Peter Memishian wrote:
> > This fast-path has been removed in the final phases of development.
> > For the initial UV putback, data packets will flow through the Nemo
> > framework as with any other Nemo driver.
>
> Should we note that an alternate (simplified) fastpath is under
> development and will be the subject of a future case? Otherwise, it seems
> to beg the question of why the fastpath is no longer needed.
Agreed; how about:
This fast-path has been removed in the final phases of development due
to to its incompatibility with an upcoming project which is making
major changes in the same area of the kernel. For the initial UV
putback, data packets will flow through the Nemo framework as with any
other Nemo driver. A simpler fastpath implementation is being
developed, and will be part of a future case.
> > * MAC_CAPAB_MDT (2006/499, Project Private) XXX
>
> XXX here can go.
Fixed.
>
> > * MAC_CAPAB_NO_NATIVEVLAN (2007/527, Consolidation Private)
>
> This has not been removed.
True; I'm not sure how it ended up on this list. Fixed.
> Also, MAC_CAPAB_NOZCOPY has been renamed to
> MAC_CAPAB_NO_ZCOPY to match MAC_CAPAB_NO_NATIVEVLAN.
Added:
2.3 MAC_CAPAB_NO_ZCOPY
The MAC_CAPAB_NOZCOPY MAC capability defined in 2006/499 is being
renamed by this case to MAC_CAPAB_NO_ZCOPY to be consistent with other
capability nomenclature.
>
> > typedef struct mac_capab_legacy_s {
> > dev_t mcl_dev;
> > uint32_t mcl_unsup_note;
> > } mac_capab_legacy_t;
>
> The above members should be ml_dev and ml_unsup_note.
Fixed.
>
> > The ml_unsup_note field is a bitfield of MAC capabilities.
>
> s/MAC capabilities/DLPI notifications/
Fixed.
> > Certain
> > "legacy" devices under softmac may not support the DLPI notifications
> > needed for softmac to provide all MAC notifications to the Nemo
> > framework. The softmac module fills in this field with the set of MAC
> > notifications that are not supported by the given softmac instance.
>
> s/MAC notifications/DLPI notifications/ ?
Yes for the last instance; Fixed.
> > 2.4 MAXLINKNAMELEN
> >
> > The MAXLINKNAMELEN constant has been added to <sys/params.h>, and it
> > defines the longest possible permitted datalink name (including the
> > terminating NULL character.) Its stability level is Committed.
>
> s/params.h/param.h/.
Fixed.
> Also, maybe we should also say something like "Note
> that the existing DLPI_LINKNAME_MAX constant remains, but is only intended
> for use by libdlpi consumers, whereas MAXLINKNAMELEN is appropriate for
> more general use."
Yes; added.
>
> > 2.5 New mac_margin_update() MAC driver runction
>
> s/runction/function/
Fixed.
> > 2.6 New Nemo MAC client functions
> >
> > In addition to the existing mac_open() function which allows MAC
> > clients to open a specific MAC by MAC name, UV introduces the
> > following functions to allow clients to open a MAC using different
> > identifers which may be available to MAC clients:
>
> s/identifer/identifier/
Fixed.
> > If the third argument is set to B_TRUE, then the caller wishes the
> > margin to remain at or above the current value, and the second
> > argument is set to the current value. If B_FALSE, the second argument
> > specifies a specific value that the caller wishes to be the minimum
> > margin.
>
> s/wish/require/ :-)
Fixed.
Thanks,
-Seb
