Re: [PATCH] platform/x86: Add IMS/ZII SCU driver
From: Andy Shevchenko
Date: Tue Jan 24 2017 - 16:08:35 EST
On Wed, Jan 11, 2017 at 11:26 PM, Florian Fainelli <f.fainelli@xxxxxxxxx> wrote:
> From: Guenter Roeck <linux@xxxxxxxxxxxx>
>
> This patch adds support for the IMS (now Zodiac Inflight Innovations)
> SCU Generation 1/2/3 platform driver. This driver registers all the
> on-module peripherals: Ethernet switches (Broadcom or Marvell), I2C to
> GPIO expanders, Kontrom CPLD/I2C master, and more.
Here additional comments to previously shallow review.
> +config SCU
> + tristate "SCU driver"
> + depends on DMI
> + ---help---
> + Enable SCU driver
Be more verbose here, list platforms it supports, etc.
> + * You should have received a copy of the GNU General Public License
> + * along with this program; if not, write to the Free Software
> + * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
Please, remove this snail address.
> +struct __packed eeprom_data {
> + unsigned short length; /* 0 - 1 */
> + unsigned char checksum; /* 2 */
> + unsigned char have_gsm_modem; /* 3 */
> + unsigned char have_cdma_modem; /* 4 */
> + unsigned char have_wifi_modem; /* 5 */
> + unsigned char have_rhdd; /* 6 */
> + unsigned char have_dvd; /* 7 */
> + unsigned char have_tape; /* 8 */
> + unsigned char have_humidity_sensor; /* 9 */
> + unsigned char have_fiber_channel; /* 10 */
> + unsigned char lru_part_number[11]; /* 11 - 21 Box Part Number */
> + unsigned char lru_revision[7]; /* 22 - 28 Box Revision */
> + unsigned char lru_serial_number[7]; /* 29 - 35 Box Serial Number */
> + unsigned char lru_date_of_manufacture[7];
> + /* 36 - 42 Box Date of Manufacture */
> + unsigned char board_part_number[11];
> + /* 43 - 53 Base Board Part Number */
> + unsigned char board_revision[7];
> + /* 54 - 60 Base Board Revision */
> + unsigned char board_serial_number[7];
> + /* 61 - 67 Base Board Serial Number */
> + unsigned char board_date_of_manufacture[7];
> + /* 68 - 74 Base Board Date of Manufacture */
> + unsigned char board_updated_date_of_manufacture[7];
> + /* 75 - 81 Updated Box Date of Manufacture */
> + unsigned char board_updated_revision[7];
> + /* 82 - 88 Updated Box Revision */
> + unsigned char dummy[7]; /* 89 - 95 spare/filler */
> +};
Perhaps gather all comments to one nice description of EEPROM structure?
> +struct scu_platform_data {
> + const char *board_type;
> + const char *lru_part_number;
> + const char *board_part_number;
> + enum scu_version version;
> + int eeprom_len;
> + struct i2c_board_info *i2c_board_info;
> + int num_i2c_board_info;
> + struct spi_board_info *spi_board_info;
> + int num_spi_board_info;
> + void (*init)(struct scu_data *);
> + void (*remove)(struct scu_data *);
> +};
> +
> +struct scu_data {
> + struct device *dev; /* SCU platform device */
> + struct net_device *netdev; /* Ethernet device */
> + struct platform_device *mdio_dev; /* MDIO device */
> + struct platform_device *dsa_dev; /* DSA device */
> + struct proc_dir_entry *rave_proc_dir;
> + struct mutex write_lock;
> + struct platform_device *leds_pdev[3];
3
> + struct i2c_client *client[10]; /* I2C clients */
10
> + struct spi_device *spidev[1]; /* SPI devices */
1
Is it by design in hardware, or it might be changed in SCU firmware?
Would be nice to have a comment for the fields.
> +/* sysfs */
Kinda useless.
