linux-kernel - Re: Simplify load_unaligned_zeropad() (was Re: [GIT PULL] Ceph updates for 5.20-rc1)

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:   Mon, 15 Aug 2022 15:49:44 -0700
From:   Linus Torvalds <torvalds@...ux-foundation.org>
To:     Peter Zijlstra <peterz@...radead.org>
Cc:     Al Viro <viro@...iv.linux.org.uk>,
        Nathan Chancellor <nathan@...nel.org>,
        Nick Desaulniers <ndesaulniers@...gle.com>,
        Jeff Layton <jlayton@...nel.org>,
        Ilya Dryomov <idryomov@...il.com>, ceph-devel@...r.kernel.org,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Matthew Wilcox <willy@...radead.org>,
        clang-built-linux <llvm@...ts.linux.dev>
Subject: Re: Simplify load_unaligned_zeropad() (was Re: [GIT PULL] Ceph
 updates for 5.20-rc1)

On Mon, Aug 15, 2022 at 1:09 PM Peter Zijlstra <peterz@...radead.org> wrote:
>
> I'm not at all suggesting we do this; but it might be
> insn_get_addr_ref() does what is needed.

.. you didn't suggest it at all, but I started doing it anyway.

It's a bit more complicated, and the fixup certainly isn't that
trivial thing any more, but you were right, it's not *that*
complicated, and it does allow us to use arbitrary 'mov' instructions.

And while it now has more added lines than deletions, and the diffstat now says

 3 files changed, 60 insertions(+), 43 deletions(-)

most of the added lines are still that block comment, and some *very*
anal but trivial sanity checks of the instruction decode.

So I could have made it smaller than it used to be by just not doing
any of those verifications, and maybe I went a bit overboard, but I
think this is such a rare case that it's better to be ridiculously
careful than to try to minimize the number of lines.

So it may be a few more lines, but I can argue that it is still at
least conceptually simpler than the conditional asm goto with outputs
code was.

And yeah, it makes some of the code generation places marginally better.

So since I was tricked into writing this patch, and it's even tested
(the second attachment has a truly stupid patch with my test-case), I
think it's worth doing.

Comments? I left your "Acked-by" from the previous version of this
thing, so holler now if you think this got too ugly in the meantime..

               Linus

View attachment "0001-x86-simplify-load_unaligned_zeropad-implementation.patch" of type "text/x-patch" (5564 bytes)

View attachment "patch.diff" of type "text/x-patch" (990 bytes)