On Wed, Nov 21, 2018 at 05:43:43PM +0100, Bartosz Golaszewski wrote:

??r., 21 lis 2018 o 17:39 Nicholas Mc Guire [off-list ref] napisa??(a):

quoted

kasprintf() may return NULL on failure of internal allocation thus the
assigned  label  is not safe if not explicitly checked. On error
mediatek_gpio_bank_probe() returns negative values so -ENOMEM in the
(unlikely) failure case should be fine here.

Signed-off-by: Nicholas Mc Guire <redacted>
Fixes: 4ba9c3afda41 ("gpio: mt7621: Add a driver for MT7621")
---

Problem located with experimental coccinelle script

Patch was compile tested with: omega2p_defconfig, SOC_MT7621=y,
GPIOLIB=y, GPIO_MT7621=y

Patch is against 4.20-rc3 (localversion-next is next-20181121)

 drivers/gpio/gpio-mt7621.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/gpio/gpio-mt7621.c b/drivers/gpio/gpio-mt7621.c
index d72af6f..42f0c67 100644
--- a/drivers/gpio/gpio-mt7621.c
+++ b/drivers/gpio/gpio-mt7621.c

@@ -244,6 +244,10 @@ mediatek_gpio_bank_probe(struct device *dev,
        rg->chip.of_xlate = mediatek_gpio_xlate;
        rg->chip.label = devm_kasprintf(dev, GFP_KERNEL, "%s-bank%d",
                                        dev_name(dev), bank);
+       if (!rg->chip.label) {
+               dev_err(dev, "Could not allocate label\n");

No need for an error message here, an -ENOMEM says enough.

I just thought its more consistent with the rest of
the code - and as it is a highly unlikely event it would
probably be hard to pinpoint as some of the other functions
can return -ENOMEM (but with a message).

....and the call site of mediatek_gpio_bank_probe() actually 
currently ignores the return value so the -ENOMEM would go
unnoticed which is a bug as well - patch for that a bit later.

anyway fine with me - I?ll send a V2 and drop the dev_err() here.

thx!
hofrat
 

Bart

quoted

+               return -ENOMEM;
+       }

        ret = devm_gpiochip_add_data(dev, &rg->chip, mtk);
        if (ret < 0) {
--
2.1.4