x86: revert X86_HT semantics change

PATCH In review Adrian Bunk 4 participants 8 messages Nov 15, 2007 – Dec 4, 2007 x86

lore

Share & tools

lore

Adrian Bunk

author

2007-11-15 16:17 UTC #0 lore raw

PATCH 1/1 x86: revert X86_HT semantics change

The x86 unification resulted in CONFIG_X86_HT no longer being set if (X86_32 && MK8).

After grep'ing through the tree I think the problem is that different places have different assumptions about the semantics of CONFIG_X86_HT, either:
- hyperthreading or
- multicore
and the SCHED_SMT and SCHED_MC dependencies are just one of the symptoms.

This should be sorted out properly, but until then we should keep the 2.6.23 status quo.

Signed-off-by Adrian Bunk <[email protected]>

---

 arch/x86/Kconfig |    7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

18fef247715449d36438fe4de59fdf430f1b5c37

arch/x86/Kconfig +4 -3

index 1eb5997..2c357f0 100644
@@ -148,7 +148,8 @@ config X86_SMP
 
 config X86_HT
 	bool
-	depends on SMP && !(X86_VISWS || X86_VOYAGER || MK8)
+	depends on SMP
+	depends on (X86_32 && !(X86_VISWS || X86_VOYAGER)) || (X86_64 && !MK8)
 	default y
 
 config X86_BIOS_REBOOT

@@ -476,7 +477,7 @@ config NR_CPUS
 
 config SCHED_SMT
 	bool "SMT (Hyperthreading) scheduler support"
-	depends on (X86_64 && SMP) || (X86_32 && X86_HT)
+	depends on X86_HT
 	help
 	  SMT scheduler support improves the CPU scheduler's decision making
 	  when dealing with Intel Pentium 4 chips with HyperThreading at a

@@ -485,7 +486,7 @@ config SCHED_SMT
 
 config SCHED_MC
 	bool "Multi-core scheduler support"
-	depends on (X86_64 && SMP) || (X86_32 && X86_HT)
+	depends on X86_HT
 	default y
 	help
 	  Multi-core scheduler support improves the CPU scheduler's decision

Sam Ravnborg

Acked-by on v1 1/1

2007-11-15 18:03 UTC #1 lore raw

Quoted 2 lines · Adrian Bunk

The x86 unification resulted in CONFIG_X86_HT no longer being set if (X86_32 && MK8).

Yup - my bad. I wrongly assumed MK8 was an X86_64 thing.

Thanks for fixing this.

Quoted 9 lines

After grep'ing through the tree I think the problem is that different places have different assumptions about the semantics of CONFIG_X86_HT, either:
- hyperthreading or
- multicore
and the SCHED_SMT and SCHED_MC dependencies are just one of the symptoms.

Quoted 4 lines

This should be sorted out properly, but until then we should keep the 2.6.23 status quo.
Signed-off-by Adrian Bunk <[email protected]>

Acked-by Sam Ravnborg <[email protected]>

I assume Thomas & Co will forward the patch.

Sam

Thomas Gleixner

2007-11-15 18:07 UTC #2 lore raw

Quoted 15 lines · Sam Ravnborg

Quoted 2 lines · Adrian Bunk
The x86 unification resulted in CONFIG_X86_HT no longer being set if (X86_32 && MK8).
Yup - my bad. I wrongly assumed MK8 was an X86_64 thing.
Thanks for fixing this.
Quoted 9 lines
After grep'ing through the tree I think the problem is that different places have different assumptions about the semantics of CONFIG_X86_HT, either:
- hyperthreading or
- multicore
and the SCHED_SMT and SCHED_MC dependencies are just one of the symptoms.

Quoted 7 lines

Quoted 4 lines
This should be sorted out properly, but until then we should keep the 2.6.23 status quo.
Signed-off-by Adrian Bunk <[email protected]>
Acked-by Sam Ravnborg <[email protected]>
I assume Thomas & Co will forward the patch.

After looking what does what and fixing it.

Right now this patch is not a 1:1 replacement of the .23 status quo, as it now makes SCHED_SMT and SCHED_HT depend on !MK8 for 64bit.

tglx

Adrian Bunk

author

2007-11-15 18:20 UTC #3 lore raw

PATCH 1/1 x86: revert X86_HT semantics change

Quoted 21 lines · Thomas Gleixner

Quoted 19 lines · Sam Ravnborg
Quoted 2 lines · Adrian Bunk
The x86 unification resulted in CONFIG_X86_HT no longer being set if (X86_32 && MK8).
Yup - my bad. I wrongly assumed MK8 was an X86_64 thing.
Thanks for fixing this.
Quoted 13 lines
After grep'ing through the tree I think the problem is that different places have different assumptions about the semantics of CONFIG_X86_HT, either:
- hyperthreading or
- multicore
and the SCHED_SMT and SCHED_MC dependencies are just one of the symptoms.
This should be sorted out properly, but until then we should keep the 2.6.23 status quo.
Signed-off-by Adrian Bunk <[email protected]>

Quoted 8 lines

Quoted 3 lines
Acked-by Sam Ravnborg <[email protected]>
I assume Thomas & Co will forward the patch.
After looking what does what and fixing it.
Right now this patch is not a 1:1 replacement of the .23 status quo, as it now makes SCHED_SMT and SCHED_HT depend on !MK8 for 64bit.

