Re: [PATCH v2 0/3] nvdimm: Add an IOCTL pass thru for DSM calls

[Jerry Hoemann <jerry.hoemann@hpe.com>]

On Tue, Nov 17, 2015 at 09:05:28AM -0800, Dan Williams wrote:
> On Tue, Nov 17, 2015 at 8:56 AM, Jerry Hoemann <jerry.hoemann@hpe.com> wrote:
> > On Mon, Nov 16, 2015 at 05:29:41PM -0800, Dan Williams wrote:
> >> On Mon, Nov 16, 2015 at 1:10 PM, Jerry Hoemann <jerry.hoemann@hpe.com> wrote:
> >> > On Mon, Nov 16, 2015 at 11:00:20AM -0800, Dan Williams wrote:
> >> >> On Mon, Nov 16, 2015 at 10:38 AM, Jerry Hoemann <jerry.hoemann@hpe.com> wrote:
> >> >> >
> >> >>
> >
> > ...
> >
> >> I want to do it in separate steps, I'd just like to see cmd number 100
> >> added to the existing __nd_ioctl and acpi_nfit_ctl routines.  That
> >
> >    Why?
> 
> Because there's no need for the intel vs passthru distinction, it's
> just yet another command.

  Yes and no.  The way the two marshal their arguments in prep for
  the copy in/copy out is different and are not compatible.

  Also, the existing upstream acpi_nfit_ctl does multiple things that
  we don't want done in the "semi" passthru case.

  To accomodate these differences, I implemented in separate functions.
  I can merge the functions together, it will not be clean.

  This approach also creates testing issues I didn't have previously.  I
  was confident w/ code inspection that I wasn't breaking the existing
  usage case.  I will need your help in testing on hardware that I don't
  have access to.

  You expressed a desire to depricate the existing ioctl commands
  and transition to the semi passthru structure.

  What do you anticipate that code looking like?

> 
> >> plus quibbling about the name "ND_CMD_PASSTHRU".  Given the plans to
> >> eventually replace the existing commands we can call it something like
> >> 'ND_DSM_GENERIC'.
> >
> >
> >   No problem.  I'll change the name for ndn_passthru_pkg in a similar fashion.
> >
> >
> >   Question:     Are you planning to add other CMDs to the IOCTL in the future?
> >                 (eg. ones not directly related to calling _dsm?)
> >
> >                 Or, is the ultimate goal to have an IOCTL that supports
> >                 only the generic DSM call?
> 
> I'm not ruling out the possibility that there may be a non-DSM command
> in the future, but I don't see any need for that on the horizon.  Why
> would it matter?

  Neither the existing upstream apci_nfit_ctl nor the semi pass thru
  marshal arguments in a traditional straight forward manner.  So likely
  the marshaling code for any new commands would be different.

  Also, since it doesn't call DSM it wouldn't be doing the evaluate dsm.



-- 

-----------------------------------------------------------------------------
Jerry Hoemann            Software Engineer      Hewlett-Packard Enterprise
-----------------------------------------------------------------------------