Spdk/lib/iscsi/iscsi_subsystem.c

1376 lines
35 KiB
C
Raw Normal View History

/*-
* BSD LICENSE
*
* Copyright (C) 2008-2012 Daisuke Aoyama <aoyama@peach.ne.jp>.
* 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/string.h"
#include "spdk/likely.h"
#include "iscsi/iscsi.h"
#include "iscsi/init_grp.h"
#include "iscsi/portal_grp.h"
#include "iscsi/conn.h"
#include "iscsi/task.h"
#include "iscsi/tgt_node.h"
#include "spdk/log.h"
struct spdk_iscsi_opts *g_spdk_iscsi_opts = NULL;
static struct spdk_thread *g_init_thread = NULL;
static spdk_iscsi_init_cb g_init_cb_fn = NULL;
static void *g_init_cb_arg = NULL;
static spdk_iscsi_fini_cb g_fini_cb_fn;
static void *g_fini_cb_arg;
#define ISCSI_DATA_BUFFER_ALIGNMENT (0x1000)
#define ISCSI_DATA_BUFFER_MASK (ISCSI_DATA_BUFFER_ALIGNMENT - 1)
static void
mobj_ctor(struct spdk_mempool *mp, __attribute__((unused)) void *arg,
void *_m, __attribute__((unused)) unsigned i)
{
struct spdk_mobj *m = _m;
m->mp = mp;
m->buf = (uint8_t *)m + sizeof(struct spdk_mobj);
m->buf = (void *)((uintptr_t)((uint8_t *)m->buf + ISCSI_DATA_BUFFER_ALIGNMENT) &
~ISCSI_DATA_BUFFER_MASK);
}
static int
iscsi_initialize_pdu_pool(void)
{
struct spdk_iscsi_globals *iscsi = &g_iscsi;
int imm_mobj_size = SPDK_BDEV_BUF_SIZE_WITH_MD(iscsi_get_max_immediate_data_size()) +
sizeof(struct spdk_mobj) + ISCSI_DATA_BUFFER_ALIGNMENT;
int dout_mobj_size = SPDK_BDEV_BUF_SIZE_WITH_MD(SPDK_ISCSI_MAX_RECV_DATA_SEGMENT_LENGTH) +
sizeof(struct spdk_mobj) + ISCSI_DATA_BUFFER_ALIGNMENT;
/* create PDU pool */
iscsi->pdu_pool = spdk_mempool_create("PDU_Pool",
iscsi->pdu_pool_size,
sizeof(struct spdk_iscsi_pdu),
256, SPDK_ENV_SOCKET_ID_ANY);
if (!iscsi->pdu_pool) {
SPDK_ERRLOG("create PDU pool failed\n");
return -1;
}
iscsi->pdu_immediate_data_pool = spdk_mempool_create_ctor("PDU_immediate_data_Pool",
iscsi->immediate_data_pool_size,
imm_mobj_size, 256,
SPDK_ENV_SOCKET_ID_ANY,
mobj_ctor, NULL);
if (!iscsi->pdu_immediate_data_pool) {
SPDK_ERRLOG("create PDU immediate data pool failed\n");
return -1;
}
iscsi->pdu_data_out_pool = spdk_mempool_create_ctor("PDU_data_out_Pool",
iscsi->data_out_pool_size,
dout_mobj_size, 256,
SPDK_ENV_SOCKET_ID_ANY,
mobj_ctor, NULL);
if (!iscsi->pdu_data_out_pool) {
SPDK_ERRLOG("create PDU data out pool failed\n");
return -1;
}
return 0;
}
static void
iscsi_sess_ctor(struct spdk_mempool *pool, void *arg, void *session_buf,
unsigned index)
{
struct spdk_iscsi_globals *iscsi = arg;
struct spdk_iscsi_sess *sess = session_buf;
iscsi->session[index] = sess;
/* tsih 0 is reserved, so start tsih values at 1. */
sess->tsih = index + 1;
}
#define DEFAULT_TASK_POOL_SIZE 32768
static int
iscsi_initialize_task_pool(void)
{
struct spdk_iscsi_globals *iscsi = &g_iscsi;
/* create scsi_task pool */
iscsi->task_pool = spdk_mempool_create("SCSI_TASK_Pool",
DEFAULT_TASK_POOL_SIZE,
sizeof(struct spdk_iscsi_task),
128, SPDK_ENV_SOCKET_ID_ANY);
if (!iscsi->task_pool) {
SPDK_ERRLOG("create task pool failed\n");
return -1;
}
return 0;
}
#define SESSION_POOL_SIZE(iscsi) (iscsi->MaxSessions)
static int
iscsi_initialize_session_pool(void)
{
struct spdk_iscsi_globals *iscsi = &g_iscsi;
iscsi->session_pool = spdk_mempool_create_ctor("Session_Pool",
SESSION_POOL_SIZE(iscsi),
sizeof(struct spdk_iscsi_sess), 0,
SPDK_ENV_SOCKET_ID_ANY,
iscsi_sess_ctor, iscsi);
if (!iscsi->session_pool) {
SPDK_ERRLOG("create session pool failed\n");
return -1;
}
return 0;
}
static int
iscsi_initialize_all_pools(void)
{
if (iscsi_initialize_pdu_pool() != 0) {
return -1;
}
if (iscsi_initialize_session_pool() != 0) {
return -1;
}
if (iscsi_initialize_task_pool() != 0) {
return -1;
}
return 0;
}
static void
iscsi_check_pool(struct spdk_mempool *pool, size_t count)
{
if (pool && spdk_mempool_count(pool) != count) {
SPDK_ERRLOG("spdk_mempool_count(%s) == %zu, should be %zu\n",
spdk_mempool_get_name(pool), spdk_mempool_count(pool), count);
}
}
static void
iscsi_check_pools(void)
{
struct spdk_iscsi_globals *iscsi = &g_iscsi;
iscsi_check_pool(iscsi->pdu_pool, iscsi->pdu_pool_size);
iscsi_check_pool(iscsi->session_pool, SESSION_POOL_SIZE(iscsi));
iscsi_check_pool(iscsi->pdu_immediate_data_pool, iscsi->immediate_data_pool_size);
iscsi_check_pool(iscsi->pdu_data_out_pool, iscsi->data_out_pool_size);
iscsi_check_pool(iscsi->task_pool, DEFAULT_TASK_POOL_SIZE);
}
static void
iscsi_free_pools(void)
{
struct spdk_iscsi_globals *iscsi = &g_iscsi;
spdk_mempool_free(iscsi->pdu_pool);
spdk_mempool_free(iscsi->session_pool);
spdk_mempool_free(iscsi->pdu_immediate_data_pool);
spdk_mempool_free(iscsi->pdu_data_out_pool);
spdk_mempool_free(iscsi->task_pool);
}
void iscsi_put_pdu(struct spdk_iscsi_pdu *pdu)
{
if (!pdu) {
return;
}
assert(pdu->ref > 0);
pdu->ref--;
if (pdu->ref == 0) {
lib/iscsi: Merge multiple Data-OUT PDUs up to 64KB in a sequence Some iSCSI initiators send a Data-OUT PDU sequence whose PDUs do not have block size multiples data. SPDK iSCSI target had replied SCSI write error to such initiators because previously we had sent a write subtask per Data-OUT PDU. SPDK SCSI library had rejected the write subtask because its data was not block size multiples. This patch fixes the issue. The idea is to aggregate multiple Data-OUT PDUs into a single write subtask up to 64KB or until F bit is set. MaxRecvDataSegmentLength is 64KB but MaxBurstLength is 1MB. Hence one Data-OUT PDU data may be split into multiple data buffers, but the maximum number of split is two. When processing the data segment of the Data-OUT PDU, save the data buffer of the current PDU to the current task if the data buffer is not full and F bit is not set. In this case, write subtask is not submitted. When processing the header of the Data-OUT PDU, if the current task saves the data buffer from the last Data-OUT PDU, it passes the data buffer to the Data-OUT PDU. When reading the data segment of the current PDU, attach the second data buffer to the current PDU if the first data buffer becomes full. These are enabled only if DIF is disabled. Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Change-Id: Ib9cfb53fe8c0807a63e58c61bed3bb52f60f4830 Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6439 Reviewed-by: Jim Harris <james.r.harris@intel.com> Reviewed-by: Changpeng Liu <changpeng.liu@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Community-CI: Mellanox Build Bot
2021-02-22 04:14:49 +00:00
if (pdu->mobj[0]) {
iscsi_datapool_put(pdu->mobj[0]);
}
if (pdu->mobj[1]) {
iscsi_datapool_put(pdu->mobj[1]);
}
if (pdu->data && !pdu->data_from_mempool) {
free(pdu->data);
}
spdk_mempool_put(g_iscsi.pdu_pool, (void *)pdu);
}
}
struct spdk_iscsi_pdu *iscsi_get_pdu(struct spdk_iscsi_conn *conn)
{
struct spdk_iscsi_pdu *pdu;
assert(conn != NULL);
pdu = spdk_mempool_get(g_iscsi.pdu_pool);
if (!pdu) {
SPDK_ERRLOG("Unable to get PDU\n");
abort();
}
/* we do not want to zero out the last part of the structure reserved for AHS and sense data */
memset(pdu, 0, offsetof(struct spdk_iscsi_pdu, ahs));
pdu->ref = 1;
pdu->conn = conn;
pdu->crc32c = SPDK_CRC32C_INITIAL;
return pdu;
}
static void
iscsi_log_globals(void)
{
SPDK_DEBUGLOG(iscsi, "AuthFile %s\n",
g_iscsi.authfile ? g_iscsi.authfile : "(none)");
SPDK_DEBUGLOG(iscsi, "NodeBase %s\n", g_iscsi.nodebase);
SPDK_DEBUGLOG(iscsi, "MaxSessions %d\n", g_iscsi.MaxSessions);
SPDK_DEBUGLOG(iscsi, "MaxConnectionsPerSession %d\n",
g_iscsi.MaxConnectionsPerSession);
SPDK_DEBUGLOG(iscsi, "MaxQueueDepth %d\n", g_iscsi.MaxQueueDepth);
SPDK_DEBUGLOG(iscsi, "DefaultTime2Wait %d\n",
g_iscsi.DefaultTime2Wait);
SPDK_DEBUGLOG(iscsi, "DefaultTime2Retain %d\n",
g_iscsi.DefaultTime2Retain);
SPDK_DEBUGLOG(iscsi, "FirstBurstLength %d\n",
g_iscsi.FirstBurstLength);
SPDK_DEBUGLOG(iscsi, "ImmediateData %s\n",
