[compiler-rt] [libunwind] [AArch64] Fix nofp regressions in compiler-rt and libunwind (PR #111235)

[Keith Packard via cfe-commits]

https://github.com/keith-packard updated https://github.com/llvm/llvm-project/pull/111235

>From 08ed820b2bc6235998e47c49eb3029b7945b5fec Mon Sep 17 00:00:00 2001
From: Keith Packard <keithp at keithp.com>
Date: Fri, 4 Oct 2024 21:06:37 -0700
Subject: [PATCH 1/2] [libunwind] Support aarch64 without FPU

Skip save/restore of FPU registers on targets without them.

Signed-off-by: Keith Packard <keithp at keithp.com>
---
 libunwind/src/UnwindRegistersRestore.S | 4 ++--
 libunwind/src/UnwindRegistersSave.S    | 2 ++
 2 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/libunwind/src/UnwindRegistersRestore.S b/libunwind/src/UnwindRegistersRestore.S
index 180a66582f41b5..1702d016c368ba 100644
--- a/libunwind/src/UnwindRegistersRestore.S
+++ b/libunwind/src/UnwindRegistersRestore.S
@@ -658,7 +658,7 @@ DEFINE_LIBUNWIND_FUNCTION(__libunwind_Registers_arm64_jumpto)
   ldp    x26,x27, [x0, #0x0D0]
   ldp    x28,x29, [x0, #0x0E0]
   ldr    x30,     [x0, #0x100]  // restore pc into lr
-
+#if defined(__ARM_FP) && __ARM_FP != 0
   ldp    d0, d1,  [x0, #0x110]
   ldp    d2, d3,  [x0, #0x120]
   ldp    d4, d5,  [x0, #0x130]
@@ -676,7 +676,7 @@ DEFINE_LIBUNWIND_FUNCTION(__libunwind_Registers_arm64_jumpto)
   ldp    d28,d29, [x0, #0x1F0]
   ldr    d30,     [x0, #0x200]
   ldr    d31,     [x0, #0x208]
-
+#endif
   // Finally, restore sp. This must be done after the last read from the
   // context struct, because it is allocated on the stack, and an exception
   // could clobber the de-allocated portion of the stack after sp has been
diff --git a/libunwind/src/UnwindRegistersSave.S b/libunwind/src/UnwindRegistersSave.S
index fab234fcd6f318..a489a8ba6df159 100644
--- a/libunwind/src/UnwindRegistersSave.S
+++ b/libunwind/src/UnwindRegistersSave.S
@@ -746,6 +746,7 @@ DEFINE_LIBUNWIND_FUNCTION(__unw_getcontext)
   str    x1,      [x0, #0x0F8]
   str    x30,     [x0, #0x100]    // store return address as pc
   // skip cpsr
+#if defined(__ARM_FP) && __ARM_FP != 0
   stp    d0, d1,  [x0, #0x110]
   stp    d2, d3,  [x0, #0x120]
   stp    d4, d5,  [x0, #0x130]
@@ -763,6 +764,7 @@ DEFINE_LIBUNWIND_FUNCTION(__unw_getcontext)
   stp    d28,d29, [x0, #0x1F0]
   str    d30,     [x0, #0x200]
   str    d31,     [x0, #0x208]
+#endif
   mov    x0, #0                   // return UNW_ESUCCESS
   ret
 

>From 474ae35242eb23f37689a1d933d018c74f2597d5 Mon Sep 17 00:00:00 2001
From: Keith Packard <keithp at keithp.com>
Date: Fri, 4 Oct 2024 21:08:17 -0700
Subject: [PATCH 2/2] [compiler-rt] Support aarch64 targets without FPU

Fall back to the old C implementation of __arm_sc_memset when
the target doesn't have an FPU.

Signed-off-by: Keith Packard <keithp at keithp.com>
---
 .../lib/builtins/aarch64/sme-libc-mem-routines.S     |  5 ++---
 compiler-rt/lib/builtins/aarch64/sme-libc-routines.c | 12 ++++++++++++
 2 files changed, 14 insertions(+), 3 deletions(-)

diff --git a/compiler-rt/lib/builtins/aarch64/sme-libc-mem-routines.S b/compiler-rt/lib/builtins/aarch64/sme-libc-mem-routines.S
index 0318d9a6f1ebd2..6e13a03691cfd6 100644
--- a/compiler-rt/lib/builtins/aarch64/sme-libc-mem-routines.S
+++ b/compiler-rt/lib/builtins/aarch64/sme-libc-mem-routines.S
@@ -6,8 +6,6 @@
 
 #include "../assembly.h"
 
-#ifdef __aarch64__
-
 #define L(l) .L ## l
 
 //
@@ -238,7 +236,8 @@ END_COMPILERRT_OUTLINE_FUNCTION(__arm_sc_memcpy)
 
 DEFINE_COMPILERRT_FUNCTION_ALIAS(__arm_sc_memmove, __arm_sc_memcpy)
 
-
+// This version uses FP registers. Use this only on targets with them
+#if defined(__aarch64__) && __ARM_FP != 0
 //
 //  __arm_sc_memset
 //
diff --git a/compiler-rt/lib/builtins/aarch64/sme-libc-routines.c b/compiler-rt/lib/builtins/aarch64/sme-libc-routines.c
index 315490e73ea2b1..07d6681485556b 100644
--- a/compiler-rt/lib/builtins/aarch64/sme-libc-routines.c
+++ b/compiler-rt/lib/builtins/aarch64/sme-libc-routines.c
@@ -1,5 +1,17 @@
 #include <stddef.h>
 
+/* The asm version uses FP registers. Use this on targets without them */
+#if __ARM_FP == 0
+void *__arm_sc_memset(void *dest, int c, size_t n) __arm_streaming_compatible {
+  unsigned char *destp = (unsigned char *)dest;
+  unsigned char c8 = (unsigned char)c;
+  for (size_t i = 0; i < n; ++i)
+    destp[i] = c8;
+
+  return dest;
+}
+#endif
+
 const void *__arm_sc_memchr(const void *src, int c,
                             size_t n) __arm_streaming_compatible {
   const unsigned char *srcp = (const unsigned char *)src;