linux-scsi.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Nicholas A. Bellinger" <nab@linux-iscsi.org>
To: linux-kernel <linux-kernel@vger.kernel.org>,
	linux-scsi <linux-scsi@vger.kernel.org>,
	James Bottomley <James.Bottomley@HansenPartnership.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
	Christoph Hellwig <hch@lst.de>,
	Kiran Patil <kiran.patil@intel.com>,
	"Robert W. Love" <robert.w.love@intel.com>,
	FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>,
	Bart Van Assche <bvanassche@acm.org>,
	Hannes Reinecke <hare@suse.de>,
	Mike Christie <michaelc@cs.wisc.edu>,
	Andy Grover <agrover@redhat.com>,
	Nicholas Bellinger <nab@linux-iscsi.org>
Subject: [PATCH 1/3] target: Fix multi task->task_sg[] chaining logic bug
Date: Fri,  6 May 2011 19:54:56 -0700	[thread overview]
Message-ID: <1304736898-10830-2-git-send-email-nab@linux-iscsi.org> (raw)
In-Reply-To: <1304736898-10830-1-git-send-email-nab@linux-iscsi.org>

From: Nicholas Bellinger <nab@linux-iscsi.org>

This patch fixes a bug in transport_do_task_sg_chain() used by HW target
mode modules with sg_chain() to provide a single sg_next() walkable memory
layout for use with pci_map_sg() and friends.  This patch addresses an
issue with mapping multiple small block max_sector tasks across multiple
struct se_task->task_sg[] mappings for HW target mode operation.

This was causing OOPs with (cmd->t_task->t_tasks_no > 1) I/O traffic for
HW target drivers using transport_do_task_sg_chain(), and has been tested
so far with tcm_fc(openfcoe), tcm_qla2xxx, and ib_srpt fabrics with
t_tasks_no > 1 IBLOCK backends using a smaller max_sectors to trigger the
original issue.

Reported-by: Kiran Patil <kiran.patil@intel.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
---
 drivers/target/target_core_transport.c |   26 +++++++++++++++-----------
 1 files changed, 15 insertions(+), 11 deletions(-)

diff --git a/drivers/target/target_core_transport.c b/drivers/target/target_core_transport.c
index 9583b23..fefe10a 100644
--- a/drivers/target/target_core_transport.c
+++ b/drivers/target/target_core_transport.c
@@ -4776,18 +4776,20 @@ void transport_do_task_sg_chain(struct se_cmd *cmd)
 				sg_end_cur->page_link &= ~0x02;
 
 				sg_chain(sg_head, task_sg_num, sg_head_cur);
-				sg_count += (task->task_sg_num + 1);
-			} else
 				sg_count += task->task_sg_num;
+				task_sg_num = (task->task_sg_num + 1);
+			} else {
+				sg_chain(sg_head, task_sg_num, sg_head_cur);
+				sg_count += task->task_sg_num;
+				task_sg_num = task->task_sg_num;
+			}
 
 			sg_head = sg_head_cur;
 			sg_link = sg_link_cur;
-			task_sg_num = task->task_sg_num;
 			continue;
 		}
 		sg_head = sg_first = &task->task_sg[0];
 		sg_link = &task->task_sg[task->task_sg_num];
-		task_sg_num = task->task_sg_num;
 		/*
 		 * Check for single task..
 		 */
@@ -4798,9 +4800,12 @@ void transport_do_task_sg_chain(struct se_cmd *cmd)
 			 */
 			sg_end = &task->task_sg[task->task_sg_num - 1];
 			sg_end->page_link &= ~0x02;
-			sg_count += (task->task_sg_num + 1);
-		} else
 			sg_count += task->task_sg_num;
+			task_sg_num = (task->task_sg_num + 1);
+		} else {
+			sg_count += task->task_sg_num;
+			task_sg_num = task->task_sg_num;
+		}
 	}
 	/*
 	 * Setup the starting pointer and total t_tasks_sg_linked_no including
@@ -4809,21 +4814,20 @@ void transport_do_task_sg_chain(struct se_cmd *cmd)
 	T_TASK(cmd)->t_tasks_sg_chained = sg_first;
 	T_TASK(cmd)->t_tasks_sg_chained_no = sg_count;
 
-	DEBUG_CMD_M("Setup T_TASK(cmd)->t_tasks_sg_chained: %p and"
-		" t_tasks_sg_chained_no: %u\n", T_TASK(cmd)->t_tasks_sg_chained,
+	DEBUG_CMD_M("Setup cmd: %p T_TASK(cmd)->t_tasks_sg_chained: %p and"
+		" t_tasks_sg_chained_no: %u\n", cmd, T_TASK(cmd)->t_tasks_sg_chained,
 		T_TASK(cmd)->t_tasks_sg_chained_no);
 
 	for_each_sg(T_TASK(cmd)->t_tasks_sg_chained, sg,
 			T_TASK(cmd)->t_tasks_sg_chained_no, i) {
 
-		DEBUG_CMD_M("SG: %p page: %p length: %d offset: %d\n",
-			sg, sg_page(sg), sg->length, sg->offset);
+		DEBUG_CMD_M("SG[%d]: %p page: %p length: %d offset: %d, magic: 0x%08x\n",
+			i, sg, sg_page(sg), sg->length, sg->offset, sg->sg_magic);
 		if (sg_is_chain(sg))
 			DEBUG_CMD_M("SG: %p sg_is_chain=1\n", sg);
 		if (sg_is_last(sg))
 			DEBUG_CMD_M("SG: %p sg_is_last=1\n", sg);
 	}
-
 }
 EXPORT_SYMBOL(transport_do_task_sg_chain);
 
-- 
1.7.5.1


  reply	other threads:[~2011-05-07  2:55 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-05-07  2:54 [PATCH 0/3] target: Bugfixes for .39-rc7 Nicholas A. Bellinger
2011-05-07  2:54 ` Nicholas A. Bellinger [this message]
2011-05-07  2:54 ` [PATCH 2/3] target: Fix interrupt context bug with stats_lock and core_tmr_alloc_req Nicholas A. Bellinger
2011-05-07  2:54 ` [PATCH 3/3] target: Fix bug with task_sg chained transport_free_dev_tasks release Nicholas A. Bellinger

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=1304736898-10830-2-git-send-email-nab@linux-iscsi.org \
    --to=nab@linux-iscsi.org \
    --cc=James.Bottomley@HansenPartnership.com \
    --cc=agrover@redhat.com \
    --cc=bvanassche@acm.org \
    --cc=fujita.tomonori@lab.ntt.co.jp \
    --cc=hare@suse.de \
    --cc=hch@lst.de \
    --cc=kiran.patil@intel.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-scsi@vger.kernel.org \
    --cc=michaelc@cs.wisc.edu \
    --cc=robert.w.love@intel.com \
    --cc=torvalds@linux-foundation.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).