unicorn/qemu
Aurelien Jarno 5f67ab74e7
tcg/optimize: fix constant signedness
By convention, on a 64-bit host TCG internally stores 32-bit constants
as sign-extended. This is not the case in the optimizer when a 32-bit
constant is folded.

This doesn't seem to have more consequences than suboptimal code
generation. For instance the x86 backend assumes sign-extended constants,
and in some rare cases uses a 32-bit unsigned immediate 0xffffffff
instead of a 8-bit signed immediate 0xff for the constant -1. This is
with a ppc guest:

before
------

 ---- 0x9f29cc
 movi_i32 tmp1,$0xffffffff
 movi_i32 tmp2,$0x0
 add2_i32 tmp0,CA,CA,tmp2,r6,tmp2
 add2_i32 tmp0,CA,tmp0,CA,tmp1,tmp2
 mov_i32 r10,tmp0

0x7fd8c7dfe90c:  xor    %ebp,%ebp
0x7fd8c7dfe90e:  mov    %ebp,%r11d
0x7fd8c7dfe911:  mov    0x18(%r14),%r9d
0x7fd8c7dfe915:  add    %r9d,%r10d
0x7fd8c7dfe918:  adc    %ebp,%r11d
0x7fd8c7dfe91b:  add    $0xffffffff,%r10d
0x7fd8c7dfe922:  adc    %ebp,%r11d
0x7fd8c7dfe925:  mov    %r11d,0x134(%r14)
0x7fd8c7dfe92c:  mov    %r10d,0x28(%r14)

after
-----

 ---- 0x9f29cc
 movi_i32 tmp1,$0xffffffffffffffff
 movi_i32 tmp2,$0x0
 add2_i32 tmp0,CA,CA,tmp2,r6,tmp2
 add2_i32 tmp0,CA,tmp0,CA,tmp1,tmp2
 mov_i32 r10,tmp0

0x7f37010d490c:  xor    %ebp,%ebp
0x7f37010d490e:  mov    %ebp,%r11d
0x7f37010d4911:  mov    0x18(%r14),%r9d
0x7f37010d4915:  add    %r9d,%r10d
0x7f37010d4918:  adc    %ebp,%r11d
0x7f37010d491b:  add    $0xffffffffffffffff,%r10d
0x7f37010d491f:  adc    %ebp,%r11d
0x7f37010d4922:  mov    %r11d,0x134(%r14)
0x7f37010d4929:  mov    %r10d,0x28(%r14)

