linux-kernel - Re: [PATCH] kexec jump: fix compiling warning on xchg(&kexec_lock, 0) in kernel

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: <20080813130749.c406ab6c.akpm@linux-foundation.org>
Date:	Wed, 13 Aug 2008 13:07:49 -0700
From:	Andrew Morton <akpm@...ux-foundation.org>
To:	Linus Torvalds <torvalds@...ux-foundation.org>
Cc:	ebiederm@...ssion.com, ying.huang@...el.com, pavel@....cz,
	nigel@...el.suspend2.net, rjw@...k.pl, vgoyal@...hat.com,
	mingo@...e.hu, linux-kernel@...r.kernel.org,
	kexec@...ts.infradead.org
Subject: Re: [PATCH] kexec jump: fix compiling warning on xchg(&kexec_lock,
 0) in kernel_kexec()

On Wed, 13 Aug 2008 12:50:57 -0700 (PDT)
Linus Torvalds <torvalds@...ux-foundation.org> wrote:

> 
> 
> On Wed, 13 Aug 2008, Andrew Morton wrote:
> > - * in interrupt context :)
> > + * Return true if we acquired the lock
> >   */
> > -static int kexec_lock;
> > +static inline bool kexec_trylock(void)
> > +{
> > +	return !test_and_set_bit(0, &kexec_bitlock);
> 
> Nope. That needs to be an "unsigned long".

It is.

> But more importantl, why not just make it a lock in the first place?
> 
> 	static DEFINE_SPINLOCK(kexec_lock);
> 
> 	#define kexec_trylock() spin_trylock(&kexec_lock)
> 	#define kexec_unlock() spin_unlock(&kexec_lock)
> 
> and then you get it all right and clear and obvious.

Used a bitop to preserve the runtime checking in there.  spin_unlock()
doesn't return the previous lockedness.

Presumably lockdep will whine about spun_unlock(unlocked_lock) though.

> Yeah, and I didn't check whether there is anything that is supposed to be 
> able to sleep. If there is, use a mutex instead of a spinlock, of course.

Yes, it does sleepy things inside the lock.


A bitop seems a better fit to me.  We never spin on that lock (it
always uses test_and_set), so why use a "spin"lock?
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/