Re: [PATCH] Cleanup subarch definitions in Linux/i386

From: Zachary Amsden
Date: Thu Mar 30 2006 - 14:22:25 EST

Next message: Matthew Wilcox: "Re: [PATCH] ioremap_cached()"
Previous message: Mingming Cao: "Re: [RFC][PATCH 0/2]Extend ext3 filesystem limit from 8TB to 16TB"
In reply to: Chris Wright: "Re: [PATCH] Cleanup subarch definitions in Linux/i386"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Chris Wright wrote:

* Zachary Amsden (zach@xxxxxxxxxx) wrote:

Comments, suggestions, anything welcome. I think this is a much cleaner approach, and both new and existing sub-architectures will benefit. I am sorry this patch is so large, but it is very difficult to separate into multiple steps that still allow all the subarches to compile.

Thanks for looking over these. There is a bit more cleanup on top of this, mostly due to the include mess. I have attached another patch.

Zach, looks nice. Saves Xen a partial copy of setup.c. Did you have
further/similar consolidations in mind?

Exactly.

--- linux-2.6.16.1.orig/arch/i386/Makefile 2006-03-29 19:38:47.000000000 -0800
+++ linux-2.6.16.1/arch/i386/Makefile 2006-03-29 19:38:54.000000000 -0800
@@ -45,37 +45,32 @@ CFLAGS += $(shell if [ $(call cc-vers
CFLAGS += $(cflags-y)
-# Default subarch .c files
-mcore-y := mach-default
+# Default subarch .c files (none)
+mcore-y := # Voyager subarch support
mflags-$(CONFIG_X86_VOYAGER) := -Iinclude/asm-i386/mach-voyager
-mcore-$(CONFIG_X86_VOYAGER) := mach-voyager
+mcore-$(CONFIG_X86_VOYAGER) := arch/i386/mach-voyager/

Is this intended to make way for possible fine tuning? Smth like:

mcore-$(CONFIG_X86_VOYAGER) += arch/i386/another_default.o
(hmm, not sure if that would even work)

Or just an aesthetic change?

No - it was required in order to leave the subarch blank for the default case. Including "arch/i386//" wasn't so happy in the makefile.

--- linux-2.6.16.1.orig/include/asm-i386/acpi.h 2006-03-29 19:38:47.000000000 -0800
+++ linux-2.6.16.1/include/asm-i386/acpi.h 2006-03-29 19:38:54.000000000 -0800
@@ -31,6 +31,7 @@
#include <acpi/pdc_intel.h>
#include <asm/system.h> /* defines cmpxchg */
+#include <asm/processor.h> /* defines boot_cpu_data */

that one necessary?

Needed for mach-generic subarch to compile with CPU defined as I386 -- the definition of cmpxchg on older CPUs depends on boot_cpu_data. Removing this opens a giant rat hole. This ginourmous rat hole is quite annoying. What goes in system.h vs. processor.h is so poorly defined. I had to hoist boot_cpu_data into system.h as the cleanest way to fix this without introducing circular dependencies - processor.h already depends on system.h for alternative_input(), read_cr4(), write_cr4(), write_cr3(), and there is no end to this mess. Trying to move dependent processor.h pieces into system.h didn't work, since the include/linux headers make the assumption that they can include asm/processor.h to determine if the architecture already has a prefetch instruction defined. What a nightmare. Trying to move system.h pieces into processor.h also doesn't work, since you basically need to move the whole file.

Maybe that is the way to go, but I would rather not go there now.

#define COMPILER_DEPENDENT_INT64 long long
#define COMPILER_DEPENDENT_UINT64 unsigned long long
Index: linux-2.6.16.1/include/asm-i386/arch_hooks.h
===================================================================
--- linux-2.6.16.1.orig/include/asm-i386/arch_hooks.h 2006-03-29 19:38:47.000000000 -0800
+++ linux-2.6.16.1/include/asm-i386/arch_hooks.h 2006-03-29 19:38:54.000000000 -0800
@@ -1,7 +1,13 @@
#ifndef _ASM_ARCH_HOOKS_H
#define _ASM_ARCH_HOOKS_H
+#include <linux/config.h>
+#include <linux/smp.h>
+#include <linux/init.h>
#include <linux/interrupt.h>
+#include <asm/acpi.h>
+#include <asm/arch_hooks.h>

extraneous include

Yes, somewhat extraneous.

