Re: [PATCH 3/39] NLKD - early/late CPU up/down notification

From: Randy.Dunlap
Date: Sat Nov 12 2005 - 16:08:49 EST


On Fri, 11 Nov 2005 08:52:38 +0100 Jan Beulich wrote:

> >>> Sam Ravnborg <sam@xxxxxxxxxxxx> 10.11.05 21:59:31 >>>
> >> I understand that. But you don't see my point, so I'll try to
> explain
> >> the background: When discovering the reason for the kallsyms change
> >> (also posted with the other NLKD patches) not functioning with
> >> CONFIG_MODVERSIONS and binutils between 2.16.90 and 2.16.91.0.3 I
> >> realized that the warning messages from the modpost build stage are
> very
> >> easy to overlook (in fact, all reporters of the problem overlooked
> them
> >> as well as I did on the first build attempting to reproduce the
> >> problem). This basically means these messages are almost useless,
> and
> >> detection of the problem will likely be deferred to the first
> attempt to
> >> load an offending module (which, as in the case named, may lead to
> an
> >> unusable kernel). Hence, at least until this build problem gets
> >> addressed I continue to believe that adding the preprocessor
> conditional
> >> is the better way of dealing with potential issues.
> >
> >Can you elaborate a little what you like to have done to the build
> >process.
>
> As said above, I'd like to see the messages from modpost deferred or
> re-issued at the end of the module building process (i.e. after the
> perhaps hundreds of *.mod.c files got compiled and the same number of
> *.ko got linked, which on a typical distribution will easily scroll off
> the original warnings even with a 1000 line history, as seems to be the
> default in many cases).

Lots of such error or warning messages scroll away. All that it takes
is use of egrep "foo|bar" to look for the interesting messages
or egrep -v "foo|bar" to discard the uninteresting (normal) messages.
I do that all the time and I think that it's the right thing to do,
not have the tools regurgitate something that they have already said.

---
~Randy
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/