[staging:staging-testing 41/59] drivers/staging/wfx/main.c:47:14-21: ERROR: PTR_ERR applied after initialization to constant on line 42

[Jerome Pouiller]

On Wednesday 9 October 2019 08:51:10 CEST Rong Chen wrote:
> On 10/7/19 4:36 PM, Jerome Pouiller wrote:
> > On Friday 4 October 2019 12:48:32 CEST kbuild test robot wrote:
> > [...]
> >>>> drivers/staging/wfx/main.c:47:14-21: ERROR: PTR_ERR applied after initialization to constant on line 42
> >> vim +47 drivers/staging/wfx/main.c
> >>
> >>      30
> >>      31  struct gpio_desc *wfx_get_gpio(struct device *dev, int override, const char *label)
> >>      32  {
> >>      33          struct gpio_desc *ret;
> >>      34          char label_buf[256];
> >>      35
> >>      36          if (override >= 0) {
> >>      37                  snprintf(label_buf, sizeof(label_buf), "wfx_%s", label);
> >>      38                  ret = ERR_PTR(devm_gpio_request_one(dev, override, GPIOF_OUT_INIT_LOW, label_buf));
> >>      39                  if (!ret)
> >>      40                          ret = gpio_to_desc(override);
> >>      41          } else if (override == -1) {
> >>    > 42                  ret = NULL;
> >>      43          } else {
> >>      44                  ret = devm_gpiod_get(dev, label, GPIOD_OUT_LOW);
> >>      45          }
> >>      46          if (IS_ERR(ret) || !ret) {
> >>    > 47                  if (!ret || PTR_ERR(ret) == -ENOENT)
> >>      48                          dev_warn(dev, "gpio %s is not defined\n", label);
> >>      49                  else
> >>      50                          dev_warn(dev, "error while requesting gpio %s\n", label);
> >>      51                  ret = NULL;
> >>      52          } else {
> >>      53                  dev_dbg(dev, "using gpio %d for %s\n", desc_to_gpio(ret), label);
> >>      54          }
> >>      55          return ret;
> >>      56  }
> >>      57
> > I think that this report is a false positive or I missed something?
> >
> Sorry for the inconvenience, but we confirmed that the error first
> appeared since commit 0096214a59.

Hi Rong,

Err... I continue to not understand the meaning of this warning. If
override != -1 then ret is not constant, isn't?

-- 
Jérôme Pouiller