Re: [PATCH] block: use the request length for iov alignment

[Kevin Wolf]

Am 13.09.2022 um 15:12 hat Keith Busch geschrieben:
> On Thu, Sep 08, 2022 at 09:45:26AM -0700, Keith Busch wrote:
> > From: Keith Busch <kbusch@kernel.org>
> > 
> > An iov length needs to be aligned to the logical block size, which may
> > be larger than the memory alignment.
> 
> [cc'ing some other interested folks]
> 
> Any thoughts on this patch? It is fixing an observed IO error  when running
> virtio-blk with the default 512b logical block size backed by a drive 
> formatted
> with 4k logical block.

I need to take a real look after KVM Forum, but my first thought was
that we might be overloading request_alignment with multiple meanings
now (file offset alignment and memory address alignment), and the values
just happen to be the same for files on Linux.

Did you consider a separate iov_alignment or similar and intentionally
decided against it, or is it something you just didn't think about?

Kevin

> > ---
> >  block/io.c | 3 ++-
> >  1 file changed, 2 insertions(+), 1 deletion(-)
> > 
> > diff --git a/block/io.c b/block/io.c
> > index 0a8cbefe86..296d4b49a7 100644
> > --- a/block/io.c
> > +++ b/block/io.c
> > @@ -3243,13 +3243,14 @@ bool bdrv_qiov_is_aligned(BlockDriverState *bs, 
> > QEMUIOVector *qiov)
> >  {
> >      int i;
> >      size_t alignment = bdrv_min_mem_align(bs);
> > +    size_t len = bs->bl.request_alignment;
> >      IO_CODE();
> >  
> >      for (i = 0; i < qiov->niov; i++) {
> >          if ((uintptr_t) qiov->iov[i].iov_base % alignment) {
> >              return false;
> >          }
> > -        if (qiov->iov[i].iov_len % alignment) {
> > +        if (qiov->iov[i].iov_len % len) {
> >              return false;
> >          }
> >      }
> > -- 
> > 2.30.2