* [TRIVIAL] [resend patch] CONFIG_X86_GENERIC description fixup
@ 2003-07-07 7:57 Rusty Trivial Russell
0 siblings, 0 replies; 6+ messages in thread
From: Rusty Trivial Russell @ 2003-07-07 7:57 UTC (permalink / raw)
To: Linus Torvalds; +Cc: linux-kernel
From: Stewart Smith <stewart@linux.org.au>
as per thread on lkml a little while ago, a better explanation
of the X86_GENERIC config option follows. The person who questioned
it originally seemed to like this improved version, so that's one point :)
--- trivial-2.5.74-bk4/arch/i386/Kconfig.orig 2003-07-07 17:36:54.000000000 +1000
+++ trivial-2.5.74-bk4/arch/i386/Kconfig 2003-07-07 17:36:54.000000000 +1000
@@ -303,9 +303,13 @@
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
--
What is this? http://www.kernel.org/pub/linux/kernel/people/rusty/trivial/
Don't blame me: the Monkey is driving
File: Stewart Smith <stewart@linux.org.au>: [resend patch] CONFIG_X86_GENERIC description fixup
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [TRIVIAL] [resend patch] CONFIG_X86_GENERIC description fixup
2003-09-04 12:20 ` Dave Jones
@ 2003-09-05 0:15 ` Rusty Russell
0 siblings, 0 replies; 6+ messages in thread
From: Rusty Russell @ 2003-09-05 0:15 UTC (permalink / raw)
To: Dave Jones; +Cc: Linus Torvalds, linux-kernel
In message <20030904122029.GA3357@redhat.com> you write:
> On Thu, Sep 04, 2003 at 01:26:41PM +1000, Rusty Trivial Russell wrote:
>
> > as per thread on lkml a little while ago, a better explanation
> > of the X86_GENERIC config option follows. The person who questioned
> > it originally seemed to like this improved version, so that's one point :)
>
> How about explaining _exactly_ what it does? This is still a somewhat
> mysterious description. "generic optimisations" what is that ?
Whatever it says will go out of date too quickly 8(. I think this
patch will be fine.
Rusty.
--
Anyone who quotes me in their sig is an idiot. -- Rusty Russell.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [TRIVIAL] [resend patch] CONFIG_X86_GENERIC description fixup
2003-09-04 3:26 Rusty Trivial Russell
@ 2003-09-04 12:20 ` Dave Jones
2003-09-05 0:15 ` Rusty Russell
0 siblings, 1 reply; 6+ messages in thread
From: Dave Jones @ 2003-09-04 12:20 UTC (permalink / raw)
To: Rusty Trivial Russell; +Cc: Linus Torvalds, linux-kernel
On Thu, Sep 04, 2003 at 01:26:41PM +1000, Rusty Trivial Russell wrote:
> as per thread on lkml a little while ago, a better explanation
> of the X86_GENERIC config option follows. The person who questioned
> it originally seemed to like this improved version, so that's one point :)
How about explaining _exactly_ what it does? This is still a somewhat
mysterious description. "generic optimisations" what is that ?
> --- trivial-2.6.0-test4-bk5/arch/i386/Kconfig.orig 2003-09-04 13:02:02.000000000 +1000
> +++ trivial-2.6.0-test4-bk5/arch/i386/Kconfig 2003-09-04 13:02:02.000000000 +1000
> @@ -303,9 +303,13 @@
> 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.
All that it seems to do right now is set the cacheline size to the
same as it would if the kernel was compiled for P4, regardless of
what target CPU is selected, so that for eg, an i686 kernel won't
perform any worse on a P4, saving vendors shipping seperate P4 kernels.
it should also note what performance impact (if any) users of this
option will see if they run such a kernel on a box with a smaller
cacheline size.
If this option ever does anything else, that too should get documented here.
