Update Linux to v5.4.2
Change-Id: Idf6911045d9d382da2cfe01b1edff026404ac8fd
diff --git a/net/smc/af_smc.c b/net/smc/af_smc.c
index 80e2119..737b499 100644
--- a/net/smc/af_smc.c
+++ b/net/smc/af_smc.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: GPL-2.0-only
/*
* Shared Memory Communications over RDMA (SMC-R) and RoCE
*
@@ -30,6 +31,10 @@
#include <net/smc.h>
#include <asm/ioctls.h>
+#include <net/net_namespace.h>
+#include <net/netns/generic.h>
+#include "smc_netns.h"
+
#include "smc.h"
#include "smc_clc.h"
#include "smc_llc.h"
@@ -42,8 +47,11 @@
#include "smc_rx.h"
#include "smc_close.h"
-static DEFINE_MUTEX(smc_create_lgr_pending); /* serialize link group
- * creation
+static DEFINE_MUTEX(smc_server_lgr_pending); /* serialize link group
+ * creation on server
+ */
+static DEFINE_MUTEX(smc_client_lgr_pending); /* serialize link group
+ * creation on client
*/
static void smc_tcp_listen_work(struct work_struct *);
@@ -115,6 +123,48 @@
};
EXPORT_SYMBOL_GPL(smc_proto6);
+static void smc_restore_fallback_changes(struct smc_sock *smc)
+{
+ smc->clcsock->file->private_data = smc->sk.sk_socket;
+ smc->clcsock->file = NULL;
+}
+
+static int __smc_release(struct smc_sock *smc)
+{
+ struct sock *sk = &smc->sk;
+ int rc = 0;
+
+ if (!smc->use_fallback) {
+ rc = smc_close_active(smc);
+ sock_set_flag(sk, SOCK_DEAD);
+ sk->sk_shutdown |= SHUTDOWN_MASK;
+ } else {
+ if (sk->sk_state != SMC_LISTEN && sk->sk_state != SMC_INIT)
+ sock_put(sk); /* passive closing */
+ if (sk->sk_state == SMC_LISTEN) {
+ /* wake up clcsock accept */
+ rc = kernel_sock_shutdown(smc->clcsock, SHUT_RDWR);
+ }
+ sk->sk_state = SMC_CLOSED;
+ sk->sk_state_change(sk);
+ smc_restore_fallback_changes(smc);
+ }
+
+ sk->sk_prot->unhash(sk);
+
+ if (sk->sk_state == SMC_CLOSED) {
+ if (smc->clcsock) {
+ release_sock(sk);
+ smc_clcsock_release(smc);
+ lock_sock(sk);
+ }
+ if (!smc->use_fallback)
+ smc_conn_free(&smc->conn);
+ }
+
+ return rc;
+}
+
static int smc_release(struct socket *sock)
{
struct sock *sk = sock->sk;
@@ -127,9 +177,9 @@
smc = smc_sk(sk);
/* cleanup for a dangling non-blocking connect */
+ if (smc->connect_nonblock && sk->sk_state == SMC_INIT)
+ tcp_abort(smc->clcsock->sk, ECONNABORTED);
flush_work(&smc->connect_work);
- kfree(smc->connect_info);
- smc->connect_info = NULL;
if (sk->sk_state == SMC_LISTEN)
/* smc_close_non_accepted() is called and acquires
@@ -139,30 +189,13 @@
else
lock_sock(sk);
- if (!smc->use_fallback) {
- rc = smc_close_active(smc);
- sock_set_flag(sk, SOCK_DEAD);
- sk->sk_shutdown |= SHUTDOWN_MASK;
- }
- if (smc->clcsock) {
- sock_release(smc->clcsock);
- smc->clcsock = NULL;
- }
- if (smc->use_fallback) {
- if (sk->sk_state != SMC_LISTEN && sk->sk_state != SMC_INIT)
- sock_put(sk); /* passive closing */
- sk->sk_state = SMC_CLOSED;
- sk->sk_state_change(sk);
- }
+ rc = __smc_release(smc);
/* detach socket */
sock_orphan(sk);
sock->sk = NULL;
- if (!smc->use_fallback && sk->sk_state == SMC_CLOSED)
- smc_conn_free(&smc->conn);
release_sock(sk);
- sk->sk_prot->unhash(sk);
sock_put(sk); /* final sock_put */
out:
return rc;
@@ -203,6 +236,7 @@
spin_lock_init(&smc->conn.send_lock);
sk->sk_prot->hash(sk);
sk_refcnt_debug_inc(sk);
+ mutex_init(&smc->clcsock_release_lock);
return sk;
}
@@ -236,7 +270,7 @@
/* Check if socket is already active */
rc = -EINVAL;
- if (sk->sk_state != SMC_INIT)
+ if (sk->sk_state != SMC_INIT || smc->connect_nonblock)
goto out_rel;
smc->clcsock->sk->sk_reuse = sk->sk_reuse;
@@ -280,7 +314,8 @@
(1UL << SOCK_RXQ_OVFL) | \
(1UL << SOCK_WIFI_STATUS) | \
(1UL << SOCK_NOFCS) | \
- (1UL << SOCK_FILTER_LOCKED))
+ (1UL << SOCK_FILTER_LOCKED) | \
+ (1UL << SOCK_TSTAMP_NEW))
/* copy only relevant settings and flags of SOL_SOCKET level from smc to
* clc socket (since smc is not called for these options from net/core)
*/
@@ -299,14 +334,17 @@
smc_copy_sock_settings(&smc->sk, smc->clcsock->sk, SK_FLAGS_CLC_TO_SMC);
}
-/* register a new rmb, optionally send confirm_rkey msg to register with peer */
