Re: [PATCH 2/2] hwmon: (ucd9000) Add debugfs attributes to provide mfr_status

[Guenter Roeck]

On Tue, Mar 13, 2018 at 02:01:51PM -0500, Eddie James wrote:
> 
> 
> On 03/10/2018 10:50 AM, Guenter Roeck wrote:
> >On 03/09/2018 11:19 AM, Eddie James wrote:
> >>From: Christopher Bostic <cbostic@xxxxxxxxxxxxxxxxxx>
> >>
> >>Expose the gpiN_fault fields of mfr_status as individual debugfs
> >>attributes. This provides a way for users to be easily notified of gpi
> >>faults. Also provide the whole mfr_status register in debugfs.
> >>
> >>Signed-off-by: Christopher Bostic <cbostic@xxxxxxxxxxxxxxxxxx>
> >>Signed-off-by: Andrew Jeffery <andrew@xxxxxxxx>
> >>Signed-off-by: Eddie James <eajames@xxxxxxxxxxxxxxxxxx>
> >>---
> >>  drivers/hwmon/pmbus/ucd9000.c | 172
> >>+++++++++++++++++++++++++++++++++++++++++-
> >>  1 file changed, 171 insertions(+), 1 deletion(-)
> >>
> >>diff --git a/drivers/hwmon/pmbus/ucd9000.c
> >>b/drivers/hwmon/pmbus/ucd9000.c
> >>index e3a507f..297da0e 100644
> >>--- a/drivers/hwmon/pmbus/ucd9000.c
> >>+++ b/drivers/hwmon/pmbus/ucd9000.c
> >>@@ -19,6 +19,7 @@
> >>   * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
> >>   */
> >>  +#include <linux/debugfs.h>
> >>  #include <linux/kernel.h>
> >>  #include <linux/module.h>
> >>  #include <linux/of_device.h>
> >>@@ -36,6 +37,7 @@
> >>  #define UCD9000_NUM_PAGES        0xd6
> >>  #define UCD9000_FAN_CONFIG_INDEX    0xe7
> >>  #define UCD9000_FAN_CONFIG        0xe8
> >>+#define UCD9000_MFR_STATUS        0xf3
> >>  #define UCD9000_GPIO_SELECT        0xfa
> >>  #define UCD9000_GPIO_CONFIG        0xfb
> >>  #define UCD9000_DEVICE_ID        0xfd
> >>@@ -63,13 +65,22 @@
> >>  #define UCD901XX_NUM_GPIOS    26
> >>  #define UCD90910_NUM_GPIOS    26
> >>  +#define UCD9000_DEBUGFS_NAME_LEN    24
> >>+#define UCD9000_GPI_COUNT        8
> >>+
> >>  struct ucd9000_data {
> >>      u8 fan_data[UCD9000_NUM_FAN][I2C_SMBUS_BLOCK_MAX];
> >>      struct pmbus_driver_info info;
> >>      struct gpio_chip gpio;
> >>+    struct dentry *debugfs;
> >>  };
> >>  #define to_ucd9000_data(_info) container_of(_info, struct
> >>ucd9000_data, info)
> >>  +struct ucd9000_debugfs_entry {
> >>+    struct i2c_client *client;
> >>+    u8 index;
> >>+};
> >>+
> >>  static int ucd9000_get_fan_config(struct i2c_client *client, int fan)
> >>  {
> >>      int fan_config = 0;
> >>@@ -328,6 +339,156 @@ static int ucd9000_gpio_direction_output(struct
> >>gpio_chip *gc,
> >>                        val);
> >>  }
> >>  +#if IS_ENABLED(CONFIG_DEBUG_FS)
> >>+static int ucd9000_get_mfr_status(struct i2c_client *client, u8
> >>*buffer)
> >>+{
> >>+    int ret = pmbus_set_page(client, 0);
> >>+
> >>+    if (ret < 0)
> >>+        return ret;
> >>+
> >>+    /*
> >>+     * With the ucd90120 and ucd90124 devices, this command
> >>[MFR_STATUS]
> >>+     * is 2 bytes long (bits 0-15).  With the ucd90240 this command is
> >>5
> >>+     * bytes long.  With all other devices, it is 4 bytes long.
> >>+     */
> >>+    return i2c_smbus_read_block_data(client, UCD9000_MFR_STATUS,
> >>buffer);
> >>+}
> >>+
> >>+static int ucd9000_debugfs_show_mfr_status_bit(void *data, u64 *val)
> >>+{
> >>+    struct ucd9000_debugfs_entry *entry = data;
> >>+    struct i2c_client *client = entry->client;
> >>+    u8 buffer[4];
> >>+    int ret;
> >>+
> >>+    /*
> >>+     * This attribute is only created for devices that return 4 bytes
> >>for
> >>+     * status_mfr, so it's safe to call with 4-byte buffer.
> >>+     */
> >>+    ret = ucd9000_get_mfr_status(client, buffer);
> >>+    if (ret < 0) {
> >>+        dev_err(&client->dev, "Failed to read mfr status. rc:%d\n",
> >>+            ret);
> >>+
> >>+        return ret;
> >>+    }
> >>+
> >>+    /*
> >>+     * Attribute only created for devices with gpi fault bits at bits
> >>+     * 16-23, which is the second byte of the response.
> >>+     */
> >>+    *val = !!(buffer[1] & BIT(entry->index));
> >>+
> >>+    return 0;
> >>+}
> >>+DEFINE_DEBUGFS_ATTRIBUTE(ucd9000_debugfs_mfr_status_bit,
> >>+             ucd9000_debugfs_show_mfr_status_bit, NULL, "%1lld\n");
> >>+
> >>+static int ucd9000_debugfs_show_mfr_status_word2(void *data, u64 *val)
> >>+{
> >>+    struct i2c_client *client = data;
> >>+    __be16 buffer;
> >>+    int ret;
> >>+
> >>+    ret = ucd9000_get_mfr_status(client, (u8 *)&buffer);
> >>+    if (ret < 0) {
> >>+        dev_err(&client->dev, "Failed to read mfr status. rc:%d\n",
> >>+            ret);
> >>+
> >>+        return ret;
> >>+    }
> >>+
> >>+    *val = be16_to_cpu(buffer);
> >>+
> >>+    return 0;
> >>+}
> >>+DEFINE_DEBUGFS_ATTRIBUTE(ucd9000_debugfs_mfr_status_word2,
> >>+             ucd9000_debugfs_show_mfr_status_word2, NULL,
> >>+             "%04llx\n");
> >>+
> >>+static int ucd9000_debugfs_show_mfr_status_word4(void *data, u64 *val)
> >>+{
> >>+    struct i2c_client *client = data;
> >>+    __be32 buffer;
> >>+    int ret;
> >>+
> >>+    ret = ucd9000_get_mfr_status(client, (u8 *)&buffer);
> >>+    if (ret < 0) {
> >>+        dev_err(&client->dev, "Failed to read mfr status. rc:%d\n",
> >>+            ret);
> >>+
> >>+        return ret;
> >>+    }
> >>+
> >>+    *val = be32_to_cpu(buffer);
> >>+
> >>+    return 0;
> >>+}
> >>+DEFINE_DEBUGFS_ATTRIBUTE(ucd9000_debugfs_mfr_status_word4,
> >>+             ucd9000_debugfs_show_mfr_status_word4, NULL,
> >>+             "%08llx\n");
> >>+
> >>+static int ucd9000_init_debugfs(struct i2c_client *client,
> >>+                const struct i2c_device_id *mid,
> >>+                struct ucd9000_data *data)
> >>+{
> >>+    struct dentry *debugfs;
> >>+    struct ucd9000_debugfs_entry *entries;
> >>+    int i;
> >>+    char name[UCD9000_DEBUGFS_NAME_LEN];
> >>+
> >>+    debugfs = pmbus_get_debugfs_dir(client);
> >>+    if (!debugfs)
> >>+        return -ENOENT;
> >>+
> >>+    data->debugfs = debugfs_create_dir(client->name, debugfs);
> >>+    if (!data->debugfs)
> >>+        return -ENOENT;
> >>+
> >>+    /*
> >>+     * Of the chips this driver supports, only the UCD9090, UCD90160,
> >>+     * and UCD90910 report GPI faults in their MFR_STATUS register, so
> >>only
> >>+     * create the GPI fault debugfs attributes for those chips.
> >>+     */
> >>+    if (mid->driver_data == ucd9090 || mid->driver_data == ucd90160 ||
> >>+        mid->driver_data == ucd90910) {
> >>+        entries = devm_kzalloc(&client->dev,
> >>+                       sizeof(*entries) * UCD9000_GPI_COUNT,
> >>+                       GFP_KERNEL);
> >>+        if (!entries)
> >>+            return -ENOMEM;
> >>+
> >>+        for (i = 0; i < UCD9000_GPI_COUNT; i++) {
> >>+            entries[i].client = client;
> >>+            entries[i].index = i;
> >>+            scnprintf(name, UCD9000_DEBUGFS_NAME_LEN,
> >>+                  "gpi%d_alarm", i + 1);
> >>+            debugfs_create_file(name, 0444, data->debugfs,
> >>+                        &entries[i],
> >>+                        &ucd9000_debugfs_mfr_status_bit);
> >>+        }
> >>+
> >>+        scnprintf(name, UCD9000_DEBUGFS_NAME_LEN, "mfr_status");
> >>+        debugfs_create_file(name, 0444, data->debugfs, client,
> >>+                    &ucd9000_debugfs_mfr_status_word4);
> >>+    } else if (mid->driver_data == ucd90120 ||
> >>+           mid->driver_data == ucd90124) {
> >>+        scnprintf(name, UCD9000_DEBUGFS_NAME_LEN, "mfr_status");
> >>+        debugfs_create_file(name, 0444, data->debugfs, client,
> >>+                    &ucd9000_debugfs_mfr_status_word2);
> >>+    }
> >>+
> >>+    return 0;
> >>+}
> >
> >Is all that complexity really worth it ? Why not just read the
> >manufacturing
> >status as byte string into a buffer and use hexdump to pront it, no matter
> >how
> >many bytes are actually returned ? This would also be less error prone,
> >and
> >automatically support future chips.
> 
> Hm, well then we have the additional complexity of setting up custom debugfs
> file operations to show the binary data instead of just using
> DEFINE_DEBUGFS_ATTRIBUTE. Plus, at some point someone has to interpret it as
> either a word, half-word, or 5 bytes chunk. Where better to do it than the
> driver, as this is hw-dependent?
> 
I can't exactly follow your logic. You mean it is acceptable for the user to
have to look into the datasheet to find out what the 1/2/4 byte hex value means,
but it is unacceptable to expect the user to have to use the datasheet to
identify what a 1..5 byte hex string, displayed in the order received from the
chip, means ? I am having difficulties understanding the difference. How is
12345678 different from, say, 12 34 56 78 (which you could display as 12345678
as well) ?

