maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/MC/SystemZ/insn-good-z196.s

285 lines
11 KiB
ArmAsm
Raw Normal View History

[SystemZ] Start adding z196 and zEC12 support This first step just adds definitions for SLLK, SRLK and SRAK. The next patch will actually make use of them during codegen. insn-bad.s tests that some form of error is reported when using these instructions on z10. More work is needed to get the "instruction requires: distinct-ops" that we'd ideally like, so I've stubbed that part out for now. I'll come back and make it mandatory once the necessary changes are in. llvm-svn: 186680
2013-07-20 00:09:03 +08:00
# For z196 and above.
# RUN: llvm-mc -triple s390x-linux-gnu -mcpu=z196 -show-encoding %s | FileCheck %s
[SystemZ] Add AHIK and AGHIK I did these as a separate patch because it uses a slightly different form of RIE layout. llvm-svn: 186687
2013-07-20 00:32:12 +08:00
#CHECK: aghik %r0, %r0, -32768 # encoding: [0xec,0x00,0x80,0x00,0x00,0xd9]
#CHECK: aghik %r0, %r0, -1 # encoding: [0xec,0x00,0xff,0xff,0x00,0xd9]
#CHECK: aghik %r0, %r0, 0 # encoding: [0xec,0x00,0x00,0x00,0x00,0xd9]
#CHECK: aghik %r0, %r0, 1 # encoding: [0xec,0x00,0x00,0x01,0x00,0xd9]
#CHECK: aghik %r0, %r0, 32767 # encoding: [0xec,0x00,0x7f,0xff,0x00,0xd9]
#CHECK: aghik %r0, %r15, 0 # encoding: [0xec,0x0f,0x00,0x00,0x00,0xd9]
#CHECK: aghik %r15, %r0, 0 # encoding: [0xec,0xf0,0x00,0x00,0x00,0xd9]
#CHECK: aghik %r7, %r8, -16 # encoding: [0xec,0x78,0xff,0xf0,0x00,0xd9]
aghik %r0, %r0, -32768
aghik %r0, %r0, -1
aghik %r0, %r0, 0
aghik %r0, %r0, 1
aghik %r0, %r0, 32767
aghik %r0, %r15, 0
aghik %r15, %r0, 0
aghik %r7, %r8, -16
[SystemZ] Add ARK, AGRK, SRK and SGRK The testsuite changes follow the same lines as for r186683. llvm-svn: 186686
2013-07-20 00:26:39 +08:00
#CHECK: agrk %r0, %r0, %r0 # encoding: [0xb9,0xe8,0x00,0x00]
#CHECK: agrk %r0, %r0, %r15 # encoding: [0xb9,0xe8,0xf0,0x00]
#CHECK: agrk %r0, %r15, %r0 # encoding: [0xb9,0xe8,0x00,0x0f]
#CHECK: agrk %r15, %r0, %r0 # encoding: [0xb9,0xe8,0x00,0xf0]
#CHECK: agrk %r7, %r8, %r9 # encoding: [0xb9,0xe8,0x90,0x78]
agrk %r0,%r0,%r0
agrk %r0,%r0,%r15
agrk %r0,%r15,%r0
agrk %r15,%r0,%r0
agrk %r7,%r8,%r9
[SystemZ] Add AHIK and AGHIK I did these as a separate patch because it uses a slightly different form of RIE layout. llvm-svn: 186687
2013-07-20 00:32:12 +08:00
#CHECK: ahik %r0, %r0, -32768 # encoding: [0xec,0x00,0x80,0x00,0x00,0xd8]
#CHECK: ahik %r0, %r0, -1 # encoding: [0xec,0x00,0xff,0xff,0x00,0xd8]
#CHECK: ahik %r0, %r0, 0 # encoding: [0xec,0x00,0x00,0x00,0x00,0xd8]
#CHECK: ahik %r0, %r0, 1 # encoding: [0xec,0x00,0x00,0x01,0x00,0xd8]
