[Buildroot] [autobuild.buildroot.net] Your build results for 2019-02-14

[Thomas Petazzoni]

Arnout,

On Tue, 19 Feb 2019 23:14:33 +0100
Arnout Vandecappelle <arnout at mind.be> wrote:

> > Thomas Petazzoni <thomas.petazzoni at bootlin.com> writes:  
> >>          arm |                runc-v1.0.0-rc6 | http://autobuild.buildroot.net/results/63e9d88ae5177541be463f1e2aafec59aa410479
> >>         i686 |                runc-v1.0.0-rc6 | http://autobuild.buildroot.net/results/f71ddcd17f4991a3caeaac42c81ed97bdf847718
> >>          arm |                runc-v1.0.0-rc6 | http://autobuild.buildroot.net/results/64ecdb1e007106fdb05979b10b42b90591255504
> >>          arm |                runc-v1.0.0-rc6 | http://autobuild.buildroot.net/results/c359954d671e338997615d3a2f3aca168a53cb91
> >>  
> > 
> > This is relating to O_TMPFILE not being available with this configuration:
> > 
> > libcontainer/nsenter/cloned_binary.c:206
> > #ifdef HAVE_MEMFD_CREATE
> > 	memfd = memfd_create(RUNC_MEMFD_COMMENT, MFD_CLOEXEC | MFD_ALLOW_SEALING);
> > #else
> > 	memfd = open("/tmp", O_TMPFILE | O_EXCL | O_RDWR | O_CLOEXEC, 0711);
> > #endif
> > 
> > I'm not aware of what the best alternative to O_TMPFILE in this system
> > setup would be. Do you (or anyone else) know a good solution to patch in
> > here for compatibility?  
> 
>  O_TMPFILE was added in Linux 3.11, memfd some time later IIRC. So depending on
> BR2_TOOLCHAIN_HEADERS_AT_LEAST_3_11 seems the correct approach.

Those failures are happening with a toolchain that uses 3.12 kernel
headers.

Best regards,

Thomas
-- 
Thomas Petazzoni, CTO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com