Re: [PATCH] i8k: Add support for temperature sensor labels

[Pali RohÃr]

On Saturday 29 November 2014 18:51:52 Greg Kroah-Hartman wrote:
> On Sat, Nov 29, 2014 at 06:49:43PM +0100, Pali RohÃr wrote:
> > On Saturday 29 November 2014 18:43:15 Greg Kroah-Hartman 
wrote:
> > > On Sat, Nov 29, 2014 at 05:04:07PM +0100, Pali RohÃr wrote:
> > > > This patch adds labels for temperature sensors if SMM
> > > > function with EAX register 0x11a3 reports it. These
> > > > informations was taken from DOS binary NBSVC.MDM.
> > > > 
> > > > Signed-off-by: Pali RohÃr <pali.rohar@xxxxxxxxx>
> > > > ---
> > > > 
> > > >  drivers/char/i8k.c |  110
> > > >  +++++++++++++++++++++++++++++++++++++++++----------- 1
> > > >  file changed, 88 insertions(+), 22 deletions(-)
> > > > 
> > > > diff --git a/drivers/char/i8k.c b/drivers/char/i8k.c
> > > > index e34a019..77af46b 100644
> > > > --- a/drivers/char/i8k.c
> > > > +++ b/drivers/char/i8k.c
> > > > @@ -42,6 +42,7 @@
> > > > 
> > > >  #define I8K_SMM_GET_FAN		0x00a3
> > > >  #define I8K_SMM_GET_SPEED	0x02a3
> > > >  #define I8K_SMM_GET_TEMP	0x10a3
> > > > 
> > > > +#define I8K_SMM_GET_TEMP_TYPE	0x11a3
> > > > 
> > > >  #define I8K_SMM_GET_DELL_SIG1	0xfea3
> > > >  #define I8K_SMM_GET_DELL_SIG2	0xffa3
> > > > 
> > > > @@ -288,6 +289,14 @@ static int i8k_set_fan(int fan, int
> > > > speed)
> > > > 
> > > >  	return i8k_smm(&regs) ? : i8k_get_fan_status(fan);
> > > >  
> > > >  }
> > > > 
> > > > +static int i8k_get_temp_type(int sensor)
> > > > +{
> > > > +	struct smm_regs regs = { .eax = I8K_SMM_GET_TEMP_TYPE,
> > > > }; +
> > > > +	regs.ebx = sensor & 0xff;
> > > > +	return i8k_smm(&regs) ? : regs.eax & 0xff;
> > > > +}
> > > > +
> > > > 
> > > >  /*
> > > >  
> > > >   * Read the cpu temperature.
> > > >   */
> > > > 
> > > > @@ -493,6 +502,29 @@ static int i8k_open_fs(struct inode
> > > > *inode, struct file *file)
> > > > 
> > > >   * Hwmon interface
> > > >   */
> > > > 
> > > > +static ssize_t i8k_hwmon_show_temp_label(struct device
> > > > *dev, +				   struct device_attribute *devattr,
> > > > +				   char *buf)
> > > > +{
> > > > +	static const char * const labels[] = {
> > > > +		"CPU",
> > > > +		"GPU",
> > > > +		"SODIMM",
> > > > +		"Other",
> > > > +		"Ambient",
> > > > +		"Other",
> > > > +	};
> > > > +	int index = to_sensor_dev_attr(devattr)->index;
> > > > +	int type;
> > > > +
> > > > +	type = i8k_get_temp_type(index);
> > > > +	if (type < 0)
> > > > +		return type;
> > > > +	if (type >= ARRAY_SIZE(labels))
> > > > +		type = ARRAY_SIZE(labels) - 1;
> > > > +	return sprintf(buf, "%s\n", labels[type]);
> > > > +}
> > > 
> > > No Documentation/ABI/ entry for your new sysfs file?
> > 
> > It is standard hwmon sysfs entry which is used by other
> > hwmon drivers... It is not i8k.ko driver specific.
> 
> Ok, then why is this driver not in a "hwmon-standard" location
> in the kernel tree?
> 
> greg k-h

This is question not for me but for maintainers of hwmon and char 
trees... I just contributed patch for existing driver.

But my idea is that because this driver exported temperature data 
for a long time only via /proc/i8k character file (since 2.4? or 
earlier?) and only 3 years ago was added standard hwmon interface 
to existing code (949a9d70020defd7c241607ab3ed037ea88f551c).

So now there are both interfaces. Standard hwmon and old /proc/.

CCing Jean Delvare (author of hwmon i8k code)

-- 
Pali RohÃr
pali.rohar@xxxxxxxxx

Attachment: signature.asc
Description: This is a digitally signed message part.