(virtio_scsi, bdev_virtio_initialize, bdev_virtio_finish,
NULL, bdev_virtio_get_ctx_size, NULL)
SPDK_BDEV_MODULE_ASYNC_INIT(virtio_scsi)
static void
bdev_virtio_rw(struct spdk_io_channel *ch, struct spdk_bdev_io *bdev_io)
{
struct virtio_req *vreq;
struct virtio_scsi_cmd_req *req;
struct virtio_scsi_cmd_resp *resp;
struct virtio_scsi_disk *disk = (struct virtio_scsi_disk *)bdev_io->bdev;
struct virtio_scsi_io_ctx *io_ctx = (struct virtio_scsi_io_ctx *)bdev_io->driver_ctx;
bool is_read = (bdev_io->type == SPDK_BDEV_IO_TYPE_READ);
vreq = &io_ctx->vreq;
req = &io_ctx->req;
resp = &io_ctx->resp;
vreq->iov_req.iov_base = (void *)req;
vreq->iov_req.iov_len = sizeof(*req);
vreq->iov_resp.iov_base = (void *)resp;
vreq->iov_resp.iov_len = sizeof(*resp);
vreq->is_write = !is_read;
memset(req, 0, sizeof(*req));
req->lun[0] = 1;
req->lun[1] = 0;
if (is_read) {
vreq->iov = bdev_io->u.read.iovs;
vreq->iovcnt = bdev_io->u.read.iovcnt;
req->cdb[0] = SPDK_SBC_READ_10;
to_be32(&req->cdb[2], bdev_io->u.read.offset_blocks);
to_be16(&req->cdb[7], bdev_io->u.read.num_blocks);
} else {
vreq->iov = bdev_io->u.write.iovs;
vreq->iovcnt = bdev_io->u.write.iovcnt;
req->cdb[0] = SPDK_SBC_WRITE_10;
to_be32(&req->cdb[2], bdev_io->u.write.offset_blocks);
to_be16(&req->cdb[7], bdev_io->u.write.num_blocks);
}
virtio_xmit_pkts(disk->hw->vqs[2], vreq);
}
static int _bdev_virtio_submit_request(struct spdk_io_channel *ch, struct spdk_bdev_io *bdev_io)
{
switch (bdev_io->type) {
case SPDK_BDEV_IO_TYPE_READ:
spdk_bdev_io_get_buf(bdev_io, bdev_virtio_rw);
return 0;
case SPDK_BDEV_IO_TYPE_WRITE:
bdev_virtio_rw(ch, bdev_io);
return 0;
case SPDK_BDEV_IO_TYPE_RESET:
spdk_bdev_io_complete(bdev_io, SPDK_BDEV_IO_STATUS_SUCCESS);
return 0;
case SPDK_BDEV_IO_TYPE_FLUSH:
case SPDK_BDEV_IO_TYPE_UNMAP:
default:
return -1;
}
return 0;
}
static void bdev_virtio_submit_request(struct spdk_io_channel *ch, struct spdk_bdev_io *bdev_io)
{
if (_bdev_virtio_submit_request(ch, bdev_io) < 0) {
spdk_bdev_io_complete(bdev_io, SPDK_BDEV_IO_STATUS_FAILED);
}
}
static bool
bdev_virtio_io_type_supported(void *ctx, enum spdk_bdev_io_type io_type)
{
switch (io_type) {
case SPDK_BDEV_IO_TYPE_READ:
case SPDK_BDEV_IO_TYPE_WRITE:
case SPDK_BDEV_IO_TYPE_FLUSH:
case SPDK_BDEV_IO_TYPE_RESET:
case SPDK_BDEV_IO_TYPE_UNMAP:
return true;
default:
return false;
}
}
static struct spdk_io_channel *
bdev_virtio_get_io_channel(void *ctx)
{
struct virtio_scsi_disk *disk = ctx;
return spdk_get_io_channel(&disk->hw);
}
static int
bdev_virtio_destruct(void *ctx)
{
return 0;
}
static const struct spdk_bdev_fn_table virtio_fn_table = {
.destruct = bdev_virtio_destruct,
.submit_request = bdev_virtio_submit_request,
.io_type_supported = bdev_virtio_io_type_supported,
.get_io_channel = bdev_virtio_get_io_channel,
};
static void
get_scsi_status(struct virtio_scsi_cmd_resp *resp, int *sk, int *asc, int *ascq)
{
/* see spdk_scsi_task_build_sense_data() for sense data details */
*sk = 0;
*asc = 0;
*ascq = 0;
if (resp->sense_len < 3) {
return;
}
*sk = resp->sense[2] & 0xf;
if (resp->sense_len < 13) {
return;
}
*asc = resp->sense[12];
if (resp->sense_len < 14) {
return;
}
*ascq = resp->sense[13];
}
static void
bdev_virtio_io_cpl(struct virtio_req *req)
{
struct virtio_scsi_io_ctx *io_ctx = SPDK_CONTAINEROF(req, struct virtio_scsi_io_ctx, vreq);
struct spdk_bdev_io *bdev_io = spdk_bdev_io_from_ctx(io_ctx);
