linux-kernel - Re: [PATCH v3] blk-cgroup: Replace u64 sync with spinlock for iostat update

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: <1b19b68adb34410bf6dc8fd3f50e4b82c1a014e4.camel@mediatek.com>
Date: Wed, 17 Jul 2024 04:09:53 +0000
From: Boy Wu (吳勃誼) <Boy.Wu@...iatek.com>
To: "tj@...nel.org" <tj@...nel.org>
CC: "boris@....io" <boris@....io>, "linux-block@...r.kernel.org"
	<linux-block@...r.kernel.org>, "linux-kernel@...r.kernel.org"
	<linux-kernel@...r.kernel.org>, "cgroups@...r.kernel.org"
	<cgroups@...r.kernel.org>, "linux-mediatek@...ts.infradead.org"
	<linux-mediatek@...ts.infradead.org>, "axboe@...nel.dk" <axboe@...nel.dk>,
	Iverlin Wang (王苳霖) <Iverlin.Wang@...iatek.com>,
	"josef@...icpanda.com" <josef@...icpanda.com>,
	"linux-arm-kernel@...ts.infradead.org"
	<linux-arm-kernel@...ts.infradead.org>, "matthias.bgg@...il.com"
	<matthias.bgg@...il.com>, "angelogioacchino.delregno@...labora.com"
	<angelogioacchino.delregno@...labora.com>
Subject: Re: [PATCH v3] blk-cgroup: Replace u64 sync with spinlock for iostat
 update

On Wed, 2024-07-17 at 11:44 +0800, Boy.Wu wrote:
> 
> I think there is another way to fix it. Maybe we should address the
> root cause, which is that u64_stats_update_begin_irqsave
> and u64_stats_update_end_irqrestore do not protect the seqcount in 32
> bit SMP systems. This can be fix by the following change.
> 
> diff --git a/include/linux/u64_stats_sync.h
> b/include/linux/u64_stats_sync.h
> index 46040d66334a..94dd74b4fb2c 100644
> --- a/include/linux/u64_stats_sync.h
> +++ b/include/linux/u64_stats_sync.h
> @@ -64,6 +64,7 @@
>  struct u64_stats_sync {
>  #if BITS_PER_LONG == 32
>         seqcount_t      seq;
> +       spinlock_t      spinlock;
>  #endif
>  };
> 
> @@ -138,6 +139,7 @@ static inline void u64_stats_inc(u64_stats_t *p)
>  static inline void u64_stats_init(struct u64_stats_sync *syncp)
>  {
>         seqcount_init(&syncp->seq);
> +       spin_lock_init(&syncp->spinlock)
>  }
> 
>  static inline void __u64_stats_update_begin(struct u64_stats_sync
> *syncp)
> @@ -191,6 +193,7 @@ static inline unsigned long
> u64_stats_update_begin_irqsave(struct u64_stats_sync
>  {
>         unsigned long flags = __u64_stats_irqsave();
> 
> +       spin_lock_irq(&syncp->spinlock);
>         __u64_stats_update_begin(syncp);
>         return flags;
>  }
> @@ -199,6 +202,7 @@ static inline void
> u64_stats_update_end_irqrestore(struct u64_stats_sync *syncp,
>                                                    unsigned long
> flags)
>  {
>         __u64_stats_update_end(syncp);
> +       spin_unlock_irq(&syncp->spinlock);
>         __u64_stats_irqrestore(flags);
>  }
> 
> --
> boy.wu
> 

Never mind, there is a usage in u64_stats_sync.h

 * Usage :
 *
 * Stats producer (writer) should use following template granted it
already got
 * an exclusive access to counters (a lock is already taken, or per cpu
 * data is used [in a non preemptable context])
 *
 *   spin_lock_bh(...) or other synchronization to get exclusive access
 *   ...
 *   u64_stats_update_begin(&stats->syncp);
 *   u64_stats_add(&stats->bytes64, len); // non atomic operation
 *   u64_stats_inc(&stats->packets64);    // non atomic operation
 *   u64_stats_update_end(&stats->syncp);

--
boy.wu