[PATCH] Kill CONFIG_4KSTACKS

[PATCH] Kill CONFIG_4KSTACKS

[Chris Wedgwood]

http://linux.bkbits.net:8080/linux-2.5/cset@407af9a3SmWwuO0CEQwLmAZoLXAcCA
seems to indicate 8K stacks are deprecated.  This removes the option
completely so were now use 4k-process + 4k-irq stacks all the time.

Probably a bit premature...



 arch/i386/Kconfig.debug        |    9 ---------
 arch/i386/defconfig            |    1 -
 arch/i386/kernel/irq.c         |   24 +-----------------------
 include/asm-i386/irq.h         |    4 ----
 include/asm-i386/module.h      |    8 +-------
 include/asm-i386/thread_info.h |    4 ----
 6 files changed, 2 insertions(+), 48 deletions(-)


===== arch/i386/Kconfig.debug 1.3 vs edited =====
--- 1.3/arch/i386/Kconfig.debug	2004-08-31 00:55:12 -07:00
+++ edited/arch/i386/Kconfig.debug	2004-09-11 13:11:39 -07:00
@@ -46,15 +46,6 @@
 	  This results in a large slowdown, but helps to find certain types
 	  of memory corruptions.
 
-config 4KSTACKS
-	bool "Use 4Kb for kernel stacks instead of 8Kb"
-	help
-	  If you say Y here the kernel will use a 4Kb stacksize for the
-	  kernel stack attached to each process/thread. This facilitates
-	  running more threads on a system and also reduces the pressure
-	  on the VM subsystem for higher order allocations. This option
-	  will also use IRQ stacks to compensate for the reduced stackspace.
-
 config SCHEDSTATS
 	bool "Collect scheduler statistics"
 	depends on DEBUG_KERNEL && PROC_FS
===== arch/i386/defconfig 1.113 vs edited =====
--- 1.113/arch/i386/defconfig	2004-06-27 00:19:31 -07:00
+++ edited/arch/i386/defconfig	2004-09-11 12:58:09 -07:00
@@ -1221,7 +1221,6 @@
 CONFIG_EARLY_PRINTK=y
 CONFIG_DEBUG_SPINLOCK_SLEEP=y
 # CONFIG_FRAME_POINTER is not set
-CONFIG_4KSTACKS=y
 CONFIG_X86_FIND_SMP_CONFIG=y
 CONFIG_X86_MPPARSE=y
 
===== arch/i386/kernel/irq.c 1.59 vs edited =====
--- 1.59/arch/i386/kernel/irq.c	2004-08-31 00:57:57 -07:00
+++ edited/arch/i386/kernel/irq.c	2004-09-11 13:15:23 -07:00
@@ -76,10 +76,8 @@
 /*
  * per-CPU IRQ handling stacks
  */
-#ifdef CONFIG_4KSTACKS
 union irq_ctx *hardirq_ctx[NR_CPUS];
 union irq_ctx *softirq_ctx[NR_CPUS];
-#endif
 
 /*
  * Special irq handlers.
@@ -435,7 +433,7 @@
 	irq_enter();
 
 #ifdef CONFIG_DEBUG_STACKOVERFLOW
-	/* Debugging check for stack overflow: is there less than 1KB free? */
+	/* Debugging check for stack overflow */
 	{
 		long esp;
 
@@ -489,7 +487,6 @@
 	 * useful for irq hardware that does not mask cleanly in an
 	 * SMP environment.
 	 */
-#ifdef CONFIG_4KSTACKS
 
 	for (;;) {
 		irqreturn_t action_ret;
@@ -542,23 +539,6 @@
 		desc->status &= ~IRQ_PENDING;
 	}
 
