Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <CAHP4M8Ww0-VqCBKX=iLd=zy1AcDoNdzTOqJuaqRxCGZsMhoX9w@mail.gmail.com>
 <CAHP4M8UcZ=ttB8jbN1yOY6YH8SiQ27NhdEKi9SDH1CWG-GY6eg@mail.gmail.com>
 <6b58a3e1-f2ea-cc4c-03b2-06334b559373@gmail.com> <CAHP4M8Vs8a8u98enuHXaBcC7D4fCZzCOtEq06VnvuPUqhqPK=Q@mail.gmail.com>
 <9717b429-597f-7778-c880-94361bcdee7f@gmail.com> <CAHP4M8XtFiAa1kF5A_rPbcui3DP8L6iyfP8GbwgLLzo0Bo+TNQ@mail.gmail.com>
 <65c45951-08ba-26bb-f96b-3d4442b1d4d4@gmail.com> <08c9e717-4367-5316-87cd-90b5ceb13ed9@kernel.org>
 <1305bb43-b4bf-e129-af6e-957d1f30f269@gmail.com> <CAHP4M8XfCJ8btBCf42GEZGWm_4ywhPKyXtxoBbHR4U190=gg_A@mail.gmail.com>
 <YYuLmMxbKLqHD+ZW@kroah.com>
In-Reply-To: <YYuLmMxbKLqHD+ZW@kroah.com>
From:   Ajay Garg <ajaygargnsit@gmail.com>
Date:   Wed, 10 Nov 2021 15:02:52 +0530
Message-ID: <CAHP4M8VjsZqncVKjaLqE0Hb0b5pia7qo6Vz0hHnBUcbJm1fNeg@mail.gmail.com>
Subject: Re: [PATCH] tty: vt: keyboard: do not copy an extra-byte in copy_to_user
To:     Greg KH <gregkh@linuxfoundation.org>
Cc:     Pavel Skripkin <paskripkin@gmail.com>,
        Jiri Slaby <jirislaby@kernel.org>,
        Andy Shevchenko <andy.shevchenko@gmail.com>, kernel@esmil.dk,
        David Laight <David.Laight@aculab.com>,
        "linux-serial@vger.kernel.org" <linux-serial@vger.kernel.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

> >
> > That's right Pavel.
> > Every function must work correctly as it "advertises", instead of
> > relying on "chancy correctness" of the calls leading to the method.
>
> That is not how the kernel works, sorry.  Otherwise every function would
> have to always verify all parameters passed to them, causing slow downs
> and redundant checks everywhere.
>

Hmm, agreed. Every cycle saved in the kernel is performance gained.

That's why, the RFC for strlscpy [1] makes all the more sense, as it
would save cpu cycles by removing the requirement to check the
return-value for overflows/underflows (including the "issue" I am
trying to address in this particular thread, and which actually lead
to the RFC for strlscpy].

P.S. :

I am not an egoistic person, who wants to get into unnecessary fights
just to upheld one's ego.
All I am trying is to suggest improvements, that

* make things faster.
* keeps code to as minimum as possible.
* makes developers' lives as comfortable as possible.


[1]
https://lore.kernel.org/linux-hardening/CAHP4M8WnLA0780yN+bpuuCtir+DLJRxe0atAiLbZO0bTGf6J-Q@mail.gmail.com/T/#m4a3f524eefe283a42430905fa4c0dfc2c37b2819


Thanks and Regards,
Ajay