Skip to contentSkip to conversation

s390/dasd patches

PATCH In review Stefan Haberland 3 participants 7 messages Dec 19, 2019

Subsystems (4)
lore
Share & tools
lore

Stefan Haberland

author maintainer
#0 lore raw

cover s390/dasd patches

Hi Jens,

please see the following three patches that

- fix a dead code path for DASD
- fix a memleak in DASD error case
- fix typo in copyright statement

Regards,
Stefan

Jan Höppner (1):
s390/dasd/cio: Interpret ccw_device_get_mdc return value correctly

Stefan Haberland (2):
s390/dasd: fix memleak in path handling error case s390/dasd: fix typo in copyright statement

 drivers/s390/block/dasd_eckd.c | 28 +++++++---------------------
 drivers/s390/block/dasd_fba.h  |  2 +-
 drivers/s390/block/dasd_proc.c |  2 +-
 drivers/s390/cio/device_ops.c  |  2 +-
 4 files changed, 10 insertions(+), 24 deletions(-)
Signature · 1 line
2.17.1
author maintainer
#1 lore raw

PATCH 1/3 s390/dasd/cio: Interpret ccw_device_get_mdc return value correctly

From: Jan Höppner <[email protected]>

The max data count (mdc) is an unsigned 16-bit integer value as per AR documentation and is received via ccw_device_get_mdc() for a specific path mask from the CIO layer. The function itself also always returns a positive mdc value or 0 in case mdc isn't supported or couldn't be determined.

Though, the comment for this function describes a negative return value to indicate failures.

As a result, the DASD device driver interprets the return value of ccw_device_get_mdc() incorrectly. The error case is essentially a dead code path.

To fix this behaviour, check explicitly for a return value of 0 and change the comment for ccw_device_get_mdc() accordingly.

This fix merely enables the error code path in the DASD functions get_fcx_max_data() and verify_fcx_max_data(). The actual functionality stays the same and is still correct.

---
 drivers/s390/block/dasd_eckd.c | 9 +++++----
 drivers/s390/cio/device_ops.c  | 2 +-
 2 files changed, 6 insertions(+), 5 deletions(-)
drivers/s390/block/dasd_eckd.c +5 -4
index c94184d080f8..f5622f4a2ecf 100644
@@ -1128,7 +1128,8 @@ static u32 get_fcx_max_data(struct dasd_device *device)
 {
 	struct dasd_eckd_private *private = device->private;
 	int fcx_in_css, fcx_in_gneq, fcx_in_features;
-	int tpm, mdc;
+	unsigned int mdc;
+	int tpm;
 
 	if (dasd_nofcx)
 		return 0;
@@ -1142,7 +1143,7 @@ static u32 get_fcx_max_data(struct dasd_device *device)
 		return 0;
 
 	mdc = ccw_device_get_mdc(device->cdev, 0);
-	if (mdc < 0) {
+	if (mdc == 0) {
 		dev_warn(&device->cdev->dev, "Detecting the maximum supported data size for zHPF requests failed\n");
 		return 0;
 	} else {
@@ -1153,12 +1154,12 @@ static u32 get_fcx_max_data(struct dasd_device *device)
 static int verify_fcx_max_data(struct dasd_device *device, __u8 lpm)
 {
 	struct dasd_eckd_private *private = device->private;
-	int mdc;
+	unsigned int mdc;
 	u32 fcx_max_data;
 
 	if (private->fcx_max_data) {
 		mdc = ccw_device_get_mdc(device->cdev, lpm);
-		if ((mdc < 0)) {
+		if (mdc == 0) {
 			dev_warn(&device->cdev->dev,
 				 "Detecting the maximum data size for zHPF "
 				 "requests failed (rc=%d) for a new path %x\n",
drivers/s390/cio/device_ops.c +1 -1
index 65841af15748..ccecf6b9504e 100644
@@ -635,7 +635,7 @@ EXPORT_SYMBOL(ccw_device_tm_start_timeout);
  * @mask: mask of paths to use
  *
  * Return the number of 64K-bytes blocks all paths at least support
- * for a transport command. Return values <= 0 indicate failures.
+ * for a transport command. Return value 0 indicates failure.
  */
 int ccw_device_get_mdc(struct ccw_device *cdev, u8 mask)
 {
Signature · 1 line
2.17.1
Reviewed-by on v1 1/3
#2 lore raw

Re: [PATCH 1/3] s390/dasd/cio: Interpret ccw_device_get_mdc return value correctly

Quoted 10 lines · On Thu, 19 Dec 2019 09:43:50 +0100 Stefan Haberland <[email protected]> wrote:

From: Jan Höppner <[email protected]>

The max data count (mdc) is an unsigned 16-bit integer value as per AR documentation and is received via ccw_device_get_mdc() for a specific path mask from the CIO layer. The function itself also always returns a positive mdc value or 0 in case mdc isn't supported or couldn't be determined.

