linux-kernel - Re: [PATCH 2/3] rtc: Add support for the MediaTek MT2712 RTC

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:   Wed, 17 Jul 2019 11:06:55 +0200
From:   Alexandre Belloni <alexandre.belloni@...tlin.com>
To:     Ran Bi <ran.bi@...iatek.com>
Cc:     Rob Herring <robh+dt@...nel.org>,
        Matthias Brugger <matthias.bgg@...il.com>,
        Alessandro Zummo <a.zummo@...ertech.it>,
        Mark Rutland <mark.rutland@....com>, linux-rtc@...r.kernel.org,
        devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-arm-kernel@...ts.infradead.org,
        linux-mediatek@...ts.infradead.org, srv_heupstream@...iatek.com,
        YT Shen <yt.shen@...iatek.com>,
        Eddie Huang <eddie.huang@...iatek.com>,
        Yingjoe Chen <yingjoe.chen@...iatek.com>,
        Flora Fu <flora.fu@...iatek.com>,
        Sean Wang <sean.wang@...iatek.com>
Subject: Re: [PATCH 2/3] rtc: Add support for the MediaTek MT2712 RTC

On 17/07/2019 16:54:54+0800, Ran Bi wrote:
> > > +
> > > +/* we map HW YEAR 0 to 1968 not 1970 because 2000 is the leap year */
> > > +#define RTC_MIN_YEAR		1968
> > > +#define RTC_BASE_YEAR		1900
> > > +#define RTC_MIN_YEAR_OFFSET	(RTC_MIN_YEAR - RTC_BASE_YEAR)
> > 
> > Do not do that. If this RTC range starts in 200, ths is what the driver
> > has to support, you should not care about dates before 2000. Note that
> > the RTC core can still properly shift the range if it is absolutely
> > necessary.
> > 
> 
> Do we need to care about default alarm date 1970-01-01? Or can I just
> set it to 2000-01-01?
> 

You never have to set a default value. It doesn't add any value versus
an unknown value.

> > > +
> > > +static inline u32 rtc_readl(struct mt2712_rtc *rtc, u32 reg)
> > 
> > Please use a more descriptive prefix than just rtc_.
> > 
> 
> Do you mean it's better to use prefix "mt2712_rtc_"?
> 

Yes.

> > > +
> > > +	/*
> > > +	 * register status was not correct,
> > > +	 * need set time and alarm to default
> > > +	 */
> > > +	if (p1 != RTC_POWERKEY1_KEY || p2 != RTC_POWERKEY2_KEY
> > > +	    || !valid_rtc_time(rtc)) {
> > > +		reset_rtc_time(rtc);
> > 
> > Do not do that. This is valuable information. If the time is invalid,
> > report it as such in read_time and read_alarm. Resetting the time here
> > will lead to more issues later (i.e. userspace is not able to know
> > whether the time is set correctly or not).
> > 
> 
> When RTC's power run out, RTC will lost it's registers value and time
> data at next boot up. We even cannot know what the date and time it
> shows. We want to check this state here and set a default RTC date. Do
> you think it's no need here and the date should be set by system?
> 

If I understand correctly, the POWERKEY register will lose their value.
This means that you know that the time is incorrect. instead of setting
it to a default value and losing that valuable information, simply check
for that in read_time and return EINVAL in that case. then on the next
set_time invocation, you can set the POWERKEY registers and set the time
to a known value.


-- 
Alexandre Belloni, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com