Re: [PATCH 2/2] MIPS: Convert update_persistent_clock() to update_persistent_clock64()

From: Baolin Wang <baolin.wang@linaro.org>
To: Arnd Bergmann <arnd@arndb.de>
Cc: "Maciej W. Rozycki" <macro@linux-mips.org>,
	Ralf Baechle <ralf@linux-mips.org>,
	James Hogan <jhogan@kernel.org>,
	chenhc@lemote.com, Kate Stewart <kstewart@linuxfoundation.org>,
	gregkh <gregkh@linuxfoundation.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	Philippe Ombredanne <pombredanne@nexb.com>,
	Mark Brown <broonie@kernel.org>,
	Paul Burton <paul.burton@mips.com>,
	Heiko Stuebner <heiko@sntech.de>,
	Daniel Lezcano <daniel.lezcano@linaro.org>,
	Viresh Kumar <viresh.kumar@linaro.org>,
	"open list:RALINK MIPS ARCHITECTURE" <linux-mips@linux-mips.org>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 2/2] MIPS: Convert update_persistent_clock() to update_persistent_clock64()
Date: Fri, 4 May 2018 14:23:13 +0800	[thread overview]
Message-ID: <CAMz4ku+WbtBGGcmsYmMexxcBTJONHF3J-T8Sxo7wgHoZZCVt1w@mail.gmail.com> (raw)
In-Reply-To: <CAK8P3a3TJ-5+22_CSYjdtL3pXEVHC7t_KzaOX6PG4X_1R1bMxQ@mail.gmail.com>

On 4 May 2018 at 06:31, Arnd Bergmann <arnd@arndb.de> wrote:
> On Wed, May 2, 2018 at 10:53 PM, Baolin Wang <baolin.wang@linaro.org> wrote:
>> diff --git a/arch/mips/include/asm/time.h b/arch/mips/include/asm/time.h
>> index 17d4cd2..c4e2a1a 100644
>> --- a/arch/mips/include/asm/time.h
>> +++ b/arch/mips/include/asm/time.h
>> @@ -27,8 +27,8 @@
>>   *     rtc_mips_set_mmss - similar to rtc_set_time, but only min and sec need
>>   *                     to be set.  Used by RTC sync-up.
>>   */
>> -extern int rtc_mips_set_time(unsigned long);
>> -extern int rtc_mips_set_mmss(unsigned long);
>> +extern int rtc_mips_set_time(time64_t);
>> +extern int rtc_mips_set_mmss(time64_t);
>>
>
> I think these should just get removed, and each implementation replaced
> with a direct update_persistent_clock64() function.

I thought this was one minor modification that will reduce the risk of
introducing other issues, but as you suggested we can do some complete
cleanup by removing set_mmss/set_time. OK, I will do that.

>
>> -int update_persistent_clock(struct timespec now)
>> +int update_persistent_clock64(struct timespec64 now)
>>  {
>>         return rtc_mips_set_mmss(now.tv_sec);
>>  }
>
> And this one also removed

Sure.

>
>> @@ -69,7 +69,7 @@ int proc_dolasatrtc(struct ctl_table *table, int write,
>>                 if (rtctmp < 0)
>>                         rtctmp = 0;
>>         }
>> -       r = proc_dointvec(table, write, buffer, lenp, ppos);
>> +       r = proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
>>         if (r)
>>                 return r;
>>
>> @@ -224,7 +224,7 @@ int proc_lasat_prid(struct ctl_table *table, int write,
>>         {
>>                 .procname       = "rtc",
>>                 .data           = &rtctmp,
>> -               .maxlen         = sizeof(int),
>> +               .maxlen         = sizeof(time64_t),
>>                 .mode           = 0644,
>>                 .proc_handler   = proc_dolasatrtc,
>>         },
>
> Something seems wrong here: time64_t is not the same as 'unsigned long',
> and the 'rtctmp' variable is still 'unsigned int'. Not sure what the right fix
> would be (we don't seem to have a sysctl helper for s64), but the change
> here makes it worse.

After checking again, I agree with you. So I will keep the original code here.

-- 
Baolin.wang
Best Regards