linux-kernel - Re: [PATCH] mm/page_alloc: Occasionally relinquish zone lock in batch freeing

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20250820132033.1267317-1-joshua.hahnjy@gmail.com>
Date: Wed, 20 Aug 2025 06:20:32 -0700
From: Joshua Hahn <joshua.hahnjy@...il.com>
To: Andrew Morton <akpm@...ux-foundation.org>
Cc: Johannes Weiner <hannes@...xchg.org>,
	Chris Mason <clm@...com>,
	Vlastimil Babka <vbabka@...e.cz>,
	Suren Baghdasaryan <surenb@...le.com>,
	Michal Hocko <mhocko@...e.com>,
	Brendan Jackman <jackmanb@...gle.com>,
	Zi Yan <ziy@...dia.com>,
	linux-mm@...ck.org,
	linux-kernel@...r.kernel.org,
	kernel-team@...a.com,
	Peter Zijlstra <peterz@...radead.org>
Subject: Re: [PATCH] mm/page_alloc: Occasionally relinquish zone lock in batch freeing

On Tue, 19 Aug 2025 14:44:21 -0700 Andrew Morton <akpm@...ux-foundation.org> wrote:

> On Tue, 19 Aug 2025 08:18:45 -0700 Joshua Hahn <joshua.hahnjy@...il.com> wrote:
> 
> > > Pretty this isn't.
> > > 
> > > Sigh, we do so much stuff here and in __free_one_page().
> > > 
> > > What sort of guarantee do we have that the contending task will be able
> > > to get in and grab the spinlock in that tiny time window?
> > 
> > Thank you for pointing this out -- I don't think there is any guarantee.
> > Kiryl suggested that I put a cond_resched() here, in order to guarantee that
> > the contending tasks will be able to grab the spinlock. I think that's a great
> > idea -- I'll make this change in v2.

Hello Andrew, thank you for your review!

> cond_resched() might help because it takes more CPU cycles and expands
> the window. A udelay() would of course do this more nicely.

I was wondering if we could rely on the spinlock implementation here in order
to allow some fairness in who grabs the lock. From what I have gathered, on
a lot of architectures, the default implementation for spin locks use a
queued spin lock (on x86 and arm64, among others, just by doing a quick
grep for "select ARCH_USE_QUEUED_SPINLOCKS"). This means that whoever was
waiting the longest for the spin lock will be able to grab it, guaranteeing
that this function doesn't immediately lock again.

With that said, I understand that the solution should be generic and work for
all architectures. I wonder if it would make sense to change the zone lock
into an explicit queued spin lock?

> But the contending task is already in state TASK_RUNNING so a
> cond_resched() won't have any effect on it?
> 
> Also, callers hold pcp->lock, so cond_resched() cannot be called.

Ah yes, that makes sense.

> Sigh, I dunno, it's all very nasty.  I have vague memories of there
> being a way of relinquishing a lock to some other task which is
> spinning on that lock.  Or at least, a proposal.  Or I dreamed it. 
> peterz would be a good person to ask.

Cc-ing Peter, please let us know if you have any thoughts about all of this!

Thank you Andrew, I hope you have a great day!
Joshua