linux-kernel - Re: linux-next: build failure after merge of the reset tree

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] [day] [month] [year] [list]

Message-ID: <CACMJSeuE9-t+wK3WY96oO9zJA0vv=vrtW+VFKtuKTrxXeT=mKg@mail.gmail.com>
Date: Fri, 21 Nov 2025 10:15:29 +0100
From: Bartosz Golaszewski <bartosz.golaszewski@...aro.org>
To: Philipp Zabel <p.zabel@...gutronix.de>
Cc: Stephen Rothwell <sfr@...b.auug.org.au>, 
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>, 
	Linux Next Mailing List <linux-next@...r.kernel.org>
Subject: Re: linux-next: build failure after merge of the reset tree

On Fri, 21 Nov 2025 at 09:22, Philipp Zabel <p.zabel@...gutronix.de> wrote:
>
> > >
> > > Caused by commit
> > >
> > >   d7cdbbc93c56 ("software node: allow referencing firmware nodes")
> > >
> > > I have used the reset tree from next-20251120 for today.
> >
> > This same commit is also in the gpio-brgl tree, so I have used that
> > tree from next-20251120 as well.
>
> Commit d7cdbbc93c56 ("software node: allow referencing firmware nodes")
> renames the 'node' field in software_node_ref_args to 'swnode', so the
> trivial build fix would be:
>
> ----------8<----------
> diff --git a/drivers/platform/x86/intel/chtwc_int33fe.c b/drivers/platform/x86/intel/chtwc_int33fe.c
> index 29e8b5432f4c..96400dec0baf 100644
> --- a/drivers/platform/x86/intel/chtwc_int33fe.c
> +++ b/drivers/platform/x86/intel/chtwc_int33fe.c
> @@ -77,7 +77,7 @@ static const struct software_node max17047_node = {
>   * software node.
>   */
>  static struct software_node_ref_args fusb302_mux_refs[] = {
> -       { .node = NULL },
> +       { .swnode = NULL },
>  };
>
>  static const struct property_entry fusb302_properties[] = {
> @@ -190,9 +190,9 @@ static void cht_int33fe_remove_nodes(struct cht_int33fe_data *data)
>  {
>         software_node_unregister_node_group(node_group);
>
> -       if (fusb302_mux_refs[0].node) {
> -               fwnode_handle_put(software_node_fwnode(fusb302_mux_refs[0].node));
> -               fusb302_mux_refs[0].node = NULL;
> +       if (fusb302_mux_refs[0].swnode) {
> +               fwnode_handle_put(software_node_fwnode(fusb302_mux_refs[0].swnode));
> +               fusb302_mux_refs[0].swnode = NULL;
>         }
>
>         if (data->dp) {
> @@ -222,7 +222,7 @@ static int cht_int33fe_add_nodes(struct cht_int33fe_data *data)
>          * rely on software_node_register_node_group() to use the original
>          * instance of properties instead of copying them.
>          */
> -       fusb302_mux_refs[0].node = mux_ref_node;
> +       fusb302_mux_refs[0].swnode = mux_ref_node;
>
>         ret = software_node_register_node_group(node_group);
>         if (ret)
>
> ---------->8----------
>
> I assume it was expected that drivers don't modify contents of struct
> software_node_ref_args directly, but rather assign them with
> SOFTWARE_NODE_REFERENCE(), so maybe this is not the correct fix?
>

Yeah, driver definitely should not be doing that.

The comment in this file:

/*
 * We are not using inline property here because those are constant,
 * and we need to adjust this one at runtime to point to real
 * software node.
 */

Indicates something's wrong. It looks like it's waiting for a software
node called "intel-xhci-usb-sw" to appear in the system (as in: get
registered as a firmware node). This is clearly an abuse of the whole
concept. I'll try to figure out a way to improve it.

Bart