Auditing pmem driver teardown operations, while developing get_user_pages() support for dax [1], revealed that we can trivially crash the kernel by triggering new i/o requests after unbinding the pmem driver. In fact, any bio-based driver is susceptible to this crash because the queue draining done at shutdown uses in flight 'struct request' objects to pin the queue active.
Solve the problem generically for all drivers and export the new blk_queue_enter() and blk_queue_exit() helpers for dax to indicate when the "request queue" is busy (i.e. we are actively using an address returned by ->direct_access()). [1]: https://lists.01.org/pipermail/linux-nvdimm/2015-September/002199.html --- Dan Williams (2): block: generic request_queue reference counting block, dax: fix lifetime of in-kernel dax mappings block/blk-core.c | 71 +++++++++++++++++++++++--- block/blk-mq-sysfs.c | 6 -- block/blk-mq.c | 80 +++++++++--------------------- block/blk-sysfs.c | 3 - block/blk.h | 12 ++++ fs/dax.c | 130 +++++++++++++++++++++++++++++++----------------- include/linux/blk-mq.h | 1 include/linux/blkdev.h | 4 + 8 files changed, 185 insertions(+), 122 deletions(-) -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
