Re: v3.18-stable-queue and v4.1.y-stable-queue build failures

[Greg Kroah-Hartman <gregkh@linuxfoundation.org>]

On Tue, Jun 07, 2016 at 11:24:40AM -0700, Guenter Roeck wrote:
> On Tue, Jun 07, 2016 at 12:46:16PM -0400, Sasha Levin wrote:
> > >>> v4.1, all mips builds:
> > >>>
> > >>> arch/mips/mm/cache.c: In function '__update_cache':
> > >>> arch/mips/mm/cache.c:134:26: error: implicit declaration of function 'kmap_atomic'
> > >>> arch/mips/mm/cache.c:142:4: error: implicit declaration of function '__kunmap_atomic'
> > >>>
> > >>> arch/mips/kernel/setup.c: In function 'arch_mem_init':
> > >>> arch/mips/kernel/setup.c:690:2: error: implicit declaration of function 'reserve_bootmem_region'
> > >>>
> > >>> Guenter
> > >>
> > >> Thanks Guenter, all should be fixed now.
> > >>
> > > Confirmed.
> > > 
> > > Can you let Greg know how you fixed the mips problem in mm/cache.c ? It now affects
> > > 4.4, 4.5, and 4.6 as well.
> > 
> > So the commit that was supposed to get this fix was only introduced in 4.2, so
> > I just removed it from the 3.18/4.1 trees, but I suspect Greg will need to do
> > something else here.
> > 
> I ended up reverting 'MIPS: Sync icache & dcache in set_pte_at' as well as
> 'MIPS: Handle highmem pages in __update_cache'. That results in a passing build
> at least for 4.4. No idea if it has side effects, though.

Hm, I did something else, I guess if anyone from MIPS cares about the
stable trees, they will let us know...

thanks,

greg k-h