Though, the comment for this function describes a negative return value to indicate failures.

Note that this used to be true before 040495d110ba ("s390/cio: make use of newly added format 1 channel-path data").

Quoted 4 lines

As a result, the DASD device driver interprets the return value of ccw_device_get_mdc() incorrectly. The error case is essentially a dead code path.

To be pedantic: It did not check for <= 0 (as documented) either :)

Quoted 60 lines

To fix this behaviour, check explicitly for a return value of 0 and change the comment for ccw_device_get_mdc() accordingly.

This fix merely enables the error code path in the DASD functions get_fcx_max_data() and verify_fcx_max_data(). The actual functionality stays the same and is still correct.

---
 drivers/s390/block/dasd_eckd.c | 9 +++++----
 drivers/s390/cio/device_ops.c  | 2 +-
 2 files changed, 6 insertions(+), 5 deletions(-)
drivers/s390/block/dasd_eckd.c +5 -4
index c94184d080f8..f5622f4a2ecf 100644
@@ -1128,7 +1128,8 @@ static u32 get_fcx_max_data(struct dasd_device *device)
 {
 	struct dasd_eckd_private *private = device->private;
 	int fcx_in_css, fcx_in_gneq, fcx_in_features;
-	int tpm, mdc;
+	unsigned int mdc;
+	int tpm;
 
 	if (dasd_nofcx)
 		return 0;
@@ -1142,7 +1143,7 @@ static u32 get_fcx_max_data(struct dasd_device *device)
 		return 0;
 
 	mdc = ccw_device_get_mdc(device->cdev, 0);
-	if (mdc < 0) {
+	if (mdc == 0) {
 		dev_warn(&device->cdev->dev, "Detecting the maximum supported data size for zHPF requests failed\n");
 		return 0;
 	} else {
@@ -1153,12 +1154,12 @@ static u32 get_fcx_max_data(struct dasd_device *device)
 static int verify_fcx_max_data(struct dasd_device *device, __u8 lpm)
 {
 	struct dasd_eckd_private *private = device->private;
-	int mdc;
+	unsigned int mdc;
 	u32 fcx_max_data;
 
 	if (private->fcx_max_data) {
 		mdc = ccw_device_get_mdc(device->cdev, lpm);
-		if ((mdc < 0)) {
+		if (mdc == 0) {
 			dev_warn(&device->cdev->dev,
 				 "Detecting the maximum data size for zHPF "
 				 "requests failed (rc=%d) for a new path %x\n",
drivers/s390/cio/device_ops.c +0 -0
index 65841af15748..ccecf6b9504e 100644

drivers/s390/cio/device_ops.c EXPORT_SYMBOL(ccw_device_tm_start_timeout); · line 638

Quoted 5 lines
drivers/s390/cio/device_ops.c +1 -1
@@ -635,7 +635,7 @@ EXPORT_SYMBOL(ccw_device_tm_start_timeout);
  * @mask: mask of paths to use
  *
  * Return the number of 64K-bytes blocks all paths at least support
- * for a transport command. Return values <= 0 indicate failures.
+ * for a transport command. Return value 0 indicates failure.

s/Return value/The return value/ ?

drivers/s390/cio/device_ops.c EXPORT_SYMBOL(ccw_device_tm_start_timeout); · line 641

Quoted 3 lines
drivers/s390/cio/device_ops.c +0 -0
  */
 int ccw_device_get_mdc(struct ccw_device *cdev, u8 mask)
 {
author maintainer
#3 lore raw

PATCH 2/3 s390/dasd: fix memleak in path handling error case

In reply to Stefan Haberland

If for whatever reason the dasd_eckd_check_characteristics() function exits after at least some paths have their configuration data allocated those data is never freed again. In the error case the device->private pointer is set to NULL and dasd_eckd_uncheck_device() will exit without freeing the path data because of this NULL pointer.

Fix by calling dasd_eckd_clear_conf_data() for error cases.

Also use dasd_eckd_clear_conf_data() in dasd_eckd_uncheck_device() to avoid code duplication.

