Bug 22556

Summary: GENERIC kernel config should indicate SMC PCI NIC (device tx) as using common MII code.
Product: Base System Reporter: hodgkina <hodgkina>
Component: kernAssignee: semenu <semenu>
Status: Closed FIXED    
Severity: Affects Only Me    
Priority: Normal    
Version: Unspecified   
Hardware: Any   
OS: Any   

Description hodgkina 2000-11-02 22:30:00 UTC
In existing GENERIC, SMC 9432TX cards (device tx) are not designated as using the common MII bus controller code.  However, as of Sun Sep 10 19:51:07 2000 UTC, the driver was converted to use newbus and miibus.  See http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/pci/if_tx.c and refer to revision 1.34.2.3.

Fix: 

Move 'device tx' line from current section (#PCI Ethernet NICs) to MII Bus section (#PCI Ethernet NICs that use common MII bus controller code.)
How-To-Repeat: 1. Comment out 'device miibus' in GENERIC
2. config <kernel>
3. cd to compile directory and 'make depend'.  The following error is then received:
../../pci/if_tx.c:80: miibus_if.h: No such file or directory
mkdep: compile failed
*** Error code 1
Comment 1 Johan Karlsson freebsd_committer freebsd_triage 2000-11-03 21:48:58 UTC
Responsible Changed
From-To: freebsd-bugs->semenu

Over to tx(4) maintainer.
Comment 2 semenu freebsd_committer freebsd_triage 2000-11-07 20:28:41 UTC
State Changed
From-To: open->closed

With the efforts of Bill Fumerola <billf@FreeBSD.org> and a bit 
of mine, problem is fixed and PR is closed.
Comment 3 Andrew.Hodgkins 2000-12-26 20:55:05 UTC
Just a brief followup: I noticed that this change has been made in LINT, but
it hasn't been committed to GENERIC yet (at least not in the STABLE
branch...I'm not tracking CURRENT).  Just wondering if it might be an
oversight, since the problem can still be reproduced using the previously
described method in the original PR. 

uname -a: 
FreeBSD josephine.hurlburt.af.mil 4.2-STABLE FreeBSD 4.2-STABLE #0: Mon Dec
18 19:55:45 CST 2000
root@josephine.hurlburt.af.mil:/usr/obj/usr/src/sys/JOSEPHINE  i386  

I know this is only a doc bug and as such is ultra-low priority, but at
least it's a simple fix.  ;)

--Andy