On Thu, 2018-12-06 at 15:21 -0700, Jens Axboe wrote: +AD4 On 12/6/18 3:20 PM, Jens Axboe wrote: +AD4 +AD4 After the direct dispatch corruption fix, we permanently disallow direct +AD4 +AD4 dispatch of non read/write requests. This works fine off the normal IO +AD4 +AD4 path, as they will be retried like any other failed direct dispatch +AD4 +AD4 request. But for the blk+AF8-insert+AF8-cloned+AF8-request() that only DM uses to +AD4 +AD4 bypass the bottom level scheduler, we always first attempt direct +AD4 +AD4 dispatch. For some types of requests, that's now a permanent failure, +AD4 +AD4 and no amount of retrying will make that succeed. +AD4 +AD4 +AD4 +AD4 Don't use direct dispatch off the cloned insert path, always just use +AD4 +AD4 bypass inserts. This still bypasses the bottom level scheduler, which is +AD4 +AD4 what DM wants. +AD4 +AD4 +AD4 +AD4 Fixes: ffe81d45322c (+ACI-blk-mq: fix corruption with direct issue+ACI) +AD4 +AD4 Signed-off-by: Jens Axboe +ADw-axboe+AEA-kernel.dk+AD4 +AD4 +AD4 Bart, I'll add your reported-by here of course, and also a stable CC +AD4 since the original patch went into stable. Feel free to add the following: Tested-by: Bart Van Assche +ADw-bvanassche+AEA-acm.org+AD4