--- linux-2.6.16.1.orig/include/asm-i386/mach-default/mach_hooks.h 2006-03-29 19:38:54.000000000 -0800
+++ linux-2.6.16.1/include/asm-i386/mach-default/mach_hooks.h 2006-03-29 19:38:54.000000000 -0800
@@ -0,0 +1,6 @@
+#ifndef _MACH_HOOKS_H
+#define _MACH_HOOKS_H

should probably be consistent (_MACH_HOOKS_H vs. MACH_HOOKS_H)

Patch attached. Tidy up the subarch cleanup patch with nits noticed by Chris Wright.

Signed-off-by: Zachary Amsden <zach@xxxxxxxxxx>

Index: linux-2.6.16.1/include/asm-i386/arch_hooks.h
===================================================================
--- linux-2.6.16.1.orig/include/asm-i386/arch_hooks.h 2006-03-30 10:53:45.000000000 -0800
+++ linux-2.6.16.1/include/asm-i386/arch_hooks.h 2006-03-30 10:54:27.000000000 -0800
@@ -1,13 +1,9 @@
#ifndef _ASM_ARCH_HOOKS_H
#define _ASM_ARCH_HOOKS_H

-#include <linux/config.h>
-#include <linux/smp.h>
-#include <linux/init.h>
#include <linux/interrupt.h>
#include <asm/acpi.h>
#include <asm/arch_hooks.h>
-#include <asm/desc.h>

/*
* linux/include/asm/arch_hooks.h
Index: linux-2.6.16.1/include/asm-i386/acpi.h
===================================================================
--- linux-2.6.16.1.orig/include/asm-i386/acpi.h 2006-03-30 10:53:45.000000000 -0800
+++ linux-2.6.16.1/include/asm-i386/acpi.h 2006-03-30 10:54:27.000000000 -0800
@@ -31,7 +31,6 @@
#include <acpi/pdc_intel.h>

#include <asm/system.h> /* defines cmpxchg */
-#include <asm/processor.h> /* defines boot_cpu_data */

#define COMPILER_DEPENDENT_INT64 long long
#define COMPILER_DEPENDENT_UINT64 unsigned long long
Index: linux-2.6.16.1/include/asm-i386/processor.h
===================================================================
--- linux-2.6.16.1.orig/include/asm-i386/processor.h 2006-03-30 10:53:45.000000000 -0800
+++ linux-2.6.16.1/include/asm-i386/processor.h 2006-03-30 11:03:16.000000000 -0800
@@ -16,7 +16,6 @@
#include <asm/cpufeature.h>
#include <asm/msr.h>
#include <asm/system.h>
-#include <linux/cache.h>
#include <linux/config.h>
#include <linux/threads.h>
#include <asm/percpu.h>
@@ -39,82 +38,13 @@ struct desc_struct {
*/
#define current_text_addr() ({ void *pc; __asm__("movl $1f,%0\n1:":"=g" (pc)); pc; })

