On 3/29/19 9:00 AM, Bob Liu wrote: > On 3/29/19 10:52 PM, Jens Axboe wrote: >> On 3/29/19 8:51 AM, Martin K. Petersen wrote: >>> >>> Jens, >>> >>>> The integrity stuff still has that nasty pointer in there. It'd be >>>> nice to get rid of that as well, and hide it all in a parent container >>>> of the bio. >>> >>> That's fine with me. We really didn't have that option a decade ago. >> >> Right, but now we do. Hence trying to nudge things in that direction, >> instead of adding another pointer to struct bio for a new use case. I'd >> much rather go that route, instead of unionizing with another pointer >> that should be going away instead. >> > > No problem, will update in next version. > Let me wait a while for feedback about the verify callback. > Thank you! I think you're still missing my point. If you go the bio_set and container-of-bio route, then you can do whatever you want. You will not need a callback in the bio, you will just have a private end_io function for that particular bio that does the verification. -- Jens Axboe
