linux-rdma.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Bart Van Assche <bart.vanassche-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
To: Doug Ledford <dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
Cc: Christoph Hellwig <hch-jcswGhMUV9g@public.gmane.org>,
	Sagi Grimberg <sagig-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>,
	Alex Estrin <alex.estrin-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>,
	"linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org"
	<linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org>
Subject: [PATCH v4 21/21] IB/srpt: Fix wait list processing
Date: Thu, 11 Feb 2016 11:10:09 -0800	[thread overview]
Message-ID: <56BCDC91.4060307@sandisk.com> (raw)
In-Reply-To: <56BCDAAD.7030906-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>

Since the wait list is not protected against concurrent access
it must be processed from the context of the completion handler.
Replace the wait list processing code in the IB CM RTU callback
handler by code that triggers a completion handler. This patch
fixes the following rare crash:

WARNING: CPU: 2 PID: 78656 at lib/list_debug.c:53 __list_del_entry+0x67/0xd0()
list_del corruption, ffff88041ae404b8->next is LIST_POISON1 (dead000000000100)
Call Trace:
 [<ffffffff81251c6b>] dump_stack+0x4f/0x74
 [<ffffffff810574ab>] warn_slowpath_common+0x8b/0xd0
 [<ffffffff81057591>] warn_slowpath_fmt+0x41/0x70
 [<ffffffff8126f007>] __list_del_entry+0x67/0xd0
 [<ffffffff8126f081>] list_del+0x11/0x40
 [<ffffffffa0265242>] srpt_cm_handler+0x172/0x1a4 [ib_srpt]
 [<ffffffffa0370370>] cm_process_work+0x20/0xf0 [ib_cm]
 [<ffffffffa0370dae>] cm_establish_handler+0xbe/0x110 [ib_cm]
 [<ffffffffa03733e7>] cm_work_handler+0x67/0xd0 [ib_cm]
 [<ffffffff8107184d>] process_one_work+0x1bd/0x460
 [<ffffffff81073148>] worker_thread+0x118/0x420
 [<ffffffff81078444>] kthread+0xe4/0x100
 [<ffffffff8151caff>] ret_from_fork+0x3f/0x70

Signed-off-by: Bart Van Assche <bart.vanassche-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
Reviewed-by: Christoph Hellwig <hch-jcswGhMUV9g@public.gmane.org>
Reviewed-by: Alex Estrin <alex.estrin-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Cc: Sagi Grimberg <sagig-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---
 drivers/infiniband/ulp/srpt/ib_srpt.c | 24 +++++++++++++++---------
 1 file changed, 15 insertions(+), 9 deletions(-)

diff --git a/drivers/infiniband/ulp/srpt/ib_srpt.c b/drivers/infiniband/ulp/srpt/ib_srpt.c
index 9d365ea..d0e14d4 100644
--- a/drivers/infiniband/ulp/srpt/ib_srpt.c
+++ b/drivers/infiniband/ulp/srpt/ib_srpt.c
@@ -96,7 +96,7 @@ static void srpt_free_ch(struct kref *kref);
 static int srpt_queue_status(struct se_cmd *cmd);
 static void srpt_recv_done(struct ib_cq *cq, struct ib_wc *wc);
 static void srpt_send_done(struct ib_cq *cq, struct ib_wc *wc);
-static void srpt_zerolength_write_done(struct ib_cq *cq, struct ib_wc *wc);
+static void srpt_process_wait_list(struct srpt_rdma_ch *ch);
 
 /*
  * The only allowed channel state changes are those that change the channel
@@ -833,12 +833,14 @@ static void srpt_zerolength_write_done(struct ib_cq *cq, struct ib_wc *wc)
 {
 	struct srpt_rdma_ch *ch = cq->cq_context;
 
-	WARN(wc->status == IB_WC_SUCCESS, "%s-%d: QP not in error state\n",
-	     ch->sess_name, ch->qp->qp_num);
-	if (srpt_set_ch_state(ch, CH_DISCONNECTED))
-		schedule_work(&ch->release_work);
-	else
-		WARN_ONCE("%s-%d\n", ch->sess_name, ch->qp->qp_num);
+	if (wc->status == IB_WC_SUCCESS) {
+		srpt_process_wait_list(ch);
+	} else {
+		if (srpt_set_ch_state(ch, CH_DISCONNECTED))
+			schedule_work(&ch->release_work);
+		else
+			WARN_ONCE("%s-%d\n", ch->sess_name, ch->qp->qp_num);
+	}
 }
 
