Re: ixg(4) and Intel X540?

[Hauke Fath <hauke%Espresso.Rhein-Neckar.DE@localhost>]

On Fri, 20 Feb 2015 15:53:00 +0000 (UTC), Christos Zoulas wrote:
> Or he could try to merge the diffs from FreeBSD, the code was similar
> last time I checked.

Well, if this hadn't been intimidating enough...

% ls -l 
src/sys/dev/pci/ixgbe/                                                              
total 820
drwxr-xr-x  2 hf  wheel     512 Feb 20 12:31 CVS
-rw-r--r--  1 hf  wheel    1885 Aug 12  2011 LICENSE
-rw-r--r--  1 hf  wheel    8252 Aug 12  2011 README
-rw-r--r--  1 hf  wheel  169689 Feb 12 12:48 ixgbe.c
-rw-r--r--  1 hf  wheel   15294 Feb 12 12:48 ixgbe.h
-rw-r--r--  1 hf  wheel   41859 Aug 12  2011 ixgbe_82598.c
-rw-r--r--  1 hf  wheel   69297 Oct 17 15:34 ixgbe_82599.c
-rw-r--r--  1 hf  wheel   33715 Oct 17 15:34 ixgbe_api.c
-rw-r--r--  1 hf  wheel    7882 Aug 12  2011 ixgbe_api.h
-rw-r--r--  1 hf  wheel   99878 Oct 17 15:34 ixgbe_common.c
-rw-r--r--  1 hf  wheel    6342 Aug 12  2011 ixgbe_common.h
-rw-r--r--  1 hf  wheel   19924 Aug 12  2011 ixgbe_mbx.c
-rw-r--r--  1 hf  wheel    5795 Aug 12  2011 ixgbe_mbx.h
-rw-r--r--  1 hf  wheel    5990 Feb 12 12:48 ixgbe_netbsd.c
-rw-r--r--  1 hf  wheel    3373 Oct 17 15:34 ixgbe_netbsd.h
-rw-r--r--  1 hf  wheel    5736 Oct 17 15:34 ixgbe_osdep.h
-rw-r--r--  1 hf  wheel   51520 Oct 17 15:34 ixgbe_phy.c
-rw-r--r--  1 hf  wheel    6537 Aug 12  2011 ixgbe_phy.h
-rw-r--r--  1 hf  wheel  123536 Oct 17 15:34 ixgbe_type.h
-rw-r--r--  1 hf  wheel   15689 Aug 12  2011 ixgbe_vf.c
-rw-r--r--  1 hf  wheel    4443 Aug 12  2011 ixgbe_vf.h
-rw-r--r--  1 hf  wheel  108801 Oct 17 15:34 ixv.c
-rw-r--r--  1 hf  wheel   11158 Dec  3  2012 ixv.h
%

... this surely would have been:

% foreach ff ( src/sys/dev/pci/ixgbe/*.[ch] )
foreach? test -f ../other/freebsd-svn/sys/dev/ixgbe/`basename $ff` || 
continue
foreach? echo -n "$ff   "
foreach? diff -u $ff ../other/freebsd-svn/sys/dev/ixgbe/`basename $ff` 
| wc -l
foreach? end
src/sys/dev/pci/ixgbe/ixgbe.c       7112
src/sys/dev/pci/ixgbe/ixgbe.h        492
src/sys/dev/pci/ixgbe/ixgbe_82598.c        810
src/sys/dev/pci/ixgbe/ixgbe_82599.c       2266
src/sys/dev/pci/ixgbe/ixgbe_api.c        906
src/sys/dev/pci/ixgbe/ixgbe_api.h        188
src/sys/dev/pci/ixgbe/ixgbe_common.c       3615
src/sys/dev/pci/ixgbe/ixgbe_common.h        154
src/sys/dev/pci/ixgbe/ixgbe_mbx.c        163
src/sys/dev/pci/ixgbe/ixgbe_mbx.h        145
src/sys/dev/pci/ixgbe/ixgbe_osdep.h        199
src/sys/dev/pci/ixgbe/ixgbe_phy.c       1550
src/sys/dev/pci/ixgbe/ixgbe_phy.h        210
src/sys/dev/pci/ixgbe/ixgbe_type.h       5118
src/sys/dev/pci/ixgbe/ixgbe_vf.c        500
src/sys/dev/pci/ixgbe/ixgbe_vf.h        147
src/sys/dev/pci/ixgbe/ixv.c        742
src/sys/dev/pci/ixgbe/ixv.h        166
%

Plus, there is

%  wc -l ../other/freebsd-svn/sys/dev/ixgbe/ixgbe_x540.?
    1000 ../other/freebsd-svn/sys/dev/ixgbe/ixgbe_x540.c
      67 ../other/freebsd-svn/sys/dev/ixgbe/ixgbe_x540.h
    1067 total
%

I tried porting openbsd's changes to msk(4) a while back (kern/49270), 
and even got it into ifconfig(8), but never a packet on the wire. And 
that's a far cry in complexity compared to the above. So I am kind of 
discouraged.

Cheerio,
hauke

-- 
Hauke Fath                        <hauke%Espresso.Rhein-Neckar.DE@localhost>
Ernst-Ludwig-Straße 15
64625 Bensheim
Germany