linux-kernel - Re: [PATCH v3 3/5] locking/qspinlock: Introduce CNA into the slow path of qspinlock

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>] [day] [month] [year] [list]

Date:   Wed, 17 Jul 2019 17:09:13 +0200
From:   Peter Zijlstra <peterz@...radead.org>
To:     Alex Kogan <alex.kogan@...cle.com>
Cc:     linux@...linux.org.uk, mingo@...hat.com, will.deacon@....com,
        arnd@...db.de, longman@...hat.com, linux-arch@...r.kernel.org,
        linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
        tglx@...utronix.de, bp@...en8.de, hpa@...or.com, x86@...nel.org,
        guohanjun@...wei.com, jglauber@...vell.com,
        steven.sistare@...cle.com, daniel.m.jordan@...cle.com,
        dave.dice@...cle.com, rahul.x.yadav@...cle.com
Subject: Re: [PATCH v3 3/5] locking/qspinlock: Introduce CNA into the slow
 path of qspinlock

On Wed, Jul 17, 2019 at 10:27:30AM -0400, Alex Kogan wrote:
> > On Jul 17, 2019, at 4:39 AM, Peter Zijlstra <peterz@...radead.org> wrote:

> > static void cna_splice_tail(struct cna_node *cn, struct cna_node *head, struct cna_node *tail)
> > {
> > 	struct cna_node *list;
> > 
> > 	/* remove [head,tail] */
> > 	WRITE_ONCE(cn->mcs.next, tail->mcs.next);
> > 	tail->mcs.next = NULL;
> > 
> > 	/* stick [head,tail] on the secondary list tail */
> > 	if (cn->mcs.locked <= 1) {
> > 		/* create secondary list */
> > 		head->tail = tail;
> > 		cn->mcs.locked = head->encoded_tail;
> > 	} else {
> > 		/* add to tail */
> > 		list = (struct cna_node *)decode_tail(cn->mcs.locked);
> > 		list->tail->next = head;
> > 		list->tail = tail;
> > 	}
> > }
> > 
> > static struct cna_node *cna_find_next(struct mcs_spinlock *node)
> > {
> > 	struct cna_node *cni, *cn = (struct cna_node *)node;
> > 	struct cna_node *head, *tail = NULL;
> > 
> > 	/* find any next lock from 'our' node */
> > 	for (head = cni = (struct cna_node *)READ_ONCE(cn->mcs.next);
> > 	     cni && cni->node != cn->node;
> > 	     tail = cni, cni = (struct cna_node *)READ_ONCE(cni->mcs.next))
> > 		;
> > 
> > 	/* when found, splice any skipped locks onto the secondary list */
> > 	if (cni && tail)
> > 		cna_splice_tail(cn, head, tail);
> > 
> > 	return cni;
> > }
> > 
> > How's that?
> 
> This is almost perfect!! :)
> 
> The above should work, but I think we should have a specialized fast-path for 
> checking the immediate next node in the main queue. This would be the common
> case, once we splice ‘other’ nodes onto the secondary queue. In the above we
> would go through four branches before returning from cna_find_next(). In the 
> following we would have just one:

Right, but can you measure a difference? ;-) Anyway, no real objection,
just playing devils advocate here.

> > static struct cna_node *cna_find_next(struct mcs_spinlock *node)
> > {
> > 	struct cna_node *cn = (struct cna_node *)node;
> 	   struct cna_node *cni = (struct cna_node *)READ_ONCE(cn->mcs.next);
> > 
> > 	struct cna_node *head, *tail = NULL;
> > 
> 	   /* fast path */
> 	   if (cni->node == cn->node) 
> 		return cni;
> 
> > 	/* find any next lock from 'our' node */
> > 	for (head = cn->mcs.next;
	     head = cni,

you just did that load :-)

> > 	     cni && cni->node != cn->node;
> > 	     tail = cni, cni = (struct cna_node *)READ_ONCE(cni->mcs.next))
> > 		;
> > 
> > 	/* when found, splice any skipped locks onto the secondary list */
> > 	if (cni && tail)
> > 		cna_splice_tail(cn, head, tail);
> > 
> > 	return cni;
> > }
> 
> 
> Also, any reason you say ‘lock’ instead of ’node’ in the comments?
> I.e., I think "when found, splice any skipped locks onto the secondary list” should be
> "when found, splice any skipped nodes onto the secondary list”.

Due to the confusion between lock waiter node and numa node :-)