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>,
"linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org"
<linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org>
Subject: [PATCH 11/15] IB/srpt: Fix how aborted commands are processed
Date: Tue, 5 Jan 2016 15:25:49 +0100 [thread overview]
Message-ID: <568BD26D.9080003@sandisk.com> (raw)
In-Reply-To: <568BD0FC.70207-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
srpt_abort_cmd() must not be called in state SRPT_STATE_DATA_IN. Issue
a warning if this occurs.
srpt_abort_cmd() must not invoke target_put_sess_cmd() for commands
in state SRPT_STATE_DONE because the srpt_abort_cmd() callers already
do this when necessary. Hence remove this call.
If an RDMA read fails the corresponding SCSI command must fail. Hence
add a transport_generic_request_failure() call.
Remove an incorrect srpt_abort_cmd() call from srpt_rdma_write_done().
Avoid that srpt_send_done() calls srpt_abort_cmd() for finished SCSI
commands.
Signed-off-by: Bart Van Assche <bart.vanassche-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
Cc: Christoph Hellwig <hch-jcswGhMUV9g@public.gmane.org>
---
drivers/infiniband/ulp/srpt/ib_srpt.c | 39 ++++++++++++++---------------------
1 file changed, 15 insertions(+), 24 deletions(-)
diff --git a/drivers/infiniband/ulp/srpt/ib_srpt.c b/drivers/infiniband/ulp/srpt/ib_srpt.c
index 395bc1b..2ae3c1b 100644
--- a/drivers/infiniband/ulp/srpt/ib_srpt.c
+++ b/drivers/infiniband/ulp/srpt/ib_srpt.c
@@ -1273,25 +1273,17 @@ static int srpt_abort_cmd(struct srpt_send_ioctx *ioctx)
case SRPT_STATE_NEED_DATA:
ioctx->state = SRPT_STATE_DATA_IN;
break;
- case SRPT_STATE_DATA_IN:
case SRPT_STATE_CMD_RSP_SENT:
case SRPT_STATE_MGMT_RSP_SENT:
ioctx->state = SRPT_STATE_DONE;
break;
default:
+ WARN_ONCE(true, "%s: unexpected I/O context state %d\n",
+ __func__, state);
break;
}
spin_unlock_irqrestore(&ioctx->spinlock, flags);
- if (state == SRPT_STATE_DONE) {
- struct srpt_rdma_ch *ch = ioctx->ch;
-
- BUG_ON(ch->sess == NULL);
-
- target_put_sess_cmd(&ioctx->cmd);
- goto out;
- }
-
pr_debug("Aborting cmd with state %d and tag %lld\n", state,
ioctx->cmd.tag);
@@ -1299,14 +1291,16 @@ static int srpt_abort_cmd(struct srpt_send_ioctx *ioctx)
case SRPT_STATE_NEW:
case SRPT_STATE_DATA_IN:
case SRPT_STATE_MGMT:
+ case SRPT_STATE_DONE:
/*
* Do nothing - defer abort processing until
* srpt_queue_response() is invoked.
*/
- WARN_ON(!(ioctx->cmd.transport_state & CMD_T_ABORTED));
break;
case SRPT_STATE_NEED_DATA:
- /* DMA_TO_DEVICE (write) - RDMA read error. */
+ pr_debug("tag %#llx: RDMA read error\n", ioctx->cmd.tag);
+ transport_generic_request_failure(&ioctx->cmd,
+ TCM_CHECK_CONDITION_ABORT_CMD);
break;
case SRPT_STATE_CMD_RSP_SENT:
/*
@@ -1314,18 +1308,16 @@ static int srpt_abort_cmd(struct srpt_send_ioctx *ioctx)
* not been received in time.
*/
srpt_unmap_sg_to_ib_sge(ioctx->ch, ioctx);
- target_put_sess_cmd(&ioctx->cmd);
+ transport_generic_free_cmd(&ioctx->cmd, 0);
break;
case SRPT_STATE_MGMT_RSP_SENT:
- srpt_set_cmd_state(ioctx, SRPT_STATE_DONE);
- target_put_sess_cmd(&ioctx->cmd);
+ transport_generic_free_cmd(&ioctx->cmd, 0);
break;
default:
WARN(1, "Unexpected command state (%d)", state);
break;
}
-out:
return state;
}
@@ -1365,9 +1357,14 @@ static void srpt_rdma_write_done(struct ib_cq *cq, struct ib_wc *wc)
container_of(wc->wr_cqe, struct srpt_send_ioctx, rdma_cqe);
if (unlikely(wc->status != IB_WC_SUCCESS)) {
+ /*
+ * Note: if an RDMA write error completion is received that
+ * means that a SEND also has been posted. Defer further
+ * processing of the associated command until the send error
+ * completion has been received.
+ */
pr_info("RDMA_WRITE for ioctx 0x%p failed with status %d\n",
ioctx, wc->status);
- srpt_abort_cmd(ioctx);
}
}
@@ -1716,15 +1713,10 @@ static void srpt_send_done(struct ib_cq *cq, struct ib_wc *wc)
atomic_inc(&ch->sq_wr_avail);
- if (wc->status != IB_WC_SUCCESS) {
+ if (wc->status != IB_WC_SUCCESS)
pr_info("sending response for ioctx 0x%p failed"
" with status %d\n", ioctx, wc->status);
- atomic_dec(&ch->req_lim);
- srpt_abort_cmd(ioctx);
- goto out;
- }
-
if (state != SRPT_STATE_DONE) {
srpt_unmap_sg_to_ib_sge(ch, ioctx);
transport_generic_free_cmd(&ioctx->cmd, 0);
@@ -1733,7 +1725,6 @@ static void srpt_send_done(struct ib_cq *cq, struct ib_wc *wc)
" wr_id = %u.\n", ioctx->ioctx.index);
}
-out:
while (!list_empty(&ch->cmd_wait_list) &&
ch->state == CH_LIVE &&
(ioctx = srpt_get_send_ioctx(ch)) != NULL) {
--
2.1.4
--
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
next prev parent reply other threads:[~2016-01-05 14:25 UTC|newest]
Thread overview: 57+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-01-05 14:19 [PATCH 00/15] Various ib_srpt patches Bart Van Assche
[not found] ` <568BD0FC.70207-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-01-05 14:20 ` [PATCH 01/15] IB/srpt: Add parentheses around sizeof argument Bart Van Assche
[not found] ` <568BD142.3070900-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-01-06 4:58 ` Christoph Hellwig
2016-01-06 13:58 ` Sagi Grimberg
2016-01-05 14:21 ` [PATCH 02/15] IB/srpt: Inline srpt_sdev_name() Bart Van Assche
[not found] ` <568BD15F.6010909-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-01-06 4:58 ` Christoph Hellwig
2016-01-06 13:58 ` Sagi Grimberg
2016-01-05 14:21 ` [PATCH 03/15] IB/srpt: Inline srpt_get_ch_state() Bart Van Assche
[not found] ` <568BD181.5040009-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-01-06 4:58 ` Christoph Hellwig
2016-01-06 13:59 ` Sagi Grimberg
2016-01-05 14:22 ` [PATCH 04/15] IB/srpt: Introduce target_reverse_dma_direction() Bart Van Assche
[not found] ` <568BD19C.6060404-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-01-06 4:59 ` Christoph Hellwig
2016-01-06 14:00 ` Sagi Grimberg
2016-01-05 14:22 ` [PATCH 05/15] IB/srpt: Use scsilun_to_int() Bart Van Assche
[not found] ` <568BD1B6.4090308-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-01-06 4:59 ` Christoph Hellwig
2016-01-06 14:01 ` Sagi Grimberg
2016-01-05 14:23 ` [PATCH 06/15] IB/srpt: Simplify srpt_handle_tsk_mgmt() Bart Van Assche
[not found] ` <568BD1D2.1030609-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-01-06 5:00 ` Christoph Hellwig
2016-01-06 14:02 ` Sagi Grimberg
2016-01-26 16:57 ` Estrin, Alex
[not found] ` <F3529576D8E232409F431C309E29399328F6B06A-8k97q/ur5Z1cIJlls4ac1rfspsVTdybXVpNB7YpNyf8@public.gmane.org>
2016-01-26 17:32 ` Bart Van Assche
