Re: [PATCH] spi: spi_mpc8xxx.c: fix potential memory corruption.

[Anton Vorontsov <cbouatmailru-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>]

On Wed, May 12, 2010 at 06:28:51PM +0200, Joakim Tjernlund wrote:
> Anton Vorontsov <cbouatmailru-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote on 2010/05/12 18:22:31:
> >
> > On Wed, May 12, 2010 at 05:50:50PM +0200, Joakim Tjernlund wrote:
> > > tx_dma/rx_dma are already set to a dummy buffer when no
> > > tx/rx buffer and t->tx_dma/t->rx_dma does not contain a dma
> > > address, but NULL.
> > > This may lead to corruption of kernel memory. Fix this by
> > > leaving tx_dma/rx_dma alone.
> > >
> > > Do not INIT_TX_RX while controller is enabled, this is bad according
> > > to the MPC8321 manual.
> > >
> > > Signed-off-by: Joakim Tjernlund <Joakim.Tjernlund-SNLAxHN9vbcOP4wsBPIw7w@public.gmane.org>
> >
> > Acked-by: Anton Vorontsov <cbouatmailru-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
> >
> > Thanks!
> :)
> 
> Who will make sure this goes to Linus?

I guess it's Grant. Cc'ed.

>  Jocke
> 
> BTW, I am having another problem too. An ADC doesn't work when in QE/CPM mode(CPU mode is fine)
> The only thing I can think of ATM is if word size is handled differently? We
> are using 16 bits words.

No idea. Have you tried loopback mode? Create a dummy spidev
device, and try Documentation/spi/spidev_test.c's --loop switch.

Personally I tested QE/CPM mode with M25Pxx flashes. IIRC, I
also tested it with the spi_mmc driver on MPC8323E-RDB board.
Both drivers work with 8 bits words.

-- 
Anton Vorontsov
email: cbouatmailru-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org
irc://irc.freenode.net/bd2

------------------------------------------------------------------------------