g_iscsi.ImmediateData ? "Yes" : "No");
SPDK_DEBUGLOG(iscsi, "AllowDuplicateIsid %s\n",
g_iscsi.AllowDuplicateIsid ? "Yes" : "No");
SPDK_DEBUGLOG(iscsi, "ErrorRecoveryLevel %d\n",
g_iscsi.ErrorRecoveryLevel);
SPDK_DEBUGLOG(iscsi, "Timeout %d\n", g_iscsi.timeout);
SPDK_DEBUGLOG(iscsi, "NopInInterval %d\n",
g_iscsi.nopininterval);
if (g_iscsi.disable_chap) {
SPDK_DEBUGLOG(iscsi,
"DiscoveryAuthMethod None\n");
} else if (!g_iscsi.require_chap) {
SPDK_DEBUGLOG(iscsi,
"DiscoveryAuthMethod Auto\n");
} else {
SPDK_DEBUGLOG(iscsi,
"DiscoveryAuthMethod %s %s\n",
g_iscsi.require_chap ? "CHAP" : "",
g_iscsi.mutual_chap ? "Mutual" : "");
}
if (g_iscsi.chap_group == 0) {
SPDK_DEBUGLOG(iscsi,
"DiscoveryAuthGroup None\n");
} else {
SPDK_DEBUGLOG(iscsi,
"DiscoveryAuthGroup AuthGroup%d\n",
g_iscsi.chap_group);
}
lib/iscsi: Make the max number of read subtasks for large read I/O configurable For some use case that there is heavy large read I/O, the performance bottleneck due to MAX_LARGE_DATAIN_PER_CONNECTION was reported. The following assumes that all I/Os are large read. Large read primary task whose I/O size is more than SPDK_BDEV_LARGE_BUF_MAX_SIZE (=64KB) is split into multiple read subtasks. spdk_iscsi_globals::MaxQueueDepth limits maximum number of outstanding read primary tasks, and MAX_LARGE_DATAIN_PER_CONNECTION (=64) limits maximum number of outstanding read subtasks. MAX_LARGE_DATAIN_PER_CONNECTION is also used to calculate PDU pool. To remove the performance bottleneck, change the macro constant MAX_LARGE_DATAIN_PER_CONNECTION to a global variable spdk_iscsi_globals::MaxLargeDataInPerConnection. We don't see any negative side effect if we set spdk_iscsi_globals::MaxLargeDataInPerConnection to 64. The use case that reported the performance issue will change the value of spdk_iscsi_globals::MaxLargeDataInPerConnection by its own responsibility. The next patch will add the value of spdk_iscsi_globals::MaxLargeDataInPerConnection to iSCSI options, and make it configurable by JSON RPC. Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Change-Id: Ifc30cdb8e00d50f4d3755ff399263cf5d0b681b6 Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3755 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Changpeng Liu <changpeng.liu@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2020-08-11 15:29:25 +00:00
SPDK_DEBUGLOG(iscsi, "MaxLargeDataInPerConnection %d\n",
lib/iscsi: Make the max number of read subtasks for large read I/O configurable For some use case that there is heavy large read I/O, the performance bottleneck due to MAX_LARGE_DATAIN_PER_CONNECTION was reported. The following assumes that all I/Os are large read. Large read primary task whose I/O size is more than SPDK_BDEV_LARGE_BUF_MAX_SIZE (=64KB) is split into multiple read subtasks. spdk_iscsi_globals::MaxQueueDepth limits maximum number of outstanding read primary tasks, and MAX_LARGE_DATAIN_PER_CONNECTION (=64) limits maximum number of outstanding read subtasks. MAX_LARGE_DATAIN_PER_CONNECTION is also used to calculate PDU pool. To remove the performance bottleneck, change the macro constant MAX_LARGE_DATAIN_PER_CONNECTION to a global variable spdk_iscsi_globals::MaxLargeDataInPerConnection. We don't see any negative side effect if we set spdk_iscsi_globals::MaxLargeDataInPerConnection to 64. The use case that reported the performance issue will change the value of spdk_iscsi_globals::MaxLargeDataInPerConnection by its own responsibility. The next patch will add the value of spdk_iscsi_globals::MaxLargeDataInPerConnection to iSCSI options, and make it configurable by JSON RPC. Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Change-Id: Ifc30cdb8e00d50f4d3755ff399263cf5d0b681b6 Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/3755 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Changpeng Liu <changpeng.liu@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2020-08-11 15:29:25 +00:00
g_iscsi.MaxLargeDataInPerConnection);
SPDK_DEBUGLOG(iscsi, "MaxR2TPerConnection %d\n",
g_iscsi.MaxR2TPerConnection);
}
#define NUM_PDU_PER_CONNECTION(opts) (2 * (opts->MaxQueueDepth + \
opts->MaxLargeDataInPerConnection + \
2 * opts->MaxR2TPerConnection + 8))
#define PDU_POOL_SIZE(opts) (opts->MaxSessions * NUM_PDU_PER_CONNECTION(opts))
#define IMMEDIATE_DATA_POOL_SIZE(opts) (opts->MaxSessions * 128)
#define DATA_OUT_POOL_SIZE(opts) (opts->MaxSessions * MAX_DATA_OUT_PER_CONNECTION)
static void
iscsi_opts_init(struct spdk_iscsi_opts *opts)
{
opts->MaxSessions = DEFAULT_MAX_SESSIONS;
opts->MaxConnectionsPerSession = DEFAULT_MAX_CONNECTIONS_PER_SESSION;
opts->MaxQueueDepth = DEFAULT_MAX_QUEUE_DEPTH;
opts->DefaultTime2Wait = DEFAULT_DEFAULTTIME2WAIT;
opts->DefaultTime2Retain = DEFAULT_DEFAULTTIME2RETAIN;
opts->FirstBurstLength = SPDK_ISCSI_FIRST_BURST_LENGTH;
opts->ImmediateData = DEFAULT_IMMEDIATEDATA;
opts->AllowDuplicateIsid = false;
opts->ErrorRecoveryLevel = DEFAULT_ERRORRECOVERYLEVEL;
opts->timeout = DEFAULT_TIMEOUT;
opts->nopininterval = DEFAULT_NOPININTERVAL;
opts->disable_chap = false;
opts->require_chap = false;
opts->mutual_chap = false;
opts->chap_group = 0;
opts->authfile = NULL;
opts->nodebase = NULL;
opts->MaxLargeDataInPerConnection = DEFAULT_MAX_LARGE_DATAIN_PER_CONNECTION;
opts->MaxR2TPerConnection = DEFAULT_MAXR2T;
opts->pdu_pool_size = PDU_POOL_SIZE(opts);
opts->immediate_data_pool_size = IMMEDIATE_DATA_POOL_SIZE(opts);
opts->data_out_pool_size = DATA_OUT_POOL_SIZE(opts);
}
struct spdk_iscsi_opts *
iscsi_opts_alloc(void)
{
struct spdk_iscsi_opts *opts;
opts = calloc(1, sizeof(*opts));
if (!opts) {
SPDK_ERRLOG("calloc() failed for iscsi options\n");
return NULL;
}
iscsi_opts_init(opts);
return opts;
}
void
iscsi_opts_free(struct spdk_iscsi_opts *opts)
{
free(opts->authfile);
free(opts->nodebase);
free(opts);
}
/* Deep copy of spdk_iscsi_opts */
struct spdk_iscsi_opts *
iscsi_opts_copy(struct spdk_iscsi_opts *src)
{
struct spdk_iscsi_opts *dst;
dst = calloc(1, sizeof(*dst));
if (!dst) {
SPDK_ERRLOG("calloc() failed for iscsi options\n");
return NULL;
}
if (src->authfile) {
dst->authfile = strdup(src->authfile);
if (!dst->authfile) {
free(dst);
SPDK_ERRLOG("failed to strdup for auth file %s\n", src->authfile);
return NULL;
}
}
if (src->nodebase) {
dst->nodebase = strdup(src->nodebase);
if (!dst->nodebase) {
free(dst->authfile);
free(dst);
SPDK_ERRLOG("failed to strdup for nodebase %s\n", src->nodebase);
return NULL;
}
}
dst->MaxSessions = src->MaxSessions;
dst->MaxConnectionsPerSession = src->MaxConnectionsPerSession;
dst->MaxQueueDepth = src->MaxQueueDepth;
dst->DefaultTime2Wait = src->DefaultTime2Wait;
dst->DefaultTime2Retain = src->DefaultTime2Retain;
dst->FirstBurstLength = src->FirstBurstLength;
dst->ImmediateData = src->ImmediateData;
dst->AllowDuplicateIsid = src->AllowDuplicateIsid;
dst->ErrorRecoveryLevel = src->ErrorRecoveryLevel;
dst->timeout = src->timeout;
dst->nopininterval = src->nopininterval;
dst->disable_chap = src->disable_chap;
dst->require_chap = src->require_chap;
dst->mutual_chap = src->mutual_chap;
dst->chap_group = src->chap_group;
dst->MaxLargeDataInPerConnection = src->MaxLargeDataInPerConnection;
dst->MaxR2TPerConnection = src->MaxR2TPerConnection;
dst->pdu_pool_size = src->pdu_pool_size;
dst->immediate_data_pool_size = src->immediate_data_pool_size;
dst->data_out_pool_size = src->data_out_pool_size;
return dst;
}
static int
iscsi_opts_verify(struct spdk_iscsi_opts *opts)
{
if (!opts->nodebase) {
opts->nodebase = strdup(SPDK_ISCSI_DEFAULT_NODEBASE);
if (opts->nodebase == NULL) {
SPDK_ERRLOG("strdup() failed for default nodebase\n");
return -ENOMEM;
}
}
if (opts->MaxSessions == 0 || opts->MaxSessions > 65535) {
SPDK_ERRLOG("%d is invalid. MaxSessions must be more than 0 and no more than 65535\n",
