The generic implementation of of_node_to_nid is EXPORT_SYMBOL.
The powerpc implementation added by following commit is EXPORT_SYMBOL_GPL.
commit 953039c8df7b ("[PATCH] powerpc: Allow devices to register with numa
topology")
This creates an inconsistency for of_node_to_nid callers across
architectures.
Update the powerpc implementation to be exported consistently with the
generic implementation.
Signed-off-by: Shailendra Singh <redacted>
Reviewed-by: Andy Ritger <redacted>
---
arch/powerpc/mm/numa.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/powerpc/mm/numa.c b/arch/powerpc/mm/numa.c
index b1099cb2f393..8aa4ca3c84c9 100644
--- a/arch/powerpc/mm/numa.c
+++ b/arch/powerpc/mm/numa.c
@@ -290,7 +290,7 @@ int of_node_to_nid(struct device_node *device)
return nid;
}
-EXPORT_SYMBOL_GPL(of_node_to_nid);
+EXPORT_SYMBOL(of_node_to_nid);
static int __init find_min_common_depth(void)
{--
2.4.11
Shailendra Singh [off-list ref] writes:
The generic implementation of of_node_to_nid is EXPORT_SYMBOL.
True. Added in 298535c00a2c, in April 2016.
The powerpc implementation added by following commit is EXPORT_SYMBOL_GPL.
commit 953039c8df7b ("[PATCH] powerpc: Allow devices to register with numa
topology")
Which was merged in May 2006.
So the powerpc version has almost 10 years precedence.
This creates an inconsistency for of_node_to_nid callers across
architectures.
Update the powerpc implementation to be exported consistently with the
generic implementation.
But I guess it's a pretty boring API. So I'll merge this unless anyone
else objects.
cheers
quoted hunk
Signed-off-by: Shailendra Singh <redacted>
Reviewed-by: Andy Ritger <redacted>
---
arch/powerpc/mm/numa.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/powerpc/mm/numa.c b/arch/powerpc/mm/numa.c
index b1099cb2f393..8aa4ca3c84c9 100644
--- a/arch/powerpc/mm/numa.c
+++ b/arch/powerpc/mm/numa.c
@@ -290,7 +290,7 @@ int of_node_to_nid(struct device_node *device)
return nid;
}
-EXPORT_SYMBOL_GPL(of_node_to_nid);
+EXPORT_SYMBOL(of_node_to_nid);
static int __init find_min_common_depth(void)
{--
2.4.11
Hi Michael,
quoted
quoted
Which was merged in May 2006. So the powerpc version has almost 10 yea=
rs precedence.
=20Agreed but keeping the licensing consistent will help drivers making s=
eamless use of it on all architectures.
quoted
quoted
But I guess it's a pretty boring API. So I'll merge this unless anyon=
e else objects.
Thanks. Appreciate your quick response and help.
Thanks,
Shailendra
-----Original Message-----
From: Michael Ellerman [mailto:mpe@ellerman.id.au]=20
Sent: Thursday, February 09, 2017 4:19 AM
To: Shailendra Singh <redacted>; linuxppc-dev@lists.ozlabs.=
org
Cc: jk@ozlabs.org; apopple@au1.ibm.com; balbirs@au1.ibm.com; Andy Ritger =
[off-list ref]; John Hubbard [off-list ref]; Sherry Cheung <=
SCheung@nvidia.com>; Aruna Manjunatha [off-list ref]; John McKe=
nna [off-list ref]; david.daney@cavium.com; robh@kernel.org; will.=
deacon@arm.com; Arnd Bergmann [off-list ref]
Subject: Re: [PATCH] powerpc: Fix inconsistent of_node_to_nid EXPORT_SYMB=
OL handling
Shailendra Singh [off-list ref] writes:
The generic implementation of of_node_to_nid is EXPORT_SYMBOL.
True. Added in 298535c00a2c, in April 2016.
The powerpc implementation added by following commit is EXPORT_SYMBOL_G=
PL.
commit 953039c8df7b ("[PATCH] powerpc: Allow devices to register with=20
numa
topology")
Which was merged in May 2006.
So the powerpc version has almost 10 years precedence.
[Shailendra] - Agreed. Keeping the licensing consistent will help drivers=
=20making use of it on all architectures.
This creates an inconsistency for of_node_to_nid callers across=20
architectures.
Update the powerpc implementation to be exported consistently with the =
generic implementation.
But I guess it's a pretty boring API. So I'll merge this unless anyone el=
se objects.
[Shailendra] - Thanks. Appreciate your quick response.
cheers
quoted hunk
Signed-off-by: Shailendra Singh <redacted>
Reviewed-by: Andy Ritger <redacted>
---
arch/powerpc/mm/numa.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/powerpc/mm/numa.c b/arch/powerpc/mm/numa.c index=20
b1099cb2f393..8aa4ca3c84c9 100644
--- a/arch/powerpc/mm/numa.c
+++ b/arch/powerpc/mm/numa.c
@@ -290,7 +290,7 @@ int of_node_to_nid(struct device_node *device)
=20
return nid;
}
-EXPORT_SYMBOL_GPL(of_node_to_nid);
+EXPORT_SYMBOL(of_node_to_nid);
=20
static int __init find_min_common_depth(void) {
--
2.4.11
-------------------------------------------------------------------------=
----------
This email message is for the sole use of the intended recipient(s) and m=
ay contain
confidential information. Any unauthorized review, use, disclosure or di=
stribution
is prohibited. If you are not the intended recipient, please contact the=
=20sender by
reply email and destroy all copies of the original message.
-------------------------------------------------------------------------=
----------
On Wed, 2017-02-01 at 22:52:42 UTC, Shailendra Singh wrote:
The generic implementation of of_node_to_nid is EXPORT_SYMBOL.
The powerpc implementation added by following commit is EXPORT_SYMBOL_GPL.
commit 953039c8df7b ("[PATCH] powerpc: Allow devices to register with numa
topology")
This creates an inconsistency for of_node_to_nid callers across
architectures.
Update the powerpc implementation to be exported consistently with the
generic implementation.
Signed-off-by: Shailendra Singh <redacted>
Applied to powerpc next, thanks.
https://git.kernel.org/powerpc/c/be9ba9ff93cc3e44dc46da9ed25655
cheers