Looks very good! Just a couple minor comments below. Was this tested with both
single and dual adapter configs? SIS-32 and SIS-64?

Thanks,

Brian

On 11/18/2015 11:26 AM, Heitor Ricardo Alves de Siqueira wrote:
> @@ -9986,3 +9986,47 @@ int ipr_query_io_dev_port(struct ipr_dev *dev, struct 
> ipr_query_io_port *io_port
>       return rc;
>  }
> 
> +/**
> + * ipr_jbod_sysfs_bind -
> + * @dev:             ipr dev struct
> + * @op:                      sysfs operation
> + *
> + * Returns:
> + *   0 if success / non-zero on failure
> + **/
> +int ipr_jbod_sysfs_bind(struct ipr_dev *dev, u8 op)
> +{
> +     struct ipr_dev *mp_dev;
> +     int rc, fd, size;
> +     char *sysfs_device_name;
> +
> +     sysfs_device_name = dev->scsi_dev_data->sysfs_device_name;
> +     size = strnlen(sysfs_device_name, sizeof(sysfs_device_name));
> +
> +     if (op == IPR_JBOD_SYSFS_BIND) {
> +             fd = open("/sys/bus/scsi/drivers/sd/bind", O_WRONLY);
> +     } else if (op == IPR_JBOD_SYSFS_UNBIND) {
> +             fd = open("/sys/bus/scsi/drivers/sd/unbind", O_WRONLY);
> +     } else {
> +             fd = -1;
> +             errno = ENOTSUP;
> +     }
> +     if (fd < 0)
> +             return errno;
> +
> +     rc = write(fd, sysfs_device_name, size);
> +     if (rc < 0)

Need a close(fd) here.

> +             return errno;
> +
> +     mp_dev = find_multipath_jbod(dev);
> +     if (mp_dev) {
> +             sysfs_device_name = mp_dev->scsi_dev_data->sysfs_device_name;
> +             size = strnlen(sysfs_device_name, sizeof(sysfs_device_name));
> +             rc = write(fd, sysfs_device_name, size);
> +             if (rc < 0)

Ditto

> +                     return errno;
> +     }
> +
> +     close(fd);
> +     return 0;
> +}


-- 
Brian King
Power Linux I/O
IBM Linux Technology Center


------------------------------------------------------------------------------
_______________________________________________
Iprdd-devel mailing list
Iprdd-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/iprdd-devel

Reply via email to