Re: [PATCH v5] hostmem-file: add offset option

[David Hildenbrand]

On 11.04.23 13:46, Igor Mammedov wrote:
> On Wed, 5 Apr 2023 15:58:31 +0200
> David Hildenbrand <david@redhat.com> wrote:
>
> > On 04.04.23 16:36, Peter Xu wrote:
> > > On Mon, Apr 03, 2023 at 10:14:21PM +0000, Alexander Graf wrote:
> > > > Add an option for hostmem-file to start the memory object at an offset
> > > > into the target file. This is useful if multiple memory objects reside
> > > > inside the same target file, such as a device node.
> > > >
> > > > In particular, it's useful to map guest memory directly into /dev/mem
> > > > for experimentation.
> > > >
> > > > To make this work consistently, also fix up all places in QEMU that
> > > > expect fd offsets to be 0.
> > > >
> > > > Signed-off-by: Alexander Graf <graf@amazon.com>
> > >
> > > Acked-by: Peter Xu <peterx@redhat.com>
> > >
> > > I also agree it'll be nicer to split the fix into separate patch, though.
> > > The only affected part IIUC is multi-process QEMU since 6.0.0.  Copying the
> > > maintainers too so they'll be aware.
> > >
> > > Corresponds to the tag:
> > >
> > > Fixes: ed5d001916 ("multi-process: setup memory manager for remote device")
> > >    
> >
> > If there are no options on splitting out the fix, I'll route this via my
> > tree.
>
> Having fixes as separate prep patch is much more preferable.

Right. Question is if it's really worth it. (it's quite a corner cases I 
guess ...)

> Another question is if we should also check that provided
> offset honors 'align' option?

Good point, but I wonder if it's a real requirement (IOW, what would go 
wrong)? The important part seems to be that the offset is aligned to the 
underlying page size.

I'll be on vacation until 2. Mai, planning on picking it up by then to 
send it upstream.

--
Thanks,

David / dhildenb