From mboxrd@z Thu Jan 1 00:00:00 1970 From: Bart Van Assche Subject: [PATCH 05/12] IB/srp: Fix reconnection failure handling Date: Thu, 30 Apr 2015 10:58:57 +0200 Message-ID: <5541EED1.8020702@sandisk.com> References: <5541EE21.3050809@sandisk.com> Mime-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <5541EE21.3050809@sandisk.com> Sender: linux-scsi-owner@vger.kernel.org To: Doug Ledford Cc: James Bottomley , Sagi Grimberg , Sebastian Parschauer , linux-rdma , "linux-scsi@vger.kernel.org" List-Id: linux-rdma@vger.kernel.org Although it is possible to let SRP I/O continue if a reconnect results in a reduction of the number of channels, the current code does not handle this scenario correctly. Instead of making the reconnect code more complex, consider this as a reconnection failure. Signed-off-by: Bart Van Assche Cc: Sagi Grimberg Cc: Sebastian Parschauer Cc: #v3.19 --- drivers/infiniband/ulp/srp/ib_srp.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/drivers/infiniband/ulp/srp/ib_srp.c b/drivers/infiniband/ulp/srp/ib_srp.c index 0eb07d3..5d7d790 100644 --- a/drivers/infiniband/ulp/srp/ib_srp.c +++ b/drivers/infiniband/ulp/srp/ib_srp.c @@ -1246,11 +1246,8 @@ static int srp_rport_reconnect(struct srp_rport *rport) } for (i = 0; i < target->ch_count; i++) { ch = &target->ch[i]; - if (ret || !ch->target) { - if (i > 1) - ret = 0; + if (ret) break; - } ret = srp_connect_ch(ch, multich); multich = true; } -- 2.1.4