2003-09-02 14:36:05

by Matthew Wilcox

[permalink] [raw]
Subject: CONFIG_64_BIT


What do people think of CONFIG_64_BIT? It saves us from using
!(IA64 || MIPS64 || PARISC64 || S390X || SPARC64 || X86_64) or
the X86_64 people deciding their architecture is more important.

I also considered CONFIG_ILP32 vs CONFIG_LP64 (since that's the real
problem with, eg, megaraid), but that requires more explanation and
offers people several ways to get it wrong (should I depend on ILP32
or !LP64?)

Index: arch/alpha/Kconfig
===================================================================
RCS file: /var/cvs/linux-2.6/arch/alpha/Kconfig,v
retrieving revision 1.3
diff -u -p -r1.3 Kconfig
--- arch/alpha/Kconfig 23 Aug 2003 02:46:12 -0000 1.3
+++ arch/alpha/Kconfig 2 Sep 2003 14:24:46 -0000
@@ -11,6 +11,9 @@ config ALPHA
now Hewlett-Packard. The Alpha Linux project has a home page at
<http://www.alphalinux.org/>.

+config 64_BIT
+ def_bool y
+
config MMU
bool
default y
Index: arch/ia64/Kconfig
===================================================================
RCS file: /var/cvs/linux-2.6/arch/ia64/Kconfig,v
retrieving revision 1.2
diff -u -p -r1.2 Kconfig
--- arch/ia64/Kconfig 12 Aug 2003 19:10:49 -0000 1.2
+++ arch/ia64/Kconfig 2 Sep 2003 14:24:46 -0000
@@ -18,6 +18,9 @@ config IA64
page at <http://www.linuxia64.org/> and a mailing list at
[email protected].

+config 64_BIT
+ def_bool y
+
config MMU
bool
default y
Index: arch/mips/Kconfig
===================================================================
RCS file: /var/cvs/linux-2.6/arch/mips/Kconfig,v
retrieving revision 1.2
diff -u -p -r1.2 Kconfig
--- arch/mips/Kconfig 12 Aug 2003 19:10:50 -0000 1.2
+++ arch/mips/Kconfig 2 Sep 2003 14:24:47 -0000
@@ -11,6 +11,9 @@ config MIPS64
64-bit processing, otherwise say N. You must say Y for kernels for
SGI IP27 (Origin 200 and 2000). If in doubt say N.

+config 64_BIT
+ def_bool MIPS64
+
config MIPS32
bool
depends on MIPS64 = 'n'
Index: arch/parisc/Kconfig
===================================================================
RCS file: /var/cvs/linux-2.6/arch/parisc/Kconfig,v
retrieving revision 1.6
diff -u -p -r1.6 Kconfig
--- arch/parisc/Kconfig 23 Aug 2003 02:56:56 -0000 1.6
+++ arch/parisc/Kconfig 2 Sep 2003 14:24:47 -0000
@@ -103,6 +103,9 @@ config PARISC64
enable this option otherwise. The 64bit kernel is significantly bigger
and slower than the 32bit one.

+config 64_BIT
+ def_bool PARISC64
+
config PDC_NARROW
bool "32-bit firmware"
depends on PARISC64
Index: arch/ppc64/Kconfig
===================================================================
RCS file: /var/cvs/linux-2.6/arch/ppc64/Kconfig,v
retrieving revision 1.3
diff -u -p -r1.3 Kconfig
--- arch/ppc64/Kconfig 23 Aug 2003 02:46:30 -0000 1.3
+++ arch/ppc64/Kconfig 2 Sep 2003 14:24:47 -0000
@@ -3,6 +3,9 @@
# see Documentation/kbuild/kconfig-language.txt.
#

+config 64_BIT
+ def_bool y
+
config MMU
bool
default y
Index: arch/s390/Kconfig
===================================================================
RCS file: /var/cvs/linux-2.6/arch/s390/Kconfig,v
retrieving revision 1.2
diff -u -p -r1.2 Kconfig
--- arch/s390/Kconfig 12 Aug 2003 19:10:57 -0000 1.2
+++ arch/s390/Kconfig 2 Sep 2003 14:24:47 -0000
@@ -40,6 +40,9 @@ config ARCH_S390X
Select this option if you have a 64 bit IBM zSeries machine
and want to use the 64 bit addressing mode.

+config 64_BIT
+ def_bool ARCH_S390X
+
config ARCH_S390_31
bool
depends on ARCH_S390X = 'n'
Index: arch/sparc64/Kconfig
===================================================================
RCS file: /var/cvs/linux-2.6/arch/sparc64/Kconfig,v
retrieving revision 1.3
diff -u -p -r1.3 Kconfig
--- arch/sparc64/Kconfig 12 Aug 2003 19:10:57 -0000 1.3
+++ arch/sparc64/Kconfig 2 Sep 2003 14:24:47 -0000
@@ -5,6 +5,9 @@

mainmenu "Linux/UltraSPARC Kernel Configuration"

+config 64_BIT
+ def_bool y
+
config MMU
bool
default y
Index: arch/x86_64/Kconfig
===================================================================
RCS file: /var/cvs/linux-2.6/arch/x86_64/Kconfig,v
retrieving revision 1.2
diff -u -p -r1.2 Kconfig
--- arch/x86_64/Kconfig 12 Aug 2003 19:10:59 -0000 1.2
+++ arch/x86_64/Kconfig 2 Sep 2003 14:24:47 -0000
@@ -16,6 +16,9 @@ config X86_64
Port to the x86-64 architecture. x86-64 is a 64-bit extension to the
classical 32-bit x86 architecture. For details see http://www.x86-64.org