int sk, asc, ascq;
get_scsi_status(&io_ctx->resp, &sk, &asc, &ascq);
spdk_bdev_io_complete_scsi_status(bdev_io, io_ctx->resp.status, sk, asc, ascq);
}
static void
bdev_virtio_poll(void *arg)
{
struct bdev_virtio_io_channel *ch = arg;
struct virtio_req *req[32];
uint16_t i, cnt;
cnt = virtio_recv_pkts(ch->hw->vqs[2], req, SPDK_COUNTOF(req));
for (i = 0; i < cnt; ++i) {
bdev_virtio_io_cpl(req[i]);
}
}
static int
bdev_virtio_create_cb(void *io_device, void *ctx_buf)
{
struct virtio_hw **hw = io_device;
struct bdev_virtio_io_channel *ch = ctx_buf;
ch->hw = *hw;
spdk_bdev_poller_start(&ch->poller, bdev_virtio_poll, ch,
spdk_env_get_current_core(), 0);
return 0;
}
static void
bdev_virtio_destroy_cb(void *io_device, void *ctx_buf)
{
struct bdev_virtio_io_channel *io_channel = ctx_buf;
spdk_bdev_poller_stop(&io_channel->poller);
}
static void
scan_target_finish(struct virtio_scsi_scan_base *base)
{
struct virtio_scsi_disk *disk;
assert(base->refcount > 0);
base->refcount--;
if (base->refcount > 0) {
return;
}
spdk_bdev_poller_stop(&base->scan_poller);
while ((disk = TAILQ_FIRST(&base->found_disks))) {
TAILQ_REMOVE(&base->found_disks, disk, link);
spdk_io_device_register(&disk->hw, bdev_virtio_create_cb, bdev_virtio_destroy_cb,
sizeof(struct bdev_virtio_io_channel));
spdk_bdev_register(&disk->bdev);
}
spdk_dma_free(base);
spdk_bdev_module_init_done(SPDK_GET_BDEV_MODULE(virtio_scsi));
}
static int
process_scan_inquiry(struct virtio_scsi_scan_base *base, struct virtio_req *vreq)
{
struct iovec *iov = vreq->iov;
struct virtio_scsi_cmd_req *req = vreq->iov_req.iov_base;
struct virtio_scsi_cmd_resp *resp = vreq->iov_resp.iov_base;
uint8_t lun_id;
if (resp->response != VIRTIO_SCSI_S_OK || resp->status != SPDK_SCSI_STATUS_GOOD) {
return -1;
}
lun_id = req->lun[1];
/* reuse vreq for next request */
memset(req, 0, sizeof(*req));
req->lun[0] = 1;
req->lun[1] = lun_id;
req->cdb[0] = SPDK_SPC_SERVICE_ACTION_IN_16;
req->cdb[1] = SPDK_SBC_SAI_READ_CAPACITY_16;
iov[0].iov_len = 32;
to_be32(&req->cdb[10], iov[0].iov_len);
virtio_xmit_pkts(base->hw->vqs[2], vreq);
return 0;
}
static int
process_read_cap(struct virtio_scsi_scan_base *base, struct virtio_req *vreq)
{
struct virtio_scsi_disk *disk;
struct spdk_bdev *bdev;
struct virtio_scsi_cmd_req *req = vreq->iov_req.iov_base;
struct virtio_scsi_cmd_resp *resp = vreq->iov_resp.iov_base;
if (resp->response != VIRTIO_SCSI_S_OK || resp->status != SPDK_SCSI_STATUS_GOOD) {
SPDK_ERRLOG("read capacity failed for target %"PRIu8".\n", req->lun[1]);
return -1;
}
disk = calloc(1, sizeof(*disk));
if (disk == NULL) {
SPDK_ERRLOG("could not allocate disk\n");
return -1;
}
disk->num_blocks = from_be64((uint64_t *)(vreq->iov[0].iov_base)) + 1;
disk->block_size = from_be32((uint32_t *)(vreq->iov[0].iov_base + 8));
disk->hw = base->hw;
bdev = &disk->bdev;
bdev->name = spdk_sprintf_alloc("Virtio0");
bdev->product_name = "Virtio SCSI Disk";
bdev->write_cache = 0;
bdev->blocklen = disk->block_size;
bdev->blockcnt = disk->num_blocks;
bdev->ctxt = disk;
bdev->fn_table = &virtio_fn_table;
bdev->module = SPDK_GET_BDEV_MODULE(virtio_scsi);
TAILQ_INSERT_TAIL(&base->found_disks, disk, link);
scan_target_finish(base);
return 0;
}
static void
process_scan_resp(struct virtio_scsi_scan_base *base, struct virtio_req *vreq)
{
struct virtio_scsi_cmd_req *req = vreq->iov_req.iov_base;
int rc;
