Date: Mon, 2 Nov 1998 13:06:06 -0000 From: Paul Richards <paul@originative.co.uk> To: "'Mike Smith'" <mike@smith.net.au> Cc: "'Luigi Rizzo'" <luigi@FreeBSD.ORG>, cvs-committers@FreeBSD.ORG, cvs-all@FreeBSD.ORG Subject: RE: cvs commit: src/sys/i386/isa if_lnc.c Message-ID: <E40CBF0361C7D111914000C0F0303D108792@OCTOPUS>
next in thread | raw e-mail | index | archive | help
> -----Original Message----- > From: Mike Smith [mailto:mike@smith.net.au] > Sent: Friday, October 30, 1998 5:47 PM > To: Paul Richards > Cc: 'Luigi Rizzo'; cvs-committers@FreeBSD.org; cvs-all@FreeBSD.org > Subject: Re: cvs commit: src/sys/i386/isa if_lnc.c > > > > > -----Original Message----- > > > From: Luigi Rizzo [mailto:luigi@FreeBSD.ORG] > > > Sent: Friday, October 30, 1998 11:54 AM > > > To: cvs-committers@FreeBSD.org; cvs-all@FreeBSD.org > > > Subject: cvs commit: src/sys/i386/isa if_lnc.c > > > > > > > > > luigi 1998/10/30 03:53:58 PST > > > > > > Modified files: > > > sys/i386/isa if_lnc.c > > > Log: > > > Fix a typo, and remove verbose error reporting when in > > > promisc mode (they are annoying in normal mode > > > as well so i am really tempted to remove them > unconditionally...) > > > > > > Revision Changes Path > > > 1.48 +4 -2 src/sys/i386/isa/if_lnc.c > > > > Well this is 50% ok, you're right about the spelling mistake but > > disabling the error message is a bad thing. > > > > The error messages you've disabled signify packet > corruption resulting > > in the packet being dropped (CRC and Framing errors), you'd > really want > > to know about that sort of thing happening. This is hardly > verbose error > > reporting! > > None of the other drivers report this (the 'de'driver used to, but it > was so annoying that it was disabled), and both CRC and framing errors > are _very_ common on "real world" networks. > > > Please reverse that part of the commit. If you're seeing so many of > > these errors that they're annoying you then you should find > the cause, > > not disable the error message. > > Often the cause is not fixable, or the net error rate is sufficiently > low that they can be ignored. The right thing to do is to increment > the input error count, and leave these messages as tunable options. I agree that error messages should be a tunable parameter, I'll look in to doing that. I don't think error messages should be brushed under the carpet though, I developed that driver on a busy network that showed few errors so it's not a truism that "real world" networks have lots of errors. I've also worked on networks that did show lots of errors, in those cases there were major network problems that needed fixing since network lag was a serious problem to those of us working there. These days I find a lot of network admins aren't really very knowledgable and spitting out errors at them is the only thing they'll take notice of. Telling them the network is slow when they can't see any problems on their console doesn't convince them :-( Paul Richards Ph.D. Originative Solutions Ltd To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?E40CBF0361C7D111914000C0F0303D108792>