> The m5 binary on the disk image that we distribute doesn't support the > pin command. You'll need to compile it yourself. Additionally, I don't > believe libc on the disk image supports sched_setaffinity (needed by > pin). This is something we need to fix, but it unfortunately involves > distributing a new disk image to everyone and updating a bunch of > regressions. You can easily make a new disk image by starting with a > gentoo image or you can copy the libc* and libstdc++* from one of the > cross compilers on the m5sim.org website onto the disk image (you > might need to make the actual image bigger to do this).
Maybe we should just bite the bullet and do this. If we do, I think we should probably deprecate any stuff we don't like in the m5ops stuff. My inclination is to just generate a new disk image with gentoo 2008.0, recompile the binaries and move them over. Does anyone have a problem if we do this? Does anyone know if/when a 2009.0 is supposed to come out? Should we wait for it? I will probably be a many weeks before I could have time to do this, but it may at least be good to have a plan of attack so we can do the right thing when the time comes. Nate _______________________________________________ m5-dev mailing list m5-dev@m5sim.org http://m5sim.org/mailman/listinfo/m5-dev