[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

pinning down... serial port: LSR safety check engaged



> > It has worked with an explicit 'skip_test' and the implicit test, with
> > both the 2.2.15 and 2.4.3 kernels.
>
> I wouldn't be surprised if it sometimes works without skip_test.  Has it
> ever failed with it?  Is skip_test in /etc/serial.conf?

Yes, during this and the previous round.  Not right now.


This is the best I can do to pin it down...

Looking at the timestamps of older serial.conf-s I can tell that I ran
with skip_test until problems cropped up with irq7 on the same port
(turned out to be the sound card), skip_test went out while I redid
isapnp to use irq3.  At some point I put skip_test back in until the
first episode of "LSR...", by then I had moved from Potato/2.2.15|17
to Woody/2.4.3, then to Sid/Woody (although I don't think there are
many woodies on the system anymore).

The problem first hit a few weeks after adding unstable to
sources.list, closer to when I started to fiddle with setserial via
the debconf interface, not clearly associated with an upgrade (i.e.,
I have toggled the connection a few times without problems in both
instances), and only after a reboot.

I may have triggered an unlikely bug with some combination of manual
and automated serial.conf fiddling at just the wrong time, coupled
with setserial changing the infrastructure for handling serial.conf
and isapnptool upgrades.

Looking at the changelogs shows the most recent upgrade which would
directly affect ttyS3 to be that of isapnp (May 18), but if that
borked things I should have noticed earlier 'cause libc6 has been
through a few quick upgrades since then (I reboot whenever libc gets
an upgrade). To make things interesting... isapnptools was also
upgraded (Apr 17) a little before I recall having LSR problems the
first time (before the end of April, but I destroyed the timestamp
evidence, no backup, oops), and I would have rebooted when xlibs got
an upgrade (Apr 29). I don't have a clue as to exactly when or how I
fiddled with serial.conf during this timeframe, but I did fiddle with
it.

So... the best correlation I can come up with is a new isapnp followed
by a reboot, but muddied up with a changing serial.conf.


I hope anyone who made it this far got something out of it...
like maybe keep more backups, logs, and scribbled notes.  :)

Thanks for all your help and suggestions.


- Bruce





Reply to: