[bug report] phy: realtek: usb: Add driver for the Realtek SoC USB 2.0 PHY

[Dan Carpenter]

Hello Stanley Chang,

The patch 134e6d25f6bd: "phy: realtek: usb: Add driver for the
Realtek SoC USB 2.0 PHY" from Jul 25, 2023 (linux-next), leads to the
following Smatch static checker warning:

drivers/phy/realtek/phy-rtk-usb2.c:901 get_phy_data_by_efuse()
error: 'buf' dereferencing possible ERR_PTR()

drivers/phy/realtek/phy-rtk-usb3.c:460
get_phy_data_by_efuse() error: 'buf' dereferencing possible ERR_PTR()

drivers/phy/realtek/phy-rtk-usb2.c
    878 static int get_phy_data_by_efuse(struct rtk_phy *rtk_phy,
    879                                  struct phy_parameter *phy_parameter, int index)
    880 {
    881         struct phy_cfg *phy_cfg = rtk_phy->phy_cfg;
    882         u8 value = 0;
    883         struct nvmem_cell *cell;
    884         struct soc_device_attribute rtk_soc_groot[] = {
    885                     { .family = "Realtek Groot",},
    886                     { /* empty */ } };
    887 
    888         if (!phy_cfg->check_efuse)
    889                 goto out;
    890 
    891         /* Read efuse for usb dc cal */
    892         cell = nvmem_cell_get(rtk_phy->dev, "usb-dc-cal");
    893         if (IS_ERR(cell)) {
    894                 dev_dbg(rtk_phy->dev, "%s no usb-dc-cal: %ld\n",
    895                         __func__, PTR_ERR(cell));
    896         } else {
    897                 unsigned char *buf;
    898                 size_t buf_size;
    899 
    900                 buf = nvmem_cell_read(cell, &buf_size);

nvmem_cell_read() can fail.

Sometimes, it's okay ot ignore errors, like if your reads from PCI start
failing, then there is nothing we can do as a kernel.  The user needs to
buy new hardware.  We need the serenity to ignore the things we cannot
fix.

But here the allocation can fail.  So that needs to be checked and
handled correctly.

--> 901                 value = buf[0] & phy_cfg->dc_driving_mask;
                                ^^^^^^
warning

    902 
    903                 kfree(buf);
    904                 nvmem_cell_put(cell);
    905         }
    906 

regards,
dan carpenter