Re: [PATCH v5 1/2] ftrace: disable preemption when recursion locked

From: 王贇 <yun.wang@linux.alibaba.com>
To: Steven Rostedt <rostedt@goodmis.org>
Cc: Miroslav Benes <mbenes@suse.cz>, Guo Ren <guoren@kernel.org>,
	Ingo Molnar <mingo@redhat.com>,
	"James E.J. Bottomley" <James.Bottomley@HansenPartnership.com>,
	Helge Deller <deller@gmx.de>,
	Michael Ellerman <mpe@ellerman.id.au>,
	Benjamin Herrenschmidt <benh@kernel.crashing.org>,
	Paul Mackerras <paulus@samba.org>,
	Paul Walmsley <paul.walmsley@sifive.com>,
	Palmer Dabbelt <palmer@dabbelt.com>,
	Albert Ou <aou@eecs.berkeley.edu>,
	Thomas Gleixner <tglx@linutronix.de>,
	Borislav Petkov <bp@alien8.de>,
	x86@kernel.org, "H. Peter Anvin" <hpa@zytor.com>,
	Josh Poimboeuf <jpoimboe@redhat.com>,
	Jiri Kosina <jikos@kernel.org>, Petr Mladek <pmladek@suse.com>,
	Joe Lawrence <joe.lawrence@redhat.com>,
	Masami Hiramatsu <mhiramat@kernel.org>,
	"Peter Zijlstra (Intel)" <peterz@infradead.org>,
	Nicholas Piggin <npiggin@gmail.com>,
	Jisheng Zhang <jszhang@kernel.org>,
	linux-csky@vger.kernel.org, linux-kernel@vger.kernel.org,
	linux-parisc@vger.kernel.org, linuxppc-dev@lists.ozlabs.org,
	linux-riscv@lists.infradead.org, live-patching@vger.kernel.org
Subject: Re: [PATCH v5 1/2] ftrace: disable preemption when recursion locked
Date: Wed, 27 Oct 2021 09:54:13 +0800	[thread overview]
Message-ID: <3d897161-7b74-944a-f2a0-07311436fbd9@linux.alibaba.com> (raw)
In-Reply-To: <20211026080117.366137a5@gandalf.local.home>

On 2021/10/26 下午8:01, Steven Rostedt wrote:
> On Tue, 26 Oct 2021 17:48:10 +0800
> 王贇 <yun.wang@linux.alibaba.com> wrote:
> 
>>> The two comments should be updated too since Steven removed the "bit == 0" 
>>> trick.  
>>
>> Could you please give more hint on how will it be correct?
>>
>> I get the point that bit will no longer be 0, there are only -1 or > 0 now
>> so trace_test_and_set_recursion() will disable preemption on bit > 0 and
>> trace_clear_recursion() will enabled it since it should only be called when
>> bit > 0 (I remember we could use a WARN_ON here now :-P).
>>
>>>   
>>>> @@ -178,7 +187,7 @@ static __always_inline void trace_clear_recursion(int bit)
>>>>   * tracing recursed in the same context (normal vs interrupt),
>>>>   *
>>>>   * Returns: -1 if a recursion happened.
>>>> - *           >= 0 if no recursion
>>>> + *           > 0 if no recursion.
>>>>   */
>>>>  static __always_inline int ftrace_test_recursion_trylock(unsigned long ip,
>>>>  							 unsigned long parent_ip)  
>>>
>>> And this change would not be correct now.  
>>
>> I thought it will no longer return 0 so I change it to > 0, isn't that correct?
> 
> No it is not. I removed the bit + 1 return value, which means it returns the
> actual bit now. Which is 0 or more.

Ah, the return is bit not val, I must be drunk...

My apologize for the stupid comments... I'll send a v6 for this patch
only to fix that, please let me know if this is not a good way to fix
few lines of comments.

Regards,
Michael Wang

> 
> -- Steve
> 