linux-kernel - Re: [PATCH printk 1/5] printk: get new seq before enabling interrupts

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:   Wed, 23 Sep 2020 16:17:23 +0200
From:   Petr Mladek <pmladek@...e.com>
To:     John Ogness <john.ogness@...utronix.de>
Cc:     Sergey Senozhatsky <sergey.senozhatsky.work@...il.com>,
        Sergey Senozhatsky <sergey.senozhatsky@...il.com>,
        Steven Rostedt <rostedt@...dmis.org>,
        Linus Torvalds <torvalds@...ux-foundation.org>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH printk 1/5] printk: get new seq before enabling interrupts

On Tue 2020-09-22 17:44:12, John Ogness wrote:
> After copying all records to the dynamic ringbuffer, setup_log_buf()
> checks to see if any records were dropped during the switch. However,
> it needs to check before enabling interrupts since new records could
> arrive in an interrupt, thus causing setup_log_buf() to erroneously
> think that it had dropped messages.

Have you seen the problem, please?

> Signed-off-by: John Ogness <john.ogness@...utronix.de>
> ---
>  kernel/printk/printk.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c
> index 1fe3d0cb2fe0..00bc1fce3299 100644
> --- a/kernel/printk/printk.c
> +++ b/kernel/printk/printk.c
> @@ -1181,12 +1181,12 @@ void __init setup_log_buf(int early)
>  	 */
>  	prb = &printk_rb_dynamic;
>  
> +	seq = prb_next_seq(&printk_rb_static) - seq;
> +
>  	logbuf_unlock_irqrestore(flags);
>  
> -	if (seq != prb_next_seq(&printk_rb_static)) {

I can't see how these two values could get modified after enabling interrupts.

  + @seq is set in the for-cycle before the interrupts are enabled.

  + @prb is updated before the interrupts are  enabled. So that
    the static buffer should not longer be used after that.

> -		pr_err("dropped %llu messages\n",
> -		       prb_next_seq(&printk_rb_static) - seq);
> -	}
> +	if (seq)
> +		pr_err("dropped %llu messages\n", seq);
>  
>  	pr_info("log_buf_len: %u bytes\n", log_buf_len);
>  	pr_info("early log buf free: %u(%u%%)\n",

Best Regards,
Petr