linux-kernel - Re: [PATCH 5.10-rt] printk: ignore that console preempted by irq/softirq

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: <20240527092403.x2m7SpvL@linutronix.de>
Date: Mon, 27 May 2024 11:24:03 +0200
From: Sebastian Andrzej Siewior <bigeasy@...utronix.de>
To: xu.xin16@....com.cn
Cc: john.ogness@...utronix.de, rostedt@...dmis.org, zhang.yunkai@....com.cn,
	yang.yang29@....com.cn, liu.chun2@....com.cn, si.hao@....com.cn,
	linux-kernel@...r.kernel.org, linux-rt-users@...r.kernel.org
Subject: Re: [PATCH  5.10-rt] printk: ignore that console preempted by
 irq/softirq

On 2024-05-23 23:55:37 [+0800], xu.xin16@....com.cn wrote:
> diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c
> index 7f27cfee283e..faab85dd4439 100644
> --- a/kernel/printk/printk.c
> +++ b/kernel/printk/printk.c
> @@ -3735,6 +3735,14 @@ bool pr_flush(int timeout_ms, bool reset_on_progress)
> diff = 0;
> 
> for_each_console(con) {
> + /*
> + * When we're in the unpreemptible context on the same cpu
> + * with which the thread of console locates on, we should
> + * ignore this console, because it's a vain.
> + */
> + if (!preemptible() && con->thread &&
> + task_cpu(con->thread) == smp_processor_id())
> + continue;
> if (!(con->flags & CON_ENABLED))
> continue;
> printk_seq = read_console_seq(con);

This does not apply.
There is `may_sleep' set earlier.

There is no console_lock() around for each…

The other question is which kernel started enforcing might_sleep() for
pr_flush(). This should be applied to all kernel or none so we don't
have random behaviour across kernels (5.4 yes, 5.10 no, 5.15 yes).

This is a delay of max 1 sec during bug() and panic(). Not sure how
"critical" this is…

Sebastian