[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#696038: ITP: compat-wireless-3.5.4 -- backported linux wireless drivers



On Sat, 2012-12-22 at 20:36 +1100, Jayen Ashar wrote:
> On Sat, Dec 22, 2012 at 12:18 AM, Ben Hutchings <ben@decadent.org.uk> wrote:
> > On Mon, 2012-12-17 at 12:41 +1100, Jayen Ashar wrote:
> >> On Mon, Dec 17, 2012 at 12:07 PM, Ben Hutchings <ben@decadent.org.uk> wrote:
> >> > On Mon, 2012-12-17 at 11:14 +1100, Jayen Ashar wrote:
> >> >> Hi kernel team,
> >> >>
> >> >> I intend to package the stable versions of compat-wireless and
> >> >> compat-drivers (when available) as module-assistant packages, starting
> >> >> with compat-wireless 3.5.4.  The reason I am using module-assistant
> >> >> (instead of dkms) is so that embedded users of debian can compile &
> >> >> distribute kernel module packages without needing a compiler on the
> >> >> embedded device.
> >> >
> >> > This is not a reason to avoid DKMS, as you can use 'dkms mkdeb' to build
> >> > a package for installation elsewhere.
> >>
> >> didn't know about mkdeb.  will start looking at dkms (at least for our
> >> internal use).  thanks.
> >>
> >> >> I am starting with 3.5.4 because I have already
> >> >> packaged it for my company and have tested it internally with the
> >> >> squeeze kernel.
> >> >
> >> > Since wheezy is now frozen, this cannot be added to wheezy and therefore
> >> > cannot be added to squeeze-backports either.  So this will not be a very
> >> > effective way to help other squeeze users.
> >>
> >> Could it be added to wheezy-backports?
> >
> > Yes, once it's in testing for jessie.
> >
> >> > The official kernel package (linux-2.6/linux source package) does get
> >> > stable updates to extend hardware support, but not as many as I would
> >> > like.  I would really like to see compat-drivers integrated into the
> >> > linux source package so users don't need to look for extra package or a
> >> > special installer.  But we will need to ensure that there is adequate
> >> > regression testing whenever we update compat-drivers in stable.
> >> >
> >> > Would you be interesting in working on this?  It would help many more
> >> > users than a separate package.
> >>
> >> Are you asking for work on adding it to linux-source, or doing some
> >> testing?  I'm not sure I could provide adequate testing on any version
> >> other than the one's my company is running.
> >
> > I'm requesting that you work on adding it to the linux source package
> > and do what testing you can.  In any case we'll have to ask users for
> > beta testing.
> >
> > Before you do that work, though, it would need to be agreed among the
> > kernel and release teams that we should update drivers in this way.  But
> > I hope that's not too controversial.
> 
> Do you think there is value in having many versions available?
> (Ubuntu seems to have this.)  I was planning to separately package
> 3.3, 3.4, 3.5, and 3.6, so users could have the version they wanted.

Users will want the version that works, which they likely won't know in
advance.  We should take responsibility for picking a good version and
any necessary bug fixes on top of it.

Ben.

> (I was planning on earlier versions as well, but if there's no hope of
> getting into squeeze-backports, I see little purpose.)

-- 
Ben Hutchings
Every program is either trivial or else contains at least one bug

Attachment: signature.asc
Description: This is a digitally signed message part


Reply to: