Account for MacOS X ABI reserved space in linkage area (Andreas Faerber)
git-svn-id: svn://svn.savannah.nongnu.org/qemu/trunk@4985 c046a42c-6fe2-441c-8c8c-71466251a162
Preliminary MacOS X on PPC32 support
Big thanks to BlueSwirl for Sparc failure analysis.
git-svn-id: svn://svn.savannah.nongnu.org/qemu/trunk@4984 c046a42c-6fe2-441c-8c8c-71466251a162
On ppc32 make tb_set_jmp_target1 behave like it does on a ppc64
Avoids nasty warnings about flush_icache_range from gcc4 and inabilityto compile [cpu-]exec.c with gcc3 and -O, also the function is muchtoo large to be candidate for inlining anyway.
git-svn-id: svn://svn.savannah.nongnu.org/qemu/trunk@4974 c046a42c-6fe2-441c-8c8c-71466251a162
Do not try handle "special" arguments of and/or/xor/shl/shr, upper level does it
git-svn-id: svn://svn.savannah.nongnu.org/qemu/trunk@4961 c046a42c-6fe2-441c-8c8c-71466251a162
Provide extNs_M instructions
git-svn-id: svn://svn.savannah.nongnu.org/qemu/trunk@4934 c046a42c-6fe2-441c-8c8c-71466251a162
Fuse EQ and NE handling in tcg_out_brcond2
git-svn-id: svn://svn.savannah.nongnu.org/qemu/trunk@4845 c046a42c-6fe2-441c-8c8c-71466251a162
Mask LL portion of B to 24 bits in tcg_out_b (Thanks to Thiemo Seufer)
git-svn-id: svn://svn.savannah.nongnu.org/qemu/trunk@4841 c046a42c-6fe2-441c-8c8c-71466251a162
According to gcc-4.3.0/gcc/config/rs6000/crtsavres.asm R13 is volatile
git-svn-id: svn://svn.savannah.nongnu.org/qemu/trunk@4779 c046a42c-6fe2-441c-8c8c-71466251a162
Shuffle contents of tcg_target_reg_alloc_order
Move reserved/volatile registers down. Currently qemu_ld/stXX aremarked with TCG_OPF_CALL_CLOBBER and since memory accesses arefrequent and R3 through R12 are volatile moving this down results inless spills and tighter generated code....
Save LR into proper place on callers stack frame
git-svn-id: svn://svn.savannah.nongnu.org/qemu/trunk@4745 c046a42c-6fe2-441c-8c8c-71466251a162
View revisions
Also available in: Atom