[libc-commits] [libc] [libc] Add support for MVE to Arm startup code (PR #167338)

[Victor Campos via libc-commits]

https://github.com/vhscampos created https://github.com/llvm/llvm-project/pull/167338

In order to have MVE support, the same bits of the CPACR register that enable the floating-point extension must be set.

>From 69e5bf37d264695aa1e684492f58300252c5fa49 Mon Sep 17 00:00:00 2001
From: Victor Campos <victor.campos at arm.com>
Date: Mon, 10 Nov 2025 16:33:19 +0000
Subject: [PATCH] [libc] Add support for MVE to Arm startup code

In order to have MVE support, the same bits of the CPACR register that
enable the floating-point extension must be set.
---
 libc/startup/baremetal/arm/start.cpp | 16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)

diff --git a/libc/startup/baremetal/arm/start.cpp b/libc/startup/baremetal/arm/start.cpp
index 4740067722022..b22529f214216 100644
--- a/libc/startup/baremetal/arm/start.cpp
+++ b/libc/startup/baremetal/arm/start.cpp
@@ -131,20 +131,23 @@ namespace LIBC_NAMESPACE_DECL {
   __arm_wsr("CPSR_c", 0x13); // SVC
 #endif
 
-#ifdef __ARM_FP
-// Enable FPU
-#if __ARM_ARCH_PROFILE == 'M'
+#if __ARM_ARCH_PROFILE == 'M' &&                                               \
+    (defined(__ARM_FP) || defined(__ARM_FEATURE_MVE))
+  // Enable FPU and MVE. They can't be enabled independently: the two are
+  // governed by the same bits in CPACR.
   // Based on
   // https://developer.arm.com/documentation/dui0646/c/Cortex-M7-Peripherals/Floating-Point-Unit/Enabling-the-FPU
-  // Set CPACR cp10 and cp11
+  // Set CPACR cp10 and cp11.
   auto cpacr = (volatile uint32_t *const)0xE000ED88;
   *cpacr |= (0xF << 20);
   __dsb(0xF);
   __isb(0xF);
-#elif __ARM_ARCH_PROFILE == 'A' || __ARM_ARCH_PROFILE == 'R'
+#elif (__ARM_ARCH_PROFILE == 'A' || __ARM_ARCH_PROFILE == 'R') &&              \
+    defined(__ARM_FP)
+  // Enable FPU.
   // Based on
   // https://developer.arm.com/documentation/dui0472/m/Compiler-Coding-Practices/Enabling-NEON-and-FPU-for-bare-metal
-  // Set CPACR cp10 and cp11
+  // Set CPACR cp10 and cp11.
   uint32_t cpacr = __arm_rsr("p15:0:c1:c0:2");
   cpacr |= (0xF << 20);
   __arm_wsr("p15:0:c1:c0:2", cpacr);
@@ -154,7 +157,6 @@ namespace LIBC_NAMESPACE_DECL {
   __asm__ __volatile__("vmrs %0, FPEXC" : "=r"(fpexc) : :);
   fpexc |= (1 << 30);
   __asm__ __volatile__("vmsr FPEXC, %0" : : "r"(fpexc) :);
-#endif
 #endif
 
   // Perform the equivalent of scatterloading