[RISCV] Fix double counting CSRs with Zcmp in RISCVFrameLowering::getFrameIndexReference. (#117207)

The Zcmp callee saved registers are already accounted for in
getCalleeSavedStackSize(). Subtracting RVPushStackSize subtracts
them a second time leading to incorrect stack offsets during frame
index elimination.
    
This should have been removed in
0de2b26
when Zcmp handling was changed. Prior to that, RVPushStackSize was
not included in getCalleeSavedStackSize(). The commit message at the
time noted that Zcmp+RVV was likely broken.

Author: topperc

llvm/lib/Target/RISCV/RISCVFrameLowering.cpp

@@ -1205,7 +1205,6 @@ RISCVFrameLowering::getFrameIndexReference(const MachineFunction &MF, int FI,
     // alignment padding.
     int ScalarLocalVarSize = MFI.getStackSize() -
                              RVFI->getCalleeSavedStackSize() -
-                             RVFI->getRVPushStackSize() -
                              RVFI->getVarArgsSaveSize() + RVFI->getRVVPadding();
     Offset += StackOffset::get(ScalarLocalVarSize, RVFI->getRVVStackSize());
   }

llvm/test/CodeGen/RISCV/rvv/rv32-spill-vector-csr.ll

@@ -3,6 +3,8 @@
 ; RUN:    | FileCheck --check-prefix=SPILL-O0 %s
 ; RUN: llc -mtriple=riscv32 -mattr=+v,+d -O2 < %s \
 ; RUN:    | FileCheck --check-prefix=SPILL-O2 %s
+; RUN: llc -mtriple=riscv32 -mattr=+v,+d,+zcmp -O2 < %s \
+; RUN:    | FileCheck --check-prefix=SPILL-O2-ZCMP %s
 
 @.str = private unnamed_addr constant [6 x i8] c"hello\00", align 1
 
@@ -82,6 +84,37 @@ define <vscale x 1 x double> @foo(<vscale x 1 x double> %a, <vscale x 1 x double
 ; SPILL-O2-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
 ; SPILL-O2-NEXT:    addi sp, sp, 32
 ; SPILL-O2-NEXT:    ret
+;
+; SPILL-O2-ZCMP-LABEL: foo:
+; SPILL-O2-ZCMP:       # %bb.0:
+; SPILL-O2-ZCMP-NEXT:    cm.push {ra, s0}, -32
+; SPILL-O2-ZCMP-NEXT:    csrr a1, vlenb
+; SPILL-O2-ZCMP-NEXT:    slli a1, a1, 1
+; SPILL-O2-ZCMP-NEXT:    sub sp, sp, a1
+; SPILL-O2-ZCMP-NEXT:    mv s0, a0
+; SPILL-O2-ZCMP-NEXT:    addi a1, sp, 16
+; SPILL-O2-ZCMP-NEXT:    vs1r.v v8, (a1) # Unknown-size Folded Spill
+; SPILL-O2-ZCMP-NEXT:    vsetvli zero, a0, e64, m1, ta, ma
+; SPILL-O2-ZCMP-NEXT:    vfadd.vv v9, v8, v9
+; SPILL-O2-ZCMP-NEXT:    csrr a0, vlenb
+; SPILL-O2-ZCMP-NEXT:    add a0, a0, sp
+; SPILL-O2-ZCMP-NEXT:    addi a0, a0, 16
+; SPILL-O2-ZCMP-NEXT:    vs1r.v v9, (a0) # Unknown-size Folded Spill
+; SPILL-O2-ZCMP-NEXT:    lui a0, %hi(.L.str)
+; SPILL-O2-ZCMP-NEXT:    addi a0, a0, %lo(.L.str)
+; SPILL-O2-ZCMP-NEXT:    call puts
+; SPILL-O2-ZCMP-NEXT:    csrr a0, vlenb
+; SPILL-O2-ZCMP-NEXT:    add a0, a0, sp
+; SPILL-O2-ZCMP-NEXT:    addi a0, a0, 16
+; SPILL-O2-ZCMP-NEXT:    vl1r.v v8, (a0) # Unknown-size Folded Reload
+; SPILL-O2-ZCMP-NEXT:    addi a0, sp, 16
+; SPILL-O2-ZCMP-NEXT:    vl1r.v v9, (a0) # Unknown-size Folded Reload
+; SPILL-O2-ZCMP-NEXT:    vsetvli zero, s0, e64, m1, ta, ma
+; SPILL-O2-ZCMP-NEXT:    vfadd.vv v8, v9, v8
+; SPILL-O2-ZCMP-NEXT:    csrr a0, vlenb
+; SPILL-O2-ZCMP-NEXT:    slli a0, a0, 1
+; SPILL-O2-ZCMP-NEXT:    add sp, sp, a0
+; SPILL-O2-ZCMP-NEXT:    cm.popret {ra, s0}, 32
 {
    %x = call <vscale x 1 x double> @llvm.riscv.vfadd.nxv1f64.nxv1f64(<vscale x 1 x double> undef, <vscale x 1 x double> %a, <vscale x 1 x double> %b, i32 7, i32 %gvl)
    %call = call signext i32 @puts(ptr @.str)

