Current code assumes abort will remove the original command from the active
list where scsi_done will not be called. Instead, the eh_abort thread will
do the scsi_done. That is not the case. Instead, we have a double
scsi_done calls triggering use after free.
Abort will tell FW to release the command from FW possesion. The original
command will return to ULP with error in its normal fashion via scsi_done.
eh_abort path would wait for the original command completion before
returning. eh_abort path will not perform the scsi_done call.
Fixes: 219d27d7147e0 ("scsi: qla2xxx: Fix race conditions in the code for aborting SCSI commands")
Cc: stable@vger.kernel.org # 5.2
Link: https://lore.kernel.org/r/20191105150657.8092-6-hmadhani@marvell.com
Reviewed-by: Ewan D. Milne <emilne@redhat.com>
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Arun Easi <aeasi@marvell.com>
Signed-off-by: Himanshu Madhani <hmadhani@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
[ commit f45bca8c5052e8c59bab64ee90c44441678b9a52 upstream ]
git-svn-id: http://svn.code.sf.net/p/scst/svn/trunk@8715 d57e44dd-8a1f-0410-8b47-8ef2f437770f
Current code will send PRLI with FC-NVMe bit set for the targets which
support only FCP. This may result into issue with targets which do not
understand NVMe and will go into a strange state. This patch would restart
the login process by going back to PLOGI state. The PLOGI state will force
the target to respond to correct PRLI request.
Fixes: c76ae845ea836 ("scsi: qla2xxx: Add error handling for PLOGI ELS passthrough")
Cc: stable@vger.kernel.org # 5.4
Link: https://lore.kernel.org/r/20191105150657.8092-2-hmadhani@marvell.com
Reviewed-by: Ewan D. Milne <emilne@redhat.com>
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Himanshu Madhani <hmadhani@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
[ commit 983f127603fac650fa34ee69db363e4615eaf9e7 upstream ]
git-svn-id: http://svn.code.sf.net/p/scst/svn/trunk@8711 d57e44dd-8a1f-0410-8b47-8ef2f437770f
This patch prevents that the following warning is reported when running the
libiscsi tests against the vdisk_blockio handler:
Refused bio with invalid length 4080 and/or offset 16.
WARNING: CPU: 3 PID: 16022 at /home/bvanassche/software/scst.git/scst/src/dev_handlers/scst_vdisk.c:6180 blockio_exec_rw+0xc89/0xcf0 [scst_vdisk]
Call Trace:
blockio_exec_write+0x9c/0xe0 [scst_vdisk]
vdev_do_job+0xe8/0x220 [scst_vdisk]
blockio_exec+0x140/0x370 [scst_vdisk]
scst_do_real_exec+0xf4/0x610 [scst]
scst_exec_check_blocking+0x24e/0x6b0 [scst]
scst_exec_check_sn+0x222/0x6c0 [scst]
scst_process_active_cmd+0xdff/0x31c0 [scst]
scst_cmd_thread+0x36f/0xb60 [scst]
kthread+0x1bc/0x210
ret_from_fork+0x24/0x30
See also https://github.com/sahlberg/libiscsi/issues/302.
git-svn-id: http://svn.code.sf.net/p/scst/svn/trunk@8700 d57e44dd-8a1f-0410-8b47-8ef2f437770f
Since this function is only used by the vdisk_fileio handler, reflect this
in the function name.
git-svn-id: http://svn.code.sf.net/p/scst/svn/trunk@8699 d57e44dd-8a1f-0410-8b47-8ef2f437770f
The exec_req_fifo patches have been removed some time ago. Hence also remove
the code that depends on these patches having been applied. See also commit
78d6da4517 ("scst: Remove obsolete exec_req_fifo kernel patches"; r8022).
git-svn-id: http://svn.code.sf.net/p/scst/svn/trunk@8698 d57e44dd-8a1f-0410-8b47-8ef2f437770f
pr_warning() has been removed from kernel v5.5. Hence use pr_warn() instead
of pr_warning().
git-svn-id: http://svn.code.sf.net/p/scst/svn/trunk@8697 d57e44dd-8a1f-0410-8b47-8ef2f437770f
Instead of trying to fix the regressions introduced upstream, stay close
to the upstream code and let the upstream maintainers fix regressions.
git-svn-id: http://svn.code.sf.net/p/scst/svn/trunk@8688 d57e44dd-8a1f-0410-8b47-8ef2f437770f
Copying Module.symvers is the oldest supported method for building external
kernel modules that depend on another external kernel module. Since support
for that method will be removed from Linux kernel v5.5, use
KBUILD_EXTRA_SYMBOLS instead to specify the external symbols SCST kernel
modules depend on. Support for KBUILD_EXTRA_SYMBOLS was introduced in
upstream kernel v2.6.26 so this patch drops support for kernels before
v2.6.26.
git-svn-id: http://svn.code.sf.net/p/scst/svn/trunk@8686 d57e44dd-8a1f-0410-8b47-8ef2f437770f
This patch adds support for kernel module namespaces. This patch is essential
when using SCST in combination with kernel version v5.4 or later.
git-svn-id: http://svn.code.sf.net/p/scst/svn/trunk@8673 d57e44dd-8a1f-0410-8b47-8ef2f437770f
Instead of only checking the mlnx-ofed-kernel-dkms, also check the
mlnx-ofed-kernel-modules package name.
Signed-off-by: Ville Skytta <ville.skytta@iki.fi>
[ bvanassche: Combined two patches into this single change ]
git-svn-id: http://svn.code.sf.net/p/scst/svn/trunk@8668 d57e44dd-8a1f-0410-8b47-8ef2f437770f
Instead of only checking the mlnx-ofed-kernel-dkms package name, also check
the mlnx-ofed-kernel-modules package name.
Signed-off-by: Ville Skytta <ville.skytta@iki.fi>
[ bvanassche: extracted this change from a larger patch ]
git-svn-id: http://svn.code.sf.net/p/scst/svn/trunk@8667 d57e44dd-8a1f-0410-8b47-8ef2f437770f
This patch fixes the following Coverity complaint:
7. assignment: Assigning: usn_len = max_len. The value of usn_len is now 252.
CID 326814 (#1 of 1): Out-of-bounds access (OVERRUN)
8. overrun-buffer-arg: Overrunning array virt_dev->usn of 21 bytes by
passing it to a function which accesses it at byte offset 251 using
argument usn_len (which evaluates to 252).
Fixes: ca180efeec ("scst_vdisk: Avoid that vdisk_usn_vpd() can trigger a buffer overflow")
git-svn-id: http://svn.code.sf.net/p/scst/svn/trunk@8665 d57e44dd-8a1f-0410-8b47-8ef2f437770f
