On Wed, Aug 23, 2017 at 09:29:15PM +0800, Adam Steen wrote:
> On Thu, Aug 17, 2017 at 12:19:28PM +0800, Adam Steen wrote:
> > On Tue, Aug 8, 2017 at 10:12 PM, Mike Belopuhov <[email protected]> wrote:
> > > On Tue, Aug 08, 2017 at 08:18 +0800, Adam Steen wrote:
> > >> On Mon, Jul 31, 2017 at 3:58 PM, Mike Belopuhov <[email protected]>
> > >> wrote:
> > >> > On Mon, Jul 31, 2017 at 09:48 +0800, Adam Steen wrote:
> > >> >> Ted Unangst wrote:
> > >> >> > we don't currently export this info, but we could add some sysctls.
> > >> >> > there's
> > >> >> > some cpufeatures stuff there, but generally stuff isn't exported
> > >> >> > until
> > >> >> > somebody finds a use for it... it shouldn't be too hard to add
> > >> >> > something to
> > >> >> > amd64/machdep.c sysctl if you're interested.
> > >> >>
> > >> >> I am interested, as i need the info, i will look into it and hopefully
> > >> >> come back with a patch.
> > >> >
> > >> > This is a bad idea because TSC as the time source is only usable
> > >> > by OpenBSD on Skylake and Kaby Lake CPUs since they encode the TSC
> > >> > frequency in the CPUID. All older CPUs have their TSCs measured
> > >> > against the PIT. Currently the measurement done by the kernel isn't
> > >> > very precise and if TSC is selected as a timecounter, the machine
> > >> > would be gaining time on a pace that cannot be corrected by our NTP
> > >> > daemon. (IIRC, about an hour a day on my Haswell running with NTP).
> > >> >
> > >> > To be able to use TSC as a timecounter source on OpenBSD or Solo5
> > >> > you'd have to improve the in-kernel measurement of the TSC frequency
> > >> > first. I've tried to perform 10 measurements and take an average and
> > >> > it does improve accuracy, however I believe we need to poach another
> > >> > bit from Linux and re-calibrate TSC via HPET:
> > >> >
> > >> >
> > >> > http://elixir.free-electrons.com/linux/v4.12.4/source/arch/x86/kernel/tsc.c#L409
> > >> >
> > >> > I think this is the most sane thing we can do. Here's a complete
> > >> > procedure that Linux kernel undertakes:
> > >> >
> > >> >
> > >> > http://elixir.free-electrons.com/linux/v4.12.4/source/arch/x86/kernel/tsc.c#L751
> > >> >
> > >> > Regards,
> > >> > Mike
> > >>
> > >> Hi Mike/All
> > >>
> > >> I would like to improve the accuracy of TSC frequency calibration as
> > >> Mike B. describes above.
> > >>
> > >> I initially thought the calibration would take place at line 470 of
> > >> amd64/identcpu.c
> > >> (https://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sys/arch/amd64/amd64/identcpu.c?annotate=1.87)
> > >>
> > >
> > > Indeed, it cannot happen there simply because you don't know at
> > > that point whether or not HPET actually exists.
> > >
> > >> But I looked into using the acpihpet directly but it is never exposed
> > >> outside of acpihpet.c.
> > >>
> > >
> > > And it shouldn't be.
> > >
> > >> Could someone point me to were if would be appropriate to complete
> > >> this calibration and how to use the acpihpet?
> > >
> > > The way I envision this is a multi-step approach:
> > >
> > > 1) TSC frequency is approximated with the PIT (possibly performing
> > > multiple measurements and averaging them out; also keep in mind that
> > > doing it 8 times means you can shift the sum right by 3 instead of
> > > using actual integer division). This is what should happen around
> > > the line 470 of identcpu.c
> > >
> > > 2) A function can be provided by identcpu.c to further adjust the
> > > TSC frequency once acpitimer(4) (this is a PM timer) and acpihpet(4)
> > > (or any other timer for that matter) are attached.
> > >
> > > 3) Once acpitimer(4) or acpihpet(4) or any other timecounter source
> > > are attached and are verified to be operating correctly, they can
> > > perform TSC re-calibration and update the TSC frequency with their
> > > measurements. The idea here is that the function (or functions) that
> > > facilitate this must abstract enough logic so that you don't have to
> > > duplicate it in the acpitimer or acpihpet themselves.
> > >
> > >> (Will it need to be
> > >> exposed like i8254_delay/delay_func/delay in machdep.c and cpu.h)
> > >>
> > >
> > > No it won't.
> > >
> > >> Lastly should the calibration be done using both delay(i8254 pit) and
> > >> hpet timers similar to Linux described above or just using the hpet?
