Hi Jens This patch set fixes some defects during update hw queues.
1st patch refactor the blk-mq debugfs and sysfs register during blk_mq_update_nr_hw_queues. 2nd patch change the GFP_KERNEL to GFP_NOIO during blk_mq_realloc_hw_ctxs, 3rd patch try to realloc hctx when this hctx is mapped to a different node. 4th patch should be able to fix the panic reported in follwing link https://marc.info/?l=linux-block&m=153799465603723&w=2 V2: - Newly add 2nd and 3rd patch. - Use blk_mq_map_queues when __blk_mq_update_nr_hw_queues fallbacks to previous nr_hw_queues to avoid driver's .map_queues leak mapping of some cpus due the fallback nr_hw_queues. - Add Ming's Reviewed-by in 1st patch. Jianchao Wang[4] blk-mq: adjust debugfs and sysfs register when blk-mq: change gfp flags to GFP_NOIO in blk-mq: realloc hctx when hw queue is mapped to blk-mq: fallback to previous nr_hw_queues when block/blk-core.c | 2 +- block/blk-flush.c | 6 +-- block/blk-mq.c | 155 ++++++++++++++++++++++++++++++++++-------------------- block/blk.h | 2 +- 4 files changed, 102 insertions(+), 63 deletions(-)
