Skip to contentSkip to conversation

fix 2.6.19 data digest calculation bug

PATCH New [email protected] 1 participant 5 messages Dec 17, 2006

Subsystems (2)
SCSI iSCSI
lore
Share & tools
lore

[email protected]

author
#0 lore raw

PATCH 1/5 fix 2.6.19 data digest calculation bug

From: Arne Redlich <[email protected]>

The transition from crypto_digest_*() to the crypto_hash_*() family introduced a bug into the data digest calculation: crypto_hash_update() is called with the number of S/G elements instead of the S/G lists data size.

Signed-off-by Arne Redlich <[email protected]>

-

Signed-off-by Mike Christie <[email protected]>
---
 drivers/scsi/iscsi_tcp.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
drivers/scsi/iscsi_tcp.c +1 -1
index d0b139c..61e3a61 100644
@@ -749,7 +749,7 @@ static int iscsi_scsi_data_in(struct isc
 				if (!offset)
 					crypto_hash_update(
 							&tcp_conn->rx_hash,
-							&sg[i], 1);
+							&sg[i], sg[i].length);
 				else
 					partial_sg_digest_update(
 							&tcp_conn->rx_hash,
Signature · 1 line
1.4.1.1
author
#1 lore raw

PATCH 2/5 iscsi: fix crypto_alloc_hash() error check

From: Akinobu Mita <[email protected]>

The return value of crypto_alloc_hash() should be checked by IS_ERR().

Cc Dmitry Yusupov <[email protected]>
Cc Alex Aizman <[email protected]>
Signed-off-by Akinobu Mita <[email protected]>
Cc James Bottomley <[email protected]>
Signed-off-by Andrew Morton <[email protected]>
Signed-off-by Mike Christie <[email protected]>
---
 drivers/scsi/iscsi_tcp.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)
drivers/scsi/iscsi_tcp.c +2 -2
index 61e3a61..8889686 100644
@@ -1777,13 +1777,13 @@ iscsi_tcp_conn_create(struct iscsi_cls_s
 	tcp_conn->tx_hash.tfm = crypto_alloc_hash("crc32c", 0,
 						  CRYPTO_ALG_ASYNC);
 	tcp_conn->tx_hash.flags = 0;
-	if (!tcp_conn->tx_hash.tfm)
+	if (IS_ERR(tcp_conn->tx_hash.tfm))
 		goto free_tcp_conn;
 
 	tcp_conn->rx_hash.tfm = crypto_alloc_hash("crc32c", 0,
 						  CRYPTO_ALG_ASYNC);
 	tcp_conn->rx_hash.flags = 0;
-	if (!tcp_conn->rx_hash.tfm)
+	if (IS_ERR(tcp_conn->rx_hash.tfm))
 		goto free_tx_tfm;
 
 	return cls_conn;
Signature · 1 line
1.4.1.1
author
#2 lore raw

PATCH 3/5 newline in iscsi printk

From: Meelis Roos <[email protected]>

This patch against 2.6.19 cures two runtogether printk messages in iSCSI driver.

Signed-off-by Mike Christie <[email protected]>
---
 drivers/scsi/scsi_transport_iscsi.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
drivers/scsi/scsi_transport_iscsi.c +1 -1
index 9b25124..2648dbd 100644
@@ -1414,7 +1414,7 @@ static __init int iscsi_transport_init(v
 {
 	int err;
 
-	printk(KERN_INFO "Loading iSCSI transport class v%s.",
+	printk(KERN_INFO "Loading iSCSI transport class v%s.\n",
 		ISCSI_TRANSPORT_VERSION);
 
 	err = class_register(&iscsi_transport_class);
Signature · 1 line
1.4.1.1
author
#3 lore raw

PATCH 4/5 IPv6 and Linux iSCSI initiator

From: FUJITA Tomonori <[email protected]>

By the way, here's a minor patch to simplify IP address print.

Signed-off-by FUJITA Tomonori <[email protected]>
Signed-off-by Mike Christie <[email protected]>
---
 drivers/scsi/iscsi_tcp.c |    6 ++----
 1 files changed, 2 insertions(+), 4 deletions(-)
drivers/scsi/iscsi_tcp.c +2 -4
index 8889686..4376840 100644
@@ -2044,13 +2044,11 @@ iscsi_tcp_conn_get_param(struct iscsi_cl
 		sk = tcp_conn->sock->sk;
 		if (sk->sk_family == PF_INET) {
 			inet = inet_sk(sk);
-			len = sprintf(buf, "%u.%u.%u.%u\n",
+			len = sprintf(buf, NIPQUAD_FMT "\n",
 				      NIPQUAD(inet->daddr));
 		} else {
 			np = inet6_sk(sk);
-			len = sprintf(buf,
-				"%04x:%04x:%04x:%04x:%04x:%04x:%04x:%04x\n",
-				NIP6(np->daddr));
+			len = sprintf(buf, NIP6_FMT "\n", NIP6(np->daddr));
 		}
 		mutex_unlock(&conn->xmitmutex);
 		break;
Signature · 1 line
1.4.1.1
author
#4 lore raw

PATCH 5/5 libiscsi: fix senselen calculation

From: Mike Christie <[email protected]>

Yanling Qi, noted that when the sense data length of
a check-condition is greater than 0x7f (127), senselen = (data[0] << 8) | data[1] will become negative. It causes different kinds of panics from GPF, spin_lock deadlock to spin_lock recursion.

We were also swapping this value on big endien machines.

This patch fixes both issues by using be16_to_cpu().

Signed-off-by Mike Christie <[email protected]>
---
 drivers/scsi/libiscsi.c |    6 +++---
 1 files changed, 3 insertions(+), 3 deletions(-)
drivers/scsi/libiscsi.c +3 -3
index 5d88621..c58ae3b 100644
@@ -260,7 +260,7 @@ static int iscsi_scsi_cmd_rsp(struct isc
 	}
 
 	if (rhdr->cmd_status == SAM_STAT_CHECK_CONDITION) {
-		int senselen;
+		uint16_t senselen;
 
 		if (datalen < 2) {
 invalid_datalen:

@@ -270,12 +270,12 @@ invalid_datalen:
 			goto out;
 		}
 
-		senselen = (data[0] << 8) | data[1];
+		senselen = be16_to_cpu(*(uint16_t *)data);
 		if (datalen < senselen)
 			goto invalid_datalen;
 
 		memcpy(sc->sense_buffer, data + 2,
-		       min(senselen, SCSI_SENSE_BUFFERSIZE));
+		       min_t(uint16_t, senselen, SCSI_SENSE_BUFFERSIZE));
 		debug_scsi("copied %d bytes of sense\n",
 			   min(senselen, SCSI_SENSE_BUFFERSIZE));
 	}
Signature · 1 line
1.4.1.1