+config 64_BIT
+ def_bool y
+
config X86
bool
default y
Index: drivers/scsi/Kconfig
===================================================================
RCS file: /var/cvs/linux-2.6/drivers/scsi/Kconfig,v
retrieving revision 1.3
diff -u -p -r1.3 Kconfig
--- drivers/scsi/Kconfig 25 Aug 2003 19:49:06 -0000 1.3
+++ drivers/scsi/Kconfig 2 Sep 2003 14:24:48 -0000
@@ -355,7 +355,7 @@ source "drivers/scsi/aic7xxx/Kconfig.aic
# All the I2O code and drivers do not seem to be 64bit safe.
config SCSI_DPT_I2O
tristate "Adaptec I2O RAID support "
- depends on !X86_64 && SCSI
+ depends on !64_BIT && SCSI
help
This driver supports all of Adaptec's I2O based RAID controllers as
well as the DPT SmartRaid V cards. This is an Adaptec maintained
@@ -416,7 +416,7 @@ config SCSI_AM53C974

config SCSI_MEGARAID
tristate "AMI MegaRAID support"
- depends on PCI && SCSI
+ depends on PCI && SCSI && !64_BIT
help
This driver supports the AMI MegaRAID 418, 428, 438, 466, 762, 490
and 467 SCSI host adapters.

--
"It's not Hollywood. War is real, war is primarily not about defeat or
victory, it is about death. I've seen thousands and thousands of dead bodies.
Do you think I want to have an academic debate on this subject?" -- Robert Fisk


2003-09-02 17:55:42

by Jeff Garzik

[permalink] [raw]
Subject: Re: CONFIG_64_BIT

Matthew Wilcox wrote:
> What do people think of CONFIG_64_BIT? It saves us from using


Yes! Please!

Though I would of course make it CONFIG_64BIT ;-)


2003-09-02 18:08:29

by Andi Kleen

[permalink] [raw]
Subject: Re: CONFIG_64_BIT

On Tue, 2 Sep 2003 18:44:36 +0100
Matthew Wilcox <[email protected]> wrote:

> On Tue, Sep 02, 2003 at 07:35:11PM +0200, Andi Kleen wrote:
> > Matthew Wilcox <[email protected]> writes:
> >
> > > What do people think of CONFIG_64_BIT? It saves us from using
> > > !(IA64 || MIPS64 || PARISC64 || S390X || SPARC64 || X86_64) or
> >
> > For Kconfigs it may make sense, but is there any Config rule that
> > checks for all 64bit archs (opposed to checking for specific archs)?
> > I cannot thinkg of any.
>
> ... that was what the patch added.

It added a symbol that means that, but are there any users for it?

Ok, I2O and ATM and WANPIPE maybe but I assume that both are getting fixed
Still all those are non 64bit safe, so it may be better to have an
CONFIG_32BIT_ONLY or similar.

-Andi

2003-09-03 08:40:03

by Geert Uytterhoeven

[permalink] [raw]
Subject: Re: CONFIG_64_BIT

On Tue, 2 Sep 2003, Jeff Garzik wrote:
> Matthew Wilcox wrote:
> > What do people think of CONFIG_64_BIT? It saves us from using
>
> Yes! Please!
>
> Though I would of course make it CONFIG_64BIT ;-)

Or CONFIG_LINUX64? :-)

> Matthew Wilcox wrote:
> > What do people think of CONFIG_64_BIT? It saves us from using

Why do some many mails from lkml have duplicated contents?

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

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

2003-09-03 10:10:45

by Andreas Schwab

[permalink] [raw]
Subject: Re: CONFIG_64_BIT

Geert Uytterhoeven <[email protected]> writes:

> Why do some many mails from lkml have duplicated contents?

Might be a problem on your side, I didn't see that.

Andreas.

--
Andreas Schwab, SuSE Labs, [email protected]
SuSE Linux AG, Deutschherrnstr. 15-19, D-90429 N?rnberg
Key fingerprint = 58CA 54C7 6D53 942B 1756 01D3 44D5 214B 8276 4ED5
"And now for something completely different."

2003-09-03 10:45:19

by Wade

[permalink] [raw]
Subject: Re: CONFIG_64_BIT

Andreas Schwab wrote:
> Geert Uytterhoeven <[email protected]> writes:
>
>
>>Why do some many mails from lkml have duplicated contents?
>
>
> Might be a problem on your side, I didn't see that.
>
> Andreas.
>

No, see the "vger's today hickups" thread.


2003-09-03 13:40:43

by Alan

[permalink] [raw]
Subject: Re: CONFIG_64_BIT

On Maw, 2003-09-02 at 18:52, Andi Kleen wrote:
> Ok, I2O and ATM and WANPIPE maybe but I assume that both are getting fixed
> Still all those are non 64bit safe, so it may be better to have an
> CONFIG_32BIT_ONLY or similar.

I2O is currently unmaintained. I fixed some of the DMA issues but there
are still a few bits of code that pass pointers in 32bit slots rather
than using array indices.