9:34 a.m.
On Tue, 2018-06-26 at 09:32 -0700, Dan Williams wrote:
On Tue, Jun 26, 2018 at 9:27 AM, Verma, Vishal L
<vishal.l.verma(a)intel.com> wrote:
> On Tue, 2018-06-26 at 10:29 -0600, Keith Busch wrote:
> > On Tue, Jun 26, 2018 at 09:19:28AM -0700, Verma, Vishal L wrote:
> > > On Tue, 2018-06-26 at 09:37 -0600, Keith Busch wrote:
> > > > The available_size attribute returns all the unused regions, but
> > > > a
> > > > namespace has to use contiguous free regions. This patch uses the
> > > > attribute returning the largest capacity that can be created for
> > > > determining if the namespace can be created.
> > > >
> > > > Signed-off-by: Keith Busch <keith.busch(a)intel.com>
> > > > ---
> > > > ndctl/lib/libndctl.c | 30 ++++++++++++++++++++++++++++++
> > > > ndctl/lib/libndctl.sym | 1 +
> > > > ndctl/libndctl.h | 2 ++
> > > > ndctl/namespace.c | 2 +-
> > > > 4 files changed, 34 insertions(+), 1 deletion(-)
> > >
> > > Hi Keith,
> > >
> > > The patch looks good, but just a couple of 'meta' comments.
> > > 1. We typically send ndctl patches separately from kernel patches
> > > (i.e.
> > > not
> > > thraded together).
> > > 2. for ndctl patches, an 'ndctl PATCH' prefix is recommended.
You
> > > can
> > > set a
> > > repo local config parameter for doing this automatically on git
> > > format-
> > > patch.
> > > git config format.subjectprefix "ndctl PATCH"
> > >
> > > I'm thinking the kernel changes will be queued for 4.19, which
> > > means
> > > the
> > > ndctl changes will go into v62.
> >
> > Thanks for the info. I'll make those changes for next time.
> >
> > I think I may need to send a v2 for this. Should we have this fall
> > back
> > to
> > the available_size for the older kernels where the
> > max_available_extents
> > attribute is not provided? I actually had that in my repo and used a
> > slightly older patch here, but I'm not sure if its okay to strongly
> > couple an ndctl release to a kernel version.
>
> I was thinking that too. Typically we don't guarantee ndctl to work
> with
> old kernels, but this does seem like a bit of an invasive change.
>
> Dan, thoughts?
It should fall back if the attribute is not available. Our *tests*
aren't guaranteed to pass on older kernels, but ndctl proper should
try it's best to accommodate old kernels.
Ah yes makes sense.