[lkml]   [1998]   [Jan]   [11]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
    SubjectRe: devfs
    Michael Neuffer writes:
    > On Sun, 11 Jan 1998, Richard Gooch wrote:
    > > > No not controllers. This is buses on the controller. 4 bits for this
    > > > should be the lower limit. The smallest SSA that you can buy from
    > > > SUN for example has already 6 buses.
    > >
    > > That's interesting. Since you seem to know something about SCSI host
    > > adapters with multiple SCSI buses, can you please tell us how they
    > > show that in /dev? I'm aware of the following format:
    > >
    > > cCtTdDsS
    > >
    > > where <C> is the controller number (host adaptor), <T> is the SCSI
    > > target (SCSI ID), <D> is the device number (LUN) and <S> is the stripe
    > > (partition).
    > >
    > > Where do they put the per-host/controller SCSI bus ID?
    > Solaris resorts to an ugly trick to encode the buses of the SSAs.
    > When speaking to an SSA the LUN gets dropped and the fields shift
    > right to make room for the channel (bus).
    > Now the meaning is follows:
    > cCtBdTsS
    > Where <C> is the controller, <B> is the bus/channel, <T> is the target and
    > <S> is the slice.
    > This allows then to stay compatible with their naming sheme but has the
    > drawback of dropping support for LUNs. This only works properly for the
    > special case of the SSAs since you only have disk drives in them.
    > However I do not think that we should make the same mistake as them
    > and cut ourselves off from supporting multi LUN devices.

    Agreed! Considering this, I'm open to alternative naming schemes, but
    only if they preserve the information.



     \ /
      Last update: 2005-03-22 13:41    [W:0.019 / U:7.956 seconds]
    ©2003-2017 Jasper Spaans. hosted at Digital OceanAdvertise on this site