Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <e41a58eb-c950-acff-0880-a693515aa650@bytedance.com>
Date:   Wed, 12 Apr 2023 15:10:59 +0800
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:102.0)
 Gecko/20100101 Thunderbird/102.9.1
Subject: Re: [PATCH 2/2] x86: make __get_wchan() use arch_stack_walk()
Content-Language: en-US
To:     Josh Poimboeuf <jpoimboe@kernel.org>
Cc:     peterz@infradead.org, keescook@chromium.org,
        dave.hansen@linux.intel.com, bp@alien8.de, mingo@redhat.com,
        tglx@linutronix.de, rostedt@goodmis.org, x86@kernel.org,
        linux-kernel@vger.kernel.org
References: <20230330081552.54178-1-zhengqi.arch@bytedance.com>
 <20230330081552.54178-3-zhengqi.arch@bytedance.com>
 <20230408050822.ezdbhc6j7zuvutrm@treble>
 <fffd59ff-eaa8-bba7-03f2-9fc27620bd1c@bytedance.com>
 <20230408221206.mi7nsquoxklgword@treble>
 <4aba3e74-a205-5dd1-69bc-9186c04e733d@bytedance.com>
 <20230412052037.nz75jeoeqibx6otr@treble>
From:   Qi Zheng <zhengqi.arch@bytedance.com>
In-Reply-To: <20230412052037.nz75jeoeqibx6otr@treble>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk


On 2023/4/12 13:20, Josh Poimboeuf wrote:
> On Sun, Apr 09, 2023 at 02:30:23PM +0800, Qi Zheng wrote:
>>
>>
>> On 2023/4/9 06:12, Josh Poimboeuf wrote:
>>> On Sat, Apr 08, 2023 at 01:36:06PM +0800, Qi Zheng wrote:
>>>>
>>>>
>>>> On 2023/4/8 13:08, Josh Poimboeuf wrote:
>>>>> On Thu, Mar 30, 2023 at 04:15:52PM +0800, Qi Zheng wrote:
>>>>>> Make __get_wchan() use arch_stack_walk() directly to
>>>>>> avoid open-coding of unwind logic.
>>>>>>
>>>>>> Signed-off-by: Qi Zheng <zhengqi.arch@bytedance.com>
>>>>>
>>>>> Can we just have a shared version of __get_wchan() for all
>>>>> CONFIG_ARCH_STACKWALK arches?
>>>>
>>>>   From a quick glance, I think it's ok, but we still need to define
>>>> the arch's own get_wchan_cb(). I will try to do it.
>>>
>>> Hm, why would we need to do that?
>>
>> Because I see checks for count++ < 16 exist in __get_wchan() for some
>> arches and some don't. So I'm not sure if this check can be discarded
>> after using arch_stack_walk(). And I see that this check is retained in
>> arm64, see [1] for details.
>>
>> [1]. https://github.com/torvalds/linux/commit/4f62bb7cb165f3e7b0a91279fe9dd5c56daf3457
> 
> That difference seems to have nothing to do with individual arch
> differences.
> 
> The 16-check limit looks like some ancient cargo cult ritual which was
> copy-pasted decades ago, presumably to avoid some kind of infinite stack
> recursion loop in scheduler code, which should never happen.  That
> should definitely be removed.

Got it.

> 
> Another good reason to unify them, to get rid of cruft like that.

OK, will try to make a shared version of __get_wchan() for all
CONFIG_ARCH_STACKWALK arches.

Thanks,
Qi

>