From 3d499ba2dce4ac8716f2c9bb07f66f45d0555a19 Mon Sep 17 00:00:00 2001
From: The-going <48602507+The-going@users.noreply.github.com>
Date: Fri, 12 Jul 2024 14:20:15 +0300
Subject: arm: patch: call 'patch_unmap' after flushing dcache and icache

Problem: Linux kernel hangs in early boot on 32-bit ARM platform,
when ftrace 4-byte "mcount" function call location for
"_raw_spin_unlock_irqrestore" function straddles icache lines.

1. The ftrace location for "_raw_spin_unlock_irqrestore" is NOT
   4-byte aligned and 4 bytes at this location straddle the instruction
   cache line (0x20) boundaries.

The problem is present for (cross-compiler) GCC 10, 11, 12.
It does not happen when the kernel is compiled with GCC 9,
even when condition (1) is satisfied.

Detailed description:
https://forum.armbian.com/topic/41339-linux-image-legacy-sunxi2451-kernel-6192-is-broken-stuck-at-starting-kernel/?do=findComment&comment=196528

Author: @mikhailai
---
 arch/arm/kernel/patch.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/arch/arm/kernel/patch.c b/arch/arm/kernel/patch.c
index e9e828b6bb30..ce0fd3aeb575 100644
--- a/arch/arm/kernel/patch.c
+++ b/arch/arm/kernel/patch.c
@@ -101,11 +101,18 @@ void __kprobes __patch_text_real(void *addr, unsigned int insn, bool remap)
 
 	if (waddr != addr) {
 		flush_kernel_vmap_range(waddr, twopage ? size / 2 : size);
-		patch_unmap(FIX_TEXT_POKE0, &flags);
 	}
 
 	flush_icache_range((uintptr_t)(addr),
 			   (uintptr_t)(addr) + size);
+
+	/* Can only call 'patch_unmap' after flushing dcache and icache,
+	 * because it calls 'raw_spin_unlock_irqrestore', but that may
+	 * happen to be the very function we're currently patching
+	 * (as it happens during the ftrace init).
+	 */
+	if (waddr != addr)
+		patch_unmap(FIX_TEXT_POKE0, &flags);
 }
 
 static int __kprobes patch_text_stop_machine(void *data)
-- 
2.35.3