linux-kernel - Re: [PATCH] vc_screen: break from vcs_read() while loop if vcs

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAHk-=wi5h32VBgzYgFy8KoXbcDMa9K_ihDjfxD-iScy7L+M=QQ@mail.gmail.com>
Date:   Mon, 6 Feb 2023 10:12:43 -0800
From:   Linus Torvalds <torvalds@...ux-foundation.org>
To:     George Kennedy <george.kennedy@...cle.com>
Cc:     gregkh@...uxfoundation.org, jslaby@...e.cz, sfr@...b.auug.org.au,
        akpm@...ux-foundation.org, linux-kernel@...r.kernel.org,
        linux-serial@...r.kernel.org
Subject: Re: [PATCH] vc_screen: break from vcs_read() while loop if vcs_vc()
 returns NULL

On Mon, Feb 6, 2023 at 9:34 AM George Kennedy <george.kennedy@...cle.com> wrote:
>
>
> -               ret = -ENXIO;
>                 vc = vcs_vc(inode, &viewed);
> -               if (!vc)
> +               if (!vc) {
> +                       if (read)
> +                               break;
> +                       ret = -ENXIO;
>                         goto unlock_out;
> +               }

That works, but the whole "if (read)" thing is already done after the
loop, so instead of essentially duplicating that logic, I really think
the patch should be just a plain

                vc = vcs_vc(inode, &viewed);
                if (!vc)
-                       goto unlock_out;
+                       break;

and nothing else.

And yes, the pre-existing vcs_size() error handling has that same ugly pattern.

It might be worth cleaning up too, although right now that

                size = vcs_size(vc, attr, uni_mode);
                if (size < 0) {
                        if (read)
                                break;

pattern means that if we 'break' there, 'read' is non-zero, so 'ret'
doesn't matter. Which is also ugly, but works.

I *think* it could all be rewritten to just use 'break' everywhere in
the loop, and make 'ret' handling be saner.

Something like the attached patch, but while I tried to think about
it, I didn't spend a lot of effort on it, and I certainly didn't test
it. So I'm sending this out as a "Hmm. This _looks_ better to me, but
whatever" patch.

               Linus

View attachment "patch.diff" of type "text/x-patch" (864 bytes)