From: Spencer Baugh <sbaugh@catern.com>
To: "Nicholas A. Bellinger" <nab@linux-iscsi.org>,
Sagi Grimberg <sagig@mellanox.com>,
Christoph Hellwig <hch@lst.de>,
Bart Van Assche <bart.vanassche@sandisk.com>,
Alexei Potashnik <alexei@purestorage.com>,
Andy Grover <agrover@redhat.com>,
Christophe Vu-Brugier <cvubrugier@fastmail.fm>,
Roland Dreier <roland@purestorage.com>,
"open list:TARGET SUBSYSTEM" <linux-scsi@vger.kernel.org>,
"open list:TARGET SUBSYSTEM" <target-devel@vger.kernel.org>,
open list <linux-kernel@vger.kernel.org>
Cc: Joern Engel <joern@purestorage.com>,
Spencer Baugh <Spencer.baugh@purestorage.com>,
Spencer Baugh <sbaugh@catern.com>
Subject: [PATCH] target: fix crash in cmd tracing when cmd didn't match a LUN
Date: Thu, 23 Jul 2015 15:19:32 -0700 [thread overview]
Message-ID: <1437689973-22991-1-git-send-email-sbaugh@catern.com> (raw)
From: Alexei Potashnik <alexei@purestorage.com>
If command didn't match a LUN and we're sending check condition, the
target_cmd_complete ftrace point will crash because it assumes that
cmd->t_task_cdb has been set.
The fix will temporarily set t_task_cdb to the se_cmd buffer
and copy first 6 bytes of cdb in there as soon as possible.
At a later point t_task_cdb is reset to the correct buffer,
but until then traces and printks don't cause a crash.
Signed-off-by: Alexei Potashnik <alexei@purestorage.com>
Signed-off-by: Spencer Baugh <sbaugh@catern.com>
---
drivers/target/iscsi/iscsi_target.c | 4 ++--
drivers/target/target_core_device.c | 11 +++++++++--
drivers/target/target_core_transport.c | 9 +++++----
include/target/target_core_fabric.h | 2 +-
4 files changed, 17 insertions(+), 9 deletions(-)
diff --git a/drivers/target/iscsi/iscsi_target.c b/drivers/target/iscsi/iscsi_target.c
index f615d75..98899f1 100644
--- a/drivers/target/iscsi/iscsi_target.c
+++ b/drivers/target/iscsi/iscsi_target.c
@@ -1003,8 +1003,8 @@ int iscsit_setup_scsi_cmd(struct iscsi_conn *conn, struct iscsi_cmd *cmd,
target_get_sess_cmd(&cmd->se_cmd, true);
- cmd->sense_reason = transport_lookup_cmd_lun(&cmd->se_cmd,
- scsilun_to_int(&hdr->lun));
+ cmd->sense_reason = transport_lookup_cmd_lun_cdb(&cmd->se_cmd,
+ scsilun_to_int(&hdr->lun), hdr->cdb);
if (cmd->sense_reason)
goto attach_cmd;
diff --git a/drivers/target/target_core_device.c b/drivers/target/target_core_device.c
index c4a8db6..acf84cf 100644
--- a/drivers/target/target_core_device.c
+++ b/drivers/target/target_core_device.c
@@ -56,13 +56,20 @@ static struct se_hba *lun0_hba;
struct se_device *g_lun0_dev;
sense_reason_t
-transport_lookup_cmd_lun(struct se_cmd *se_cmd, u64 unpacked_lun)
+transport_lookup_cmd_lun_cdb(struct se_cmd *se_cmd, u64 unpacked_lun, unsigned char *cdb)
{
struct se_lun *se_lun = NULL;
struct se_session *se_sess = se_cmd->se_sess;
struct se_node_acl *nacl = se_sess->se_node_acl;
struct se_dev_entry *deve;
+ /* Temporarily set t_task_cdb to the se_cmd buffer and save a portion
+ * of cdb in there (fabrics must provide at least 6 bytes). t_task_cdb
+ * will be correctly replaced in target_setup_cmd_from_cdb. Until then
+ * tracing and printks can access t_task_cdb without causing a crash. */
+ se_cmd->t_task_cdb = se_cmd->__t_task_cdb;
+ memcpy(se_cmd->t_task_cdb, cdb, 6);
+
rcu_read_lock();
deve = target_nacl_find_deve(nacl, unpacked_lun);
if (deve) {
@@ -142,7 +149,7 @@ transport_lookup_cmd_lun(struct se_cmd *se_cmd, u64 unpacked_lun)
return 0;
}
-EXPORT_SYMBOL(transport_lookup_cmd_lun);
+EXPORT_SYMBOL(transport_lookup_cmd_lun_cdb);
int transport_lookup_tmr_lun(struct se_cmd *se_cmd, u64 unpacked_lun)
{
diff --git a/drivers/target/target_core_transport.c b/drivers/target/target_core_transport.c
index f6626bb..1f761a3 100644
--- a/drivers/target/target_core_transport.c
+++ b/drivers/target/target_core_transport.c
@@ -1210,15 +1210,16 @@ target_setup_cmd_from_cdb(struct se_cmd *cmd, unsigned char *cdb)
* setup the pointer from __t_task_cdb to t_task_cdb.