opts->MaxSessions);
return -EINVAL;
}
if (opts->MaxConnectionsPerSession == 0 || opts->MaxConnectionsPerSession > 65535) {
SPDK_ERRLOG("%d is invalid. MaxConnectionsPerSession must be more than 0 and no more than 65535\n",
opts->MaxConnectionsPerSession);
return -EINVAL;
}
if (opts->MaxQueueDepth == 0 || opts->MaxQueueDepth > 256) {
SPDK_ERRLOG("%d is invalid. MaxQueueDepth must be more than 0 and no more than 256\n",
opts->MaxQueueDepth);
return -EINVAL;
}
if (opts->DefaultTime2Wait > 3600) {
SPDK_ERRLOG("%d is invalid. DefaultTime2Wait must be no more than 3600\n",
opts->DefaultTime2Wait);
return -EINVAL;
}
if (opts->DefaultTime2Retain > 3600) {
SPDK_ERRLOG("%d is invalid. DefaultTime2Retain must be no more than 3600\n",
opts->DefaultTime2Retain);
return -EINVAL;
}
if (opts->FirstBurstLength >= SPDK_ISCSI_MIN_FIRST_BURST_LENGTH) {
if (opts->FirstBurstLength > SPDK_ISCSI_MAX_BURST_LENGTH) {
SPDK_ERRLOG("FirstBurstLength %d shall not exceed MaxBurstLength %d\n",
opts->FirstBurstLength, SPDK_ISCSI_MAX_BURST_LENGTH);
return -EINVAL;
}
} else {
SPDK_ERRLOG("FirstBurstLength %d shall be no less than %d\n",
opts->FirstBurstLength, SPDK_ISCSI_MIN_FIRST_BURST_LENGTH);
return -EINVAL;
}
if (opts->ErrorRecoveryLevel > 2) {
SPDK_ERRLOG("ErrorRecoveryLevel %d is not supported.\n", opts->ErrorRecoveryLevel);
return -EINVAL;
}
if (opts->timeout < 0) {
SPDK_ERRLOG("%d is invalid. timeout must not be less than 0\n", opts->timeout);
return -EINVAL;
}
if (opts->nopininterval < 0 || opts->nopininterval > MAX_NOPININTERVAL) {
SPDK_ERRLOG("%d is invalid. nopinterval must be between 0 and %d\n",
opts->nopininterval, MAX_NOPININTERVAL);
return -EINVAL;
}
if (!iscsi_check_chap_params(opts->disable_chap, opts->require_chap,
opts->mutual_chap, opts->chap_group)) {
SPDK_ERRLOG("CHAP params in opts are illegal combination\n");
return -EINVAL;
}
if (opts->MaxLargeDataInPerConnection == 0) {
SPDK_ERRLOG("0 is invalid. MaxLargeDataInPerConnection must be more than 0\n");
return -EINVAL;
}
if (opts->MaxR2TPerConnection == 0) {
SPDK_ERRLOG("0 is invalid. MaxR2TPerConnection must be more than 0\n");
return -EINVAL;
}
if (opts->pdu_pool_size == 0) {
SPDK_ERRLOG("0 is invalid. pdu_pool_size must be more than 0\n");
return -EINVAL;
}
if (opts->immediate_data_pool_size == 0) {
SPDK_ERRLOG("0 is invalid. immediate_data_pool_size must be more than 0\n");
return -EINVAL;
}
if (opts->data_out_pool_size == 0) {
SPDK_ERRLOG("0 is invalid. data_out_pool_size must be more than 0\n");
return -EINVAL;
}
return 0;
}
static int
iscsi_set_global_params(struct spdk_iscsi_opts *opts)
{
int rc;
rc = iscsi_opts_verify(opts);
if (rc != 0) {
SPDK_ERRLOG("spdk_iscsi_opts_verify() failed\n");
return rc;
}
if (opts->authfile != NULL) {
g_iscsi.authfile = strdup(opts->authfile);
if (!g_iscsi.authfile) {
SPDK_ERRLOG("failed to strdup for auth file %s\n", opts->authfile);
return -ENOMEM;
}
}
g_iscsi.nodebase = strdup(opts->nodebase);
if (!g_iscsi.nodebase) {
SPDK_ERRLOG("failed to strdup for nodebase %s\n", opts->nodebase);
return -ENOMEM;
}
g_iscsi.MaxSessions = opts->MaxSessions;
g_iscsi.MaxConnectionsPerSession = opts->MaxConnectionsPerSession;
g_iscsi.MaxQueueDepth = opts->MaxQueueDepth;
g_iscsi.DefaultTime2Wait = opts->DefaultTime2Wait;
g_iscsi.DefaultTime2Retain = opts->DefaultTime2Retain;
g_iscsi.FirstBurstLength = opts->FirstBurstLength;
g_iscsi.ImmediateData = opts->ImmediateData;
g_iscsi.AllowDuplicateIsid = opts->AllowDuplicateIsid;
g_iscsi.ErrorRecoveryLevel = opts->ErrorRecoveryLevel;
g_iscsi.timeout = opts->timeout;
g_iscsi.nopininterval = opts->nopininterval;
g_iscsi.disable_chap = opts->disable_chap;
g_iscsi.require_chap = opts->require_chap;
g_iscsi.mutual_chap = opts->mutual_chap;
g_iscsi.chap_group = opts->chap_group;
g_iscsi.MaxLargeDataInPerConnection = opts->MaxLargeDataInPerConnection;
g_iscsi.MaxR2TPerConnection = opts->MaxR2TPerConnection;
g_iscsi.pdu_pool_size = opts->pdu_pool_size;
g_iscsi.immediate_data_pool_size = opts->immediate_data_pool_size;
g_iscsi.data_out_pool_size = opts->data_out_pool_size;
iscsi_log_globals();
return 0;
}
int
iscsi_set_discovery_auth(bool disable_chap, bool require_chap, bool mutual_chap,
int32_t chap_group)
{
if (!iscsi_check_chap_params(disable_chap, require_chap, mutual_chap,
chap_group)) {
SPDK_ERRLOG("CHAP params are illegal combination\n");
return -EINVAL;
}
pthread_mutex_lock(&g_iscsi.mutex);
g_iscsi.disable_chap = disable_chap;
g_iscsi.require_chap = require_chap;
g_iscsi.mutual_chap = mutual_chap;
g_iscsi.chap_group = chap_group;
pthread_mutex_unlock(&g_iscsi.mutex);
return 0;
}
int
iscsi_auth_group_add_secret(struct spdk_iscsi_auth_group *group,
const char *user, const char *secret,
const char *muser, const char *msecret)
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
{
struct spdk_iscsi_auth_secret *_secret;
size_t len;
if (user == NULL || secret == NULL) {
SPDK_ERRLOG("user and secret must be specified\n");
return -EINVAL;
}
if (muser != NULL && msecret == NULL) {
SPDK_ERRLOG("msecret must be specified with muser\n");
return -EINVAL;
}
TAILQ_FOREACH(_secret, &group->secret_head, tailq) {
if (strcmp(_secret->user, user) == 0) {
SPDK_ERRLOG("user for secret is duplicated\n");
return -EEXIST;
}
}
_secret = calloc(1, sizeof(*_secret));
if (_secret == NULL) {
SPDK_ERRLOG("calloc() failed for CHAP secret\n");
return -ENOMEM;
}
len = strnlen(user, sizeof(_secret->user));
if (len > sizeof(_secret->user) - 1) {
SPDK_ERRLOG("CHAP user longer than %zu characters: %s\n",
sizeof(_secret->user) - 1, user);
free(_secret);
return -EINVAL;
}
memcpy(_secret->user, user, len);
len = strnlen(secret, sizeof(_secret->secret));
if (len > sizeof(_secret->secret) - 1) {
SPDK_ERRLOG("CHAP secret longer than %zu characters: %s\n",
sizeof(_secret->secret) - 1, secret);
free(_secret);
return -EINVAL;
}
memcpy(_secret->secret, secret, len);
if (muser != NULL) {
len = strnlen(muser, sizeof(_secret->muser));
if (len > sizeof(_secret->muser) - 1) {
SPDK_ERRLOG("Mutual CHAP user longer than %zu characters: %s\n",
sizeof(_secret->muser) - 1, muser);
free(_secret);
return -EINVAL;
}
memcpy(_secret->muser, muser, len);
len = strnlen(msecret, sizeof(_secret->msecret));
if (len > sizeof(_secret->msecret) - 1) {
SPDK_ERRLOG("Mutual CHAP secret longer than %zu characters: %s\n",
sizeof(_secret->msecret) - 1, msecret);
free(_secret);
return -EINVAL;
}
memcpy(_secret->msecret, msecret, len);
}
TAILQ_INSERT_TAIL(&group->secret_head, _secret, tailq);
return 0;
}
int
iscsi_auth_group_delete_secret(struct spdk_iscsi_auth_group *group,
const char *user)
{
struct spdk_iscsi_auth_secret *_secret;
if (user == NULL) {
SPDK_ERRLOG("user must be specified\n");
return -EINVAL;
}
TAILQ_FOREACH(_secret, &group->secret_head, tailq) {
if (strcmp(_secret->user, user) == 0) {
break;
}
}
if (_secret == NULL) {
SPDK_ERRLOG("secret is not found\n");
return -ENODEV;
}
TAILQ_REMOVE(&group->secret_head, _secret, tailq);
free(_secret);
return 0;
}
int
iscsi_add_auth_group(int32_t tag, struct spdk_iscsi_auth_group **_group)
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
{
struct spdk_iscsi_auth_group *group;
TAILQ_FOREACH(group, &g_iscsi.auth_group_head, tailq) {
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
if (group->tag == tag) {
SPDK_ERRLOG("Auth group (%d) already exists\n", tag);
return -EEXIST;
}
}
group = calloc(1, sizeof(*group));
if (group == NULL) {
SPDK_ERRLOG("calloc() failed for auth group\n");
return -ENOMEM;
}
TAILQ_INIT(&group->secret_head);
group->tag = tag;
TAILQ_INSERT_TAIL(&g_iscsi.auth_group_head, group, tailq);
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
*_group = group;
return 0;
}
void
