[RFC] ppc: don't override CONFIG_PPC_PSERIES_DEBUG

These files undef DEBUG, but I think they were added before the ability
to control this from Kconfig. It's really annoying to only get some of
the debug messages!

Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>

---
Because the lpar and pci_dlpar code is pretty low-level & verbose,
perhaps it makes sense to add another Kconfig variable for really
low-level stuff? But it's annoying to have DEBUG *somewhat* effective,
especially in the EEH area when doing PCI stuff.
---
 arch/powerpc/platforms/pseries/eeh.c       |    2 --
 arch/powerpc/platforms/pseries/lpar.c      |    3 ---
 arch/powerpc/platforms/pseries/pci_dlpar.c |    2 --
 3 files changed, 0 insertions(+), 7 deletions(-)

On Thu, 2010-10-14 at 10:48 -0700, Nishanth Aravamudan wrote:
> These files undef DEBUG, but I think they were added before the ability
> to control this from Kconfig. 

Perhaps. Some people, *cough*, have a tendency to merge those back in
again from time to time :)

> It's really annoying to only get some of the debug messages!

True, but ..

> Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
> 
> ---
> Because the lpar and pci_dlpar code is pretty low-level & verbose,
> perhaps it makes sense to add another Kconfig variable for really
> low-level stuff? But it's annoying to have DEBUG *somewhat* effective,
> especially in the EEH area when doing PCI stuff.

I really don't think you want to enable the lpar debug by default. Have
you tried it? It can make for a pretty unusable system, just because of
the console spam.

Also these days there is CONFIG_DYNAMIC_DEBUG which is much smarter than
all this, but requires setup at runtime.

cheers

On 15.10.2010 [11:14:23 +1100], Michael Ellerman wrote:
> On Thu, 2010-10-14 at 10:48 -0700, Nishanth Aravamudan wrote:
> > These files undef DEBUG, but I think they were added before the ability
> > to control this from Kconfig. 
> 
> Perhaps. Some people, *cough*, have a tendency to merge those back in
> again from time to time :)
> 
> > It's really annoying to only get some of the debug messages!
> 
> True, but ..
> 
> > Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
> > 
> > ---
> > Because the lpar and pci_dlpar code is pretty low-level & verbose,
> > perhaps it makes sense to add another Kconfig variable for really
> > low-level stuff? But it's annoying to have DEBUG *somewhat* effective,
> > especially in the EEH area when doing PCI stuff.
> 
> I really don't think you want to enable the lpar debug by default.
> Have you tried it? It can make for a pretty unusable system, just
> because of the console spam.

Yeah, you're right. After enabling it, I had to kill my boot and start
over w/o the lpar DEBUG on. I assume dlpar_pci is similar?

I dunno, would a patch to a least remove the EEH one be ok? Seems like
it isn't super-verbose, and does have some handy output.

> Also these days there is CONFIG_DYNAMIC_DEBUG which is much smarter than
> all this, but requires setup at runtime.

True, I started looking into it, but only realized today that eeh.c had
that #undef! :)

Thanks,
Nish

On Thu, 2010-10-14 at 17:23 -0700, Nishanth Aravamudan wrote:
> On 15.10.2010 [11:14:23 +1100], Michael Ellerman wrote:
> > On Thu, 2010-10-14 at 10:48 -0700, Nishanth Aravamudan wrote:
> > > Because the lpar and pci_dlpar code is pretty low-level & verbose,
> > > perhaps it makes sense to add another Kconfig variable for really
> > > low-level stuff? But it's annoying to have DEBUG *somewhat* effective,
> > > especially in the EEH area when doing PCI stuff.
> > 
> > I really don't think you want to enable the lpar debug by default.
> > Have you tried it? It can make for a pretty unusable system, just
> > because of the console spam.
> 
> Yeah, you're right. After enabling it, I had to kill my boot and start
> over w/o the lpar DEBUG on. 

:)

> I assume dlpar_pci is similar?

That should be OK to enable I think. Suck it and see I guess.

> I dunno, would a patch to a least remove the EEH one be ok? Seems like
> it isn't super-verbose, and does have some handy output.

Yeah definitely. That undef was merged as part of a cleanup/fix patch
but shouldn't have been.

cheers

On 14 October 2010 12:48, Nishanth Aravamudan <nacc@us.ibm.com> wrote:
> These files undef DEBUG, but I think they were added before the ability
> to control this from Kconfig.

Right.

> It's really annoying to only get some of
> the debug messages!

I don't get the big picture.  Will there be some CONFIG_DEBUG_EEH in Kconfig?
or just some option to turn on DEBUG for all powerpc-related files?

Or maybe I am demonstrating my utter ignorance of some new whiz-bang
Kconfig technology?

Anyway, I see no harm in the EEH portion of the patch.

--linas