Re: Re%3A [PATCH v4 1%2F2] hw%2Fwatchdog%3A Implement SBSA watchdog device&In-Reply-To=<CAD8XO3YmTC7fzxttg%3DQLMpJbg7uLogatxN7q7vf-iGZjyQLjjQ%40mail.gmail.com>

[Maxim Uvarov]

> +static uint64_t sbsa_gwdt_rread(void *opaque, hwaddr addr, unsigned int size)
> +{
> +    uint32_t ret;
> +
> +    if (addr == SBSA_GWDT_WRR) {
> +        /* watch refresh read has no effect and returns 0 */
> +        ret = 0;
> +    } else {
> +        qemu_log_mask(LOG_GUEST_ERROR, "bad address in refresh frame read :"
> +

Then it has to be  uint32_t ret = 0;

On Fri, 9 Oct 2020 at 20:51, Peter Maydell <peter.maydell@linaro.org> wrote:
>
> On Fri, 9 Oct 2020 at 18:30, Shashi Mallela <shashi.mallela@linaro.org> wrote:
> >
> > The value being returned here is 0 (initialized to 0 at the beginning of 
> > read function).
> > I have seen similar practices being followed in other qemu implementations 
> > like for example bcm2835_dma_read() in 
> > qemu/hw/dma/bcm2835_dma.c,a9_scu_read() in qemu/hw/misc/a9scu.c.
> >
> > Please confirm if you would still like to add specific value like 
> > 0xdeadbeef for bad read offset.
>
> Judging by the subject line you've mangled the headers on this
> email so I'm not sure exactly what this was a reply to,
> but in general QEMU doesn't do that kind of "return
> specific marker values for bad register offsets". We
> typically log it with a qemu_log_mask(LOG_GUEST_ERROR) and
> return 0 (unless the h/w does something else).
>
> thanks
> -- PMM