[ltt-dev] trace-clock-userspace.patch for ARM/PowerPC ?
Abbas Raza
abbas_raza at mentor.com
Thu Aug 18 03:39:48 EDT 2011
Hello Mathieu,
Sorry for late reply. Yes it can be used under LGPLv2.1/GPLv2 dual-license.
Thanks!
Abbas Raza
On 08/16/2011 03:55 PM, Mathieu Desnoyers wrote:
> Just to make sure this email did not get lost, I'm still waiting on the
> OK to use this code under LGPLv2.1/GPLv2 dual-license.
>
> Thanks,
>
> Mathieu
>
> * Abbas Raza (abbas_raza at mentor.com) wrote:
>> On 07/28/2011 10:26 PM, Mathieu Desnoyers wrote:
>>> * Abbas Raza (abbas_raza at mentor.com) wrote:
>>>> On 07/26/2011 11:02 PM, Mathieu Desnoyers wrote:
>>>>> * Abbas Raza (abbas_raza at mentor.com) wrote:
>>>>>> On 07/26/2011 10:17 PM, Mathieu Desnoyers wrote:
>>>>>>> * Abbas Raza (abbas_raza at mentor.com) wrote:
>>>>>>> [...]
>>>>>>>> Hi Mathieu,
>>>>>>>>
>>>>>>>> Some confusion here :(
>>>>>>>>
>>>>>>>> In case of LTTng 0.x, when we take kernel traces then timestamps for
>>>>>>>> these traces is based on TSC/TB . So after that we view these traces by
>>>>>>>> lttv and it shows timestamps with all events which occurred.
>>>>>>>> I print the timestamp in kernel ltt function 'ltt_trace_alloc()' and it
>>>>>>>> comes out to be 22944596907 (by calling trace_clock_read64() ). While
>>>>>>>> lttv show Birth sec of first event to 367 second. How this conversion is
>>>>>>>> made from TSC to the Birth sec/nsec value displayed in lttv? just want
>>>>>>>> to confirm that whether timestamps shown in lttv (Birth sec/nsec) are
>>>>>>>> based on TSC values... if yes then how TSC values are converted to Birth
>>>>>>>> sec/nsec and i no then from where lttv gets timestamps to be displayed?
>>>>>>>> Or in simple words, do kernel traces which we get contain timestamps
>>>>>>>> based simply on TSC values or it is something else?
>>>>>>> Yes, we write, in the lttng 0.x headers, the following information:
>>>>>>>
>>>>>>> struct ltt_subbuffer_header {
>>>>>>> [...]
>>>>>>> uint64_t start_freq; /*
>>>>>>> * Frequency at trace start,
>>>>>>> * used all along the trace.
>>>>>>> */
>>>>>>> uint32_t freq_scale; /* Frequency scaling (divisor) */
>>>>>>> [...]
>>>>>>> }
>>>>>>>
>>>>>>> so by using the start frequency and the scale, we can convert from TSC
>>>>>>> values to nanoseconds. Please note that LTTng 0.x requires the clock
>>>>>>> source to appear as if it has a constant rate.
>>>>>>>
>>>>>>> These values map to trace_clock_frequency() and trace_clock_freq_scale()
>>>>>>> trace clock functions.
>>>>>>>
>>>>>>> Best regards,
>>>>>>>
>>>>>>> Mathieu
>>>>>>>
>>>>>>>> Thanks a lot again for help :)
>>>>>>>>
>>>>>>>>
>>>>>>>> Abbas Raza
>>>>>>>>
>>>>>> Just to confirm that TSC's just have raw values of clock cycles and lttv
>>>>>> converts these values into sec/nsec by reading ltt_subbuffer_header ?
>>>>> Yes, that's it.
>>>>>
>>>>> Thanks,
>>>>>
>>>>> Mathieu
>>>>>
>>>>>> Thanks!
>>>>>>
>>>>>> Abbas Raza
>>>>>>
>>>> Hello Mathieu,
>>>>
>>>> Just wanted to ask that trace clock userspace patch for x86 adds
>>>> functionality to register posix clocks like
>>>>
>>>> register_posix_clock(CLOCK_TRACE,&clock_trace);
>>>> register_posix_clock(CLOCK_TRACE_FREQ,&clock_trace_freq);
>>>>
>>>> What we understood is that these clocks are registered to provide system
>>>> call functionality for CLOCK_TRACE and CLOCK_TRACE_FREQ clocks just like
>>>> other posix clocks. Right?