+/* register a new rmb, send confirm_rkey msg to register with peer */
static int smc_reg_rmb(struct smc_link *link, struct smc_buf_desc *rmb_desc,
bool conf_rkey)
{
- /* register memory region for new rmb */
- if (smc_wr_reg_send(link, rmb_desc->mr_rx[SMC_SINGLE_LINK])) {
- rmb_desc->regerr = 1;
- return -EFAULT;
+ if (!rmb_desc->wr_reg) {
+ /* register memory region for new rmb */
+ if (smc_wr_reg_send(link, rmb_desc->mr_rx[SMC_SINGLE_LINK])) {
+ rmb_desc->regerr = 1;
+ return -EFAULT;
+ }
+ rmb_desc->wr_reg = 1;
}
if (!conf_rkey)
return 0;
@@ -335,8 +373,8 @@
struct smc_clc_msg_decline dclc;
rc = smc_clc_wait_msg(smc, &dclc, sizeof(dclc),
- SMC_CLC_DECLINE);
- return rc;
+ SMC_CLC_DECLINE, CLC_WAIT_TIME_SHORT);
+ return rc == -EAGAIN ? SMC_CLC_DECL_TIMEOUT_CL : rc;
}
if (link->llc_confirm_rc)
@@ -363,8 +401,8 @@
struct smc_clc_msg_decline dclc;
rc = smc_clc_wait_msg(smc, &dclc, sizeof(dclc),
- SMC_CLC_DECLINE);
- return rc;
+ SMC_CLC_DECLINE, CLC_WAIT_TIME_SHORT);
+ return rc == -EAGAIN ? SMC_CLC_DECL_TIMEOUT_AL : rc;
}
/* send add link reject message, only one link supported for now */
@@ -423,12 +461,22 @@
link->peer_mtu = clc->qp_mtu;
}
+static void smc_switch_to_fallback(struct smc_sock *smc)
+{
+ smc->use_fallback = true;
+ if (smc->sk.sk_socket && smc->sk.sk_socket->file) {
+ smc->clcsock->file = smc->sk.sk_socket->file;
+ smc->clcsock->file->private_data = smc->clcsock;
+ }
+}
+
/* fall back during connect */
static int smc_connect_fallback(struct smc_sock *smc, int reason_code)
{
- smc->use_fallback = true;
+ smc_switch_to_fallback(smc);
smc->fallback_rsn = reason_code;
smc_copy_sock_settings_to_clc(smc);
+ smc->connect_nonblock = 0;
if (smc->sk.sk_state == SMC_INIT)
smc->sk.sk_state = SMC_ACTIVE;
return 0;
@@ -461,48 +509,48 @@
{
if (local_contact == SMC_FIRST_CONTACT)
smc_lgr_forget(smc->conn.lgr);
- mutex_unlock(&smc_create_lgr_pending);
+ if (smc->conn.lgr->is_smcd)
+ /* there is only one lgr role for SMC-D; use server lock */
+ mutex_unlock(&smc_server_lgr_pending);
+ else
+ mutex_unlock(&smc_client_lgr_pending);
+
smc_conn_free(&smc->conn);
+ smc->connect_nonblock = 0;
return reason_code;
}
/* check if there is a rdma device available for this connection. */
/* called for connect and listen */
-static int smc_check_rdma(struct smc_sock *smc, struct smc_ib_device **ibdev,
- u8 *ibport, unsigned short vlan_id, u8 gid[])
+static int smc_find_rdma_device(struct smc_sock *smc, struct smc_init_info *ini)
{
- int reason_code = 0;
-
/* PNET table look up: search active ib_device and port
* within same PNETID that also contains the ethernet device
* used for the internal TCP socket
*/
- smc_pnet_find_roce_resource(smc->clcsock->sk, ibdev, ibport, vlan_id,
- gid);
- if (!(*ibdev))
- reason_code = SMC_CLC_DECL_CNFERR; /* configuration error */
-
- return reason_code;
+ smc_pnet_find_roce_resource(smc->clcsock->sk, ini);
+ if (!ini->ib_dev)
+ return SMC_CLC_DECL_NOSMCRDEV;
+ return 0;
}
/* check if there is an ISM device available for this connection. */
/* called for connect and listen */
-static int smc_check_ism(struct smc_sock *smc, struct smcd_dev **ismdev)
+static int smc_find_ism_device(struct smc_sock *smc, struct smc_init_info *ini)
{
/* Find ISM device with same PNETID as connecting interface */
- smc_pnet_find_ism_resource(smc->clcsock->sk, ismdev);
- if (!(*ismdev))
- return SMC_CLC_DECL_CNFERR; /* configuration error */
+ smc_pnet_find_ism_resource(smc->clcsock->sk, ini);
+ if (!ini->ism_dev)
+ return SMC_CLC_DECL_NOSMCDDEV;
return 0;
}
/* Check for VLAN ID and register it on ISM device just for CLC handshake */
static int smc_connect_ism_vlan_setup(struct smc_sock *smc,
- struct smcd_dev *ismdev,
- unsigned short vlan_id)
+ struct smc_init_info *ini)
{
- if (vlan_id && smc_ism_get_vlan(ismdev, vlan_id))
- return SMC_CLC_DECL_CNFERR;
+ if (ini->vlan_id && smc_ism_get_vlan(ini->ism_dev, ini->vlan_id))
+ return SMC_CLC_DECL_ISMVLANERR;
return 0;
}
@@ -510,12 +558,11 @@
* used, the VLAN ID will be registered again during the connection setup.
