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] powerpc/time: use feature fixup in __USE_RTC() instead of
 cpu feature.
To:     Michael Ellerman <mpe@ellerman.id.au>,
        Benjamin Herrenschmidt <benh@kernel.crashing.org>,
        Paul Mackerras <paulus@samba.org>
Cc:     linux-kernel@vger.kernel.org, linuxppc-dev@lists.ozlabs.org
References: <55c267ac6e0cd289970accfafbf9dda11a324c2e.1566802736.git.christophe.leroy@c-s.fr>
 <87blwc40i4.fsf@concordia.ellerman.id.au>
From:   Christophe Leroy <christophe.leroy@c-s.fr>
Message-ID: <ba2a7a1f-8dcc-ec44-81d4-ef11a9e29377@c-s.fr>
Date:   Mon, 26 Aug 2019 13:48:41 +0200
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:60.0) Gecko/20100101
 Thunderbird/60.8.0
MIME-Version: 1.0
In-Reply-To: <87blwc40i4.fsf@concordia.ellerman.id.au>
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


Le 26/08/2019 à 13:41, Michael Ellerman a écrit :
> Christophe Leroy <christophe.leroy@c-s.fr> writes:
>> sched_clock(), used by printk(), calls __USE_RTC() to know
>> whether to use realtime clock or timebase.
>>
>> __USE_RTC() uses cpu_has_feature() which is initialised by
>> machine_init(). Before machine_init(), __USE_RTC() returns true,
>> leading to a program check exception on CPUs not having realtime
>> clock.
>>
>> In order to be able to use printk() earlier, use feature fixup.
>> Feature fixups are applies in early_init(), enabling the use of
>> printk() earlier.
>>
>> Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
>> ---
>>   arch/powerpc/include/asm/time.h | 9 ++++++++-
>>   1 file changed, 8 insertions(+), 1 deletion(-)
> 
> The other option would be just to make this a compile time decision, eg.
> add CONFIG_PPC_601 and use that to gate whether we use RTC.

Euh ... yes OK, why not. And that would help simplify many places in the 
code. I can propose something in that direction, but it will be a bigger 
change.


> 
> Given how many 601 users there are, maybe 1?, I think that would be a
> simpler option and avoids complicating the code / binary for everyone
> else.

However this patch doesn't complicate things more than it was with 
cpu_has_feature() which does exactly the same but using static keys, 
does it ?

Christophe


> 
> cheers
> 
>> diff --git a/arch/powerpc/include/asm/time.h b/arch/powerpc/include/asm/time.h
>> index 54f4ec1f9fab..3455cb54c333 100644
>> --- a/arch/powerpc/include/asm/time.h
>> +++ b/arch/powerpc/include/asm/time.h
>> @@ -42,7 +42,14 @@ struct div_result {
>>   /* Accessor functions for the timebase (RTC on 601) registers. */
>>   /* If one day CONFIG_POWER is added just define __USE_RTC as 1 */
>>   #ifdef CONFIG_PPC_BOOK3S_32
>> -#define __USE_RTC()	(cpu_has_feature(CPU_FTR_USE_RTC))
>> +static inline bool __USE_RTC(void)
>> +{
>> +	asm_volatile_goto(ASM_FTR_IFCLR("nop;", "b %1;", %0) ::
>> +			  "i" (CPU_FTR_USE_RTC) :: l_use_rtc);
>> +	return false;
>> +l_use_rtc:
>> +	return true;
>> +}
>>   #else
>>   #define __USE_RTC()	0
>>   #endif
>> -- 
>> 2.13.3