From: Hans-Joachim Picht <hpicht@redhat.com> Date: Tue, 22 Sep 2009 22:13:37 +0200 Subject: [s390] dasd: fail requests when dev state is not ready Message-id: 20090922201337.GC32244@blc4eb509856389.ibm.com O-Subject: [RHEL5 U5 PATCH 1/1] s390 - dasd: fail requests when device state is less then ready Bugzilla: 523219 RH-Acked-by: Pete Zaitcev <zaitcev@redhat.com> Description =========== Device mapper multipath device blocks when one path is set to state 'basic' In certain device mapper multipath / PPRC setups a DASD device gets quiesced and then set to state basic to flush it's queue and return all already queued requests back to device mapper. Sometimes it can happen that a request is queued after the device was set to basic, so that requests stays queued, is not processed and device mapper is blocked waiting for it. Solution: A DASD device that is not ready or online has no defined disk layout, so all requests that arrive in such a state need to be returned as failed. Bugzilla ======== BZ 523219 https://bugzilla.redhat.com/show_bug.cgi?id=523219 Upstream status of the patch ============================ The patch will be upstream as of kernel version 2.6.32 http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=97f604b07473669f4da52b1f859ee45e8d48d42a Test status =========== The patch has been tested and fixes the problem. The fix has been verified by the IBM test department. Please ACK. With best regards, Hans diff --git a/drivers/s390/block/dasd.c b/drivers/s390/block/dasd.c index 17c7224..2ad10fb 100644 --- a/drivers/s390/block/dasd.c +++ b/drivers/s390/block/dasd.c @@ -267,12 +267,12 @@ dasd_state_ready_to_basic(struct dasd_device * device) rc = dasd_flush_ccw_queue(device, 0); if (rc) return rc; + device->state = DASD_STATE_BASIC; dasd_destroy_partitions(device); dasd_flush_request_queue(device); device->blocks = 0; device->bp_block = 0; device->s2b_shift = 0; - device->state = DASD_STATE_BASIC; return 0; } @@ -1210,9 +1210,14 @@ __dasd_process_blk_queue(struct dasd_device * device) * for that. State DASD_STATE_ONLINE is normal block device * operation. */ - if (device->state != DASD_STATE_READY && - device->state != DASD_STATE_ONLINE) + if (device->state < DASD_STATE_READY) { + while ((req = elv_next_request(queue))) { + blkdev_dequeue_request(req); + dasd_end_request(req, 0); + } return; + } + nr_queued = 0; /* Now we try to fetch requests from the request queue */ list_for_each_entry(cqr, &device->ccw_queue, list)