Re: [Qemu-devel] [PATCH v9 07/13] xilinx_spips: Add support for RX discard and RX drain

[francisco iglesias]

On Saturday, 13 January 2018, Peter Maydell <address@hidden>
wrote:

> On 13 January 2018 at 01:04, francisco iglesias
> <address@hidden> wrote:
> > CID 1383841 (#2 of 4): Uninitialized scalar variable (UNINIT)29.
> > uninit_use_in_call: Using uninitialized value (uint32_t)tx_rx[0] when
> > calling
> > ssi_transfer.
> >
> > This is correct, tx_rx is used uninitialized but since we are
> transmitting
> > dummy cycles the transmitted value (tx_rx[0] in this case) is not used
> (by
> > the flashes), this the reason the code looks like that. Would you like
> me to
> > create a patch for quieting coverity here anyway?
>
>   tx_rx[0] = ssi_transfer(s->spi[bus], (uint32_t)tx_rx[0]);
>
> is effectively using the value (since you don't know what the
> thing on the other end of that function is going to do with the
> value you pass it. You really don't want to make this code's
> correctness depend on the called function never looking at
> its argument. If you want to pass a dummy value why not
> just use constant 0 or something rather than an uninitialized
> variable ?
>
> I think quieting coverity is probably a good idea anyway for
> the other false positives (it sounds like it hasn't been able
> to determine that paths in the rest of the loop can only occur
> if the top of the loop went through a path initializing the
> variable. Code like that is also hard for humans to comprehend,
> though...)
>
> thanks
> -- PMM
>

Hi Peter,

I'll create a patch attempting to correct the coverity complains and
comeback!

Bets regards,
Francisco Iglesias