2.6.18-rc5-mm1, make oldconfig from 2.6.18-rc5 destroys LVM

2.6.18-rc5-mm1, make oldconfig from 2.6.18-rc5 destroys LVM

[John Stoffel]

Andrew,

When I do a make oldconfig under 2.6.18-rc5-mm1, with my working
2.6.18-rc5 .config file (appended below), all the LVM and MD stuff
gets blown away silently.  It looks like the drivers/md/Kconfig didn't
get updated properly when it was tweaked, possibly the new options to
get rid of BLOCK devices, or the ripping out of LVM1 stuff.

It looks like instead of 'if CONFIG_BLOCK' at the top, it just needs
to be 'if BLOCK' instead.  

And I'd really suggest that it NOT be this silly name BLOCK, something
more meaningful, like USE_BLOCK_DEVICES or something equally useful to
parse.

John

-- 
VGER BF report: U 0.499836

Re: 2.6.18-rc5-mm1, make oldconfig from 2.6.18-rc5 destroys LVM

[Andrew Morton]

On Sat, 2 Sep 2006 20:49:04 -0400
"John Stoffel" <john@stoffel.org> wrote:

> 
> Andrew,
> 
> When I do a make oldconfig under 2.6.18-rc5-mm1, with my working
> 2.6.18-rc5 .config file (appended below), all the LVM and MD stuff
> gets blown away silently.  It looks like the drivers/md/Kconfig didn't
> get updated properly when it was tweaked, possibly the new options to
> get rid of BLOCK devices, or the ripping out of LVM1 stuff.
> 
> It looks like instead of 'if CONFIG_BLOCK' at the top, it just needs
> to be 'if BLOCK' instead.  

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.18-rc5/2.6.18-rc5-mm1/hot-fixes/
contains a fix for this.

> And I'd really suggest that it NOT be this silly name BLOCK, something
> more meaningful, like USE_BLOCK_DEVICES or something equally useful to
> parse.

mm...  I think CONFIG_BLOCK is a reasonable compromise between the needs for
brevity and understandability.

-- 
VGER BF report: H 0.277375

Re: 2.6.18-rc5-mm1, make oldconfig from 2.6.18-rc5 destroys LVM

[John Stoffel]

>>>>> "Andrew" == Andrew Morton <akpm@osdl.org> writes:

>> And I'd really suggest that it NOT be this silly name BLOCK, something
>> more meaningful, like USE_BLOCK_DEVICES or something equally useful to
>> parse.

Andrew> mm...  I think CONFIG_BLOCK is a reasonable compromise between
Andrew> the needs for brevity and understandability.

Well, maybe the hotfix patch converts all the uses of 'if BLOCK' to
'if CONFIG_BLOCK', but I still think that 'if USE_BLOCK_DEVS' would be
more clear.  It's not like it should be all *that* used outside of
config stuff.

Oh well, thanks for your reply, now I can test 2.6.18-rc5-mm1 and see
if my IRQ problem is really there or not.

John

-- 
VGER BF report: U 0.5