[<prev] [next>] [day] [month] [year] [list]
Date: Fri, 5 May 2017 11:52:32 -0700
From: Vineet Gupta <Vineet.Gupta1@...opsys.com>
To: <linux-snps-arc@...ts.infradead.org>
CC: <noreply@...erman.id.au>, lkml <linux-kernel@...r.kernel.org>,
<Alexey.Brodkin@...opsys.com>,
"linux-mm@...ck.org" <linux-mm@...ck.org>,
"linux-arch@...r.kernel.org" <linux-arch@...r.kernel.org>,
Michal Hocko <mhocko@...e.cz>,
Anton Kolesov <Anton.Kolesov@...opsys.com>
Subject: asm/mmu.h using NR_CPUS (was Re: kisskb: FAILED
linux-next/axs101_defconfig/arcompact Thu May 04, 18:53)
+CC Michal, linux-arch as it potentially affects other arches !
Hi,
On 05/04/2017 01:53 AM, noreply@...erman.id.au wrote:
> FAILED linux-next/axs101_defconfig/arcompact Thu May 04, 18:53
>
> http://kisskb.ellerman.id.au/kisskb/buildresult/13022475/
>
> Commit: Add linux-next specific files for 20170504
> ef57eb19c96fa099a578aeaed9b9d0dbcc4fe069
> Compiler: arc-buildroot-linux-uclibc-gcc (Buildroot 2015.08.1) 4.8.4
>
> Possible errors
> ---------------
>
> arch/arc/include/asm/mmu.h:75:21: error: 'NR_CPUS' undeclared here (not in a function)
> make[3]: *** [arch/arc/mm/ioremap.o] Error 1
> make[2]: *** [arch/arc/mm] Error 2
> make[1]: *** [arch/arc] Error 2
> arch/arc/include/asm/mmu.h:75:21: error: 'NR_CPUS' undeclared here (not in a function)
> make[3]: *** [kernel/irq/generic-chip.o] Error 1
> make[2]: *** [kernel/irq] Error 2
> make[1]: *** [kernel] Error 2
> arch/arc/include/asm/mmu.h:75:21: error: 'NR_CPUS' undeclared here (not in a function)
> make[2]: *** [mm/vmalloc.o] Error 1
> make[1]: *** [mm] Error 2
> make: *** [sub-make] Error 2
I bisected the linux-next ARC build failure (!SMP only) to a subtle side effect of
a950a30220657d ("mm, vmalloc: properly track vmalloc users") which includes
asm/pgtable.h causing the final include chain (for ARC ioremap.c) to have
asm/mmu.h (using NR_CPUS) before linux/threads.h (defining it)
The quick fix is to include linux/threads.h and this is just a heads up other
arches might run into same - although xtensa and mips with similar mm_context_t
seem to build just fine !
-Vineet
Powered by blists - more mailing lists