From 312df5f1a1da780e084b328bcabb02a6dcd044c3 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Mon, 16 May 2005 21:53:28 -0700 Subject: [PATCH] x86_64: Add pmtimer support There are unfortunately more and more multi processor Opteron systems which don't have HPET timer support in the southbridge. This covers in particular Nvidia and VIA chipsets. They also don't guarantee that the TSCs are synchronized between CPUs; and especially with MP powernow the systems are nearly unusable because the time gets very inconsistent between CPUs. The timer code for x86-64 was originally written under the assumption that we could fall back to the HPET timer on such systems. But this doesn't work there. Another alternative is to use the ACPI PM timer as primary time source. This patch does that. The kernel only uses PM timer when there is no other choice because it has some disadvantages. Ported over from i386. It should be faster than the i386 version because I dropped the "read three times" workaround, but is still considerable slower than HPET and also does not work together with vsyscalls which have to be disabled. Cc: Signed-off-by: Andi Kleen Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/x86_64/kernel/time.c | 62 ++++++++++++++++++++++++++++++++++------------- 1 file changed, 45 insertions(+), 17 deletions(-) (limited to 'arch/x86_64/kernel/time.c') diff --git a/arch/x86_64/kernel/time.c b/arch/x86_64/kernel/time.c index 0e3b60e1ac5..735b6767c8e 100644 --- a/arch/x86_64/kernel/time.c +++ b/arch/x86_64/kernel/time.c @@ -26,6 +26,8 @@ #include #include #include +#include +#include /* for PM timer frequency */ #include #include #include @@ -396,6 +398,10 @@ static irqreturn_t timer_interrupt(int irq, void *dev_id, struct pt_regs *regs) (offset - vxtime.last)*(NSEC_PER_SEC/HZ) / hpet_tick; vxtime.last = offset; +#ifdef CONFIG_X86_PM_TIMER + } else if (vxtime.mode == VXTIME_PMTMR) { + lost = pmtimer_mark_offset(); +#endif } else { offset = (((tsc - vxtime.last_tsc) * vxtime.tsc_quot) >> 32) - (USEC_PER_SEC / HZ); @@ -898,6 +904,13 @@ void __init time_init(void) hpet_period; cpu_khz = hpet_calibrate_tsc(); timename = "HPET"; +#ifdef CONFIG_X86_PM_TIMER + } else if (pmtmr_ioport) { + vxtime_hz = PM_TIMER_FREQUENCY; + timename = "PM"; + pit_init(); + cpu_khz = pit_calibrate_tsc(); +#endif } else { pit_init(); cpu_khz = pit_calibrate_tsc(); @@ -922,6 +935,27 @@ void __init time_init(void) #endif } +/* + * Make an educated guess if the TSC is trustworthy and synchronized + * over all CPUs. + */ +static __init int unsynchronized_tsc(void) +{ +#ifdef CONFIG_SMP + if (oem_force_hpet_timer()) + return 1; + /* Intel systems are normally all synchronized. Exceptions + are handled in the OEM check above. */ + if (boot_cpu_data.x86_vendor == X86_VENDOR_INTEL) + return 0; + /* All in a single socket - should be synchronized */ + if (cpus_weight(cpu_core_map[0]) == num_online_cpus()) + return 0; +#endif + /* Assume multi socket systems are not synchronized */ + return num_online_cpus() > 1; +} + /* * Decide after all CPUs are booted what mode gettimeofday should use. */ @@ -929,29 +963,23 @@ void __init time_init_gtod(void) { char *timetype; - /* - * AMD systems with more than one CPU don't have fully synchronized - * TSCs. Always use HPET gettimeofday for these, although it is slower. - * Intel SMP systems usually have synchronized TSCs, so use always - * the TSC. - * - * Exceptions: - * IBM Summit2 checked by oem_force_hpet_timer(). - * AMD dual core may also not need HPET. Check me. - * - * Can be turned off with "notsc". - */ - if (num_online_cpus() > 1 && - boot_cpu_data.x86_vendor == X86_VENDOR_AMD) - notsc = 1; - /* Some systems will want to disable TSC and use HPET. */ - if (oem_force_hpet_timer()) + if (unsynchronized_tsc()) notsc = 1; if (vxtime.hpet_address && notsc) { timetype = "HPET"; vxtime.last = hpet_readl(HPET_T0_CMP) - hpet_tick; vxtime.mode = VXTIME_HPET; do_gettimeoffset = do_gettimeoffset_hpet; +#ifdef CONFIG_X86_PM_TIMER + /* Using PM for gettimeofday is quite slow, but we have no other + choice because the TSC is too unreliable on some systems. */ + } else if (pmtmr_ioport && !vxtime.hpet_address && notsc) { + timetype = "PM"; + do_gettimeoffset = do_gettimeoffset_pm; + vxtime.mode = VXTIME_PMTMR; + sysctl_vsyscall = 0; + printk(KERN_INFO "Disabling vsyscall due to use of PM timer\n"); +#endif } else { timetype = vxtime.hpet_address ? "HPET/TSC" : "PIT/TSC"; vxtime.mode = VXTIME_TSC; -- cgit v1.2.3