/*
- * Copyright (C) 2016 Red Hat, Inc.
+ * Copyright (C) 2016-2017 Red Hat, Inc.
*
* Network Block Device Server Side
#include "qemu/osdep.h"
#include "qapi/error.h"
+#include "trace.h"
#include "nbd-internal.h"
static int system_errno_to_nbd_errno(int err)
case EFBIG:
case ENOSPC:
return NBD_ENOSPC;
+ case EOVERFLOW:
+ return NBD_EOVERFLOW;
case ESHUTDOWN:
return NBD_ESHUTDOWN;
case EINVAL:
int refcount;
void (*close_fn)(NBDClient *client, bool negotiated);
- bool no_zeroes;
NBDExport *exp;
QCryptoTLSCreds *tlscreds;
char *tlsaclname;
/* Send a reply header, including length, but no payload.
* Return -errno on error, 0 on success. */
static int nbd_negotiate_send_rep_len(QIOChannel *ioc, uint32_t type,
- uint32_t opt, uint32_t len)
+ uint32_t opt, uint32_t len, Error **errp)
{
uint64_t magic;
- TRACE("Reply opt=%" PRIx32 " type=%" PRIx32 " len=%" PRIu32,
- type, opt, len);
+ trace_nbd_negotiate_send_rep_len(opt, nbd_opt_lookup(opt),
+ type, nbd_rep_lookup(type), len);
+ assert(len < NBD_MAX_BUFFER_SIZE);
magic = cpu_to_be64(NBD_REP_MAGIC);
- if (nbd_write(ioc, &magic, sizeof(magic), NULL) < 0) {
- LOG("write failed (rep magic)");
+ if (nbd_write(ioc, &magic, sizeof(magic), errp) < 0) {
+ error_prepend(errp, "write failed (rep magic): ");
return -EINVAL;
}
+
opt = cpu_to_be32(opt);
- if (nbd_write(ioc, &opt, sizeof(opt), NULL) < 0) {
- LOG("write failed (rep opt)");
+ if (nbd_write(ioc, &opt, sizeof(opt), errp) < 0) {
+ error_prepend(errp, "write failed (rep opt): ");
return -EINVAL;
}
+
type = cpu_to_be32(type);
- if (nbd_write(ioc, &type, sizeof(type), NULL) < 0) {
- LOG("write failed (rep type)");
+ if (nbd_write(ioc, &type, sizeof(type), errp) < 0) {
+ error_prepend(errp, "write failed (rep type): ");
return -EINVAL;
}
+
len = cpu_to_be32(len);
- if (nbd_write(ioc, &len, sizeof(len), NULL) < 0) {
- LOG("write failed (rep data length)");
+ if (nbd_write(ioc, &len, sizeof(len), errp) < 0) {
+ error_prepend(errp, "write failed (rep data length): ");
return -EINVAL;
}
return 0;
/* Send a reply header with default 0 length.
* Return -errno on error, 0 on success. */
-static int nbd_negotiate_send_rep(QIOChannel *ioc, uint32_t type, uint32_t opt)
+static int nbd_negotiate_send_rep(QIOChannel *ioc, uint32_t type, uint32_t opt,
+ Error **errp)
{
- return nbd_negotiate_send_rep_len(ioc, type, opt, 0);
+ return nbd_negotiate_send_rep_len(ioc, type, opt, 0, errp);
}
/* Send an error reply.
* Return -errno on error, 0 on success. */
-static int GCC_FMT_ATTR(4, 5)
+static int GCC_FMT_ATTR(5, 6)
nbd_negotiate_send_rep_err(QIOChannel *ioc, uint32_t type,
- uint32_t opt, const char *fmt, ...)
+ uint32_t opt, Error **errp, const char *fmt, ...)
{
va_list va;
char *msg;
va_end(va);
len = strlen(msg);
assert(len < 4096);
- TRACE("sending error message \"%s\"", msg);
- ret = nbd_negotiate_send_rep_len(ioc, type, opt, len);
+ trace_nbd_negotiate_send_rep_err(msg);
+ ret = nbd_negotiate_send_rep_len(ioc, type, opt, len, errp);
if (ret < 0) {
goto out;
}
- if (nbd_write(ioc, msg, len, NULL) < 0) {
- LOG("write failed (error message)");
+ if (nbd_write(ioc, msg, len, errp) < 0) {
+ error_prepend(errp, "write failed (error message): ");
ret = -EIO;
} else {
ret = 0;
}
+
out:
g_free(msg);
return ret;
/* Send a single NBD_REP_SERVER reply to NBD_OPT_LIST, including payload.
* Return -errno on error, 0 on success. */
-static int nbd_negotiate_send_rep_list(QIOChannel *ioc, NBDExport *exp)
+static int nbd_negotiate_send_rep_list(QIOChannel *ioc, NBDExport *exp,
+ Error **errp)
{
size_t name_len, desc_len;
uint32_t len;
const char *desc = exp->description ? exp->description : "";
int ret;
- TRACE("Advertising export name '%s' description '%s'", name, desc);
+ trace_nbd_negotiate_send_rep_list(name, desc);
name_len = strlen(name);
desc_len = strlen(desc);
len = name_len + desc_len + sizeof(len);
- ret = nbd_negotiate_send_rep_len(ioc, NBD_REP_SERVER, NBD_OPT_LIST, len);
+ ret = nbd_negotiate_send_rep_len(ioc, NBD_REP_SERVER, NBD_OPT_LIST, len,
+ errp);
if (ret < 0) {
return ret;
}
len = cpu_to_be32(name_len);
- if (nbd_write(ioc, &len, sizeof(len), NULL) < 0) {
- LOG("write failed (name length)");
+ if (nbd_write(ioc, &len, sizeof(len), errp) < 0) {
+ error_prepend(errp, "write failed (name length): ");
return -EINVAL;
}
- if (nbd_write(ioc, name, name_len, NULL) < 0) {
- LOG("write failed (name buffer)");
+
+ if (nbd_write(ioc, name, name_len, errp) < 0) {
+ error_prepend(errp, "write failed (name buffer): ");
return -EINVAL;
}
- if (nbd_write(ioc, desc, desc_len, NULL) < 0) {
- LOG("write failed (description buffer)");
+
+ if (nbd_write(ioc, desc, desc_len, errp) < 0) {
+ error_prepend(errp, "write failed (description buffer): ");
return -EINVAL;
}
+
return 0;
}
/* Process the NBD_OPT_LIST command, with a potential series of replies.
