Date: Fri, 2 Mar 2007 14:28:10 +0000 From: "Joao Barros" <joao.barros@gmail.com> To: "Gavin Atkinson" <gavin.atkinson@ury.york.ac.uk> Cc: freebsd-sparc64@freebsd.org Subject: Re: Ultrasparc 3 support Message-ID: <70e8236f0703020628u13d2b03al923ae1098a154d3b@mail.gmail.com> In-Reply-To: <1172667463.10762.12.camel@buffy.york.ac.uk> References: <45E3B8A5.108@uchicago.edu> <1172667463.10762.12.camel@buffy.york.ac.uk>
next in thread | previous in thread | raw e-mail | index | archive | help
On 2/28/07, Gavin Atkinson <gavin.atkinson@ury.york.ac.uk> wrote: > On Mon, 2007-02-26 at 22:50 -0600, Nathan Whitehorn wrote: > > I've been given a pair of Sunfire v210s and a v440 (dual and quad > > UltraSPARC 3, respectively), and would love to put FreeBSD on them. > > Sadly, of course, the schizo chipset isn't supported. > > > > I'm planning on trying to port over OpenBSD's work in this area, but am > > somewhat out of my depth (the only work I've ever done on the FreeBSD > > kernel involves some fixes to nve(4)). Because we have no immediate need > > to put any of these systems into production (and certainly not all > > three), I currently can use any or all of them for experimentation and > > development work for at least the next few months. I believe some > > members of this list have done their own work in this regard in the > > past, and was wondering if anyone wanted to collaborate or had done > > anything that would serve as a starting point. Similarly, if anyone more > > competent than I feels they just can just do it themselves, but needs > > access to the hardware, I would be happy to provide that. > > A fair bit of work has already been put into supporting the Schizo > chipset under FreeBSD, but it's a long way off being complete. With the > code that myself and Marius Strobl have published in the past, you can > get US3 based machines booting multiuser off root stored on NFS or USB > key. > > The single biggest part of the problem are the MMUs. Actually routing > read/writes and interrupts over the Schizo chipset isn't hard - but I > believe a lot of work would be needed to support the IOMMU found in > post-USII chipsets, along with the associated work to the PMAP code etc. > This is not an area I've looked into at all. > > What would be interesting from my point of view is to see how well the > OpenBSD code works. I've heard a few reports that it is unusable, and a > few reports that it sort of works. Knowing whether you can use DMA > under OpenBSD may be useful as it would give us another code base to > gain information from. > > Gavin I got this on my RSS today, Hope it shed's some more light on this subject: http://undeadly.org/cgi?action=article&sid=20070301230514 -- Joao Barros
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?70e8236f0703020628u13d2b03al923ae1098a154d3b>