 /**
@@ -2323,9 +2325,13 @@ static void srpt_cm_rtu_recv(struct srpt_rdma_ch *ch)
 	if (srpt_set_ch_state(ch, CH_LIVE)) {
 		ret = srpt_ch_qp_rts(ch, ch->qp);
 
-		srpt_process_wait_list(ch);
-		if (ret)
+		if (ret == 0) {
+			/* Trigger wait list processing. */
+			ret = srpt_zerolength_write(ch);
+			WARN_ONCE(ret < 0, "%d\n", ret);
+		} else {
 			srpt_close_ch(ch);
+		}
 	}
 }
 
-- 
2.7.0

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

  parent reply	other threads:[~2016-02-11 19:10 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-02-11 19:02 [PATCH v4 00/21] IB/srpt patches for Linux kernel v4.6 Bart Van Assche
     [not found] ` <56BCDAAD.7030906-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-02-11 19:03   ` [PATCH v4 01/21] IB/srpt: Simplify srpt_handle_tsk_mgmt() Bart Van Assche
2016-02-11 19:03   ` [PATCH v4 02/21] IB/srpt: Add parentheses around sizeof argument Bart Van Assche
2016-02-11 19:04   ` [PATCH v4 03/21] IB/srpt: Remove struct srpt_node_acl Bart Van Assche
2016-02-11 19:04   ` [PATCH v4 04/21] IB/srpt: Inline srpt_sdev_name() Bart Van Assche
2016-02-11 19:04   ` [PATCH v4 05/21] IB/srpt: Inline srpt_get_ch_state() Bart Van Assche
2016-02-11 19:05   ` [PATCH v4 06/21] IB/srpt: Introduce target_reverse_dma_direction() Bart Van Assche
2016-02-11 19:05   ` [PATCH v4 07/21] IB/srpt: Use scsilun_to_int() Bart Van Assche
2016-02-11 19:05   ` [PATCH v4 08/21] IB/srpt: Simplify channel state management Bart Van Assche
2016-02-11 19:05   ` [PATCH v4 09/21] IB/srpt: Simplify srpt_shutdown_session() Bart Van Assche
2016-02-11 19:06   ` [PATCH v4 10/21] IB/srpt: Fix srpt_close_session() Bart Van Assche
2016-02-11 19:06   ` [PATCH v4 11/21] IB/srpt: Fix srpt_handle_cmd() error paths Bart Van Assche
2016-02-11 19:07   ` [PATCH v4 12/21] IB/srpt: Fix how aborted commands are processed Bart Van Assche
2016-02-11 19:07   ` [PATCH v4 13/21] IB/srpt: Inline trivial CM callback functions Bart Van Assche
2016-02-11 19:07   ` [PATCH v4 14/21] IB/srpt: Eliminate srpt_find_channel() Bart Van Assche
2016-02-11 19:08   ` [PATCH v4 15/21] IB/srpt: Log private data associated with REJ Bart Van Assche
     [not found]     ` <56BCDC1C.107-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-02-17 15:57       ` Doug Ledford
     [not found]         ` <56C49854.6010204-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2016-02-17 23:24           ` Bart Van Assche
     [not found]             ` <56C50110.7060003-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-02-18 15:14               ` Doug Ledford
2016-02-11 19:08   ` [PATCH v4 16/21] IB/srpt: Use a mutex to protect the channel list Bart Van Assche
2016-02-11 19:08   ` [PATCH v4 17/21] IB/srpt: Detect session shutdown reliably Bart Van Assche
2016-02-11 19:09   ` [PATCH v4 18/21] IB/srpt: Fix srpt_write_pending() Bart Van Assche
2016-02-11 19:09   ` [PATCH v4 19/21] IB/srpt: Log out all initiators if a port is disabled Bart Van Assche
2016-02-11 19:09   ` [PATCH v4 20/21] IB/srpt: Introduce srpt_process_wait_list() Bart Van Assche
2016-02-11 19:10   ` Bart Van Assche [this message]
2016-02-17 16:18   ` [PATCH v4 00/21] IB/srpt patches for Linux kernel v4.6 Doug Ledford

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=56BCDC91.4060307@sandisk.com \
    --to=bart.vanassche-xdaiopvojttbdgjk7y7tuq@public.gmane.org \
    --cc=alex.estrin-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org \
    --cc=dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org \
    --cc=hch-jcswGhMUV9g@public.gmane.org \
    --cc=linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \
    --cc=sagig-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).