iscsi_delete_auth_group(struct spdk_iscsi_auth_group *group)
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
{
struct spdk_iscsi_auth_secret *_secret, *tmp;
TAILQ_REMOVE(&g_iscsi.auth_group_head, group, tailq);
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
TAILQ_FOREACH_SAFE(_secret, &group->secret_head, tailq, tmp) {
TAILQ_REMOVE(&group->secret_head, _secret, tailq);
free(_secret);
}
free(group);
}
struct spdk_iscsi_auth_group *
iscsi_find_auth_group_by_tag(int32_t tag)
{
struct spdk_iscsi_auth_group *group;
TAILQ_FOREACH(group, &g_iscsi.auth_group_head, tailq) {
if (group->tag == tag) {
return group;
}
}
return NULL;
}
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
static void
iscsi_auth_groups_destroy(void)
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
{
struct spdk_iscsi_auth_group *group, *tmp;
TAILQ_FOREACH_SAFE(group, &g_iscsi.auth_group_head, tailq, tmp) {
iscsi_delete_auth_group(group);
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
}
}
static int
iscsi_parse_auth_group(struct spdk_conf_section *sp)
{
int rc;
int i;
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
int tag;
const char *val, *user, *secret, *muser, *msecret;
struct spdk_iscsi_auth_group *group = NULL;
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
val = spdk_conf_section_get_val(sp, "Comment");
if (val != NULL) {
SPDK_DEBUGLOG(iscsi, "Comment %s\n", val);
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
}
tag = spdk_conf_section_get_num(sp);
rc = iscsi_add_auth_group(tag, &group);
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
if (rc != 0) {
SPDK_ERRLOG("Failed to add auth group\n");
return rc;
}
for (i = 0; ; i++) {
val = spdk_conf_section_get_nval(sp, "Auth", i);
if (val == NULL) {
break;
}
user = spdk_conf_section_get_nmval(sp, "Auth", i, 0);
secret = spdk_conf_section_get_nmval(sp, "Auth", i, 1);
muser = spdk_conf_section_get_nmval(sp, "Auth", i, 2);
msecret = spdk_conf_section_get_nmval(sp, "Auth", i, 3);
rc = iscsi_auth_group_add_secret(group, user, secret, muser, msecret);
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
if (rc != 0) {
SPDK_ERRLOG("Failed to add secret to auth group\n");
iscsi_delete_auth_group(group);
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
return rc;
}
}
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
return 0;
}
static int
iscsi_parse_auth_info(void)
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
{
struct spdk_conf *config;
struct spdk_conf_section *sp;
int rc;
config = spdk_conf_allocate();
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
if (!config) {
SPDK_ERRLOG("Failed to allocate config file\n");
return -ENOMEM;
}
rc = spdk_conf_read(config, g_iscsi.authfile);
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
if (rc != 0) {
SPDK_INFOLOG(iscsi, "Failed to load auth file\n");
spdk_conf_free(config);
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
return rc;
}
sp = spdk_conf_first_section(config);
while (sp != NULL) {
if (spdk_conf_section_match_prefix(sp, "AuthGroup")) {
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
if (spdk_conf_section_get_num(sp) == 0) {
SPDK_ERRLOG("Group 0 is invalid\n");
iscsi_auth_groups_destroy();
spdk_conf_free(config);
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
return -EINVAL;
}
rc = iscsi_parse_auth_group(sp);
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
if (rc != 0) {
SPDK_ERRLOG("parse_auth_group() failed\n");
iscsi_auth_groups_destroy();
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
spdk_conf_free(config);
return rc;
}
}
sp = spdk_conf_next_section(sp);
}
spdk_conf_free(config);
return 0;
}
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
static struct spdk_iscsi_auth_secret *
iscsi_find_auth_secret(const char *authuser, int ag_tag)
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
{
struct spdk_iscsi_auth_group *group;
struct spdk_iscsi_auth_secret *_secret;
TAILQ_FOREACH(group, &g_iscsi.auth_group_head, tailq) {
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
if (group->tag == ag_tag) {
TAILQ_FOREACH(_secret, &group->secret_head, tailq) {
if (strcmp(_secret->user, authuser) == 0) {
return _secret;
}
}
}
}
return NULL;
}
int
iscsi_chap_get_authinfo(struct iscsi_chap_auth *auth, const char *authuser,
int ag_tag)
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
{
struct spdk_iscsi_auth_secret *_secret;
if (authuser == NULL) {
return -EINVAL;
}
if (auth->user[0] != '\0') {
memset(auth->user, 0, sizeof(auth->user));
memset(auth->secret, 0, sizeof(auth->secret));
memset(auth->muser, 0, sizeof(auth->muser));
memset(auth->msecret, 0, sizeof(auth->msecret));
}
pthread_mutex_lock(&g_iscsi.mutex);
_secret = iscsi_find_auth_secret(authuser, ag_tag);
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
if (_secret == NULL) {
pthread_mutex_unlock(&g_iscsi.mutex);
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
SPDK_ERRLOG("CHAP secret is not found: user:%s, tag:%d\n",
authuser, ag_tag);
return -ENOENT;
}
memcpy(auth->user, _secret->user, sizeof(auth->user));
memcpy(auth->secret, _secret->secret, sizeof(auth->secret));
if (_secret->muser[0] != '\0') {
memcpy(auth->muser, _secret->muser, sizeof(auth->muser));
memcpy(auth->msecret, _secret->msecret, sizeof(auth->msecret));
}
pthread_mutex_unlock(&g_iscsi.mutex);
iscsi: Load CHAP secrets from file once at boot and use them in memory For secure iSCSI targets, dynamic reconfiguration of CHAP secrets is must to have. Currently CHAP secrets are loaded for every CHAP authentication operation. The current implementation will not work correctly when CHAP secrets in the file are changed dynamically. If SPDK loads CHAP secrets from the file only at boot and they can be configured by RPCs, user can change CHAP secrets safely during run time. Even if there are any users who expect dynamic reconfiguration of CHAP secrets based on the current implementation, if we provide this better alternative based on RPCs, they will be able to continue to satisfy their requirement. This patch change the current implementation so that SPDK loads CHAP secrets from the file once at boot and uses them in memory hereafter. Besides, use fixed size buffers to hold CHAP secrets. Previously dynamically allocated buffers by strdup() had been used, but it required many nomem checks and should be avoided. Other iSCSI targets/initiators have used fixed size buffers and SPDK follows others. Set the size of buffers for both user names and secrets to 256 (the last byte is for NULL termination). 256 is sufficiently large compared with others. CHANGELOG will be updated in the separate patch because new RPCs will be added instead. Change-Id: I499e792817c2ed01c3d970bbd3d34a6b1fccf65b Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/421463 Chandler-Test-Pool: SPDK Automated Test System <sys_sgsw@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2018-08-20 01:32:30 +00:00
return 0;
}
static int
iscsi_initialize_global_params(void)
{
int rc;
if (!g_spdk_iscsi_opts) {
g_spdk_iscsi_opts = iscsi_opts_alloc();
if (!g_spdk_iscsi_opts) {
SPDK_ERRLOG("iscsi_opts_alloc_failed() failed\n");
return -ENOMEM;
}
}
rc = iscsi_set_global_params(g_spdk_iscsi_opts);
if (rc != 0) {
SPDK_ERRLOG("iscsi_set_global_params() failed\n");
}
iscsi_opts_free(g_spdk_iscsi_opts);
g_spdk_iscsi_opts = NULL;
return rc;
}
static void
iscsi_init_complete(int rc)
{
spdk_iscsi_init_cb cb_fn = g_init_cb_fn;
void *cb_arg = g_init_cb_arg;
g_init_cb_fn = NULL;
g_init_cb_arg = NULL;
cb_fn(cb_arg, rc);
}
static void
iscsi_parse_configuration(void)
{
int rc = 0;
if (g_iscsi.authfile != NULL) {
if (access(g_iscsi.authfile, R_OK) == 0) {
rc = iscsi_parse_auth_info();
if (rc < 0) {
SPDK_ERRLOG("iscsi_parse_auth_info() failed\n");
