Skip to contentSkip to conversation

SCSI sd: udev accessing an uninitialized scsi_disk results in a crash

PATCH In review Nagendra Singh Tomar 4 participants 4 messages Feb 2, 2007 – Feb 3, 2007 SCSI

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 SCSI sd: udev accessing an uninitialized scsi_disk results in a crash

Hi,
sd_probe() calls class_device_add() even before initializing the
sdkp->device variable. class_device_add() eventually results in the user mode udev program to be called. udev program can read the the allow_restart attribute of the newly created scsi device. This is resulting in a crash as the show function for allow_restart (i.e sd_show_allow_restart) returns the attribute value by reading the sdkp->device->allow_restart variable. As the sdkp->device is not initialized before calling the user mode hotplug helper, this results in a crash.
The patch below solves it by calling class_device_add() only after the
necessary fields in the scsi_disk structure are initialized properly.

Thanx,
Tomar

---

drivers/scsi/sd.c renamed from drivers/scsi/sd.c.orig +10 -10
@@ -1646,16 +1646,6 @@ static int sd_probe(struct device *dev)
 	if (error)
 		goto out_put;
 
-	class_device_initialize(&sdkp->cdev);
-	sdkp->cdev.dev = &sdp->sdev_gendev;
-	sdkp->cdev.class = &sd_disk_class;
-	strncpy(sdkp->cdev.class_id, sdp->sdev_gendev.bus_id, BUS_ID_SIZE);
-
-	if (class_device_add(&sdkp->cdev))
-		goto out_put;
-
-	get_device(&sdp->sdev_gendev);
-
 	sdkp->device = sdp;
 	sdkp->driver = &sd_template;
 	sdkp->disk = gd;
@@ -1669,6 +1659,16 @@ static int sd_probe(struct device *dev)
 			sdp->timeout = SD_MOD_TIMEOUT;
 	}
 
+	class_device_initialize(&sdkp->cdev);
+	sdkp->cdev.dev = &sdp->sdev_gendev;
+	sdkp->cdev.class = &sd_disk_class;
+	strncpy(sdkp->cdev.class_id, sdp->sdev_gendev.bus_id, BUS_ID_SIZE);
+
+	if (class_device_add(&sdkp->cdev))
+		goto out_put;
+
+	get_device(&sdp->sdev_gendev);
+
 	gd->major = sd_major((index & 0xf0) >> 4);
 	gd->first_minor = ((index & 0xf) << 4) | (index & 0xfff00);
 	gd->minors = 16;
#1 lore raw
Quoted 40 lines · On Fri, 2 Feb 2007 17:34:56 +0530 Nagendra Singh Tomar…

Hi,
sd_probe() calls class_device_add() even before initializing the
sdkp->device variable. class_device_add() eventually results in the user mode udev program to be called. udev program can read the the allow_restart attribute of the newly created scsi device. This is resulting in a crash as the show function for allow_restart (i.e sd_show_allow_restart) returns the attribute value by reading the sdkp->device->allow_restart variable. As the sdkp->device is not initialized before calling the user mode hotplug helper, this results in a crash.
The patch below solves it by calling class_device_add() only after the
necessary fields in the scsi_disk structure are initialized properly.

drivers/scsi/sd.c renamed from drivers/scsi/sd.c.orig +3 -10
@@ -1646,16 +1646,6 @@ static int sd_probe(struct device *dev)
 	if (error)
 		goto out_put;
 
-	class_device_initialize(&sdkp->cdev);
-	sdkp->cdev.dev = &sdp->sdev_gendev;
-	sdkp->cdev.class = &sd_disk_class;
-	strncpy(sdkp->cdev.class_id, sdp->sdev_gendev.bus_id, BUS_ID_SIZE);
-
-	if (class_device_add(&sdkp->cdev))
-		goto out_put;
-
-	get_device(&sdp->sdev_gendev);
-
 	sdkp->device = sdp;
 	sdkp->driver = &sd_template;
 	sdkp->disk = gd;
@@ -1669,6 +1659,16 @@ static int sd_probe(struct device *dev)
 			sdp->timeout = SD_MOD_TIMEOUT;
 	}
 