-/*
- * CPU type and hardware bug flags. Kept separately for each CPU.
- * Members of this structure are referenced in head.S, so think twice
- * before touching them. [mj]
- */
-
-struct cpuinfo_x86 {
- __u8 x86; /* CPU family */
- __u8 x86_vendor; /* CPU vendor */
- __u8 x86_model;
- __u8 x86_mask;
- char wp_works_ok; /* It doesn't on 386's */
- char hlt_works_ok; /* Problems on some 486Dx4's and old 386's */
- char hard_math;
- char rfu;
- int cpuid_level; /* Maximum supported CPUID level, -1=no CPUID */
- unsigned long x86_capability[NCAPINTS];
- char x86_vendor_id[16];
- char x86_model_id[64];
- int x86_cache_size; /* in KB - valid for CPUS which support this
- call */
- int x86_cache_alignment; /* In bytes */
- char fdiv_bug;
- char f00f_bug;
- char coma_bug;
- char pad0;
- int x86_power;
- unsigned long loops_per_jiffy;
- unsigned char x86_max_cores; /* cpuid returned max cores value */
- unsigned char booted_cores; /* number of cores as seen by OS */
- unsigned char apicid;
-} __attribute__((__aligned__(SMP_CACHE_BYTES)));
-
-#define X86_VENDOR_INTEL 0
-#define X86_VENDOR_CYRIX 1
-#define X86_VENDOR_AMD 2
-#define X86_VENDOR_UMC 3
-#define X86_VENDOR_NEXGEN 4
-#define X86_VENDOR_CENTAUR 5
-#define X86_VENDOR_RISE 6
-#define X86_VENDOR_TRANSMETA 7
-#define X86_VENDOR_NSC 8
-#define X86_VENDOR_NUM 9
-#define X86_VENDOR_UNKNOWN 0xff
-
-/*
- * capabilities of CPUs
- */
-
-extern struct cpuinfo_x86 boot_cpu_data;
-extern struct cpuinfo_x86 new_cpu_data;
extern struct tss_struct doublefault_tss;
DECLARE_PER_CPU(struct tss_struct, init_tss);

-#ifdef CONFIG_SMP
-extern struct cpuinfo_x86 cpu_data[];
-#define current_cpu_data cpu_data[smp_processor_id()]
-#else
-#define cpu_data (&boot_cpu_data)
-#define current_cpu_data boot_cpu_data
-#endif
-
extern int phys_proc_id[NR_CPUS];
extern int cpu_core_id[NR_CPUS];
extern char ignore_fpu_irq;

-extern void identify_cpu(struct cpuinfo_x86 *);
-extern void print_cpu_info(struct cpuinfo_x86 *);
-extern unsigned int init_intel_cacheinfo(struct cpuinfo_x86 *c);
-
-#ifdef CONFIG_X86_HT
-extern void detect_ht(struct cpuinfo_x86 *c);
-#else
-static inline void detect_ht(struct cpuinfo_x86 *c) {}
-#endif
-
/*
* EFLAGS bits
*/
@@ -720,8 +650,6 @@ static inline void prefetchw(const void

extern void select_idle_routine(const struct cpuinfo_x86 *c);

-#define cache_line_size() (boot_cpu_data.x86_cache_alignment)
-
extern unsigned long boot_option_idle_override;
extern void enable_sep_cpu(void);
extern int sysenter_setup(void);
Index: linux-2.6.16.1/include/asm-i386/system.h
===================================================================
--- linux-2.6.16.1.orig/include/asm-i386/system.h 2006-03-30 10:53:45.000000000 -0800
+++ linux-2.6.16.1/include/asm-i386/system.h 2006-03-30 11:04:28.000000000 -0800
@@ -5,12 +5,80 @@
#include <linux/kernel.h>
#include <asm/segment.h>
#include <asm/cpufeature.h>
+#include <linux/cache.h>
#include <linux/bitops.h> /* for LOCK_PREFIX */