}
} else {
SPDK_INFOLOG(iscsi, "CHAP secret file is not found in the path %s\n",
g_iscsi.authfile);
}
}
iscsi_init_complete(rc);
}
static int
iscsi_poll_group_poll(void *ctx)
{
struct spdk_iscsi_poll_group *group = ctx;
struct spdk_iscsi_conn *conn, *tmp;
int rc;
if (spdk_unlikely(STAILQ_EMPTY(&group->connections))) {
return SPDK_POLLER_IDLE;
}
rc = spdk_sock_group_poll(group->sock_group);
if (rc < 0) {
SPDK_ERRLOG("Failed to poll sock_group=%p\n", group->sock_group);
}
STAILQ_FOREACH_SAFE(conn, &group->connections, pg_link, tmp) {
if (conn->state == ISCSI_CONN_STATE_EXITING) {
iscsi_conn_destruct(conn);
}
}
return rc != 0 ? SPDK_POLLER_BUSY : SPDK_POLLER_IDLE;
}
static int
iscsi_poll_group_handle_nop(void *ctx)
{
struct spdk_iscsi_poll_group *group = ctx;
struct spdk_iscsi_conn *conn, *tmp;
STAILQ_FOREACH_SAFE(conn, &group->connections, pg_link, tmp) {
iscsi_conn_handle_nop(conn);
}
return SPDK_POLLER_BUSY;
}
iscsi: Assign connections to poll groups instead of lcores. This patch binds poll groups to SPDK thread through IO channel and assigns connections to poll groups instead of cores. iSCSI subsystem registers iSCSI global object as an IO device, and create poll groups as context of IO channels of the IO device. Each portal get and hold portal group on which the corresponding acceptor is running. When a connection is constructed, iSCSI subsystem assigns a poll group to the connection by getting it from the corresponding portal. When a connection enters full-feature phase, iSCSI subsystem schedules the connection to a poll group by round-robin. Then, each connection can know its running SPDK thread directly and can use SPDK message passing infrastructure instead of SPDK event framework. By this change, iSCSI connections are binded to SPDK thread, and not binded to processor core anymore. Some other changes in this patch are - core ID is removed from the output of get_iscsi_connections RPC. The upcoming patches will change the RPC to use spdk_for_each_channel and can access SPDK thread safely, and add SPDK thread ID instead. - utilize UT multithread framework added by the last patch to test iSCSI poll groups by UT. Change-Id: Iec73c778aa413bcabdb63141cc41d4160911ea0e Signed-off-by: Ben Walker <benjamin.walker@intel.com> Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/463359 Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-07-30 03:00:15 +00:00
static int
iscsi_poll_group_create(void *io_device, void *ctx_buf)
{
iscsi: Assign connections to poll groups instead of lcores. This patch binds poll groups to SPDK thread through IO channel and assigns connections to poll groups instead of cores. iSCSI subsystem registers iSCSI global object as an IO device, and create poll groups as context of IO channels of the IO device. Each portal get and hold portal group on which the corresponding acceptor is running. When a connection is constructed, iSCSI subsystem assigns a poll group to the connection by getting it from the corresponding portal. When a connection enters full-feature phase, iSCSI subsystem schedules the connection to a poll group by round-robin. Then, each connection can know its running SPDK thread directly and can use SPDK message passing infrastructure instead of SPDK event framework. By this change, iSCSI connections are binded to SPDK thread, and not binded to processor core anymore. Some other changes in this patch are - core ID is removed from the output of get_iscsi_connections RPC. The upcoming patches will change the RPC to use spdk_for_each_channel and can access SPDK thread safely, and add SPDK thread ID instead. - utilize UT multithread framework added by the last patch to test iSCSI poll groups by UT. Change-Id: Iec73c778aa413bcabdb63141cc41d4160911ea0e Signed-off-by: Ben Walker <benjamin.walker@intel.com> Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/463359 Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-07-30 03:00:15 +00:00
struct spdk_iscsi_poll_group *pg = ctx_buf;
STAILQ_INIT(&pg->connections);
pg->sock_group = spdk_sock_group_create(NULL);
assert(pg->sock_group != NULL);
pg->poller = SPDK_POLLER_REGISTER(iscsi_poll_group_poll, pg, 0);
/* set the period to 1 sec */
pg->nop_poller = SPDK_POLLER_REGISTER(iscsi_poll_group_handle_nop, pg, 1000000);
iscsi: Assign connections to poll groups instead of lcores. This patch binds poll groups to SPDK thread through IO channel and assigns connections to poll groups instead of cores. iSCSI subsystem registers iSCSI global object as an IO device, and create poll groups as context of IO channels of the IO device. Each portal get and hold portal group on which the corresponding acceptor is running. When a connection is constructed, iSCSI subsystem assigns a poll group to the connection by getting it from the corresponding portal. When a connection enters full-feature phase, iSCSI subsystem schedules the connection to a poll group by round-robin. Then, each connection can know its running SPDK thread directly and can use SPDK message passing infrastructure instead of SPDK event framework. By this change, iSCSI connections are binded to SPDK thread, and not binded to processor core anymore. Some other changes in this patch are - core ID is removed from the output of get_iscsi_connections RPC. The upcoming patches will change the RPC to use spdk_for_each_channel and can access SPDK thread safely, and add SPDK thread ID instead. - utilize UT multithread framework added by the last patch to test iSCSI poll groups by UT. Change-Id: Iec73c778aa413bcabdb63141cc41d4160911ea0e Signed-off-by: Ben Walker <benjamin.walker@intel.com> Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/463359 Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-07-30 03:00:15 +00:00
return 0;
}
static void
iscsi: Assign connections to poll groups instead of lcores. This patch binds poll groups to SPDK thread through IO channel and assigns connections to poll groups instead of cores. iSCSI subsystem registers iSCSI global object as an IO device, and create poll groups as context of IO channels of the IO device. Each portal get and hold portal group on which the corresponding acceptor is running. When a connection is constructed, iSCSI subsystem assigns a poll group to the connection by getting it from the corresponding portal. When a connection enters full-feature phase, iSCSI subsystem schedules the connection to a poll group by round-robin. Then, each connection can know its running SPDK thread directly and can use SPDK message passing infrastructure instead of SPDK event framework. By this change, iSCSI connections are binded to SPDK thread, and not binded to processor core anymore. Some other changes in this patch are - core ID is removed from the output of get_iscsi_connections RPC. The upcoming patches will change the RPC to use spdk_for_each_channel and can access SPDK thread safely, and add SPDK thread ID instead. - utilize UT multithread framework added by the last patch to test iSCSI poll groups by UT. Change-Id: Iec73c778aa413bcabdb63141cc41d4160911ea0e Signed-off-by: Ben Walker <benjamin.walker@intel.com> Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/463359 Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-07-30 03:00:15 +00:00
iscsi_poll_group_destroy(void *io_device, void *ctx_buf)
{