+	class_device_initialize(&sdkp->cdev);
+	sdkp->cdev.dev = &sdp->sdev_gendev;
+	sdkp->cdev.class = &sd_disk_class;

drivers/scsi/sd.c static int sd_probe(struct device *dev) · line 1674

Quoted 10 lines
drivers/scsi/sd.c renamed from drivers/scsi/sd.c.orig +7 -0
+	strncpy(sdkp->cdev.class_id, sdp->sdev_gendev.bus_id, BUS_ID_SIZE);
+
+	if (class_device_add(&sdkp->cdev))
+		goto out_put;
+
+	get_device(&sdp->sdev_gendev);
+
 	gd->major = sd_major((index & 0xf0) >> 4);
 	gd->first_minor = ((index & 0xf) << 4) | (index & 0xfff00);
 	gd->minors = 16;

Thanks - I'll queue this up for 2.6.20 also.

#2 lore raw
Quoted 40 lines · Andrew Morton
Quoted 37 lines · On Fri, 2 Feb 2007 17:34:56 +0530 Nagendra Singh Tomar…

Hi,
sd_probe() calls class_device_add() even before initializing the
sdkp->device variable. class_device_add() eventually results in the user mode udev program to be called. udev program can read the the allow_restart attribute of the newly created scsi device. This is resulting in a crash as the show function for allow_restart (i.e sd_show_allow_restart) returns the attribute value by reading the sdkp->device->allow_restart variable. As the sdkp->device is not initialized before calling the user mode hotplug helper, this results in a crash.
The patch below solves it by calling class_device_add() only after the
necessary fields in the scsi_disk structure are initialized properly.

drivers/scsi/sd.c renamed from drivers/scsi/sd.c.orig +0 -10
@@ -1646,16 +1646,6 @@ static int sd_probe(struct device *dev)
 	if (error)
 		goto out_put;
 
-	class_device_initialize(&sdkp->cdev);
-	sdkp->cdev.dev = &sdp->sdev_gendev;
-	sdkp->cdev.class = &sd_disk_class;
-	strncpy(sdkp->cdev.class_id, sdp->sdev_gendev.bus_id, BUS_ID_SIZE);
-
-	if (class_device_add(&sdkp->cdev))
-		goto out_put;
-
-	get_device(&sdp->sdev_gendev);
-
 	sdkp->device = sdp;
 	sdkp->driver = &sd_template;
 	sdkp->disk = gd;
@@ -1669,6 +1659,16 @@ static int sd_probe(struct device *dev)
 			sdp->timeout = SD_MOD_TIMEOUT;
 	}
 
Quoted 15 lines

drivers/scsi/sd.c static int sd_probe(struct device *dev) · line 1674

Quoted 13 lines
drivers/scsi/sd.c renamed from drivers/scsi/sd.c.orig +10 -0
+	class_device_initialize(&sdkp->cdev);
+	sdkp->cdev.dev = &sdp->sdev_gendev;
+	sdkp->cdev.class = &sd_disk_class;
+	strncpy(sdkp->cdev.class_id, sdp->sdev_gendev.bus_id, BUS_ID_SIZE);
+
+	if (class_device_add(&sdkp->cdev))
+		goto out_put;
+
+	get_device(&sdp->sdev_gendev);
+
 	gd->major = sd_major((index & 0xf0) >> 4);
 	gd->first_minor = ((index & 0xf) << 4) | (index & 0xfff00);
 	gd->minors = 16;

Thanks - I'll queue this up for 2.6.20 also.

No objection from me, as long as James says this is ok.

I wonder why we haven't noticed this in the past?

thanks,

greg k-h

#3 lore raw
Quoted 5 lines · Greg KH
Quoted 1 line

Thanks - I'll queue this up for 2.6.20 also.

No objection from me, as long as James says this is ok.

I wonder why we haven't noticed this in the past?

Because the race is so small ...

I'll queue it in the rc-fixes tree .. I have three others for 2.6.20

James