> > >>
> > >
> > > Well, that's what I was arguing for. As I said in my initial mail
> > > on misc (not quoted here), the TSC must be calibrated using separate
> > > known clocks sources.
> >
> > Hi Mike
> >
> > Please see the below diff to improve the accuracy of the TSC
> > frequency. It is model after the linux calibration you linked to
> > earlier. https://marc.info/?l=openbsd-misc&m=150148792804747&w=2
> >
> > I feel like i don't know enough about the kernel internals, the
> > consistency of the results across reboots are not as close as i would
> > have liked, i feel the call to do the actual calibration should be
> > later in the boot cycle, when things have calmed down a little, but
> > couldn't figure out the best way of doing this.
> >
> > please bear with me i haven't been programming c for long, but the
> > only way to get things done is to do it your self.
> >
> > Cheers
> > Adam
>
> <SNIP>
>
> Thank you Mike on the feedback on the last patch, please see the diff
> below, update with your input and style(9)
>
> I have continued to use tsc as my timecounter and /var/db/ntpd.driff
> has stayed under 10.
>
> cat /var/db/ntpd.drift
> 6.665
>
> ntpctl -s all
> 4/4 peers valid, constraint offset -1s, clock synced, stratum 3
>
> peer
> wt tl st next poll offset delay jitter
> 144.48.166.166 from pool pool.ntp.org
> 1 10 2 4s 32s -3.159ms 87.723ms 10.389ms
> 13.55.50.68 from pool pool.ntp.org
> 1 10 3 11s 32s -3.433ms 86.053ms 18.095ms
> 14.202.204.182 from pool pool.ntp.org
> 1 10 1 14s 32s 1.486ms 86.545ms 16.483ms
> 27.124.125.250 from pool pool.ntp.org
> * 1 10 2 12s 30s -10.275ms 54.156ms 70.389ms
>
> Cheers
> Adam
IIRC you have an x220, right?
If so, could you try letting the clock run for a bit (while using tsc
timecounter selection) after apm -L to drop the speed? (make sure
apm shows that it dropped).
Even though my x230 supposedly has a constant/invar TSC (according to
cpuid), the TSC drops from 2.5GHz to 1.2GHz when apm -L runs, which
causes time to run too slowly when tsc is selected there.
-ml
> Index: sys/arch/amd64/amd64/acpi_machdep.c
> ===================================================================
> RCS file: /cvs/src/sys/arch/amd64/amd64/acpi_machdep.c,v
> retrieving revision 1.78
> diff -u -p -u -r1.78 acpi_machdep.c
> --- sys/arch/amd64/amd64/acpi_machdep.c 27 Mar 2017 18:32:53 -0000
> 1.78
> +++ sys/arch/amd64/amd64/acpi_machdep.c 23 Aug 2017 11:43:25 -0000
> @@ -69,6 +69,17 @@ extern int acpi_savecpu(void) __returns_
>
> u_int8_t *acpi_scan(struct acpi_mem_map *, paddr_t, size_t);
>
> +#define RECALIBRATE_MAX_RETRIES 5
> +#define RECALIBRATE_SMI_THRESHOLD 50000
> +#define RECALIBRATE_DELAY_THRESHOLD 20
> +struct timecounter *acpitimer_recalibrate_timecounter;
> +struct timecounter *acpihpet_recalibrate_timecounter;
> +uint64_t acpi_calculate_timecounter_delay(struct timecounter *, uint64_t,
> + uint64_t);
> +uint64_t acpi_calculate_tsc_frequency(uint64_t, uint64_t, int);
> +uint64_t acpi_calibrate_tsc_frequency();
> +int acpi_get_tsc_and_timecount(struct timecounter *, uint64_t *, uint64_t *);
> +
> int
> acpi_map(paddr_t pa, size_t len, struct acpi_mem_map *handle)
> {
> @@ -438,6 +449,123 @@ acpi_resume_cpu(struct acpi_softc *sc)
> /* Re-initialise memory range handling on BSP */
> if (mem_range_softc.mr_op != NULL)