#CHECK: ahik %r0, %r0, 32767 # encoding: [0xec,0x00,0x7f,0xff,0x00,0xd8]
#CHECK: ahik %r0, %r15, 0 # encoding: [0xec,0x0f,0x00,0x00,0x00,0xd8]
#CHECK: ahik %r15, %r0, 0 # encoding: [0xec,0xf0,0x00,0x00,0x00,0xd8]
#CHECK: ahik %r7, %r8, -16 # encoding: [0xec,0x78,0xff,0xf0,0x00,0xd8]
ahik %r0, %r0, -32768
ahik %r0, %r0, -1
ahik %r0, %r0, 0
ahik %r0, %r0, 1
ahik %r0, %r0, 32767
ahik %r0, %r15, 0
ahik %r15, %r0, 0
ahik %r7, %r8, -16
[SystemZ] Add ALRK, AGLRK, SLRK and SGLRK Follows the same lines as r186686, but much more limited, since we only use ADD LOGICAL for multi-i64 additions. llvm-svn: 186689
2013-07-20 00:37:00 +08:00
#CHECK: algrk %r0, %r0, %r0 # encoding: [0xb9,0xea,0x00,0x00]
#CHECK: algrk %r0, %r0, %r15 # encoding: [0xb9,0xea,0xf0,0x00]
#CHECK: algrk %r0, %r15, %r0 # encoding: [0xb9,0xea,0x00,0x0f]
#CHECK: algrk %r15, %r0, %r0 # encoding: [0xb9,0xea,0x00,0xf0]
#CHECK: algrk %r7, %r8, %r9 # encoding: [0xb9,0xea,0x90,0x78]
algrk %r0,%r0,%r0
algrk %r0,%r0,%r15
algrk %r0,%r15,%r0
algrk %r15,%r0,%r0
algrk %r7,%r8,%r9
#CHECK: alrk %r0, %r0, %r0 # encoding: [0xb9,0xfa,0x00,0x00]
#CHECK: alrk %r0, %r0, %r15 # encoding: [0xb9,0xfa,0xf0,0x00]
#CHECK: alrk %r0, %r15, %r0 # encoding: [0xb9,0xfa,0x00,0x0f]
#CHECK: alrk %r15, %r0, %r0 # encoding: [0xb9,0xfa,0x00,0xf0]
#CHECK: alrk %r7, %r8, %r9 # encoding: [0xb9,0xfa,0x90,0x78]
alrk %r0,%r0,%r0
alrk %r0,%r0,%r15
alrk %r0,%r15,%r0
alrk %r15,%r0,%r0
alrk %r7,%r8,%r9
[SystemZ] Add ARK, AGRK, SRK and SGRK The testsuite changes follow the same lines as for r186683. llvm-svn: 186686
2013-07-20 00:26:39 +08:00
#CHECK: ark %r0, %r0, %r0 # encoding: [0xb9,0xf8,0x00,0x00]
#CHECK: ark %r0, %r0, %r15 # encoding: [0xb9,0xf8,0xf0,0x00]
#CHECK: ark %r0, %r15, %r0 # encoding: [0xb9,0xf8,0x00,0x0f]
#CHECK: ark %r15, %r0, %r0 # encoding: [0xb9,0xf8,0x00,0xf0]
#CHECK: ark %r7, %r8, %r9 # encoding: [0xb9,0xf8,0x90,0x78]
ark %r0,%r0,%r0
ark %r0,%r0,%r15
ark %r0,%r15,%r0
ark %r15,%r0,%r0
ark %r7,%r8,%r9
[SystemZ] Add NGRK, OGRK and XGRK Like r186683, but for 64 bits. llvm-svn: 186685
2013-07-20 00:24:22 +08:00
#CHECK: ngrk %r0, %r0, %r0 # encoding: [0xb9,0xe4,0x00,0x00]
#CHECK: ngrk %r0, %r0, %r15 # encoding: [0xb9,0xe4,0xf0,0x00]
#CHECK: ngrk %r0, %r15, %r0 # encoding: [0xb9,0xe4,0x00,0x0f]
#CHECK: ngrk %r15, %r0, %r0 # encoding: [0xb9,0xe4,0x00,0xf0]
#CHECK: ngrk %r7, %r8, %r9 # encoding: [0xb9,0xe4,0x90,0x78]
ngrk %r0,%r0,%r0
ngrk %r0,%r0,%r15
ngrk %r0,%r15,%r0
ngrk %r15,%r0,%r0
ngrk %r7,%r8,%r9
[SystemZ] Add NRK, ORK and XRK The atomic tests assume the two-operand forms, so I've restricted them to z10. Running and-01.ll, or-01.ll and xor-01.ll for z196 as well as z10 shows why using convertToThreeAddress() is better than exposing the three-operand forms first and then converting back to two operands where possible (which is what I'd originally tried). Using the three-operand form first stops us from taking advantage of NG, OG and XG for spills. llvm-svn: 186683
2013-07-20 00:21:55 +08:00
#CHECK: nrk %r0, %r0, %r0 # encoding: [0xb9,0xf4,0x00,0x00]
#CHECK: nrk %r0, %r0, %r15 # encoding: [0xb9,0xf4,0xf0,0x00]
#CHECK: nrk %r0, %r15, %r0 # encoding: [0xb9,0xf4,0x00,0x0f]
#CHECK: nrk %r15, %r0, %r0 # encoding: [0xb9,0xf4,0x00,0xf0]
#CHECK: nrk %r7, %r8, %r9 # encoding: [0xb9,0xf4,0x90,0x78]
nrk %r0,%r0,%r0
nrk %r0,%r0,%r15
nrk %r0,%r15,%r0
nrk %r15,%r0,%r0
nrk %r7,%r8,%r9
[SystemZ] Add NGRK, OGRK and XGRK Like r186683, but for 64 bits. llvm-svn: 186685
2013-07-20 00:24:22 +08:00
#CHECK: ogrk %r0, %r0, %r0 # encoding: [0xb9,0xe6,0x00,0x00]
#CHECK: ogrk %r0, %r0, %r15 # encoding: [0xb9,0xe6,0xf0,0x00]
#CHECK: ogrk %r0, %r15, %r0 # encoding: [0xb9,0xe6,0x00,0x0f]
#CHECK: ogrk %r15, %r0, %r0 # encoding: [0xb9,0xe6,0x00,0xf0]
#CHECK: ogrk %r7, %r8, %r9 # encoding: [0xb9,0xe6,0x90,0x78]
ogrk %r0,%r0,%r0
ogrk %r0,%r0,%r15
ogrk %r0,%r15,%r0
ogrk %r15,%r0,%r0
ogrk %r7,%r8,%r9
[SystemZ] Add NRK, ORK and XRK The atomic tests assume the two-operand forms, so I've restricted them to z10. Running and-01.ll, or-01.ll and xor-01.ll for z196 as well as z10 shows why using convertToThreeAddress() is better than exposing the three-operand forms first and then converting back to two operands where possible (which is what I'd originally tried). Using the three-operand form first stops us from taking advantage of NG, OG and XG for spills. llvm-svn: 186683
2013-07-20 00:21:55 +08:00
#CHECK: ork %r0, %r0, %r0 # encoding: [0xb9,0xf6,0x00,0x00]
#CHECK: ork %r0, %r0, %r15 # encoding: [0xb9,0xf6,0xf0,0x00]
#CHECK: ork %r0, %r15, %r0 # encoding: [0xb9,0xf6,0x00,0x0f]
#CHECK: ork %r15, %r0, %r0 # encoding: [0xb9,0xf6,0x00,0xf0]
#CHECK: ork %r7, %r8, %r9 # encoding: [0xb9,0xf6,0x90,0x78]
ork %r0,%r0,%r0
ork %r0,%r0,%r15
ork %r0,%r15,%r0
ork %r15,%r0,%r0
ork %r7,%r8,%r9
[SystemZ] Add ARK, AGRK, SRK and SGRK The testsuite changes follow the same lines as for r186683. llvm-svn: 186686
2013-07-20 00:26:39 +08:00
#CHECK: sgrk %r0, %r0, %r0 # encoding: [0xb9,0xe9,0x00,0x00]
#CHECK: sgrk %r0, %r0, %r15 # encoding: [0xb9,0xe9,0xf0,0x00]
#CHECK: sgrk %r0, %r15, %r0 # encoding: [0xb9,0xe9,0x00,0x0f]
#CHECK: sgrk %r15, %r0, %r0 # encoding: [0xb9,0xe9,0x00,0xf0]
#CHECK: sgrk %r7, %r8, %r9 # encoding: [0xb9,0xe9,0x90,0x78]
sgrk %r0,%r0,%r0
