On Mon, May 15, 2017 at 5:57 AM, Arnd Bergmann <arnd@xxxxxxxx> wrote: > On Sun, May 14, 2017 at 3:48 PM, kernelci.org bot <bot@xxxxxxxxxxxx> wrote: >> >> stable/linux-4.10.y build: 203 builds: 3 failed, 200 passed, 3 errors, 5 warnings (v4.10.16) >> Full Build Summary: https://kernelci.org/build/stable/branch/linux-4.10.y/kernel/v4.10.16/ >> Tree: stable >> Branch: linux-4.10.y >> Git Describe: v4.10.16 >> Git Commit: 6e8e9958691907e8d7eb3b2107619dddbdaeb175 >> Git URL: http://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git >> Built: 4 unique architectures >> >> Build Failures Detected: >> >> arm: gcc version 5.3.1 20160412 (Linaro GCC 5.3-2016.05) >> spear3xx_defconfig FAIL >> spear6xx_defconfig FAIL >> tct_hammer_defconfig FAIL >> Errors summary: >> 3 net/core/skbuff.c:1575:37: error: 'sock_edemux' undeclared (first use in this function) >> Warnings summary: > > This is a regression against v4.10.15, caused by the backport of > c21b48cc1bbf ("net: adjust skb->truesize in ___pskb_trim()") by Eric > Dumazet. > > Part of another commit that Eric did earlier fixes it: > > 158f323b9868 ("net: adjust skb->truesize in pskb_expand_head()") > --- a/include/net/sock.h > +++ b/include/net/sock.h > @@ -1534,7 +1534,7 @@ void sock_efree(struct sk_buff *skb); > #ifdef CONFIG_INET > void sock_edemux(struct sk_buff *skb); > #else > -#define sock_edemux(skb) sock_efree(skb) > +#define sock_edemux sock_efree > #endif > > int sock_setsockopt(struct socket *sock, int level, int op, > > This commit is not marked 'Cc: stable' upstream, but is referenced > in the one that was backported and looks like it might be appropriate > for stable as well. Eric, can you clarify? > Hi Arnd Yes, it looks like stable would need this small extract from 158f323b9868, removing the parameter from sock_edemux macro Thanks.