On Wed, 2015-03-04 at 16:00 +0100, Olaf Hering wrote: > On Fri, Feb 27, Nicholas A. Bellinger wrote: > > > On Thu, 2015-02-26 at 10:26 +0100, Olaf Hering wrote: > > > > [ 145.661468] IP: [<ffffffffa024f17b>] pscsi_get_device_type+0xb/0x20 > > > [target_core_pscsi] > > Looks like a PSCSI specific NULL pointer dereference when reading the > > dev_type attribute if a backend device was not able to be successfully > > configured -> enabled. > > > > Here's a quick patch to address this case. Please confirm on your end. > > Unfortunately I can not reproduce it anymore. Maybe it depends on the .config? >
So the way to reproduce is to read the scsi_lu/dev_type configfs attribute when a PSCSI backend device has not yet been configured -> enabled. Eg: root@scsi-mq:/usr/src/target-pending.git# modprobe target_core_mod root@scsi-mq:/usr/src/target-pending.git# mkdir -p /sys/kernel/config/target/core/pscsi_0/foo/ root@scsi-mq:/usr/src/target-pending.git# cat /sys/kernel/config/target/core/pscsi_0/foo/statistics/scsi_lu/dev_type Killed I've confirmed the previous patch addresses this case, which now returns TYPE_NO_LUN. (0x7f) root@scsi-mq:/usr/src/target-pending.git# modprobe target_core_mod root@scsi-mq:/usr/src/target-pending.git# mkdir -p /sys/kernel/config/target/core/pscsi_0/foo root@scsi-mq:/usr/src/target-pending.git# cat /sys/kernel/config/target/core/pscsi_0/foo/statistics/scsi_lu/dev_type 127 Queuing up this patch to target-pending/master now, and including a CC' to stable as well. Thank you, --nab -- To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