*/
static int smc_connect_ism_vlan_cleanup(struct smc_sock *smc, bool is_smcd,
- struct smcd_dev *ismdev,
- unsigned short vlan_id)
+ struct smc_init_info *ini)
{
if (!is_smcd)
return 0;
- if (vlan_id && smc_ism_put_vlan(ismdev, vlan_id))
+ if (ini->vlan_id && smc_ism_put_vlan(ini->ism_dev, ini->vlan_id))
return SMC_CLC_DECL_CNFERR;
return 0;
}
@@ -523,39 +570,37 @@
/* CLC handshake during connect */
static int smc_connect_clc(struct smc_sock *smc, int smc_type,
struct smc_clc_msg_accept_confirm *aclc,
- struct smc_ib_device *ibdev, u8 ibport,
- u8 gid[], struct smcd_dev *ismdev)
+ struct smc_init_info *ini)
{
int rc = 0;
/* do inband token exchange */
- rc = smc_clc_send_proposal(smc, smc_type, ibdev, ibport, gid, ismdev);
+ rc = smc_clc_send_proposal(smc, smc_type, ini);
if (rc)
return rc;
/* receive SMC Accept CLC message */
- return smc_clc_wait_msg(smc, aclc, sizeof(*aclc), SMC_CLC_ACCEPT);
+ return smc_clc_wait_msg(smc, aclc, sizeof(*aclc), SMC_CLC_ACCEPT,
+ CLC_WAIT_TIME);
}
/* setup for RDMA connection of client */
static int smc_connect_rdma(struct smc_sock *smc,
struct smc_clc_msg_accept_confirm *aclc,
- struct smc_ib_device *ibdev, u8 ibport)
+ struct smc_init_info *ini)
{
- int local_contact = SMC_FIRST_CONTACT;
struct smc_link *link;
int reason_code = 0;
- mutex_lock(&smc_create_lgr_pending);
- local_contact = smc_conn_create(smc, false, aclc->hdr.flag, ibdev,
- ibport, &aclc->lcl, NULL, 0);
- if (local_contact < 0) {
- if (local_contact == -ENOMEM)
- reason_code = SMC_CLC_DECL_MEM;/* insufficient memory*/
- else if (local_contact == -ENOLINK)
- reason_code = SMC_CLC_DECL_SYNCERR; /* synchr. error */
- else
- reason_code = SMC_CLC_DECL_INTERR; /* other error */
- return smc_connect_abort(smc, reason_code, 0);
+ ini->is_smcd = false;
+ ini->ib_lcl = &aclc->lcl;
+ ini->ib_clcqpn = ntoh24(aclc->qpn);
+ ini->srv_first_contact = aclc->hdr.flag;
+
+ mutex_lock(&smc_client_lgr_pending);
+ reason_code = smc_conn_create(smc, ini);
+ if (reason_code) {
+ mutex_unlock(&smc_client_lgr_pending);
+ return reason_code;
}
link = &smc->conn.lgr->lnk[SMC_SINGLE_LINK];
@@ -563,46 +608,48 @@
/* create send buffer and rmb */
if (smc_buf_create(smc, false))
- return smc_connect_abort(smc, SMC_CLC_DECL_MEM, local_contact);
+ return smc_connect_abort(smc, SMC_CLC_DECL_MEM,
+ ini->cln_first_contact);
- if (local_contact == SMC_FIRST_CONTACT)
+ if (ini->cln_first_contact == SMC_FIRST_CONTACT)
smc_link_save_peer_info(link, aclc);
if (smc_rmb_rtoken_handling(&smc->conn, aclc))
return smc_connect_abort(smc, SMC_CLC_DECL_ERR_RTOK,
- local_contact);
+ ini->cln_first_contact);
smc_close_init(smc);
smc_rx_init(smc);
- if (local_contact == SMC_FIRST_CONTACT) {
+ if (ini->cln_first_contact == SMC_FIRST_CONTACT) {
if (smc_ib_ready_link(link))
return smc_connect_abort(smc, SMC_CLC_DECL_ERR_RDYLNK,
- local_contact);
+ ini->cln_first_contact);
} else {
- if (!smc->conn.rmb_desc->reused &&
- smc_reg_rmb(link, smc->conn.rmb_desc, true))
+ if (smc_reg_rmb(link, smc->conn.rmb_desc, true))
return smc_connect_abort(smc, SMC_CLC_DECL_ERR_REGRMB,
- local_contact);
+ ini->cln_first_contact);
}
smc_rmb_sync_sg_for_device(&smc->conn);
reason_code = smc_clc_send_confirm(smc);
if (reason_code)
- return smc_connect_abort(smc, reason_code, local_contact);
+ return smc_connect_abort(smc, reason_code,
+ ini->cln_first_contact);
smc_tx_init(smc);
- if (local_contact == SMC_FIRST_CONTACT) {
+ if (ini->cln_first_contact == SMC_FIRST_CONTACT) {
/* QP confirmation over RoCE fabric */
reason_code = smc_clnt_conf_first_link(smc);
if (reason_code)
return smc_connect_abort(smc, reason_code,
- local_contact);
+ ini->cln_first_contact);
}
- mutex_unlock(&smc_create_lgr_pending);
+ mutex_unlock(&smc_client_lgr_pending);
smc_copy_sock_settings_to_clc(smc);
+ smc->connect_nonblock = 0;
if (smc->sk.sk_state == SMC_INIT)
smc->sk.sk_state = SMC_ACTIVE;
@@ -612,20 +659,26 @@
/* setup for ISM connection of client */
static int smc_connect_ism(struct smc_sock *smc,
struct smc_clc_msg_accept_confirm *aclc,
- struct smcd_dev *ismdev)
+ struct smc_init_info *ini)
{
- int local_contact = SMC_FIRST_CONTACT;
int rc = 0;
- mutex_lock(&smc_create_lgr_pending);
- local_contact = smc_conn_create(smc, true, aclc->hdr.flag, NULL, 0,
- NULL, ismdev, aclc->gid);
- if (local_contact < 0)