#ifdef __KERNEL__

-struct task_struct; /* one of the stranger aspects of C forward declarations.. */
-extern struct task_struct * FASTCALL(__switch_to(struct task_struct *prev, struct task_struct *next));
+/*
+ * CPU type and hardware bug flags. Kept separately for each CPU.
+ * Members of this structure are referenced in head.S, so think twice
+ * before touching them. [mj]
+ */
+
+struct cpuinfo_x86 {
+ __u8 x86; /* CPU family */
+ __u8 x86_vendor; /* CPU vendor */
+ __u8 x86_model;
+ __u8 x86_mask;
+ char wp_works_ok; /* It doesn't on 386's */
+ char hlt_works_ok; /* Problems on some 486Dx4's and old 386's */
+ char hard_math;
+ char rfu;
+ int cpuid_level; /* Maximum supported CPUID level, -1=no CPUID */
+ unsigned long x86_capability[NCAPINTS];
+ char x86_vendor_id[16];
+ char x86_model_id[64];
+ int x86_cache_size; /* in KB - valid for CPUS which support this
+ call */
+ int x86_cache_alignment; /* In bytes */
+ char fdiv_bug;
+ char f00f_bug;
+ char coma_bug;
+ char pad0;
+ int x86_power;
+ unsigned long loops_per_jiffy;
+ unsigned char x86_max_cores; /* cpuid returned max cores value */
+ unsigned char booted_cores; /* number of cores as seen by OS */
+ unsigned char apicid;
+} __attribute__((__aligned__(SMP_CACHE_BYTES)));
+
+#define X86_VENDOR_INTEL 0
+#define X86_VENDOR_CYRIX 1
+#define X86_VENDOR_AMD 2
+#define X86_VENDOR_UMC 3
+#define X86_VENDOR_NEXGEN 4
+#define X86_VENDOR_CENTAUR 5
+#define X86_VENDOR_RISE 6
+#define X86_VENDOR_TRANSMETA 7
+#define X86_VENDOR_NSC 8
+#define X86_VENDOR_NUM 9
+#define X86_VENDOR_UNKNOWN 0xff
+
+/*
+ * capabilities of CPUs
+ */
+extern struct cpuinfo_x86 boot_cpu_data;
+extern struct cpuinfo_x86 new_cpu_data;
+#define cache_line_size() (boot_cpu_data.x86_cache_alignment)
+
+#ifdef CONFIG_SMP
+extern struct cpuinfo_x86 cpu_data[];
+#define current_cpu_data cpu_data[smp_processor_id()]
+#else
+#define cpu_data (&boot_cpu_data)
+#define current_cpu_data boot_cpu_data
+#endif
+
+extern void identify_cpu(struct cpuinfo_x86 *);
+extern void print_cpu_info(struct cpuinfo_x86 *);
+extern unsigned int init_intel_cacheinfo(struct cpuinfo_x86 *c);
+
+#ifdef CONFIG_X86_HT
+extern void detect_ht(struct cpuinfo_x86 *c);
+#else
+static inline void detect_ht(struct cpuinfo_x86 *c) {}
+#endif

#define switch_to(prev,next,last) do { \
unsigned long esi,edi; \
Index: linux-2.6.16.1/include/asm-i386/mach-visws/mach_hooks.h
===================================================================
--- linux-2.6.16.1.orig/include/asm-i386/mach-visws/mach_hooks.h 2006-03-30 10:53:45.000000000 -0800
+++ linux-2.6.16.1/include/asm-i386/mach-visws/mach_hooks.h 2006-03-30 10:54:27.000000000 -0800
@@ -1,5 +1,5 @@
-#ifndef MACH_HOOKS_H
-#define MACH_HOOKS_H
+#ifndef _MACH_HOOKS_H
+#define _MACH_HOOKS_H

#define pre_intr_init_hook() init_VISWS_APIC_irqs()

Index: linux-2.6.16.1/include/asm-i386/mach-voyager/mach_hooks.h
===================================================================
--- linux-2.6.16.1.orig/include/asm-i386/mach-voyager/mach_hooks.h 2006-03-30 10:53:45.000000000 -0800
+++ linux-2.6.16.1/include/asm-i386/mach-voyager/mach_hooks.h 2006-03-30 10:54:27.000000000 -0800
@@ -1,5 +1,5 @@
-#ifndef MACH_HOOKS_H
-#define MACH_HOOKS_H
+#ifndef _MACH_HOOKS_H
+#define _MACH_HOOKS_H

/**
* intr_init_hook - post gate setup interrupt initialisation

Next message: Matthew Wilcox: "Re: [PATCH] ioremap_cached()"
Previous message: Mingming Cao: "Re: [RFC][PATCH 0/2]Extend ext3 filesystem limit from 8TB to 16TB"
In reply to: Chris Wright: "Re: [PATCH] Cleanup subarch definitions in Linux/i386"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]