LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

Re: Error building 2.6.10 kernel with ip_vs_nfct patch - doesanyone else

To: LVS List <lvs-users@xxxxxxxxxxxxxxxxxxxxxx>
Subject: Re: Error building 2.6.10 kernel with ip_vs_nfct patch - doesanyone else get this?
From: "Vince W." <listacct1@xxxxxxxxxx>
Date: Sat, 05 Feb 2005 06:12:14 -0500
Vince W. wrote:

CC [M]  net/ipv4/ipvs/ip_vs_nfct.o

net/ipv4/ipvs/ip_vs_nfct.c: In function `ip_vs_nfct_conn_drop':
include/linux/netfilter_ipv4/ip_conntrack.h:248: sorry, unimplemented: inlining failed in call to 'ip_conntrack_put': function body not available
net/ipv4/ipvs/ip_vs_nfct.c:385: sorry, unimplemented: called from here
make[3]: *** [net/ipv4/ipvs/ip_vs_nfct.o] Error 1
make[2]: *** [net/ipv4/ipvs] Error 2

Has anyone else seen this, using this patch?

Allright well from what I've seen of others seeing error messages like this trying to compile stuff, it looks like the common thread is a 3.4.2-series GCC (GNU C Compiler), which is apparently less forgiving about calls to inline functions if the function body is not available. I found one result googling around where someone said they got a successful compile by removing "inline" from the function being called. But I suspect that removing "inline" from the 'ip_conntrack_put' function declaration in ip_conntrack.h would have various detrimental effects on the intended design of the function everywhere else it is used, and considering that the rest of the kernel seems to be able to compile just fine with "inline" in there, I am inclined to suspect something isn't kosher with the patch. But I'm not a programmer, so I wouldn't really know. It just seems odd that a patch written for such a new kernel wouldn't compile with a compiler of the same relative vintage.

Oh well. Such is life. I should focus on attainable goals. Next project....









<Prev in Thread] Current Thread [Next in Thread>