linux-kernel - Re: [PATCH bpf-next v6 01/11] kernel: Implement try_module_get

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: <YfBGetHxW+qdk8WD@bombadil.infradead.org>
Date:   Tue, 25 Jan 2022 10:50:34 -0800
From:   Luis Chamberlain <mcgrof@...nel.org>
To:     Kumar Kartikeya Dwivedi <memxor@...il.com>
Cc:     bpf@...r.kernel.org, Alexei Starovoitov <ast@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>,
        Andrii Nakryiko <andrii@...nel.org>, netdev@...r.kernel.org,
        netfilter-devel@...r.kernel.org, Jessica Yu <jeyu@...nel.org>,
        linux-kernel@...r.kernel.org, linux-modules@...r.kernel.org,
        Martin KaFai Lau <kafai@...com>,
        Song Liu <songliubraving@...com>, Yonghong Song <yhs@...com>,
        John Fastabend <john.fastabend@...il.com>,
        Maxim Mikityanskiy <maximmi@...dia.com>,
        Pablo Neira Ayuso <pablo@...filter.org>,
        Florian Westphal <fw@...len.de>,
        Jesper Dangaard Brouer <brouer@...hat.com>,
        Toke Høiland-Jørgensen <toke@...hat.com>
Subject: Re: [PATCH bpf-next v6 01/11] kernel: Implement try_module_get_live

On Sun, Jan 02, 2022 at 09:51:05PM +0530, Kumar Kartikeya Dwivedi wrote:
> Refactor shared functionality between strong_try_module_get and
> try_module_get into a common helper, and expose try_module_get_live
> that returns a bool similar to try_module_get.
> 
> It will be used in the next patch for btf_try_get_module, to eliminate a
> race between module __init function invocation and module_put from BPF
> side.
> 
> Cc: Luis Chamberlain <mcgrof@...nel.org>
> Cc: Jessica Yu <jeyu@...nel.org>
> Cc: linux-kernel@...r.kernel.org
> Cc: linux-modules@...r.kernel.org
> Signed-off-by: Kumar Kartikeya Dwivedi <memxor@...il.com>
> ---
>  include/linux/module.h | 26 +++++++++++++++++++-------
>  kernel/module.c        | 20 ++++++++------------
>  2 files changed, 27 insertions(+), 19 deletions(-)
> 
> diff --git a/include/linux/module.h b/include/linux/module.h
> index c9f1200b2312..eb83aaeaa76e 100644
> --- a/include/linux/module.h
> +++ b/include/linux/module.h
> @@ -608,17 +608,17 @@ void symbol_put_addr(void *addr);
>  /* Sometimes we know we already have a refcount, and it's easier not
>     to handle the error case (which only happens with rmmod --wait). */
>  extern void __module_get(struct module *module);
> -
> -/* This is the Right Way to get a module: if it fails, it's being removed,
> - * so pretend it's not there. */
> -extern bool try_module_get(struct module *module);
> -
> +extern int __try_module_get(struct module *module, bool strong);
>  extern void module_put(struct module *module);
>  
>  #else /*!CONFIG_MODULE_UNLOAD*/
> -static inline bool try_module_get(struct module *module)
> +static inline int __try_module_get(struct module *module, bool strong)
>  {
> -	return !module || module_is_live(module);
> +	if (module && !module_is_live(module))
> +		return -ENOENT;
> +	if (strong && module && module->state == MODULE_STATE_COMING)
> +		return -EBUSY;
> +	return 0;
>  }

The bool return is clear here before on try_module_get().

>  static inline void module_put(struct module *module)
>  {
> @@ -631,6 +631,18 @@ static inline void __module_get(struct module *module)
>  
>  #endif /* CONFIG_MODULE_UNLOAD */
>  
> +/* This is the Right Way to get a module: if it fails, it's being removed,
> + * so pretend it's not there. */
> +static inline bool try_module_get(struct module *module)
> +{
> +	return !__try_module_get(module, false);

Now you're making it negate an int return... 

> +}
> +/* Only take reference for modules which have fully initialized */
> +static inline bool try_module_get_live(struct module *module)
> +{
> +	return !__try_module_get(module, true);
> +}
> +
>  /* This is a #define so the string doesn't get put in every .o file */
>  #define module_name(mod)			\
>  ({						\
> diff --git a/kernel/module.c b/kernel/module.c
> index 84a9141a5e15..a9bb0a5576c8 100644
> --- a/kernel/module.c
> +++ b/kernel/module.c
> @@ -318,12 +318,7 @@ EXPORT_SYMBOL(unregister_module_notifier);
>  static inline int strong_try_module_get(struct module *mod)
>  {
>  	BUG_ON(mod && mod->state == MODULE_STATE_UNFORMED);
> -	if (mod && mod->state == MODULE_STATE_COMING)
> -		return -EBUSY;
> -	if (try_module_get(mod))
> -		return 0;
> -	else
> -		return -ENOENT;

Before this change, this check had no disabled preemption
prior to the first branch, now we are having it moved with
preemption disabled. That's an OK change, but it is a
small functional change.

Because of these two things NACK on this patch for now.
Please split the patch up if you intend to make a new
functional change. And this patch should be easy to read,
this is not.

  Luis