if (vreq->iov_req.iov_len < sizeof(struct virtio_scsi_cmd_req) ||
vreq->iov_resp.iov_len < sizeof(struct virtio_scsi_cmd_resp)) {
SPDK_ERRLOG("Received target scan message with invalid length.\n");
scan_target_finish(base);
return;
}
switch (req->cdb[0]) {
case SPDK_SPC_INQUIRY:
rc = process_scan_inquiry(base, vreq);
break;
case SPDK_SPC_SERVICE_ACTION_IN_16:
rc = process_read_cap(base, vreq);
break;
default:
SPDK_ERRLOG("Received invalid target scan message: cdb[0] = %"PRIu8".\n", req->cdb[0]);
rc = -1;
break;
}
if (rc < 0) {
scan_target_finish(base);
}
}
static void
bdev_scan_poll(void *arg)
{
struct virtio_scsi_scan_base *base = arg;
struct virtio_req *req[32];
uint16_t i, cnt;
cnt = virtio_recv_pkts(base->hw->vqs[2], req, SPDK_COUNTOF(req));
if (cnt > base->refcount) {
SPDK_ERRLOG("Received too many virtio messages. Got %"PRIu16", expected %u",
cnt, base->refcount);
return;
}
for (i = 0; i < cnt; ++i) {
process_scan_resp(base, req[i]);
}
}
static void
scan_target(struct virtio_scsi_scan_base *base, uint8_t target)
{
struct iovec *iov;
struct virtio_req *vreq;
struct virtio_scsi_cmd_req *req;
struct virtio_scsi_cmd_resp *resp;
struct spdk_scsi_cdb_inquiry *cdb;
vreq = &base->buf[target].io_ctx.vreq;
req = &base->buf[target].io_ctx.req;
resp = &base->buf[target].io_ctx.resp;
iov = &base->buf[target].iov;
vreq->iov = iov;
vreq->iovcnt = 1;
vreq->is_write = 0;
vreq->iov_req.iov_base = (void *)req;
vreq->iov_req.iov_len = sizeof(*req);
vreq->iov_resp.iov_base = (void *)resp;
vreq->iov_resp.iov_len = sizeof(*resp);
iov[0].iov_base = (void *)&base->buf[target].payload;
iov[0].iov_len = BDEV_VIRTIO_SCAN_PAYLOAD_SIZE;
req->lun[0] = 1;
req->lun[1] = target;
cdb = (struct spdk_scsi_cdb_inquiry *)req->cdb;
cdb->opcode = SPDK_SPC_INQUIRY;
cdb->alloc_len[1] = 255;
virtio_xmit_pkts(base->hw->vqs[2], vreq);
}
static int
bdev_virtio_initialize(void)
{
struct spdk_conf_section *sp = spdk_conf_find_section(NULL, "Virtio");
struct virtio_scsi_scan_base *base;
struct virtio_hw *hw = NULL;
char *type, *path;
uint32_t i;
int rc = 0;
if (sp == NULL) {
goto out;
}
for (i = 0; spdk_conf_section_get_nval(sp, "Dev", i) != NULL; i++) {
type = spdk_conf_section_get_nmval(sp, "Dev", i, 0);
if (type == NULL) {
SPDK_ERRLOG("No type specified for index %d\n", i);
continue;
}
if (!strcmp("User", type)) {
path = spdk_conf_section_get_nmval(sp, "Dev", i, 1);
if (path == NULL) {
SPDK_ERRLOG("No path specified for index %d\n", i);
continue;
}
hw = virtio_user_dev_init(path, 1, 512);
} else if (!strcmp("Pci", type)) {
hw = get_pci_virtio_hw();
} else {
SPDK_ERRLOG("Invalid type %s specified for index %d\n", type, i);
continue;
}
}
if (hw == NULL) {
goto out;
}
base = spdk_dma_zmalloc(sizeof(*base), 64, NULL);
if (base == NULL) {
SPDK_ERRLOG("couldn't allocate memory for scsi target scan.\n");
rc = -1;
goto out;
}
/* TODO check rc, add virtio_dev_deinit() */
eth_virtio_dev_init(hw);
virtio_dev_start(hw);
base->hw = hw;
base->refcount = BDEV_VIRTIO_MAX_TARGET;
TAILQ_INIT(&base->found_disks);
spdk_bdev_poller_start(&base->scan_poller, bdev_scan_poll, base,
spdk_env_get_current_core(), 0);
for (i = 0; i < BDEV_VIRTIO_MAX_TARGET; i++) {
scan_target(base, i);
}
return 0;
out:
spdk_bdev_module_init_done(SPDK_GET_BDEV_MODULE(virtio_scsi));
return rc;
}
static void bdev_virtio_finish(void)
{
}
SPDK_LOG_REGISTER_TRACE_FLAG("virtio", SPDK_TRACE_VIRTIO)