kernel/linux.git/drivers/s390, branch linux-4.4.y

scsi: zfcp: Fix failed recovery on gone remote port with non-NPIV FCP devices

2022-02-03T08:27:52+00:00

commit 8c9db6679be4348b8aae108e11d4be2f83976e30 upstream. Suppose we have an environment with a number of non-NPIV FCP devices (virtual HBAs / FCP devices / zfcp "adapter"s) sharing the same physical FCP channel (HBA port) and its I_T nexus. Plus a number of storage target ports zoned to such shared channel. Now one target port logs out of the fabric causing an RSCN. Zfcp reacts with an ADISC ELS and subsequent port recovery depending on the ADISC result. This happens on all such FCP devices (in different Linux images) concurrently as they all receive a copy of this RSCN. In the following we look at one of those FCP devices. Requests other than FSF_QTCB_FCP_CMND can be slow until they get a response. Depending on which requests are affected by slow responses, there are different recovery outcomes. Here we want to fix failed recoveries on port or adapter level by avoiding recovery requests that can be slow. We need the cached N_Port_ID for the remote port "link" test with ADISC. Just before sending the ADISC, we now intentionally forget the old cached N_Port_ID. The idea is that on receiving an RSCN for a port, we have to assume that any cached information about this port is stale. This forces a fresh new GID_PN [FC-GS] nameserver lookup on any subsequent recovery for the same port. Since we typically can still communicate with the nameserver efficiently, we now reach steady state quicker: Either the nameserver still does not know about the port so we stop recovery, or the nameserver already knows the port potentially with a new N_Port_ID and we can successfully and quickly perform open port recovery. For the one case, where ADISC returns successfully, we re-initialize port->d_id because that case does not involve any port recovery. This also solves a problem if the storage WWPN quickly logs into the fabric again but with a different N_Port_ID. Such as on virtual WWPN takeover during target NPIV failover. [https://www.redbooks.ibm.com/abstracts/redp5477.html] In that case the RSCN from the storage FDISC was ignored by zfcp and we could not successfully recover the failover. On some later failback on the storage, we could have been lucky if the virtual WWPN got the same old N_Port_ID from the SAN switch as we still had cached. Then the related RSCN triggered a successful port reopen recovery. However, there is no guarantee to get the same N_Port_ID on NPIV FDISC. Even though NPIV-enabled FCP devices are not affected by this problem, this code change optimizes recovery time for gone remote ports as a side effect. The timely drop of cached N_Port_IDs prevents unnecessary slow open port attempts. While the problem might have been in code before v2.6.32 commit 799b76d09aee ("[SCSI] zfcp: Decouple gid_pn requests from erp") this fix depends on the gid_pn_work introduced with that commit, so we mark it as culprit to satisfy fix dependencies. Note: Point-to-point remote port is already handled separately and gets its N_Port_ID from the cached peer_d_id. So resetting port->d_id in general does not affect PtP. Link: https://lore.kernel.org/r/20220118165803.3667947-1-maier@linux.ibm.com Fixes: 799b76d09aee ("[SCSI] zfcp: Decouple gid_pn requests from erp") Cc: #2.6.32+ Suggested-by: Benjamin Block Reviewed-by: Benjamin Block Signed-off-by: Steffen Maier Signed-off-by: Martin K. Petersen Signed-off-by: Greg Kroah-Hartman