* Return -errno on error, 0 on success. */
-static int nbd_negotiate_handle_list(NBDClient *client, uint32_t length)
+static int nbd_negotiate_handle_list(NBDClient *client, uint32_t length,
+ Error **errp)
{
NBDExport *exp;
if (length) {
- if (nbd_drop(client->ioc, length, NULL) < 0) {
+ if (nbd_drop(client->ioc, length, errp) < 0) {
return -EIO;
}
return nbd_negotiate_send_rep_err(client->ioc,
NBD_REP_ERR_INVALID, NBD_OPT_LIST,
+ errp,
"OPT_LIST should not have length");
}
/* For each export, send a NBD_REP_SERVER reply. */
QTAILQ_FOREACH(exp, &exports, next) {
- if (nbd_negotiate_send_rep_list(client->ioc, exp)) {
+ if (nbd_negotiate_send_rep_list(client->ioc, exp, errp)) {
return -EINVAL;
}
}
/* Finish with a NBD_REP_ACK. */
- return nbd_negotiate_send_rep(client->ioc, NBD_REP_ACK, NBD_OPT_LIST);
+ return nbd_negotiate_send_rep(client->ioc, NBD_REP_ACK, NBD_OPT_LIST, errp);
}
-static int nbd_negotiate_handle_export_name(NBDClient *client, uint32_t length)
+/* Send a reply to NBD_OPT_EXPORT_NAME.
+ * Return -errno on error, 0 on success. */
+static int nbd_negotiate_handle_export_name(NBDClient *client, uint32_t length,
+ uint16_t myflags, bool no_zeroes,
+ Error **errp)
{
char name[NBD_MAX_NAME_SIZE + 1];
+ char buf[NBD_REPLY_EXPORT_NAME_SIZE] = "";
+ size_t len;
+ int ret;
/* Client sends:
[20 .. xx] export name (length bytes)
+ Server replies:
+ [ 0 .. 7] size
+ [ 8 .. 9] export flags
+ [10 .. 133] reserved (0) [unless no_zeroes]
*/
- TRACE("Checking length");
+ trace_nbd_negotiate_handle_export_name();
if (length >= sizeof(name)) {
- LOG("Bad length received");
+ error_setg(errp, "Bad length received");
return -EINVAL;
}
- if (nbd_read(client->ioc, name, length, NULL) < 0) {
- LOG("read failed");
+ if (nbd_read(client->ioc, name, length, errp) < 0) {
+ error_prepend(errp, "read failed: ");
return -EINVAL;
}
name[length] = '\0';
- TRACE("Client requested export '%s'", name);
+ trace_nbd_negotiate_handle_export_name_request(name);
client->exp = nbd_export_find(name);
if (!client->exp) {
- LOG("export not found");
+ error_setg(errp, "export not found");
return -EINVAL;
}
+ trace_nbd_negotiate_new_style_size_flags(client->exp->size,
+ client->exp->nbdflags | myflags);
+ stq_be_p(buf, client->exp->size);
+ stw_be_p(buf + 8, client->exp->nbdflags | myflags);
+ len = no_zeroes ? 10 : sizeof(buf);
+ ret = nbd_write(client->ioc, buf, len, errp);
+ if (ret < 0) {
+ error_prepend(errp, "write failed: ");
+ return ret;
+ }
+
QTAILQ_INSERT_TAIL(&client->exp->clients, client, next);
nbd_export_get(client->exp);
return 0;
}
+/* Send a single NBD_REP_INFO, with a buffer @buf of @length bytes.
+ * The buffer does NOT include the info type prefix.
+ * Return -errno on error, 0 if ready to send more. */
+static int nbd_negotiate_send_info(NBDClient *client, uint32_t opt,
+ uint16_t info, uint32_t length, void *buf,
+ Error **errp)
+{
+ int rc;
+
+ trace_nbd_negotiate_send_info(info, nbd_info_lookup(info), length);
+ rc = nbd_negotiate_send_rep_len(client->ioc, NBD_REP_INFO, opt,
+ sizeof(info) + length, errp);
+ if (rc < 0) {
+ return rc;
+ }
+ cpu_to_be16s(&info);
+ if (nbd_write(client->ioc, &info, sizeof(info), errp) < 0) {
+ return -EIO;
+ }
+ if (nbd_write(client->ioc, buf, length, errp) < 0) {
+ return -EIO;
+ }
+ return 0;
+}
+
+/* Handle NBD_OPT_INFO and NBD_OPT_GO.
