I don't really agree with the rationale, but I'm fine with reverting to it this late in the cycle, so: Acked-by: Christoph Hellwig <hch@xxxxxx> But looking at your slightly convoluted handling of the errors in dm-think, what do you think about splitting out these early checks into a separate helper, at which point __blkdev_issue_discard doesn't need a return value any more? -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel