linux-kernel - Re: [PATCH] module: check for exit sections in layout_sections() instead of module_init

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] [day] [month] [year] [list]

Message-ID: <YJ4i3XNa3V7ifCih@p200300cbcf361a0029e37a38368d6727.dip0.t-ipconnect.de>
Date:   Fri, 14 May 2021 09:12:29 +0200
From:   Jessica Yu <jeyu@...nel.org>
To:     Russell King - ARM Linux admin <linux@...linux.org.uk>
Cc:     linux-kernel@...r.kernel.org, Peter Zijlstra <peterz@...radead.org>
Subject: Re: [PATCH] module: check for exit sections in layout_sections()
 instead of module_init_section()

+++ Russell King - ARM Linux admin [12/05/21 17:06 +0100]:
>Hi,
>
>On Wed, May 12, 2021 at 04:46:53PM +0200, Jessica Yu wrote:
>> diff --git a/kernel/module.c b/kernel/module.c
>> index 173a09175511..a5c9842371b1 100644
>> --- a/kernel/module.c
>> +++ b/kernel/module.c
>> @@ -2430,6 +2430,9 @@ static void layout_sections(struct module *mod, struct load_info *info)
>>  			if ((s->sh_flags & masks[m][0]) != masks[m][0]
>>  			    || (s->sh_flags & masks[m][1])
>>  			    || s->sh_entsize != ~0UL
>> +#ifndef CONFIG_MODULE_UNLOAD
>> +			    || module_exit_section(sname)
>> +#endif
>>  			    || module_init_section(sname))
>
>How about a helper to make this a bit easier in both these places to
>make the code more undertsandable? I think the great value comes from
>the resulting change in the second hunk.
>
>static bool module_evictable_section(const char *sname)
>{
>#ifndef CONFIG_MODULE_UNLOAD
>	if (module_exit_section(sname))
>		return true;
>#endif
>	return module_init_section(sname);
>}
>
>and then just use that above?
>
>>  				continue;
>>  			s->sh_entsize = get_offset(mod, &mod->core_layout.size, s, i);
>> @@ -2463,7 +2466,11 @@ static void layout_sections(struct module *mod, struct load_info *info)
>>  			if ((s->sh_flags & masks[m][0]) != masks[m][0]
>>  			    || (s->sh_flags & masks[m][1])
>>  			    || s->sh_entsize != ~0UL
>> +#ifndef CONFIG_MODULE_UNLOAD
>> +			    || (!module_init_section(sname) && !module_exit_section(sname)))
>> +#else
>>  			    || !module_init_section(sname))
>> +#endif
>
>I find this a tad confusing, and this is the reason for my suggestion
>above. With that, this becomes:
>
>			    || !module_evictable_section(sname))
>
>which can be clearly seen is the opposite condition from the above
>without doing mental logic gymnastics.

Thanks Russell for the feedback! Yeah, agreed that it could be made
easier to read - will respin with a helper function.

Jessica