linux-kernel - Re: [PATCH] PCI: Fix reference count leak in pci_create

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

Date:   Fri, 29 May 2020 07:45:38 +0200
From:   Markus Elfring <Markus.Elfring@....de>
To:     Qiushi Wu <wu000273@....edu>, linux-pci@...r.kernel.org
Cc:     kernel-janitors@...r.kernel.org, linux-kernel@...r.kernel.org,
        Alex Chiang <achiang@...com>,
        Bjorn Helgaas <bhelgaas@...gle.com>,
        Jesse Barnes <jbarnes@...tuousgeek.org>,
        Kangjie Lu <kjlu@....edu>,
        Kenji Kaneshige <kaneshige.kenji@...fujitsu.com>
Subject: Re: [PATCH] PCI: Fix reference count leak in pci_create_slot()

Please add a prefix to the patch subject.

> kobject_init_and_add() takes reference even when it fails.

I suggest to extend this description another bit.
Which object is affected here?

> If this function returns an error, kobject_put() must be called to
> properly clean up the memory associated with the object.

Such a copy from the function description of this programming interface
can be helpful.

> Thus, when call of kobject_init_and_add() fail,

I propose to avoid the repetition of this condition.

> we should call kobject_put() instead of kfree().

How do you think about a wording variant like the following?

   Replace a call of the function “kfree” by “kobject_put”
   because of using kernel objects in the proper way.

> Previous commit "b8eb718348b8" fixed a similar problem.

I wonder if such information is really relevant for the commit message.

Regards,
Markus