On Tue, Nov 28, 2017 at 06:33:55PM +0000, Will Deacon wrote:
> On Sat, Nov 25, 2017 at 12:41:27PM +0300, Yury Norov wrote:
> > It was discovered during LTO-enabled compilation with gcc/ld.bfd.
>
> What was discovered? Could you provide a bit more information in the
> changelog, please? I'm happy to take this as a fix if it's actually fixing
> something.
Yes it does. There's inconsistency in variable declaration and
section type. GCC doesn't throw error for usual build, but if LTO
enabled, build becomes broken, like this:
mm/percpu.c:2168:20: error: pcpu_fc_names causes a section type conflict
with dt_supported_cpu_ops
const char * const pcpu_fc_names[PCPU_FC_NR] __initconst = {
^
arch/arm64/kernel/cpu_ops.c:34:37: note: ‘dt_supported_cpu_ops’ was declared here
static const struct cpu_operations *dt_supported_cpu_ops[] __initconst = {
And so on. Let me know if you need full error log, then I'll resend the
patch with it.
You can also try it yourself, the very dirty and unfinished branch is
here:
https://github.com/norov/linux/tree/lto
Yury
From 1585335853128098173@xxx Tue Nov 28 18:34:46 +0000 2017
X-GM-THRID: 1585030578918772458
X-Gmail-Labels: Inbox,Category Forums,HistoricalUnread