>>>>
>>>> And same we will have to do for powerpc (even if we have got vdso
>>>> functionality for these clocks)?
>>> Yes, exactly. The kernel always needs to provide a system call fallback
>>> for vDSO, just in case the vDSO cannot be used.
>>>
>>> Thanks
>>>
>>> Mathieu
>>>
>>>> Thanks !
>>>>
>>>> --
>>>> Abbas Raza
>>>>
>> Hello Mathieu,
>>
>> We have added userspace trace clock support for powerpc architecture. We
>> have verified our changes and userspace, kernel traces appear to be in
>> sync for powerpc now. Patches are attached for kernel and UST.
>>
>> Thanks!
>>
>> Abbas Raza
>>
>> From 018e85ce1a15b9c02719de29663824e1e4f8bc2b Mon Sep 17 00:00:00 2001
>> From: Abbas Raza<Abbas_Raza at mentor.com>
>> Date: Thu, 4 Aug 2011 16:16:28 +0500
>> Subject: [PATCH] Make use of TRACE_CLOCK_FREQ functionality provided by clock_gettime for powerpc.
>>
>> Signed-off-by: Abbas Raza<Abbas_Raza at mentor.com>
>> ---
>> include/ust/clock.h | 8 ++++----
>> 1 files changed, 4 insertions(+), 4 deletions(-)
>>
>> diff --git a/include/ust/clock.h b/include/ust/clock.h
>> index 5e8f755..13e49ab 100644
>> --- a/include/ust/clock.h
>> +++ b/include/ust/clock.h
>> @@ -41,7 +41,7 @@
>> precision and monotonicity.
>> */
>>
>> -/* Only available for x86 arch */
>> +/* Only available for x86/powerpc arch */
>> #define CLOCK_TRACE_FREQ 14
>> #define CLOCK_TRACE 15
>> union lttng_timespec {
>> @@ -76,7 +76,7 @@ static __inline__ uint64_t trace_clock_read64(void)
>> return retval;
>> }
>>
>> -#if __i386__ || __x86_64__
>> +#if __i386__ || __x86_64__ || __powerpc__
>> static __inline__ uint64_t trace_clock_frequency(void)
>> {
>> struct timespec ts;
>> @@ -88,12 +88,12 @@ static __inline__ uint64_t trace_clock_frequency(void)
>> }
>> return 1000000000LL;
>> }
>> -#else /* #if __i386__ || __x86_64__ */
>> +#else /* #if __i386__ || __x86_64__ || __powerpc__ */
>> static __inline__ uint64_t trace_clock_frequency(void)
>> {
>> return 1000000000LL;
>> }
>> -#endif /* #else #if __i386__ || __x86_64__ */
>> +#endif /* #else #if __i386__ || __x86_64__ || __powerpc__ */
>>
>> static __inline__ uint32_t trace_clock_freq_scale(void)
>> {
>> --
>> 1.7.0.4
>>
>> From ac11b1d5235216be99d25f95da30c8b1e2dbb8ee Mon Sep 17 00:00:00 2001
>> From: Abbas Raza<Abbas_Raza at mentor.com>
>> Date: Thu, 4 Aug 2011 16:05:49 +0500
>> Subject: [PATCH] TRACE_CLOCK and TRACE_CLOCK_FREQ in clock_gettime for powerpc
>>
>> These new options to clock_gettime allows the user to retreive the TB frequency and the current TB from userspace.
