x86: fix system die when load with "reservetop" parameter

The system will die if the kernel is booted with "reservetop" parameter, in present code, parse "reservetop" parameter after early_ioremap_init(), and some function still use early_ioremap() after it.

The problem is, "reservetop" parameter can modify 'FIXADDR_TOP', then the virtual address got by early_ioremap() is base on old 'FIXADDR_TOP', but the page mapping is base on new 'FIXADDR_TOP', it will occur page fault, and the IDT is not prepare yet, so, the system is dead.

So, put parse_early_param() in the front of early_ioremap_init() in this patch.

---
 arch/x86/kernel/setup.c |   10 +++++-----
 1 files changed, 5 insertions(+), 5 deletions(-)

index 26fce55..aad6149 100644
@@ -713,6 +713,11 @@ void __init setup_arch(char **cmdline_p)
 	printk(KERN_INFO "Command line: %s\n", boot_command_line);
 #endif
 
+	strlcpy(command_line, boot_command_line, COMMAND_LINE_SIZE);
+	*cmdline_p = command_line;
+
+	parse_early_param();
+
 	/* VMI may relocate the fixmap; do this before touching ioremap area */
 	vmi_init();
 

@@ -795,11 +800,6 @@ void __init setup_arch(char **cmdline_p)
 #endif
 #endif
 
-	strlcpy(command_line, boot_command_line, COMMAND_LINE_SIZE);
-	*cmdline_p = command_line;
-
-	parse_early_param();
-
 #ifdef CONFIG_X86_64
 	check_efer();
 #endif

1.6.1.2

btw., what are you using the 'reservetop' boot option for?

Does this bug trigger in 2.6.30 too?

I'm quite nervous about doing this change so late in the .31 cycle, we've got a hundred early parameters that now get executed much earlier than before. No way can i test all of them and others testing it (like in your case) takes time to trickle through.

So unless this is a .31 regression i'd be inclined to queue it up for .32.

Ingo

Commit-ID:  8126dec32738421afa362114337331337b4be17f
Gitweb:     http://git.kernel.org/tip/8126dec32738421afa362114337331337b4be17f
Author:     Xiao Guangrong <[email protected]>
AuthorDate: Thu, 20 Aug 2009 20:23:11 +0800
Committer:  Ingo Molnar <[email protected]>
CommitDate: Fri, 21 Aug 2009 16:40:30 +0200

x86: Fix system crash when loading with "reservetop" parameter

The system will die if the kernel is booted with "reservetop" parameter, in present code, parse "reservetop" parameter after early_ioremap_init(), and some function still use early_ioremap() after it.

The problem is, "reservetop" parameter can modify 'FIXADDR_TOP', then the virtual address got by early_ioremap() is base on old 'FIXADDR_TOP', but the page mapping is base on new 'FIXADDR_TOP', it will occur page fault, and the IDT is not prepare yet, so, the system is dead.

So, put parse_early_param() in the front of
early_ioremap_init() in this patch.

LKML-Reference: <[email protected]>

---
 arch/x86/kernel/setup.c |   10 +++++-----
 1 files changed, 5 insertions(+), 5 deletions(-)

index 63f32d2..02643cc 100644
@@ -711,6 +711,11 @@ void __init setup_arch(char **cmdline_p)
 	printk(KERN_INFO "Command line: %s\n", boot_command_line);
 #endif
 
+	strlcpy(command_line, boot_command_line, COMMAND_LINE_SIZE);
+	*cmdline_p = command_line;
+
+	parse_early_param();
+
 	/* VMI may relocate the fixmap; do this before touching ioremap area */
 	vmi_init();
 

@@ -793,11 +798,6 @@ void __init setup_arch(char **cmdline_p)
 #endif
 #endif
 
-	strlcpy(command_line, boot_command_line, COMMAND_LINE_SIZE);
-	*cmdline_p = command_line;
-
-	parse_early_param();
-
 #ifdef CONFIG_X86_64
 	check_efer();
 #endif

Hi Ingo,

Actually, this bug is detected by my review first, then confirm it by loading with "reservetop" parameter

OK, this parameter is introduced in v2.6.27, but It seems like less people use it and no one report this bug before. So, I think we can queue it up for .32

Thanks,
Xiao