linux-kernel - Re: [PATCH 09/10] x86/rcu: Add rcu_preempt

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]

Date: Tue, 23 Apr 2024 17:02:35 +0800
From: Lai Jiangshan <jiangshanlai@...il.com>
To: Frederic Weisbecker <frederic@...nel.org>
Cc: linux-kernel@...r.kernel.org, rcu@...r.kernel.org, x86@...nel.org, 
	Lai Jiangshan <jiangshan.ljs@...group.com>, "Paul E. McKenney" <paulmck@...nel.org>, 
	Peter Zijlstra <peterz@...radead.org>, Thomas Gleixner <tglx@...utronix.de>, 
	Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>, 
	Dave Hansen <dave.hansen@...ux.intel.com>, "H. Peter Anvin" <hpa@...or.com>, 
	Uros Bizjak <ubizjak@...il.com>, Josh Poimboeuf <jpoimboe@...nel.org>, Nadav Amit <namit@...are.com>, 
	Breno Leitao <leitao@...ian.org>, Kent Overstreet <kent.overstreet@...ux.dev>, 
	Pawan Gupta <pawan.kumar.gupta@...ux.intel.com>, 
	Rick Edgecombe <rick.p.edgecombe@...el.com>, Vegard Nossum <vegard.nossum@...cle.com>, 
	Daniel Sneddon <daniel.sneddon@...ux.intel.com>, Nikolay Borisov <nik.borisov@...e.com>
Subject: Re: [PATCH 09/10] x86/rcu: Add rcu_preempt_count

Hello, Frederic

Thanks for reviewing.

On Mon, Apr 22, 2024 at 7:05 PM Frederic Weisbecker <frederic@...nel.org> wrote:

> > +
> > +/* We use the MSB mostly because its available */
>
> I think you can safely remove the "We " from all the comments :-)

The file is mainly copied from arch/x86/include/asm/preempt.h.
I will rephrase sentences in later iterations.

>
> > +#define RCU_PREEMPT_UNLOCK_SPECIAL_INVERTED  0x80000000
>
> How about RCU_PREEMPT_UNLOCK_FASTPATH ?


I'm not good at naming. But the MSB really means exactly the opposite
of current->rcu_read_unlock_special and I think "UNLOCK_SPECIAL_INVERTED"
fits the meaning.

>
> > +
> > +/*
> > + * We use the RCU_PREEMPT_UNLOCK_SPECIAL_INVERTED bit as an inverted
> > + * current->rcu_read_unlock_special.s such that a decrement hitting 0
> > + * means we can and should call rcu_read_unlock_special().
> > + */
> > +#define RCU_PREEMPT_INIT     (0 + RCU_PREEMPT_UNLOCK_SPECIAL_INVERTED)
>
> Or simply:
>
> #define RCU_PREEMPT_INIT     RCU_PREEMPT_UNLOCK_FASTPATH
>
> Or you can even remove RCU_PREEMPT_INIT and use RCU_PREEMPT_UNLOCK_FASTPATH directly.

"0" means the initial rcu_preempt_count is 0 for the initial task.

> > +
> > +#endif // #ifdef CONFIG_PCPU_RCU_PREEMPT_COUNT
> > +
> > +#endif /* __ASM_RCU_PREEMPT_H */
> > diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
> > index ba8cf5e9ce56..0b204a649442 100644
> > --- a/arch/x86/kernel/cpu/common.c
> > +++ b/arch/x86/kernel/cpu/common.c
> > @@ -1992,9 +1992,10 @@ static __init int setup_clearcpuid(char *arg)
> >  __setup("clearcpuid=", setup_clearcpuid);
> >
> >  DEFINE_PER_CPU_ALIGNED(struct pcpu_hot, pcpu_hot) = {
> > -     .current_task   = &init_task,
> > -     .preempt_count  = INIT_PREEMPT_COUNT,
> > -     .top_of_stack   = TOP_OF_INIT_STACK,
> > +     .current_task           = &init_task,
> > +     .preempt_count          = INIT_PREEMPT_COUNT,
> > +     .top_of_stack           = TOP_OF_INIT_STACK,
> > +     .rcu_preempt_count      = RCU_PREEMPT_INIT,
>
> #ifdef CONFIG_PCPU_RCU_PREEMPT_COUNT ?
>
> Thanks.

Fixed in V2:
https://lore.kernel.org/lkml/20240407090558.3395-1-jiangshanlai@gmail.com/

Thanks
Lai

>
>
> >  };
> >  EXPORT_PER_CPU_SYMBOL(pcpu_hot);
> >  EXPORT_PER_CPU_SYMBOL(const_pcpu_hot);
> > --
> > 2.19.1.6.gb485710b
> >