- return smc_connect_abort(smc, SMC_CLC_DECL_MEM, 0);
+ ini->is_smcd = true;
+ ini->ism_gid = aclc->gid;
+ ini->srv_first_contact = aclc->hdr.flag;
+
+ /* there is only one lgr role for SMC-D; use server lock */
+ mutex_lock(&smc_server_lgr_pending);
+ rc = smc_conn_create(smc, ini);
+ if (rc) {
+ mutex_unlock(&smc_server_lgr_pending);
+ return rc;
+ }
/* Create send and receive buffers */
if (smc_buf_create(smc, true))
- return smc_connect_abort(smc, SMC_CLC_DECL_MEM, local_contact);
+ return smc_connect_abort(smc, SMC_CLC_DECL_MEM,
+ ini->cln_first_contact);
smc_conn_save_peer_info(smc, aclc);
smc_close_init(smc);
@@ -634,10 +687,11 @@
rc = smc_clc_send_confirm(smc);
if (rc)
- return smc_connect_abort(smc, rc, local_contact);
- mutex_unlock(&smc_create_lgr_pending);
+ return smc_connect_abort(smc, rc, ini->cln_first_contact);
+ mutex_unlock(&smc_server_lgr_pending);
smc_copy_sock_settings_to_clc(smc);
+ smc->connect_nonblock = 0;
if (smc->sk.sk_state == SMC_INIT)
smc->sk.sk_state = SMC_ACTIVE;
@@ -649,15 +703,9 @@
{
bool ism_supported = false, rdma_supported = false;
struct smc_clc_msg_accept_confirm aclc;
- struct smc_ib_device *ibdev;
- struct smcd_dev *ismdev;
- u8 gid[SMC_GID_SIZE];
- unsigned short vlan;
+ struct smc_init_info ini = {0};
int smc_type;
int rc = 0;
- u8 ibport;
-
- sock_hold(&smc->sk); /* sock put in passive closing */
if (smc->use_fallback)
return smc_connect_fallback(smc, smc->fallback_rsn);
@@ -670,20 +718,21 @@
if (using_ipsec(smc))
return smc_connect_decline_fallback(smc, SMC_CLC_DECL_IPSEC);
- /* check for VLAN ID */
- if (smc_vlan_by_tcpsk(smc->clcsock, &vlan))
- return smc_connect_decline_fallback(smc, SMC_CLC_DECL_CNFERR);
+ /* get vlan id from IP device */
+ if (smc_vlan_by_tcpsk(smc->clcsock, &ini))
+ return smc_connect_decline_fallback(smc,
+ SMC_CLC_DECL_GETVLANERR);
/* check if there is an ism device available */
- if (!smc_check_ism(smc, &ismdev) &&
- !smc_connect_ism_vlan_setup(smc, ismdev, vlan)) {
+ if (!smc_find_ism_device(smc, &ini) &&
+ !smc_connect_ism_vlan_setup(smc, &ini)) {
/* ISM is supported for this connection */
ism_supported = true;
smc_type = SMC_TYPE_D;
}
/* check if there is a rdma device available */
- if (!smc_check_rdma(smc, &ibdev, &ibport, vlan, gid)) {
+ if (!smc_find_rdma_device(smc, &ini)) {
/* RDMA is supported for this connection */
rdma_supported = true;
if (ism_supported)
@@ -697,25 +746,25 @@
return smc_connect_decline_fallback(smc, SMC_CLC_DECL_NOSMCDEV);
/* perform CLC handshake */
- rc = smc_connect_clc(smc, smc_type, &aclc, ibdev, ibport, gid, ismdev);
+ rc = smc_connect_clc(smc, smc_type, &aclc, &ini);
if (rc) {
- smc_connect_ism_vlan_cleanup(smc, ism_supported, ismdev, vlan);
+ smc_connect_ism_vlan_cleanup(smc, ism_supported, &ini);
return smc_connect_decline_fallback(smc, rc);
}
/* depending on previous steps, connect using rdma or ism */
if (rdma_supported && aclc.hdr.path == SMC_TYPE_R)
- rc = smc_connect_rdma(smc, &aclc, ibdev, ibport);
+ rc = smc_connect_rdma(smc, &aclc, &ini);
else if (ism_supported && aclc.hdr.path == SMC_TYPE_D)
- rc = smc_connect_ism(smc, &aclc, ismdev);
+ rc = smc_connect_ism(smc, &aclc, &ini);
else
rc = SMC_CLC_DECL_MODEUNSUPP;
if (rc) {
- smc_connect_ism_vlan_cleanup(smc, ism_supported, ismdev, vlan);
+ smc_connect_ism_vlan_cleanup(smc, ism_supported, &ini);
return smc_connect_decline_fallback(smc, rc);
}
- smc_connect_ism_vlan_cleanup(smc, ism_supported, ismdev, vlan);
+ smc_connect_ism_vlan_cleanup(smc, ism_supported, &ini);
return 0;
}
@@ -723,17 +772,31 @@
{
struct smc_sock *smc = container_of(work, struct smc_sock,
connect_work);
- int rc;
+ long timeo = smc->sk.sk_sndtimeo;
+ int rc = 0;
- lock_sock(&smc->sk);
- rc = kernel_connect(smc->clcsock, &smc->connect_info->addr,
- smc->connect_info->alen, smc->connect_info->flags);
+ if (!timeo)
+ timeo = MAX_SCHEDULE_TIMEOUT;
+ lock_sock(smc->clcsock->sk);
if (smc->clcsock->sk->sk_err) {
smc->sk.sk_err = smc->clcsock->sk->sk_err;
- goto out;
+ } else if ((1 << smc->clcsock->sk->sk_state) &
+ (TCPF_SYN_SENT | TCP_SYN_RECV)) {
+ rc = sk_stream_wait_connect(smc->clcsock->sk, &timeo);
+ if ((rc == -EPIPE) &&
+ ((1 << smc->clcsock->sk->sk_state) &
+ (TCPF_ESTABLISHED | TCPF_CLOSE_WAIT)))
+ rc = 0;
}
- if (rc < 0) {
- smc->sk.sk_err = -rc;
+ release_sock(smc->clcsock->sk);