-#else
-
-	for (;;) {
-		irqreturn_t action_ret;
-
-		spin_unlock(&desc->lock);
-
-		action_ret = handle_IRQ_event(irq, &regs, action);
-
-		spin_lock(&desc->lock);
-		if (!noirqdebug)
-			note_interrupt(irq, desc, action_ret);
-		if (likely(!(desc->status & IRQ_PENDING)))
-			break;
-		desc->status &= ~IRQ_PENDING;
-	}
-#endif
 	desc->status &= ~IRQ_INPROGRESS;
 
 out:
@@ -1077,7 +1057,6 @@
 }
 
 
-#ifdef CONFIG_4KSTACKS
 /*
  * These should really be __section__(".bss.page_aligned") as well, but
  * gcc's 3.0 and earlier don't handle that correctly.
@@ -1155,4 +1134,3 @@
 }
 
 EXPORT_SYMBOL(do_softirq);
-#endif
===== include/asm-i386/irq.h 1.11 vs edited =====
--- 1.11/include/asm-i386/irq.h	2004-04-12 10:54:45 -07:00
+++ edited/include/asm-i386/irq.h	2004-09-11 12:58:19 -07:00
@@ -31,7 +31,6 @@
 #define ARCH_HAS_NMI_WATCHDOG		/* See include/linux/nmi.h */
 #endif
 
-#ifdef CONFIG_4KSTACKS
 /*
  * per-CPU IRQ handling contexts (thread information and stack)
  */
@@ -46,9 +45,6 @@
 extern void irq_ctx_init(int cpu);
 
 #define __ARCH_HAS_DO_SOFTIRQ
-#else
-#define irq_ctx_init(cpu) do { ; } while (0)
-#endif
 
 struct irqaction;
 struct pt_regs;
===== include/asm-i386/module.h 1.11 vs edited =====
--- 1.11/include/asm-i386/module.h	2004-04-12 10:54:45 -07:00
+++ edited/include/asm-i386/module.h	2004-09-11 13:01:43 -07:00
@@ -60,12 +60,6 @@
 #define MODULE_REGPARM ""
 #endif
 
-#ifdef CONFIG_4KSTACKS
-#define MODULE_STACKSIZE "4KSTACKS "
-#else
-#define MODULE_STACKSIZE ""
-#endif
-
-#define MODULE_ARCH_VERMAGIC MODULE_PROC_FAMILY MODULE_REGPARM MODULE_STACKSIZE
+#define MODULE_ARCH_VERMAGIC MODULE_PROC_FAMILY MODULE_REGPARM
 
 #endif /* _ASM_I386_MODULE_H */
===== include/asm-i386/thread_info.h 1.20 vs edited =====
--- 1.20/include/asm-i386/thread_info.h	2004-08-23 01:14:45 -07:00
+++ edited/include/asm-i386/thread_info.h	2004-09-11 13:01:43 -07:00
@@ -52,11 +52,7 @@
 #endif
 
 #define PREEMPT_ACTIVE		0x4000000
-#ifdef CONFIG_4KSTACKS
 #define THREAD_SIZE            (4096)
-#else
-#define THREAD_SIZE		(8192)
-#endif
 
 #define STACK_WARN             (THREAD_SIZE/8)
 /*

Re: [PATCH] Kill CONFIG_4KSTACKS

[William Lee Irwin III]

On Sat, Sep 11, 2004 at 01:41:25PM -0700, Chris Wedgwood wrote:
> http://linux.bkbits.net:8080/linux-2.5/cset@407af9a3SmWwuO0CEQwLmAZoLXAcCA
> seems to indicate 8K stacks are deprecated.  This removes the option
> completely so were now use 4k-process + 4k-irq stacks all the time.
> Probably a bit premature...

Another, distinct patch to add warnings or errors for all versions of
gcc prior to the stack fix commit might be helpful.


-- wli

Re: [PATCH] Kill CONFIG_4KSTACKS

[Chris Wedgwood]

On Sat, Sep 11, 2004 at 10:00:30PM -0700, William Lee Irwin III wrote:

> Another, distinct patch to add warnings or errors for all versions
> of gcc prior to the stack fix commit might be helpful.

It's only for comments right now, but sure yes.  And if we are going
to do that we should drop support for ancient gcc versions[1] and fix
up some 'sections' presently used (I thought we had some hacks in
there now for gcc-2.95, I should recheck).


 --cw

[1] it used to be sparc64 was a problem when this was discussed
    before, is that still the case?