[for-next][PATCH 3/4] x86-32, ftrace: Fix static ftrace when early microcode is enabled

[Steven Rostedt <rostedt@goodmis.org>]

[-- Attachment #1: 0003-x86-32-ftrace-Fix-static-ftrace-when-early-microcode.patch --]
[-- Type: text/plain, Size: 1160 bytes --]

From: "H. Peter Anvin" <hpa@linux.intel.com>

Early microcode loading runs C code before paging is enabled on 32
bits.  Since ftrace puts a hook into every function, that hook needs
to be safe to execute in the pre-paging environment.  This is
currently true for dynamic ftrace but not for static ftrace.

Static ftrace is obsolescent and assumed to not be
performance-critical, so we can simply test that the stack pointer
falls within the valid range of kernel addresses.

Reported-by: Jan Kiszka <jan.kiszka@siemens.com>
Tested-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 arch/x86/kernel/entry_32.S |    3 +++
 1 file changed, 3 insertions(+)

diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S
index 2cfbc3a..f0dcb0c 100644
--- a/arch/x86/kernel/entry_32.S
+++ b/arch/x86/kernel/entry_32.S
@@ -1176,6 +1176,9 @@ ftrace_restore_flags:
 #else /* ! CONFIG_DYNAMIC_FTRACE */
 
 ENTRY(mcount)
+	cmpl $__PAGE_OFFSET, %esp
+	jb ftrace_stub		/* Paging not enabled yet? */
+
 	cmpl $0, function_trace_stop
 	jne  ftrace_stub
 
-- 
1.7.10.4