+ lock_sock(&smc->sk);
+ if (rc != 0 || smc->sk.sk_err) {
+ smc->sk.sk_state = SMC_CLOSED;
+ if (rc == -EPIPE || rc == -EAGAIN)
+ smc->sk.sk_err = EPIPE;
+ else if (signal_pending(current))
+ smc->sk.sk_err = -sock_intr_errno(timeo);
+ sock_put(&smc->sk); /* passive closing */
goto out;
}
@@ -742,12 +805,14 @@
smc->sk.sk_err = -rc;
out:
- if (smc->sk.sk_err)
- smc->sk.sk_state_change(&smc->sk);
- else
- smc->sk.sk_write_space(&smc->sk);
- kfree(smc->connect_info);
- smc->connect_info = NULL;
+ if (!sock_flag(&smc->sk, SOCK_DEAD)) {
+ if (smc->sk.sk_err) {
+ smc->sk.sk_state_change(&smc->sk);
+ } else { /* allow polling before and after fallback decision */
+ smc->clcsock->sk->sk_write_space(smc->clcsock->sk);
+ smc->sk.sk_write_space(&smc->sk);
+ }
+ }
release_sock(&smc->sk);
}
@@ -780,26 +845,20 @@
smc_copy_sock_settings_to_clc(smc);
tcp_sk(smc->clcsock->sk)->syn_smc = 1;
+ if (smc->connect_nonblock) {
+ rc = -EALREADY;
+ goto out;
+ }
+ rc = kernel_connect(smc->clcsock, addr, alen, flags);
+ if (rc && rc != -EINPROGRESS)
+ goto out;
+
+ sock_hold(&smc->sk); /* sock put in passive closing */
if (flags & O_NONBLOCK) {
- if (smc->connect_info) {
- rc = -EALREADY;
- goto out;
- }
- smc->connect_info = kzalloc(alen + 2 * sizeof(int), GFP_KERNEL);
- if (!smc->connect_info) {
- rc = -ENOMEM;
- goto out;
- }
- smc->connect_info->alen = alen;
- smc->connect_info->flags = flags ^ O_NONBLOCK;
- memcpy(&smc->connect_info->addr, addr, alen);
- schedule_work(&smc->connect_work);
+ if (schedule_work(&smc->connect_work))
+ smc->connect_nonblock = 1;
rc = -EINPROGRESS;
} else {
- rc = kernel_connect(smc->clcsock, addr, alen, flags);
- if (rc)
- goto out;
-
rc = __smc_connect(smc);
if (rc < 0)
goto out;
@@ -818,7 +877,7 @@
struct socket *new_clcsock = NULL;
struct sock *lsk = &lsmc->sk;
struct sock *new_sk;
- int rc;
+ int rc = -EINVAL;
release_sock(lsk);
new_sk = smc_sock_alloc(sock_net(lsk), NULL, lsk->sk_protocol);
@@ -831,16 +890,19 @@
}
*new_smc = smc_sk(new_sk);
- rc = kernel_accept(lsmc->clcsock, &new_clcsock, 0);
+ mutex_lock(&lsmc->clcsock_release_lock);
+ if (lsmc->clcsock)
+ rc = kernel_accept(lsmc->clcsock, &new_clcsock, 0);
+ mutex_unlock(&lsmc->clcsock_release_lock);
lock_sock(lsk);
if (rc < 0)
lsk->sk_err = -rc;
if (rc < 0 || lsk->sk_state == SMC_CLOSED) {
+ new_sk->sk_prot->unhash(new_sk);
if (new_clcsock)
sock_release(new_clcsock);
new_sk->sk_state = SMC_CLOSED;
sock_set_flag(new_sk, SOCK_DEAD);
- new_sk->sk_prot->unhash(new_sk);
sock_put(new_sk); /* final */
*new_smc = NULL;
goto out;
@@ -891,16 +953,21 @@
smc_accept_unlink(new_sk);
if (new_sk->sk_state == SMC_CLOSED) {
+ new_sk->sk_prot->unhash(new_sk);
if (isk->clcsock) {
sock_release(isk->clcsock);
isk->clcsock = NULL;
}
- new_sk->sk_prot->unhash(new_sk);
sock_put(new_sk); /* final */
continue;
}
- if (new_sock)
+ if (new_sock) {
sock_graft(new_sk, new_sock);
+ if (isk->use_fallback) {
+ smc_sk(new_sk)->clcsock->file = new_sock->file;
+ isk->clcsock->file->private_data = isk->clcsock;
+ }
+ }
return new_sk;
}
return NULL;
@@ -915,27 +982,8 @@
if (!sk->sk_lingertime)
/* wait for peer closing */
sk->sk_lingertime = SMC_MAX_STREAM_WAIT_TIMEOUT;
- if (!smc->use_fallback) {
- smc_close_active(smc);
- sock_set_flag(sk, SOCK_DEAD);
- sk->sk_shutdown |= SHUTDOWN_MASK;
- }
- if (smc->clcsock) {
- struct socket *tcp;
-
- tcp = smc->clcsock;
- smc->clcsock = NULL;
- sock_release(tcp);
- }
- if (smc->use_fallback) {
- sock_put(sk); /* passive closing */
- sk->sk_state = SMC_CLOSED;
- } else {
- if (sk->sk_state == SMC_CLOSED)
- smc_conn_free(&smc->conn);
- }
+ __smc_release(smc);
release_sock(sk);
- sk->sk_prot->unhash(sk);
sock_put(sk); /* final sock_put */
}
@@ -965,8 +1013,8 @@
struct smc_clc_msg_decline dclc;
rc = smc_clc_wait_msg(smc, &dclc, sizeof(dclc),
- SMC_CLC_DECLINE);
- return rc;
+ SMC_CLC_DECLINE, CLC_WAIT_TIME_SHORT);
+ return rc == -EAGAIN ? SMC_CLC_DECL_TIMEOUT_CL : rc;
}
if (link->llc_confirm_resp_rc)
@@ -986,8 +1034,8 @@
struct smc_clc_msg_decline dclc;
rc = smc_clc_wait_msg(smc, &dclc, sizeof(dclc),
- SMC_CLC_DECLINE);
- return rc;
+ SMC_CLC_DECLINE, CLC_WAIT_TIME_SHORT);
+ return rc == -EAGAIN ? SMC_CLC_DECL_TIMEOUT_AL : rc;
}
smc_llc_link_active(link, net->ipv4.sysctl_tcp_keepalive_time);
@@ -1001,13 +1049,13 @@
struct smc_sock *lsmc = new_smc->listen_smc;
struct sock *newsmcsk = &new_smc->sk;
- lock_sock_nested(&lsmc->sk, SINGLE_DEPTH_NESTING);
if (lsmc->sk.sk_state == SMC_LISTEN) {
+ lock_sock_nested(&lsmc->sk, SINGLE_DEPTH_NESTING);
smc_accept_enqueue(&lsmc->sk, newsmcsk);
+ release_sock(&lsmc->sk);
} else { /* no longer listening */
smc_close_non_accepted(newsmcsk);
}
- release_sock(&lsmc->sk);
/* Wake up accept */
lsmc->sk.sk_data_ready(&lsmc->sk);
@@ -1051,7 +1099,7 @@
return;
}
smc_conn_free(&new_smc->conn);
- new_smc->use_fallback = true;
+ smc_switch_to_fallback(new_smc);
new_smc->fallback_rsn = reason_code;
if (reason_code && reason_code != SMC_CLC_DECL_PEERDECL) {
if (smc_clc_send_decline(new_smc, reason_code) < 0) {
@@ -1063,7 +1111,7 @@
}
/* listen worker: check prefixes */
-static int smc_listen_rdma_check(struct smc_sock *new_smc,
+static int smc_listen_prfx_check(struct smc_sock *new_smc,
struct smc_clc_msg_proposal *pclc)
{
struct smc_clc_msg_proposal_prefix *pclc_prfx;
@@ -1071,25 +1119,21 @@
pclc_prfx = smc_clc_proposal_get_prefix(pclc);
if (smc_clc_prfx_match(newclcsock, pclc_prfx))
- return SMC_CLC_DECL_CNFERR;
+ return SMC_CLC_DECL_DIFFPREFIX;
return 0;
}
/* listen worker: initialize connection and buffers */
static int smc_listen_rdma_init(struct smc_sock *new_smc,
- struct smc_clc_msg_proposal *pclc,
- struct smc_ib_device *ibdev, u8 ibport,
- int *local_contact)
+ struct smc_init_info *ini)
{
+ int rc;
+
/* allocate connection / link group */
- *local_contact = smc_conn_create(new_smc, false, 0, ibdev, ibport,
- &pclc->lcl, NULL, 0);
- if (*local_contact < 0) {
- if (*local_contact == -ENOMEM)
- return SMC_CLC_DECL_MEM;/* insufficient memory*/
- return SMC_CLC_DECL_INTERR; /* other error */
- }
+ rc = smc_conn_create(new_smc, ini);
+ if (rc)
+ return rc;
/* create send buffer and rmb */
if (smc_buf_create(new_smc, false))
@@ -1101,33 +1145,30 @@
/* listen worker: initialize connection and buffers for SMC-D */
static int smc_listen_ism_init(struct smc_sock *new_smc,
struct smc_clc_msg_proposal *pclc,
- struct smcd_dev *ismdev,
- int *local_contact)
+ struct smc_init_info *ini)
{
struct smc_clc_msg_smcd *pclc_smcd;
+ int rc;
pclc_smcd = smc_get_clc_msg_smcd(pclc);
- *local_contact = smc_conn_create(new_smc, true, 0, NULL, 0, NULL,
- ismdev, pclc_smcd->gid);
- if (*local_contact < 0) {
- if (*local_contact == -ENOMEM)
- return SMC_CLC_DECL_MEM;/* insufficient memory*/
- return SMC_CLC_DECL_INTERR; /* other error */
- }
+ ini->ism_gid = pclc_smcd->gid;
+ rc = smc_conn_create(new_smc, ini);
+ if (rc)
+ return rc;
/* Check if peer can be reached via ISM device */
if (smc_ism_cantalk(new_smc->conn.lgr->peer_gid,
new_smc->conn.lgr->vlan_id,
new_smc->conn.lgr->smcd)) {
- if (*local_contact == SMC_FIRST_CONTACT)
+ if (ini->cln_first_contact == SMC_FIRST_CONTACT)
smc_lgr_forget(new_smc->conn.lgr);
smc_conn_free(&new_smc->conn);
- return SMC_CLC_DECL_CNFERR;
+ return SMC_CLC_DECL_SMCDNOTALK;
}
/* Create send and receive buffers */
if (smc_buf_create(new_smc, true)) {
- if (*local_contact == SMC_FIRST_CONTACT)
+ if (ini->cln_first_contact == SMC_FIRST_CONTACT)
smc_lgr_forget(new_smc->conn.lgr);
smc_conn_free(&new_smc->conn);
return SMC_CLC_DECL_MEM;
@@ -1142,10 +1183,8 @@
struct smc_link *link = &new_smc->conn.lgr->lnk[SMC_SINGLE_LINK];
if (local_contact != SMC_FIRST_CONTACT) {
- if (!new_smc->conn.rmb_desc->reused) {
- if (smc_reg_rmb(link, new_smc->conn.rmb_desc, true))
- return SMC_CLC_DECL_ERR_REGRMB;
- }
+ if (smc_reg_rmb(link, new_smc->conn.rmb_desc, true))
+ return SMC_CLC_DECL_ERR_REGRMB;
}
smc_rmb_sync_sg_for_device(&new_smc->conn);
@@ -1181,7 +1220,6 @@
return 0;
decline:
- mutex_unlock(&smc_create_lgr_pending);
smc_listen_decline(new_smc, reason_code, local_contact);
return reason_code;
}
@@ -1194,15 +1232,13 @@
struct socket *newclcsock = new_smc->clcsock;
struct smc_clc_msg_accept_confirm cclc;
struct smc_clc_msg_proposal *pclc;
- struct smc_ib_device *ibdev;
+ struct smc_init_info ini = {0};
bool ism_supported = false;
- struct smcd_dev *ismdev;
u8 buf[SMC_CLC_MAX_LEN];
- int local_contact = 0;
- unsigned short vlan;
- int reason_code = 0;
int rc = 0;
- u8 ibport;
+
+ if (new_smc->listen_smc->sk.sk_state != SMC_LISTEN)
+ return smc_listen_out_err(new_smc);
if (new_smc->use_fallback) {
smc_listen_out_connected(new_smc);
@@ -1211,7 +1247,7 @@
