linux-kernel - Unexplained long boot delays [Was Re: [GIT PULL] RCU changes for v6.9]

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: <4274be61-60bd-4e1e-9c16-26e6e5e06f65@gmail.com>
Date: Tue, 12 Mar 2024 13:32:03 -0700
From: Florian Fainelli <f.fainelli@...il.com>
To: Boqun Feng <boqun.feng@...il.com>, torvalds@...ux-foundation.org
Cc: linux-kernel@...r.kernel.org, kernel-team@...a.com, paulmck@...nel.org,
 mingo@...nel.org, tglx@...utronix.de, rcu@...r.kernel.org,
 joel@...lfernandes.org, neeraj.upadhyay@....com, urezki@...il.com,
 qiang.zhang1211@...il.com, frederic@...nel.org, bigeasy@...utronix.de,
 anna-maria@...utronix.de, chenzhongjin@...wei.com, yangjihong1@...wei.com,
 rostedt@...dmis.org
Subject: Unexplained long boot delays [Was Re: [GIT PULL] RCU changes for
 v6.9]

Hi Boqun,

On 3/8/24 09:15, Boqun Feng wrote:
> Hi Linus,
> 
> Please pull this for the RCU changes of v6.9:
> 
> The following changes since commit 41bccc98fb7931d63d03f326a746ac4d429c1dd3:
> 
>    Linux 6.8-rc2 (2024-01-28 17:01:12 -0800)
> 
> are available in the Git repository at:
> 
>    git://git.kernel.org/pub/scm/linux/kernel/git/boqun/linux.git tags/rcu.next.v6.9
> 
> for you to fetch changes up to 3add00be5fe5810d7aa5ec3af8b6a245ef33144b:
> 
>    Merge branches 'rcu-doc.2024.02.14a', 'rcu-nocb.2024.02.14a', 'rcu-exp.2024.02.14a', 'rcu-tasks.2024.02.26a' and 'rcu-misc.2024.02.14a' into rcu.2024.02.26a (2024-02-26 17:37:25 -0800)
> 
> 
> Two merge conflicts were detected by linux-next:
> 
> * https://lore.kernel.org/lkml/20240226135745.12ac854d@canb.auug.org.au/
> * https://lore.kernel.org/lkml/20240227125522.2bdbe6be@canb.auug.org.au/
> 
> These conflict resolutions from linux-next look good to me, plus I made
> my own resolutions at branch merge/rcu.2024.02.27a for your reference.
> 
> 
> Some highlights of the changes:
> 
> * Eliminates deadlocks involving do_exit() and RCU tasks, by Paul:
>    Instead of SRCU read side critical sections, now a percpu list is used
>    in do_exit() for scaning yet-to-exit tasks.
> 
> * Fixes a deadlock due to the dependency between workqueue and RCU
>    expedited grace period, reported by Anna-Maria Behnsen and Thomas
>    Gleixner and fixed by Frederic: Now RCU expedited always uses its own
>    kthread worker instead of a workqueue.

At least one device in my test farm (ARM 32-bit) has consistently shown 
a very long boot, and some others are intermittently affected. This 
consistently looks like this on most of my devices:

[    2.450351] bcmgenet f0480000.ethernet: GENET 5.0 EPHY: 0x0000
[    2.547562] ata1: SATA link down (SStatus 0 SControl 300)
[  162.107264] unimac-mdio unimac-mdio.0: Broadcom UniMAC MDIO bus

this gets flagged by my boot script as a boot failure since we exceeded 
the 30 seconds timeout given to boot a kernel to a prompt.

It has been somewhat difficult to get a reliable bisection going on, but 
what I am sure of is that e5a3878c947ceef7b6ab68fdc093f3848059842c~1 
does not expose the problem for 10 consecutive boots, while I *might* 
see it at e5a3878c947ceef7b6ab68fdc093f3848059842c and beyond.

Any clues what is going on here?

Thanks!
-- 
Florian