[PATCH v2] riscv: tracing: Improve hardirq tracing message

Yipeng Zou zouyipeng at huawei.com
Fri Sep 16 04:48:57 PDT 2022


在 2022/9/15 22:04, Conor.Dooley at microchip.com 写道:
> On 15/09/2022 14:36, Yipeng Zou wrote:
>> EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
>>
>> Use trace_hardirqs_on_caller to improve irq_tracing message.
>>
>> lockdep log in riscv showing the last {enabled,disabled} at
>> __trace_hardirqs_{on,off} all the time(if called by).
>> But that's not what we want to see, the caller is what we want.
>>
>> Before this commit:
>> [   57.853175] hardirqs last  enabled at (2519): __trace_hardirqs_on+0xc/0x14
>> [   57.853848] hardirqs last disabled at (2520): __trace_hardirqs_off+0xc/0x14
>>
>> After this commit
>> [   53.781428] hardirqs last  enabled at (2595): restore_all+0xe/0x66
>> [   53.782185] hardirqs last disabled at (2596): ret_from_exception+0xa/0x10
>>
>> Fixes: 22e2100b1b07 ("riscv: fix oops caused by irqsoff latency tracer")
>> Signed-off-by: Yipeng Zou <zouyipeng at huawei.com>
>> Acked-by: Guo Ren <guoren at kernel.org>
>> ---
>> v2: add tags
>>
>>    arch/riscv/kernel/trace_irq.c | 4 ++--
>>    include/linux/irqflags.h      | 2 ++
>>    2 files changed, 4 insertions(+), 2 deletions(-)
>>
>> diff --git a/arch/riscv/kernel/trace_irq.c b/arch/riscv/kernel/trace_irq.c
>> index 095ac976d7da..7ca24b26e19f 100644
>> --- a/arch/riscv/kernel/trace_irq.c
>> +++ b/arch/riscv/kernel/trace_irq.c
>> @@ -16,12 +16,12 @@
>>
>>    void __trace_hardirqs_on(void)
>>    {
>> -       trace_hardirqs_on();
>> +       trace_hardirqs_on_caller(CALLER_ADDR0);
> Out of curiosity, why does riscv need to use this when
> other archs don;t?

At the start we use trace_hardirqs_{on,off} directly in entry.S, but it 
casue an oops here.

Because of "The trace_hardirqs_{on,off}() require the caller to setup 
frame pointer properly.",

and then we add these function.

But I notice that guoren at kernel.org is working on convert riscv to use 
the generic entry infrastructure,

at that time i think we can remove__trace_hardirqs_{on,off}.

>>    }
>>    NOKPROBE_SYMBOL(__trace_hardirqs_on);
>>
>>    void __trace_hardirqs_off(void)
>>    {
>> -       trace_hardirqs_off();
>> +       trace_hardirqs_off_caller(CALLER_ADDR0);
>>    }
>>    NOKPROBE_SYMBOL(__trace_hardirqs_off);
>> diff --git a/include/linux/irqflags.h b/include/linux/irqflags.h
>> index 5ec0fa71399e..46774fa85cde 100644
>> --- a/include/linux/irqflags.h
>> +++ b/include/linux/irqflags.h
>> @@ -53,6 +53,8 @@ extern void trace_hardirqs_on_prepare(void);
>>    extern void trace_hardirqs_off_finish(void);
>>    extern void trace_hardirqs_on(void);
>>    extern void trace_hardirqs_off(void);
>> +extern void trace_hardirqs_on_caller(unsigned long caller_addr);
>> +extern void trace_hardirqs_off_caller(unsigned long caller_addr);
>>
>>    # define lockdep_hardirq_context()     (raw_cpu_read(hardirq_context))
>>    # define lockdep_softirq_context(p)    ((p)->softirq_context)
>> --
>> 2.17.1
>>
>>
>> _______________________________________________
>> linux-riscv mailing list
>> linux-riscv at lists.infradead.org
>> http://lists.infradead.org/mailman/listinfo/linux-riscv

-- 
Regards,
Yipeng Zou




More information about the linux-riscv mailing list