linux-kernel - Re: [PATCH] locking: type cleanup when accessing fast_read

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: <20150519111105.GB3644@twins.programming.kicks-ass.net>
Date:	Tue, 19 May 2015 13:11:05 +0200
From:	Peter Zijlstra <peterz@...radead.org>
To:	Nicholas Mc Guire <hofrat@...dl.org>
Cc:	Ingo Molnar <mingo@...hat.com>, linux-kernel@...r.kernel.org,
	Oleg Nesterov <oleg@...hat.com>
Subject: Re: [PATCH] locking: type cleanup when accessing fast_read_ctr

On Mon, May 18, 2015 at 07:48:02PM +0200, Nicholas Mc Guire wrote:
> static code checking with coccinelle was unhappy with:
> ./kernel/locking/percpu-rwsem.c:113 WARNING: return of wrong type
> 	int != unsigned int
> 
> 
> current state:
> 
> include/linux/percpu-rwsem.h:
> struct percpu_rw_semaphore {
>   unsigned int __percpu   *fast_read_ctr;
>   ...
>   atomic_t                slow_read_ctr;
> 
> allocated as int:
> 
> int __percpu_init_rwsem(struct percpu_rw_semaphore *brw,
>                         const char *name, struct lock_class_key *rwsem_key)
> {
> 	brw->fast_read_ctr = alloc_percpu(int); 
> 
> used compliant with type int in:
> 
> static bool update_fast_ctr(struct percpu_rw_semaphore *brw, 
> 			    unsigned int val)
> usage (2):
>   update_fast_ctr(brw, +1)
>   update_fast_ctr(brw, -1)
>  
> so val should be int here not unsigned int 
> 
> static int clear_fast_ctr(struct percpu_rw_semaphore *brw)
> usage (1):
>   atomic_add(clear_fast_ctr(brw), &brw->slow_read_ctr);
> 
> slow_read_ctr is atomic_t which is int so it would be consistent here to
> have fast_read_ctr being changed to int as well.
> 
> This patch changes:
>   fast_read_ctr to int
>   clear_fast_ctr():sum to int
>   update_fast_ctr():val to int
> to make usage of fast_read_ctr consistent with respect to type.
> 
>   ( Patch was build tested with x86_64_defconfig + 
>     CONFIG_UPROBE_EVENT (implies CONFIG_PERCPU_RWSEM=y))
> 
> Signed-off-by: Nicholas Mc Guire <hofrat@...dl.org>

So the value is unsigned by purpose; that said we should never cross the
2G I think, so it really doesn't matter much.

Cc'ed Oleg who wrote this code, left the rest of the msg for his
convencience.

That said; Oleg, should I post those patches implementing percpu-rwsem
with the stuff we did for the hotplug lock a while back? I think I
actually have those around somewhere.

> ---
>  include/linux/percpu-rwsem.h  |    2 +-
>  kernel/locking/percpu-rwsem.c |    4 ++--
>  2 files changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/include/linux/percpu-rwsem.h b/include/linux/percpu-rwsem.h
> index 3e88c9a..ba37dbe 100644
> --- a/include/linux/percpu-rwsem.h
> +++ b/include/linux/percpu-rwsem.h
> @@ -8,7 +8,7 @@
>  #include <linux/lockdep.h>
>  
>  struct percpu_rw_semaphore {
> -	unsigned int __percpu	*fast_read_ctr;
> +	int __percpu		*fast_read_ctr;
>  	atomic_t		write_ctr;
>  	struct rw_semaphore	rw_sem;
>  	atomic_t		slow_read_ctr;
> diff --git a/kernel/locking/percpu-rwsem.c b/kernel/locking/percpu-rwsem.c
> index 652a8ee..002837d 100644
> --- a/kernel/locking/percpu-rwsem.c
> +++ b/kernel/locking/percpu-rwsem.c
> @@ -52,7 +52,7 @@ void percpu_free_rwsem(struct percpu_rw_semaphore *brw)
>   * reader inside the critical section. See the comments in down_write and
>   * up_write below.
>   */
> -static bool update_fast_ctr(struct percpu_rw_semaphore *brw, unsigned int val)
> +static bool update_fast_ctr(struct percpu_rw_semaphore *brw, int val)
>  {
>  	bool success = false;
>  
> @@ -102,7 +102,7 @@ void percpu_up_read(struct percpu_rw_semaphore *brw)
>  
>  static int clear_fast_ctr(struct percpu_rw_semaphore *brw)
>  {
> -	unsigned int sum = 0;
> +	int sum = 0;
>  	int cpu;
>  
>  	for_each_possible_cpu(cpu) {
> -- 
> 1.7.10.4
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/