*/
if (scsi_command_size(cdb) > sizeof(cmd->__t_task_cdb)) {
- cmd->t_task_cdb = kzalloc(scsi_command_size(cdb),
- GFP_KERNEL);
- if (!cmd->t_task_cdb) {
+ unsigned char *ptr = kzalloc(scsi_command_size(cdb),
+ GFP_KERNEL);
+ if (!ptr) {
pr_err("Unable to allocate cmd->t_task_cdb"
" %u > sizeof(cmd->__t_task_cdb): %lu ops\n",
scsi_command_size(cdb),
(unsigned long)sizeof(cmd->__t_task_cdb));
return TCM_OUT_OF_RESOURCES;
}
+ cmd->t_task_cdb = ptr;
} else
cmd->t_task_cdb = &cmd->__t_task_cdb[0];
/*
@@ -1404,7 +1405,7 @@ int target_submit_cmd_map_sgls(struct se_cmd *se_cmd, struct se_session *se_sess
/*
* Locate se_lun pointer and attach it to struct se_cmd
*/
- rc = transport_lookup_cmd_lun(se_cmd, unpacked_lun);
+ rc = transport_lookup_cmd_lun_cdb(se_cmd, unpacked_lun, cdb);
if (rc) {
transport_send_check_condition_and_sense(se_cmd, rc, 0);
target_put_sess_cmd(se_cmd);
diff --git a/include/target/target_core_fabric.h b/include/target/target_core_fabric.h
index 18afef9..bfa6368 100644
--- a/include/target/target_core_fabric.h
+++ b/include/target/target_core_fabric.h
@@ -116,7 +116,7 @@ void transport_deregister_session(struct se_session *);
void transport_init_se_cmd(struct se_cmd *,
const struct target_core_fabric_ops *,
struct se_session *, u32, int, int, unsigned char *);
-sense_reason_t transport_lookup_cmd_lun(struct se_cmd *, u64);
+sense_reason_t transport_lookup_cmd_lun_cdb(struct se_cmd *, u64, unsigned char *);
sense_reason_t target_setup_cmd_from_cdb(struct se_cmd *, unsigned char *);
int target_submit_cmd_map_sgls(struct se_cmd *, struct se_session *,
unsigned char *, unsigned char *, u64, u32, int, int, int,
--
2.5.0.rc3
next reply other threads:[~2015-07-23 22:19 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-07-23 22:19 Spencer Baugh [this message]
2015-07-24 10:52 ` [PATCH] target: fix crash in cmd tracing when cmd didn't match a LUN Christoph Hellwig
2015-07-24 20:32 ` Nicholas A. Bellinger
2015-07-25 6:48 ` Christoph Hellwig
2015-07-25 9:28 ` Nicholas A. Bellinger
-- strict thread matches above, loose matches on Subject: below --
2015-07-21 22:07 Spencer Baugh
2015-07-21 22:13 ` Spencer Baugh
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=1437689973-22991-1-git-send-email-sbaugh@catern.com \
--to=sbaugh@catern.com \
--cc=Spencer.baugh@purestorage.com \
--cc=agrover@redhat.com \
--cc=alexei@purestorage.com \
--cc=bart.vanassche@sandisk.com \
--cc=cvubrugier@fastmail.fm \
--cc=hch@lst.de \
--cc=joern@purestorage.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-scsi@vger.kernel.org \
--cc=nab@linux-iscsi.org \
--cc=roland@purestorage.com \
--cc=sagig@mellanox.com \
--cc=target-devel@vger.kernel.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