From: Karsten Graul <kgraul@linux.ibm.com>
Subject: net/smc: short wait for late smc_clc_wait_msg
Patch-mainline: v5.0-rc1
Git-commit: 2b59f58e34e78abec2bccd2413ae9e7ea509a855
References: FATE#325698, LTC#167867, bsc#1113481
Description: net/smc: bugfix and compatibility patches
Symptom: Random hangs in smc processing:
user space application hangs in socket send() or recv() call or
does never get a notification from a select() call.
Missing compatibility to other platforms:
confirm rkey and delete rkey processing is required by the
design, but delete rkey processing is missing. This leads to
protocol failures when communicating with other platforms like
zOS. The SMC-D shutdown signal support is missing, so there is
no detection if the remote peer closed the link group.
Broken administration of available WR send payload buffers due to
a use-after-free condition.
Problem: Misbehaviour regarding the user space api can lead to hang
situations. SMC is not fully compatible to some other platforms
due to missing rkey processing and SMC-D shutdown signal support.
Solution: Fixed protocoll deficiencies by implementing the required rkey
processing. For SMC-D, the cursors are now handled atomically to
handle parallel modifications. The SMC-D shutdown signal is now
processed when received and sent to the remote peer if needed.
Prereq patches are included.
Reproduction: Run SMC on a loaded system against zOS as peer system.
Upstream-Description:
net/smc: short wait for late smc_clc_wait_msg
After sending one of the initial LLC messages CONFIRM LINK or
ADD LINK, there is already a wait for the LLC response. It does
not make sense to wait another long time for a CLC DECLINE. Thus
this patch introduces a shorter wait time for these cases.
Signed-off-by: Ursula Braun <ubraun@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Karsten Graul <kgraul@linux.ibm.com>
Acked-by: Petr Tesarik <ptesarik@suse.com>
---
net/smc/af_smc.c | 15 ++++++++-------
net/smc/smc_clc.c | 6 +++---
net/smc/smc_clc.h | 3 ++-
3 files changed, 13 insertions(+), 11 deletions(-)
--- a/net/smc/af_smc.c
+++ b/net/smc/af_smc.c
@@ -337,7 +337,7 @@ static int smc_clnt_conf_first_link(stru
struct smc_clc_msg_decline dclc;
rc = smc_clc_wait_msg(smc, &dclc, sizeof(dclc),
- SMC_CLC_DECLINE);
+ SMC_CLC_DECLINE, CLC_WAIT_TIME_SHORT);
return rc == -EAGAIN ? SMC_CLC_DECL_TIMEOUT_CL : rc;
}
@@ -365,7 +365,7 @@ static int smc_clnt_conf_first_link(stru
struct smc_clc_msg_decline dclc;
rc = smc_clc_wait_msg(smc, &dclc, sizeof(dclc),
- SMC_CLC_DECLINE);
+ SMC_CLC_DECLINE, CLC_WAIT_TIME_SHORT);
return rc == -EAGAIN ? SMC_CLC_DECL_TIMEOUT_AL : rc;
}
@@ -535,7 +535,8 @@ static int smc_connect_clc(struct smc_so
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 */
@@ -968,7 +969,7 @@ static int smc_serv_conf_first_link(stru
struct smc_clc_msg_decline dclc;
