linux-kernel - Re: [PATCH 4/9] x86/alternative: Implement .retpoline

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: <YW6XqiOkWG/avY29@hirez.programming.kicks-ass.net>
Date:   Tue, 19 Oct 2021 12:02:18 +0200
From:   Peter Zijlstra <peterz@...radead.org>
To:     Alexander Lobakin <alobakin@...me>
Cc:     Borislav Petkov <bp@...en8.de>, x86@...nel.org,
        jpoimboe@...hat.com, andrew.cooper3@...rix.com,
        linux-kernel@...r.kernel.org, alexei.starovoitov@...il.com,
        ndesaulniers@...gle.com
Subject: Re: [PATCH 4/9] x86/alternative: Implement .retpoline_sites support

On Tue, Oct 19, 2021 at 11:40:56AM +0200, Peter Zijlstra wrote:
> On Mon, Oct 18, 2021 at 11:06:35PM +0000, Alexander Lobakin wrote:

> > WARN in patch_retpoline:408: addr pcibios_scan_specific_bus+0x196/0x200, op 0xe8, reg 0xb88ca
> > WARN in patch_retpoline:408: addr xen_pv_teardown_msi_irqs+0x8d/0x120, op 0xe8, reg 0xb88ca
> > WARN in patch_retpoline:408: addr __mptcp_sockopt_sync+0x7e/0x200, op 0xe8, reg 0xb88ca
> > [...]
> > (thousands of them, but op == 0xe8 && reg == 0xb88ca are always the same)
> > 
> > I know this reg calculation is about to be replaced, but anyway ;)
> 
> Well, I was sorta hoping to keep that with something like:
> 
>   https://lkml.kernel.org/r/YWgA+vbWCdGLZhq5@hirez.programming.kicks-ass.net
> 
> Anyway, let me try and reproduce.

Using: make O=defconfig CC=clang-13 -j80 -s

( is actually defconfig+kvm_guest.config,
  and clang-13 from debian/testing )

on https://git.kernel.org/pub/scm/linux/kernel/git/peterz/queue.git/log/?h=objtool/core

It seems to 'just' work:

spectre_v2=retpoline,amd:

gets me two very sad and lonely replacements:

 [] SMP alternatives: retpoline at: 0xffffffff82e764b4 (ffffffff82e764b4) len: 5 to: __x86_indirect_thunk_rbx+0x0/0x20
 [] SMP alternatives: ffffffff82e764b4: orig: e8 07 d3 18 ff
 [] SMP alternatives: ffffffff82e764b4: repl: 0f ae e8 ff d3

 [] SMP alternatives: retpoline at: 0xffffffff82e76f39 (ffffffff82e76f39) len: 5 to: __x86_indirect_thunk_rdi+0x0/0x20
 [] SMP alternatives: ffffffff82e76f39: orig: e8 02 c9 18 ff
 [] SMP alternatives: ffffffff82e76f39: repl: 0f ae e8 ff d7

The rest is R11 and therefore doesn't actaully fit.

For spectre_v2=off everything gets replaced, and that also seems happy.

 [] SMP alternatives: retpoline at: pm_check_save_msr+0x24/0x30 (ffffffff81d2ef24) len: 5 to: __x86_indirect_thunk_r11+0x0/0x20
 [] SMP alternatives: ffffffff82603eb0: [3:5) optimized NOPs: 41 ff d3 66 90
 [] SMP alternatives: ffffffff81d2ef24: orig: e8 97 49 2d 00
 [] SMP alternatives: ffffffff81d2ef24: repl: 41 ff d3 66 90