sgrk %r0,%r0,%r15
sgrk %r0,%r15,%r0
sgrk %r15,%r0,%r0
sgrk %r7,%r8,%r9
[SystemZ] Add ALRK, AGLRK, SLRK and SGLRK Follows the same lines as r186686, but much more limited, since we only use ADD LOGICAL for multi-i64 additions. llvm-svn: 186689
2013-07-20 00:37:00 +08:00
#CHECK: slgrk %r0, %r0, %r0 # encoding: [0xb9,0xeb,0x00,0x00]
#CHECK: slgrk %r0, %r0, %r15 # encoding: [0xb9,0xeb,0xf0,0x00]
#CHECK: slgrk %r0, %r15, %r0 # encoding: [0xb9,0xeb,0x00,0x0f]
#CHECK: slgrk %r15, %r0, %r0 # encoding: [0xb9,0xeb,0x00,0xf0]
#CHECK: slgrk %r7, %r8, %r9 # encoding: [0xb9,0xeb,0x90,0x78]
slgrk %r0,%r0,%r0
slgrk %r0,%r0,%r15
slgrk %r0,%r15,%r0
slgrk %r15,%r0,%r0
slgrk %r7,%r8,%r9
#CHECK: slrk %r0, %r0, %r0 # encoding: [0xb9,0xfb,0x00,0x00]
#CHECK: slrk %r0, %r0, %r15 # encoding: [0xb9,0xfb,0xf0,0x00]
#CHECK: slrk %r0, %r15, %r0 # encoding: [0xb9,0xfb,0x00,0x0f]
#CHECK: slrk %r15, %r0, %r0 # encoding: [0xb9,0xfb,0x00,0xf0]
#CHECK: slrk %r7, %r8, %r9 # encoding: [0xb9,0xfb,0x90,0x78]
slrk %r0,%r0,%r0
slrk %r0,%r0,%r15
slrk %r0,%r15,%r0
slrk %r15,%r0,%r0
slrk %r7,%r8,%r9
[SystemZ] Start adding z196 and zEC12 support This first step just adds definitions for SLLK, SRLK and SRAK. The next patch will actually make use of them during codegen. insn-bad.s tests that some form of error is reported when using these instructions on z10. More work is needed to get the "instruction requires: distinct-ops" that we'd ideally like, so I've stubbed that part out for now. I'll come back and make it mandatory once the necessary changes are in. llvm-svn: 186680
2013-07-20 00:09:03 +08:00
#CHECK: sllk %r0, %r0, 0 # encoding: [0xeb,0x00,0x00,0x00,0x00,0xdf]
#CHECK: sllk %r15, %r1, 0 # encoding: [0xeb,0xf1,0x00,0x00,0x00,0xdf]
#CHECK: sllk %r1, %r15, 0 # encoding: [0xeb,0x1f,0x00,0x00,0x00,0xdf]
#CHECK: sllk %r15, %r15, 0 # encoding: [0xeb,0xff,0x00,0x00,0x00,0xdf]
#CHECK: sllk %r0, %r0, -524288 # encoding: [0xeb,0x00,0x00,0x00,0x80,0xdf]
#CHECK: sllk %r0, %r0, -1 # encoding: [0xeb,0x00,0x0f,0xff,0xff,0xdf]
#CHECK: sllk %r0, %r0, 1 # encoding: [0xeb,0x00,0x00,0x01,0x00,0xdf]
#CHECK: sllk %r0, %r0, 524287 # encoding: [0xeb,0x00,0x0f,0xff,0x7f,0xdf]
#CHECK: sllk %r0, %r0, 0(%r1) # encoding: [0xeb,0x00,0x10,0x00,0x00,0xdf]
#CHECK: sllk %r0, %r0, 0(%r15) # encoding: [0xeb,0x00,0xf0,0x00,0x00,0xdf]
#CHECK: sllk %r0, %r0, 524287(%r1) # encoding: [0xeb,0x00,0x1f,0xff,0x7f,0xdf]
#CHECK: sllk %r0, %r0, 524287(%r15) # encoding: [0xeb,0x00,0xff,0xff,0x7f,0xdf]
sllk %r0,%r0,0
sllk %r15,%r1,0
sllk %r1,%r15,0
sllk %r15,%r15,0
sllk %r0,%r0,-524288
sllk %r0,%r0,-1
sllk %r0,%r0,1
sllk %r0,%r0,524287
sllk %r0,%r0,0(%r1)