s390/cio: dont call css_wait_for_slow_path() inside a lock

2021-07-20T14:22:31+00:00

commit c749d8c018daf5fba6dfac7b6c5c78b27efd7d65 upstream. Currently css_wait_for_slow_path() gets called inside the chp->lock. The path-verification-loop of slowpath inside this lock could lead to deadlock as reported by the lockdep validator. The ccw_device_get_chp_desc() during the instance of a device-set-online would try to acquire the same 'chp->lock' to read the chp->desc. The instance of this function can get called from multiple scenario, like probing or setting-device online manually. This could, in some corner-cases lead to the deadlock. lockdep validator reported this as, CPU0 CPU1 ---- ---- lock(&chp->lock); lock(kn->active#43); lock(&chp->lock); lock((wq_completion)cio); The chp->lock was introduced to serialize the access of struct channel_path. This lock is not needed for the css_wait_for_slow_path() function, so invoke the slow-path function outside this lock. Fixes: b730f3a93395 ("[S390] cio: add lock to struct channel_path") Cc: Reviewed-by: Peter Oberparleiter Signed-off-by: Vineeth Vijayan Signed-off-by: Vasily Gorbik Signed-off-by: Greg Kroah-Hartman

s390/dasd: fix hanging DASD driver unbind

2021-03-17T15:07:20+00:00

commit 7d365bd0bff3c0310c39ebaffc9a8458e036d666 upstream. In case of an unbind of the DASD device driver the function dasd_generic_remove() is called which shuts down the device. Among others this functions removes the int_handler from the cdev. During shutdown the device cancels all outstanding IO requests and waits for completion of the clear request. Unfortunately the clear interrupt will never be received when there is no interrupt handler connected. Fix by moving the int_handler removal after the call to the state machine where no request or interrupt is outstanding. Cc: stable@vger.kernel.org Signed-off-by: Stefan Haberland Tested-by: Bjoern Walk Reviewed-by: Jan Hoeppner Signed-off-by: Jens Axboe Signed-off-by: Greg Kroah-Hartman

s390/dasd: fix hanging device offline processing

2021-01-09T12:34:14+00:00

[ Upstream commit 658a337a606f48b7ebe451591f7681d383fa115e ] For an LCU update a read unit address configuration IO is required. This is started using sleep_on(), which has early exit paths in case the device is not usable for IO. For example when it is in offline processing. In those cases the LCU update should fail and not be retried. Therefore lcu_update_work checks if EOPNOTSUPP is returned or not. Commit 41995342b40c ("s390/dasd: fix endless loop after read unit address configuration") accidentally removed the EOPNOTSUPP return code from read_unit_address_configuration(), which in turn might lead to an endless loop of the LCU update in offline processing. Fix by returning EOPNOTSUPP again if the device is not able to perform the request. Fixes: 41995342b40c ("s390/dasd: fix endless loop after read unit address configuration") Cc: stable@vger.kernel.org #5.3 Signed-off-by: Stefan Haberland Reviewed-by: Jan Hoeppner Signed-off-by: Jens Axboe Signed-off-by: Sasha Levin

s390/dasd: fix list corruption of lcu list

2020-12-29T12:42:42+00:00

commit 53a7f655834c7c335bf683f248208d4fbe4b47bc upstream. In dasd_alias_disconnect_device_from_lcu the device is removed from any list on the LCU. Afterwards the LCU is removed from the lcu list if it does not contain devices any longer. The lcu->lock protects the lcu from parallel updates. But to cancel all workers and wait for completion the lcu->lock has to be unlocked. If two devices are removed in parallel and both are removed from the LCU the first device that takes the lcu->lock again will delete the LCU because it is already empty but the second device also tries to free the LCU which leads to a list corruption of the lcu list. Fix by removing the device right before the lcu is checked without unlocking the lcu->lock in between. Fixes: 8e09f21574ea ("[S390] dasd: add hyper PAV support to DASD device driver, part 1") Cc: stable@vger.kernel.org Signed-off-by: Stefan Haberland Reviewed-by: Jan Hoeppner Signed-off-by: Jens Axboe Signed-off-by: Greg Kroah-Hartman

