power: supply_core: Pass pointer to battery info
Hello Linus Walleij,
The patch 25fd330370ac: "power: supply_core: Pass pointer to battery info" from Dec 15, 2021, leads to the following Smatch static checker warning:
drivers/power/supply/bq256xx_charger.c:1529 bq256xx_hw_init() error: potentially dereferencing uninitialized 'bat_info'.
drivers/power/supply/bq256xx_charger.c
1505 static int bq256xx_hw_init(struct bq256xx_device *bq)
1506 {
1507 struct power_supply_battery_info *bat_info;
1508 int wd_reg_val = BQ256XX_WATCHDOG_DIS;
1509 int ret = 0;
1510 int i;
1511
1512 for (i = 0; i < BQ256XX_NUM_WD_VAL; i++) {
1513 if (bq->watchdog_timer == bq256xx_watchdog_time[i]) {
1514 wd_reg_val = i;
1515 break;
1516 }
1517 if (bq->watchdog_timer > bq256xx_watchdog_time[i] &&
1518 bq->watchdog_timer < bq256xx_watchdog_time[i + 1])
1519 wd_reg_val = i;
1520 }
1521 ret = regmap_update_bits(bq->regmap, BQ256XX_CHARGER_CONTROL_1,
1522 BQ256XX_WATCHDOG_MASK, wd_reg_val <<
1523 BQ256XX_WDT_BIT_SHIFT);
1524
1525 ret = power_supply_get_battery_info(bq->charger, &bat_info);If the first allocation in power_supply_get_battery_info() fails then bat_info is uninitialized. (It's not really possible unless you do failure injection).
1526 if (ret) {
1527 dev_warn(bq->dev, "battery info missing, default values will be applied\n");
1528
--> 1529 bat_info->constant_charge_current_max_ua =
1530 bq->chip_info->bq256xx_def_ichg;
1531
1532 bat_info->constant_charge_voltage_max_uv =
1533 bq->chip_info->bq256xx_def_vbatreg;
1534
1535 bat_info->precharge_current_ua =
1536 bq->chip_info->bq256xx_def_iprechg;
1537
1538 bat_info->charge_term_current_ua =
1539 bq->chip_info->bq256xx_def_iterm;
1540
1541 bq->init_data.ichg_max =
1542 bq->chip_info->bq256xx_max_ichg;
1543
1544 bq->init_data.vbatreg_max =
1545 bq->chip_info->bq256xx_max_vbatreg;
1546 } else {
1547 bq->init_data.ichg_max =
1548 bat_info->constant_charge_current_max_ua;
1549
1550 bq->init_data.vbatreg_max =
1551 bat_info->constant_charge_voltage_max_uv;
1552 }
1553
1554 ret = bq->chip_info->bq256xx_set_vindpm(bq, bq->init_data.vindpm);
1555 if (ret)
1556 return ret;
1557
1558 ret = bq->chip_info->bq256xx_set_iindpm(bq, bq->init_data.iindpm);
1559 if (ret)
1560 return ret;
1561
1562 ret = bq->chip_info->bq256xx_set_ichg(bq,
1563 bat_info->constant_charge_current_max_ua);
1564 if (ret)
1565 return ret;
1566
1567 ret = bq->chip_info->bq256xx_set_iprechg(bq,
1568 bat_info->precharge_current_ua);
1569 if (ret)
1570 return ret;
1571
1572 ret = bq->chip_info->bq256xx_set_vbatreg(bq,
1573 bat_info->constant_charge_voltage_max_uv);
1574 if (ret)
1575 return ret;
1576
1577 ret = bq->chip_info->bq256xx_set_iterm(bq,
1578 bat_info->charge_term_current_ua);
1579 if (ret)
1580 return ret;
1581
1582 power_supply_put_battery_info(bq->charger, bat_info);
1583
1584 return 0;
1585 }regards,
dan carpenter
Quoted 6 lines · Dan Carpenter
The patch
25fd330370ac: "power:supply_core: Pass pointer to battery info" from Dec 15, 2021, leads to the following Smatch static checker warning:
drivers/power/supply/bq256xx_charger.c:1529bq256xx_hw_init()
error: potentially dereferencing uninitialized 'bat_info'.
(...)
Quoted 10 lines
1525 ret = power_supply_get_battery_info(bq->charger, &bat_info);If the first allocation in
power_supply_get_battery_info()fails thenbat_infois uninitialized. (It's not really possible unless you do failure injection).1526 if (ret) { 1527 dev_warn(bq->dev, "battery info missing, default values will be applied\n"); 1528--> 1529
bat_info->constant_charge_current_max_ua=
Ouch, that one was easy to miss! I sent a fixup patch.
Yours,
Linus Walleij