Hi Jean,
On Fri, Aug 26, 2011 at 07:15:52AM -0400, Jean Delvare wrote:
> Hi Guenter,
>
> On Thu, 25 Aug 2011 19:12:34 -0700, Guenter Roeck wrote:
> > ADM1275 supports a second current limit, which can be configured as either lower
> > or upper limit. Add support for it and report it as either lower or upper
> > critical current limit.
> >
> > Signed-off-by: Guenter Roeck <guenter.roeck@xxxxxxxxxxxx>
> > ---
> > Documentation/hwmon/adm1275 | 8 ++++
> > drivers/hwmon/pmbus/adm1275.c | 88 ++++++++++++++++++++++++++++++++++++++--
> > 2 files changed, 91 insertions(+), 5 deletions(-)
> >
> > diff --git a/Documentation/hwmon/adm1275 b/Documentation/hwmon/adm1275
> > index 097b3cc..c438c98 100644
> > --- a/Documentation/hwmon/adm1275
> > +++ b/Documentation/hwmon/adm1275
> > @@ -60,5 +60,13 @@ curr1_label "iout1"
> > curr1_input Measured current. From READ_IOUT register.
> > curr1_max Maximum current. From IOUT_OC_WARN_LIMIT register.
> > curr1_max_alarm Current high alarm. From IOUT_OC_WARN_LIMIT register.
> > +curr1_lcrit Critical minimum current. Depending on the chip
> > + configuration, either curr1_lcrit or curr1_crit is
> > + supported, but not both.
> > +curr1_lcrit_alarm Critical current low alarm.
> > +curr1_crit Critical maximum current. Depending on the chip
> > + configuration, either curr1_lcrit or curr1_crit is
> > + supported, but not both.
> > +curr1_crit_alarm Critical current high alarm.
> > curr1_highest Historical maximum current.
> > curr1_reset_history Write any value to reset history.
> > diff --git a/drivers/hwmon/pmbus/adm1275.c b/drivers/hwmon/pmbus/adm1275.c
> > index c936e27..061e7e7 100644
> > --- a/drivers/hwmon/pmbus/adm1275.c
> > +++ b/drivers/hwmon/pmbus/adm1275.c
> > @@ -31,14 +31,44 @@
> > #define ADM1275_VIN_VOUT_SELECT (1 << 6)
> > #define ADM1275_VRANGE (1 << 5)
> >
> > +#define ADM1275_IOUT_WARN2_LIMIT 0xd7
> > +#define ADM1275_DEVICE_CONFIG 0xd8
> > +
> > +#define ADM1275_IOUT_WARN2_SELECT (1 << 4)
> > +
> > +#define ADM1275_MFR_STATUS_IOUT_WARN2 (1 << 0)
> > +
> > +struct adm1275_data {
> > + bool have_oc_fault;
> > + struct pmbus_driver_info info;
> > +};
> > +
> > +#define to_adm1275_data(x) container_of(x, struct adm1275_data, info)
> > +
> > static int adm1275_read_word_data(struct i2c_client *client, int page, int reg)
> > {
> > + const struct pmbus_driver_info *info = pmbus_get_driver_info(client);
> > + const struct adm1275_data *data = to_adm1275_data(info);
> > int ret;
> >
> > if (page)
> > return -EINVAL;
> >
> > switch (reg) {
> > + case PMBUS_IOUT_UC_FAULT_LIMIT:
> > + if (data->have_oc_fault) {
> > + ret = -EINVAL;
> > + break;
> > + }
> > + ret = pmbus_read_word_data(client, 0, ADM1275_IOUT_WARN2_LIMIT);
> > + break;
> > + case PMBUS_IOUT_OC_FAULT_LIMIT:
> > + if (!data->have_oc_fault) {
> > + ret = -EINVAL;
> > + break;
> > + }
> > + ret = pmbus_read_word_data(client, 0, ADM1275_IOUT_WARN2_LIMIT);
> > + break;
>
> I am not familiar with the pmbus layer (is it documented anywhere?), is
> it possible for these errors to happen at all? I am a little surprised
> that you return errors here and not in adm1275_write_word_data below.
> But maybe it's OK.
>
> If you really have to return these errors, then why do you return
> -EINVAL when other unsupported features get -ENODATA?
>
Guess I'll need to document the logic here.
EINVAL:
The calling code does not try to access the real register and returns the error
to the caller. Not sure about EINVAL here; maybe I should return EIO.
ENODATA:
There is no chip specific register to return this data, but there may be
a standard register. The calling code tries to access the standard register.