Dave
--
Dave Jones http://www.codemonkey.org.uk
^ permalink raw reply [flat|nested] 6+ messages in thread
* [TRIVIAL] [resend patch] CONFIG_X86_GENERIC description fixup
@ 2003-09-04 3:26 Rusty Trivial Russell
2003-09-04 12:20 ` Dave Jones
0 siblings, 1 reply; 6+ messages in thread
From: Rusty Trivial Russell @ 2003-09-04 3:26 UTC (permalink / raw)
To: Linus Torvalds; +Cc: linux-kernel
From: Stewart Smith <stewart@linux.org.au>
as per thread on lkml a little while ago, a better explanation
of the X86_GENERIC config option follows. The person who questioned
it originally seemed to like this improved version, so that's one point :)
--- trivial-2.6.0-test4-bk5/arch/i386/Kconfig.orig 2003-09-04 13:02:02.000000000 +1000
+++ trivial-2.6.0-test4-bk5/arch/i386/Kconfig 2003-09-04 13:02:02.000000000 +1000
@@ -303,9 +303,13 @@
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
--
What is this? http://www.kernel.org/pub/linux/kernel/people/rusty/trivial/
Don't blame me: the Monkey is driving
File: Stewart Smith <stewart@linux.org.au>: [resend patch] CONFIG_X86_GENERIC description fixup
^ permalink raw reply [flat|nested] 6+ messages in thread
* [TRIVIAL] [resend patch] CONFIG_X86_GENERIC description fixup
@ 2003-07-21 11:24 Rusty Trivial Russell
0 siblings, 0 replies; 6+ messages in thread
From: Rusty Trivial Russell @ 2003-07-21 11:24 UTC (permalink / raw)
To: Linus Torvalds; +Cc: linux-kernel
[ I like the clearer explanation --RR ]
From: Stewart Smith <stewart@linux.org.au>
as per thread on lkml a little while ago, a better explanation
of the X86_GENERIC config option follows. The person who questioned
it originally seemed to like this improved version, so that's one point :)
--- trivial-2.5.75-bk3/arch/i386/Kconfig.orig 2003-07-21 21:23:02.000000000 +1000
+++ trivial-2.5.75-bk3/arch/i386/Kconfig 2003-07-21 21:23:02.000000000 +1000
@@ -303,9 +303,13 @@
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
--
What is this? http://www.kernel.org/pub/linux/kernel/people/rusty/trivial/
Don't blame me: the Monkey is driving
File: Stewart Smith <stewart@linux.org.au>: [resend patch] CONFIG_X86_GENERIC description fixup
^ permalink raw reply [flat|nested] 6+ messages in thread
* [TRIVIAL] [resend patch] CONFIG_X86_GENERIC description fixup
@ 2003-06-23 7:19 Rusty Trivial Russell
0 siblings, 0 replies; 6+ messages in thread
From: Rusty Trivial Russell @ 2003-06-23 7:19 UTC (permalink / raw)
To: torvalds, linux-kernel
From: Stewart Smith <stewart@linux.org.au>
as per thread on lkml a little while ago, a better explanation
of the X86_GENERIC config option follows. The person who questioned
it originally seemed to like this improved version, so that's one point :)
--- trivial-2.5.73/arch/i386/Kconfig.orig 2003-06-23 17:04:46.000000000 +1000
+++ trivial-2.5.73/arch/i386/Kconfig 2003-06-23 17:04:46.000000000 +1000
@@ -303,9 +303,13 @@
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
--
What is this? http://www.kernel.org/pub/linux/kernel/people/rusty/trivial/
Don't blame me: the Monkey is driving
File: Stewart Smith <stewart@linux.org.au>: [resend patch] CONFIG_X86_GENERIC description fixup
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2003-09-05 5:22 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2003-07-07 7:57 [TRIVIAL] [resend patch] CONFIG_X86_GENERIC description fixup Rusty Trivial Russell
-- strict thread matches above, loose matches on Subject: below --
2003-09-04 3:26 Rusty Trivial Russell
2003-09-04 12:20 ` Dave Jones
2003-09-05 0:15 ` Rusty Russell
2003-07-21 11:24 Rusty Trivial Russell
2003-06-23 7:19 Rusty Trivial Russell
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).