+ * Return -errno on error, 0 if ready for next option, and 1 to move
+ * into transmission phase. */
+static int nbd_negotiate_handle_info(NBDClient *client, uint32_t length,
+ uint32_t opt, uint16_t myflags,
+ Error **errp)
+{
+ int rc;
+ char name[NBD_MAX_NAME_SIZE + 1];
+ NBDExport *exp;
+ uint16_t requests;
+ uint16_t request;
+ uint32_t namelen;
+ bool sendname = false;
+ bool blocksize = false;
+ uint32_t sizes[3];
+ char buf[sizeof(uint64_t) + sizeof(uint16_t)];
+ const char *msg;
+
+ /* Client sends:
+ 4 bytes: L, name length (can be 0)
+ L bytes: export name
+ 2 bytes: N, number of requests (can be 0)
+ N * 2 bytes: N requests
+ */
+ if (length < sizeof(namelen) + sizeof(requests)) {
+ msg = "overall request too short";
+ goto invalid;
+ }
+ if (nbd_read(client->ioc, &namelen, sizeof(namelen), errp) < 0) {
+ return -EIO;
+ }
+ be32_to_cpus(&namelen);
+ length -= sizeof(namelen);
+ if (namelen > length - sizeof(requests) || (length - namelen) % 2) {
+ msg = "name length is incorrect";
+ goto invalid;
+ }
+ if (nbd_read(client->ioc, name, namelen, errp) < 0) {
+ return -EIO;
+ }
+ name[namelen] = '\0';
+ length -= namelen;
+ trace_nbd_negotiate_handle_export_name_request(name);
+
+ if (nbd_read(client->ioc, &requests, sizeof(requests), errp) < 0) {
+ return -EIO;
+ }
+ be16_to_cpus(&requests);
+ length -= sizeof(requests);
+ trace_nbd_negotiate_handle_info_requests(requests);
+ if (requests != length / sizeof(request)) {
+ msg = "incorrect number of requests for overall length";
+ goto invalid;
+ }
+ while (requests--) {
+ if (nbd_read(client->ioc, &request, sizeof(request), errp) < 0) {
+ return -EIO;
+ }
+ be16_to_cpus(&request);
+ length -= sizeof(request);
+ trace_nbd_negotiate_handle_info_request(request,
+ nbd_info_lookup(request));
+ /* We care about NBD_INFO_NAME and NBD_INFO_BLOCK_SIZE;
+ * everything else is either a request we don't know or
+ * something we send regardless of request */
+ switch (request) {
+ case NBD_INFO_NAME:
+ sendname = true;
+ break;
+ case NBD_INFO_BLOCK_SIZE:
+ blocksize = true;
+ break;
+ }
+ }
+
+ exp = nbd_export_find(name);
+ if (!exp) {
+ return nbd_negotiate_send_rep_err(client->ioc, NBD_REP_ERR_UNKNOWN,
+ opt, errp, "export '%s' not present",
+ name);
+ }
+
+ /* Don't bother sending NBD_INFO_NAME unless client requested it */
+ if (sendname) {
+ rc = nbd_negotiate_send_info(client, opt, NBD_INFO_NAME, length, name,
+ errp);
+ if (rc < 0) {
+ return rc;
+ }
+ }
+
+ /* Send NBD_INFO_DESCRIPTION only if available, regardless of
+ * client request */
+ if (exp->description) {
+ size_t len = strlen(exp->description);
+
+ rc = nbd_negotiate_send_info(client, opt, NBD_INFO_DESCRIPTION,
+ len, exp->description, errp);
+ if (rc < 0) {
+ return rc;
+ }
+ }
+
+ /* Send NBD_INFO_BLOCK_SIZE always, but tweak the minimum size
+ * according to whether the client requested it, and according to
+ * whether this is OPT_INFO or OPT_GO. */
+ /* minimum - 1 for back-compat, or 512 if client is new enough.
+ * TODO: consult blk_bs(blk)->bl.request_alignment? */
+ sizes[0] = (opt == NBD_OPT_INFO || blocksize) ? BDRV_SECTOR_SIZE : 1;
+ /* preferred - Hard-code to 4096 for now.
+ * TODO: is blk_bs(blk)->bl.opt_transfer appropriate? */
+ sizes[1] = 4096;
+ /* maximum - At most 32M, but smaller as appropriate. */
+ sizes[2] = MIN(blk_get_max_transfer(exp->blk), NBD_MAX_BUFFER_SIZE);
+ trace_nbd_negotiate_handle_info_block_size(sizes[0], sizes[1], sizes[2]);
+ cpu_to_be32s(&sizes[0]);
+ cpu_to_be32s(&sizes[1]);
+ cpu_to_be32s(&sizes[2]);
+ rc = nbd_negotiate_send_info(client, opt, NBD_INFO_BLOCK_SIZE,
+ sizeof(sizes), sizes, errp);
+ if (rc < 0) {
+ return rc;
+ }
+
+ /* Send NBD_INFO_EXPORT always */
+ trace_nbd_negotiate_new_style_size_flags(exp->size,
+ exp->nbdflags | myflags);
+ stq_be_p(buf, exp->size);
+ stw_be_p(buf + 8, exp->nbdflags | myflags);
+ rc = nbd_negotiate_send_info(client, opt, NBD_INFO_EXPORT,
+ sizeof(buf), buf, errp);
+ if (rc < 0) {
+ return rc;
+ }
+
+ /* If the client is just asking for NBD_OPT_INFO, but forgot to
+ * request block sizes, return an error.
+ * TODO: consult blk_bs(blk)->request_align, and only error if it
+ * is not 1? */
+ if (opt == NBD_OPT_INFO && !blocksize) {
+ return nbd_negotiate_send_rep_err(client->ioc,
+ NBD_REP_ERR_BLOCK_SIZE_REQD, opt,
+ errp,
+ "request NBD_INFO_BLOCK_SIZE to "
+ "use this export");
+ }
+
+ /* Final reply */
+ rc = nbd_negotiate_send_rep(client->ioc, NBD_REP_ACK, opt, errp);
+ if (rc < 0) {
+ return rc;
+ }
+
+ if (opt == NBD_OPT_GO) {
+ client->exp = exp;
+ QTAILQ_INSERT_TAIL(&client->exp->clients, client, next);
+ nbd_export_get(client->exp);
+ rc = 1;
+ }
+ return rc;
+
+ invalid:
+ if (nbd_drop(client->ioc, length, errp) < 0) {
+ return -EIO;
+ }
+ return nbd_negotiate_send_rep_err(client->ioc, NBD_REP_ERR_INVALID, opt,
+ errp, "%s", msg);
+}
+
+
/* Handle NBD_OPT_STARTTLS. Return NULL to drop connection, or else the
* new channel for all further (now-encrypted) communication. */
static QIOChannel *nbd_negotiate_handle_starttls(NBDClient *client,
- uint32_t length)
+ uint32_t length,
+ Error **errp)
{
QIOChannel *ioc;
QIOChannelTLS *tioc;
struct NBDTLSHandshakeData data = { 0 };
- TRACE("Setting up TLS");
+ trace_nbd_negotiate_handle_starttls();
ioc = client->ioc;
if (length) {
- if (nbd_drop(ioc, length, NULL) < 0) {
+ if (nbd_drop(ioc, length, errp) < 0) {
return NULL;
}
nbd_negotiate_send_rep_err(ioc, NBD_REP_ERR_INVALID, NBD_OPT_STARTTLS,
+ errp,
"OPT_STARTTLS should not have length");
return NULL;
}
if (nbd_negotiate_send_rep(client->ioc, NBD_REP_ACK,
- NBD_OPT_STARTTLS) < 0) {
+ NBD_OPT_STARTTLS, errp) < 0) {
return NULL;
}
tioc = qio_channel_tls_new_server(ioc,
client->tlscreds,
client->tlsaclname,
- NULL);
+ errp);
if (!tioc) {
return NULL;
}
qio_channel_set_name(QIO_CHANNEL(tioc), "nbd-server-tls");
- TRACE("Starting TLS handshake");
+ trace_nbd_negotiate_handle_starttls_handshake();
data.loop = g_main_loop_new(g_main_context_default(), FALSE);
qio_channel_tls_handshake(tioc,
nbd_tls_handshake,
g_main_loop_unref(data.loop);
if (data.error) {
object_unref(OBJECT(tioc));
- error_free(data.error);
+ error_propagate(errp, data.error);
return NULL;
}
return QIO_CHANNEL(tioc);
}
-
-/* Process all NBD_OPT_* client option commands.
