Thread (41 messages) flat view 41 messages, 7 authors, 2025-10-14

Re: [PATCH v7 4/8] powerpc/smp: Introduce CONFIG_SCHED_MC to guard MC scheduling bits

From: Geert Uytterhoeven <geert@linux-m68k.org>
Date: 2025-10-14 12:43:46
Also in: linux-s390, lkml

Hoi Peter,

On Tue, 14 Oct 2025 at 11:42, Peter Zijlstra [off-list ref] wrote:
On Tue, Oct 14, 2025 at 11:25:53AM +0200, Geert Uytterhoeven wrote:
quoted
On Thu, 28 Aug 2025 at 14:57, Peter Zijlstra [off-list ref] wrote:
quoted
Now, when I look at unifying those config options (there's a metric ton
of crap that's duplicated in the arch/*/Kconfig), I end up with something
like the below.

And while that isn't exact, it is the closest I could make it without
making a giant mess of things.

WDYT?
Thanks for your patch, which is now commit 7bd291abe2da09f5 ("sched:
Unify the SCHED_{SMT,CLUSTER,MC} Kconfig") in v6.18-rc1.
quoted
--- a/arch/Kconfig
+++ b/arch/Kconfig
@@ -41,6 +41,44 @@ config HOTPLUG_SMT
 config SMT_NUM_THREADS_DYNAMIC
        bool

+config ARCH_SUPPORTS_SCHED_SMT
+       bool
+
+config ARCH_SUPPORTS_SCHED_CLUSTER
+       bool
+
+config ARCH_SUPPORTS_SCHED_MC
+       bool
+
+config SCHED_SMT
+       bool "SMT (Hyperthreading) scheduler support"
+       depends on ARCH_SUPPORTS_SCHED_SMT
+       default y
This is now enabled by default everywhere, while it was disabled by
default on most architectures before...
I'm not sure ARCH_SUPPORTS_SCHED_SMT counts as everywhere, but yes.
A fair deal of the architectures had all this default yes, and I had to
pick something. Can't make an omelette without breaking an egg and all
that :/
OK.
quoted
quoted
+       help
+         Improves the CPU scheduler's decision making when dealing with
+         MultiThreading at a cost of slightly increased overhead in some
+         places. If unsure say N here.
So it should default to n?
That's just help text that got carried around. Many of the architectures
that had default y still had this text on. I suppose we can change it if
someone cares.
Please do so.
quoted
If it is really needed on some architectures or platforms, I guess
they can still select it explicitly?
There were 4 cases:

 - arch doesn't support SMT
 - arch supports SMT and lets user pick, default Y
 - arch supports SMT and lets user pick, default N
 - arch mandates SMT

Of those 3 are still possible, the one we lost is the default N case.

Old configs that have =N will continue to have N. New configs might end
up with Y.

Why is this a problem?
While old .config files will see no changes, old defconfig files do
need updates.

Thanks!

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help