| Message ID | 1460595252-2330-1-git-send-email-yang.shi@linaro.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Wed, Apr 13, 2016 at 05:54:12PM -0700, Yang Shi wrote: > SCHED_MC and SCHED_SMT are pointless when SMP is disabled. > Although SMP is rarely disabled for ARM64, it looks more consistent to have > such depend in Kconfig. You can't disable CONFIG_SMP for arm64 -- we select it unconditionally in the kconfig. Will
On 4/14/2016 1:47 AM, Will Deacon wrote: > On Wed, Apr 13, 2016 at 05:54:12PM -0700, Yang Shi wrote: >> SCHED_MC and SCHED_SMT are pointless when SMP is disabled. >> Although SMP is rarely disabled for ARM64, it looks more consistent to have >> such depend in Kconfig. > > You can't disable CONFIG_SMP for arm64 -- we select it unconditionally > in the kconfig. Thanks Will. I didn't realize ARM64 has SMP selected unconditionally, it looks the patch is pointless. A follow-up question, I know ARM64 has no UP implementation now, it sounds make sense to have SMP select unconditionally, however, it might be more flexible to have SMP like x86 and other architectures. And, it may also help to find more bugs when !SMP. Thanks, Yang > > Will >
On Thu, Apr 14, 2016 at 09:33:07AM -0700, Shi, Yang wrote: > On 4/14/2016 1:47 AM, Will Deacon wrote: > >On Wed, Apr 13, 2016 at 05:54:12PM -0700, Yang Shi wrote: > >>SCHED_MC and SCHED_SMT are pointless when SMP is disabled. > >>Although SMP is rarely disabled for ARM64, it looks more consistent to have > >>such depend in Kconfig. > > > >You can't disable CONFIG_SMP for arm64 -- we select it unconditionally > >in the kconfig. > > Thanks Will. I didn't realize ARM64 has SMP selected unconditionally, it > looks the patch is pointless. > > A follow-up question, I know ARM64 has no UP implementation now, it sounds > make sense to have SMP select unconditionally, however, it might be more > flexible to have SMP like x86 and other architectures. And, it may also help > to find more bugs when !SMP. Well, an SMP kernel would still work on a UP implementation, so I don't really see the point in carrying a whole load of barely tested #ifdef code. Top that off with the fact that we still need shareable memory for coherent DMA and we still need barrier instructions for the page table walker, and it's really diminishing returns. Will
diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index 4f43622..e5de825 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -521,6 +521,7 @@ config CPU_BIG_ENDIAN config SCHED_MC bool "Multi-core scheduler support" + depends on SMP help Multi-core scheduler support improves the CPU scheduler's decision making when dealing with multi-core CPU chips at a cost of slightly @@ -528,6 +529,7 @@ config SCHED_MC config SCHED_SMT bool "SMT scheduler support" + depends on SMP help Improves the CPU scheduler's decision making when dealing with MultiThreading at a cost of slightly increased overhead in some
SCHED_MC and SCHED_SMT are pointless when SMP is disabled. Although SMP is rarely disabled for ARM64, it looks more consistent to have such depend in Kconfig. Signed-off-by: Yang Shi <yang.shi@linaro.org> --- arch/arm64/Kconfig | 2 ++ 1 file changed, 2 insertions(+)