linux-kernel - Re: Buggy __free(kfree) usage pattern already in tree

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]

Date:   Fri, 15 Sep 2023 15:10:54 -0700
From:   Linus Torvalds <torvalds@...ux-foundation.org>
To:     Peter Zijlstra <peterz@...radead.org>
Cc:     Bartosz Golaszewski <bartosz.golaszewski@...aro.org>,
        Alexey Dobriyan <adobriyan@...il.com>,
        linux-kernel@...r.kernel.org,
        Linus Walleij <linus.walleij@...aro.org>,
        akpm@...ux-foundation.org
Subject: Re: Buggy __free(kfree) usage pattern already in tree

On Fri, 15 Sept 2023 at 14:50, Linus Torvalds
<torvalds@...ux-foundation.org> wrote:
>
> Call it "cond_guard()", and make the syntax otherwise be the same as
> "scoped_guard()", iow, using a unique ID for the guard name.

I *think* you don't even need a new "cond_guard()" macro.

This might be as simple as having our current scoped_guard() macro
have the conditional

  "scope.lock && !done"

instead of just the current "!done".

Then you introduce "trylock" variants that conditionally set "lock = 1".

I dunno. Maybe there's something really obvious I'm missing, but I
really think that gives us the option to literally do just

        scoped_guard(try_mutex, ..somelock..) {
                ...
        }

and all it requires is for us to declare a "try_mutex" lock guard class.

Ok, so the current DEFINE_LOCK_GUARD_x() macros clearly set .lock to 1
unconditionally, so there's a small matter of extending the lock side
to be conditional...

"SMOP".

                 Linus