sllk %r0,%r0,0(%r15)
sllk %r0,%r0,524287(%r1)
sllk %r0,%r0,524287(%r15)
#CHECK: srak %r0, %r0, 0 # encoding: [0xeb,0x00,0x00,0x00,0x00,0xdc]
#CHECK: srak %r15, %r1, 0 # encoding: [0xeb,0xf1,0x00,0x00,0x00,0xdc]
#CHECK: srak %r1, %r15, 0 # encoding: [0xeb,0x1f,0x00,0x00,0x00,0xdc]
#CHECK: srak %r15, %r15, 0 # encoding: [0xeb,0xff,0x00,0x00,0x00,0xdc]
#CHECK: srak %r0, %r0, -524288 # encoding: [0xeb,0x00,0x00,0x00,0x80,0xdc]
#CHECK: srak %r0, %r0, -1 # encoding: [0xeb,0x00,0x0f,0xff,0xff,0xdc]
#CHECK: srak %r0, %r0, 1 # encoding: [0xeb,0x00,0x00,0x01,0x00,0xdc]
#CHECK: srak %r0, %r0, 524287 # encoding: [0xeb,0x00,0x0f,0xff,0x7f,0xdc]
#CHECK: srak %r0, %r0, 0(%r1) # encoding: [0xeb,0x00,0x10,0x00,0x00,0xdc]
#CHECK: srak %r0, %r0, 0(%r15) # encoding: [0xeb,0x00,0xf0,0x00,0x00,0xdc]
#CHECK: srak %r0, %r0, 524287(%r1) # encoding: [0xeb,0x00,0x1f,0xff,0x7f,0xdc]
#CHECK: srak %r0, %r0, 524287(%r15) # encoding: [0xeb,0x00,0xff,0xff,0x7f,0xdc]
srak %r0,%r0,0
srak %r15,%r1,0
srak %r1,%r15,0
srak %r15,%r15,0
srak %r0,%r0,-524288
srak %r0,%r0,-1
srak %r0,%r0,1
srak %r0,%r0,524287
srak %r0,%r0,0(%r1)
srak %r0,%r0,0(%r15)
srak %r0,%r0,524287(%r1)
srak %r0,%r0,524287(%r15)
[SystemZ] Add ARK, AGRK, SRK and SGRK The testsuite changes follow the same lines as for r186683. llvm-svn: 186686
2013-07-20 00:26:39 +08:00
#CHECK: srk %r0, %r0, %r0 # encoding: [0xb9,0xf9,0x00,0x00]
#CHECK: srk %r0, %r0, %r15 # encoding: [0xb9,0xf9,0xf0,0x00]
#CHECK: srk %r0, %r15, %r0 # encoding: [0xb9,0xf9,0x00,0x0f]
#CHECK: srk %r15, %r0, %r0 # encoding: [0xb9,0xf9,0x00,0xf0]
#CHECK: srk %r7, %r8, %r9 # encoding: [0xb9,0xf9,0x90,0x78]
srk %r0,%r0,%r0
srk %r0,%r0,%r15
srk %r0,%r15,%r0
srk %r15,%r0,%r0
srk %r7,%r8,%r9
[SystemZ] Start adding z196 and zEC12 support This first step just adds definitions for SLLK, SRLK and SRAK. The next patch will actually make use of them during codegen. insn-bad.s tests that some form of error is reported when using these instructions on z10. More work is needed to get the "instruction requires: distinct-ops" that we'd ideally like, so I've stubbed that part out for now. I'll come back and make it mandatory once the necessary changes are in. llvm-svn: 186680
2013-07-20 00:09:03 +08:00
#CHECK: srlk %r0, %r0, 0 # encoding: [0xeb,0x00,0x00,0x00,0x00,0xde]
#CHECK: srlk %r15, %r1, 0 # encoding: [0xeb,0xf1,0x00,0x00,0x00,0xde]
#CHECK: srlk %r1, %r15, 0 # encoding: [0xeb,0x1f,0x00,0x00,0x00,0xde]
#CHECK: srlk %r15, %r15, 0 # encoding: [0xeb,0xff,0x00,0x00,0x00,0xde]
#CHECK: srlk %r0, %r0, -524288 # encoding: [0xeb,0x00,0x00,0x00,0x80,0xde]
