linux-kernel - Re: [PATCH 2/5] of: dynamic: Refactor action prints to not use "%pOF" inside devtree

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: <CAL_JsqLUBPaeZx8k4BoEnojauVLzs1dWJS+K0kR9U3aNF+EbPw@mail.gmail.com>
Date:   Thu, 3 Aug 2023 11:36:57 -0600
From:   Rob Herring <robh@...nel.org>
To:     Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
Cc:     Frank Rowand <frowand.list@...il.com>,
        "Enrico Weigelt, metux IT consult" <info@...ux.net>,
        "Rafael J. Wysocki" <rafael.j.wysocki@...el.com>,
        Sakari Ailus <sakari.ailus@...ux.intel.com>,
        Petr Mladek <pmladek@...e.com>,
        Geert Uytterhoeven <geert+renesas@...der.be>,
        devicetree@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 2/5] of: dynamic: Refactor action prints to not use "%pOF"
 inside devtree_lock

On Wed, Aug 2, 2023 at 3:33 PM Rob Herring <robh@...nel.org> wrote:
>
> On Tue, Aug 1, 2023 at 9:35 PM Andy Shevchenko
> <andriy.shevchenko@...ux.intel.com> wrote:
> >
> > On Wed, Aug 02, 2023 at 06:28:48AM +0300, Andy Shevchenko wrote:
> > > On Tue, Aug 01, 2023 at 03:54:45PM -0600, Rob Herring wrote:
> > > > While originally it was fine to format strings using "%pOF" while
> > > > holding devtree_lock, this now causes a deadlock.  Lockdep reports:
> > > >
> > > >     of_get_parent from of_fwnode_get_parent+0x18/0x24
> > > >     ^^^^^^^^^^^^^
> > > >     of_fwnode_get_parent from fwnode_count_parents+0xc/0x28
> > > >     fwnode_count_parents from fwnode_full_name_string+0x18/0xac
> > > >     fwnode_full_name_string from device_node_string+0x1a0/0x404
> > > >     device_node_string from pointer+0x3c0/0x534
> > > >     pointer from vsnprintf+0x248/0x36c
> > > >     vsnprintf from vprintk_store+0x130/0x3b4
> > > >
> > > > To fix this, move the printing in __of_changeset_entry_apply() outside the
> > > > lock. As there's already similar printing of the same changeset actions,
> > > > refactor all of them to use a common action print function. This has the
> > > > side benefit of getting rid of some ifdefs.
> >
> > ...
> >
> > > > v3:
> > > >  - Add missing 'static' reported by 0-day
> > >
> > > It reported two issues (at least what I see).
>
> Indeed. I missed the 2nd one.
>
> > ...
> >
> > > > +   if (pr_debug("notify "))
> > >
> > > This is weird. How did you compile it?
>
> I agree it's a weird pattern...
>
> > Urgh, you need to fix dynamic debug macros to return an error code.
>
> Or adding a 'pr_debug_cont' macro would do it. I'm inclined to wrap it
> in an "#ifdef DEBUG" and be done with it.

Here's what I've come up with instead:

#define _do_print(func, prefix, action, node, prop, ...) ({ \
  if (prop) \
    func(prefix "%-15s %pOF:%s\n", ##__VA_ARGS__,
action_names[action], node, prop); \
  else \
    func(prefix "%-15s %pOF\n", ##__VA_ARGS__, action_names[action], node);\
})
#define of_changeset_action_err(...) _do_print(pr_err, __VA_ARGS__)
#define of_changeset_action_debug(...) _do_print(pr_debug, __VA_ARGS__)

Rob