> mem_range_softc.mr_op->initAP(&mem_range_softc);
> +}
> +
> +int
> +acpi_get_tsc_and_timecount(struct timecounter *tc, uint64_t *tsc,
> + uint64_t *count)
> +{
> + uint64_t n, tsc1, tsc2;
> + int i;
> +
> + for (i = 0; i < RECALIBRATE_MAX_RETRIES; i++) {
> + tsc1 = rdtsc();
> + n = (tc->tc_get_timecount(tc) & tc->tc_counter_mask);
> + tsc2 = rdtsc();
> +
> + if ((tsc2 - tsc1) < RECALIBRATE_SMI_THRESHOLD) {
> + *count = n;
> + *tsc = tsc2;
> + return (0);
> + }
> + }
> + return 1;
> +}
> +
> +uint64_t
> +acpi_calculate_timecounter_delay(struct timecounter *tc, uint64_t count1,
> + uint64_t count2)
> +{
> + uint64_t delta;
> +
> + if (count2 < count1)
> + count2 += tc->tc_counter_mask;
> +
> + delta = (count2 - count1);
> + return delta * 1000000 / tc->tc_frequency;
> +}
> +
> +uint64_t
> +acpi_calculate_tsc_frequency(uint64_t tsc1, uint64_t tsc2, int usec)
> +{
> + uint64_t delta;
> +
> + delta = (tsc2 - tsc1);
> + return delta * 1000000 / usec;
> +}
> +
> +void
> +acpitimer_recalibrate_timecounter_init(struct timecounter *tc)
> +{
> + acpitimer_recalibrate_timecounter = tc;
> + cpu_adjust_tsc_frequency(&acpi_calibrate_tsc_frequency);
> +}
> +
> +void
> +acpihpet_recalibrate_timecounter_init(struct timecounter *tc)
> +{
> + acpihpet_recalibrate_timecounter = tc;
> + cpu_adjust_tsc_frequency(&acpi_calibrate_tsc_frequency);
> +}
> +
> +uint64_t
> +acpi_calibrate_tsc_frequency()
> +{
> + struct timecounter *tc;
> + uint64_t count1, count2, frequency, min_freq, tsc1, tsc2;
> + u_long ef;
> + int delay_usec, i, err1, err2, usec;
> +
> + if (acpihpet_recalibrate_timecounter != NULL) {
> + tc = acpihpet_recalibrate_timecounter;
> + } else if (acpitimer_recalibrate_timecounter != NULL) {
> + tc = acpitimer_recalibrate_timecounter;
> + } else {
> + printf("Failed to recalibrate timecounter \"tsc\"");
> + printf(" No timcounter set");
> + return 0;
> + }
> +
> + /* warmup the timers */
> + for (i =0; i < 3; i++) {
> + (void)tc->tc_get_timecount(tc);
> + (void)rdtsc();
> + }
> +
> + min_freq = ULLONG_MAX;
> +
> + delay_usec = 100000;
> + for (i = 0; i < 3; i++) {
> + ef = read_rflags();
> + disable_intr();
> +
> + err1 = acpi_get_tsc_and_timecount(tc, &tsc1, &count1);
> + delay(delay_usec);
> + err2 = acpi_get_tsc_and_timecount(tc, &tsc2, &count2);
> +
> + write_rflags(ef);
> +
> + if (err1 || err2)
> + continue;
> +
> + usec = acpi_calculate_timecounter_delay(tc, count1, count2);
> +
> + if (usec < (delay_usec - RECALIBRATE_DELAY_THRESHOLD))
> + continue;
> +
> + if (usec > (delay_usec + RECALIBRATE_DELAY_THRESHOLD))
> + continue;
> +
> + frequency = acpi_calculate_tsc_frequency(tsc1, tsc2, usec);
> +
> + min_freq = ulmin(min_freq, frequency);
> + }
> +
> + if (min_freq == 0) {
> + printf("Failed to recalibrate timecounter \"tsc\"");
> + printf(" with timecounter \"%s\"\n", tc->tc_name);
> + }
> + return min_freq;
> }
>
> #ifdef MULTIPROCESSOR
> Index: sys/arch/amd64/amd64/identcpu.c
> ===================================================================
> RCS file: /cvs/src/sys/arch/amd64/amd64/identcpu.c,v
> retrieving revision 1.87
> diff -u -p -u -r1.87 identcpu.c
> --- sys/arch/amd64/amd64/identcpu.c 20 Jun 2017 05:34:41 -0000 1.87
> +++ sys/arch/amd64/amd64/identcpu.c 23 Aug 2017 11:43:26 -0000
> @@ -63,6 +63,10 @@ struct timecounter tsc_timecounter = {
> tsc_get_timecount, NULL, ~0u, 0, "tsc", -1000, NULL
