Date: Wed, 12 Jun 2002 16:20:10 +1000 (EST) From: Bruce Evans <bde@zeta.org.au> To: Jacob Ritorto <jritorto@pit.comms.marconi.com> Cc: freebsd-hardware@FreeBSD.ORG Subject: Re: serial trouble on dell cpi 366 Message-ID: <20020612155759.W596-100000@gamplex.bde.org> In-Reply-To: <Pine.GSO.4.42.0206111051290.2861-100000@gravel>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 11 Jun 2002, Jacob Ritorto wrote: > I'm frequently losing bits on my laptop's > serial port using FreeBSD 4.5-RELEASE and > 5.0-CURRENT. The kernel frequently reports one > more silo overflow (xx total) on the serial port. > Both installations are generic kernels, though I > did try cutting out superfluous devices in 5.0, > but to no effect. > 5.0 Seems to drop many more bits than 4.5, > and 5.0 seems not to care how fast the transfer > happens -- it loses data even at 2400 bps, while > the 4.5-RELEASE usually only loses data at rates > above 38400. > Do people already know about this? Is > there a bug filed yet? Any resolution / > workaround? I use the machine (and camediaplay or I think this is unknown, except that interrupt latency problems are worse in -current. > gphoto -- same behaviour with both programs) to > grab digital photos from my Olympus. The scheme > worked fine on previous FreeBSD versions with my > now defunct Intel machine.. It's interesting that it used to work. The only significant change in the driver between 4.4 and 4.5 was to reduce the input fifo (silo) trigger level. In theory, this should reduce silo overruns by up to a factor of 4, but since there should be none without this change, this change should make no difference except to reduce efficiency. Silo overruns at low speeds like 38400 point to deeper problems. Try running 4.4 to see if this is a software problem. > Please advise as it seems to me that the > kernel is having trouble getting around to > servicing the serial port's requests, which I'm > assuming denotes the kernel's failure to fully > grok the uart or something like that.. Silo overflows are more often caused by a software bug, or other hardware using so many bus cycles that the serial driver can't access its hardware in time. Bruce To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hardware" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020612155759.W596-100000>