#CHECK: srlk %r0, %r0, -1 # encoding: [0xeb,0x00,0x0f,0xff,0xff,0xde]
#CHECK: srlk %r0, %r0, 1 # encoding: [0xeb,0x00,0x00,0x01,0x00,0xde]
#CHECK: srlk %r0, %r0, 524287 # encoding: [0xeb,0x00,0x0f,0xff,0x7f,0xde]
#CHECK: srlk %r0, %r0, 0(%r1) # encoding: [0xeb,0x00,0x10,0x00,0x00,0xde]
#CHECK: srlk %r0, %r0, 0(%r15) # encoding: [0xeb,0x00,0xf0,0x00,0x00,0xde]
#CHECK: srlk %r0, %r0, 524287(%r1) # encoding: [0xeb,0x00,0x1f,0xff,0x7f,0xde]
#CHECK: srlk %r0, %r0, 524287(%r15) # encoding: [0xeb,0x00,0xff,0xff,0x7f,0xde]
srlk %r0,%r0,0
srlk %r15,%r1,0
srlk %r1,%r15,0
srlk %r15,%r15,0
srlk %r0,%r0,-524288
srlk %r0,%r0,-1
srlk %r0,%r0,1
srlk %r0,%r0,524287
srlk %r0,%r0,0(%r1)
srlk %r0,%r0,0(%r15)
srlk %r0,%r0,524287(%r1)
srlk %r0,%r0,524287(%r15)
[SystemZ] Add NRK, ORK and XRK The atomic tests assume the two-operand forms, so I've restricted them to z10. Running and-01.ll, or-01.ll and xor-01.ll for z196 as well as z10 shows why using convertToThreeAddress() is better than exposing the three-operand forms first and then converting back to two operands where possible (which is what I'd originally tried). Using the three-operand form first stops us from taking advantage of NG, OG and XG for spills. llvm-svn: 186683
2013-07-20 00:21:55 +08:00
[SystemZ] Add NGRK, OGRK and XGRK Like r186683, but for 64 bits. llvm-svn: 186685
2013-07-20 00:24:22 +08:00
#CHECK: xgrk %r0, %r0, %r0 # encoding: [0xb9,0xe7,0x00,0x00]
#CHECK: xgrk %r0, %r0, %r15 # encoding: [0xb9,0xe7,0xf0,0x00]
#CHECK: xgrk %r0, %r15, %r0 # encoding: [0xb9,0xe7,0x00,0x0f]
#CHECK: xgrk %r15, %r0, %r0 # encoding: [0xb9,0xe7,0x00,0xf0]
#CHECK: xgrk %r7, %r8, %r9 # encoding: [0xb9,0xe7,0x90,0x78]
xgrk %r0,%r0,%r0
xgrk %r0,%r0,%r15
xgrk %r0,%r15,%r0
xgrk %r15,%r0,%r0
xgrk %r7,%r8,%r9
[SystemZ] Add NRK, ORK and XRK The atomic tests assume the two-operand forms, so I've restricted them to z10. Running and-01.ll, or-01.ll and xor-01.ll for z196 as well as z10 shows why using convertToThreeAddress() is better than exposing the three-operand forms first and then converting back to two operands where possible (which is what I'd originally tried). Using the three-operand form first stops us from taking advantage of NG, OG and XG for spills. llvm-svn: 186683
2013-07-20 00:21:55 +08:00
#CHECK: xrk %r0, %r0, %r0 # encoding: [0xb9,0xf7,0x00,0x00]
#CHECK: xrk %r0, %r0, %r15 # encoding: [0xb9,0xf7,0xf0,0x00]
#CHECK: xrk %r0, %r15, %r0 # encoding: [0xb9,0xf7,0x00,0x0f]
#CHECK: xrk %r15, %r0, %r0 # encoding: [0xb9,0xf7,0x00,0xf0]
#CHECK: xrk %r7, %r8, %r9 # encoding: [0xb9,0xf7,0x90,0x78]
xrk %r0,%r0,%r0
xrk %r0,%r0,%r15
xrk %r0,%r15,%r0
xrk %r15,%r0,%r0
xrk %r7,%r8,%r9