svn commit: r194918 - head/sys/net

[Navdeep Parhar]

On Thu, Jun 25, 2009 at 09:18:47AM -0400, George Neville-Neil wrote:
> 
> On Jun 24, 2009, at 17:53 , Navdeep Parhar wrote:
> ....
> >======================================================================
> >--- head/sys/net/if_media.h	Wed Jun 24 21:51:42 2009	(r194917)
> >+++ head/sys/net/if_media.h	Wed Jun 24 21:53:25 2009	(r194918)
> >@@ -149,6 +149,7 @@ uint64_t	ifmedia_baudrate(int);
> >#define IFM_10G_TWINAX_LONG	23	/* 10GBase Twinax Long copper */
> >#define IFM_10G_LRM	24		/* 10GBase-LRM 850nm Multi-mode */
> >#define IFM_UNKNOWN	25		/* media types not defined yet */
> >+#define IFM_10G_T	26		/* 10GBase-T - RJ45 */
> >
> >
> >/* note 31 is the max! */
> >@@ -358,6 +359,7 @@ struct ifmedia_description {
> >	{ IFM_10G_TWINAX,	"10Gbase-Twinax" },			\
> >	{ IFM_10G_TWINAX_LONG,	"10Gbase-Twinax-Long" },		\
> >	{ IFM_UNKNOWN,	"Unknown" },					\
> >+	{ IFM_10G_T,	"10Gbase-T" },					\
> >	{ 0, NULL },							\
> >}
> >
> >@@ -615,6 +617,7 @@ struct ifmedia_baudrate {
> >	{ IFM_ETHER | IFM_10G_TWINAX,	IF_Gbps(10ULL) },		\
> >	{ IFM_ETHER | IFM_10G_TWINAX_LONG,	IF_Gbps(10ULL) },	\
> >	{ IFM_ETHER | IFM_10G_LRM,	IF_Gbps(10ULL) },		\
> >+	{ IFM_ETHER | IFM_10G_T,	IF_Gbps(10ULL) },		\
> >									\
> >	{ IFM_TOKEN | IFM_TOK_STP4,	IF_Mbps(4) },			\
> >	{ IFM_TOKEN | IFM_TOK_STP16,	IF_Mbps(16) },			\
> 
> 
> Howdy,
> 
> John Baldwin pointed out to me that UNKNOWN really should have been  
> pushed up to the end.

Mostly for cosmetic reasons, correct?  I'm not sure if pushing it to
the end has any functional impact.  I'd deliberately left it where
it was, simply to avoid gratuitous differences between 7 and 8
#define's.  [I don't think we can reorder the #define's on 7 without
having to rebuild its ifconfig too].

> 
> Can you make a small patch and send it to me?  Increase the number and  
> move its string to the end of the array?

If you'd still like them reordered - should we simply swap
IFM_UNKNOWN/25 and IFM_10G_T/26 or should we #define IFM_UNKNOWN to be
31, the max allowed?  That way we can add more types and not have to
bother with IFM_UNKNOWN again.

Regards,
Navdeep