Vale port naming

[Matt Churchyard]

>please fetch the version from github.com/luigirizzo/netmap<http://github.com/luigirizzo/netmap> if
>you want a more recent version of the code, and email me
>directly if you have a panic.

>Both the versioni in HEAD and 10.x are lagging a bit behind.

>cheers
>luigi

I’m probably doing something wrong (I don’t build kernels by hand very often), but I’ve tried to apply the github code to a recent copy of head (297617) and I’m getting errors like the following –

/usr/src/sys/dev/netmap/if_em_netmap.h:151:22: error: incompatible pointer types initializing 'struct em_buffer *' with an expression of type 'struct em_txbuffer *' [-Werror,-Wincompatible-pointer-types]
                        struct em_buffer *txbuf = &txr->tx_buffers[nic_i];
                                          ^       ~~~~~~~~~~~~~~~~~~~~~~~
/usr/src/sys/dev/netmap/if_em_netmap.h:161:44: error: incomplete definition of type 'struct em_buffer'
                                netmap_reload_map(na, txr->txtag, txbuf->map, addr);
                                                                  ~~~~~^
/usr/src/sys/dev/netmap/if_em_netmap.h:151:11: note: forward declaration of 'struct em_buffer'
                        struct em_buffer *txbuf = &txr->tx_buffers[nic_i];
                               ^
/usr/src/sys/dev/netmap/if_em_netmap.h:169:37: error: incomplete definition of type 'struct em_buffer'
                        bus_dmamap_sync(txr->txtag, txbuf->map,
                                                    ~~~~~^
/usr/src/sys/sys/bus_dma.h:300:8: note: expanded from macro 'bus_dmamap_sync'
                if ((dmamap) != NULL)                           \
                     ^~~~~~
/usr/src/sys/dev/netmap/if_em_netmap.h:151:11: note: forward declaration of 'struct em_buffer'
                        struct em_buffer *txbuf = &txr->tx_buffers[nic_i];
                               ^
/usr/src/sys/dev/netmap/if_em_netmap.h:169:37: error: incomplete definition of type 'struct em_buffer'
                        bus_dmamap_sync(txr->txtag, txbuf->map,
                                                    ~~~~~^
/usr/src/sys/sys/bus_dma.h:301:27: note: expanded from macro 'bus_dmamap_sync'
                        _bus_dmamap_sync(dmat, dmamap, op);     \
                                               ^~~~~~
/usr/src/sys/dev/netmap/if_em_netmap.h:151:11: note: forward declaration of 'struct em_buffer'
                        struct em_buffer *txbuf = &txr->tx_buffers[nic_i];

Any ideas or am I trying to do something stupid?

Regards,
Matt

On Tue, Apr 5, 2016 at 9:59 AM, Matt Churchyard via freebsd-net <freebsd-net at freebsd.org<mailto:freebsd-net at freebsd.org>> wrote:
> On Mon, Apr 04, 2016 at 10:37:48PM +0000, Brooks Davis wrote:
> > On Mon, Apr 04, 2016 at 11:29:31AM +0000, Matt Churchyard via freebsd-net wrote:
> > > From the little information I can find on the net it seems that valeX:Y is the format for a port on a vale switch.
> > > Some examples use vale0:1, others use a letter such as valeA:0. The only details I can find is a vague reference to a 15/16? character limit,
> >  > although I don't know whether that applies to just the first part of the whole thing.
> >
> > Can anyone clarify the valid format for a vale switch/port?
> > Is there any length or character restriction on X & Y?
>
> According to the vale.4 manpage:
>
>         vale ports are named vale[bdg:][port] where vale is the prefix
>         indicating a VALE switch rather than a standard interface, bdg
>         indicates a specific switch (the colon is a separator), and port
>         indicates a port within the switch.  Bridge and port names are
>         arbitrary strings, the only constraint being that the full name
>         must fit within 16 characters.
>
> The manpage is confusing in that the name must be 15 characters plus
> the NUL character.  Given the other limits, it looks like bdg can be
> up to 8 characters in practice.

>  Having hit send, I've noticed I'm wrong here. :)  In fact, you have 10 characters to split betwen bdg and port.

> -- Brooks
Ok, so I can assume using port names like "vale9d9af:66aa9" isn't the reason that my test machine is panicing.
I might give 10.3 a go instead of head and see how I get on with that.

Matt
_______________________________________________
freebsd-net at freebsd.org<mailto:freebsd-net at freebsd.org> mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to "freebsd-net-unsubscribe at freebsd.org<mailto:freebsd-net-unsubscribe at freebsd.org>"



--
-----------------------------------------+-------------------------------
 Prof. Luigi RIZZO, rizzo at iet.unipi.it<mailto:rizzo at iet.unipi.it>  . Dip. di Ing. dell'Informazione
 http://www.iet.unipi.it/~luigi/        . Universita` di Pisa
 TEL      +39-050-2217533               . via Diotisalvi 2
 Mobile   +39-338-6809875               . 56122 PISA (Italy)
-----------------------------------------+-------------------------------