Date: Thu, 16 Jun 2011 10:11:28 -0700 From: Matt Birkholz <[email protected]>
So obviously a less borked fix would be to adjust FP_CONTROL_WORD for glibc on i386 and x86-64. But doesn't our shiny new floenv.scm takes care of all this, withOUT going behind libc's back? Can we punt the ancient fpu inits in microcode/cmpauxmd/i386.m4 and x86-64.m4? (Dare me to try it?) Maybe. I'm not sure in what floating-point environment a restored band would start up if we removed the setting of the control word in the assembly hooks. I'd suggest masking all exceptions there, and then going through every entry point into a restored band (and the cold load) to make sure it sets up the floating-point environment appropriately. (And I have no idea what happens with Windows.) Perhaps Scheme ought to expose the x86 denormalized exception (with a null implementation for all the other machines we support -- er, uh, yeah), but I'm pretty sure it shouldn't be trapped by default. There's also a flush-to-zero control bit that we might want to expose somehow. It's Unity, isn't it? You can't stand Unity? I gave it up after a couple weeks -- went back to Ubuntu Classic. Maybe I should go back to Debian! No, it's more that I can't fit Ubuntu into my head, whereas I can fit NetBSD into my head, and NetBSD is much easier to fix when it's broken. _______________________________________________ MIT-Scheme-devel mailing list [email protected] https://lists.gnu.org/mailman/listinfo/mit-scheme-devel
