Re: [PATCH 3/4] Move CONFIG_XFS handling to meson.build

[Thomas Huth]

On 10/12/2021 09.39, Paolo Bonzini wrote:
> On 12/10/21 08:53, Thomas Huth wrote:
> > On 02/11/2021 12.34, Paolo Bonzini wrote:
> > > On 28/10/21 20:59, Thomas Huth wrote:
> > > > Checking for xfsctl() can be done more easily in meson.build. Also,
> > > > this is not a "real" feature like the other features that we provide
> > > > with the "--enable-xxx" and "--disable-xxx" switches for the
> > > > configure script, since this does not influence lots of code (it's
> > > > only about one call to xfsctl() in file-posix.c), so people don't
> > > > gain much with the ability to disable this with "--disable-xfsctl".
> > > > Let's rather treat this like the other cc.has_function() checks in
> > > > meson.build, i.e. don't add a new option for this in meson_options.txt.
> > > >
> > > > Signed-off-by: Thomas Huth <thuth@redhat.com>
> > >
> > > I think we should just use ioctl and copy the relevant definitions from 
> > > Linux:
> > >
> > > struct dioattr {
> > >          u32           d_mem;          /* data buffer memory alignment */
> > >          u32           d_miniosz;      /* min xfer size                */
> > >          u32           d_maxiosz;      /* max xfer size                */
> > > };
> > >
> > > #define XFS_IOC_DIOINFO        _IOR ('X', 30, struct dioattr)
> >
> > I've now had a closer look at this idea, but it's getting messy: We'd 
> > additionally also need the platform_test_xfs_fd() function that is called 
> > from file-posix.c ...
>
> platform_test_xfs_fd() is only used to decide whether to invoke 
> XFS_IOC_DIOINFO; but failures of XFS_IOC_DIOINFO are ignored anyway, so we 
> can get rid of is_xfs in BDRVRawState, too.

After staring at the code for a while, I wonder why we're not simply using 
fstat() here instead to get the st_blksize value... wouldn't that be better 
anyway since it also works with other file system types?

 Thomas