Re: [PATCH][v4] hung_task: Panic when there are more than N hung tasks at the same time
Lance Yang <[email protected]> Thu, 16 Oct 2025 13:07:23 +0800
| Newsgroups | com.zx2c4.lists.wireguard,org.infradead.lists.linux-arm-kernel,org.kernel.vger.linux-doc,org.kernel.vger.linux-kernel,org.kernel.vger.linux-kselftest,org.kernel.vger.netdev,org.ozlabs.lists.linux-aspeed |
|---|---|
| Message-ID | <[email protected]> |
LGTM. It works as expected, thanks! On 2025/10/15 14:36, lirongqing wrote: > From: Li RongQing <[email protected]> For the commit message, I'd suggest the following for better clarity: ``` The hung_task_panic sysctl is currently a blunt instrument: it's all or nothing. Panicking on a single hung task can be an overreaction to a transient glitch. A more reliable indicator of a systemic problem is when multiple tasks hang simultaneously. Extend hung_task_panic to accept an integer threshold, allowing the kernel to panic only when N hung tasks are detected in a single scan. This provides finer control to distinguish between isolated incidents and system-wide failures. The accepted values are: - 0: Don't panic (unchanged) - 1: Panic on the first hung task (unchanged) - N > 1: Panic after N hung tasks are detected in a single scan The original behavior is preserved for values 0 and 1, maintaining full backward compatibility. ``` If you agree, likely no need to resend - Andrew could pick it up directly when applying :) > > Currently, when 'hung_task_panic' is enabled, the kernel panics > immediately upon detecting the first hung task. However, some hung > tasks are transient and allow system recovery, while persistent hangs > should trigger a panic when accumulating beyond a threshold. > > Extend the 'hung_task_panic' sysctl to accept a threshold value > specifying the number of hung tasks that must be detected before > triggering a kernel panic. This provides finer control for environments > where transient hangs may occur but persistent hangs should be fatal. > > The sysctl now accepts: > - 0: don't panic (maintains original behavior) > - 1: panic on first hung task (maintains original behavior) > - N > 1: panic after N hung tasks are detected in a single scan > > This maintains backward compatibility while providing flexibility for > different hang scenarios. > > Signed-off-by: Li RongQing <[email protected]> > Cc: Andrew Jeffery <[email protected]> > Cc: Anshuman Khandual <[email protected]> > Cc: Arnd Bergmann <[email protected]> > Cc: David Hildenbrand <[email protected]> > Cc: Florian Wesphal <[email protected]> > Cc: Jakub Kacinski <[email protected]> > Cc: Jason A. Donenfeld <[email protected]> > Cc: Joel Granados <[email protected]> > Cc: Joel Stanley <[email protected]> > Cc: Jonathan Corbet <[email protected]> > Cc: Kees Cook <[email protected]> > Cc: Lance Yang <[email protected]> > Cc: Liam Howlett <[email protected]> > Cc: Lorenzo Stoakes <[email protected]> > Cc: "Masami Hiramatsu (Google)" <[email protected]> > Cc: "Paul E . McKenney" <[email protected]> > Cc: Pawan Gupta <[email protected]> > Cc: Petr Mladek <[email protected]> > Cc: Phil Auld <[email protected]> > Cc: Randy Dunlap <[email protected]> > Cc: Russell King <[email protected]> > Cc: Shuah Khan <[email protected]> > Cc: Simon Horman <[email protected]> > Cc: Stanislav Fomichev <[email protected]> > Cc: Steven Rostedt <[email protected]> > --- So: Reviewed-by: Lance Yang <[email protected]> Tested-by: Lance Yang <[email protected]> Cheers, Lance