]> git.ipfire.org Git - thirdparty/valgrind.git/commit
mips: allow Valgrind to be compiled for soft-float
authorPetar Jovanovic <mips32r2@gmail.com>
Wed, 12 Oct 2016 15:13:27 +0000 (15:13 +0000)
committerPetar Jovanovic <mips32r2@gmail.com>
Wed, 12 Oct 2016 15:13:27 +0000 (15:13 +0000)
commited74dc8a27e2d5f65380bd94c3ddc488c55673c0
tree03bb3d2d343e4486a295f714d827a581d75a3415
parent517d43c27820c2cba0e79ab856a021943cdf4226
mips: allow Valgrind to be compiled for soft-float

Force "hardfloat" mode for inline assembly that uses FPU instructions,
but pop original mode at the end of the assembly. Unhandled FPU
instructions will be handled by the signal handler (env_unsup_insn).

Skip MIPS specific tests for FPU if the code gets compiled for soft-
float.

This should allow Valgrind to be compiled as a soft-float binary, but
that executable should be used for soft-float systems only.

Related VEX change - r3261.
Related issue - BZ#351282.

Patch by Aleksandar Rikalo.

git-svn-id: svn://svn.valgrind.org/valgrind/trunk@16039
19 files changed:
coregrind/m_machine.c
none/tests/mips32/FPUarithmetic.c
none/tests/mips32/FPUarithmetic.vgtest
none/tests/mips32/MoveIns.c
none/tests/mips32/MoveIns.vgtest
none/tests/mips32/change_fp_mode.c
none/tests/mips32/change_fp_mode.vgtest
none/tests/mips32/fpu_branches.c
none/tests/mips32/fpu_branches.vgtest
none/tests/mips32/round.c
none/tests/mips32/round.vgtest
none/tests/mips32/round_fpu64.c
none/tests/mips32/round_fpu64.vgtest
none/tests/mips32/test_fcsr.c
none/tests/mips32/test_fcsr.vgtest
none/tests/mips32/vfp.c
none/tests/mips32/vfp.vgtest
none/tests/mips64/change_fp_mode.c
none/tests/mips64/change_fp_mode.vgtest