linux-kernel - Re: [PATCH v5 02/10] locking/mutex: introduce devm_mutex

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 Mar 2024 21:10:16 -0400
From: Waiman Long <longman@...hat.com>
To: George Stark <gnstark@...utedevices.com>, Marek Behún
 <marek.behun@....cz>
Cc: andy.shevchenko@...il.com, pavel@....cz, lee@...nel.org,
 vadimp@...dia.com, mpe@...erman.id.au, npiggin@...il.com,
 christophe.leroy@...roup.eu, hdegoede@...hat.com, mazziesaccount@...il.com,
 peterz@...radead.org, mingo@...hat.com, will@...nel.org,
 boqun.feng@...il.com, nikitos.tr@...il.com, kabel@...nel.org,
 linux-leds@...r.kernel.org, linux-kernel@...r.kernel.org,
 linuxppc-dev@...ts.ozlabs.org, kernel@...utedevices.com
Subject: Re: [PATCH v5 02/10] locking/mutex: introduce devm_mutex_init

On 3/11/24 19:47, George Stark wrote:
> Hello Waiman, Marek
>
> Thanks for the review.
>
> I've never used lockdep for debug but it seems preferable to
> keep that feature working. It could be look like this:
>
> diff --git a/include/linux/mutex.h b/include/linux/mutex.h
> index f7611c092db7..574f6de6084d 100644
> --- a/include/linux/mutex.h
> +++ b/include/linux/mutex.h
> @@ -22,6 +22,8 @@
>  #include <linux/cleanup.h>
>  #include <linux/mutex_types.h>
>
> +struct device;
> +
>  #ifdef CONFIG_DEBUG_LOCK_ALLOC
>  # define __DEP_MAP_MUTEX_INITIALIZER(lockname)            \
>          , .dep_map = {                    \
> @@ -115,10 +117,31 @@ do {                            \
>
>  #ifdef CONFIG_DEBUG_MUTEXES
>
> +int debug_devm_mutex_init(struct device *dev, struct mutex *lock);
> +
> +#define devm_mutex_init(dev, mutex)            \
> +({                            \
> +    int ret;                    \
> +    mutex_init(mutex);                \
> +    ret = debug_devm_mutex_init(dev, mutex);    \
> +    ret;                        \
> +})

The int ret variable is not needed. The macro can just end with 
debug_devm_mutex_init().


> +
>  void mutex_destroy(struct mutex *lock);
>
>  #else
>
> +/*
> +* When CONFIG_DEBUG_MUTEXES is off mutex_destroy is just a nop so
> +* there's no really need to register it in devm subsystem.
"no really need"?
> +*/
> +#define devm_mutex_init(dev, mutex)            \
> +({                            \
> +    typecheck(struct device *, dev);        \
> +    mutex_init(mutex);                \
> +    0;                        \
> +})

Do we need a typecheck() here? Compilation will fail with 
CONFIG_DEBUG_MUTEXES if dev is not a device pointer.


> +
>  static inline void mutex_destroy(struct mutex *lock) {}
>
>  #endif
> diff --git a/kernel/locking/mutex-debug.c b/kernel/locking/mutex-debug.c
> index bc8abb8549d2..967a5367c79a 100644
> --- a/kernel/locking/mutex-debug.c
> +++ b/kernel/locking/mutex-debug.c
> @@ -19,6 +19,7 @@
>  #include <linux/kallsyms.h>
>  #include <linux/interrupt.h>
>  #include <linux/debug_locks.h>
> +#include <linux/device.h>
>
>  #include "mutex.h"
>
> @@ -89,6 +90,16 @@ void debug_mutex_init(struct mutex *lock, const 
> char *name,
>      lock->magic = lock;
>  }
>
> +static void devm_mutex_release(void *res)
> +{
> +    mutex_destroy(res);
> +}
> +
> +int debug_devm_mutex_init(struct device *dev, struct mutex *lock)
> +{
> +    return devm_add_action_or_reset(dev, devm_mutex_release, lock);
> +}
> +
>  /***
>   * mutex_destroy - mark a mutex unusable
>   * @lock: the mutex to be destroyed