llvm/test/CodeGen/RISCV/rvv/rv64-spill-vector-csr.ll

@@ -5,6 +5,8 @@
 ; RUN:    | FileCheck --check-prefix=SPILL-O2 %s
 ; RUN: llc -mtriple=riscv64 -mattr=+v,+d -mattr=+d -riscv-v-vector-bits-max=128 -O2 < %s \
 ; RUN:    | FileCheck --check-prefix=SPILL-O2-VLEN128 %s
+; RUN: llc -mtriple=riscv64 -mattr=+v,+d,+zcmp -O2 < %s \
+; RUN:    | FileCheck --check-prefix=SPILL-O2-ZCMP %s
 
 
 @.str = private unnamed_addr constant [6 x i8] c"hello\00", align 1
@@ -113,6 +115,37 @@ define <vscale x 1 x double> @foo(<vscale x 1 x double> %a, <vscale x 1 x double
 ; SPILL-O2-VLEN128-NEXT:    ld s0, 16(sp) # 8-byte Folded Reload
 ; SPILL-O2-VLEN128-NEXT:    addi sp, sp, 32
 ; SPILL-O2-VLEN128-NEXT:    ret
+;
+; SPILL-O2-ZCMP-LABEL: foo:
+; SPILL-O2-ZCMP:       # %bb.0:
+; SPILL-O2-ZCMP-NEXT:    cm.push {ra, s0}, -32
+; SPILL-O2-ZCMP-NEXT:    csrr a1, vlenb
+; SPILL-O2-ZCMP-NEXT:    slli a1, a1, 1
+; SPILL-O2-ZCMP-NEXT:    sub sp, sp, a1
+; SPILL-O2-ZCMP-NEXT:    mv s0, a0
+; SPILL-O2-ZCMP-NEXT:    addi a1, sp, 16
+; SPILL-O2-ZCMP-NEXT:    vs1r.v v8, (a1) # Unknown-size Folded Spill
+; SPILL-O2-ZCMP-NEXT:    vsetvli zero, a0, e64, m1, ta, ma
+; SPILL-O2-ZCMP-NEXT:    vfadd.vv v9, v8, v9
+; SPILL-O2-ZCMP-NEXT:    csrr a0, vlenb
+; SPILL-O2-ZCMP-NEXT:    add a0, a0, sp
+; SPILL-O2-ZCMP-NEXT:    addi a0, a0, 16
+; SPILL-O2-ZCMP-NEXT:    vs1r.v v9, (a0) # Unknown-size Folded Spill
+; SPILL-O2-ZCMP-NEXT:    lui a0, %hi(.L.str)
+; SPILL-O2-ZCMP-NEXT:    addi a0, a0, %lo(.L.str)
+; SPILL-O2-ZCMP-NEXT:    call puts
+; SPILL-O2-ZCMP-NEXT:    csrr a0, vlenb
+; SPILL-O2-ZCMP-NEXT:    add a0, a0, sp
+; SPILL-O2-ZCMP-NEXT:    addi a0, a0, 16
+; SPILL-O2-ZCMP-NEXT:    vl1r.v v8, (a0) # Unknown-size Folded Reload
+; SPILL-O2-ZCMP-NEXT:    addi a0, sp, 16
+; SPILL-O2-ZCMP-NEXT:    vl1r.v v9, (a0) # Unknown-size Folded Reload
+; SPILL-O2-ZCMP-NEXT:    vsetvli zero, s0, e64, m1, ta, ma
+; SPILL-O2-ZCMP-NEXT:    vfadd.vv v8, v9, v8
+; SPILL-O2-ZCMP-NEXT:    csrr a0, vlenb
+; SPILL-O2-ZCMP-NEXT:    slli a0, a0, 1
+; SPILL-O2-ZCMP-NEXT:    add sp, sp, a0
+; SPILL-O2-ZCMP-NEXT:    cm.popret {ra, s0}, 32
 {
    %x = call <vscale x 1 x double> @llvm.riscv.vfadd.nxv1f64.nxv1f64(<vscale x 1 x double> undef, <vscale x 1 x double> %a, <vscale x 1 x double> %b, i64 7, i64 %gvl)
    %call = call signext i32 @puts(ptr @.str)