linux-kernel - Re: [PATCH v2] device property: fix of node refcount leak in fwnode_graph_get_next

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]

Date:   Tue, 22 Nov 2022 16:22:35 +0200
From:   Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
To:     Yang Yingliang <yangyingliang@...wei.com>
Cc:     linux-acpi@...r.kernel.org, linux-kernel@...r.kernel.org,
        djrscally@...il.com, heikki.krogerus@...ux.intel.com,
        sakari.ailus@...ux.intel.com, gregkh@...uxfoundation.org,
        rafael@...nel.org
Subject: Re: [PATCH v2] device property: fix of node refcount leak in
 fwnode_graph_get_next_endpoint()

On Tue, Nov 22, 2022 at 04:07:10PM +0200, Andy Shevchenko wrote:
> On Tue, Nov 22, 2022 at 09:41:28PM +0800, Yang Yingliang wrote:
> > On 2022/11/22 21:16, Andy Shevchenko wrote:
> > > On Tue, Nov 22, 2022 at 09:12:41PM +0800, Yang Yingliang wrote:
> > > > On 2022/11/22 20:54, Andy Shevchenko wrote:
> > > > > On Tue, Nov 22, 2022 at 08:00:39PM +0800, Yang Yingliang wrote:

...

> > > > > It seems too complicated for the simple fix.
> > > > > 
> > > > > As I said, just drop const qualifier and add fwnode_handle_get() in the 'else'
> > > > > branch. This will allow you to drop if (prev) at the end.
> > > > fwnode is const, fwnode_handle_get doesn't accept this type.
> > > I'm talking about parent.
> > You suggested this:
> > 
> > "Instead you might consider to replace
> > 
> > 	parent = fwnode;
> > 
> > by
> > 
> > 	parent = fwnode_handle_get(fwnode);"
> > 
> > 
> > It has compile warning:
> > drivers/base/property.c: In function ‘fwnode_graph_get_next_endpoint’:
> > drivers/base/property.c:1004:30: warning: passing argument 1 of ‘fwnode_handle_get’ discards ‘const’ qualifier from pointer target type [-Wdiscarded-qualifiers]
> >    parent = fwnode_handle_get(fwnode);
> >                               ^~~~~~
> > drivers/base/property.c:809:63: note: expected ‘struct fwnode_handle *’ but argument is of type ‘const struct fwnode_handle *’
> >  struct fwnode_handle *fwnode_handle_get(struct fwnode_handle *fwnode)
> > 
> > ~~~~~~~~~~~~~~~~~~~~~~^~~~~~
> 
> I see what you mean. Thank you for clarification.
> 
> So, it seems a bit twisted.
> 
> If prev == NULL, can the
> 
>         ep = fwnode_call_ptr_op(parent, graph_get_next_endpoint, NULL);
> 
> return NULL?
> 
> If no, we may move this case directly to the 'else' branch and return from there.

Answering to my own question: unfortunately it's the case when we have no
endpoints for the fwnode, but might have for the secondary one.

Okay, let's proceed with your slightly modified version 2 (label) for now.

-- 
With Best Regards,
Andy Shevchenko