> +static int scu_update_checksum(struct scu_data *data)
> +{
> + unsigned char checksum;
> + int ret;
> +
> + data->eeprom.checksum = 0;
> + checksum = scu_get_checksum((unsigned char *)&data->eeprom,
> + data->pdata->eeprom_len);
> + data->eeprom.checksum = ~checksum + 1;
> + ret = nvmem_device_write(data->nvmem,
> + 0x300 + offsetof(struct eeprom_data, checksum),
Magic.
> + sizeof(data->eeprom.checksum),
> + &data->eeprom.checksum);
> + if (ret <= 0)
> + return ret < 0 ? ret : -EIO;
> + return 0;
if (ret < 0)
return ret;
return ret ? 0 : -EIO;
?
Though I think ret == 0 case would be handled by caller.
> +static ssize_t scu_object_store(struct scu_data *data, int offset,
> + const char *in, char *out, ssize_t len)
> +{
> + char buffer[12] = { 0 };
Do you need initialization for it?
> + strlcpy(buffer, in, sizeof(buffer));
> + /* do not copy newline from input string */
> + cp = strchr(buffer, '\n');
> + if (cp)
> + *cp = '\0';
Perhaps better to move it after you get new len.
> + if (len > sizeof(buffer))
> + len = sizeof(buffer);
> +
> + mutex_lock(&data->write_lock);
> + strncpy(out, buffer, len);
Does it need to be protected?
> +
> + /* Write entire eeprom if it was marked invalid */
> + if (!data->eeprom_valid) {
> + offset = 0;
> + /* Write checksumed and non checksumed data */
> + len = sizeof(data->eeprom);
> + out = (char *)&data->eeprom;
Maybe just do a call with different parameters?
> + }
> +
> + ret = nvmem_device_write(data->nvmem, 0x300 + offset, len, out);
Magic.
> + if (ret <= 0) {
> + data->eeprom_valid = false;
> + if (ret == 0)
> + ret = -EIO;
Since you are using this idiom you might consider a wrapper for this.
It would perhaps hide magic number as well.
> + goto error;
> + }
> + if (offset < data->pdata->eeprom_len) {
> + /*
> + * Write to checksummed area of eeprom
> + * Update checksum
> + */
> + ret = scu_update_checksum(data);
> + if (ret < 0) {
> + data->eeprom_valid = false;
> + goto error;
> + }
> + }
> + data->eeprom_valid = true;
> +error:
error_unlock: ?
> + mutex_unlock(&data->write_lock);
> + return ret;
> +}
> +static umode_t scu_attr_is_visible(struct kobject *kobj, struct attribute *attr,
> + int index)
> +{
> + struct device *dev = container_of(kobj, struct device, kobj);
> + struct scu_data *data = dev_get_drvdata(dev);
> + umode_t mode = attr->mode;
> +
> + /*
> + * If the eeprom has not been processed, disable its attributes.
> + * If it has been processed but is not accessible, disable
> + * write accesses to it.
> + */
> + if (index >= 1 && !data->eeprom_accessible)
> + mode &= 0444;
return mode & 0444; ?
> +
> + if (index >= 4 && data->pdata->version == scu1)
> + return 0;
> +
> + return mode;
> +}
> +/* platform data */
Useless.
> + for (i = 0; i < ARRAY_SIZE(data->leds_pdev); i++)
> + platform_device_unregister(data->leds_pdev[i]);
Sign of MFD.
> +static int scu_gpio_common_setup(unsigned int gpio_base, unsigned int ngpio,
> + u32 mask, u32 is_input, u32 is_active,
> + u32 active_low)
> +{
> + int i;
> + unsigned long flags;
> +
> + for (i = 0; i < ngpio; i++) {
> + if (!(mask & (1 << i)))
> + continue;
for_each_set_bit()
> + flags = GPIOF_EXPORT_DIR_FIXED;
> + if (is_input & (1 << i)) {
> + flags |= GPIOF_DIR_IN;
> + } else {
> + flags |= GPIOF_DIR_OUT;
> + if (is_active & (1 << i))
> + flags |= GPIOF_INIT_HIGH;
> + }
> + if (active_low & (1 << i))
> + flags |= GPIOF_ACTIVE_LOW;
> + gpio_request_one(gpio_base + i, flags, NULL);
> + }
Use BIT() macro
> + return 0;
> +}
I might assume that GPIO library has some helper for that.
