linux-kernel - Re: [PATCH] MIPS: Ignore relocs against __ex

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 for Android: free password hash cracker in your pocket

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

Message-ID: <Z8A6lBXX8LZPyEDS@alpha.franken.de>
Date: Thu, 27 Feb 2025 11:12:36 +0100
From: Thomas Bogendoerfer <tsbogend@...ha.franken.de>
To: Xi Ruoyao <xry111@...111.site>
Cc: Steven Rostedt <rostedt@...dmis.org>,
	Jiaxun Yang <jiaxun.yang@...goat.com>,
	Matt Redfearn <matt.redfearn@...ize.com>,
	linux-mips@...r.kernel.org, linux-kbuild@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH] MIPS: Ignore relocs against __ex_table for relocatable
 kernel

On Wed, Feb 26, 2025 at 09:28:41PM +0800, Xi Ruoyao wrote:
> Since commit 6f2c2f93a190 ("scripts/sorttable: Remove unneeded
> Elf_Rel"), sorttable no longer clears relocs against __ex_table,
> claiming "it was never used."  But in fact MIPS relocatable kernel had
> been implicitly depending on this behavior, so after this commit the
> MIPS relocatable kernel has started to spit oops like:
> 
> 	CPU 1 Unable to handle kernel paging request at virtual address 000000fffbbdbff8, epc == ffffffff818f9a6c, ra == ffffffff813ad7d0
> 	... ...
> 	Call Trace:
> 	[<ffffffff818f9a6c>] __raw_copy_from_user+0x48/0x2fc
> 	[<ffffffff813ad7d0>] cp_statx+0x1a0/0x1e0
> 	[<ffffffff813ae528>] do_statx_fd+0xa8/0x118
> 	[<ffffffff813ae670>] sys_statx+0xd8/0xf8
> 	[<ffffffff81156cc8>] syscall_common+0x34/0x58
> 
> So ignore those relocs on our own to fix the issue.
> 
> Fixes: 6f2c2f93a190 ("scripts/sorttable: Remove unneeded Elf_Rel")
> Signed-off-by: Xi Ruoyao <xry111@...111.site>
> ---
>  arch/mips/boot/tools/relocs.c | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/arch/mips/boot/tools/relocs.c b/arch/mips/boot/tools/relocs.c
> index a88d66c46d7f..9863e1d5c62e 100644
> --- a/arch/mips/boot/tools/relocs.c
> +++ b/arch/mips/boot/tools/relocs.c
> @@ -468,6 +468,8 @@ static void walk_relocs(int (*process)(struct section *sec, Elf_Rel *rel,
>  			Elf_Sym *sym, const char *symname))
>  {
>  	int i;
> +	struct section *extab_sec = sec_lookup("__ex_table");
> +	int extab_index = extab_sec ? extab_sec - secs : -1;
>  
>  	/* Walk through the relocations */
>  	for (i = 0; i < ehdr.e_shnum; i++) {
> @@ -480,6 +482,9 @@ static void walk_relocs(int (*process)(struct section *sec, Elf_Rel *rel,
>  		if (sec->shdr.sh_type != SHT_REL_TYPE)
>  			continue;
>  
> +		if (sec->shdr.sh_info == extab_index)
> +			continue;
> +
>  		sec_symtab  = sec->link;
>  		sec_applies = &secs[sec->shdr.sh_info];
>  		if (!(sec_applies->shdr.sh_flags & SHF_ALLOC))
> -- 
> 2.48.1

applied to mips-fixes.

Thomas.

-- 
Crap can work. Given enough thrust pigs will fly, but it's not necessarily a
good idea.                                                [ RFC1925, 2.3 ]