linux-kernel - Re: [PATCH] x86: Put trampoline

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <4B19B53B.60605@zytor.com>
Date:	Fri, 04 Dec 2009 17:19:55 -0800
From:	"H. Peter Anvin" <hpa@...or.com>
To:	Rakib Mullick <rakib.mullick@...il.com>
CC:	Ingo Molnar <mingo@...e.hu>, Thomas Gleixner <tglx@...utronix.de>,
	LKML <linux-kernel@...r.kernel.org>, x86@...nel.org,
	"Cihula, Joseph" <joseph.cihula@...el.com>,
	Shane Wang <shane.wang@...el.com>
Subject: Re: [PATCH] x86: Put trampoline_data into readonly section.

On 12/04/2009 08:24 AM, Rakib Mullick wrote:
> WARNING: arch/x86/kernel/built-in.o(.text+0x13c77): Section mismatch
> in reference from the function tboot_setup_sleep() to the variable
> .cpuinit.rodata:trampoline_end
> The function tboot_setup_sleep() references
> the variable __cpuinitconst trampoline_end.
> This is often because tboot_setup_sleep lacks a __cpuinitconst
> annotation or the annotation of trampoline_end is wrong.
> 
> WARNING: arch/x86/kernel/built-in.o(.text+0x13c84): Section mismatch
> in reference from the function tboot_setup_sleep() to the variable
> .cpuinit.rodata:trampoline_data
> The function tboot_setup_sleep() references
> the variable __cpuinitconst trampoline_data.
> This is often because tboot_setup_sleep lacks a __cpuinitconst
> annotation or the annotation of trampoline_data is wrong.
> 
> --- linus/arch/x86/kernel/trampoline_32.S	2009-12-04 17:17:51.000000000 +0600
> +++ rakib/arch/x86/kernel/trampoline_32.S	2009-12-04 17:24:44.000000000 +0600
> @@ -32,8 +32,13 @@
>  #include <asm/segment.h>
>  #include <asm/page_types.h>
> 
> +#ifdef CONFIG_ACPI_SLEEP
> +.section .rodata, "a", @progbits
> +#else
>  /* We can free up trampoline after bootup if cpu hotplug is not supported. */
>  __CPUINITRODATA
> +#endif
> +

This is a false positive.

The reference in question is actually:

	add_mac_region(virt_to_phys(trampoline_base), TRAMPOLINE_SIZE);

... which in turn is defined as ...

#define TRAMPOLINE_SIZE roundup(trampoline_end - trampoline_data, PAGE_SIZE)

This causes "references" of two symbols which aren't actually referenced
at all.

Perhaps the better thing would be to define trampoline_size as an
absolute symbol in arch/x86/kernel/trampoline_*.S, especially since it
probably generates bad code to subtract two global symbols like that.

	-hpa
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/