---
 drivers/s390/block/dasd_eckd.c | 19 ++-----------------
 1 file changed, 2 insertions(+), 17 deletions(-)
drivers/s390/block/dasd_eckd.c +2 -17
index f5622f4a2ecf..a28b9ff82378 100644
@@ -2074,7 +2074,7 @@ dasd_eckd_check_characteristics(struct dasd_device *device)
 	dasd_free_block(device->block);
 	device->block = NULL;
 out_err1:
-	kfree(private->conf_data);
+	dasd_eckd_clear_conf_data(device);
 	kfree(device->private);
 	device->private = NULL;
 	return rc;
@@ -2083,7 +2083,6 @@ dasd_eckd_check_characteristics(struct dasd_device *device)
 static void dasd_eckd_uncheck_device(struct dasd_device *device)
 {
 	struct dasd_eckd_private *private = device->private;
-	int i;
 
 	if (!private)
 		return;
@@ -2093,21 +2092,7 @@ static void dasd_eckd_uncheck_device(struct dasd_device *device)
 	private->sneq = NULL;
 	private->vdsneq = NULL;
 	private->gneq = NULL;
-	private->conf_len = 0;
-	for (i = 0; i < 8; i++) {
-		kfree(device->path[i].conf_data);
-		if ((__u8 *)device->path[i].conf_data ==
-		    private->conf_data) {
-			private->conf_data = NULL;
-			private->conf_len = 0;
-		}
-		device->path[i].conf_data = NULL;
-		device->path[i].cssid = 0;
-		device->path[i].ssid = 0;
-		device->path[i].chpid = 0;
-	}
-	kfree(private->conf_data);
-	private->conf_data = NULL;
+	dasd_eckd_clear_conf_data(device);
 }
 
 static struct dasd_ccw_req *
Signature · 1 line
2.17.1
author maintainer
#4 lore raw

PATCH 3/3 s390/dasd: fix typo in copyright statement

In reply to Stefan Haberland

coypright -> copyright

---
 drivers/s390/block/dasd_fba.h  | 2 +-
 drivers/s390/block/dasd_proc.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
drivers/s390/block/dasd_fba.h +1 -1
index 8f75df06e893..45ddabec4017 100644
@@ -2,7 +2,7 @@
 /*
  * Author(s)......: Holger Smolinski <[email protected]>
  * Bugreports.to..: <[email protected]>
- * Coypright IBM Corp. 1999, 2000
+ * Copyright IBM Corp. 1999, 2000
  *
  */
 
drivers/s390/block/dasd_proc.c +1 -1
index 1770b99f607e..8d4d69ea5baf 100644
@@ -5,7 +5,7 @@
  *		    Carsten Otte <[email protected]>
  *		    Martin Schwidefsky <[email protected]>
  * Bugreports.to..: <[email protected]>
- * Coypright IBM Corp. 1999, 2002
+ * Copyright IBM Corp. 1999, 2002
  *
  * /proc interface for the dasd driver.
  *
Signature · 1 line
2.17.1
#5 lore raw

Re: [PATCH 3/3] s390/dasd: fix typo in copyright statement

Quoted 14 lines · On Thu, 19 Dec 2019 09:43:52 +0100 Stefan Haberland <[email protected]> wrote:

coypright -> copyright

---
 drivers/s390/block/dasd_fba.h  | 2 +-
 drivers/s390/block/dasd_proc.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
drivers/s390/block/dasd_fba.h +0 -0
index 8f75df06e893..45ddabec4017 100644

drivers/s390/block/dasd_fba.h line 4

Quoted 3 lines
drivers/s390/block/dasd_fba.h +0 -0
@@ -2,7 +2,7 @@
 /*
  * Author(s)......: Holger Smolinski <[email protected]>
  * Bugreports.to..: <[email protected]>

No comment on the actual patch; but do you really still want people to send bug reports to that address?

(A quick grep shows that it is mostly present in the dasd code, but also in some other code parts.)

Quoted 18 lines
drivers/s390/block/dasd_fba.h +1 -1
- * Coypright IBM Corp. 1999, 2000
+ * Copyright IBM Corp. 1999, 2000
  *
  */
 
drivers/s390/block/dasd_proc.c +1 -1
index 1770b99f607e..8d4d69ea5baf 100644
@@ -5,7 +5,7 @@
  *		    Carsten Otte <[email protected]>
  *		    Martin Schwidefsky <[email protected]>
  * Bugreports.to..: <[email protected]>
- * Coypright IBM Corp. 1999, 2002
+ * Copyright IBM Corp. 1999, 2002
  *
  * /proc interface for the dasd driver.
  *
maintainer
#6 lore raw
In reply to Stefan Haberland
Quoted 7 lines · Stefan Haberland

Hi Jens,

please see the following three patches that

- fix a dead code path for DASD
- fix a memleak in DASD error case
- fix typo in copyright statement

Applied for 5.5, thanks

Signature · 1 line
Jens Axboe