Skip to contentSkip to conversation

some fixups for the X86_NUMAQ dependencies

PATCH In review Adrian Bunk 2 participants 6 messages Feb 19, 2006 – Mar 3, 2006

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 some fixups for the X86_NUMAQ dependencies

You must always ensure to fulfill the dependencies of what you are select'ing.

---

 arch/i386/Kconfig |    7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)
arch/i386/Kconfig renamed from arch/i386/Kconfig.old +3 -4
@@ -84,6 +84,7 @@
 
 config X86_NUMAQ
 	bool "NUMAQ (IBM/Sequent)"
+	select SMP
 	select NUMA
 	help
 	  This option is used for getting Linux to run on a (IBM/Sequent) NUMA
@@ -419,6 +420,7 @@
 
 config NOHIGHMEM
 	bool "off"
+	depends on !X86_NUMAQ
 	---help---
 	  Linux can use up to 64 Gigabytes of physical memory on x86 systems.
 	  However, the address space of 32-bit x86 processors is only 4
@@ -455,6 +457,7 @@
 
 config HIGHMEM4G
 	bool "4GB"
+	depends on !X86_NUMAQ
 	help
 	  Select this if you have a 32-bit processor and between 1 and 4
 	  gigabytes of physical RAM.
@@ -522,10 +525,6 @@
 	default n if X86_PC
 	default y if (X86_NUMAQ || X86_SUMMIT)
 
-# Need comments to help the hapless user trying to turn on NUMA support
-comment "NUMA (NUMA-Q) requires SMP, 64GB highmem support"
-	depends on X86_NUMAQ && (!HIGHMEM64G || !SMP)
-
 comment "NUMA (Summit) requires SMP, 64GB highmem support, ACPI"
 	depends on X86_SUMMIT && (!HIGHMEM64G || !ACPI)
 
#1 lore raw

arch/i386/Kconfig

Quoted 21 lines · Adrian Bunk

You must always ensure to fulfill the dependencies of what you are select'ing.

---

 arch/i386/Kconfig |    7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)
arch/i386/Kconfig renamed from arch/i386/Kconfig.old +1 -0
@@ -84,6 +84,7 @@
 
 config X86_NUMAQ
 	bool "NUMAQ (IBM/Sequent)"
+	select SMP
 	select NUMA
 	help
 	  This option is used for getting Linux to run on a (IBM/Sequent) NUMA
@@ -419,6 +420,7 @@

Surely NUMA should select SMP, not NUMA-Q?

Quoted 15 lines
arch/i386/Kconfig renamed from arch/i386/Kconfig.old +2 -0
 config NOHIGHMEM
 	bool "off"
+	depends on !X86_NUMAQ
 	---help---
 	  Linux can use up to 64 Gigabytes of physical memory on x86 systems.
 	  However, the address space of 32-bit x86 processors is only 4
@@ -455,6 +457,7 @@
 
 config HIGHMEM4G
 	bool "4GB"
+	depends on !X86_NUMAQ
 	help
 	  Select this if you have a 32-bit processor and between 1 and 4
 	  gigabytes of physical RAM.

arch/i386/Kconfig line 528

Quoted 7 lines
arch/i386/Kconfig renamed from arch/i386/Kconfig.old +0 -4
@@ -522,10 +525,6 @@
 	default n if X86_PC
 	default y if (X86_NUMAQ || X86_SUMMIT)
 
-# Need comments to help the hapless user trying to turn on NUMA support
-comment "NUMA (NUMA-Q) requires SMP, 64GB highmem support"
-	depends on X86_NUMAQ && (!HIGHMEM64G || !SMP)
-

Hmm. ISTR the reason we put that in there in the first place was that
NUMA-Q got mysteriously hidden by other deps before, and it wasn't clear how to select it. Perhaps we just had some of the deps backwards.

M.

#2 lore raw
Quoted 18 lines · Martin J. Bligh
Quoted 17 lines · Adrian Bunk

You must always ensure to fulfill the dependencies of what you are select'ing.

---

arch/i386/Kconfig | 7 +++----

1 file changed, 3 insertions(+), 4 deletions(-)

--- linux-2.6.16-rc3-mm1-full/arch/i386/Kconfig.old 2006-02-20
00:12:50.000000000 +0100
+++ linux-2.6.16-rc3-mm1-full/arch/i386/Kconfig 2006-02-20
00:17:57.000000000 +0100

(quoted diff) +0 -0
@@ -84,6 +84,7 @@
 
Quoted 10 lines

Review

Quoted 8 lines

config X86_NUMAQ
bool "NUMAQ (IBM/Sequent)"

arch/i386/Kconfig renamed from arch/i386/Kconfig.old +1 -0
+	select SMP

select NUMA
help
This option is used for getting Linux to run on a (IBM/Sequent) NUMA

(quoted diff) +0 -0
@@ -419,6 +420,7 @@

Surely NUMA should select SMP, not NUMA-Q?

NUMA depends on SMP.

Therefore, if you select NUMA, you have to ensure that SMP is enabled.

