> On Fri, Oct 10, 2008 at 12:59 PM, Robert Healey > <[EMAIL PROTECTED]> wrote: > > Does anyone else have a large installation of sun4m > architecture > > machines it would be nice to run something other > than Linux/NetBSD on? > > > > SS20, SS5, SS10, various server 4m machines. > > Solaris 10 and later only support 64-bit on sparc - a > 32 bit sparc > kernel does not exist. Further, the (MMU of the) > UltraSPARC I CPU is > not supported. As such, the oldest/slowest sparc > machine that you can > have a reasonable expectation of support on is an > Ultra 2.
Hi! Only a small correction: * a 32-bit kernel does exist and is still part of OS/Net (namely for x86 IA32) * the reason why it isn't useful for 32-bit SPARC anymore is, because all 32-bit SPARC iommu-code has been dropped from OS/Net in one of the earliest (then still) internal-only builds of Solaris 10 in 2003 (circa build 10, don't recall). There is nothing special in this context to the Ultra1's 32bit iommu. The Ultra1 is a normal type_a 64-bit sun4u machine that can be booted either in 32-bit or 64-bit mode (from IIe, IIe+, III and IIIi on only 64-bit kernels can be booted, let's call those type_b). * The reason why the U1 64-bit support has beeen artificially disabled inside Solaris has to do with security reasons, because of (very special and unlikely to happen in daily use) incorrect calculation results when U1's run a 64-bit kernel. However, this bug is documented can be exploited. To avoid legal problems Sun decided as early as in Solaris 2.7 (the first release with 64-bit kernel) to disable 64-bit support on U1's, and to abort the boot process fatally and print out an error message (or at first maybe a non-fatal warning msg). HOWEVER: If you let it, it will just boot and work on Ultra1 based systems, even Solaris 10_current or Nevada_current. You also have to change build target settings in the Makefiles from their current default ultra2 to ultra1. I'm not sure if Studio12 still supports the original v9 and v8plus targets (rather than v9[a-c]... or v8plus8[a-c] ...). But if you use gcc as primary compiler and disable shadow compilation (and hence the then secondary Studio compiler), then you are safe anyway. Rainer Orth has researched the Ultra1 revival relevant aspects as early as more than two years ago: http://www.opensolaris.org/jive/thread.jspa?messageID=35672識 http://www.opensolaris.org/jive/thread.jspa?threadID=8028&tstart=0 He also posted diffs to his homepage. I talked about it during OSDEVCONF2007, here is my script: http://www.martux.org/MartUX__OSDevCon2007.pdf Be aware that the legacy MartUX LiveDVD ( http://www.genunix.org/distributions/martux/marTux_0.2/ ) supports Ultra1 systems without problems. My current Natamar LiveCD however doesn't boot on my Ultra1's anymore, presumably because of the new hsfs bootblk. But one could most probably solve this issue. Regards, Martin Bochnig Natamar OpenSolaris distro project developer && Sun-certified network admin -- This message posted from opensolaris.org _______________________________________________ indiana-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/indiana-discuss
