Uhg, gcc is doing the idiotic llvm thing of maximally-hoisting large stack usage, and also float usage (saving call-saved float regs). This surprisingly leads to hardfloat code SIGILL'ing on nofpu machines even when the float code is never reached.
-
-
Failure to shrinkwrap float register save/restore is a fairly big performance flaw on archs where the fpu might be missing and the kernel is expected to trap-and-emulate. Suddenly non-float codepaths could become 10000x slower just because of the gratuitous save/restore.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Would you prefer a user to not bring up an issue they have with musl when they are unsure whether it's already known? Would you prefer them not bring it up when they are sure it was known, but was last discussed couple of years back as far as they know?
-
No, I'm looking into it before posting though.
-
I want to construct a minimal mips example that saves and restores fpu regs in a non-float code path.
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.