linux-kernel - Re: [PATCH v3 3/3] module: Make __tracepoints

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: <20190411200610.GB75190@google.com>
Date:   Thu, 11 Apr 2019 16:06:10 -0400
From:   Joel Fernandes <joel@...lfernandes.org>
To:     Steven Rostedt <rostedt@...dmis.org>
Cc:     linux-kernel@...r.kernel.org, paulmck@...ux.vnet.ibm.com,
        keescook@...omium.org, mathieu.desnoyers@...icios.com,
        Jessica Yu <jeyu@...nel.org>,
        kernel-hardening@...ts.openwall.com, kernel-team@...roid.com,
        rcu@...r.kernel.org
Subject: Re: [PATCH v3 3/3] module: Make __tracepoints_ptrs as read-only

On Thu, Apr 11, 2019 at 09:19:55AM -0400, Steven Rostedt wrote:
> On Thu, 11 Apr 2019 04:21:06 -0400
> Joel Fernandes <joel@...lfernandes.org> wrote:
> 
> > Patch 2/3 and 3/3 would not be nececessary if this works out. 1/3 may be a
> > nice clean up but is not something urgent and we could do that in the future
> > if needed.
> 
> Well, jump_labels is "special" because it requires sorting the RO data
> and is done via module notify. The only other user that had to modify
> RO data on module load is ftrace. It had to do the nop conversions in
> the text area. It use to do it via module notify, but because of the
> hardening of the kernel, doing it there was no longer possible because
> everything was RO then. The solution was to call into ftrace directly
> from the module code instead of a notifier. This was done before
> sections were made RO.
> 
> One option is to do the same with jump_labels and just have a call to
> the sorting before the notifiers and before the section gets turned
> into RO. Or I would say just leave it as is. As I stated, jump_labels
> are "special" and adding a loop of one section where I don't envision
> any other sections needing to do the same thing for a long time to
> come. I would save that patch for if there is another section that
> comes along that needs to be modify at module notify.

Sounds good, thanks for the detailed history of this. It sounds like you are
Ok with making of the srcu pointer array as const, which I will test shortly
and send it out for your review. And we can drop this patch series for now.

thanks,

 - Joel