linux-kernel - Re: [GIT PULL] integrity subsystem updates for v5.8

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:   Sun, 7 Jun 2020 12:59:01 -0700
From:   Linus Torvalds <torvalds@...ux-foundation.org>
To:     Mimi Zohar <zohar@...ux.ibm.com>
Cc:     linux-integrity <linux-integrity@...r.kernel.org>,
        linux-kernel <linux-kernel@...r.kernel.org>
Subject: Re: [GIT PULL] integrity subsystem updates for v5.8

On Fri, Jun 5, 2020 at 10:03 AM Mimi Zohar <zohar@...ux.ibm.com> wrote:
>
> The main changes are extending the TPM 2.0 PCR banks with bank
> specific file hashes, calculating the "boot_aggregate" based on other
> TPM PCR banks, using the default IMA hash algorithm, instead of SHA1,
> as the basis for the cache hash table key, and preventing the mprotect
> syscall to circumvent an IMA mmap appraise policy rule.

I'm not sure why I didn't notice this during my test builds, but this
results in a new warning:

  WARNING: modpost: vmlinux.o(.text+0x496264): Section mismatch in
reference from the function ima_calc_boot_aggregate() to the function
.init.text:ima_pcrread()
  The function ima_calc_boot_aggregate() references
  the function __init ima_pcrread().
  This is often because ima_calc_boot_aggregate lacks a __init
  annotation or the annotation of ima_pcrread is wrong.

and I think the warning is real - and the problem is real, not just a
missing annotation.

It looks like ima_calc_boot_aggregate() may be called not only at init
time (ima_eventdigest_init() is referenced from "struct
ima_template_field supported_fields[]", for example)

So calling an __init function would be very wrong, because it might be
gone by that time.

Hmm?

                Linus