mwester at dls.net
Thu Nov 20 01:27:06 CET 2008
Werner Almesberger wrote:
> Mike (mwester) wrote:
>> The new kernel version is really going to be quite the shock to
>> user-space, I can see.
> I think it will greatly contribute towards a deeper understanding
> of all the sysfs paths that are used in there :-)
That it will -- and flush out the non-portable constructs that have
crept into all of the distros. Perhaps a wiki page needs to be created
with a table of the old/new mappings, and any discussion on any nuances
that may be associated with the new kernel switch.
> What's actually the best way to make the switch ? Do we just
> change the build and the various user spaces will use a frozen
> 2.6.24 kernel until they've updated the sysfs paths ?
Users have become very accustomed (in general) to flashing new kernels
-- yet there is no real interest in kernels built from the
stable-tracking branch at present. I think the need for suspend/resume
and correct behavior of the Glamo have been gating issues for users.
The distros will follow user demand, I think, and user demand for the
kernel switch will happen when the improved features (like the wifi
improvements) outweigh the remaining broken bits (Glamo).
So.... are we there yet? Can we (or do we already) have a list of the
outstanding bits of badness that need to be fixed so that we know when
it's time to encourage a switch?
More information about the openmoko-kernel