Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [PATCH 1/1] arch/powerpc: Rework local_paca to avoid LTO warnings
To:     Alastair D'Silva <alastair@au1.ibm.com>
Cc:     Alastair D'Silva <alastair@d-silva.org>,
        Benjamin Herrenschmidt <benh@kernel.crashing.org>,
        Paul Mackerras <paulus@samba.org>,
        Michael Ellerman <mpe@ellerman.id.au>,
        Nicholas Piggin <npiggin@gmail.com>,
        Mahesh Salgaonkar <mahesh@linux.vnet.ibm.com>,
        "Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com>,
        Michal Hocko <mhocko@suse.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        Mike Rapoport <rppt@linux.vnet.ibm.com>,
        linuxppc-dev@lists.ozlabs.org, linux-kernel@vger.kernel.org
References: <20190313034208.13134-1-alastair@au1.ibm.com>
From:   Christophe Leroy <christophe.leroy@c-s.fr>
Message-ID: <bd858b12-7fab-057d-0065-2ccb98cbf648@c-s.fr>
Date:   Wed, 13 Mar 2019 10:06:21 +0100
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:60.0) Gecko/20100101
 Thunderbird/60.5.3
MIME-Version: 1.0
In-Reply-To: <20190313034208.13134-1-alastair@au1.ibm.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: fr
Content-Transfer-Encoding: 8bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Hello,

Le 13/03/2019 à 04:42, Alastair D'Silva a écrit :
> From: Alastair D'Silva <alastair@d-silva.org>
> 
> When building an LTO kernel, the existing code generates warnings:
>      ./arch/powerpc/include/asm/paca.h:37:30: warning: register of
>          ‘local_paca’ used for multiple global register variables
>       register struct paca_struct *local_paca asm("r13");
>                                ^
>      ./arch/powerpc/include/asm/paca.h:37:30: note: conflicts with
>          ‘local_paca’

How do you build a LTO kernel ?

> 
> This patch reworks local_paca into an inline getter & setter function,
> which addresses the warning.

This patch adds sparse warnings, see 
https://patchwork.ozlabs.org/patch/1055875/

> 
> Generated ASM from this patch is broadly similar (addresses have
> changed and the compiler uses different GPRs in some places).

Your text might be confusion. When I read it the first time I thought 
you were saying that the compiler was now using another GPR than r13.

> 
> Signed-off-by: Alastair D'Silva <alastair@d-silva.org>

I guess the same has to be done for current, see 
arch/powerpc/include/asm/current.h :

/*
  * We keep `current' in r2 for speed.
  */
register struct task_struct *current asm ("r2");

> ---
>   arch/powerpc/include/asm/paca.h | 44 +++++++++++++++++++++++----------
>   arch/powerpc/kernel/paca.c      |  2 +-
>   2 files changed, 32 insertions(+), 14 deletions(-)
> 
> diff --git a/arch/powerpc/include/asm/paca.h b/arch/powerpc/include/asm/paca.h
> index e843bc5d1a0f..9c9e2dea0f9b 100644
> --- a/arch/powerpc/include/asm/paca.h
> +++ b/arch/powerpc/include/asm/paca.h
> @@ -34,19 +34,6 @@
>   #include <asm/cpuidle.h>
>   #include <asm/atomic.h>
>   
> -register struct paca_struct *local_paca asm("r13");
> -
> -#if defined(CONFIG_DEBUG_PREEMPT) && defined(CONFIG_SMP)
> -extern unsigned int debug_smp_processor_id(void); /* from linux/smp.h */
> -/*
> - * Add standard checks that preemption cannot occur when using get_paca():
> - * otherwise the paca_struct it points to may be the wrong one just after.
> - */
> -#define get_paca()	((void) debug_smp_processor_id(), local_paca)
> -#else
> -#define get_paca()	local_paca
> -#endif
> -
>   #ifdef CONFIG_PPC_PSERIES
>   #define get_lppaca()	(get_paca()->lppaca_ptr)
>   #endif
> @@ -266,6 +253,37 @@ struct paca_struct {
>   #endif
>   } ____cacheline_aligned;
>   
> +#if defined(CONFIG_DEBUG_PREEMPT) && defined(CONFIG_SMP)
> +extern unsigned int debug_smp_processor_id(void); /* from linux/smp.h */
> +#endif

Why moving this down, why not leaving at the same place as before ?

If you really need to move it, you should remove the 'extern' at the 
same time to make checkpatch happy.

> +
> +static inline struct paca_struct *get_paca_no_preempt_check(void)
> +{
> +	register struct paca_struct *paca asm("r13");

Should be a blank line there.

> +	return paca;
> +}
> +
> +static inline struct paca_struct *get_paca(void)
> +{
> +#if defined(CONFIG_DEBUG_PREEMPT) && defined(CONFIG_SMP)
> +	/*
> +	 * Add standard checks that preemption cannot occur when using get_paca():
> +	 * otherwise the paca_struct it points to may be the wrong one just after.
> +	 */
> +	debug_smp_processor_id();
> +#endif
> +	return get_paca_no_preempt_check();
> +}
> +
> +#define local_paca	get_paca_no_preempt_check()
> +
> +static inline void set_paca(struct paca_struct *new)
> +{
> +	register struct paca_struct *paca asm("r13");

Blank line should be added here.

> +	paca = new;
> +}
> +
> +
>   extern void copy_mm_to_paca(struct mm_struct *mm);
>   extern struct paca_struct **paca_ptrs;
>   extern void initialise_paca(struct paca_struct *new_paca, int cpu);
> diff --git a/arch/powerpc/kernel/paca.c b/arch/powerpc/kernel/paca.c
> index 913bfca09c4f..ae5c243f9d5a 100644
> --- a/arch/powerpc/kernel/paca.c
> +++ b/arch/powerpc/kernel/paca.c
> @@ -172,7 +172,7 @@ void __init initialise_paca(struct paca_struct *new_paca, int cpu)
>   void setup_paca(struct paca_struct *new_paca)
>   {
>   	/* Setup r13 */
> -	local_paca = new_paca;
> +	set_paca(new_paca);
>   
>   #ifdef CONFIG_PPC_BOOK3E
>   	/* On Book3E, initialize the TLB miss exception frames */
> 

Christophe