Re: [PATCH] more helpful WARN_ON and BUG_ON messages

[Adrian Bunk <bunk@stusta.de>]

On Fri, Oct 20, 2006 at 11:23:54AM -0500, Eric Sandeen wrote:
> After a few bugs I encountered in FC6 in buffer.c, with output like:
> 
> Kernel BUG at fs/buffer.c: 2791
> 
> where buffer.c contains:
> 
> ...
>         BUG_ON(!buffer_locked(bh));
>         BUG_ON(!buffer_mapped(bh));
>         BUG_ON(!bh->b_end_io);
> ...
> 
> around line 2790, it's awfully tedious to go get the exact failing kernel tree
> just to see -which- BUG_ON was encountered.
> 
> Printing out the failing condition as a string would make this more helpful IMHO.
> 
> This is mostly just compile-tested... comments?
>...

Who really needs this considering it implies a size increase of the 
kernel image?

Using a kernel tree so unusual that you can't locate the source anymore 
sounds like an extremely rare and unintelligent situation, not something 
that must be handled.

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed