Re: [PATCH nft 2/2] tests: shell: expand get command test with open intervals

On Tue, Jul 07, 2026 at 04:49:55PM +0200, Phil Sutter wrote:
> On Tue, Jul 07, 2026 at 04:16:01PM +0200, Pablo Neira Ayuso wrote:
> > Hi Phil,
> > 
> > On Tue, Jul 07, 2026 at 03:26:20PM +0200, Phil Sutter wrote:
> > > Hi Pablo,
> > > On Thu, Jul 02, 2026 at 02:36:34PM +0200, Pablo Neira Ayuso wrote:
> > > > Extend the existing test to cover get commands with open internals.
> > > 
> > > This test fails on a Big Endian testing machine running Fedora Rawhide
> > > running 7.2.0-0.rc2.21.fc45. Does this perhaps test a recent kernel fix
> > > or something?
> > 
> > Does this kernel contain this kernel fix?
> > 
> > https://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf.git/commit/?id=d63611cbe8af99dd61b118ee6e5b5e3e518250b2
> 
> Ah, thanks for the pointer. Rawhide kernels are "best effort" by early
> rebase vs. excessive backporting, so a fix in nf.git must be rather
> critical to hit it before linus.git.
> 
> > as well as this userspace fix:
> > 
> > https://git.netfilter.org/nftables/commit/?id=4eafc1a2a9ef5a827b1b4e58cb3b2832d2eb1650
> 
> Yes, this is present. To give you a bit of context: I have a script
> which builds current upstream HEAD of iptables, nftables and
> dependencies on a fresh Rawhide for running the test suites. I reserve a
> RH-internal Big Endian machine (s390x VM) for it. The goal is to detect
> Big Endian breakers early.

Thanks for explaining.

> Thanks for clarifying!

OK, then, all good, it is a matter of waiting for the kernel fix to
propagate to the rawhide kernel tree.

Ping back if problem persists by then, thanks.