*searches brown paperbag*

Thanks for spotting, fixed patch below.

Quoted 1 line

tglx

cu
Adrian

<-- snip -->

The x86 unification resulted in CONFIG_X86_HT no longer being set if (X86_32 && MK8).

After grep'ing through the tree I think the problem is that different places have different assumptions about the semantics of CONFIG_X86_HT, either
- hyperthreading or
- multicore

This should be sorted out properly, but until then we should keep the 2.6.23 status quo.

Signed-off-by Adrian Bunk <[email protected]>

---
480a5ae48036f6825a5a111db87da05e298327d6

arch/x86/Kconfig +2 -1

index af1b1ca..5871b44 100644
@@ -148,7 +148,8 @@ config X86_SMP
 
 config X86_HT
 	bool
-	depends on SMP && !(X86_VISWS || X86_VOYAGER || MK8)
+	depends on SMP
+	depends on (X86_32 && !(X86_VISWS || X86_VOYAGER)) || (X86_64 && !MK8)
 	default y
 
 config X86_BIOS_REBOOT

Ingo Molnar

2007-12-04 13:29 UTC #4 lore raw

Quoted 9 lines · Adrian Bunk

The x86 unification resulted in CONFIG_X86_HT no longer being set if (X86_32 && MK8).
After grep'ing through the tree I think the problem is that different places have different assumptions about the semantics of CONFIG_X86_HT, either
- hyperthreading or
- multicore

Quoted 4 lines

This should be sorted out properly, but until then we should keep the 2.6.23 status quo.
Signed-off-by Adrian Bunk <[email protected]>

thanks, applied. This is for 2.6.24, right?

Ingo

Adrian Bunk

author

2007-12-04 13:58 UTC #5 lore raw

Quoted 12 lines · Ingo Molnar

Quoted 9 lines · Adrian Bunk
The x86 unification resulted in CONFIG_X86_HT no longer being set if (X86_32 && MK8).
After grep'ing through the tree I think the problem is that different places have different assumptions about the semantics of CONFIG_X86_HT, either
- hyperthreading or
- multicore

Quoted 6 lines

Quoted 4 lines
This should be sorted out properly, but until then we should keep the 2.6.23 status quo.
Signed-off-by Adrian Bunk <[email protected]>
thanks, applied. This is for 2.6.24, right?

Yes.

Quoted 1 line

Ingo

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

Sam Ravnborg

2007-11-15 18:41 UTC #6 lore raw

In reply to Thomas Gleixner

Quoted 21 lines · Thomas Gleixner

Quoted 19 lines · Sam Ravnborg
Quoted 2 lines · Adrian Bunk
The x86 unification resulted in CONFIG_X86_HT no longer being set if (X86_32 && MK8).
Yup - my bad. I wrongly assumed MK8 was an X86_64 thing.
Thanks for fixing this.
Quoted 13 lines
After grep'ing through the tree I think the problem is that different places have different assumptions about the semantics of CONFIG_X86_HT, either:
- hyperthreading or
- multicore
and the SCHED_SMT and SCHED_MC dependencies are just one of the symptoms.
This should be sorted out properly, but until then we should keep the 2.6.23 status quo.
Signed-off-by Adrian Bunk <[email protected]>

Quoted 8 lines

Quoted 3 lines
Acked-by Sam Ravnborg <[email protected]>
I assume Thomas & Co will forward the patch.
After looking what does what and fixing it.
Right now this patch is not a 1:1 replacement of the .23 status quo, as it now makes SCHED_SMT and SCHED_HT depend on !MK8 for 64bit.

I assumed this was intentional - Adrian?

Sam

Adrian Bunk

author

2007-11-15 18:48 UTC #7 lore raw

Quoted 24 lines · Sam Ravnborg

Quoted 23 lines · Thomas Gleixner
Quoted 21 lines · Sam Ravnborg
Quoted 2 lines · Adrian Bunk
The x86 unification resulted in CONFIG_X86_HT no longer being set if (X86_32 && MK8).
Yup - my bad. I wrongly assumed MK8 was an X86_64 thing.
Thanks for fixing this.
Quoted 9 lines
After grep'ing through the tree I think the problem is that different places have different assumptions about the semantics of CONFIG_X86_HT, either:
- hyperthreading or
- multicore
and the SCHED_SMT and SCHED_MC dependencies are just one of the symptoms.
Quoted 4 lines
This should be sorted out properly, but until then we should keep the 2.6.23 status quo.
Signed-off-by Adrian Bunk <[email protected]>
Acked-by Sam Ravnborg <[email protected]>

Quoted 8 lines

Quoted 6 lines
Quoted 1 line
I assume Thomas & Co will forward the patch.
After looking what does what and fixing it.
Right now this patch is not a 1:1 replacement of the .23 status quo, as it now makes SCHED_SMT and SCHED_HT depend on !MK8 for 64bit.
I assumed this was intentional - Adrian?

No, it was just buggy...

Quoted 1 line

Sam

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