- * Return -errno on error, 0 on success. */
-static int nbd_negotiate_options(NBDClient *client)
+/* nbd_negotiate_options
+ * Process all NBD_OPT_* client option commands, during fixed newstyle
+ * negotiation.
+ * Return:
+ * -errno on error, errp is set
+ * 0 on successful negotiation, errp is not set
+ * 1 if client sent NBD_OPT_ABORT, i.e. on valid disconnect,
+ * errp is not set
+ */
+static int nbd_negotiate_options(NBDClient *client, uint16_t myflags,
+ Error **errp)
{
uint32_t flags;
bool fixedNewstyle = false;
+ bool no_zeroes = false;
/* Client sends:
[ 0 .. 3] client flags
+ Then we loop until NBD_OPT_EXPORT_NAME or NBD_OPT_GO:
[ 0 .. 7] NBD_OPTS_MAGIC
[ 8 .. 11] NBD option
[12 .. 15] Data length
... Rest of request
*/
- if (nbd_read(client->ioc, &flags, sizeof(flags), NULL) < 0) {
- LOG("read failed");
+ if (nbd_read(client->ioc, &flags, sizeof(flags), errp) < 0) {
+ error_prepend(errp, "read failed: ");
return -EIO;
}
- TRACE("Checking client flags");
be32_to_cpus(&flags);
+ trace_nbd_negotiate_options_flags(flags);
if (flags & NBD_FLAG_C_FIXED_NEWSTYLE) {
- TRACE("Client supports fixed newstyle handshake");
fixedNewstyle = true;
flags &= ~NBD_FLAG_C_FIXED_NEWSTYLE;
}
if (flags & NBD_FLAG_C_NO_ZEROES) {
- TRACE("Client supports no zeroes at handshake end");
- client->no_zeroes = true;
+ no_zeroes = true;
flags &= ~NBD_FLAG_C_NO_ZEROES;
}
if (flags != 0) {
- TRACE("Unknown client flags 0x%" PRIx32 " received", flags);
- return -EIO;
+ error_setg(errp, "Unknown client flags 0x%" PRIx32 " received", flags);
+ return -EINVAL;
}
while (1) {
int ret;
- uint32_t clientflags, length;
+ uint32_t option, length;
uint64_t magic;
- if (nbd_read(client->ioc, &magic, sizeof(magic), NULL) < 0) {
- LOG("read failed");
+ if (nbd_read(client->ioc, &magic, sizeof(magic), errp) < 0) {
+ error_prepend(errp, "read failed: ");
return -EINVAL;
}
- TRACE("Checking opts magic");
- if (magic != be64_to_cpu(NBD_OPTS_MAGIC)) {
- LOG("Bad magic received");
+ magic = be64_to_cpu(magic);
+ trace_nbd_negotiate_options_check_magic(magic);
+ if (magic != NBD_OPTS_MAGIC) {
+ error_setg(errp, "Bad magic received");
return -EINVAL;
}
- if (nbd_read(client->ioc, &clientflags,
- sizeof(clientflags), NULL) < 0)
- {
- LOG("read failed");
+ if (nbd_read(client->ioc, &option,
+ sizeof(option), errp) < 0) {
+ error_prepend(errp, "read failed: ");
return -EINVAL;
}
- clientflags = be32_to_cpu(clientflags);
+ option = be32_to_cpu(option);
- if (nbd_read(client->ioc, &length, sizeof(length), NULL) < 0) {
- LOG("read failed");
+ if (nbd_read(client->ioc, &length, sizeof(length), errp) < 0) {
+ error_prepend(errp, "read failed: ");
return -EINVAL;
}
length = be32_to_cpu(length);
- TRACE("Checking option 0x%" PRIx32, clientflags);
+ trace_nbd_negotiate_options_check_option(option,
+ nbd_opt_lookup(option));
if (client->tlscreds &&
client->ioc == (QIOChannel *)client->sioc) {
QIOChannel *tioc;
if (!fixedNewstyle) {
- TRACE("Unsupported option 0x%" PRIx32, clientflags);
+ error_setg(errp, "Unsupported option 0x%" PRIx32, option);
return -EINVAL;
}
- switch (clientflags) {
+ switch (option) {
case NBD_OPT_STARTTLS:
- tioc = nbd_negotiate_handle_starttls(client, length);
+ tioc = nbd_negotiate_handle_starttls(client, length, errp);
if (!tioc) {
return -EIO;
}
+ ret = 0;
object_unref(OBJECT(client->ioc));
client->ioc = QIO_CHANNEL(tioc);
break;
case NBD_OPT_EXPORT_NAME:
/* No way to return an error to client, so drop connection */
- TRACE("Option 0x%x not permitted before TLS", clientflags);
+ error_setg(errp, "Option 0x%x not permitted before TLS",
+ option);
return -EINVAL;
default:
- if (nbd_drop(client->ioc, length, NULL) < 0) {
+ if (nbd_drop(client->ioc, length, errp) < 0) {
return -EIO;
}
ret = nbd_negotiate_send_rep_err(client->ioc,
NBD_REP_ERR_TLS_REQD,
- clientflags,
+ option, errp,
"Option 0x%" PRIx32
"not permitted before TLS",
- clientflags);
- if (ret < 0) {
- return ret;
- }
- /* Let the client keep trying, unless they asked to quit */
- if (clientflags == NBD_OPT_ABORT) {
- return -EINVAL;
+ option);
+ /* Let the client keep trying, unless they asked to
+ * quit. In this mode, we've already sent an error, so
+ * we can't ack the abort. */
+ if (option == NBD_OPT_ABORT) {
+ return 1;
}
break;
}
} else if (fixedNewstyle) {
- switch (clientflags) {
+ switch (option) {
case NBD_OPT_LIST:
- ret = nbd_negotiate_handle_list(client, length);
- if (ret < 0) {
- return ret;
- }
+ ret = nbd_negotiate_handle_list(client, length, errp);
break;
case NBD_OPT_ABORT:
/* NBD spec says we must try to reply before
* disconnecting, but that we must also tolerate
* guests that don't wait for our reply. */
- nbd_negotiate_send_rep(client->ioc, NBD_REP_ACK, clientflags);
- return -EINVAL;
+ nbd_negotiate_send_rep(client->ioc, NBD_REP_ACK, option, NULL);
+ return 1;
case NBD_OPT_EXPORT_NAME:
- return nbd_negotiate_handle_export_name(client, length);
+ return nbd_negotiate_handle_export_name(client, length,
+ myflags, no_zeroes,
+ errp);
+
+ case NBD_OPT_INFO:
+ case NBD_OPT_GO:
+ ret = nbd_negotiate_handle_info(client, length, option,
+ myflags, errp);
+ if (ret == 1) {
+ assert(option == NBD_OPT_GO);
+ return 0;
+ }
+ break;
case NBD_OPT_STARTTLS:
- if (nbd_drop(client->ioc, length, NULL) < 0) {
+ if (nbd_drop(client->ioc, length, errp) < 0) {
return -EIO;
}
if (client->tlscreds) {
ret = nbd_negotiate_send_rep_err(client->ioc,
NBD_REP_ERR_INVALID,
- clientflags,
+ option, errp,
"TLS already enabled");
} else {
ret = nbd_negotiate_send_rep_err(client->ioc,
NBD_REP_ERR_POLICY,
- clientflags,
+ option, errp,
"TLS not configured");
}
- if (ret < 0) {
- return ret;
- }
break;
default:
- if (nbd_drop(client->ioc, length, NULL) < 0) {
+ if (nbd_drop(client->ioc, length, errp) < 0) {
return -EIO;
}
ret = nbd_negotiate_send_rep_err(client->ioc,
NBD_REP_ERR_UNSUP,
- clientflags,
+ option, errp,
"Unsupported option 0x%"
- PRIx32,
- clientflags);
- if (ret < 0) {
- return ret;
- }
+ PRIx32 " (%s)", option,
+ nbd_opt_lookup(option));
break;
}
} else {
* If broken new-style we should drop the connection
* for anything except NBD_OPT_EXPORT_NAME
*/
- switch (clientflags) {
+ switch (option) {
case NBD_OPT_EXPORT_NAME:
- return nbd_negotiate_handle_export_name(client, length);
+ return nbd_negotiate_handle_export_name(client, length,
+ myflags, no_zeroes,
+ errp);
default:
- TRACE("Unsupported option 0x%" PRIx32, clientflags);
+ error_setg(errp, "Unsupported option 0x%" PRIx32 " (%s)",
+ option, nbd_opt_lookup(option));
return -EINVAL;
}
}
+ if (ret < 0) {
+ return ret;
+ }
}
}
-static coroutine_fn int nbd_negotiate(NBDClient *client)
+/* nbd_negotiate
+ * Return:
+ * -errno on error, errp is set
+ * 0 on successful negotiation, errp is not set
+ * 1 if client sent NBD_OPT_ABORT, i.e. on valid disconnect,
+ * errp is not set
+ */
+static coroutine_fn int nbd_negotiate(NBDClient *client, Error **errp)
{
- char buf[8 + 8 + 8 + 128];
+ char buf[NBD_OLDSTYLE_NEGOTIATE_SIZE] = "";
int ret;
const uint16_t myflags = (NBD_FLAG_HAS_FLAGS | NBD_FLAG_SEND_TRIM |
NBD_FLAG_SEND_FLUSH | NBD_FLAG_SEND_FUA |
NBD_FLAG_SEND_WRITE_ZEROES);
bool oldStyle;
- size_t len;
- /* Old style negotiation header without options
+ /* Old style negotiation header, no room for options
[ 0 .. 7] passwd ("NBDMAGIC")
[ 8 .. 15] magic (NBD_CLIENT_MAGIC)
[16 .. 23] size
- [24 .. 25] server flags (0)
- [26 .. 27] export flags
+ [24 .. 27] export flags (zero-extended)
[28 .. 151] reserved (0)
- New style negotiation header with options
+ New style negotiation header, client can send options
[ 0 .. 7] passwd ("NBDMAGIC")
[ 8 .. 15] magic (NBD_OPTS_MAGIC)
[16 .. 17] server flags (0)
- ....options sent....
- [18 .. 25] size
- [26 .. 27] export flags
- [28 .. 151] reserved (0, omit if no_zeroes)
+ ....options sent, ending in NBD_OPT_EXPORT_NAME or NBD_OPT_GO....