>>
>> Signed-off-by: Abbas Raza<Abbas_Raza at mentor.com>
>> ---
>> arch/powerpc/include/asm/trace-clock.h | 8 +++
>> arch/powerpc/kernel/Makefile | 2 +-
>> arch/powerpc/kernel/asm-offsets.c | 2 +
>> arch/powerpc/kernel/trace-clock.c | 56 ++++++++++++++++++++++++
>> arch/powerpc/kernel/vdso32/gettimeofday.S | 67 ++++++++++++++++++++++++++---
>> 5 files changed, 127 insertions(+), 8 deletions(-)
>> create mode 100644 arch/powerpc/kernel/trace-clock.c
>>
>> diff --git a/arch/powerpc/include/asm/trace-clock.h b/arch/powerpc/include/asm/trace-clock.h
>> index 05facc3..9ed1108 100644
>> --- a/arch/powerpc/include/asm/trace-clock.h
>> +++ b/arch/powerpc/include/asm/trace-clock.h
>> @@ -13,6 +13,14 @@
>> #include<linux/time.h>
>> #include<asm/time.h>
>>
>> +#define TRACE_CLOCK_MIN_PROBE_DURATION 200
>> +#define TRACE_CLOCK_RES TRACE_CLOCK_MIN_PROBE_DURATION
>> +
>> +union lttng_timespec {
>> + struct timespec ts;
>> + u64 lttng_ts;
>> +};
>> +
>> static inline u32 trace_clock_read32(void)
>> {
>> return get_tbl();
>> diff --git a/arch/powerpc/kernel/Makefile b/arch/powerpc/kernel/Makefile
>> index 3bb2a3e..788a96d 100644
>> --- a/arch/powerpc/kernel/Makefile
>> +++ b/arch/powerpc/kernel/Makefile
>> @@ -120,7 +120,7 @@ obj-$(CONFIG_8XX_MINIMAL_FPEMU) += softemu8xx.o
>> ifneq ($(CONFIG_PPC_INDIRECT_IO),y)
>> obj-y += iomap.o
>> endif
>> -
>> +obj-$(CONFIG_HAVE_TRACE_CLOCK) += trace-clock.o
>> obj-$(CONFIG_PPC64) += $(obj64-y)
>> obj-$(CONFIG_PPC32) += $(obj32-y)
>>
>> diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c
>> index 23e6a93..efb7ead 100644
>> --- a/arch/powerpc/kernel/asm-offsets.c
>> +++ b/arch/powerpc/kernel/asm-offsets.c
>> @@ -370,6 +370,8 @@ int main(void)
>> /* Other bits used by the vdso */
>> DEFINE(CLOCK_REALTIME, CLOCK_REALTIME);
>> DEFINE(CLOCK_MONOTONIC, CLOCK_MONOTONIC);
>> + DEFINE(CLOCK_TRACE, CLOCK_TRACE);
>> + DEFINE(CLOCK_TRACE_FREQ, CLOCK_TRACE_FREQ);
>> DEFINE(NSEC_PER_SEC, NSEC_PER_SEC);
>> DEFINE(CLOCK_REALTIME_RES, MONOTONIC_RES_NSEC);
>>
>> diff --git a/arch/powerpc/kernel/trace-clock.c b/arch/powerpc/kernel/trace-clock.c
>> new file mode 100644
>> index 0000000..5592416
>> --- /dev/null
>> +++ b/arch/powerpc/kernel/trace-clock.c
>> @@ -0,0 +1,56 @@
>> +/*
>> + * arch/powerpc/kernel/trace-clock.c
>> + *
>> + * Trace clock for powerpc.
>> + *
>> + * Abbas Raza<Abbas_Raza at mentor.com>, July 2011
>> + */
>> +
>> +#include<linux/module.h>
>> +#include<linux/trace-clock.h>
>> +#include<linux/timer.h>
>> +#include<linux/posix-timers.h>
>> +
>> +static int posix_get_trace(clockid_t which_clock, struct timespec *tp)
>> +{
>> + union lttng_timespec *lts = (union lttng_timespec *) tp;
>> +
>> + lts->lttng_ts = trace_clock_read64();
>> + return 0;
>> +}
>> +
>> +static int posix_get_trace_freq(clockid_t which_clock, struct timespec *tp)
>> +{
>> + union lttng_timespec *lts = (union lttng_timespec *) tp;
>> +
>> + lts->lttng_ts = trace_clock_frequency();
>> + return 0;
>> +}
>> +
>> +static int posix_get_trace_res(const clockid_t which_clock, struct timespec *tp)
>> +{
>> + union lttng_timespec *lts = (union lttng_timespec *) tp;
>> +
>> + lts->lttng_ts = TRACE_CLOCK_RES;
>> + return 0;
>> +}
>> +
>> +static __init int init_trace_clock(void)
>> +{
>> +
>> + struct k_clock clock_trace = {
>> + .clock_getres = posix_get_trace_res,
>> + .clock_get = posix_get_trace,
>> + };
>> + struct k_clock clock_trace_freq = {
>> + .clock_getres = posix_get_trace_res,
>> + .clock_get = posix_get_trace_freq,
>> + };
>> +
>> + register_posix_clock(CLOCK_TRACE,&clock_trace);
>> + register_posix_clock(CLOCK_TRACE_FREQ,&clock_trace_freq);
>> +
>> + return 0;
>> +}
>> +
>> +early_initcall(init_trace_clock);
>> diff --git a/arch/powerpc/kernel/vdso32/gettimeofday.S b/arch/powerpc/kernel/vdso32/gettimeofday.S
>> index 4ee09ee..dce1344 100644
>> --- a/arch/powerpc/kernel/vdso32/gettimeofday.S
>> +++ b/arch/powerpc/kernel/vdso32/gettimeofday.S
>> @@ -75,7 +75,7 @@ V_FUNCTION_BEGIN(__kernel_clock_gettime)
>> cmpli cr0,r3,CLOCK_REALTIME
>> cmpli cr1,r3,CLOCK_MONOTONIC
>> cror cr0*4+eq,cr0*4+eq,cr1*4+eq
>> - bne cr0,99f
>> + bne cr0,80f
>>
>> mflr r12 /* r12 saves lr */
>> .cfi_register lr,r12
>> @@ -85,7 +85,7 @@ V_FUNCTION_BEGIN(__kernel_clock_gettime)
>> lis r7,NSEC_PER_SEC at h /* want nanoseconds */
>> ori r7,r7,NSEC_PER_SEC at l
>> 50: bl __do_get_tspec at local /* get sec/nsec from tb& kernel */
>> - bne cr1,80f /* not monotonic -> all done */
>> + bne cr1,99f /* not monotonic -> all done */
>>
>> /*
>> * CLOCK_MONOTONIC
>> @@ -122,25 +122,53 @@ V_FUNCTION_BEGIN(__kernel_clock_gettime)
>> blt 1f
>> subf r4,r7,r4
>> addi r3,r3,1
>> -1: bge cr1,80f
>> +1: bge cr1,99f
>> addi r3,r3,-1
>> add r4,r4,r7
>> + b 99f
>> +
>> +80:
>> + cmpli cr0,r3,CLOCK_TRACE
>> + bne cr0,85f
>> +
>> + mflr r12 /* r12 saves lr */
>> + .cfi_register lr,r12
>> + mr r11,r4 /* r11 saves tp */
>> + bl __get_datapage at local /* get data page */
>> + mr r9,r3 /* datapage ptr in r9 */
>> +
>> + bl __do_get_tb at local /* get sec/nsec from tb& kernel */
>> + b 99f
>> +
>> +85:
>> + cmpli cr0,r3,CLOCK_TRACE_FREQ
>> + bne cr0,100f
>>
>> -80: stw r3,TSPC32_TV_SEC(r11)
>> + mflr r5
>> + .cfi_register lr,r5
>> +
>> + mr r11,r4
>> + bl __kernel_get_tbfreq at local
>> + mr r12,r5
>> +
>> +99:
>> + stw r3,TSPC32_TV_SEC(r11)
>> stw r4,TSPC32_TV_NSEC(r11)
>>
>> mtlr r12
>> crclr cr0*4+so
>> li r3,0
>> +
>> blr
>>
>> /*
>> * syscall fallback
>> */
>> -99:
>> - li r0,__NR_clock_gettime
>> +100:
>> +
>> + li r0,__NR_clock_gettime
>> sc
>> - blr
>> + blr
>> .cfi_endproc
>> V_FUNCTION_END(__kernel_clock_gettime)
>>
>> @@ -264,3 +292,28 @@ __do_get_tspec:
>>
>> blr
>> .cfi_endproc
>> +
>> +__do_get_tb:
>> + .cfi_startproc
>> + /* Check for update count& load values. We use the low
>> + * order 32 bits of the update count
>> + */
>> +1: lwz r8,(CFG_TB_UPDATE_COUNT+LOPART)(r9)
>> + andi. r0,r8,1 /* pending update ? loop */
>> + bne- 1b
>> + xor r0,r8,r8 /* create dependency */
>> + add r9,r9,r0
>> +
>> + /* Load orig stamp (offset to TB) */
>> + lwz r5,CFG_TB_ORIG_STAMP(r9)
>> + lwz r6,(CFG_TB_ORIG_STAMP+4)(r9)
>> +
>> + /* Get a stable TB value */
>> +2: mftbu r3
>> + mftbl r4
>> + mftbu r0
>> + cmpl cr0,r3,r0
>> + bne- 2b
>> +
>> + blr
>> + .cfi_endproc
>> --
>> 1.7.0.4
>>
>
More information about the lttng-dev
mailing list