The skd driver has never handled discards reliably. The kernel will BUG as a result of issuing discards to the skd device. Disable the skd driver's discard support until it is proven reliable.
The device-mapper-test-suite test that exposed this bug just issues a discard that covers a portion of the skd device that was previously written through a dm-thin device. The discard spans the entire 1GB thin device (logical sector 0 through 2097152). dmtest run --profile stec --suite thin-provisioning -n /discard_fully_provisioned_device/ associated device-mapper-test-suite ruby test code follows: def test_discard_fully_provisioned_device with_standard_pool(@size) do |pool| with_new_thins(pool, @volume_size, 0, 1) do |thin, thin2| wipe_device(thin) wipe_device(thin2) assert_used_blocks(pool, 2 * @blocks_per_dev) thin.discard(0, @volume_size) assert_used_blocks(pool, @blocks_per_dev) end end ... Signed-off-by: Mike Snitzer <snit...@redhat.com> --- drivers/block/skd_main.c | 3 +++ 1 files changed, 3 insertions(+), 0 deletions(-) diff --git a/drivers/block/skd_main.c b/drivers/block/skd_main.c index eb6e1e0..5dadecc 100644 --- a/drivers/block/skd_main.c +++ b/drivers/block/skd_main.c @@ -4441,12 +4441,15 @@ static int skd_cons_disk(struct skd_device *skdev) /* set sysfs ptimal_io_size to 8K */ blk_queue_io_opt(q, 8192); +#if 0 + /* FIXME: Disable discard support until it no longer BUGs */ /* DISCARD Flag initialization. */ q->limits.discard_granularity = 8192; q->limits.discard_alignment = 0; q->limits.max_discard_sectors = UINT_MAX >> 9; q->limits.discard_zeroes_data = 1; queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, q); +#endif queue_flag_set_unlocked(QUEUE_FLAG_NONROT, q); spin_lock_irqsave(&skdev->lock, flags); -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/