Hi, On Mon, 25 Jul 2005, Steve Langasek wrote:
On Mon, Jul 25, 2005 at 02:06:46PM -0700, Blars Blarson wrote:In my opinion, we should drop support of all 32-bit sparc systems from Etch due to lack of people willing to spend the time to support them. This doesn't mean that we should delibaratly break things for them, but that the interest in continuing to support them is below what is needed to keep them as a viable part of Debian.Support of sun4c and sun4d was effectivly dropped from Sarge. The only reports trying d-i on this hardware that I remember seeing were failures, and noone bother to try to fix it. Upgrades from Woody may work, but were not well tested either.Were there actually install reports on sun4c and sun4d? I don't remember seeing any. Anyway, AIUI BenC killed these off years ago by changes to how gilbc was compiled.
I have built an unoptimized glibc and made it available. A couple of people have attempted the woody -> sarge upgrades on sun4c machines using it and instructions at http://wiki.debian.net/?SparcSun4c with results which I would describe as "mostly successful" :-).
Sun4m is the last supported 32-bit sparc architecture. Reportedly, the 2.6 kernel does not work in multi-processor mode on them, and dropping support of 2.4 from Etch is being discussed.Reportedly, current 2.6 kernels do not work *at all* on sun4m. This according to Jurij Smakov, who appears to currently be the sparc kernel maintainer in Debian.
Yes, indeed I could not make it work with neither 2.6.11 or 2.6.12 kernels. It would not boot with initrd, debugging showed that basic memory copying routines are broken. Removing initrd support (initial plan for 2.6.12) actually made it possible to boot, but it did not eliminate the problems, so I was still occasionally triggering a filesystem corruption in my tests. In the end we decided not to build the sparc32 kernel images for 2.6.12 release, so the first step towards dropping it has already been done. It also appears that the success or failure to boot is correlated with positions of memory chips in the slots, not really an acceptable situation.
Note that lack of hardware is not the problem, if anyone wants some sun4m systems (located in Los Angeles) let me know before they wind up in the recycle pile.
Blars, if you can keep one machine (fastest/most memory) until at least the end of September, I should be able to pick it up.
I have one here; works fine under sarge with a 2.4 kernel. I have no intention of spending large amounts of my own time to keep 2.6 viable on this architecture, though, when as it stands the box I have is only powered up for use as a porting machine and it can't even be used to build Debian kernels because depmod bombs out.
I am willing to fiddle with kernel options and available patches, but I'm really not hardcore enough to keep the sparc32 afloat. So unless someone upstream will start actively working on it again, I see dropping support for it as inevitable.
Best regards, Jurij Smakov jurij@wooyd.org Key: http://www.wooyd.org/pgpkey/ KeyID: C99E03CC