From 0a4d8a472f645d99f86303db1462b64e371b090d Mon Sep 17 00:00:00 2001 From: Glauber Costa Date: Tue, 24 Jun 2008 09:34:08 -0300 Subject: x86: provide delay loop for x86_64. This is for consistency with i386. We call use_tsc_delay() at tsc initialization for x86_64, so we'll be always using it. Signed-off-by: Glauber Costa Signed-off-by: H. Peter Anvin Signed-off-by: Ingo Molnar --- arch/x86/lib/delay_64.c | 44 ++++++++++++++++++++++++++++++++++++++++---- 1 file changed, 40 insertions(+), 4 deletions(-) (limited to 'arch/x86/lib/delay_64.c') diff --git a/arch/x86/lib/delay_64.c b/arch/x86/lib/delay_64.c index 4c441be92641..d0326d07c845 100644 --- a/arch/x86/lib/delay_64.c +++ b/arch/x86/lib/delay_64.c @@ -22,13 +22,28 @@ #include #endif -int __devinit read_current_timer(unsigned long *timer_value) +/* simple loop based delay: */ +static void delay_loop(unsigned long loops) { - rdtscll(*timer_value); - return 0; + asm volatile( + " test %0,%0 \n" + " jz 3f \n" + " jmp 1f \n" + + ".align 16 \n" + "1: jmp 2f \n" + + ".align 16 \n" + "2: dec %0 \n" + " jnz 2b \n" + "3: dec %0 \n" + + : /* we don't need output */ + :"a" (loops) + ); } -void __delay(unsigned long loops) +static void delay_tsc(unsigned long loops) { unsigned bclock, now; int cpu; @@ -63,6 +78,27 @@ void __delay(unsigned long loops) } preempt_enable(); } + +static void (*delay_fn)(unsigned long) = delay_loop; + +void use_tsc_delay(void) +{ + delay_fn = delay_tsc; +} + +int __devinit read_current_timer(unsigned long *timer_value) +{ + if (delay_fn == delay_tsc) { + rdtscll(*timer_value); + return 0; + } + return -1; +} + +void __delay(unsigned long loops) +{ + delay_fn(loops); +} EXPORT_SYMBOL(__delay); inline void __const_udelay(unsigned long xloops) -- cgit v1.2.1 From 7e58818d32c18197602d1869b22cfda99efd05fe Mon Sep 17 00:00:00 2001 From: Glauber Costa Date: Tue, 24 Jun 2008 10:21:25 -0300 Subject: x86: explicitly use edx in const delay function. For x86_64, we can't just use %0, as it would generate a mul against rdx, which is not really what we want (note the ">> 32" in x86_64 version). Using a u64 variable with a shift in i386 generates bad code, so the solution is to explicitly use %%edx in inline assembly for both. Signed-off-by: Glauber Costa Signed-off-by: H. Peter Anvin Signed-off-by: Ingo Molnar --- arch/x86/lib/delay_64.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) (limited to 'arch/x86/lib/delay_64.c') diff --git a/arch/x86/lib/delay_64.c b/arch/x86/lib/delay_64.c index d0326d07c845..ff3dfecdb6f9 100644 --- a/arch/x86/lib/delay_64.c +++ b/arch/x86/lib/delay_64.c @@ -103,9 +103,16 @@ EXPORT_SYMBOL(__delay); inline void __const_udelay(unsigned long xloops) { - __delay(((xloops * HZ * - cpu_data(raw_smp_processor_id()).loops_per_jiffy) >> 32) + 1); + int d0; + xloops *= 4; + __asm__("mull %%edx" + :"=d" (xloops), "=&a" (d0) + :"1" (xloops), "0" + (cpu_data(raw_smp_processor_id()).loops_per_jiffy * (HZ/4))); + + __delay(++xloops); } + EXPORT_SYMBOL(__const_udelay); void __udelay(unsigned long usecs) -- cgit v1.2.1 From f0fbf0abc093ec8bf64506eee4ede9e5daf40ffd