Received-SPF: pass (google.com: domain of linux-kernel+bounces-188820-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Precedence: bulk
MIME-Version: 1.0
References: <o89373n4-3oq5-25qr-op7n-55p9657r96o8@vanv.qr> <CAHk-=wjxdtkFMB8BPYpU3JedjAsva3XXuzwxtzKoMwQ2e8zRzw@mail.gmail.com>
 <ZkvO-h7AsWnj4gaZ@slm.duckdns.org> <CALOAHbCYpV1ubO3Z3hjMWCQnSmGd9-KYARY29p9OnZxMhXKs4g@mail.gmail.com>
 <CAHk-=wj9gFa31JiMhwN6aw7gtwpkbAJ76fYvT5wLL_tMfRF77g@mail.gmail.com>
 <CALOAHbAmHTGxTLVuR5N+apSOA29k08hky5KH9zZDY8yg2SAG8Q@mail.gmail.com>
 <CAHk-=wjAmmHUg6vho1KjzQi2=psR30+CogFd4aXrThr2gsiS4g@mail.gmail.com>
 <CALOAHbAAAU9MTQFc56GYoYWR3TsLbkncp5QrrwHMbqJ9SECivw@mail.gmail.com>
 <CAHk-=whwtEFJnDVrkkMtb6SWcmBQMK8+qXGtqvBO+xH8y2i6nA@mail.gmail.com> <CALOAHbD0LdbQTWyvDiLcgGupcQJKmadzWhoZiUTj126Rqqn6fQ@mail.gmail.com>
In-Reply-To: <CALOAHbD0LdbQTWyvDiLcgGupcQJKmadzWhoZiUTj126Rqqn6fQ@mail.gmail.com>
From: Linus Torvalds <torvalds@linux-foundation.org>
Date: Fri, 24 May 2024 07:58:26 -0700
Message-ID: <CAHk-=wivfrF0_zvf+oj6==Sh=-npJooP8chLPEfaFV0oNYTTBA@mail.gmail.com>
Subject: Re: [PATCH workqueue/for-6.10-fixes] workqueue: Refactor worker ID
 formatting and make wq_worker_comm() use full ID string
To: Yafang Shao <laoar.shao@gmail.com>
Cc: bpf <bpf@vger.kernel.org>, Tejun Heo <tj@kernel.org>, Jan Engelhardt <jengelh@inai.de>, 
	Craig Small <csmall@enc.com.au>, linux-kernel@vger.kernel.org, 
	Lai Jiangshan <jiangshanlai@gmail.com>
Content-Type: text/plain; charset="UTF-8"

On Fri, 24 May 2024 at 00:43, Yafang Shao <laoar.shao@gmail.com> wrote:
>
> Actually, there are already helpers for this: get_task_comm() and
> __get_task_comm(). We can simply replace the memcpy() with one of
> these

No. We should get rid of those horrendous helpers.

> If the task_lock() in __get_task_comm() is a concern, we could
> consider adding a new __get_current_comm().

The task_lock is indeed the problem - it generates locking problems
and basically means that most places cannot use them. Certainly not
things like tracing etc.

The locking is also entirely pointless\, since absolutely nobody
cares. If somebody is changing the name at the same time - which
doesn't happen in practice - getting some halfway result is fine as
long as you get a proper NUL terminated result.

Even for non-current, they are largely useless. They were a mistake.

So those functions should never be used for any normal thing. Instead
of locking, the function should literally just do a "copy a couple of
words and make sure the end result still has a NUL at the end".

That's literally what selinuxfs.c wants, for example - it copies the
thing to a local buffer not because it cares about some locking issue,
but because it wants one stable value. But by using 'memcpy()' and
that fixed size, it means that we can't sanely extend the source size
because now it wouldn't be NUL-terminated. But selinux never wanted a
lock, and never wanted any kind of *consistent* result, it just wanted
a *stable* result.

Since user space can randomly change their names anyway, using locking
was always wrong for readers (for writers it probably does make sense
to have some lock - although practically speaking nobody cares there
either, but at least for a writer some kind of race could have
long-term mixed results)

Oh well.

                Linus