> http://www.opensolaris.org/os/project/clearview/nemo-mactype-plugins.txt > http://www.opensolaris.org/os/project/clearview/nemo-binary-compatibility.txt
Nice job! In the second document, this: mc_callbacks is a flags field that drivers use to denote which _optional_ callbacks are set in the structure. The last two callbacks defined in this structure are currently optional (mc_ioctl and mc_capab_get), and thus the only two possible flags are: ... appears to be stale (mc_resources should also be named). Also, regarding names: wouldn't mc_getcapab be more consistent with mc_getstat and mc_setpromisc? (Likewise, mac_getcapab_t and MC_GETCAPAB seem preferable.) In addition, I presume that a plugin author should set m_plugin_data_size should be set to zero if m_plugin_data is NULL? If so, I'd make that explicit in the proposal. Regarding the DLPI notifications: it might be worth including a reference to the PSARC case which introduced DL_NOTE_FASTPATH_FLUSH. Also, what's our rationale for making DL_NOTE_PHYS_ADDR/DL_CURR_DEST_ADDR not be Evolving like DL_NOTE_PHYS_ADDR/DL_CURR_PHYS_ADDR? Finally: why are the Section 4 interfaces not in the interface table? -- meem _______________________________________________ networking-discuss mailing list [email protected]