Mon Sep 17 00:00:00 2001 From: Glauber Costa Date: Thu, 3 Jul 2008 12:35:41 -0300 Subject: x86: integrate delay functions. delay_32.c, delay_64.c are now equal, and are integrated into delay.c. Signed-off-by: Glauber Costa Signed-off-by: H. Peter Anvin Signed-off-by: Ingo Molnar --- arch/x86/lib/delay_64.c | 128 ------------------------------------------------ 1 file changed, 128 deletions(-) delete mode 100644 arch/x86/lib/delay_64.c (limited to 'arch/x86/lib/delay_64.c') diff --git a/arch/x86/lib/delay_64.c b/arch/x86/lib/delay_64.c deleted file mode 100644 index ff3dfecdb6f9..000000000000 --- a/arch/x86/lib/delay_64.c +++ /dev/null @@ -1,128 +0,0 @@ -/* - * Precise Delay Loops for x86-64 - * - * Copyright (C) 1993 Linus Torvalds - * Copyright (C) 1997 Martin Mares - * - * The __delay function must _NOT_ be inlined as its execution time - * depends wildly on alignment on many x86 processors. - */ - -#include -#include -#include -#include -#include -#include - -#include -#include - -#ifdef CONFIG_SMP -#include -#endif - -/* simple loop based delay: */ -static void delay_loop(unsigned long loops) -{ - asm volatile( - " test %0,%0 \n" - " jz 3f \n" - " jmp 1f \n" - - ".align 16 \n" - "1: jmp 2f \n" - - ".align 16 \n" - "2: dec %0 \n" - " jnz 2b \n" - "3: dec %0 \n" - - : /* we don't need output */ - :"a" (loops) - ); -} - -static void delay_tsc(unsigned long loops) -{ - unsigned bclock, now; - int cpu; - - preempt_disable(); - cpu = smp_processor_id(); - rdtscl(bclock); - for (;;) { - rdtscl(now); - if ((now - bclock) >= loops) - break; - - /* Allow RT tasks to run */ - preempt_enable(); - rep_nop(); - preempt_disable(); - - /* - * It is possible that we moved to another CPU, and - * since TSC's are per-cpu we need to calculate - * that. The delay must guarantee that we wait "at - * least" the amount of time. Being moved to another - * CPU could make the wait longer but we just need to - * make sure we waited long enough. Rebalance the - * counter for this CPU. - */ - if (unlikely(cpu != smp_processor_id())) { - loops -= (now - bclock); - cpu = smp_processor_id(); - rdtscl(bclock); - } - } - preempt_enable(); -} - -static void (*delay_fn)(unsigned long) = delay_loop; - -void use_tsc_delay(void) -{ - delay_fn = delay_tsc; -} - -int __devinit read_current_timer(unsigned long *timer_value) -{ - if (delay_fn == delay_tsc) { - rdtscll(*timer_value); - return 0; - } - return -1; -} - -void __delay(unsigned long loops) -{ - delay_fn(loops); -} -EXPORT_SYMBOL(__delay); - -inline void __const_udelay(unsigned long xloops) -{ - int d0; - xloops *= 4; - __asm__("mull %%edx" - :"=d" (xloops), "=&a" (d0) - :"1" (xloops), "0" - (cpu_data(raw_smp_processor_id()).loops_per_jiffy * (HZ/4))); - - __delay(++xloops); -} - -EXPORT_SYMBOL(__const_udelay); - -void __udelay(unsigned long usecs) -{ - __const_udelay(usecs * 0x000010c7); /* 2**32 / 1000000 (rounded up) */ -} -EXPORT_SYMBOL(__udelay); - -void __ndelay(unsigned long nsecs) -{ - __const_udelay(nsecs * 0x00005); /* 2**32 / 1000000000 (rounded up) */ -} -EXPORT_SYMBOL(__ndelay); -- cgit v1.2.1