Messages in this thread | | | Date | Mon, 15 May 2006 20:49:35 +0200 | From | Ingo Molnar <> | Subject | Re: [PATCH] x86 NUMA panic compile error |
| |
* Andi Kleen <ak@suse.de> wrote:
> > So shouldn't such a patch remove that code rather than panicing? > > I would be for remove, but apparently we have one or two users in IBM > that run their x440s (32bit only) with CONFIG_NUMA. No distributions > do so though and I would expect x440s to usually run distributions > because they are quite expensive machines. > > My arguments for remove: > - The code is very hackish - it was written before the proper ACPI > infrastructure is in place - and NUMA on 32bit in general needs a lot > of hacks because of the limited ZONE_NORMAL.
works fine here now. The whole NUMA code is still quite hackish in general, (including most of arch/x86_64/*/*.c), so i'd not judge based on that.
> - NUMA on 32bit is kind of broken by design.
well. 32bit itself is broken by design, if you consider RAM larger than say 1GB.
> - It isn't used much.
it's an enabler of a feature-set that i couldnt test on these boxes otherwise. Look at it like the highmem= boot option. Or consider it a primitive form of NUMA emulation.
> - It breaks often
Martin says he's daily testing it in his grid.
> - It tends to not work on Opterons and hits the users who try it > there.
maybe due to the zone alignment problem?
Ingo - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |