OSDN Git Service

net: wireless: mt76: fix similar warning reported by kbuild test robot
authorPetr Štetiar <ynezz@true.cz>
Fri, 10 May 2019 09:35:18 +0000 (11:35 +0200)
committerDavid S. Miller <davem@davemloft.net>
Fri, 10 May 2019 22:14:29 +0000 (15:14 -0700)
commit1b9705d9713b67996608763e2317673d8378b03c
tree2b2fa498721cf6fbe217e1bd99cf40b63378394f
parent2d2924af96881f3300239156d6ff316a1a3e48e5
net: wireless: mt76: fix similar warning reported by kbuild test robot

This patch fixes following (similar) warning reported by kbuild test robot:

 In function ‘memcpy’,
  inlined from ‘smsc75xx_init_mac_address’ at drivers/net/usb/smsc75xx.c:778:3,
  inlined from ‘smsc75xx_bind’ at drivers/net/usb/smsc75xx.c:1501:2:
  ./include/linux/string.h:355:9: warning: argument 2 null where non-null expected [-Wnonnull]
  return __builtin_memcpy(p, q, size);
         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
  drivers/net/usb/smsc75xx.c: In function ‘smsc75xx_bind’:
  ./include/linux/string.h:355:9: note: in a call to built-in function ‘__builtin_memcpy’

I've replaced the offending memcpy with ether_addr_copy, because I'm
100% sure, that of_get_mac_address can't return NULL as it returns valid
pointer or ERR_PTR encoded value, nothing else.

I'm hesitant to just change IS_ERR into IS_ERR_OR_NULL check, as this
would make the warning disappear also, but it would be confusing to
check for impossible return value just to make a compiler happy.

I'm now changing all occurencies of memcpy to ether_addr_copy after the
of_get_mac_address call, as it's very likely, that we're going to get
similar reports from kbuild test robot in the future.

Fixes: d31a36b5f407 ("net: wireless: support of_get_mac_address new ERR_PTR error")
Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Petr Štetiar <ynezz@true.cz>
Signed-off-by: David S. Miller <davem@davemloft.net>
drivers/net/wireless/mediatek/mt76/eeprom.c