iscsi: Assign connections to poll groups instead of lcores. This patch binds poll groups to SPDK thread through IO channel and assigns connections to poll groups instead of cores. iSCSI subsystem registers iSCSI global object as an IO device, and create poll groups as context of IO channels of the IO device. Each portal get and hold portal group on which the corresponding acceptor is running. When a connection is constructed, iSCSI subsystem assigns a poll group to the connection by getting it from the corresponding portal. When a connection enters full-feature phase, iSCSI subsystem schedules the connection to a poll group by round-robin. Then, each connection can know its running SPDK thread directly and can use SPDK message passing infrastructure instead of SPDK event framework. By this change, iSCSI connections are binded to SPDK thread, and not binded to processor core anymore. Some other changes in this patch are - core ID is removed from the output of get_iscsi_connections RPC. The upcoming patches will change the RPC to use spdk_for_each_channel and can access SPDK thread safely, and add SPDK thread ID instead. - utilize UT multithread framework added by the last patch to test iSCSI poll groups by UT. Change-Id: Iec73c778aa413bcabdb63141cc41d4160911ea0e Signed-off-by: Ben Walker <benjamin.walker@intel.com> Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/463359 Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-07-30 03:00:15 +00:00
struct spdk_iscsi_poll_group *pg = ctx_buf;
struct spdk_io_channel *ch;
struct spdk_thread *thread;
assert(pg->poller != NULL);
assert(pg->sock_group != NULL);
spdk_sock_group_close(&pg->sock_group);
spdk_poller_unregister(&pg->poller);
spdk_poller_unregister(&pg->nop_poller);
ch = spdk_io_channel_from_ctx(pg);
thread = spdk_io_channel_get_thread(ch);
assert(thread == spdk_get_thread());
spdk_thread_exit(thread);
}
static void
_iscsi_init_thread_done(void *ctx)
{
struct spdk_iscsi_poll_group *pg = ctx;
TAILQ_INSERT_TAIL(&g_iscsi.poll_group_head, pg, link);
if (--g_iscsi.refcnt == 0) {
iscsi_parse_configuration();
}
}
static void
iscsi: Assign connections to poll groups instead of lcores. This patch binds poll groups to SPDK thread through IO channel and assigns connections to poll groups instead of cores. iSCSI subsystem registers iSCSI global object as an IO device, and create poll groups as context of IO channels of the IO device. Each portal get and hold portal group on which the corresponding acceptor is running. When a connection is constructed, iSCSI subsystem assigns a poll group to the connection by getting it from the corresponding portal. When a connection enters full-feature phase, iSCSI subsystem schedules the connection to a poll group by round-robin. Then, each connection can know its running SPDK thread directly and can use SPDK message passing infrastructure instead of SPDK event framework. By this change, iSCSI connections are binded to SPDK thread, and not binded to processor core anymore. Some other changes in this patch are - core ID is removed from the output of get_iscsi_connections RPC. The upcoming patches will change the RPC to use spdk_for_each_channel and can access SPDK thread safely, and add SPDK thread ID instead. - utilize UT multithread framework added by the last patch to test iSCSI poll groups by UT. Change-Id: Iec73c778aa413bcabdb63141cc41d4160911ea0e Signed-off-by: Ben Walker <benjamin.walker@intel.com> Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/463359 Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-07-30 03:00:15 +00:00
_iscsi_init_thread(void *ctx)
{
iscsi: Assign connections to poll groups instead of lcores. This patch binds poll groups to SPDK thread through IO channel and assigns connections to poll groups instead of cores. iSCSI subsystem registers iSCSI global object as an IO device, and create poll groups as context of IO channels of the IO device. Each portal get and hold portal group on which the corresponding acceptor is running. When a connection is constructed, iSCSI subsystem assigns a poll group to the connection by getting it from the corresponding portal. When a connection enters full-feature phase, iSCSI subsystem schedules the connection to a poll group by round-robin. Then, each connection can know its running SPDK thread directly and can use SPDK message passing infrastructure instead of SPDK event framework. By this change, iSCSI connections are binded to SPDK thread, and not binded to processor core anymore. Some other changes in this patch are - core ID is removed from the output of get_iscsi_connections RPC. The upcoming patches will change the RPC to use spdk_for_each_channel and can access SPDK thread safely, and add SPDK thread ID instead. - utilize UT multithread framework added by the last patch to test iSCSI poll groups by UT. Change-Id: Iec73c778aa413bcabdb63141cc41d4160911ea0e Signed-off-by: Ben Walker <benjamin.walker@intel.com> Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/463359 Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-07-30 03:00:15 +00:00
struct spdk_io_channel *ch;
struct spdk_iscsi_poll_group *pg;
ch = spdk_get_io_channel(&g_iscsi);
iscsi: Assign connections to poll groups instead of lcores. This patch binds poll groups to SPDK thread through IO channel and assigns connections to poll groups instead of cores. iSCSI subsystem registers iSCSI global object as an IO device, and create poll groups as context of IO channels of the IO device. Each portal get and hold portal group on which the corresponding acceptor is running. When a connection is constructed, iSCSI subsystem assigns a poll group to the connection by getting it from the corresponding portal. When a connection enters full-feature phase, iSCSI subsystem schedules the connection to a poll group by round-robin. Then, each connection can know its running SPDK thread directly and can use SPDK message passing infrastructure instead of SPDK event framework. By this change, iSCSI connections are binded to SPDK thread, and not binded to processor core anymore. Some other changes in this patch are - core ID is removed from the output of get_iscsi_connections RPC. The upcoming patches will change the RPC to use spdk_for_each_channel and can access SPDK thread safely, and add SPDK thread ID instead. - utilize UT multithread framework added by the last patch to test iSCSI poll groups by UT. Change-Id: Iec73c778aa413bcabdb63141cc41d4160911ea0e Signed-off-by: Ben Walker <benjamin.walker@intel.com> Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/463359 Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-07-30 03:00:15 +00:00
pg = spdk_io_channel_get_ctx(ch);
spdk_thread_send_msg(g_init_thread, _iscsi_init_thread_done, pg);
iscsi: Assign connections to poll groups instead of lcores. This patch binds poll groups to SPDK thread through IO channel and assigns connections to poll groups instead of cores. iSCSI subsystem registers iSCSI global object as an IO device, and create poll groups as context of IO channels of the IO device. Each portal get and hold portal group on which the corresponding acceptor is running. When a connection is constructed, iSCSI subsystem assigns a poll group to the connection by getting it from the corresponding portal. When a connection enters full-feature phase, iSCSI subsystem schedules the connection to a poll group by round-robin. Then, each connection can know its running SPDK thread directly and can use SPDK message passing infrastructure instead of SPDK event framework. By this change, iSCSI connections are binded to SPDK thread, and not binded to processor core anymore. Some other changes in this patch are - core ID is removed from the output of get_iscsi_connections RPC. The upcoming patches will change the RPC to use spdk_for_each_channel and can access SPDK thread safely, and add SPDK thread ID instead. - utilize UT multithread framework added by the last patch to test iSCSI poll groups by UT. Change-Id: Iec73c778aa413bcabdb63141cc41d4160911ea0e Signed-off-by: Ben Walker <benjamin.walker@intel.com> Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/463359 Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-07-30 03:00:15 +00:00
