MIME-Version: 1.0
References: <Y9+Yv1CQKNP0Bwqq@kroah.com>
In-Reply-To: <Y9+Yv1CQKNP0Bwqq@kroah.com>
From:   Linus Torvalds <torvalds@linux-foundation.org>
Date:   Sun, 5 Feb 2023 12:17:05 -0800
Message-ID: <CAHk-=whSQFHVzgQ+s8cbCSyE_yrd+BFuYktHe2=0h-nLXBi__Q@mail.gmail.com>
Subject: Re: [GIT PULL] TTY/Serial driver fixes for 6.2-rc7
To:     Greg KH <gregkh@linuxfoundation.org>,
        George Kennedy <george.kennedy@oracle.com>
Cc:     Jiri Slaby <jslaby@suse.cz>,
        Stephen Rothwell <sfr@canb.auug.org.au>,
        Andrew Morton <akpm@linux-foundation.org>,
        linux-kernel@vger.kernel.org, linux-serial@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Sun, Feb 5, 2023 at 3:53 AM Greg KH <gregkh@linuxfoundation.org> wrote:
>
> George Kennedy (1):
>       vc_screen: move load of struct vc_data pointer in vcs_read() to avoid UAF

Hmm.

I think the "goto unlock_out" should be a "break". A partial read
should return the partial success, not an error.

That situation didn't exist when the check was outside the loop, but does now.

Now, I don't think this matters for any sane situation, but since I
looked at this patch I thought I'd mention it.

                   Linus