[PATCH net] vxlan: fix sparse warnings

[PATCH net] vxlan: fix sparse warnings

[Jiri Benc]

Sparse reports false positives for the header manipulation inlines. Annotate
them correctly.

Tested by sparse on a little endian and big endian machine.

Fixes: 54bfd872bf16d ("vxlan: keep flags and vni in network byte order")
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Jiri Benc <jbenc@redhat.com>
---
As a side note, I tried to simplify all of these functions by using nested
be32_to_cpu(cpu_to_be32()) calls but GCC seems not to be clever enough to
optimize the byteswaps out.
---
 include/net/vxlan.h | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)

diff --git a/include/net/vxlan.h b/include/net/vxlan.h
index a763c96ecde4..73ed2e951c02 100644
--- a/include/net/vxlan.h
+++ b/include/net/vxlan.h
@@ -271,36 +271,36 @@ static inline struct vxlanhdr *vxlan_hdr(struct sk_buff *skb)
 static inline __be32 vxlan_vni(__be32 vni_field)
 {
 #if defined(__BIG_ENDIAN)
-	return vni_field >> 8;
+	return (__force __be32)((__force u32)vni_field >> 8);
 #else
-	return (vni_field & VXLAN_VNI_MASK) << 8;
+	return (__force __be32)((__force u32)(vni_field & VXLAN_VNI_MASK) << 8);
 #endif
 }
 
 static inline __be32 vxlan_vni_field(__be32 vni)
 {
 #if defined(__BIG_ENDIAN)
-	return vni << 8;
+	return (__force __be32)((__force u32)vni << 8);
 #else
-	return vni >> 8;
+	return (__force __be32)((__force u32)vni >> 8);
 #endif
 }
 
 static inline __be32 vxlan_tun_id_to_vni(__be64 tun_id)
 {
 #if defined(__BIG_ENDIAN)
-	return tun_id;
+	return (__force __be32)tun_id;
 #else
-	return tun_id >> 32;
+	return (__force __be32)((__force u64)tun_id >> 32);
 #endif
 }
 
 static inline __be64 vxlan_vni_to_tun_id(__be32 vni)
 {
 #if defined(__BIG_ENDIAN)
-	return (__be64)vni;
+	return (__force __be64)vni;
 #else
-	return (__be64)vni << 32;
+	return (__force __be64)((u64)(__force u32)vni << 32);
 #endif
 }
 
-- 
1.8.3.1

Re: [PATCH net] vxlan: fix sparse warnings

[David Miller]

From: Jiri Benc <jbenc@redhat.com>
Date: Mon, 21 Mar 2016 17:39:18 +0100

> Sparse reports false positives for the header manipulation inlines. Annotate
> them correctly.
> 
> Tested by sparse on a little endian and big endian machine.
> 
> Fixes: 54bfd872bf16d ("vxlan: keep flags and vni in network byte order")
> Reported-by: kbuild test robot <fengguang.wu@intel.com>
> Signed-off-by: Jiri Benc <jbenc@redhat.com>

Applied, thanks Jiri.

> As a side note, I tried to simplify all of these functions by using nested
> be32_to_cpu(cpu_to_be32()) calls but GCC seems not to be clever enough to
> optimize the byteswaps out.

Ok, thanks for the note, thanks again.