}
static void
initialize_iscsi_poll_group(void)
iscsi: Assign connections to poll groups instead of lcores. This patch binds poll groups to SPDK thread through IO channel and assigns connections to poll groups instead of cores. iSCSI subsystem registers iSCSI global object as an IO device, and create poll groups as context of IO channels of the IO device. Each portal get and hold portal group on which the corresponding acceptor is running. When a connection is constructed, iSCSI subsystem assigns a poll group to the connection by getting it from the corresponding portal. When a connection enters full-feature phase, iSCSI subsystem schedules the connection to a poll group by round-robin. Then, each connection can know its running SPDK thread directly and can use SPDK message passing infrastructure instead of SPDK event framework. By this change, iSCSI connections are binded to SPDK thread, and not binded to processor core anymore. Some other changes in this patch are - core ID is removed from the output of get_iscsi_connections RPC. The upcoming patches will change the RPC to use spdk_for_each_channel and can access SPDK thread safely, and add SPDK thread ID instead. - utilize UT multithread framework added by the last patch to test iSCSI poll groups by UT. Change-Id: Iec73c778aa413bcabdb63141cc41d4160911ea0e Signed-off-by: Ben Walker <benjamin.walker@intel.com> Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/463359 Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-07-30 03:00:15 +00:00
{
struct spdk_cpuset tmp_cpumask = {};
uint32_t i;
char thread_name[32];
struct spdk_thread *thread;
spdk_io_device_register(&g_iscsi, iscsi_poll_group_create, iscsi_poll_group_destroy,
iscsi: Assign connections to poll groups instead of lcores. This patch binds poll groups to SPDK thread through IO channel and assigns connections to poll groups instead of cores. iSCSI subsystem registers iSCSI global object as an IO device, and create poll groups as context of IO channels of the IO device. Each portal get and hold portal group on which the corresponding acceptor is running. When a connection is constructed, iSCSI subsystem assigns a poll group to the connection by getting it from the corresponding portal. When a connection enters full-feature phase, iSCSI subsystem schedules the connection to a poll group by round-robin. Then, each connection can know its running SPDK thread directly and can use SPDK message passing infrastructure instead of SPDK event framework. By this change, iSCSI connections are binded to SPDK thread, and not binded to processor core anymore. Some other changes in this patch are - core ID is removed from the output of get_iscsi_connections RPC. The upcoming patches will change the RPC to use spdk_for_each_channel and can access SPDK thread safely, and add SPDK thread ID instead. - utilize UT multithread framework added by the last patch to test iSCSI poll groups by UT. Change-Id: Iec73c778aa413bcabdb63141cc41d4160911ea0e Signed-off-by: Ben Walker <benjamin.walker@intel.com> Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/463359 Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-07-30 03:00:15 +00:00
sizeof(struct spdk_iscsi_poll_group), "iscsi_tgt");
/* Create threads for CPU cores active for this application, and send a
* message to each thread to create a poll group on it.
*/
g_init_thread = spdk_get_thread();
assert(g_init_thread != NULL);
assert(g_iscsi.refcnt == 0);
SPDK_ENV_FOREACH_CORE(i) {
spdk_cpuset_zero(&tmp_cpumask);
spdk_cpuset_set_cpu(&tmp_cpumask, i, true);
snprintf(thread_name, sizeof(thread_name), "iscsi_poll_group_%u", i);
thread = spdk_thread_create(thread_name, &tmp_cpumask);
assert(thread != NULL);
g_iscsi.refcnt++;
spdk_thread_send_msg(thread, _iscsi_init_thread, NULL);
}
}
static int
iscsi_parse_globals(void)
{
int rc;
rc = iscsi_initialize_global_params();
if (rc != 0) {
SPDK_ERRLOG("iscsi_initialize_iscsi_global_params() failed\n");
return rc;
}
g_iscsi.session = calloc(1, sizeof(struct spdk_iscsi_sess *) * g_iscsi.MaxSessions);
if (!g_iscsi.session) {
SPDK_ERRLOG("calloc() failed for session array\n");
return -1;
}
/*
* For now, just support same number of total connections, rather
* than MaxSessions * MaxConnectionsPerSession. After we add better
* handling for low resource conditions from our various buffer
* pools, we can bump this up to support more connections.
*/
g_iscsi.MaxConnections = g_iscsi.MaxSessions;
rc = iscsi_initialize_all_pools();
if (rc != 0) {
SPDK_ERRLOG("initialize_all_pools() failed\n");
free(g_iscsi.session);
g_iscsi.session = NULL;
return -1;
}
rc = initialize_iscsi_conns();
if (rc < 0) {
SPDK_ERRLOG("initialize_iscsi_conns() failed\n");
free(g_iscsi.session);
g_iscsi.session = NULL;
return rc;
}
initialize_iscsi_poll_group();
return 0;
}
void
spdk_iscsi_init(spdk_iscsi_init_cb cb_fn, void *cb_arg)
{
int rc;
assert(cb_fn != NULL);
g_init_cb_fn = cb_fn;
g_init_cb_arg = cb_arg;
rc = iscsi_parse_globals();
if (rc < 0) {
SPDK_ERRLOG("iscsi_parse_globals() failed\n");
iscsi_init_complete(-1);
}
/*
* iscsi_parse_configuration() will be called as the callback to
* spdk_initialize_iscsi_poll_group() and will complete iSCSI
* subsystem initialization.
*/
}
void
spdk_iscsi_fini(spdk_iscsi_fini_cb cb_fn, void *cb_arg)
{
g_fini_cb_fn = cb_fn;
g_fini_cb_arg = cb_arg;
iscsi_portal_grp_close_all();
shutdown_iscsi_conns();
}
static void
iscsi_fini_done(void *io_device)
{
g_fini_cb_fn(g_fini_cb_arg);
}
static void
_iscsi_fini_dev_unreg(struct spdk_io_channel_iter *i, int status)
{
iscsi_check_pools();
iscsi_free_pools();
free(g_iscsi.session);
assert(TAILQ_EMPTY(&g_iscsi.poll_group_head));
iscsi: Assign connections to poll groups instead of lcores. This patch binds poll groups to SPDK thread through IO channel and assigns connections to poll groups instead of cores. iSCSI subsystem registers iSCSI global object as an IO device, and create poll groups as context of IO channels of the IO device. Each portal get and hold portal group on which the corresponding acceptor is running. When a connection is constructed, iSCSI subsystem assigns a poll group to the connection by getting it from the corresponding portal. When a connection enters full-feature phase, iSCSI subsystem schedules the connection to a poll group by round-robin. Then, each connection can know its running SPDK thread directly and can use SPDK message passing infrastructure instead of SPDK event framework. By this change, iSCSI connections are binded to SPDK thread, and not binded to processor core anymore. Some other changes in this patch are - core ID is removed from the output of get_iscsi_connections RPC. The upcoming patches will change the RPC to use spdk_for_each_channel and can access SPDK thread safely, and add SPDK thread ID instead. - utilize UT multithread framework added by the last patch to test iSCSI poll groups by UT. Change-Id: Iec73c778aa413bcabdb63141cc41d4160911ea0e Signed-off-by: Ben Walker <benjamin.walker@intel.com> Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/463359 Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-07-30 03:00:15 +00:00
iscsi_shutdown_tgt_nodes();
iscsi_init_grps_destroy();
iscsi_portal_grps_destroy();
iscsi_auth_groups_destroy();
free(g_iscsi.authfile);
free(g_iscsi.nodebase);
pthread_mutex_destroy(&g_iscsi.mutex);
if (g_init_thread != NULL) {
/* g_init_thread is set just after the io_device is
* registered, so we can use it to determine if it
* needs to be unregistered (in cases where iscsi init
* fails).
