Re: [dm-devel] [PATCH 0/6] dax poison recovery with RWF_RECOVERY_DATA flag

From: Dan Williams
Date: Thu Nov 04 2021 - 21:35:26 EST


On Thu, Nov 4, 2021 at 5:46 PM Dan Williams <dan.j.williams@xxxxxxxxx> wrote:
>
> On Thu, Nov 4, 2021 at 1:27 PM Jane Chu <jane.chu@xxxxxxxxxx> wrote:
> >
> > On 11/4/2021 12:00 PM, Dan Williams wrote:
> >
> > >>
> > >> If this understanding is in the right direction, then I'd like to
> > >> propose below changes to
> > >> dax_direct_access(), dax_copy_to/from_iter(), pmem_copy_to/from_iter()
> > >> and the dm layer copy_to/from_iter, dax_iomap_iter().
> > >>
> > >> 1. dax_iomap_iter() rely on dax_direct_access() to decide whether there
> > >> is likely media error: if the API without DAX_F_RECOVERY returns
> > >> -EIO, then switch to recovery-read/write code. In recovery code,
> > >> supply DAX_F_RECOVERY to dax_direct_access() in order to obtain
> > >> 'kaddr', and then call dax_copy_to/from_iter() with DAX_F_RECOVERY.
> > >
> > > I like it. It allows for an atomic write+clear implementation on
> > > capable platforms and coordinates with potentially unmapped pages. The
> > > best of both worlds from the dax_clear_poison() proposal and my "take
> > > a fault and do a slow-path copy".
> > >
> > >> 2. the _copy_to/from_iter implementation would be largely the same
> > >> as in my recent patch, but some changes in Christoph's
> > >> 'dax-devirtualize' maybe kept, such as DAX_F_VIRTUAL, obviously
> > >> virtual devices don't have the ability to clear poison, so no need
> > >> to complicate them. And this also means that not every endpoint
> > >> dax device has to provide dax_op.copy_to/from_iter, they may use the
> > >> default.
> > >
> > > Did I miss this series or are you talking about this one?
> > > https://lore.kernel.org/all/20211018044054.1779424-1-hch@xxxxxx/
> >
> > I was referring to
> >
> > http://git.infradead.org/users/hch/misc.git/shortlog/refs/heads/dax-devirtualize
> > that has not come out yet, I said early on that I'll rebase on it,
> > but looks like we still need pmem_copy_to/from_iter(), so.
>
> Yeah, since the block-layer divorce gets rid of the old poison
> clearing path, then we're back to pmem_copy_to_iter() (or something
> like it) needing to pick up the slack for poison clearing. I do agree
> it would be nice to clean up all the unnecessary boilerplate, but the
> error-list coordination requires a driver specific callback. At least
> the DAX_F_VIRTUAL flag can eliminate the virtiofs and fuse callbacks.

Although, if error management is generically implemented relative to a
'struct dax_device' then there wouldn't be a need to call all the way
back into the driver, and the cleanup could still move forward.