On Mon, Aug 22, 2011 at 09:01:07PM -0700, Guenter Roeck wrote: > On Fri, Aug 19, 2011 at 07:05:50PM -0400, Guenter Roeck wrote: > > Hi Jean, > > > > On Fri, Aug 19, 2011 at 05:16:24PM -0400, Jean Delvare wrote: > > > Hi Guenter, > > > > > > On Thu, 18 Aug 2011 10:49:31 -0700, Guenter Roeck wrote: > > > > Executing BUG as a result of an internal driver error seems to be a bit harsh. > > > > > > Well, that's pretty much what BUG was designed for: internal errors > > > that aren't supposed to happen. > > > > > Problem with BUG is that it can be undefined for some configurations, > > in which case it doesn't do anything. > > > > Also, while it may make sense to bring down the system if there is a severe bug > > which makes it impossible to continue, that seems to be a bit excessive > > for a hwmon driver. > > > > > > Replace it with WARN and return -ENODEV if the condition is seen. > > > > > > I'm not sure what is the benefit of WARN over pr_warn() in this case. > > > > > WARN generates a traceback, pr_warn doesn't. I think the scope is a bit different > > - one warns about a programming error, the other warns about a system consition. > > > Hi all, > > As I don't see a consensus for my proposed patch, I'll drop it. Sorry, I'm just slow at testing things sometimes. :( In general I like the idea of replacing BUG with WARN when possible, particularly with hardware detection. Your patch seems to load ok on all the systems I can still claw together, so, Acked-by: Darrick J. Wong <djwong@xxxxxxxxxx> --D > > Guenter > > _______________________________________________ > lm-sensors mailing list > lm-sensors@xxxxxxxxxxxxxx > http://lists.lm-sensors.org/mailman/listinfo/lm-sensors _______________________________________________ lm-sensors mailing list lm-sensors@xxxxxxxxxxxxxx http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