> +static void scu_gpio_common_teardown(unsigned int gpio_base, int ngpio,
> + u32 mask)
> +{
> + int i;
> +
> + for (i = 0; i < ngpio; i++) {
> + if (mask & (1 << i)) {
for_each_set_bit()
> + gpio_unexport(gpio_base + i);
> + gpio_free(gpio_base + i);
> + }
> + }
> +}
And something telling me that it would much easier to do this.
Existing helpers / drivers?
> +static int scu_instantiate_i2c(struct scu_data *data, int base,
> + struct i2c_board_info *info, int count)
> +{
> + int i;
> +
> + for (i = 0; i < count; i++) {
> + data->client[base + i] = i2c_new_device(data->adapter, info);
> + if (!data->client[base + i]) {
> + /*
> + * Unfortunately this call does not tell us
> + * why it failed. Pick the most likely reason.
> + */
> + return -EBUSY;
> + }
Shouldn't i2c core do this for you whenever adapter is probed?
> + info++;
> + }
> + return 0;
> +}
> +static int scu_instantiate_spi(struct scu_data *data,
> + struct spi_board_info *info, int count)
> +{
> + struct spi_master *master;
> + int i;
> +
> + /* SPI bus number matches i2c bus number (set by sc18is602 driver) */
> + master = spi_busnum_to_master(data->adapter->nr);
> + if (!master) {
> + dev_err(data->dev, "Failed to find SPI adapter\n");
> + return -ENODEV;
> + }
> + data->master = master;
> +
> + for (i = 0; i < count; i++) {
> + info->bus_num = master->bus_num;
> + /* ignore errors */
> + data->spidev[i] = spi_new_device(master, info);
> + info++;
> + }
Ditto (spi core).
> + return 0;
> +}
> +/*
> + * This is the callback function when a a specifc at24 eeprom is found.
> + * Its reads out the eeprom contents via the read function passed back in via
> + * struct memory_accessor. It then calls part_number_proc, serial_number_proc,
> + * and dom_proc to populate the procfs entries for each specific field.
> + */
Hmm... Sounds like a stuff on top of EEPROM chip(s).
Can it be separate module?
> +static void populate_unit_info(struct nvmem_device *nvmem,
> + void *context)
> +{
> + const struct scu_platform_data *pdata = &scu_platform_data[unknown];
> + struct scu_data *data = context;
> + unsigned char *ptr;
> + int i, len;
> +
> + data->nvmem = nvmem;
> +
> + ptr = (unsigned char *)&data->eeprom;
> + /* Read Data structure from EEPROM */
> + if (nvmem_device_read(nvmem, 0x300, sizeof(data->eeprom), ptr) <= 0) {
> + dev_err(data->dev, "Failed to read eeprom data\n");
> + goto error;
> + }
> +
> + /* EEPROM is accessible, permit write access to it */
> + data->eeprom_accessible = true;
> +
> + /* Special case - eeprom not programmed */
> + if (data->eeprom.length == 0xffff && data->eeprom.checksum == 0xff) {
> + /* Assume it is SCU3, but report different board type */
> + memset(&data->eeprom, '\0', sizeof(data->eeprom));
> + data->eeprom.length = cpu_to_le16(SCU_EEPROM_LEN_EEPROM);
> + goto unprogrammed;
> + }
> +
> + /* Sanity check */
> + if (le16_to_cpu(data->eeprom.length) != SCU_EEPROM_LEN_EEPROM) {
> + dev_err(data->dev,
> + "Bad eeprom data length: Expected %d, got %d\n",
> + SCU_EEPROM_LEN_EEPROM,
> + le16_to_cpu(data->eeprom.length));
> + goto error;
> + }
> +
> + /* Update platform data based on part number retrieved from EEPROM */
> + for (i = 0; i < ARRAY_SIZE(scu_platform_data); i++) {
> + const struct scu_platform_data *tpdata = &scu_platform_data[i];
> +
> + if (tpdata->lru_part_number == NULL)
> + continue;
> + if (!strncmp(data->eeprom.lru_part_number,
> + tpdata->lru_part_number,
> + strlen(tpdata->lru_part_number))) {
> + pdata = tpdata;
> + break;
> + }
> + }
> +
> +unprogrammed:
> + data->pdata = pdata;
> + /*
> + * We know as much as we will ever find out about the platform.
> + * Perform final platform initialization and instantiate additional
> + * I2C devices as well as SPI devices now, prior to validating
> + * the EEPROM checksum.
> + */
> + if (pdata->init)
> + pdata->init(data);
> +
> + if (pdata->i2c_board_info)
> + scu_instantiate_i2c(data, ARRAY_SIZE(scu_i2c_info_common),
> + pdata->i2c_board_info,
> + pdata->num_i2c_board_info);
> +
> + if (pdata->spi_board_info)
> + scu_instantiate_spi(data, pdata->spi_board_info,
> + pdata->num_spi_board_info);
> +
> + len = data->pdata->eeprom_len;
> +
> + /* Validate checksum */
> + if (scu_get_checksum(ptr, len)) {
> + dev_err(data->dev,
> + "EEPROM data checksum error: expected 0, got 0x%x [len=%d]\n",
> + scu_get_checksum(ptr, len), len);
> + /* TBD: Do we want to clear the eeprom in this case ? */
> + goto error_noclean;
> + }
> +
> + /* Create sysfs attributes based on retrieved platform data */
> + data->eeprom_valid = true;
> + goto done;
> +
> +error:
> + memset(&data->eeprom, '\0', sizeof(data->eeprom));
> + data->eeprom.length = cpu_to_le16(SCU_EEPROM_LEN_EEPROM);
> +error_noclean:
> + data->eeprom_valid = false;
> +done:
> + if (sysfs_create_group(&data->dev->kobj, &scu_eeprom_group))
> + ;
> + if (sysfs_create_bin_file(&data->dev->kobj,
> + &scu_eeprom_test_scratchpad_file))
> + ;
> +}
> +static int scu_probe(struct platform_device *pdev)
> +{
> + struct proc_dir_entry *rave_board_type;
> + struct device *dev = &pdev->dev;
> + struct i2c_adapter *adapter;
> + struct net_device *ndev;
> + struct scu_data *data;
> + int i, ret;
> +
> + scu_request_modules(true);
> +
> + data = devm_kzalloc(dev, sizeof(*data), GFP_KERNEL);
> + if (!data)
> + return -ENOMEM;
> +
> + platform_set_drvdata(pdev, data);
> +
> + data->dev = dev;
> +
> + mutex_init(&data->write_lock);
> +
> + /* look for ethernet device attached to 'e1000e' driver */
> + rtnl_lock();
> + for_each_netdev(&init_net, ndev) {
> + if (ndev->dev.parent && ndev->dev.parent->driver &&
> + !strcmp(ndev->dev.parent->driver->name, "e1000e")) {
> + data->netdev = ndev;
> + break;
> + }
> + }
> + rtnl_unlock();
> +
> + if (!data->netdev)
> + return -EPROBE_DEFER;
Can it be that it never appears?
> +
> + /*
> + * The adapter driver should have been loaded by now.
> + * If not, try again later.
> + */
> + adapter = scu_find_i2c_adapter("i2c-kempld");
> + if (!adapter) {
> + ret = -EPROBE_DEFER;
> + goto error_put_net;
Ditto.
> + .owner = THIS_MODULE,
Do we need this?
> +static struct platform_device *scu_pdev;
--
With Best Regards,
Andy Shevchenko