Re: [PATCH 3.2 108/221] spi: dw-mid: avoid potential NULL dereference

[Andy Shevchenko]

On Wed, 2015-05-06 at 16:45 +0100, Ben Hutchings wrote:
> On Wed, 2015-05-06 at 11:25 +0100, Luis Henriques wrote:
> > On Tue, May 05, 2015 at 02:16:39AM +0100, Ben Hutchings wrote:
> > > 3.2.69-rc1 review patch.  If anyone has any objections, please let me know.
> > > 
> > > ------------------
> > > 
> > > From: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
> > > 
> > > commit c9dafb27c84412fe4b17c3b94cc4ffeef5df1833 upstream.
> > > 
> > > When DMA descriptor allocation fails we should not try to assign any fields in
> > > the bad descriptor. The patch adds the necessary checks for that.
> > > 
> > > Fixes: 7063c0d942a1 (spi/dw_spi: add DMA support)
> > > Signed-off-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
> > > Signed-off-by: Mark Brown <broonie@xxxxxxxxxx>
> > > [bwh: Backported to 3.2: adjust context]
> > > Signed-off-by: Ben Hutchings <ben@xxxxxxxxxxxxxxx>
> > > ---
> > >  drivers/spi/spi-dw-mid.c | 6 ++++++
> > >  1 file changed, 6 insertions(+)
> > > 
> > > --- a/drivers/spi/spi-dw-mid.c
> > > +++ b/drivers/spi/spi-dw-mid.c
> > > @@ -155,6 +155,9 @@ static int mid_spi_dma_transfer(struct d
> > >  				1,
> > >  				DMA_TO_DEVICE,
> > >  				DMA_PREP_INTERRUPT | DMA_COMPL_SKIP_DEST_UNMAP);
> > > +	if (!txdesc)
> > > +		return NULL;
> > > +
> > 
> > This backport doesn't look right to me as the function returns an int,
> > not a pointer.  OTOH, it seems like this return value is not checked
> > anywhere, which is why I dropped this patch in the 3.16 kernel.
> 
> In mainline the DMA descriptor setup is split out into the
> dw_spi_dma_prepare_tx() and dw_spi_dma_prepare_rx() functions.  If they
> return NULL then mid_spi_dma_transfer() doesn't attempt to submit
> adescriptor to the respective channel, but it still returns 0.

Seems like error handling is not good enough.
I hope someone in the future can fix this (currently it looks like DMA
is used on Intel SoCs only, though I have no time to amend the code).

> 
> So not only is this backport pretty useless, but so is the upstream
> 'fix'.

> 
> Ben.
> 
> > Cheers,
> > --
> > LuÃs
> > 
> > >  	txdesc->callback = dw_spi_dma_done;
> > >  	txdesc->callback_param = dws;
> > >  
> > > @@ -177,6 +180,9 @@ static int mid_spi_dma_transfer(struct d
> > >  				1,
> > >  				DMA_FROM_DEVICE,
> > >  				DMA_PREP_INTERRUPT | DMA_COMPL_SKIP_DEST_UNMAP);
> > > +	if (!rxdesc)
> > > +		return NULL;
> > > +
> > >  	rxdesc->callback = dw_spi_dma_done;
> > >  	rxdesc->callback_param = dws;
> > >  
> > > 
> > > --
> > > To unsubscribe from this list: send the line "unsubscribe stable" in
> > > the body of a message to majordomo@xxxxxxxxxxxxxxx
> > > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 


-- 
Andy Shevchenko <andriy.shevchenko@xxxxxxxxx>
Intel Finland Oy

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/