2003-06-01 18:11:49

by khromy

[permalink] [raw]
Subject: Generic x86 support in 2.5.70-bk

Generic x86 support (X86_GENERIC) [N/y/?] (NEW) ?

Including some tuning for non selected x86 CPUs too.
when it has moderate overhead. This is intended for generic
distributions kernels.

^-- Am I the only one confused by this description?

--
L1: khromy ;khromy (at) lnuxlab.ath.cx


2003-06-01 21:42:58

by William Lee Irwin III

[permalink] [raw]
Subject: Re: Generic x86 support in 2.5.70-bk

On Sun, Jun 01, 2003 at 02:25:32PM -0400, khromy wrote:
> Generic x86 support (X86_GENERIC) [N/y/?] (NEW) ?
> Including some tuning for non selected x86 CPUs too.
> when it has moderate overhead. This is intended for generic
> distributions kernels.
> ^-- Am I the only one confused by this description?

It'll boot on a "wider" variety of machines, for some unspecified but
valuable-to-distros value of "wider".


-- wli

2003-06-06 16:17:40

by Bill Davidsen

[permalink] [raw]
Subject: Re: Generic x86 support in 2.5.70-bk

On Sun, 1 Jun 2003, khromy wrote:

> Generic x86 support (X86_GENERIC) [N/y/?] (NEW) ?
>
> Including some tuning for non selected x86 CPUs too.
> when it has moderate overhead. This is intended for generic
> distributions kernels.
>
> ^-- Am I the only one confused by this description?

The second sentence was mangled when it was machine translated from the
original Klingon ;-)

Seriously, it doesn't parse. I can't decide if "when" in that sentence is
the wrong word, or if it snuck in uncapitalized and shouldn't be there at
all. I don't feel to "too" in the first sentence belongs, it doesn't refer
to several things.

--
bill davidsen <[email protected]>
CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.

2003-06-07 07:04:37

by Stewart Smith

[permalink] [raw]
Subject: [TRIV-PATCH] Better CONFIG_X86_GENERIC description (was: Re: Generic x86 support in 2.5.70-bk)


On Sun, Jun 01, 2003 at 02:25:32PM -0400, khromy wrote:
> Generic x86 support (X86_GENERIC) [N/y/?] (NEW) ?
> ^-- Am I the only one confused by this description?

No, it's a bit funny (no, not ha-ha funny). Maybe this is a bit clearer?


--- linux-2.5.70-bk11-orig/arch/i386/Kconfig 2003-06-06 23:55:43.000000000 +1000
+++ linux-2.5.70-bk11stew1/arch/i386/Kconfig 2003-06-07 17:11:16.000000000 +1000
@@ -289,9 +289,13 @@
config X86_GENERIC
bool "Generic x86 support"
help
- Including some tuning for non selected x86 CPUs too.
- when it has moderate overhead. This is intended for generic
- distributions kernels.
+ Instead of just including optimizations for the selected
+ x86 variant (e.g. PII, Crusoe or Athlon), include some more
+ generic optimizations as well. This will make the kernel
+ perform better on x86 CPUs other than that selected.
+
+ This is really intended for distributors who need more
+ generic optimizations.

#
# Define implied options from the CPU selection here

2003-06-07 08:30:17

by Margit Schubert-While

[permalink] [raw]
Subject: Re:[TRIV-PATCH] Better CONFIG_X86_GENERIC description (was: Re: Generic x86 support in 2.5.70-bk)

Hmm, as a side note, looking at arch/i386/Kconfig, the following
looks very suspect :

config X86_L1_CACHE_SHIFT
int
default "7" if MPENTIUM4 || X86_GENERIC

Margit

Subject: Re:[TRIV-PATCH] Better CONFIG_X86_GENERIC description (was: Re: Generic x86 support in 2.5.70-bk)


Its okay. You want the biggest possible X86_L1_CACHE_SHIFT for
X86_GENERIC, and it is the one for MPENTIUM4 (128 bytes).
Otherwise you will degrade performance on P4 when running generic kernel.

Regards,
--
Bartlomiej

On Sat, 7 Jun 2003, Margit Schubert-While wrote:

> Hmm, as a side note, looking at arch/i386/Kconfig, the following
> looks very suspect :
>
> config X86_L1_CACHE_SHIFT
> int
> default "7" if MPENTIUM4 || X86_GENERIC
>
> Margit