linux-kernel - Re: [RFC] apparent bogosity in unregister_ftrace_function_probe

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, 28 Jan 2018 05:31:56 -0500
From:   Steven Rostedt <rostedt@...dmis.org>
To:     Al Viro <viro@...IV.linux.org.uk>
Cc:     Dmitry Safonov <0x7f454c46@...il.com>,
        open list <linux-kernel@...r.kernel.org>,
        Masami Hiramatsu <mhiramat@...nel.org>
Subject: Re: [RFC] apparent bogosity in
 unregister_ftrace_function_probe_func()

On Sat, 27 Jan 2018 17:07:48 +0000
Al Viro <viro@...IV.linux.org.uk> wrote:

Hi Al,

> On Sat, Jan 27, 2018 at 01:59:56PM +0000, Dmitry Safonov wrote:
> >   
> > > Incidentally, shouldn't filter_parse_regex("*[ab]", 5, &s, &not)
> > > end up with s = "*[ab]"?  We are returning MATCH_GLOB, after all,
> > > so we want the entire pattern there...  I would've assumed that
> > > this is what the code in unregister_ftrace_function_probe_func()
> > > is trying to compensate for, the first oddity predates MATCH_GLOB...  
> > 
> > No, I don't think filter_parse_regex() should return the full regex..
> > ftrace_match() expects search would be processed string, not a glob.
> > So, this unnecessary assignment broke unregistering multiple kprobs
> > with a middle/end pattern..  
> 
> For substring - sure, but what about something like "*a*b" and "a*b"?
> AFAICS, filter_parse_regex() ends up with identical results in both
> cases - MATCH_GLOB and *search = "a*b".  And no way for the caller
> to tell one from another.
> 
> IOW, it's a different bug sometimes obscured by the one in
> unregister_ftrace_function_probe_func().  filter_parse_regex()
> ought to revert to *search = buff; when it decides to return
> MATCH_GLOB.  Or something like
>         for (i = 0; i < len; i++) {
>                 if (buff[i] == '*') {
>                         if (!i) {
>                                 type = MATCH_END_ONLY;
>                         } else if (i == len - 1) {
>                                 if (type == MATCH_END_ONLY)
>                                         type = MATCH_MIDDLE_ONLY;
>                                 else
>                                         type = MATCH_FRONT_ONLY;
>                                 buff[i] = 0;
>                                 break;
>                         } else {        /* pattern continues, use full glob */
>                                 return MATCH_GLOB;
>                         }
>                 } else if (strchr("[?\\", buff[i])) {
>                         return MATCH_GLOB;
>                 }
>         }
>         if (buff[0] == '*')
>                 *search = buff + 1;
> for that matter - i.e. delay that "we want everything past the first character"
> until we are certain it's not a MATCH_GLOB.
> 
> That one was introduced by "ftrace: Support full glob matching" in 2016, AFAICS...

Yep, I totally agree. This code is one of those places that haven't had
the loving it deserved. It was one of our neglected children.

Thanks for taking a look here. I'm a bit embarrassed by this, and
should have audited it more. I'll have to rip into it and see what else
may be incorrect.

-- Steve