sparc64 machfb: (EE) Unable to map mmio aperture. Invalid argument (22)

Marius Strobl marius at alchemy.franken.de
Wed Jun 26 13:15:55 UTC 2013


On Wed, Jun 26, 2013 at 08:42:31AM +0100, Anton Shterenlikht wrote:
> 
> Finally, I gather from what I read in ports@
> in the last couple of weeks that the new Xorg is
> unlikely to work on sparc64 ever and the

"Ever" is the wrong word here; along with once a again fixing "old"
Xorg as of before the last set of port updates, I made "new" Xorg
build on !x86 but the result didn't look to promising on sparc64.
I plan to look at getting the current "new" Xorg to work there once
the patch for fixing the "old" one is committed. Some relatively
simple things are good candidates for why "new" Xorg failed that
miserably last time. On the other hand, something more fundamental
might also be the culprit so it's unclear at this point how much
effort it'll take to make it work on sparc64.

> old (current) will no longer be supported
> after a while, so future, say 2-3 years ahead
> of X on sparc64 looks bleak, right?
> 

IMO, the future of Xorg on FreeBSD looks bleak in general, mainly
due to the fact that we currently lack the necessary passionate
_and_ knowledgeable people for maintaining it. Don't get this
wrong; kib@ certainly spent a tremendous effort on getting KMS
to work for Intel graphics hardware. However, AFAIK the version
that went into FreeBSD isn't exactly complete and was already
rather "old" compared to what was in Linux at that time, with
things having changed there considerably since then. Also, KMS
in FreeBSD is largely unmaintained after the initial drop. The
folks that currently maintain the Xorg ports certainly are rather
active but, unfortunately, they simply are not of the same grade
as f. e. anholt@ and rnoland@ were in the past. In other words,
FreeBSD already is seriously behind Linux regarding infrastructure
needed for Xorg and it would take quite some efforts to just
catch up, let alone constantly keeping up with them. At least for
me, the two main reasons demotivating work on Xorg are that a) the
future of Xorg itself is questionable and b) Xorg constantly keeps
removing support for older generations of hardware upstream. IMO,
if we were to spend manpower on improving desktop support in FreeBSD
effectively, we should leave Xorg alone and either work on getting
Wayland to work the right way, i. e. without x86-specific hacks
and shortcuts, _now_. Alternatively, porting OpenBSD's Xorg "fork"
Xenocara in order to piggy-back on their superior support for
desktop hardware - similar to all the drivers for WIFI chipsets we
ported from them - also seems to be a good way forward to me. This
is _not_ to say, that OpenBSD generally would be superior to FreeBSD,
though :)

Marius



More information about the freebsd-sparc64 mailing list