rc = smc_clc_wait_msg(smc, &dclc, sizeof(dclc),
- SMC_CLC_DECLINE);
+ SMC_CLC_DECLINE, CLC_WAIT_TIME_SHORT);
return rc == -EAGAIN ? SMC_CLC_DECL_TIMEOUT_CL : rc;
}
@@ -989,7 +990,7 @@ static int smc_serv_conf_first_link(stru
struct smc_clc_msg_decline dclc;
rc = smc_clc_wait_msg(smc, &dclc, sizeof(dclc),
- SMC_CLC_DECLINE);
+ SMC_CLC_DECLINE, CLC_WAIT_TIME_SHORT);
return rc == -EAGAIN ? SMC_CLC_DECL_TIMEOUT_AL : rc;
}
@@ -1225,7 +1226,7 @@ static void smc_listen_work(struct work_
*/
pclc = (struct smc_clc_msg_proposal *)&buf;
reason_code = smc_clc_wait_msg(new_smc, pclc, SMC_CLC_MAX_LEN,
- SMC_CLC_PROPOSAL);
+ SMC_CLC_PROPOSAL, CLC_WAIT_TIME);
if (reason_code) {
smc_listen_decline(new_smc, reason_code, 0);
return;
@@ -1275,7 +1276,7 @@ static void smc_listen_work(struct work_
/* receive SMC Confirm CLC message */
reason_code = smc_clc_wait_msg(new_smc, &cclc, sizeof(cclc),
- SMC_CLC_CONFIRM);
+ SMC_CLC_CONFIRM, CLC_WAIT_TIME);
if (reason_code) {
mutex_unlock(&smc_create_lgr_pending);
smc_listen_decline(new_smc, reason_code, local_contact);
--- a/net/smc/smc_clc.c
+++ b/net/smc/smc_clc.c
@@ -264,7 +264,7 @@ out:
* clcsock error, -EINTR, -ECONNRESET, -EPROTO otherwise.
*/
int smc_clc_wait_msg(struct smc_sock *smc, void *buf, int buflen,
- u8 expected_type)
+ u8 expected_type, unsigned long timeout)
{
long rcvtimeo = smc->clcsock->sk->sk_rcvtimeo;
struct sock *clc_sk = smc->clcsock->sk;
@@ -284,7 +284,7 @@ int smc_clc_wait_msg(struct smc_sock *sm
* sizeof(struct smc_clc_msg_hdr)
*/
krflags = MSG_PEEK | MSG_WAITALL;
- smc->clcsock->sk->sk_rcvtimeo = CLC_WAIT_TIME;
+ clc_sk->sk_rcvtimeo = timeout;
iov_iter_kvec(&msg.msg_iter, READ | ITER_KVEC, &vec, 1,
sizeof(struct smc_clc_msg_hdr));
len = sock_recvmsg(smc->clcsock, &msg, krflags);
@@ -350,7 +350,7 @@ int smc_clc_wait_msg(struct smc_sock *sm
}
out:
- smc->clcsock->sk->sk_rcvtimeo = rcvtimeo;
+ clc_sk->sk_rcvtimeo = rcvtimeo;
return reason_code;
}
--- a/net/smc/smc_clc.h
+++ b/net/smc/smc_clc.h
@@ -26,6 +26,7 @@
#define SMC_TYPE_D 1 /* SMC-D only */
#define SMC_TYPE_B 3 /* SMC-R and SMC-D */
#define CLC_WAIT_TIME (6 * HZ) /* max. wait time on clcsock */
+#define CLC_WAIT_TIME_SHORT HZ /* short wait time on clcsock */
#define SMC_CLC_DECL_MEM 0x01010000 /* insufficient memory resources */
#define SMC_CLC_DECL_TIMEOUT_CL 0x02010000 /* timeout w4 QP confirm link */
#define SMC_CLC_DECL_TIMEOUT_AL 0x02020000 /* timeout w4 QP add link */
@@ -181,7 +182,7 @@ struct smcd_dev;
int smc_clc_prfx_match(struct socket *clcsock,
struct smc_clc_msg_proposal_prefix *prop);
int smc_clc_wait_msg(struct smc_sock *smc, void *buf, int buflen,
- u8 expected_type);
+ u8 expected_type, unsigned long timeout);
int smc_clc_send_decline(struct smc_sock *smc, u32 peer_diag_info);
int smc_clc_send_proposal(struct smc_sock *smc, int smc_type,
struct smc_ib_device *smcibdev, u8 ibport, u8 gid[],