linux-kernel - Re: [PATCH] lockdep: report broken irq restoration

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

Message-ID: <20210104101402.GA1482@C02TD0UTHF1T.local>
Date:   Mon, 4 Jan 2021 10:14:02 +0000
From:   Mark Rutland <mark.rutland@....com>
To:     Peter Zijlstra <peterz@...radead.org>
Cc:     linux-kernel@...r.kernel.org, Andy Lutomirski <luto@...nel.org>,
        Ingo Molnar <mingo@...hat.com>,
        Juergen Gross <jgross@...e.com>,
        Thomas Gleixner <tglx@...utronix.de>
Subject: Re: [PATCH] lockdep: report broken irq restoration

On Thu, Dec 17, 2020 at 03:36:40PM +0100, Peter Zijlstra wrote:
> On Wed, Dec 09, 2020 at 06:33:37PM +0000, Mark Rutland wrote:
> > This means that a seuence such as:
> 
> +q
> 
> > diff --git a/include/linux/irqflags.h b/include/linux/irqflags.h
> > index 3ed4e8771b64..bca3c6fa8270 100644
> > --- a/include/linux/irqflags.h
> > +++ b/include/linux/irqflags.h
> > @@ -220,10 +220,26 @@ do {						\
> >  
> >  #else /* !CONFIG_TRACE_IRQFLAGS */
> >  
> > +#ifdef CONFIG_DEBUG_IRQFLAGS
> > +extern void warn_bogus_irq_restore(bool *warned);
> > +#define check_bogus_irq_restore()				\
> > +	do {							\
> > +		static bool __section(".data.once") __warned;	\
> > +		if (unlikely(!raw_irqs_disabled()))		\
> > +			warn_bogus_irq_restore(&__warned);	\
> > +	} while (0)
> > +#else
> > +#define check_bogus_irq_restore() do { } while (0)
> > +#endif
> > +
> >  #define local_irq_enable()	do { raw_local_irq_enable(); } while (0)
> >  #define local_irq_disable()	do { raw_local_irq_disable(); } while (0)
> >  #define local_irq_save(flags)	do { raw_local_irq_save(flags); } while (0)
> > -#define local_irq_restore(flags) do { raw_local_irq_restore(flags); } while (0)
> > +#define local_irq_restore(flags)		\
> > +	do {					\
> > +		check_bogus_irq_restore();	\
> > +		raw_local_irq_restore(flags);	\
> > +	} while (0)
> 
> Shouldn't that be in raw_local_irq_restore() ?

Yup, that'd be preferable. Note that will require a refactoring here if
we want the warning to be available regardless of CONFIG_TRACE_IRQFLAGS.

I'd intended to rejig that for v2, but I didn't get the chance before
finishing work for the year, and had only done some basic rework in my
WIP (e.g. removing the per-instance bool as per Andy's comments):

https://git.kernel.org/pub/scm/linux/kernel/git/mark/linux.git/commit/?h=fuzzing/5.10-rc7&id=61336e25d1415a4ac3aaf8cf75105c2ec2eb95e7

I'll try to respin that in the next few days if I get the chance.

Thanks,
Mark.

> 
> >  #define safe_halt()		do { raw_safe_halt(); } while (0)
> >  
> >  #endif /* CONFIG_TRACE_IRQFLAGS */
> 
> --- a/include/linux/irqflags.h
> +++ b/include/linux/irqflags.h
> @@ -162,6 +162,7 @@ do {						\
>  #define raw_local_irq_restore(flags)			\
>  	do {						\
>  		typecheck(unsigned long, flags);	\
> +		check_bogus_irq_restore();		\
>  		arch_local_irq_restore(flags);		\
>  	} while (0)
>  #define raw_local_save_flags(flags)			\
> @@ -235,11 +236,7 @@ extern void warn_bogus_irq_restore(bool
>  #define local_irq_enable()	do { raw_local_irq_enable(); } while (0)
>  #define local_irq_disable()	do { raw_local_irq_disable(); } while (0)
>  #define local_irq_save(flags)	do { raw_local_irq_save(flags); } while (0)
> -#define local_irq_restore(flags)		\
> -	do {					\
> -		check_bogus_irq_restore();	\
> -		raw_local_irq_restore(flags);	\
> -	} while (0)
> +#define local_irq_restore(flags) do { raw_local_irq_restore(flags); } while (0)
>  #define safe_halt()		do { raw_safe_halt(); } while (0)
>  
>  #endif /* CONFIG_TRACE_IRQFLAGS */