/* check if peer is smc capable */
if (!tcp_sk(newclcsock->sk)->syn_smc) {
- new_smc->use_fallback = true;
+ smc_switch_to_fallback(new_smc);
new_smc->fallback_rsn = SMC_CLC_DECL_PEERNOSMC;
smc_listen_out_connected(new_smc);
return;
@@ -1221,72 +1257,101 @@
* wait for and receive SMC Proposal CLC message
*/
pclc = (struct smc_clc_msg_proposal *)&buf;
- reason_code = smc_clc_wait_msg(new_smc, pclc, SMC_CLC_MAX_LEN,
- SMC_CLC_PROPOSAL);
- if (reason_code) {
- smc_listen_decline(new_smc, reason_code, 0);
- return;
- }
+ rc = smc_clc_wait_msg(new_smc, pclc, SMC_CLC_MAX_LEN,
+ SMC_CLC_PROPOSAL, CLC_WAIT_TIME);
+ if (rc)
+ goto out_decl;
/* IPSec connections opt out of SMC-R optimizations */
if (using_ipsec(new_smc)) {
- smc_listen_decline(new_smc, SMC_CLC_DECL_IPSEC, 0);
- return;
+ rc = SMC_CLC_DECL_IPSEC;
+ goto out_decl;
}
- mutex_lock(&smc_create_lgr_pending);
+ /* check for matching IP prefix and subnet length */
+ rc = smc_listen_prfx_check(new_smc, pclc);
+ if (rc)
+ goto out_decl;
+
+ /* get vlan id from IP device */
+ if (smc_vlan_by_tcpsk(new_smc->clcsock, &ini)) {
+ rc = SMC_CLC_DECL_GETVLANERR;
+ goto out_decl;
+ }
+
+ mutex_lock(&smc_server_lgr_pending);
smc_close_init(new_smc);
smc_rx_init(new_smc);
smc_tx_init(new_smc);
/* check if ISM is available */
- if ((pclc->hdr.path == SMC_TYPE_D || pclc->hdr.path == SMC_TYPE_B) &&
- !smc_check_ism(new_smc, &ismdev) &&
- !smc_listen_ism_init(new_smc, pclc, ismdev, &local_contact)) {
- ism_supported = true;
+ if (pclc->hdr.path == SMC_TYPE_D || pclc->hdr.path == SMC_TYPE_B) {
+ ini.is_smcd = true; /* prepare ISM check */
+ rc = smc_find_ism_device(new_smc, &ini);
+ if (!rc)
+ rc = smc_listen_ism_init(new_smc, pclc, &ini);
+ if (!rc)
+ ism_supported = true;
+ else if (pclc->hdr.path == SMC_TYPE_D)
+ goto out_unlock; /* skip RDMA and decline */
}
/* check if RDMA is available */
- if (!ism_supported &&
- ((pclc->hdr.path != SMC_TYPE_R && pclc->hdr.path != SMC_TYPE_B) ||
- smc_vlan_by_tcpsk(new_smc->clcsock, &vlan) ||
- smc_check_rdma(new_smc, &ibdev, &ibport, vlan, NULL) ||
- smc_listen_rdma_check(new_smc, pclc) ||
- smc_listen_rdma_init(new_smc, pclc, ibdev, ibport,
- &local_contact) ||
- smc_listen_rdma_reg(new_smc, local_contact))) {
- /* SMC not supported, decline */
- mutex_unlock(&smc_create_lgr_pending);
- smc_listen_decline(new_smc, SMC_CLC_DECL_MODEUNSUPP,
- local_contact);
- return;
+ if (!ism_supported) { /* SMC_TYPE_R or SMC_TYPE_B */
+ /* prepare RDMA check */
+ ini.is_smcd = false;
+ ini.ism_dev = NULL;
+ ini.ib_lcl = &pclc->lcl;
+ rc = smc_find_rdma_device(new_smc, &ini);
+ if (rc) {
+ /* no RDMA device found */
+ if (pclc->hdr.path == SMC_TYPE_B)
+ /* neither ISM nor RDMA device found */
+ rc = SMC_CLC_DECL_NOSMCDEV;
+ goto out_unlock;
+ }
+ rc = smc_listen_rdma_init(new_smc, &ini);
+ if (rc)
+ goto out_unlock;
+ rc = smc_listen_rdma_reg(new_smc, ini.cln_first_contact);
+ if (rc)
+ goto out_unlock;
}
/* send SMC Accept CLC message */
- rc = smc_clc_send_accept(new_smc, local_contact);
- if (rc) {
- mutex_unlock(&smc_create_lgr_pending);
- smc_listen_decline(new_smc, rc, local_contact);
- return;
- }
+ rc = smc_clc_send_accept(new_smc, ini.cln_first_contact);
+ if (rc)
+ goto out_unlock;
+
+ /* SMC-D does not need this lock any more */
+ if (ism_supported)
+ mutex_unlock(&smc_server_lgr_pending);
/* receive SMC Confirm CLC message */
- reason_code = smc_clc_wait_msg(new_smc, &cclc, sizeof(cclc),
- SMC_CLC_CONFIRM);
- if (reason_code) {
- mutex_unlock(&smc_create_lgr_pending);
- smc_listen_decline(new_smc, reason_code, local_contact);
- return;
+ rc = smc_clc_wait_msg(new_smc, &cclc, sizeof(cclc),
+ SMC_CLC_CONFIRM, CLC_WAIT_TIME);
+ if (rc) {
+ if (!ism_supported)
+ goto out_unlock;
+ goto out_decl;
}
/* finish worker */
if (!ism_supported) {
- if (smc_listen_rdma_finish(new_smc, &cclc, local_contact))
+ rc = smc_listen_rdma_finish(new_smc, &cclc,
+ ini.cln_first_contact);
+ mutex_unlock(&smc_server_lgr_pending);
+ if (rc)
return;
}
smc_conn_save_peer_info(new_smc, &cclc);
- mutex_unlock(&smc_create_lgr_pending);
smc_listen_out_connected(new_smc);
+ return;
+
+out_unlock:
+ mutex_unlock(&smc_server_lgr_pending);
+out_decl:
+ smc_listen_decline(new_smc, rc, ini.cln_first_contact);
}
static void smc_tcp_listen_work(struct work_struct *work)
@@ -1333,7 +1398,8 @@
lock_sock(sk);
rc = -EINVAL;
- if ((sk->sk_state != SMC_INIT) && (sk->sk_state != SMC_LISTEN))
+ if ((sk->sk_state != SMC_INIT && sk->sk_state != SMC_LISTEN) ||
+ smc->connect_nonblock)
goto out;
rc = 0;
@@ -1354,7 +1420,6 @@
sk->sk_max_ack_backlog = backlog;
sk->sk_ack_backlog = 0;
sk->sk_state = SMC_LISTEN;
- INIT_WORK(&smc->tcp_listen_work, smc_tcp_listen_work);
sock_hold(sk); /* sock_hold in tcp_listen_worker */
if (!schedule_work(&smc->tcp_listen_work))
sock_put(sk);
@@ -1462,8 +1527,8 @@
goto out;
if (msg->msg_flags & MSG_FASTOPEN) {
- if (sk->sk_state == SMC_INIT) {
- smc->use_fallback = true;
+ if (sk->sk_state == SMC_INIT && !smc->connect_nonblock) {
+ smc_switch_to_fallback(smc);
smc->fallback_rsn = SMC_CLC_DECL_OPTUNSUPP;
} else {
rc = -EINVAL;
@@ -1489,6 +1554,11 @@
smc = smc_sk(sk);
lock_sock(sk);
+ if (sk->sk_state == SMC_CLOSED && (sk->sk_shutdown & RCV_SHUTDOWN)) {
+ /* socket was connected before, no more data to read */
+ rc = 0;
+ goto out;
+ }
if ((sk->sk_state == SMC_INIT) ||
(sk->sk_state == SMC_LISTEN) ||
(sk->sk_state == SMC_CLOSED))
@@ -1528,8 +1598,8 @@
poll_table *wait)
{
struct sock *sk = sock->sk;
- __poll_t mask = 0;
struct smc_sock *smc;
+ __poll_t mask = 0;
if (!sk)
return EPOLLNVAL;
@@ -1539,8 +1609,6 @@
/* delegate to CLC child sock */
mask = smc->clcsock->ops->poll(file, smc->clcsock, wait);
sk->sk_err = smc->clcsock->sk->sk_err;
- if (sk->sk_err)
- mask |= EPOLLERR;
} else {
if (sk->sk_state != SMC_CLOSED)
sock_poll_wait(file, sock, wait);
@@ -1551,9 +1619,14 @@
mask |= EPOLLHUP;
if (sk->sk_state == SMC_LISTEN) {
/* woken up by sk_data_ready in smc_listen_work() */
- mask = smc_accept_poll(sk);
+ mask |= smc_accept_poll(sk);
+ } else if (smc->use_fallback) { /* as result of connect_work()*/
+ mask |= smc->clcsock->ops->poll(file, smc->clcsock,
+ wait);
+ sk->sk_err = smc->clcsock->sk->sk_err;
} else {
- if (atomic_read(&smc->conn.sndbuf_space) ||
+ if ((sk->sk_state != SMC_INIT &&
+ atomic_read(&smc->conn.sndbuf_space)) ||
sk->sk_shutdown & SEND_SHUTDOWN) {
mask |= EPOLLOUT | EPOLLWRNORM;
} else {
@@ -1659,8 +1732,8 @@
case TCP_FASTOPEN_KEY:
case TCP_FASTOPEN_NO_COOKIE:
/* option not supported by SMC */
- if (sk->sk_state == SMC_INIT) {
- smc->use_fallback = true;
+ if (sk->sk_state == SMC_INIT && !smc->connect_nonblock) {
+ smc_switch_to_fallback(smc);
smc->fallback_rsn = SMC_CLC_DECL_OPTUNSUPP;
} else {
if (!smc->use_fallback)
@@ -1668,14 +1741,18 @@
}
break;
case TCP_NODELAY:
- if (sk->sk_state != SMC_INIT && sk->sk_state != SMC_LISTEN) {
+ if (sk->sk_state != SMC_INIT &&
+ sk->sk_state != SMC_LISTEN &&
+ sk->sk_state != SMC_CLOSED) {
if (val && !smc->use_fallback)
mod_delayed_work(system_wq, &smc->conn.tx_work,
0);
}
break;
case TCP_CORK:
- if (sk->sk_state != SMC_INIT && sk->sk_state != SMC_LISTEN) {
+ if (sk->sk_state != SMC_INIT &&
+ sk->sk_state != SMC_LISTEN &&
+ sk->sk_state != SMC_CLOSED) {
if (!val && !smc->use_fallback)
mod_delayed_work(system_wq, &smc->conn.tx_work,
0);
@@ -1824,7 +1901,11 @@
smc = smc_sk(sk);
lock_sock(sk);
-
+ if (sk->sk_state == SMC_CLOSED && (sk->sk_shutdown & RCV_SHUTDOWN)) {
+ /* socket was connected before, no more data to read */
+ rc = 0;
+ goto out;
+ }
if (sk->sk_state == SMC_INIT ||
sk->sk_state == SMC_LISTEN ||
sk->sk_state == SMC_CLOSED)
@@ -1923,14 +2004,37 @@
.create = smc_create,
};
+unsigned int smc_net_id;
+
+static __net_init int smc_net_init(struct net *net)
+{
+ return smc_pnet_net_init(net);
+}
+
+static void __net_exit smc_net_exit(struct net *net)
+{
+ smc_pnet_net_exit(net);
+}
+
+static struct pernet_operations smc_net_ops = {
+ .init = smc_net_init,
+ .exit = smc_net_exit,
+ .id = &smc_net_id,
+ .size = sizeof(struct smc_net),
+};
+
static int __init smc_init(void)
{
int rc;
- rc = smc_pnet_init();
+ rc = register_pernet_subsys(&smc_net_ops);
if (rc)
return rc;
+ rc = smc_pnet_init();
+ if (rc)
+ goto out_pernet_subsys;
+
rc = smc_llc_init();
if (rc) {
pr_err("%s: smc_llc_init fails with %d\n", __func__, rc);
@@ -1980,6 +2084,9 @@
proto_unregister(&smc_proto);
out_pnet:
smc_pnet_exit();
+out_pernet_subsys:
+ unregister_pernet_subsys(&smc_net_ops);
+
return rc;
}
@@ -1992,6 +2099,7 @@
proto_unregister(&smc_proto6);
proto_unregister(&smc_proto);
smc_pnet_exit();
+ unregister_pernet_subsys(&smc_net_ops);
}
module_init(smc_init);