Re: Repeatable, raid1+O_DIRECT, hang/warn

On Mon, Jun 15, 2026 at 05:42:42PM +0000, Dr. David Alan Gilbert wrote:
> * Keith Busch ([email protected]) wrote:
> > On Mon, Jun 15, 2026 at 04:37:39PM +0000, Dr. David Alan Gilbert wrote:
> > > Hi Keith,
> > >   Thanks for the patch, alas it doesn't seem to be helping here;
> > >  the first warn is still the same
> > > and it still hangs the test process hard and eventually BUGs at
> > > 
> > > void blk_mq_end_request(struct request *rq, blk_status_t error)
> > > {
> > >         if (blk_update_request(rq, error, blk_rq_bytes(rq)))
> > >                 BUG();
> > 
> > Oh, that was not expected.
> > 
> > What is the dma alignment requirement of your backing devices? You can
> > find the attribute for sda at /sys/block/sda/queue/dma_alignment. I'm
> > expecting 511, but just want to double check.
> 
> Yeh looks like it:
> /sys/block/sda/queue/dma_alignment:511
> /sys/block/sdb/queue/dma_alignment:511
> 
> all of the lvm also looks like it is.

Thanks for confirming.

I'm struggling to see how you're getting there with your reproducer with
the proposal included. I can see other short comings with preadv or
really large pread's, but not with a 4k pread. For those other issues
this patch can fix it:

  https://lore.kernel.org/linux-block/[email protected]/

It is currently staged for upstream, so hasn't landed yet. But again, I
don't think those conditions apply to what you're seeing, but worth a
shot on top of the previous proposal to use byte units instead of
sectors.

In the meantime, since I so far can't reproduce this after including my
previous proposal, I may have to request trying out a debug patch to get
some more visibility on what's happening if that's okay.