Re: [PATCH] moxart: fix potential use-after-free on remove path

[Greg Kroah-Hartman]

On Fri, Jan 21, 2022 at 01:41:27PM +0100, Ulf Hansson wrote:
> On Fri, 14 Jan 2022 at 08:59, Greg Kroah-Hartman
> <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
> >
> > It was reported that the mmc host structure could be accessed after it
> > was freed in moxart_remove(), so fix this by saving the base register of
> > the device and using it instead of the pointer dereference.
> >
> > Cc: Ulf Hansson <ulf.hansson@xxxxxxxxxx>
> > Cc: Xiyu Yang <xiyuyang19@xxxxxxxxxxxx>
> > Cc: Xin Xiong <xiongx18@xxxxxxxxxxxx>
> > Cc: Xin Tan <tanxin.ctf@xxxxxxxxx>
> > Cc: Tony Lindgren <tony@xxxxxxxxxxx>
> > Cc: Yang Li <yang.lee@xxxxxxxxxxxxxxxxx>
> > Cc: linux-mmc@xxxxxxxxxxxxxxx
> > Cc: stable <stable@xxxxxxxxxxxxxxx>
> > Reported-by: whitehat002 <hackyzh002@xxxxxxxxx>
> > Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> > ---
> >  drivers/mmc/host/moxart-mmc.c | 9 +++++----
> >  1 file changed, 5 insertions(+), 4 deletions(-)
> >
> > diff --git a/drivers/mmc/host/moxart-mmc.c b/drivers/mmc/host/moxart-mmc.c
> > index 16d1c7a43d33..f5d96940a9b8 100644
> > --- a/drivers/mmc/host/moxart-mmc.c
> > +++ b/drivers/mmc/host/moxart-mmc.c
> > @@ -697,6 +697,7 @@ static int moxart_remove(struct platform_device *pdev)
> >  {
> >         struct mmc_host *mmc = dev_get_drvdata(&pdev->dev);
> >         struct moxart_host *host = mmc_priv(mmc);
> > +       void __iomem *base = host->base;
> >
> >         dev_set_drvdata(&pdev->dev, NULL);
> >
> > @@ -707,10 +708,10 @@ static int moxart_remove(struct platform_device *pdev)
> >         mmc_remove_host(mmc);
> >         mmc_free_host(mmc);
> >
> > -       writel(0, host->base + REG_INTERRUPT_MASK);
> > -       writel(0, host->base + REG_POWER_CONTROL);
> > -       writel(readl(host->base + REG_CLOCK_CONTROL) | CLK_OFF,
> > -              host->base + REG_CLOCK_CONTROL);
> 
> Rather than doing it like this, I think it would be easier to move
> mmc_free_host() below this part. That's usually what mmc host drivers
> do clean up things in ->remove().
> 
> > +       writel(0, base + REG_INTERRUPT_MASK);
> > +       writel(0, base + REG_POWER_CONTROL);
> > +       writel(readl(base + REG_CLOCK_CONTROL) | CLK_OFF,
> > +              base + REG_CLOCK_CONTROL);
> >

Ok, I can do that, I didn't know if it would cause any functionality
changes, so I was trying to preserve the same logic that the driver
currently has.

Do you have this device to test this with?

thanks,

greg k-h