login
Header Space

 
 

Re: [RFC] x86: x86-64defconfig remake

Previous thread: none

Next thread: [-mm][PATCH] make inlining to __alloc_pages() by KOSAKI Motohiro on Sunday, May 4, 2008 - 12:17 pm. (1 message)
To: H. Peter Anvin <hpa@...>, Thomas Gleixner <tglx@...>, Ingo Molnar <mingo@...>, Andi Kleen <ak@...>
Cc: LKML <linux-kernel@...>
Date: Sunday, May 4, 2008 - 11:57 am

The current x86_64_defconfig contains a number of nonexistent
symbols. Lets fix it.

Signed-off-by: Cyrill Gorcunov &lt;gorcunov@gmail.com&gt;
---

WARNING! I think this patch *must* be carefully reviewed by kernel experts.
How it was done - original defconfig taken and some options (like MS DOS FS
turned on by default - was turned off). But I'm not sure for example about
I2C subsystem - it seems it has to be turned on by default. Anyway - if someone
has time to review - please do it. Comments are highly appreciated!!!

Index: linux-2.6.git/arch/x86/configs/x86_64_defconfig
===================================================================
--- linux-2.6.git.orig/arch/x86/configs/x86_64_defconfig	2008-05-01 11:07:29.000000000 +0400
+++ linux-2.6.git/arch/x86/configs/x86_64_defconfig	2008-05-04 19:39:44.000000000 +0400
@@ -1,48 +1,65 @@
 #
 # Automatically generated make config: don't edit
-# Linux kernel version: 2.6.22-git14
-# Fri Jul 20 09:53:15 2007
+# Linux kernel version: 2.6.26-rc1
+# Sun May  4 19:27:29 2008
 #
-CONFIG_X86_64=y
 CONFIG_64BIT=y
+# CONFIG_X86_32 is not set
+CONFIG_X86_64=y
 CONFIG_X86=y
+CONFIG_DEFCONFIG_LIST="arch/x86/configs/x86_64_defconfig"
+# CONFIG_GENERIC_LOCKBREAK is not set
 CONFIG_GENERIC_TIME=y
-CONFIG_GENERIC_TIME_VSYSCALL=y
 CONFIG_GENERIC_CMOS_UPDATE=y
-CONFIG_ZONE_DMA32=y
+CONFIG_CLOCKSOURCE_WATCHDOG=y
+CONFIG_GENERIC_CLOCKEVENTS=y
+CONFIG_GENERIC_CLOCKEVENTS_BROADCAST=y
 CONFIG_LOCKDEP_SUPPORT=y
 CONFIG_STACKTRACE_SUPPORT=y
-CONFIG_SEMAPHORE_SLEEPERS=y
+CONFIG_HAVE_LATENCYTOP_SUPPORT=y
+CONFIG_FAST_CMPXCHG_LOCAL=y
 CONFIG_MMU=y
 CONFIG_ZONE_DMA=y
-CONFIG_QUICKLIST=y
-CONFIG_NR_QUICK=2
-CONFIG_RWSEM_GENERIC_SPINLOCK=y
-CONFIG_GENERIC_HWEIGHT=y
-CONFIG_GENERIC_CALIBRATE_DELAY=y
-CONFIG_X86_CMPXCHG=y
-CONFIG_EARLY_PRINTK=y
 CONFIG_GENERIC_ISA_DMA=y
 CONFIG_GENERIC_IOMAP=y
-CONFIG_ARCH_MAY_HAVE_PC_FDC=y
-CONFIG_ARCH_POPULATES_NODE_MAP=y
-CONFIG_DMI=y
-CONFIG_AUDIT_ARCH=y
 CONFIG_GENERIC_BUG=y
+CONFIG_GENERIC_HWEIGHT=y
+...
To: Cyrill Gorcunov <gorcunov@...>
Cc: <hpa@...>, <tglx@...>, <mingo@...>, <ak@...>, <linux-kernel@...>
Date: Sunday, May 4, 2008 - 7:13 pm

These symbols don't exist in what context?

The one I notice:

-# CONFIG_CPUSETS is not set

certainly exists, in init/Kconfig:

    config CPUSETS
	bool "Cpuset support"
	depends on SMP &amp;&amp; CGROUPS

and is enabled by default in some configs:

    arch/ia64/configs/sn2_defconfig
    arch/mips/configs/ip27_defconfig
    arch/mips/configs/sb1250-swarm_defconfig
    arch/powerpc/configs/cell_defconfig
    arch/powerpc/configs/ppc64_defconfig
    arch/powerpc/configspseries_defconfig

I could be tempted to enable cpusets by default in x86_64
as well, however I don't understand the tradeoffs of such
a change well enough to know if I should propose that.

-- 
                  I won't rest till it's the best ...
                  Programmer, Linux Scalability
                  Paul Jackson &lt;pj@sgi.com&gt; 1.940.382.4214
--
To: Paul Jackson <pj@...>
Cc: Cyrill Gorcunov <gorcunov@...>, <hpa@...>, <tglx@...>, <ak@...>, <linux-kernel@...>
Date: Monday, May 5, 2008 - 3:03 am

that is enabled in the update i did (which was pushed out to x86.git 
yesterday). You can check it at:

  http://people.redhat.com/mingo/x86.git/README

	Ingo
--
To: Ingo Molnar <mingo@...>
Cc: <gorcunov@...>, <hpa@...>, <tglx@...>, <ak@...>, <linux-kernel@...>
Date: Monday, May 5, 2008 - 3:33 am

Excellent - thanks.

-- 
                  I won't rest till it's the best ...
                  Programmer, Linux Scalability
                  Paul Jackson &lt;pj@sgi.com&gt; 1.940.382.4214
--
To: Paul Jackson <pj@...>
Cc: <hpa@...>, <tglx@...>, <mingo@...>, <linux-kernel@...>
Date: Monday, May 5, 2008 - 12:21 am

Thanks Paul for feedback, will take a look today evening.
--
To: Cyrill Gorcunov <gorcunov@...>
Cc: H. Peter Anvin <hpa@...>, Thomas Gleixner <tglx@...>, Andi Kleen <ak@...>, LKML <linux-kernel@...>
Date: Sunday, May 4, 2008 - 2:06 pm

thanks Cyrill, applied. I also fixed up the 32-bit defconfig and did a 
few more updates to both of them as well, to make it more generally 
useful.

	Ingo
--
To: Ingo Molnar <mingo@...>
Cc: H. Peter Anvin <hpa@...>, Thomas Gleixner <tglx@...>, Andi Kleen <ak@...>, LKML <linux-kernel@...>
Date: Sunday, May 4, 2008 - 2:36 pm

[Ingo Molnar - Sun, May 04, 2008 at 08:06:42PM +0200]
| 
| * Cyrill Gorcunov &lt;gorcunov@gmail.com&gt; wrote:
| 
| &gt; The current x86_64_defconfig contains a number of nonexistent
| &gt; symbols. Lets fix it.
| 
| thanks Cyrill, applied. I also fixed up the 32-bit defconfig and did a 
| few more updates to both of them as well, to make it more generally 
| useful.
| 
| 	Ingo
| 

Thanks Ingo, lets wait for feedback from community ;)
I hope they would not kill us ;)

		- Cyrill -
--
Previous thread: none

Next thread: [-mm][PATCH] make inlining to __alloc_pages() by KOSAKI Motohiro on Sunday, May 4, 2008 - 12:17 pm. (1 message)
speck-geostationary