Backports commit 29f3ff8d6cbc28f79933aeaa25805408d0984a8f from qemu
2018-02-10 21:40:20 -05:00
..
default-configs arm64eb: add support for ARM64 big endian. 2017-04-24 23:30:01 +08:00
docs cleanup qemu docs 2017-01-18 15:23:40 +08:00
fpu Automated leading tab to spaces conversion. 2017-01-21 12:28:22 +11:00
hw Arm support ported. (#736) 2017-01-23 23:30:57 +08:00
include tcg: Change translator-side labels to a pointer 2018-02-09 14:17:56 -05:00
qapi This code should now build the x86_x64-softmmu part 2. 2017-01-19 22:50:28 +11:00
qobject This code should now build the x86_x64-softmmu part 2. 2017-01-19 22:50:28 +11:00
qom cleanup after msvc port 2017-01-22 21:27:17 +08:00
scripts Save copies of generated qapi files. 2017-01-21 00:30:50 +11:00
target-arm tcg: Push merged memop+mmu_idx parameter to softmmu routines 2018-02-10 20:03:22 -05:00
target-i386 tcg: Change translator-side labels to a pointer 2018-02-09 14:17:56 -05:00
target-m68k tcg: Change translator-side labels to a pointer 2018-02-09 14:17:56 -05:00
target-mips target-mips: Use CPU_LOG_INT for logging related to interrupts 2018-02-10 21:12:41 -05:00
target-sparc tcg: Change translator-side labels to a pointer 2018-02-09 14:17:56 -05:00
tcg tcg/optimize: fix constant signedness 2018-02-10 21:40:20 -05:00
util Arm support ported. (#736) 2017-01-23 23:30:57 +08:00
aarch64.h tcg: Implement insert_op_before 2018-02-09 13:11:50 -05:00
aarch64eb.h tcg: Implement insert_op_before 2018-02-09 13:11:50 -05:00
accel.c Automated leading tab to spaces conversion. 2017-01-21 12:28:22 +11:00
arm.h tcg: Implement insert_op_before 2018-02-09 13:11:50 -05:00
armeb.h tcg: Implement insert_op_before 2018-02-09 13:11:50 -05:00
CODING_STYLE import 2015-08-21 15:04:50 +08:00
configure tcg: Drop ia64 host support 2018-02-04 18:33:02 -05:00
COPYING import 2015-08-21 15:04:50 +08:00
COPYING.LIB import 2015-08-21 15:04:50 +08:00
cpu-exec.c Only set eip to the instruction pointer after an interrupt if the interrupt was user-generated (#875) 2017-08-29 17:14:36 +07:00
cpus.c cleanup more synchronization code 2017-01-09 14:05:39 +08:00
cputlb.c This code should now build the x86_x64-softmmu part 2. 2017-01-19 22:50:28 +11:00
exec.c fix the last fix that crashes samples 2017-02-24 20:34:52 +08:00
gen_all_header.sh arm64eb: add support for ARM64 big endian. 2017-04-24 23:30:01 +08:00
glib_compat.c Merge branch 'master' into msvc2 2017-04-21 01:17:00 +08:00
HACKING import 2015-08-21 15:04:50 +08:00
header_gen.py tcg: Implement insert_op_before 2018-02-09 13:11:50 -05:00
ioport.c This code should now build the x86_x64-softmmu part 2. 2017-01-19 22:50:28 +11:00
LICENSE import 2015-08-21 15:04:50 +08:00
m68k.h tcg: Implement insert_op_before 2018-02-09 13:11:50 -05:00
Makefile cleanup qemu/default-configs/ 2017-01-19 14:52:30 +08:00
Makefile.objs cleanup qemu/Makefile.objs 2017-01-21 21:50:12 +08:00
Makefile.target tcg: Move some opcode generation functions out of line 2018-02-09 08:10:00 -05:00
memory.c merge msvc with master 2017-02-24 10:39:36 +08:00
memory_mapping.c revert to use of g_free to make future qemu integrations easier (#695) 2016-12-21 22:28:36 +08:00
mips.h tcg: Implement insert_op_before 2018-02-09 13:11:50 -05:00
mips64.h tcg: Implement insert_op_before 2018-02-09 13:11:50 -05:00
mips64el.h tcg: Implement insert_op_before 2018-02-09 13:11:50 -05:00
mipsel.h tcg: Implement insert_op_before 2018-02-09 13:11:50 -05:00
powerpc.h tcg: Implement insert_op_before 2018-02-09 13:11:50 -05:00
qapi-schema.json import 2015-08-21 15:04:50 +08:00
qemu-log.c import 2015-08-21 15:04:50 +08:00
qemu-timer.c timer is redundant 2017-01-20 16:46:58 +08:00
rules.mak import 2015-08-21 15:04:50 +08:00
softmmu_template.h tcg: Add MO_ALIGN, MO_UNALN 2018-02-10 20:18:53 -05:00
sparc.h tcg: Implement insert_op_before 2018-02-09 13:11:50 -05:00
sparc64.h tcg: Implement insert_op_before 2018-02-09 13:11:50 -05:00
tcg-runtime.c platform.h move #3 2017-01-21 00:13:21 +11:00
translate-all.c Fixed warning about {} initialisers. 2017-01-21 11:41:11 +11:00
translate-all.h import 2015-08-21 15:04:50 +08:00
unicorn_common.h This code should now build the x86_x64-softmmu part 2. 2017-01-19 22:50:28 +11:00
VERSION import 2015-08-21 15:04:50 +08:00
vl.c This code should now build the x86_x64-softmmu part 2. 2017-01-19 22:50:28 +11:00
vl.h import 2015-08-21 15:04:50 +08:00
x86_64.h tcg: Implement insert_op_before 2018-02-09 13:11:50 -05:00