Skip to contentSkip to conversation

sparc: Provide cmpxchg64 for 32-bit builds

PATCH In review Guenter Roeck 2 participants 3 messages Nov 6, 2017 – Nov 9, 2017 SPARC + UltraSPARC (sparc/sparc64)

lore
Share & tools
lore
#0 lore raw

PATCH v3 1/1 sparc: Provide cmpxchg64 for 32-bit builds

Fix the following build error, seen when building sparc32:allmodconfig.

drivers/net/ethernet/intel/i40e/i40e_ethtool.c:
In function 'i40e_set_priv_flags':
drivers/net/ethernet/intel/i40e/i40e_ethtool.c:4150:2: error:
implicit declaration of function 'cmpxchg64'

---
v3: Drop extern from cmpxchg64 declaration in include file
v2: Actually implement cmpxchg64 with code snippet provided by Dave Miller
Note: I am aware that this doesn't pass checkpatch; I found it more important to be in line with __cmpxchg_u32().

 arch/sparc/include/asm/cmpxchg_32.h |  2 ++
 arch/sparc/lib/atomic32.c           | 14 ++++++++++++++
 2 files changed, 16 insertions(+)
arch/sparc/include/asm/cmpxchg_32.h +2 -0
index 83ffb83c5397..01bb66b5d125 100644
@@ -62,6 +62,8 @@ __cmpxchg(volatile void *ptr, unsigned long old, unsigned long new_, int size)
 			(unsigned long)_n_, sizeof(*(ptr)));		\
 })
 
+u64 cmpxchg64(u64 *ptr, u64 old, u64 new);
+
 #include <asm-generic/cmpxchg-local.h>
 
 /*
arch/sparc/lib/atomic32.c +14 -0
index 2c373329d5cb..41601eb540e7 100644
@@ -172,6 +172,20 @@ unsigned long __cmpxchg_u32(volatile u32 *ptr, u32 old, u32 new)
 }
 EXPORT_SYMBOL(__cmpxchg_u32);
 
+u64 cmpxchg64(u64 *ptr, u64 old, u64 new)
+{
+	unsigned long flags;
+	u64 prev;
+
+	spin_lock_irqsave(ATOMIC_HASH(ptr), flags);
+	if ((prev = *ptr) = old)
+		*ptr = new;
+	spin_unlock_irqrestore(ATOMIC_HASH(ptr), flags);
+
+	return prev;
+}
+EXPORT_SYMBOL(cmpxchg64);
+
 unsigned long __xchg_u32(volatile u32 *ptr, u32 new)
 {
 	unsigned long flags;
Signature · 1 line
2.7.4
maintainer
#1 lore raw

From: Guenter Roeck <[email protected]>
Date: Mon, 6 Nov 2017 06:41:45 -0800

Quoted 8 lines

Fix the following build error, seen when building sparc32:allmodconfig.

drivers/net/ethernet/intel/i40e/i40e_ethtool.c:
In function 'i40e_set_priv_flags':
drivers/net/ethernet/intel/i40e/i40e_ethtool.c:4150:2: error:
implicit declaration of function 'cmpxchg64'

Guenter, are you really not seeing this commit which has been in my sparc GIT tree for a while?

commit 23198ddffb6cddb5d5824230af4dd4b46e4046a4

Author: David S. Miller <[email protected]>
Date:   Wed Sep 27 22:38:19 2017 -0700

sparc32: Add cmpxchg64().

This fixes the build with i40e driver enabled.

arch/sparc/include/asm/cmpxchg_32.h +3 -0
index 83ffb83c5397..a0101a58fd3f 100644
@@ -62,6 +62,9 @@ __cmpxchg(volatile void *ptr, unsigned long old, unsigned long new_, int size)
 			(unsigned long)_n_, sizeof(*(ptr)));		\
 })
 
+u64 __cmpxchg_u64(u64 *ptr, u64 old, u64 new);
+#define cmpxchg64(ptr, old, new)	__cmpxchg_u64(ptr, old, new)
+
 #include <asm-generic/cmpxchg-local.h>
 
 /*
arch/sparc/lib/atomic32.c +14 -0
index 2c373329d5cb..ddacb5aeb424 100644
@@ -172,6 +172,20 @@ unsigned long __cmpxchg_u32(volatile u32 *ptr, u32 old, u32 new)
 }
 EXPORT_SYMBOL(__cmpxchg_u32);
 
+u64 __cmpxchg_u64(u64 *ptr, u64 old, u64 new)
+{
+	unsigned long flags;
+	u64 prev;
+
+	spin_lock_irqsave(ATOMIC_HASH(ptr), flags);
+	if ((prev = *ptr) = old)
+		*ptr = new;
+	spin_unlock_irqrestore(ATOMIC_HASH(ptr), flags);
+
+	return prev;
+}
+EXPORT_SYMBOL(__cmpxchg_u64);
+
 unsigned long __xchg_u32(volatile u32 *ptr, u32 new)
 {
 	unsigned long flags;
#2 lore raw
Quoted 10 lines · David Miller

From: Guenter Roeck <[email protected]>
Date: Mon, 6 Nov 2017 06:41:45 -0800

Quoted 7 lines

Fix the following build error, seen when building sparc32:allmodconfig.

drivers/net/ethernet/intel/i40e/i40e_ethtool.c:
In function 'i40e_set_priv_flags':
drivers/net/ethernet/intel/i40e/i40e_ethtool.c:4150:2: error:
implicit declaration of function 'cmpxchg64'

Quoted 5 lines
Quoted 1 line

Guenter, are you really not seeing this commit which has been in my sparc GIT tree for a while?

Dave,

At least as of yesterday, the patch had not found its way into mainline.

I am running kernel build tests in my spare time, as community service. I do spend time before I submit a report or a proposed patch trying to find if the problem or its fix was already posted, just to avoid the "you dummy, this was already fixed" type feedback. However, I typically don't repeat that if it appears that the original report or submission was ignored. I just don't have the time to do that.

So, no, I did not see the patch in your tree. Sorry for the noise. I'll refrain from reporting problems on sparc to avoid such noise in the future. I'll definitely no longer send out reminders.

Guenter