linux-kernel - RE: [EXT] [PATCH] net: fec: add missed clk_disable

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: <VI1PR0402MB36008CCD634F6E320EED733EFF7E0@VI1PR0402MB3600.eurprd04.prod.outlook.com>
Date:   Tue, 5 Nov 2019 16:46:30 +0000
From:   Andy Duan <fugang.duan@....com>
To:     Chuhong Yuan <hslester96@...il.com>
CC:     "David S . Miller" <davem@...emloft.net>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: RE: [EXT] [PATCH] net: fec: add missed clk_disable_unprepare in
 remove

From: Chuhong Yuan <hslester96@...il.com> Sent: Tuesday, November 5, 2019 11:34 PM
> On Tue, Nov 5, 2019 at 9:26 AM Andy Duan <fugang.duan@....com> wrote:
> >
> > From: Chuhong Yuan <hslester96@...il.com> Sent: Monday, November 4,
> > 2019 11:50 PM
> > > This driver forgets to disable and unprepare clks when remove.
> > > Add calls to clk_disable_unprepare to fix it.
> > >
> > > Signed-off-by: Chuhong Yuan <hslester96@...il.com>
> >
> > If runtime is enabled, the patch will introduce clock count mis-match.
> > Probe->
> >     Enable clk_ipg, clk_ahb clocks
> >     ...
> >     In the end, runtime auto suspend callback disable clk_ipg, clk_ahb
> clocks.
> >
> > You should check CONFIG_PM is enabled or not in your platform, if not,
> > it can disable these two clocks by checking CONFIG_PM.
> >
> 
> Thanks for your hint!
> But I am still not very clear about the mechanism.
> In my opinion, it means that if CONFIG_PM is disabled, runtime_suspend will
> be called automatically to disable clks.
CONFIG_PM is enabled, runtime pm works. Otherwise, it doesn't work, then it
Requires .remove() to disable the clks.

Andy
> Therefore, #ifdef CONFIG_PM check should be added before disabling clks in
> remove.
> I am not sure whether this understanding is right or not?
> 
> Regards,
> Chuhong
> 
> > Regards,
> > Andy
> > > ---
> > >  drivers/net/ethernet/freescale/fec_main.c | 2 ++
> > >  1 file changed, 2 insertions(+)
> > >
> > > diff --git a/drivers/net/ethernet/freescale/fec_main.c
> > > b/drivers/net/ethernet/freescale/fec_main.c
> > > index 22c01b224baa..a9c386b63581 100644
> > > --- a/drivers/net/ethernet/freescale/fec_main.c
> > > +++ b/drivers/net/ethernet/freescale/fec_main.c
> > > @@ -3645,6 +3645,8 @@ fec_drv_remove(struct platform_device *pdev)
> > >                 regulator_disable(fep->reg_phy);
> > >         pm_runtime_put(&pdev->dev);
> > >         pm_runtime_disable(&pdev->dev);
> > > +       clk_disable_unprepare(fep->clk_ahb);
> > > +       clk_disable_unprepare(fep->clk_ipg);
> > >         if (of_phy_is_fixed_link(np))
> > >                 of_phy_deregister_fixed_link(np);
> > >         of_node_put(fep->phy_node);
> > > --
> > > 2.23.0
> >