kern/52179: Minor issue with input errors with usmsc(4) on the netbsd-7 branch after netbsd-7-nhusb pullup

[dhgutteridge%sympatico.ca@localhost]

>Number:         52179
>Category:       kern
>Synopsis:       Minor issue with input errors with usmsc(4) on the netbsd-7 branch after netbsd-7-nhusb pullup
>Confidential:   no
>Severity:       non-critical
>Priority:       medium
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Wed Apr 19 01:20:00 +0000 2017
>Originator:     David H. Gutteridge
>Release:        netbsd-7 (a.k.a. "NetBSD 7.1_STABLE" at present)
>Organization:
>Environment:
NetBSD arcusviii.nonus-porta.net 7.1_STABLE NetBSD 7.1_STABLE (ARCUSVIII) #28: Wed Apr 12 09:14:01 EDT 2017  disciple%arcusiii.nonus-porta.net@localhost:/usr/builds/netbsd-7/src/sys/arch/evbarm/compile/obj/ARCUSVIII evbarm

>Description:
After the netbsd-7-nhusb branch was pulled into netbsd-7, I've found
the usmsc(4) interface on my Raspberry Pi has started reporting
input errors, where it wasn't previously, with equivalent traffic. To
date, the numbers are quite low, roughly 0.1% of total packets, and I
haven't found any broader issue. (I recall there was a serious
problem back in June 2014 that was soon fixed.[1] I can see that fix
is still present in the code, and I haven't experienced any
filesystem corruption or panics.)

A kernel with DEBUG enabled yields repeated error messages like this:

Apr 17 01:09:07 arcusviii /netbsd: debug: rx error (hdr 0x00403400)

I also have "usbhist" data dumped via vmstat(1) available, if that's
of use.

1. http://cvsweb.netbsd.org/bsdweb.cgi/src/sys/dev/usb/if_smsc.c?rev=1.17&content-type=text/x-cvsweb-markup&only_with_tag=MAIN

>How-To-Repeat:
Upgrade to a kernel after the netbsd-7-nhusb branch merger.
>Fix:
Unknown.