Hi Zi Shen,
On 10/01/2013 12:09 AM, Zi Shen Lim wrote:
On Mon, Sep 30, 2013 at 12:27 PM, Kim Phillips kim.phillips@linaro.org wrote:
On Mon, 30 Sep 2013 12:00:36 -0700 Zi Shen Lim zishen.lim@linaro.org wrote:
On Mon, Sep 30, 2013 at 12:03 PM, Anders Roxell anders.roxell@linaro.org wrote:
Hi,
On 2013-09-30 11:48, Zi Shen Lim wrote:
Signed-off-by: Zi Shen Lim zishen.lim@linaro.org
linaro/configs/linaro-base.conf | 4 ++++ 1 file changed, 4 insertions(+)
diff --git a/linaro/configs/linaro-base.conf b/linaro/configs/linaro-base.conf index 947ca1f..0093640 100644 --- a/linaro/configs/linaro-base.conf +++ b/linaro/configs/linaro-base.conf @@ -92,3 +92,7 @@ CONFIG_HW_PERF_EVENTS=y CONFIG_FUNCTION_TRACER=y CONFIG_ENABLE_DEFAULT_TRACERS=y CONFIG_PROC_DEVICETREE=y +CONFIG_HUGETLB_PAGE=y +CONFIG_HUGETLBFS=y +CONFIG_TRANSPARENT_HUGEPAGE=y +CONFIG_TRANSPARENT_HUGEPAGE_MADVISE=y
This CONFIG_* fragments are already in linaro/configs/linaro-base64.conf right?
Isn't linaro-base64.conf meant for ARMv8?
We need hugepage support on existing 32-bit / ARMv7 platforms too, don't we?
[adding linaro-kernel, Andrey]
Yes, it seems the base64 component is a misnomer, at least for now.
Meanwhile, if this patch is targeted to the main linaro kernel, please send via the linaro-kernel list, cc'ing linaro-networking. If not, then perhaps a separate .conf file is in order? Anyway, I think we'd all prefer a resolution within the main linaro kernel for hugepage support.
Andrey, which approach do you prefer?
- applying this patch in main linaro/configs
- maintaining hugepage as a separate config fragment
I would start from a separate config fragment. Not sure if HUGETLB and TRANSPARENT_HUGEPAGE would be good for all the builds. Later, if every 32-bit build start using it, the hugepage config fragment could be merged into linaro-base.conf. linaro-base64.conf is a special case, as there is only one ARMv8 board in Linaro builds atm - this would be revisited when more ARMv8 targets arrive.
- applying this patch only in LNG
Let me know if you want me to respin the patch. Thanks.
You would need a new patch. It should create a new config fragment instead of modifying linaro-base.conf.
Thanks, Andrey
and we are building with it in our ci/job/linux-lng* scripts...
perhaps those should be patched as well.
Kim