diff options
Diffstat (limited to 'libffi/src/powerpc/linux64.S')
| -rw-r--r-- | libffi/src/powerpc/linux64.S | 111 |
1 files changed, 83 insertions, 28 deletions
diff --git a/libffi/src/powerpc/linux64.S b/libffi/src/powerpc/linux64.S index f0006fe..e92d64a 100644 --- a/libffi/src/powerpc/linux64.S +++ b/libffi/src/powerpc/linux64.S @@ -109,40 +109,70 @@ ffi_call_LINUX64: ld %r2, 8(%r29) # endif /* Now do the call. */ - /* Set up cr1 with bits 4-7 of the flags. */ - mtcrf 0x40, %r31 + /* Set up cr1 with bits 3-7 of the flags. */ + mtcrf 0xc0, %r31 /* Get the address to call into CTR. */ mtctr %r12 /* Load all those argument registers. */ - ld %r3, -32-(8*8)(%r28) - ld %r4, -32-(7*8)(%r28) - ld %r5, -32-(6*8)(%r28) - ld %r6, -32-(5*8)(%r28) + addi %r29, %r28, -32-(8*8) + ld %r3, (0*8)(%r29) + ld %r4, (1*8)(%r29) + ld %r5, (2*8)(%r29) + ld %r6, (3*8)(%r29) bf- 5, 1f - ld %r7, -32-(4*8)(%r28) - ld %r8, -32-(3*8)(%r28) - ld %r9, -32-(2*8)(%r28) - ld %r10, -32-(1*8)(%r28) + ld %r7, (4*8)(%r29) + ld %r8, (5*8)(%r29) + ld %r9, (6*8)(%r29) + ld %r10, (7*8)(%r29) 1: /* Load all the FP registers. */ bf- 6, 2f - lfd %f1, -32-(21*8)(%r28) - lfd %f2, -32-(20*8)(%r28) - lfd %f3, -32-(19*8)(%r28) - lfd %f4, -32-(18*8)(%r28) - lfd %f5, -32-(17*8)(%r28) - lfd %f6, -32-(16*8)(%r28) - lfd %f7, -32-(15*8)(%r28) - lfd %f8, -32-(14*8)(%r28) - lfd %f9, -32-(13*8)(%r28) - lfd %f10, -32-(12*8)(%r28) - lfd %f11, -32-(11*8)(%r28) - lfd %f12, -32-(10*8)(%r28) - lfd %f13, -32-(9*8)(%r28) + addi %r29, %r29, -(14*8) + lfd %f1, ( 1*8)(%r29) + lfd %f2, ( 2*8)(%r29) + lfd %f3, ( 3*8)(%r29) + lfd %f4, ( 4*8)(%r29) + lfd %f5, ( 5*8)(%r29) + lfd %f6, ( 6*8)(%r29) + lfd %f7, ( 7*8)(%r29) + lfd %f8, ( 8*8)(%r29) + lfd %f9, ( 9*8)(%r29) + lfd %f10, (10*8)(%r29) + lfd %f11, (11*8)(%r29) + lfd %f12, (12*8)(%r29) + lfd %f13, (13*8)(%r29) 2: + /* Load all the vector registers. */ + bf- 3, 3f + addi %r29, %r29, -16 + lvx %v13, 0, %r29 + addi %r29, %r29, -16 + lvx %v12, 0, %r29 + addi %r29, %r29, -16 + lvx %v11, 0, %r29 + addi %r29, %r29, -16 + lvx %v10, 0, %r29 + addi %r29, %r29, -16 + lvx %v9, 0, %r29 + addi %r29, %r29, -16 + lvx %v8, 0, %r29 + addi %r29, %r29, -16 + lvx %v7, 0, %r29 + addi %r29, %r29, -16 + lvx %v6, 0, %r29 + addi %r29, %r29, -16 + lvx %v5, 0, %r29 + addi %r29, %r29, -16 + lvx %v4, 0, %r29 + addi %r29, %r29, -16 + lvx %v3, 0, %r29 + addi %r29, %r29, -16 + lvx %v2, 0, %r29 +3: + /* Make the call. */ ld %r11, 8(%r28) bctrl @@ -160,6 +190,7 @@ ffi_call_LINUX64: bt 31, .Lstruct_return_value bt 30, .Ldone_return_value bt 29, .Lfp_return_value + bt 28, .Lvec_return_value std %r3, 0(%r30) /* Fall through... */ @@ -175,12 +206,16 @@ ffi_call_LINUX64: ld %r31, -8(%r1) blr +.Lvec_return_value: + stvx %v2, 0, %r30 + b .Ldone_return_value + .Lfp_return_value: .cfi_def_cfa_register 28 - bf 28, .Lfloat_return_value - stfd %f1, 0(%r30) mtcrf 0x02, %r31 /* cr6 */ - bf 27, .Ldone_return_value + bf 27, .Lfloat_return_value + stfd %f1, 0(%r30) + bf 26, .Ldone_return_value stfd %f2, 8(%r30) b .Ldone_return_value .Lfloat_return_value: @@ -188,8 +223,9 @@ ffi_call_LINUX64: b .Ldone_return_value .Lstruct_return_value: - bf 29, .Lsmall_struct - bf 28, .Lfloat_homog_return_value + bf 29, .Lvec_homog_or_small_struct + mtcrf 0x02, %r31 /* cr6 */ + bf 27, .Lfloat_homog_return_value stfd %f1, 0(%r30) stfd %f2, 8(%r30) stfd %f3, 16(%r30) @@ -211,6 +247,25 @@ ffi_call_LINUX64: stfs %f8, 28(%r30) b .Ldone_return_value +.Lvec_homog_or_small_struct: + bf 28, .Lsmall_struct + stvx %v2, 0, %r30 + addi %r30, %r30, 16 + stvx %v3, 0, %r30 + addi %r30, %r30, 16 + stvx %v4, 0, %r30 + addi %r30, %r30, 16 + stvx %v5, 0, %r30 + addi %r30, %r30, 16 + stvx %v6, 0, %r30 + addi %r30, %r30, 16 + stvx %v7, 0, %r30 + addi %r30, %r30, 16 + stvx %v8, 0, %r30 + addi %r30, %r30, 16 + stvx %v9, 0, %r30 + b .Ldone_return_value + .Lsmall_struct: std %r3, 0(%r30) std %r4, 8(%r30) |