Quoted 18 lines
Quoted 18 lines

config NOHIGHMEM
bool "off"

(quoted diff) +1 -0
+	depends on !X86_NUMAQ

---help---
Linux can use up to 64 Gigabytes of physical memory on x86 systems. However, the address space of 32-bit x86 processors is only 4

(quoted diff) +0 -0
@@ -455,6 +457,7 @@
 

config HIGHMEM4G

	bool "4GB"
(quoted diff) +1 -0
+	depends on !X86_NUMAQ

help
Select this if you have a 32-bit processor and between 1 and 4 gigabytes of physical RAM.

arch/i386/Kconfig renamed from arch/i386/Kconfig.old +0 -0
@@ -522,10 +525,6 @@
 	default n if X86_PC
 	default y if (X86_NUMAQ || X86_SUMMIT)
 
Quoted 8 lines

arch/i386/Kconfig line 528

Quoted 4 lines
arch/i386/Kconfig renamed from arch/i386/Kconfig.old +0 -4
-# Need comments to help the hapless user trying to turn on NUMA support
-comment "NUMA (NUMA-Q) requires SMP, 64GB highmem support"
-	depends on X86_NUMAQ && (!HIGHMEM64G || !SMP)
-

Hmm. ISTR the reason we put that in there in the first place was that NUMA-Q got mysteriously hidden by other deps before, and it wasn't clear how to select it. Perhaps we just had some of the deps backwards.

NUMAQ can't be hidden since it doesn't has any dependencies. And this isn't what this comment is talking about (note the the comment is only shown if NUMAQ was already select'ed).

NUMAQ didn't fulfill the contract that when select'ing NUMA, it has to ensure the dependencies of NUMA are fulfilled. My patch solves this properly instead of telling the user through a comment that he ran into this bug.

Quoted 1 line

M.

cu
Adrian

Signature · 4 lines
       "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
#3 lore raw
Quoted 11 lines
Quoted 10 lines

Review

Quoted 8 lines

config X86_NUMAQ
bool "NUMAQ (IBM/Sequent)"

arch/i386/Kconfig renamed from arch/i386/Kconfig.old +1 -0
+	select SMP

select NUMA
help
This option is used for getting Linux to run on a (IBM/Sequent) NUMA

(quoted diff) +0 -0
@@ -419,6 +420,7 @@

Surely NUMA should select SMP, not NUMA-Q?

Quoted 3 lines

NUMA depends on SMP.

Therefore, if you select NUMA, you have to ensure that SMP is enabled.

Yes. but that should link SMP -> NUMA -> NUMA-Q, not SMP directly to NUMA-Q, surely?

Quoted 4 lines

NUMAQ can't be hidden since it doesn't has any dependencies.
And this isn't what this comment is talking about (note the the comment is only shown if NUMAQ was already select'ed).

NUMAQ didn't fulfill the contract that when select'ing NUMA, it has to ensure the dependencies of NUMA are fulfilled. My patch solves this properly instead of telling the user through a comment that he ran into this bug.

Yes, if that works, it's much cleaner. Perhaps we just had insufficient
config-fu to figure it out ... it looks good - I suppose I'd better test it, and make sure we don't hit the same thing we did before.

m.

#4 lore raw
Quoted 11 lines · Martin J. Bligh
Quoted 11 lines
Quoted 10 lines

Review

Quoted 8 lines

config X86_NUMAQ
bool "NUMAQ (IBM/Sequent)"

arch/i386/Kconfig renamed from arch/i386/Kconfig.old +1 -0
+	select SMP

select NUMA
help
This option is used for getting Linux to run on a (IBM/Sequent) NUMA

(quoted diff) +0 -0
@@ -419,6 +420,7 @@

Surely NUMA should select SMP, not NUMA-Q?

Quoted 6 lines
Quoted 3 lines

NUMA depends on SMP.

Therefore, if you select NUMA, you have to ensure that SMP is enabled.

Yes. but that should link SMP -> NUMA -> NUMA-Q, not SMP directly to NUMA-Q, surely?

The problem is that a select bypasses the dependencies of the select'ed symbol.

Quoted 4 lines
Quoted 4 lines

NUMAQ can't be hidden since it doesn't has any dependencies. And this isn't what this comment is talking about (note the the comment is only shown if NUMAQ was already select'ed).

Quoted 8 lines
Quoted 4 lines

NUMAQ didn't fulfill the contract that when select'ing NUMA, it has to ensure the dependencies of NUMA are fulfilled. My patch solves this properly instead of telling the user through a comment that he ran into this bug.

Yes, if that works, it's much cleaner. Perhaps we just had insufficient config-fu to figure it out ... it looks good - I suppose I'd better test it, and make sure we don't hit the same thing we did before.

:-)

Quoted 1 line

m.

cu
Adrian

Signature · 4 lines
       "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
#5 lore raw
In reply to Martin J. Bligh
Quoted 5 lines · Martin J. Bligh

...

Quoted 4 lines

NUMAQ can't be hidden since it doesn't has any dependencies. And this isn't what this comment is talking about (note the the comment is only shown if NUMAQ was already select'ed).

Quoted 8 lines
Quoted 4 lines

NUMAQ didn't fulfill the contract that when select'ing NUMA, it has to ensure the dependencies of NUMA are fulfilled. My patch solves this properly instead of telling the user through a comment that he ran into this bug.

Yes, if that works, it's much cleaner. Perhaps we just had insufficient config-fu to figure it out ... it looks good - I suppose I'd better test it, and make sure we don't hit the same thing we did before.

Have you tested my patch, and if yes, is it OK?

Quoted 1 line

m.

cu
Adrian

Signature · 4 lines
       "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