All of lore.kernel.org
 help / color / mirror / Atom feed
From: Lance Yang <lance.yang@linux.dev>
To: Waiman Long <llong@redhat.com>, peterz@infradead.org
Cc: mingo@redhat.com, will@kernel.org, boqun.feng@gmail.com,
	linux-kernel@vger.kernel.org, Lance Yang <ioworker0@gmail.com>,
	Zi Li <zi.li@linux.dev>
Subject: Re: [WARN] LOCKDEP: MAX_LOCKDEP_CHAIN_HLOCKS too low
Date: Tue, 27 May 2025 19:59:26 +0800	[thread overview]
Message-ID: <33103ab3-6237-4215-b8de-bef680bef691@linux.dev> (raw)
In-Reply-To: <25e2293a-f49e-4bc9-bf80-efb66d87c7cb@redhat.com>



On 2025/5/27 13:53, Waiman Long wrote:
> MAX_LOCKDEP_CHAIN_HLOCKS is composed of 2 parts - (1 << 
> MAX_LOCKDEP_CHAINS_BITS) and AVG_LOCKDEP_CHAIN_DEPTH (5). I believe that 
> the average lock chain length is probably bigger than 5 now. We will 
> have to check the /proc/lockdep file to figure out if we should increase 
> it as well. Anyway, I think we should increase 

Yeah, just checked `/proc/lockdep_stats` on my machine, and it shows
'max locking depth: 30':

$ cat /proc/lockdep_stats
  lock-classes:                         2074 [max: 8192]
  direct dependencies:                 22596 [max: 32768]
  indirect dependencies:              124267
  all direct dependencies:            527384
  dependency chains:                   51354 [max: 65536]
  dependency chain hlocks used:       327679 [max: 327680]
  dependency chain hlocks lost:            1
  in-hardirq chains:                     209
  in-softirq chains:                    1068
  in-process chains:                   50076
  stack-trace entries:                306274 [max: 524288]
  number of stack traces:              11921
  number of stack hash chains:          8482
  combined max dependencies:      2651851138
  hardirq-safe locks:                     85
  hardirq-unsafe locks:                 1301
  softirq-safe locks:                    284
  softirq-unsafe locks:                 1123
  irq-safe locks:                        303
  irq-unsafe locks:                     1301
  hardirq-read-safe locks:                 4
  hardirq-read-unsafe locks:             312
  softirq-read-safe locks:                12
  softirq-read-unsafe locks:             307
  irq-read-safe locks:                    12
  irq-read-unsafe locks:                 312
  uncategorized locks:                   352
  unused locks:                            0
  max locking depth:                      30
  max bfs queue depth:                   379
  max lock class index:                 2073
  debug_locks:                             0

  zapped classes:                          6
  zapped lock chains:                    163
  large chain blocks:                      0

And, the average lock chain depth could be calculated as:

dependency chain hlocks used (327679) / dependency chains (51354) ~= 6.38

Seems like we should also consider bumping 'AVG_LOCKDEP_CHAIN_DEPTH' to
something like 7 or higher.

> CONFIG_LOCKDEP_CHAINS_BITS to at least 17, those we may still hit the 
> "MAX_LOCKDEP_CHAIN_HLOCKS too low" if we run a variety of different 
> workloads without reboot.

Agreed. We may still hit this issue, but tweaking these values can make
it less likely ;)

Thanks,
Lance


      reply	other threads:[~2025-05-27 11:59 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-05-27  2:02 [WARN] LOCKDEP: MAX_LOCKDEP_CHAIN_HLOCKS too low Lance Yang
2025-05-27  4:48 ` Waiman Long
2025-05-27  5:33   ` Lance Yang
2025-05-27  5:53     ` Waiman Long
2025-05-27 11:59       ` Lance Yang [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=33103ab3-6237-4215-b8de-bef680bef691@linux.dev \
    --to=lance.yang@linux.dev \
    --cc=boqun.feng@gmail.com \
    --cc=ioworker0@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=llong@redhat.com \
    --cc=mingo@redhat.com \
    --cc=peterz@infradead.org \
    --cc=will@kernel.org \
    --cc=zi.li@linux.dev \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.