Travis Downs

@andrwng - very sorry I never saw this request for logs. Please find them attached: [issue-9505-logs.tar.gz](https://github.com/redpanda-data/redpanda/files/11639028/issue-9505-logs.tar.gz) For whatever reason I don't have the logs from 3/15 for 2 of the...

We considered something like this, but I had thought was unsafe since foreign pointer only helps for deletion, but some other tb operations (e.g., `.share()`) also manipulate the reference count....

@avikivity wrote: > I think it's better to re-architect so sharing happens at the request level rather than buffer level. That is, have some object that owns the buffers, and...

@tgrabiec wrote: > You mean deleter::impl::refs? Yes. > When migrating temporary_buffer to another shard, you create a new chain of deleters which is only manipulated on the other shard, e.g.:...

> My suggestion works if the caller halts work on the data while the callee is processing. Yes, but in the scenario we are considering this is almost impossible: the...

> The native one? Or posix? posix > In either case, my second suggestion will work (and indeed is safer, you keep the refcounts always local). Yes, agreed. I just...

> It will not block, io_uring is actually async, aio is just not. Much thanks for the speedy reply! What are the mechanics of that? Doesn't io_uring rely on the...

I decided to try it, here's an example of the blocking I'm thinking about: ``` fio 3642736 [007] 359323.757309: sched:sched_switch: prev_comm=fio prev_pid=3642736 prev_prio=120 prev_state=D ==> next_comm=swapper/7 next_pid=0 next_prio=120 ffffffff8193f5d4 __traceiter_sched_switch+0x44...

Thanks Jens, your response is much appreciated. > `RQF_NOWAIT` _is_ the mechanism. If something is blocking on other IO off the submission path, then that piece is buggy. In this...

> Just back from 2 weeks away so don't have a lot of bandwidth to test this, but remind me end next week if I don't get back to you...