*/
qio_channel_set_blocking(client->ioc, false, NULL);
- TRACE("Beginning negotiation.");
- memset(buf, 0, sizeof(buf));
+ trace_nbd_negotiate_begin();
memcpy(buf, "NBDMAGIC", 8);
oldStyle = client->exp != NULL && !client->tlscreds;
if (oldStyle) {
- TRACE("advertising size %" PRIu64 " and flags %x",
- client->exp->size, client->exp->nbdflags | myflags);
+ trace_nbd_negotiate_old_style(client->exp->size,
+ client->exp->nbdflags | myflags);
stq_be_p(buf + 8, NBD_CLIENT_MAGIC);
stq_be_p(buf + 16, client->exp->size);
- stw_be_p(buf + 26, client->exp->nbdflags | myflags);
- } else {
- stq_be_p(buf + 8, NBD_OPTS_MAGIC);
- stw_be_p(buf + 16, NBD_FLAG_FIXED_NEWSTYLE | NBD_FLAG_NO_ZEROES);
- }
+ stl_be_p(buf + 24, client->exp->nbdflags | myflags);
- if (oldStyle) {
- if (client->tlscreds) {
- TRACE("TLS cannot be enabled with oldstyle protocol");
- return -EINVAL;
- }
- if (nbd_write(client->ioc, buf, sizeof(buf), NULL) < 0) {
- LOG("write failed");
+ if (nbd_write(client->ioc, buf, sizeof(buf), errp) < 0) {
+ error_prepend(errp, "write failed: ");
return -EINVAL;
}
} else {
- if (nbd_write(client->ioc, buf, 18, NULL) < 0) {
- LOG("write failed");
+ stq_be_p(buf + 8, NBD_OPTS_MAGIC);
+ stw_be_p(buf + 16, NBD_FLAG_FIXED_NEWSTYLE | NBD_FLAG_NO_ZEROES);
+
+ if (nbd_write(client->ioc, buf, 18, errp) < 0) {
+ error_prepend(errp, "write failed: ");
return -EINVAL;
}
- ret = nbd_negotiate_options(client);
+ ret = nbd_negotiate_options(client, myflags, errp);
if (ret != 0) {
- LOG("option negotiation failed");
- return ret;
- }
-
- TRACE("advertising size %" PRIu64 " and flags %x",
- client->exp->size, client->exp->nbdflags | myflags);
- stq_be_p(buf + 18, client->exp->size);
- stw_be_p(buf + 26, client->exp->nbdflags | myflags);
- len = client->no_zeroes ? 10 : sizeof(buf) - 18;
- ret = nbd_write(client->ioc, buf + 18, len, NULL);
- if (ret < 0) {
- LOG("write failed");
+ if (ret < 0) {
+ error_prepend(errp, "option negotiation failed: ");
+ }
return ret;
}
}
- TRACE("Negotiation succeeded.");
+ trace_nbd_negotiate_success();
return 0;
}
-static int nbd_receive_request(QIOChannel *ioc, NBDRequest *request)
+static int nbd_receive_request(QIOChannel *ioc, NBDRequest *request,
+ Error **errp)
{
uint8_t buf[NBD_REQUEST_SIZE];
uint32_t magic;
int ret;
- ret = nbd_read(ioc, buf, sizeof(buf), NULL);
+ ret = nbd_read(ioc, buf, sizeof(buf), errp);
if (ret < 0) {
return ret;
}
request->from = ldq_be_p(buf + 16);
request->len = ldl_be_p(buf + 24);
- TRACE("Got request: { magic = 0x%" PRIx32 ", .flags = %" PRIx16
- ", .type = %" PRIx16 ", from = %" PRIu64 ", len = %" PRIu32 " }",
- magic, request->flags, request->type, request->from, request->len);
+ trace_nbd_receive_request(magic, request->flags, request->type,
+ request->from, request->len);
if (magic != NBD_REQUEST_MAGIC) {
- LOG("invalid magic (got 0x%" PRIx32 ")", magic);
+ error_setg(errp, "invalid magic (got 0x%" PRIx32 ")", magic);
return -EINVAL;
}
return 0;
}
-static int nbd_send_reply(QIOChannel *ioc, NBDReply *reply)
-{
- uint8_t buf[NBD_REPLY_SIZE];
-
- reply->error = system_errno_to_nbd_errno(reply->error);
-
- TRACE("Sending response to client: { .error = %" PRId32
- ", handle = %" PRIu64 " }",
- reply->error, reply->handle);
-
- /* Reply
- [ 0 .. 3] magic (NBD_REPLY_MAGIC)
- [ 4 .. 7] error (0 == no error)
- [ 7 .. 15] handle
- */
- stl_be_p(buf, NBD_REPLY_MAGIC);
- stl_be_p(buf + 4, reply->error);
- stq_be_p(buf + 8, reply->handle);
-
- return nbd_write(ioc, buf, sizeof(buf), NULL);
-}
-
#define MAX_NBD_REQUESTS 16
void nbd_client_get(NBDClient *client)
NBDExport *exp = opaque;
NBDClient *client;
- TRACE("Export %s: Attaching clients to AIO context %p\n", exp->name, ctx);
+ trace_nbd_blk_aio_attached(exp->name, ctx);
exp->ctx = ctx;
NBDExport *exp = opaque;
NBDClient *client;
- TRACE("Export %s: Detaching clients from AIO context %p\n", exp->name, exp->ctx);
+ trace_nbd_blk_aio_detach(exp->name, exp->ctx);
QTAILQ_FOREACH(client, &exp->clients, next) {
qio_channel_detach_aio_context(client->ioc);
bool writethrough, BlockBackend *on_eject_blk,
Error **errp)
{
+ AioContext *ctx;
BlockBackend *blk;
- NBDExport *exp = g_malloc0(sizeof(NBDExport));
+ NBDExport *exp = g_new0(NBDExport, 1);
uint64_t perm;
int ret;
+ /*
+ * NBD exports are used for non-shared storage migration. Make sure
+ * that BDRV_O_INACTIVE is cleared and the image is ready for write
+ * access since the export could be available before migration handover.
+ */
+ ctx = bdrv_get_aio_context(bs);
+ aio_context_acquire(ctx);
+ bdrv_invalidate_cache(bs, NULL);
+ aio_context_release(ctx);
+
/* Don't allow resize while the NBD server is running, otherwise we don't
* care what happens with the node. */
perm = BLK_PERM_CONSISTENT_READ;
exp->eject_notifier.notify = nbd_eject_notifier;
blk_add_remove_bs_notifier(on_eject_blk, &exp->eject_notifier);
}
-
- /*
- * NBD exports are used for non-shared storage migration. Make sure
- * that BDRV_O_INACTIVE is cleared and the image is ready for write
- * access since the export could be available before migration handover.
