Re: [patch for-linus] memcg, kmem: fix build error when CONFIG_INET is disabled

[Fengguang Wu <fengguang.wu@intel.com>]

On Wed, Oct 10, 2012 at 01:29:35PM +0400, Glauber Costa wrote:
> On 10/10/2012 01:27 PM, Michal Hocko wrote:
> > On Wed 10-10-12 12:56:26, Glauber Costa wrote:
> >> On 10/10/2012 10:32 AM, David Rientjes wrote:
> >>> Commit e1aab161e013 ("socket: initial cgroup code.") causes a build error 
> >>> when CONFIG_INET is disabled in Linus' tree:
> >>>
> >> unlikely that something that old would cause a build bug now, specially
> >> that commit, that actually wraps things inside CONFIG_INET.
> >>
> >> More likely caused by the recently merged
> >> "memcg-cleanup-kmem-tcp-ifdefs.patch" in -mm by mhocko (CC'd)
> > 
> > Strange it didn't trigger during my (and Fenguang) build testing.
> 
> Fengguang mentioned to me while testing my kmemcg tree that were a build
> error occurring in the base tree, IOW, yours.

Yes, and the errors were sent to/cc Michal and kernel-janitors@vger.kernel.org

> Fengguang, was that this error? Why hasn't it showed up before in the
> test system?

I do find this error in the build error log:

        (.text+0x867f): undefined reference to `sock_update_memcg'
        2012-09-24 04:54:53 snb next:akpm:69921c3 x86_64-randconfig-s005 0a7f618

Unfortunately it was not reported because the build system could
miss/ignore build bugs due to various reasons/imperfections. It has
since then undergo lots of enhancements and as a result, the daily
reported errors have more than doubled. :-)

Thanks,
Fengguang

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>