s390/dasd: fix list corruption of pavgroup group list

2020-12-29T12:42:42+00:00

commit 0ede91f83aa335da1c3ec68eb0f9e228f269f6d8 upstream. dasd_alias_add_device() moves devices to the active_devices list in case of a scheduled LCU update regardless if they have previously been in a pavgroup or not. Example: device A and B are in the same pavgroup. Device A has already been in a pavgroup and the private->pavgroup pointer is set and points to a valid pavgroup. While going through dasd_add_device it is moved from the pavgroup to the active_devices list. In parallel device B might be removed from the same pavgroup in remove_device_from_lcu() which in turn checks if the group is empty and deletes it accordingly because device A has already been removed from there. When now device A enters remove_device_from_lcu() it is tried to remove it from the pavgroup again because the pavgroup pointer is still set and again the empty group will be cleaned up which leads to a list corruption. Fix by setting private->pavgroup to NULL in dasd_add_device. If the device has been the last device on the pavgroup an empty pavgroup remains but this will be cleaned up by the scheduled lcu_update which iterates over all existing pavgroups. Fixes: 8e09f21574ea ("[S390] dasd: add hyper PAV support to DASD device driver, part 1") Cc: stable@vger.kernel.org Signed-off-by: Stefan Haberland Reviewed-by: Jan Hoeppner Signed-off-by: Jens Axboe Signed-off-by: Greg Kroah-Hartman

s390/cio: add cond_resched() in the slow_eval_known_fn() loop

2020-09-03T09:19:25+00:00

[ Upstream commit 0b8eb2ee9da1e8c9b8082f404f3948aa82a057b2 ] The scanning through subchannels during the time of an event could take significant amount of time in case of platforms with lots of known subchannels. This might result in higher scheduling latencies for other tasks especially on systems with a single CPU. Add cond_resched() call, as the loop in slow_eval_known_fn() can be executed for a longer duration. Reviewed-by: Peter Oberparleiter Signed-off-by: Vineeth Vijayan Signed-off-by: Heiko Carstens Signed-off-by: Sasha Levin

s390/qeth: don't process empty bridge port events

2020-08-21T08:53:02+00:00

[ Upstream commit 02472e28b9a45471c6d8729ff2c7422baa9be46a ] Discard events that don't contain any entries. This shouldn't happen, but subsequent code relies on being able to use entry 0. So better be safe than accessing garbage. Fixes: b4d72c08b358 ("qeth: bridgeport support - basic control") Signed-off-by: Julian Wiedmann Reviewed-by: Alexandra Winter Signed-off-by: David S. Miller Signed-off-by: Sasha Levin

s390/qdio: put thinint indicator after early error

2020-06-30T00:07:50+00:00

[ Upstream commit 75e82bec6b2622c6f455b7a543fb5476a5d0eed7 ] qdio_establish() calls qdio_setup_thinint() via qdio_setup_irq(). If the subsequent qdio_establish_thinint() fails, we miss to put the DSCI again. Thus the DSCI isn't available for re-use. Given enough of such errors, we could end up with having only the shared DSCI available. Merge qdio_setup_thinint() into qdio_establish_thinint(), and deal with such an error internally. Fixes: 779e6e1c724d ("[S390] qdio: new qdio driver.") Signed-off-by: Julian Wiedmann Reviewed-by: Benjamin Block Signed-off-by: Vasily Gorbik Signed-off-by: Sasha Levin

scsi: zfcp: fix request object use-after-free in send path causing wrong traces

2020-06-03T06:12:16+00:00

[ Upstream commit 106d45f350c7cac876844dc685845cba4ffdb70b ] When tracing instances where we open and close WKA ports, we also pass the request-ID of the respective FSF command. But after successfully sending the FSF command we must not use the request-object anymore, as this might result in an use-after-free (see "zfcp: fix request object use-after-free in send path causing seqno errors" ). To fix this add a new variable that caches the request-ID before sending the request. This won't change during the hand-off to the FCP channel, and so it's safe to trace this cached request-ID later, instead of using the request object. Signed-off-by: Benjamin Block Fixes: d27a7cb91960 ("zfcp: trace on request for open and close of WKA port") Cc: #2.6.38+ Reviewed-by: Steffen Maier Reviewed-by: Jens Remus Signed-off-by: Martin K. Petersen Signed-off-by: Sasha Levin