> };
>
> +u_int64_t (*amd64_tsc_get_frequency)();
> +u_int64_t amd64_tsc_freq;
> +int amd64_tsc_requires_calibration;
> +int amd64_has_invariant_tsc;
> int amd64_has_xcrypt;
> #ifdef CRYPTO
> int amd64_has_pclmul;
> @@ -428,8 +432,9 @@ cpu_tsc_freq_ctr(struct cpu_info *ci)
> u_int64_t
> cpu_tsc_freq(struct cpu_info *ci)
> {
> - u_int64_t last_count, count;
> + u_int64_t last_count, count, sum;
> uint32_t eax, ebx, khz, dummy;
> + int i;
>
> if (!strcmp(cpu_vendor, "GenuineIntel") &&
> cpuid_level >= 0x15) {
> @@ -468,11 +473,21 @@ cpu_tsc_freq(struct cpu_info *ci)
> if (count != 0)
> return (count);
>
> - last_count = rdtsc();
> - delay(100000);
> - count = rdtsc();
> + if (amd64_tsc_get_frequency != NULL)
> + return (*amd64_tsc_get_frequency)();
>
> - return ((count - last_count) * 10);
> + if (ci->ci_flags & CPUF_PRIMARY)
> + amd64_tsc_requires_calibration = 1;
> +
> + sum = 0;
> + for (i=0; i < 8; i++) {
> + last_count = rdtsc();
> + delay(100000);
> + count = rdtsc();
> + sum += (count - last_count);
> + }
> +
> + return ((sum >> 3) * 10);
> }
>
> u_int
> @@ -482,6 +497,22 @@ tsc_get_timecount(struct timecounter *tc
> }
>
> void
> +cpu_adjust_tsc_frequency(u_int64_t (*get_frequency)())
> +{
> + if (amd64_tsc_requires_calibration) {
> + amd64_tsc_freq = (*get_frequency)();
> + tsc_timecounter.tc_frequency = amd64_tsc_freq;
> +
> + printf("Timecounter \"%s\" recalibrated",
> + tsc_timecounter.tc_name);
> + printf("with frequency %llu Hz\n",
> + tsc_timecounter.tc_frequency);
> + } else {
> + amd64_tsc_get_frequency = get_frequency;
> + }
> +}
> +
> +void
> identifycpu(struct cpu_info *ci)
> {
> u_int32_t dummy, val;
> @@ -566,9 +597,13 @@ identifycpu(struct cpu_info *ci)
> /* Check if it's an invariant TSC */
> if (cpu_apmi_edx & CPUIDEDX_ITSC)
> ci->ci_flags |= CPUF_INVAR_TSC;
> +
> + amd64_has_invariant_tsc = (ci->ci_flags & CPUF_INVAR_TSC) != 0;
> }
>
> ci->ci_tsc_freq = cpu_tsc_freq(ci);
> + if (ci->ci_flags & CPUF_PRIMARY)
> + amd64_tsc_freq = ci->ci_tsc_freq;
>
> amd_cpu_cacheinfo(ci);
>
> Index: sys/arch/amd64/amd64/machdep.c
> ===================================================================
> RCS file: /cvs/src/sys/arch/amd64/amd64/machdep.c,v
> retrieving revision 1.231
> diff -u -p -u -r1.231 machdep.c
> --- sys/arch/amd64/amd64/machdep.c 12 Jul 2017 06:26:32 -0000 1.231
> +++ sys/arch/amd64/amd64/machdep.c 23 Aug 2017 11:43:26 -0000
> @@ -425,6 +425,8 @@ int
> cpu_sysctl(int *name, u_int namelen, void *oldp, size_t *oldlenp, void *newp,
> size_t newlen, struct proc *p)
> {
> + extern u_int64_t amd64_tsc_freq;
> + extern int amd64_has_invariant_tsc;
> extern int amd64_has_xcrypt;
> dev_t consdev;
> dev_t dev;
> @@ -496,6 +498,11 @@ cpu_sysctl(int *name, u_int namelen, voi
> pckbc_release_console();
> return (error);
> #endif
> + case CPU_TSCFREQ:
> + return (sysctl_rdquad(oldp, oldlenp, newp, amd64_tsc_freq));
> + case CPU_INVARIANTTSC:
> + return (sysctl_rdint(oldp, oldlenp, newp,
> + amd64_has_invariant_tsc));
> default:
> return (EOPNOTSUPP);
> }
> Index: sys/arch/amd64/include/cpu.h
> ===================================================================
> RCS file: /cvs/src/sys/arch/amd64/include/cpu.h,v
> retrieving revision 1.114
> diff -u -p -u -r1.114 cpu.h
> --- sys/arch/amd64/include/cpu.h 11 Aug 2017 20:19:14 -0000 1.114