> > case PMBUS_VIRT_READ_IOUT_MAX:
> > ret = pmbus_read_word_data(client, 0, ADM1275_PEAK_IOUT);
> > break;
> > @@ -69,6 +99,11 @@ static int adm1275_write_word_data(struct i2c_client *client, int page, int reg,
> > return -EINVAL;
> >
> > switch (reg) {
> > + case PMBUS_IOUT_UC_FAULT_LIMIT:
> > + case PMBUS_IOUT_OC_FAULT_LIMIT:
> > + ret = pmbus_write_word_data(client, 0, ADM1275_IOUT_WARN2_LIMIT,
> > + word);
> > + break;
> > case PMBUS_VIRT_RESET_IOUT_HISTORY:
> > ret = pmbus_write_word_data(client, 0, ADM1275_PEAK_IOUT, 0);
> > break;
> > @@ -85,19 +120,49 @@ static int adm1275_write_word_data(struct i2c_client *client, int page, int reg,
> > return ret;
> > }
> >
> > +static int adm1275_read_byte_data(struct i2c_client *client, int page, int reg)
> > +{
> > + const struct pmbus_driver_info *info = pmbus_get_driver_info(client);
> > + const struct adm1275_data *data = to_adm1275_data(info);
> > + int mfr_status, ret;
>
> No check for "page" as you have in adm1275_read_word_data() and
> adm1275_write_word_data()?
>
Oversight. I'll add it.
> > +
> > + switch (reg) {
> > + case PMBUS_STATUS_IOUT:
> > + ret = pmbus_read_byte_data(client, page, PMBUS_STATUS_IOUT);
> > + if (ret < 0)
> > + break;
> > + mfr_status = pmbus_read_byte_data(client, page,
> > + PMBUS_STATUS_MFR_SPECIFIC);
> > + if (mfr_status < 0) {
> > + ret = mfr_status;
> > + break;
> > + }
> > + if (mfr_status & ADM1275_MFR_STATUS_IOUT_WARN2) {
> > + ret |= data->have_oc_fault ?
> > + PB_IOUT_OC_FAULT : PB_IOUT_UC_FAULT;
> > + }
> > + break;
> > + default:
> > + ret = -ENODATA;
> > + break;
> > + }
> > + return ret;
> > +}
> > +
> > static int adm1275_probe(struct i2c_client *client,
> > const struct i2c_device_id *id)
> > {
> > - int config;
> > + int config, device_config;
> > int ret;
> > struct pmbus_driver_info *info;
> > + struct adm1275_data *data;
> >
> > if (!i2c_check_functionality(client->adapter,
> > I2C_FUNC_SMBUS_READ_BYTE_DATA))
> > return -ENODEV;
> >
> > - info = kzalloc(sizeof(struct pmbus_driver_info), GFP_KERNEL);
> > - if (!info)
> > + data = kzalloc(sizeof(struct adm1275_data), GFP_KERNEL);
> > + if (!data)
> > return -ENOMEM;
> >
> > config = i2c_smbus_read_byte_data(client, ADM1275_PMON_CONFIG);
> > @@ -106,6 +171,14 @@ static int adm1275_probe(struct i2c_client *client,
> > goto err_mem;
> > }
> >
> > + device_config = i2c_smbus_read_byte_data(client, ADM1275_DEVICE_CONFIG);
> > + if (device_config < 0) {
> > + ret = device_config;
> > + goto err_mem;
> > + }
> > +
> > + info = &data->info;
> > +
> > info->pages = 1;
> > info->format[PSC_VOLTAGE_IN] = direct;
> > info->format[PSC_VOLTAGE_OUT] = direct;
> > @@ -116,6 +189,7 @@ static int adm1275_probe(struct i2c_client *client,
> > info->func[0] = PMBUS_HAVE_IOUT | PMBUS_HAVE_STATUS_IOUT;
> >
> > info->read_word_data = adm1275_read_word_data;
> > + info->read_byte_data = adm1275_read_byte_data;
> > info->write_word_data = adm1275_write_word_data;
> >
> > if (config & ADM1275_VRANGE) {
> > @@ -134,6 +208,9 @@ static int adm1275_probe(struct i2c_client *client,
> > info->R[PSC_VOLTAGE_OUT] = -1;
> > }
> >
> > + if (device_config & ADM1275_IOUT_WARN2_SELECT)
> > + data->have_oc_fault = true;
> > +
> > if (config & ADM1275_VIN_VOUT_SELECT)
> > info->func[0] |= PMBUS_HAVE_VOUT | PMBUS_HAVE_STATUS_VOUT;
> > else
> > @@ -145,17 +222,18 @@ static int adm1275_probe(struct i2c_client *client,
> > return 0;
> >
> > err_mem:
> > - kfree(info);
> > + kfree(data);
> > return ret;
> > }
> >
> > static int adm1275_remove(struct i2c_client *client)
> > {
> > const struct pmbus_driver_info *info = pmbus_get_driver_info(client);
> > + const struct adm1275_data *data = to_adm1275_data(info);
> > int ret;
> >
> > ret = pmbus_do_remove(client);
> > - kfree(info);
> > + kfree(data);
> > return ret;
> > }
>
> Unrelated to this patch, but this error path seems wrong. If
> pmbus_do_remove() returns an error, then presumably the device is not
> actually removed. As you return the error value up to the caller, the
> device will not be considered unbound from its driver either. In these
> conditions, freeing the data structure doesn't seem right.
>
> Thankfully it should never happen. It's even questionable why remove
> functions can return an error value at all...
>
In practive pmbus_do_remove returns what all other driver remove functions do,
ie it always returns 0. But you are right, the path is wrong. I'll submit a separate
set of patches to address this - probably I'll make pmbus_do_remove a void function.
Thanks,
Guenter
_______________________________________________
lm-sensors mailing list
lm-sensors@xxxxxxxxxxxxxx
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
