linux-kernel - Re: [PATCH] module: make it clearer when we're handling kallsyms symbols vs exported symbols

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: <alpine.LSU.2.21.1811221109220.6711@pobox.suse.cz>
Date:   Thu, 22 Nov 2018 11:19:48 +0100 (CET)
From:   Miroslav Benes <mbenes@...e.cz>
To:     Jessica Yu <jeyu@...nel.org>
cc:     linux-kernel@...r.kernel.org
Subject: Re: [PATCH] module: make it clearer when we're handling kallsyms
 symbols vs exported symbols

On Wed, 21 Nov 2018, Jessica Yu wrote:

> The module loader internally works with both exported symbols
> represented as struct kernel_symbol, as well as Elf symbols from a
> module's symbol table. It's hard to distinguish sometimes which type of
> symbol we're handling given that some helper function names are not
> consistent or helpful. Take get_ksymbol() for instance - are we
> looking for an exported symbol or a kallsyms symbol here? Or symname()
> and kernel_symbol_name() - which function handles an exported symbol and
> which one an Elf symbol?
> 
> Clean up and unify the function naming scheme a bit to make it clear
> which kind of symbol we're handling. This change only affects static
> functions internal to the module loader.
> 
> Signed-off-by: Jessica Yu <jeyu@...nel.org>

Great. It should help a lot. Pity we cannot rename find_symbol() as well.

I have only a naming nit. I think it is nice to have 
<verb>_exported_<noun> convention. New kallsyms_ names don't hold it 
though. Wouldn't it be better to be consistent and have 
find_kallsyms_symbol() instead of kallsyms_find_symbol()? Or we could do 
the opposite and have a "namespace" prefix first. That is, 
exported_<verb>_<noun>. However, I don't like it that much. 

To be honest, your approach may be the best in the end.

What do you think?

Regards,
Miroslav