[PATCH] s390: remove HAVE_MARCH_Z9?0_FEATURES

[PATCH] s390: remove HAVE_MARCH_Z9?0_FEATURES

[Paul Bolle]

The Kconfig symbols HAVE_MARCH_Z900_FEATURES and
HAVE_MARCH_Z990_FEATURES were added in v3.5. They have never been used.
Apparently they are not needed. They can safely be removed.

Signed-off-by: Paul Bolle <pebolle@tiscali.nl>
---
Tested only with "git grep".

 arch/s390/Kconfig | 10 ----------
 1 file changed, 10 deletions(-)

diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig
index 65a0775..b853d0d 100644
--- a/arch/s390/Kconfig
+++ b/arch/s390/Kconfig
@@ -152,16 +152,8 @@ source "kernel/Kconfig.freezer"
 
 menu "Processor type and features"
 
-config HAVE_MARCH_Z900_FEATURES
-	def_bool n
-
-config HAVE_MARCH_Z990_FEATURES
-	def_bool n
-	select HAVE_MARCH_Z900_FEATURES
-
 config HAVE_MARCH_Z9_109_FEATURES
 	def_bool n
-	select HAVE_MARCH_Z990_FEATURES
 
 config HAVE_MARCH_Z10_FEATURES
 	def_bool n
@@ -188,7 +180,6 @@ config MARCH_G5
 
 config MARCH_Z900
 	bool "IBM zSeries model z800 and z900"
-	select HAVE_MARCH_Z900_FEATURES if 64BIT
 	help
 	  Select this to enable optimizations for model z800/z900 (2064 and
 	  2066 series). This will enable some optimizations that are not
@@ -196,7 +187,6 @@ config MARCH_Z900
 
 config MARCH_Z990
 	bool "IBM zSeries model z890 and z990"
-	select HAVE_MARCH_Z990_FEATURES if 64BIT
 	help
 	  Select this to enable optimizations for model z890/z990 (2084 and
 	  2086 series). The kernel will be slightly faster but will not work
-- 
1.8.5.3

Re: [PATCH] s390: remove HAVE_MARCH_Z9?0_FEATURES

[Heiko Carstens]

On Sun, Feb 16, 2014 at 06:15:27PM +0100, Paul Bolle wrote:
> The Kconfig symbols HAVE_MARCH_Z900_FEATURES and
> HAVE_MARCH_Z990_FEATURES were added in v3.5. They have never been used.
> Apparently they are not needed. They can safely be removed.
> 

No, I want to keep them, so there is a config symbol for each cpu generation
and we don't have to add the config symbols again, if we actually use them.
It's a question of completeness in this case, not if they are being used.

Thanks,
Heiko

Re: [PATCH] s390: remove HAVE_MARCH_Z9?0_FEATURES

[Paul Bolle]

On Mon, 2014-02-17 at 09:20 +0100, Heiko Carstens wrote:
> On Sun, Feb 16, 2014 at 06:15:27PM +0100, Paul Bolle wrote:
> No, I want to keep them, so there is a config symbol for each cpu generation
> and we don't have to add the config symbols again, if we actually use them.

Yes, that's the idea: only add Kconfig symbols that you are actually
going to use.

> It's a question of completeness in this case, not if they are being used.

Completeness of what?

Anyhow, perhaps you can comment this stuff out if you want to have it in
the tree. See, these symbols will pop up each time that people run their
scripts to discover problems in the kconfig space. I suppose this was
probably not the first time these two were reported. And people not
aware of this thread might do that again, and again...

Thanks,


Paul Bolle

Re: [PATCH] s390: remove HAVE_MARCH_Z9?0_FEATURES

[Heiko Carstens]

On Mon, Feb 17, 2014 at 09:50:40AM +0100, Paul Bolle wrote:
> On Mon, 2014-02-17 at 09:20 +0100, Heiko Carstens wrote:
> > On Sun, Feb 16, 2014 at 06:15:27PM +0100, Paul Bolle wrote:
> > No, I want to keep them, so there is a config symbol for each cpu generation
> > and we don't have to add the config symbols again, if we actually use them.
> 
> Yes, that's the idea: only add Kconfig symbols that you are actually
> going to use.
> 
> > It's a question of completeness in this case, not if they are being used.
> 
> Completeness of what?

Each cpu generation we support is supposed to have a
"HAVE_MARCH_<GENERATION>_FEATURES" config symbol.

> Anyhow, perhaps you can comment this stuff out if you want to have it in
> the tree. See, these symbols will pop up each time that people run their
> scripts to discover problems in the kconfig space. I suppose this was
> probably not the first time these two were reported. And people not
> aware of this thread might do that again, and again...

What script did you use?

Re: [PATCH] s390: remove HAVE_MARCH_Z9?0_FEATURES

[Paul Bolle]

On Mon, 2014-02-17 at 10:10 +0100, Heiko Carstens wrote:
> On Mon, Feb 17, 2014 at 09:50:40AM +0100, Paul Bolle wrote:
> > Completeness of what?
> 
> Each cpu generation we support is supposed to have a
> "HAVE_MARCH_<GENERATION>_FEATURES" config symbol.

But apparently these two generations can do fine without it.

> > Anyhow, perhaps you can comment this stuff out if you want to have it in
> > the tree. See, these symbols will pop up each time that people run their
> > scripts to discover problems in the kconfig space. I suppose this was
> > probably not the first time these two were reported. And people not
> > aware of this thread might do that again, and again...
> 
> What script did you use?

A local piece of perl.

It started rather cute but it has grown quite big now. I never let it
leaves its cage. I just push some food between the bars every now and
then.


Paul Bolle