linux-kernel - Re: [PATCH] rtc: rs5c372: Add RTC_VL_READ, RTC_VL

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 for Android: free password hash cracker in your pocket
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <2303e635-dbd0-1730-cc6f-84021eb37223@axis.com>
Date:   Wed, 10 Nov 2021 15:03:49 +0100
From:   Camel Guo <camelg@...s.com>
To:     Alexandre Belloni <alexandre.belloni@...tlin.com>,
        Camel Guo <Camel.Guo@...s.com>
CC:     Alessandro Zummo <a.zummo@...ertech.it>, kernel <kernel@...s.com>,
        "linux-rtc@...r.kernel.org" <linux-rtc@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] rtc: rs5c372: Add RTC_VL_READ, RTC_VL_CLR ioctls



On 11/10/21 2:54 PM, Alexandre Belloni wrote:
> Hello,
> 
> On 10/11/2021 12:54:54+0100, Camel Guo wrote:
>> From: Camel Guo <camelg@...s.com>
>> 
>> In order to make it possible to get battery voltage status, this commit
>> adds RTC_VL_READ, RTC_VL_CLR ioctl commands to rtc-rs5c372.
>> 
>> Signed-off-by: Camel Guo <camelg@...s.com>
>> ---
>>  drivers/rtc/rtc-rs5c372.c | 66 +++++++++++++++++++++++++++++++++++++++
>>  1 file changed, 66 insertions(+)
>> 
>> diff --git a/drivers/rtc/rtc-rs5c372.c b/drivers/rtc/rtc-rs5c372.c
>> index 80980414890c..5a96e5d3663a 100644
>> --- a/drivers/rtc/rtc-rs5c372.c
>> +++ b/drivers/rtc/rtc-rs5c372.c
>> @@ -485,6 +485,71 @@ static int rs5c372_rtc_proc(struct device *dev, struct seq_file *seq)
>>  #define      rs5c372_rtc_proc        NULL
>>  #endif
>>  
>> +#ifdef CONFIG_RTC_INTF_DEV
>> +static int rs5c372_ioctl(struct device *dev, unsigned int cmd, unsigned long arg)
>> +{
>> +     struct rs5c372  *rs5c = i2c_get_clientdata(to_i2c_client(dev));
>> +     unsigned char   ctrl2;
>> +     int             addr;
>> +     unsigned int    flags;
>> +
>> +     dev_dbg(dev, "%s: cmd=%x\n", __func__, cmd);
>> +
>> +     addr = RS5C_ADDR(RS5C_REG_CTRL2);
>> +     ctrl2 = i2c_smbus_read_byte_data(rs5c->client, addr);
>> +
>> +     switch (cmd) {
>> +     case RTC_VL_READ:
>> +             flags = 0;
>> +
>> +             switch (rs5c->type) {
>> +             case rtc_r2025sd:
>> +             case rtc_r2221tl:
>> +                     if ((rs5c->type == rtc_r2025sd && !(ctrl2 & R2x2x_CTRL2_XSTP)) ||
>> +                             (rs5c->type == rtc_r2221tl &&  (ctrl2 & R2x2x_CTRL2_XSTP))) {
>> +                             flags |= RTC_VL_DATA_INVALID;
>> +                     }
>> +                     if (ctrl2 & R2x2x_CTRL2_VDET)
>> +                             flags |= RTC_VL_ACCURACY_LOW;
> 
> Shouldn't that be RTC_VL_BACKUP_LOW?

Some drivers (e.g: rv3029_ioctl and rv8803_ioctl) use 
RTC_VL_ACCURACY_LOW, but some other drivers (e.g: abx80x_ioctl, 
pcf2127_rtc_ioctl and pcf8523_rtc_ioctl) use RTC_VL_BACKUP_LOW instead. 
Is there any guideline or document telling the differences between them?

I can change it to RTC_VL_BACKUP_LOW of course.

> 
>> +                     break;
>> +             default:
>> +                     if (ctrl2 & RS5C_CTRL2_XSTP)
>> +                             flags |= RTC_VL_DATA_INVALID;
>> +                     break;
>> +             }
>> +
>> +             return put_user(flags, (unsigned int __user *)arg);
>> +     case RTC_VL_CLR:
>> +             /* clear rtc VDET, PON and XSTP bits */
>> +             switch (rs5c->type) {
>> +             case rtc_r2025sd:
>> +             case rtc_r2221tl:
>> +                     ctrl2 &= ~(R2x2x_CTRL2_VDET | R2x2x_CTRL2_PON);
>> +                     if (rs5c->type == rtc_r2025sd)
>> +                             ctrl2 |= R2x2x_CTRL2_XSTP;
>> +                     else
>> +                             ctrl2 &= ~R2x2x_CTRL2_XSTP;
>> +                     break;
>> +             default:
>> +                     ctrl2 &= ~RS5C_CTRL2_XSTP;
> 
> You can clear VDET but you must nt clear PON or XSTP as they are used to
> know whether the time on the RTC has been set correctly. Clearing those
> can only be done in .set_time.

Will update shortly.

> 
>> +                     break;
>> +             }
>> +
>> +             if (i2c_smbus_write_byte_data(rs5c->client, addr, ctrl2) < 0) {
>> +                     dev_dbg(&rs5c->client->dev, "%s: write error in line %i\n",
>> +                                     __func__, __LINE__);
>> +                     return -EIO;
>> +             }
>> +             return 0;
>> +     default:
>> +             return -ENOIOCTLCMD;
>> +     }
>> +     return 0;
>> +}
>> +#else
>> +#define rs5c372_ioctl        NULL
>> +#endif
>> +
>>  static const struct rtc_class_ops rs5c372_rtc_ops = {
>>        .proc           = rs5c372_rtc_proc,
>>        .read_time      = rs5c372_rtc_read_time,
>> @@ -492,6 +557,7 @@ static const struct rtc_class_ops rs5c372_rtc_ops = {
>>        .read_alarm     = rs5c_read_alarm,
>>        .set_alarm      = rs5c_set_alarm,
>>        .alarm_irq_enable = rs5c_rtc_alarm_irq_enable,
>> +     .ioctl          = rs5c372_ioctl,
>>  };
>>  
>>  #if IS_ENABLED(CONFIG_RTC_INTF_SYSFS)
>> -- 
>> 2.20.1
>> 
> 
> -- 
> Alexandre Belloni, co-owner and COO, Bootlin
> Embedded Linux and Kernel engineering
> https://bootlin.com