[not found] ` <56A7ADA3.5080507-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-01-26 18:35 ` Estrin, Alex
[not found] ` <F3529576D8E232409F431C309E29399328F6B0C3-8k97q/ur5Z1cIJlls4ac1rfspsVTdybXVpNB7YpNyf8@public.gmane.org>
2016-01-26 18:46 ` Bart Van Assche
[not found] ` <56A7BF23.1060306-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-01-26 18:52 ` Estrin, Alex
2016-01-05 14:23 ` [PATCH 07/15] IB/srpt: Simplify channel state management Bart Van Assche
[not found] ` <568BD1F1.5050107-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-01-06 5:01 ` Christoph Hellwig
2016-01-06 14:08 ` Sagi Grimberg
2016-01-05 14:24 ` [PATCH 08/15] IB/srpt: Simplify srpt_shutdown_session() Bart Van Assche
[not found] ` <568BD20F.7020101-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-01-06 5:07 ` Christoph Hellwig
2016-01-06 14:13 ` Sagi Grimberg
2016-01-05 14:24 ` [PATCH 09/15] IB/srpt: Fix srpt_close_session() Bart Van Assche
[not found] ` <568BD231.10205-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-01-06 5:08 ` Christoph Hellwig
2016-01-06 14:21 ` Sagi Grimberg
[not found] ` <568D22F6.4050803-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
2016-01-06 14:34 ` Bart Van Assche
[not found] ` <568D25D9.3050809-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-01-06 14:46 ` Sagi Grimberg
2016-01-05 14:25 ` [PATCH 10/15] IB/srpt: Fix srpt_handle_cmd() error paths Bart Van Assche
[not found] ` <568BD249.1000100-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-01-06 5:09 ` Christoph Hellwig
2016-01-06 14:31 ` Sagi Grimberg
[not found] ` <568D2553.8050302-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
2016-01-06 14:36 ` Bart Van Assche
2016-01-05 14:25 ` Bart Van Assche [this message]
[not found] ` <568BD26D.9080003-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-01-06 5:13 ` [PATCH 11/15] IB/srpt: Fix how aborted commands are processed Christoph Hellwig
[not found] ` <20160106051305.GK15574-jcswGhMUV9g@public.gmane.org>
2016-01-06 13:30 ` Bart Van Assche
2016-01-05 14:26 ` [PATCH 12/15] IB/srpt: Eliminate srpt_find_channel() Bart Van Assche
[not found] ` <568BD28E.40302-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-01-06 5:15 ` Christoph Hellwig
2016-01-06 14:33 ` Sagi Grimberg
2016-01-05 14:26 ` [PATCH 13/15] IB/srpt: Detect session shutdown reliably Bart Van Assche
[not found] ` <568BD2A9.6070600-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-01-06 5:21 ` Christoph Hellwig
[not found] ` <20160106052138.GM15574-jcswGhMUV9g@public.gmane.org>
2016-01-06 13:34 ` Bart Van Assche
2016-01-06 14:39 ` Sagi Grimberg
[not found] ` <568D2707.9020309-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
2016-01-06 14:46 ` Bart Van Assche
[not found] ` <568D28CA.8010804-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-01-06 17:04 ` Christoph Hellwig
2016-01-05 14:27 ` [PATCH 14/15] IB/srpt: Fix srpt_write_pending() Bart Van Assche
[not found] ` <568BD2C2.6090808-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-01-06 5:22 ` Christoph Hellwig
2016-01-06 14:41 ` Sagi Grimberg
2016-01-05 14:27 ` [PATCH 15/15] IB/srpt: Fix a rare crash in srpt_close_session() Bart Van Assche
[not found] ` <568BD2E1.1060701-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-01-06 5:23 ` Christoph Hellwig
2016-01-06 14:47 ` Sagi Grimberg
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=568BD26D.9080003@sandisk.com \
--to=bart.vanassche-xdaiopvojttbdgjk7y7tuq@public.gmane.org \
--cc=dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org \
--cc=hch-jcswGhMUV9g@public.gmane.org \
--cc=linux-rdma-u79uwXL29TY76Z2rM5mHXA@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).