*/
spdk_io_device_unregister(&g_iscsi, iscsi_fini_done);
} else {
iscsi_fini_done(NULL);
}
}
iscsi: Assign connections to poll groups instead of lcores. This patch binds poll groups to SPDK thread through IO channel and assigns connections to poll groups instead of cores. iSCSI subsystem registers iSCSI global object as an IO device, and create poll groups as context of IO channels of the IO device. Each portal get and hold portal group on which the corresponding acceptor is running. When a connection is constructed, iSCSI subsystem assigns a poll group to the connection by getting it from the corresponding portal. When a connection enters full-feature phase, iSCSI subsystem schedules the connection to a poll group by round-robin. Then, each connection can know its running SPDK thread directly and can use SPDK message passing infrastructure instead of SPDK event framework. By this change, iSCSI connections are binded to SPDK thread, and not binded to processor core anymore. Some other changes in this patch are - core ID is removed from the output of get_iscsi_connections RPC. The upcoming patches will change the RPC to use spdk_for_each_channel and can access SPDK thread safely, and add SPDK thread ID instead. - utilize UT multithread framework added by the last patch to test iSCSI poll groups by UT. Change-Id: Iec73c778aa413bcabdb63141cc41d4160911ea0e Signed-off-by: Ben Walker <benjamin.walker@intel.com> Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/463359 Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-07-30 03:00:15 +00:00
static void
_iscsi_fini_thread(struct spdk_io_channel_iter *i)
{
struct spdk_io_channel *ch;
struct spdk_iscsi_poll_group *pg;
ch = spdk_io_channel_iter_get_channel(i);
pg = spdk_io_channel_get_ctx(ch);
pthread_mutex_lock(&g_iscsi.mutex);
TAILQ_REMOVE(&g_iscsi.poll_group_head, pg, link);
pthread_mutex_unlock(&g_iscsi.mutex);
iscsi: Assign connections to poll groups instead of lcores. This patch binds poll groups to SPDK thread through IO channel and assigns connections to poll groups instead of cores. iSCSI subsystem registers iSCSI global object as an IO device, and create poll groups as context of IO channels of the IO device. Each portal get and hold portal group on which the corresponding acceptor is running. When a connection is constructed, iSCSI subsystem assigns a poll group to the connection by getting it from the corresponding portal. When a connection enters full-feature phase, iSCSI subsystem schedules the connection to a poll group by round-robin. Then, each connection can know its running SPDK thread directly and can use SPDK message passing infrastructure instead of SPDK event framework. By this change, iSCSI connections are binded to SPDK thread, and not binded to processor core anymore. Some other changes in this patch are - core ID is removed from the output of get_iscsi_connections RPC. The upcoming patches will change the RPC to use spdk_for_each_channel and can access SPDK thread safely, and add SPDK thread ID instead. - utilize UT multithread framework added by the last patch to test iSCSI poll groups by UT. Change-Id: Iec73c778aa413bcabdb63141cc41d4160911ea0e Signed-off-by: Ben Walker <benjamin.walker@intel.com> Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/463359 Reviewed-by: Broadcom SPDK FC-NVMe CI <spdk-ci.pdl@broadcom.com> Reviewed-by: Jim Harris <james.r.harris@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-07-30 03:00:15 +00:00
spdk_put_io_channel(ch);
spdk_for_each_channel_continue(i, 0);
}
void
shutdown_iscsi_conns_done(void)
{
spdk_for_each_channel(&g_iscsi, _iscsi_fini_thread, NULL, _iscsi_fini_dev_unreg);
}
void
iscsi_opts_info_json(struct spdk_json_write_ctx *w)
{
spdk_json_write_object_begin(w);
if (g_iscsi.authfile != NULL) {
spdk_json_write_named_string(w, "auth_file", g_iscsi.authfile);
}
spdk_json_write_named_string(w, "node_base", g_iscsi.nodebase);
spdk_json_write_named_uint32(w, "max_sessions", g_iscsi.MaxSessions);
spdk_json_write_named_uint32(w, "max_connections_per_session",
g_iscsi.MaxConnectionsPerSession);
spdk_json_write_named_uint32(w, "max_queue_depth", g_iscsi.MaxQueueDepth);
spdk_json_write_named_uint32(w, "default_time2wait", g_iscsi.DefaultTime2Wait);
spdk_json_write_named_uint32(w, "default_time2retain", g_iscsi.DefaultTime2Retain);
spdk_json_write_named_uint32(w, "first_burst_length", g_iscsi.FirstBurstLength);
spdk_json_write_named_bool(w, "immediate_data", g_iscsi.ImmediateData);
spdk_json_write_named_bool(w, "allow_duplicated_isid", g_iscsi.AllowDuplicateIsid);
spdk_json_write_named_uint32(w, "error_recovery_level", g_iscsi.ErrorRecoveryLevel);
spdk_json_write_named_int32(w, "nop_timeout", g_iscsi.timeout);
spdk_json_write_named_int32(w, "nop_in_interval", g_iscsi.nopininterval);
spdk_json_write_named_bool(w, "disable_chap", g_iscsi.disable_chap);
spdk_json_write_named_bool(w, "require_chap", g_iscsi.require_chap);
spdk_json_write_named_bool(w, "mutual_chap", g_iscsi.mutual_chap);
spdk_json_write_named_int32(w, "chap_group", g_iscsi.chap_group);
spdk_json_write_named_uint32(w, "max_large_datain_per_connection",
g_iscsi.MaxLargeDataInPerConnection);
spdk_json_write_named_uint32(w, "max_r2t_per_connection",
g_iscsi.MaxR2TPerConnection);
spdk_json_write_named_uint32(w, "pdu_pool_size", g_iscsi.pdu_pool_size);
spdk_json_write_named_uint32(w, "immediate_data_pool_size",
g_iscsi.immediate_data_pool_size);
spdk_json_write_named_uint32(w, "data_out_pool_size", g_iscsi.data_out_pool_size);
spdk_json_write_object_end(w);
}
static void
iscsi_auth_group_info_json(struct spdk_iscsi_auth_group *group,
struct spdk_json_write_ctx *w)
{
struct spdk_iscsi_auth_secret *_secret;
spdk_json_write_object_begin(w);
spdk_json_write_named_int32(w, "tag", group->tag);
spdk_json_write_named_array_begin(w, "secrets");
TAILQ_FOREACH(_secret, &group->secret_head, tailq) {
spdk_json_write_object_begin(w);
spdk_json_write_named_string(w, "user", _secret->user);
spdk_json_write_named_string(w, "secret", _secret->secret);
if (_secret->muser[0] != '\0') {
spdk_json_write_named_string(w, "muser", _secret->muser);
spdk_json_write_named_string(w, "msecret", _secret->msecret);
}
spdk_json_write_object_end(w);
}
spdk_json_write_array_end(w);
spdk_json_write_object_end(w);
}
static void
iscsi_auth_group_config_json(struct spdk_iscsi_auth_group *group,
struct spdk_json_write_ctx *w)
{
spdk_json_write_object_begin(w);
spdk_json_write_named_string(w, "method", "iscsi_create_auth_group");
spdk_json_write_name(w, "params");
iscsi_auth_group_info_json(group, w);
spdk_json_write_object_end(w);
}
void
iscsi_auth_groups_info_json(struct spdk_json_write_ctx *w)
{
struct spdk_iscsi_auth_group *group;
TAILQ_FOREACH(group, &g_iscsi.auth_group_head, tailq) {
iscsi_auth_group_info_json(group, w);
}
}
static void
iscsi_auth_groups_config_json(struct spdk_json_write_ctx *w)
{
struct spdk_iscsi_auth_group *group;
TAILQ_FOREACH(group, &g_iscsi.auth_group_head, tailq) {
iscsi_auth_group_config_json(group, w);
}
}
static void
iscsi_opts_config_json(struct spdk_json_write_ctx *w)
{
spdk_json_write_object_begin(w);
spdk_json_write_named_string(w, "method", "iscsi_set_options");
spdk_json_write_name(w, "params");
iscsi_opts_info_json(w);
spdk_json_write_object_end(w);
}
void
spdk_iscsi_config_json(struct spdk_json_write_ctx *w)
{
spdk_json_write_array_begin(w);
iscsi_opts_config_json(w);
iscsi_portal_grps_config_json(w);
iscsi_init_grps_config_json(w);
iscsi_tgt_nodes_config_json(w);
iscsi_auth_groups_config_json(w);
spdk_json_write_array_end(w);
}
SPDK_LOG_REGISTER_COMPONENT(iscsi)