Re: [PATCH 5.15 000/846] 5.15.17-rc1 review

[Greg Kroah-Hartman]

On Thu, Jan 27, 2022 at 06:01:53PM +0530, Naresh Kamboju wrote:
> On Wed, 26 Jan 2022 at 21:05, Alan Maguire <alan.maguire@xxxxxxxxxx> wrote:
> >
> > > Regressions detected on arm, arm64, i386, x86 on 5.15 and 5.10
> > >
> > > > >
> > > > > This is one from arm64:
> > > > >    /builds/linux/arch/arm64/mm/extable.c: In function 'fixup_exception':
> > > > >    /builds/linux/arch/arm64/mm/extable.c:17:13: error: implicit declaration of function 'in_bpf_jit' [-Werror=implicit-function-declaration]
> > > > >       17 |         if (in_bpf_jit(regs))
> > > > >          |             ^~~~~~~~~~
> > > > >    cc1: some warnings being treated as errors
> > > > >    make[3]: *** [/builds/linux/scripts/Makefile.build:277: arch/arm64/mm/extable.o] Error 1
> > > >
> > > > Bisection here pointed to "arm64/bpf: Remove 128MB limit for BPF JIT programs". Reverting made the build succeed.
> > >
> > > arm64/bpf: Remove 128MB limit for BPF JIT programs
> > > commit b89ddf4cca43f1269093942cf5c4e457fd45c335 upstream.
> > >
> > > Reported-by: Linux Kernel Functional Testing <lkft@xxxxxxxxxx>
> > >
> >
> > Thanks for the report!
> >
> > This one needs slightly different handling on 5.15. Russell had a 5.15
> > patch for this (where BPF exception handling was still handled separately)
> > and I've included it below. I verified it applies cleanly to the
> > linux-5.15.y branch and builds.  I'd suggest either skipping backport of
> > this fix to stable completely, or just applying the below to 5.15 and
> > skipping further backports.
> 
> Build test pass with this patch on stable/linux-5.15.y.

Great, thanks!

I'll queue this up now.

greg k-h