> +++ sys/arch/amd64/include/cpu.h 23 Aug 2017 11:43:28 -0000
> @@ -427,7 +427,9 @@ void mp_setperf_init(void);
> #define CPU_XCRYPT 12 /* supports VIA xcrypt in userland */
> #define CPU_LIDACTION 14 /* action caused by lid close */
> #define CPU_FORCEUKBD 15 /* Force ukbd(4) as console
> keyboard */
> -#define CPU_MAXID 16 /* number of valid machdep ids */
> +#define CPU_TSCFREQ 16 /* tsc frequency */
> +#define CPU_INVARIANTTSC 17 /* has invariant tsc */
> +#define CPU_MAXID 18 /* number of valid machdep ids */
>
> #define CTL_MACHDEP_NAMES { \
> { 0, 0 }, \
> @@ -446,6 +448,8 @@ void mp_setperf_init(void);
> { 0, 0 }, \
> { "lidaction", CTLTYPE_INT }, \
> { "forceukbd", CTLTYPE_INT }, \
> + { "tscfreq", CTLTYPE_QUAD }, \
> + { "invarianttsc", CTLTYPE_INT }, \
> }
>
> /*
> Index: sys/arch/amd64/include/cpuvar.h
> ===================================================================
> RCS file: /cvs/src/sys/arch/amd64/include/cpuvar.h,v
> retrieving revision 1.8
> diff -u -p -u -r1.8 cpuvar.h
> --- sys/arch/amd64/include/cpuvar.h 28 Jul 2016 21:57:57 -0000 1.8
> +++ sys/arch/amd64/include/cpuvar.h 23 Aug 2017 11:43:28 -0000
> @@ -96,5 +96,6 @@ void x86_ipi_init(int);
> void identifycpu(struct cpu_info *);
> void cpu_init(struct cpu_info *);
> void cpu_init_first(void);
> +void cpu_adjust_tsc_frequency(uint64_t (*)());
>
> #endif
> Index: sys/dev/acpi/acpihpet.c
> ===================================================================
> RCS file: /cvs/src/sys/dev/acpi/acpihpet.c,v
> retrieving revision 1.21
> diff -u -p -u -r1.21 acpihpet.c
> --- sys/dev/acpi/acpihpet.c 6 Oct 2015 20:49:32 -0000 1.21
> +++ sys/dev/acpi/acpihpet.c 23 Aug 2017 11:43:30 -0000
> @@ -265,6 +265,7 @@ acpihpet_attach(struct device *parent, s
> hpet_timecounter.tc_name = sc->sc_dev.dv_xname;
> tc_init(&hpet_timecounter);
> acpihpet_attached++;
> + acpihpet_recalibrate_timecounter_init(&hpet_timecounter);
> }
>
> u_int
> Index: sys/dev/acpi/acpitimer.c
> ===================================================================
> RCS file: /cvs/src/sys/dev/acpi/acpitimer.c,v
> retrieving revision 1.11
> diff -u -p -u -r1.11 acpitimer.c
> --- sys/dev/acpi/acpitimer.c 14 Mar 2015 03:38:46 -0000 1.11
> +++ sys/dev/acpi/acpitimer.c 23 Aug 2017 11:43:30 -0000
> @@ -96,6 +96,7 @@ acpitimerattach(struct device *parent, s
> acpi_timecounter.tc_priv = sc;
> acpi_timecounter.tc_name = sc->sc_dev.dv_xname;
> tc_init(&acpi_timecounter);
> + acpitimer_recalibrate_timecounter_init(&acpi_timecounter);
> }
>
>
> Index: sys/dev/acpi/acpivar.h
> ===================================================================
> RCS file: /cvs/src/sys/dev/acpi/acpivar.h,v
> retrieving revision 1.88
> diff -u -p -u -r1.88 acpivar.h
> --- sys/dev/acpi/acpivar.h 17 Aug 2017 05:16:27 -0000 1.88
> +++ sys/dev/acpi/acpivar.h 23 Aug 2017 11:43:30 -0000
> @@ -25,6 +25,7 @@
>
> #include <sys/timeout.h>
> #include <sys/rwlock.h>
> +#include <sys/timetc.h>
> #include <machine/biosvar.h>
>
> #include "acpipwrres.h"
> @@ -324,6 +325,9 @@ void acpi_resume_cpu(struct acpi_softc
> void acpi_resume_mp(void);
> void acpi_sleep_walk(struct acpi_softc *, int);
>
> +void acpihpet_recalibrate_timecounter_init(struct timecounter *);
> +void acpitimer_recalibrate_timecounter_init(struct timecounter *);
> +uint64_t acpi_calibrate_tsc_frequency();
>
> #define ACPI_IOREAD 0
> #define ACPI_IOWRITE 1
