linux-kernel - Re: [RFC PATCH v3 2/6] module: preserve Elf information for livepatch modules

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]

Date:	Mon, 11 Jan 2016 11:55:04 +1030
From:	Rusty Russell <rusty@...tcorp.com.au>
To:	Jessica Yu <jeyu@...hat.com>, Josh Poimboeuf <jpoimboe@...hat.com>,
	Seth Jennings <sjenning@...hat.com>,
	Jiri Kosina <jikos@...nel.org>,
	Vojtech Pavlik <vojtech@...e.com>,
	Jonathan Corbet <corbet@....net>,
	Miroslav Benes <mbenes@...e.cz>
Cc:	linux-api@...r.kernel.org, live-patching@...r.kernel.org,
	x86@...nel.org, linux-kernel@...r.kernel.org,
	linux-s390@...r.kernel.org, linux-doc@...r.kernel.org
Subject: Re: [RFC PATCH v3 2/6] module: preserve Elf information for livepatch modules

Hi Jessica,

        Nice patch series.  Minor issues below, but they're really
just nits which can be patched afterwards if you're getting sick of
rebasing :)

> +#ifdef CONFIG_LIVEPATCH
> +/*
> + * copy_module_elf - preserve Elf information about a module
> + *
> + * Copy relevant Elf information from the load_info struct.
> + * Note: not all fields from the original load_info are
> + * copied into mod->info.

This makes me nervous, to have a struct which is half-initialized.

Better would be to have a separate type for this, eg:

struct livepatch_modinfo {
        Elf_Ehdr hdr;
        Elf_Shdr *sechdrs;
        char *secstrings;
        /* FIXME: Which of these do you need? */
        struct {
        	unsigned int sym, str, mod, vers, info, pcpu;
        } index;
};

This also avoids an extra allocation as hdr is no longer a ptr.

> +	/*
> +	 * Update symtab's sh_addr to point to a valid
> +	 * symbol table, as the temporary symtab in module
> +	 * init memory will be freed
> +	 */
> +	mod->info->sechdrs[mod->info->index.sym].sh_addr = (unsigned long)mod->core_symtab;

This comment is a bit misleading: it's actually pointing into the
temporary module copy, which will be discarded.  The init section is
slightly different...

> +#ifdef CONFIG_LIVEPATCH
> +static int check_livepatch_modinfo(struct module *mod, struct load_info *info)
> +{
> +	mod->klp = get_modinfo(info, "livepatch") ? true : false;
> +
> +	return 0;
> +}
> +#else
> +static int check_livepatch_modinfo(struct module *mod, struct load_info *info)
> +{
> +	if (get_modinfo(info, "livepatch"))
> +		return -EINVAL;
> +
> +	return 0;
> +}
> +#endif

The term "check" implies no side-effects; I'd usually call this
int find_livepatch_modinfo(), and make sure you use the error code in
the caller:

> +	err = check_livepatch_modinfo(mod, info);
> +	if (err)
> +		return -ENOEXEC;
> +

Thanks,
Rusty.