Re: Release sprint results - team changes, auto-rm and arch status
(Re-posting, pressed send accidentally.)
On Friday, November 29, 2013, Aurelien Jarno wrote:
Some precision about the MIPS machines:
On Thu, Nov 28, 2013 at 09:04:56PM +0100, Niels Thykier wrote:
> * [mips, mipsel] buildds/porterboxes run on hardware which is very old or has known defects:
> - mips octeon is unstable
Better me more precise there, Octeon machines in general are very
stable. That said out of the three machines we have in Debian, two are
unstable, the other one is very stable. We never really understood why,
we only have remarked they have different CPU revision number.
> - mipsel loongson have CPU bugs
I see on http://release.debian.org/jessie/arch_qualify.html that it
concerns the "NOP implementation error" from Loongson 2F. Debian is
using Loongson 2E for buildds and porters machines, which are not
affected by this issue.
We've got anther 3A board today, which is an official product of Loongson, Inc. that can be purchased on the market. We are able to run a Sid mips64el system on the board without any visible glitch like the previous one from Lemote, so I think 3A boards like the ones we have are really good candidates for replacing old mipsel buildds and porter boxes. They are much faster (by being quad-core, and some boards have 2 CPU config), and support larger memory (standard config is 4GB). Also there are willing party that is considering donating several to Debian.
What we need to fix are their kernels and PMON:
1. Kernel support isn't upstreamed, and we have only succeeded in merging the patch set to 3.6 and some lower versions which can produce a working binary. We are running 3.6 for Lemote products and 2.6.36 for Loongson ones.
2. Different boards use quite different PMON, most of them support 2GB DDR3 RDIMM, so the maximum configuration will be limited to 4x2GB. This could be a smaller issue comparing with the first one though, we are working with Lemote to see if 8GB UDIMM can be officially supported on their boards.