Re: [PATCH] Loongson2: add a primary perf support (not applicable)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Thu, 2010-04-29 at 11:29 +0800, Deng-Cheng Zhu wrote:
> > And then you need to compile the user-space tools/perf with the following steps
> > if want to local-compile it:
> 
> And if you want to cross compile it, besides changing
> tools/perf/perf.h, tools/perf/Makefile also needs to have proper
> CFLAGS/LDFLAGS for headers and libs. Then it's OK to fire "make
> CROSS_COMPILE=$toolchain_prefix".
> 
> > Currently, seems "./perf record" and lots of software events not work, anybody
> > have interest in playing with it can refer to {tools/perf/Documentation,
> > arch/mips/kernel/perf_event*, arch/mips/include/asm/pmu.h,
> > arch/x86/kernel/cpu/perf_event*, arch/arm/kernel/perf_event* ...}.
> 
> "perf record" works fine on 24K/34K/74K cores. In addition, If you are
> seeing the message "Couldn't record kernel reference relocation
> symbol", and your kernel symbols only have _stext (without _text),
> then search "_text" in builtin-record.c and replace with "_stext".
> Here is the link: http://lkml.org/lkml/2010/1/18/177

Yeah, I have seen the message but have not looked into it, thanks!

> 
> For software events, it should be able to work without specific
> changes for loongson, I suppose. Because changes have been done to
> common MIPS code to support software events.
> 

Perhaps need to enable some related kernel options for the "kmem:..." is
not in the result of "./perf list", I will check it later.

> >  config HW_PERF_EVENTS
> >        bool "Enable hardware performance counter support for perf events"
> > -       depends on PERF_EVENTS && !MIPS_MT_SMTC && OPROFILE=n && CPU_MIPS32
> > +       depends on PERF_EVENTS && !MIPS_MT_SMTC && OPROFILE=n
> >        default y
> >        help
> >          Enable hardware performance counter support for perf events. If
> 
> How about adding CPU_LOONGSON2* instead of deleting CPU_MIPS32?
> Because we want the perf functionality to be available when we are
> able to choose it..

Okay, will apply it in the next revision.

BTW: After comparing this patch and your
arch/mips/kernel/perf_event_mipsxx.c, perhaps we can share more common
functions, such as hw_perf_event_destroy(), hw_perf_enable(),
hw_perf_disable() and handle_associated_event()...

Thanks & Regards,
	Wu Zhangjin



[Index of Archives]     [Linux MIPS Home]     [LKML Archive]     [Linux ARM Kernel]     [Linux ARM]     [Linux]     [Git]     [Yosemite News]     [Linux SCSI]     [Linux Hams]

  Powered by Linux