linux-kernel - Re: syscall_get_error() && TS

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 for Android: free password hash cracker in your pocket

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

Message-ID: <20170329185041.GA24806@redhat.com>
Date:   Wed, 29 Mar 2017 20:50:41 +0200
From:   Oleg Nesterov <oleg@...hat.com>
To:     Linus Torvalds <torvalds@...ux-foundation.org>
Cc:     Andrew Morton <akpm@...ux-foundation.org>,
        Andy Lutomirski <luto@...nel.org>,
        Denys Vlasenko <dvlasenk@...hat.com>,
        "H. Peter Anvin" <hpa@...or.com>, Ingo Molnar <mingo@...hat.com>,
        Jan Kratochvil <jan.kratochvil@...hat.com>,
        Pedro Alves <palves@...hat.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        the arch/x86 maintainers <x86@...nel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: syscall_get_error() && TS_ checks

On 03/29, Linus Torvalds wrote:
>
> That said, I'm not sure why you want to change this in the first
> place? I think the current syscall_get_error() - with explicit compat
> handling and all - is fine.

To simplify this logic. To kill TS_I386_REGS_POKED (which doesn't really
work and can't) and to remove the subtle dependency on TS_COMPAT in ret-
with-signal paths.

Again, afaics we only need these compat checks because regs->ax could be
changed by 32-bit debugger without sign-extension. And TS_I386_REGS_POKED
means that if TS_COMPAT is not set, then the debugger should have also
changed regs->orig_ax. This mostly works, but imo too fragile.

Currently we have the same check in get_nr_restart_syscall() but it is
even more broken (see the patch/changelog), so it should go away and in
this case it would be nice to avoid these checks in do_signal() path too.

Oleg.