The macro generates the file operations as part of the define, so I don't see
having to define as valid argument. One could instead add a generic debugfs
macro to display a string if that is of interest.

> I could just use one function and do a byte-swap based on data length in a
> loop within #ifdef LITTLE_ENDIAN, but that's a little messy. It will handle
> all the cases though. What do you think?
> 
Personally I don't see a problem displaying data as received. Either case, there
are functions/macros to convert from big/little endian to host byte order, so
related ifdefs in the code should never be necessary.

Guenter

> Thanks,
> Eddie
> 
> >
> >Thanks,
> >Guenter
> >
> >>+#else
> >>+static int ucd9000_init_debugfs(struct i2c_client *client,
> >>+                struct ucd9000_data *data)
> >>+{
> >>+    return 0;
> >>+}
> >>+#endif /* IS_ENABLED(CONFIG_DEBUG_FS) */
> >>+
> >>  static int ucd9000_probe(struct i2c_client *client,
> >>               const struct i2c_device_id *id)
> >>  {
> >>@@ -483,7 +644,16 @@ static int ucd9000_probe(struct i2c_client *client,
> >>          return ret;
> >>      }
> >>  -    return pmbus_do_probe(client, mid, info);
> >>+    ret = pmbus_do_probe(client, mid, info);
> >>+    if (ret)
> >>+        return ret;
> >>+
> >>+    ret = ucd9000_init_debugfs(client, mid, data);
> >>+    if (ret)
> >>+        dev_warn(&client->dev, "Failed to register debugfs: %d\n",
> >>+             ret);
> >>+
> >>+    return 0;
> >>  }
> >>    /* This is the driver that will be inserted */
> >>
> >
>