- */
- aio_context_acquire(exp->ctx);
- blk_invalidate_cache(blk, NULL);
- aio_context_release(exp->ctx);
return exp;
fail:
}
}
-static int nbd_co_send_reply(NBDRequestData *req, NBDReply *reply, int len)
+static int coroutine_fn nbd_co_send_iov(NBDClient *client, struct iovec *iov,
+ unsigned niov, Error **errp)
{
- NBDClient *client = req->client;
int ret;
g_assert(qemu_in_coroutine());
qemu_co_mutex_lock(&client->send_lock);
client->send_coroutine = qemu_coroutine_self();
- if (!len) {
- ret = nbd_send_reply(client->ioc, reply);
- } else {
- qio_channel_set_cork(client->ioc, true);
- ret = nbd_send_reply(client->ioc, reply);
- if (ret == 0) {
- ret = nbd_write(client->ioc, req->data, len, NULL);
- if (ret < 0) {
- ret = -EIO;
- }
- }
- qio_channel_set_cork(client->ioc, false);
- }
+ ret = qio_channel_writev_all(client->ioc, iov, niov, errp) < 0 ? -EIO : 0;
client->send_coroutine = NULL;
qemu_co_mutex_unlock(&client->send_lock);
+
return ret;
}
+static inline void set_be_simple_reply(NBDSimpleReply *reply, uint64_t error,
+ uint64_t handle)
+{
+ stl_be_p(&reply->magic, NBD_SIMPLE_REPLY_MAGIC);
+ stl_be_p(&reply->error, error);
+ stq_be_p(&reply->handle, handle);
+}
+
+static int nbd_co_send_simple_reply(NBDClient *client,
+ uint64_t handle,
+ uint32_t error,
+ void *data,
+ size_t len,
+ Error **errp)
+{
+ NBDSimpleReply reply;
+ int nbd_err = system_errno_to_nbd_errno(error);
+ struct iovec iov[] = {
+ {.iov_base = &reply, .iov_len = sizeof(reply)},
+ {.iov_base = data, .iov_len = len}
+ };
+
+ trace_nbd_co_send_simple_reply(handle, nbd_err, nbd_err_lookup(nbd_err),
+ len);
+ set_be_simple_reply(&reply, nbd_err, handle);
+
+ return nbd_co_send_iov(client, iov, len ? 2 : 1, errp);
+}
+
/* nbd_co_receive_request
* Collect a client request. Return 0 if request looks valid, -EIO to drop
* connection right away, and any other negative value to report an error to
* the client (although the caller may still need to disconnect after reporting
* the error).
*/
-static int nbd_co_receive_request(NBDRequestData *req, NBDRequest *request)
+static int nbd_co_receive_request(NBDRequestData *req, NBDRequest *request,
+ Error **errp)
{
NBDClient *client = req->client;
g_assert(qemu_in_coroutine());
assert(client->recv_coroutine == qemu_coroutine_self());
- if (nbd_receive_request(client->ioc, request) < 0) {
+ if (nbd_receive_request(client->ioc, request, errp) < 0) {
return -EIO;
}
- TRACE("Decoding type");
+ trace_nbd_co_receive_request_decode_type(request->handle, request->type,
+ nbd_cmd_lookup(request->type));
if (request->type != NBD_CMD_WRITE) {
/* No payload, we are ready to read the next request. */
if (request->type == NBD_CMD_DISC) {
/* Special case: we're going to disconnect without a reply,
* whether or not flags, from, or len are bogus */
- TRACE("Request type is DISCONNECT");
return -EIO;
}
* checks as possible until after reading any NBD_CMD_WRITE
* payload, so we can try and keep the connection alive. */
if ((request->from + request->len) < request->from) {
- LOG("integer overflow detected, you're probably being attacked");
+ error_setg(errp,
+ "integer overflow detected, you're probably being attacked");
return -EINVAL;
}
if (request->type == NBD_CMD_READ || request->type == NBD_CMD_WRITE) {
if (request->len > NBD_MAX_BUFFER_SIZE) {
- LOG("len (%" PRIu32" ) is larger than max len (%u)",
- request->len, NBD_MAX_BUFFER_SIZE);
+ error_setg(errp, "len (%" PRIu32" ) is larger than max len (%u)",
+ request->len, NBD_MAX_BUFFER_SIZE);
return -EINVAL;
}
req->data = blk_try_blockalign(client->exp->blk, request->len);
if (req->data == NULL) {
+ error_setg(errp, "No memory");
return -ENOMEM;
}
}
if (request->type == NBD_CMD_WRITE) {
- TRACE("Reading %" PRIu32 " byte(s)", request->len);
-
- if (nbd_read(client->ioc, req->data, request->len, NULL) < 0) {
- LOG("reading from socket failed");
+ if (nbd_read(client->ioc, req->data, request->len, errp) < 0) {
+ error_prepend(errp, "reading from socket failed: ");
return -EIO;
}
req->complete = true;
+
+ trace_nbd_co_receive_request_payload_received(request->handle,
+ request->len);
}
/* Sanity checks, part 2. */
if (request->from + request->len > client->exp->size) {
- LOG("operation past EOF; From: %" PRIu64 ", Len: %" PRIu32
- ", Size: %" PRIu64, request->from, request->len,
- (uint64_t)client->exp->size);
+ error_setg(errp, "operation past EOF; From: %" PRIu64 ", Len: %" PRIu32
+ ", Size: %" PRIu64, request->from, request->len,
+ (uint64_t)client->exp->size);
return request->type == NBD_CMD_WRITE ? -ENOSPC : -EINVAL;
}
if (request->flags & ~(NBD_CMD_FLAG_FUA | NBD_CMD_FLAG_NO_HOLE)) {
- LOG("unsupported flags (got 0x%x)", request->flags);
+ error_setg(errp, "unsupported flags (got 0x%x)", request->flags);
return -EINVAL;
}
if (request->type != NBD_CMD_WRITE_ZEROES &&
