Re: [PATCH v11 09/13] copy-on-read: skip non-guest reads if no copy needed

[Vladimir Sementsov-Ogievskiy]

22.10.2020 10:50, Andrey Shinkevich wrote:
> On 21.10.2020 23:43, Andrey Shinkevich wrote:
> > On 14.10.2020 18:22, Vladimir Sementsov-Ogievskiy wrote:
> > > 14.10.2020 15:51, Max Reitz wrote:
> > > > On 12.10.20 19:43, Andrey Shinkevich wrote:
> > > > > If the flag BDRV_REQ_PREFETCH was set, pass it further to the
> > > > > COR-driver to skip unneeded reading. It can be taken into account for
> > > > > the COR-algorithms optimization. That check is being made during the
> > > > > block stream job by the moment.
> > > > >
> > > > > Signed-off-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
> > > > > ---
>
> [...]
>
> > > > > diff --git a/block/io.c b/block/io.c
> > > > > index 11df188..bff1808 100644
> > > > > --- a/block/io.c
> > > > > +++ b/block/io.c
> > > > > @@ -1512,7 +1512,8 @@ static int coroutine_fn bdrv_aligned_preadv(BdrvChild 
> > > > > *child,
> > > > >       max_bytes = ROUND_UP(MAX(0, total_bytes - offset), align);
> > > > >       if (bytes <= max_bytes && bytes <= max_transfer) {
> > > > > -        ret = bdrv_driver_preadv(bs, offset, bytes, qiov, qiov_offset, 0);
> > > > > +        ret = bdrv_driver_preadv(bs, offset, bytes, qiov, qiov_offset,
> > > > > +                                 flags & bs->supported_read_flags);
> > >
> > >
> > > When BDRV_REQ_PREFETCH is passed, qiov may be (and generally should be) NULL. 
> > > This means, that we can't just drop the flag when call the driver that doesn't 
> > > support it.
> > >
> > > Actually, if driver doesn't support the PREFETCH flag we should do nothing.
> > >
> > >
> > > > Ah, OK.  I see.  I expected this to be a separate patch.  I still wonder
> > > > why it isn’t.
> > >
> > >
> > > Could it be part of patch 07? I mean introduce new field supported_read_flags 
> > > and handle it in generic code in one patch, prior to implementing support for 
> > > it in COR driver.
> >
> > We have to add the supported flags for the COR driver in the same patch. Or 
> > before handling the supported_read_flags at the generic layer (handling zero 
> > does not make a sence). Otherwise, the test #216 (where the COR-filter is 
> > applied) will not pass.
> >
> > Andrey
>
> I have found a workaround and am going to send all the related patches as a 
> separate series.

What is the problem?

If in a separate patch prior to modifying COR driver, we add new field 
supported_read_flags and add a support for it in generic layer, when no driver 
support it yet, nothing should be changed and all tests should pass..



--
Best regards,
Vladimir