Re: [2.6.11-rc5-mm1 patch] reiser4 Kconfig help cleanup

[Andrew Morton <akpm@osdl.org>]

Jes Sorensen <jes@wildopensource.com> wrote:
>
> >>>>> "Adrian" == Adrian Bunk <bunk@stusta.de> writes:
> 
> Adrian> The current reiser4 help texts have two disadvantages: 1. they
> Adrian> are more marketing speech than technical speech with some
> Adrian> debatable statements 2. they are too long
> 
> Excellent patch, that help description has been totally inappropriate
> since it was first introduced. I'm sure it will do fine on namesys'
> website, but not in the kernel.
> 
> Adrian> Signed-off-by: Adrian Bunk <bunk@stusta.de>
> 
> Signed-off-by: Jes Sorensen <jes@wildopensource.com>
> 
> Speaking of inappropriate components in reiser4:
> 
> [jes@tomahawk linux-2.6.11-rc5-mm1]$ grep PG_arch fs/reiser4/*.c
> fs/reiser4/page_cache.c:               page_flag_name(page, PG_arch_1),
> fs/reiser4/txnmgr.c:                    assert("vs-1448", test_and_clear_bit(PG_arch_1, &node->pg->flags));
> fs/reiser4/txnmgr.c:            ON_DEBUG(set_bit(PG_arch_1, &(copy->pg)->flags));
> 
> Someone was obviously smoking something illegal, what part of 'arch'
> did she/he not understand? I assume we can request this is fixed by
> the patch owner asap.
> 

Could the reiserfs team please comment?

If it's just debug then probably it would be better to add a new flag.

If these pages are never mmapped then it'll just happen to work, I guess. 
But a filesystem really shouldn't be dinking with PG_arch_1.