(request->flags & NBD_CMD_FLAG_NO_HOLE)) {
- LOG("unexpected flags (got 0x%x)", request->flags);
+ error_setg(errp, "unexpected flags (got 0x%x)", request->flags);
return -EINVAL;
}
NBDExport *exp = client->exp;
NBDRequestData *req;
NBDRequest request = { 0 }; /* GCC thinks it can be used uninitialized */
- NBDReply reply;
int ret;
int flags;
int reply_data_len = 0;
+ Error *local_err = NULL;
- TRACE("Reading request.");
+ trace_nbd_trip();
if (client->closing) {
nbd_client_put(client);
return;
}
req = nbd_request_get(client);
- ret = nbd_co_receive_request(req, &request);
+ ret = nbd_co_receive_request(req, &request, &local_err);
client->recv_coroutine = NULL;
nbd_client_receive_next_request(client);
if (ret == -EIO) {
goto disconnect;
}
- reply.handle = request.handle;
- reply.error = 0;
-
if (ret < 0) {
- reply.error = -ret;
goto reply;
}
switch (request.type) {
case NBD_CMD_READ:
- TRACE("Request type is READ");
-
/* XXX: NBD Protocol only documents use of FUA with WRITE */
if (request.flags & NBD_CMD_FLAG_FUA) {
ret = blk_co_flush(exp->blk);
if (ret < 0) {
- LOG("flush failed");
- reply.error = -ret;
+ error_setg_errno(&local_err, -ret, "flush failed");
break;
}
}
ret = blk_pread(exp->blk, request.from + exp->dev_offset,
req->data, request.len);
if (ret < 0) {
- LOG("reading from file failed");
- reply.error = -ret;
+ error_setg_errno(&local_err, -ret, "reading from file failed");
break;
}
reply_data_len = request.len;
- TRACE("Read %" PRIu32" byte(s)", request.len);
break;
case NBD_CMD_WRITE:
- TRACE("Request type is WRITE");
-
if (exp->nbdflags & NBD_FLAG_READ_ONLY) {
- TRACE("Server is read-only, return error");
- reply.error = EROFS;
+ error_setg(&local_err, "Export is read-only");
+ ret = -EROFS;
break;
}
- TRACE("Writing to device");
-
flags = 0;
if (request.flags & NBD_CMD_FLAG_FUA) {
flags |= BDRV_REQ_FUA;
ret = blk_pwrite(exp->blk, request.from + exp->dev_offset,
req->data, request.len, flags);
if (ret < 0) {
- LOG("writing to file failed");
- reply.error = -ret;
+ error_setg_errno(&local_err, -ret, "writing to file failed");
}
break;
case NBD_CMD_WRITE_ZEROES:
- TRACE("Request type is WRITE_ZEROES");
-
if (exp->nbdflags & NBD_FLAG_READ_ONLY) {
- TRACE("Server is read-only, return error");
- reply.error = EROFS;
+ error_setg(&local_err, "Export is read-only");
+ ret = -EROFS;
break;
}
- TRACE("Writing to device");
-
flags = 0;
if (request.flags & NBD_CMD_FLAG_FUA) {
flags |= BDRV_REQ_FUA;
ret = blk_pwrite_zeroes(exp->blk, request.from + exp->dev_offset,
request.len, flags);
if (ret < 0) {
- LOG("writing to file failed");
- reply.error = -ret;
+ error_setg_errno(&local_err, -ret, "writing to file failed");
}
break;
abort();
case NBD_CMD_FLUSH:
- TRACE("Request type is FLUSH");
-
ret = blk_co_flush(exp->blk);
if (ret < 0) {
- LOG("flush failed");
- reply.error = -ret;
+ error_setg_errno(&local_err, -ret, "flush failed");
}
break;
case NBD_CMD_TRIM:
- TRACE("Request type is TRIM");
ret = blk_co_pdiscard(exp->blk, request.from + exp->dev_offset,
request.len);
if (ret < 0) {
- LOG("discard failed");
- reply.error = -ret;
+ error_setg_errno(&local_err, -ret, "discard failed");
}
break;
default:
- LOG("invalid request type (%" PRIu32 ") received", request.type);
- reply.error = EINVAL;
+ error_setg(&local_err, "invalid request type (%" PRIu32 ") received",
+ request.type);
+ ret = -EINVAL;
}
reply:
+ if (local_err) {
+ /* If we get here, local_err was not a fatal error, and should be sent
+ * to the client. */
+ error_report_err(local_err);
+ local_err = NULL;
+ }
+
+ if (nbd_co_send_simple_reply(req->client, request.handle,
+ ret < 0 ? -ret : 0,
+ req->data, reply_data_len, &local_err) < 0)
+ {
+ error_prepend(&local_err, "Failed to send reply: ");
+ goto disconnect;
+ }
+
/* We must disconnect after NBD_CMD_WRITE if we did not
* read the payload.
*/
- if (nbd_co_send_reply(req, &reply, reply_data_len) < 0 || !req->complete) {
+ if (!req->complete) {
+ error_setg(&local_err, "Request handling failed in intermediate state");
goto disconnect;
}
- TRACE("Request/Reply complete");
-
done:
nbd_request_put(req);
nbd_client_put(client);
return;
disconnect:
+ if (local_err) {
+ error_reportf_err(local_err, "Disconnect client, due to: ");
+ }
nbd_request_put(req);
client_close(client, true);
nbd_client_put(client);
{
NBDClient *client = opaque;
NBDExport *exp = client->exp;
+ Error *local_err = NULL;
if (exp) {
nbd_export_get(exp);
}
qemu_co_mutex_init(&client->send_lock);
- if (nbd_negotiate(client)) {
+ if (nbd_negotiate(client, &local_err)) {
+ if (local_err) {
+ error_report_err(local_err);
+ }
client_close(client, false);
return;
}
NBDClient *client;
Coroutine *co;
- client = g_malloc0(sizeof(NBDClient));
+ client = g_new0(NBDClient, 1);
client->refcount = 1;
client->exp = exp;
client->tlscreds = tlscreds;
projects / qemu.git / blobdiff