Re: [PATCH 13/24] mpt fusion: [2.6.30-rc6] Firmware event implementation using seperate WorkQueue

[James Bottomley <James.Bottomley@HansenPartnership.com>]

On Fri, 2009-05-22 at 16:33 +0530, Kashyap, Desai wrote:
> --- a/drivers/message/fusion/mptsas.h
> +++ b/drivers/message/fusion/mptsas.h
> @@ -61,12 +61,30 @@ enum mptsas_hotplug_action {
>         MPTSAS_DEL_DEVICE,
>         MPTSAS_ADD_RAID,
>         MPTSAS_DEL_RAID,
> +       MPTSAS_ADD_PHYSDISK,
> +       MPTSAS_ADD_PHYSDISK_REPROBE,
> +       MPTSAS_DEL_PHYSDISK,
> +       MPTSAS_DEL_PHYSDISK_REPROBE,
>         MPTSAS_ADD_INACTIVE_VOLUME,
>         MPTSAS_IGNORE_EVENT,
>  };
>  
> +struct sas_mapping{
> +       u8                      id;
> +       u8                      channel;
> +};
> +
> +struct sas_device_info {
> +       struct list_head        list;
> +       struct sas_mapping      os;     /* operating system mapping*/
> +       struct sas_mapping      fw;     /* firmware mapping */
> +       u64                     sas_address;
> +       u32                     device_info; /* specific bits for devices */
> +       u16                     slot;           /* enclosure slot id */
> +       u64                     enclosure_logical_id; /*enclosure address */
> +};

sas_device_info and sas_mapping aren't very good names for structures in
a file you include along with the generic scsi_transport_sas.h header,
which has a lot of sas_ prefix structures and functions in it ... it's
setting us up for a namespace clash.  Why aren't these mptsas_ prefixes
like everything else in this file?

James