[patch x86/mce] ACPI, x86: Export boot_cpu_physical_apicid to modules

[patch x86/mce] ACPI, x86: Export boot_cpu_physical_apicid to modules

[David Rientjes]

"ACPI, x86: Fix extended error log driver to depend on CONFIG_X86_LOCAL_APIC" 
fixed a build error when CONFIG_X86_LOCAL_APIC was not selected and !CONFIG_SMP.
However, since CONFIG_ACPI_EXTLOG is tristate, there is a second build error:

ERROR: "boot_cpu_physical_apicid" [drivers/acpi/acpi_extlog.ko] undefined!

The symbol needs to be exported for it to be available.

Signed-off-by: David Rientjes <rientjes@google.com>
---
 arch/x86/kernel/apic/apic.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c
--- a/arch/x86/kernel/apic/apic.c
+++ b/arch/x86/kernel/apic/apic.c
@@ -62,6 +62,7 @@ unsigned disabled_cpus;
 
 /* Processor that is doing the boot up */
 unsigned int boot_cpu_physical_apicid = -1U;
+EXPORT_SYMBOL(boot_cpu_physical_apicid);
 
 /*
  * The highest APIC ID seen during enumeration.

RE: [patch x86/mce] ACPI, x86: Export boot_cpu_physical_apicid to modules

[Luck, Tony]

> ERROR: "boot_cpu_physical_apicid" [drivers/acpi/acpi_extlog.ko] undefined!
>
> The symbol needs to be exported for it to be available.

Good - but I wonder how many more useless layers there are to this onion :-(

First I had to add a "#include <asm/cpu.h>"
Then add the dependency on CONFIG_X86_LOCAL_APIC
Now we need to export boot_cpu_physical_apicid

all for the CONFIG_SMP=n case ... which I think is useless because there
is no conceivable way anyone would run a uniprocessor kernel on a system
that supports the extended error log. (Hint: multi-socket server multiplied by
high-core-count processors multiplied by 2xHyperThread) is a number bigger
than one).

Perhaps we should just have:

	depends on SMP && X86_MCE

and stop the madness?

-Tony

RE: [patch x86/mce] ACPI, x86: Export boot_cpu_physical_apicid to modules

[David Rientjes]

On Thu, 14 Nov 2013, Luck, Tony wrote:

> > ERROR: "boot_cpu_physical_apicid" [drivers/acpi/acpi_extlog.ko] undefined!
> >
> > The symbol needs to be exported for it to be available.
> 
> Good - but I wonder how many more useless layers there are to this onion :-(
> 
> First I had to add a "#include <asm/cpu.h>"
> Then add the dependency on CONFIG_X86_LOCAL_APIC
> Now we need to export boot_cpu_physical_apicid
> 

asm/cpu.h itself is also indirectly relying on asm/topology.h to include 
asm/mpspec.h to get that declaration of boot_cpu_physical_apicid in the 
first place.

> all for the CONFIG_SMP=n case ... which I think is useless because there
> is no conceivable way anyone would run a uniprocessor kernel on a system
> that supports the extended error log. (Hint: multi-socket server multiplied by
> high-core-count processors multiplied by 2xHyperThread) is a number bigger
> than one).
> 
> Perhaps we should just have:
> 
> 	depends on SMP && X86_MCE
> 
> and stop the madness?
> 

Or do some randconfig testing before patches are proposed :)  I don't have 
any particular dependency on this, so I'm happy with whatever is decided.  
This is the last error I've found related to CONFIG_ACPI_EXTLOG, however.

RE: [patch x86/mce] ACPI, x86: Export boot_cpu_physical_apicid to modules

[Luck, Tony]

> Or do some randconfig testing before patches are proposed :)  I don't have 
> any particular dependency on this, so I'm happy with whatever is decided.  
> This is the last error I've found related to CONFIG_ACPI_EXTLOG, however.

If you *promise* this is the last one ,.. then

Acked-by: Tony Luck <tony.luck@intel.com>

-Tony

[tip:x86/urgent] x86: Export 'boot_cpu_physical_apicid' to modules

[tip-bot for David Rientjes]

Commit-ID:  cc08e04c3fdcaab767b8db27527002b2b4d758cb
Gitweb:     http://git.kernel.org/tip/cc08e04c3fdcaab767b8db27527002b2b4d758cb
Author:     David Rientjes <rientjes@google.com>
AuthorDate: Thu, 14 Nov 2013 15:05:32 -0800
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Fri, 15 Nov 2013 08:38:30 +0100

x86: Export 'boot_cpu_physical_apicid' to modules

Commit 9ebddac7ea2a "ACPI, x86: Fix extended error log driver to depend on
CONFIG_X86_LOCAL_APIC" fixed a build error when CONFIG_X86_LOCAL_APIC was not
selected and !CONFIG_SMP.

However, since CONFIG_ACPI_EXTLOG is tristate, there is a second build error:

  ERROR: "boot_cpu_physical_apicid" [drivers/acpi/acpi_extlog.ko] undefined!

The symbol needs to be exported for it to be available.

Signed-off-by: David Rientjes <rientjes@google.com>
Acked-by: Tony Luck <tony.luck@intel.com>
Cc: Chen Gong <gong.chen@linux.intel.com>
Cc: Rafael J. Wysocki <rjw@rjwysocki.net>
Link: http://lkml.kernel.org/r/alpine.DEB.2.02.1311141504080.30112@chino.kir.corp.google.com
[ Changed it to a _GPL() export. ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/kernel/apic/apic.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c
index a7eb82d..befe498 100644
--- a/arch/x86/kernel/apic/apic.c
+++ b/arch/x86/kernel/apic/apic.c
@@ -62,6 +62,7 @@ unsigned disabled_cpus;
 
 /* Processor that is doing the boot up */
 unsigned int boot_cpu_physical_apicid = -1U;
+EXPORT_SYMBOL_GPL(boot_cpu_physical_apicid);
 
 /*
  * The highest APIC ID seen during enumeration.