Stephen Hahn wrote:
* Garrett D'Amore <gdam...@sun.com> [2009-10-28 17:46]:
Stephen Hahn wrote:
* Garrett D'Amore <gdam...@sun.com> [2009-10-28 16:12]:
   * containing platform support and drivers not in ON
   * drivers might also include userland components (X11 modules?)
  Are these only drivers that have been de-supported, or are you
  encouraging new driver development as well?
I don't see why new drivers couldn't be done here... eventually such drivers might transition to ON, but this would be a low barrier way to get stuff "out there".

  I would rather see an effort to continue to work on easing the ON path
  for new driver support than introduce a competitor.

That will be impossible until you allow non-Sun folks and non-Sun-business needs to determine ON. I don't see that as realistic at this point.

The idea here is not to compete with ON, but to provide an *alternative* when getting into ON is unrealistic.

3) Sun would have no official support for this new consolidation.

   * no support in bugster for this stuff
   * not part of any official Sun distro
   * caveat emptor
  If there's no SCAs required, then source in this consolidation will
  have an extra hurdle (potentially insurmountable) if, after a period
  of time, its maintainers want to port it to an SCA-requiring
  consolidation.  (My suspicion is that a revision sequence without SCAs
  covering each step will result in a longer diligence check.  Also, you
  haven't stipulated the set of software licenses allowed.)
We import stuff from other FOSS sources into ON all the time. It might be that the combination of CDDL and no SCA is toxic, but other licenses (e.g. BSD licenses) don't have this problem.

  Without the SCA, it means that every import will go through the OSR
  process, which is the diligence I mentioned above.  Keep good notes,
  or require an SCA.  (The BSD license isn't a panacea, as it lacks any
  coverage of applicable patents.)

  It might be interesting to have the output of this effort distributed
  via the contrib/ repository--that work has a similar "at own risk"
  aspect, and has a submission and acceptance policy already in place.
  (So the Maintained Open Legacy Drivers (?) consolidation could focus
  on code, building, and testing.)
That's a good idea. Although at present I have been unable to get any information from the IPS folks about what is required to deliver a *driver* via IPS. (Specifically there is the question of handling the postinstall requirements for drivers -- you have to run add_drv.)

  Image packaging appears to successfully install drivers with some
  regularity.  Let's see...

Okay, I'll have to research this. The problem isn't *installing* such an IPS package, its figuring out how to *construct* one. (At least for me.)

  $ man pkg.5
  /Driver
  ...
    Driver actions
     The 'driver' action represents a device driver.  It does not reference
     a payload: the driver files themselves must be installed as file
     actions.  The following attributes are recognized (see add_drv(1m) for
     more information):

     name         The name of the driver.  This is usually, but not always,
                  the filename of the driver binary.  This is the driver
                  action's key attribute.

     alias        This represents an alias for the driver.  There may be
                  more than one alias attribute for any given driver.
                  There are no special quoting rules necessary.

     class        This represents a driver class.  There may be more than
                  one class attribute for any given driver.

     perms        This represents the filesystem permissions for the
                  driver's device nodes.

     clone_perms  This represents the filesystem permissions for the
                  "clone" driver's minor nodes for this driver.

     policy       This specifies additional security policy for the device.
                  There may be more than one policy attribute for any given
                  driver, but no minor device specification may be present in
                  more than one attribute.

     privs        This specifies privileges used by the driver.  There may
                  be more than one privs attribute for any given driver.

     devlink      This specifies an entry in /etc/devlink.tab.  The value
                  is the exact line to go into the file, with tabs denoted
                  by "\t".  See devlinks(1M) for more information.  There
                  may be more than one devlink attribute for any given
                  driver.

  For a long example driver action:

  $ pkg contents -m SUNWintgige | egrep ^driver

  You can also grep for driver in /var/pkg/pkg on a typical OpenSolaris
  system for other examples.

We also need to make a consolidation available in tradition form so other folks can build LiveCDs of their own, or other distros can import it. (Imagine a legacy version of milax, for example.)

  If you mean a proto area, then, yes, that's polite.

Yes, basically a proto area -- although it has to include at least meta data required to install the bits. I'd envisioned providing SVR4 packages as well, but if no distribution needs them and IPS works for everyone, then we could just do that. I don't have a strong feeling one way or the other -- I lean towards SVR4 just because that's what I know.

   - Garrett

_______________________________________________
opensolaris-code mailing list
opensolaris-code@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/opensolaris-code

Reply via email to