Hi!
Using 'ethtool -d […]' on an i.MX6UL leads to a kernel crash:
Unhandled fault: external abort on non-linefetch (0x1008) at […]
due to this SoC has less registers in its FEC implementation compared to other i.MX6 variants. Thus, a run-time decision is required to avoid access to non-existing registers.
Ok, but the code is rather creative:
@@ -2275,7 +2300,24 @@ static void fec_enet_get_regs(struct net_device *ndev, u32 *buf = (u32 *)regbuf; u32 i, off; int ret; +#if defined(CONFIG_M523x) || defined(CONFIG_M527x) || defined(CONFIG_M528x) || \
- defined(CONFIG_M520x) || defined(CONFIG_M532x) || defined(CONFIG_ARM) || \
- defined(CONFIG_ARM64) || defined(CONFIG_COMPILE_TEST)
- u32 *reg_list;
- u32 reg_cnt;
- if (!of_machine_is_compatible("fsl,imx6ul")) {
reg_list = fec_enet_register_offset;reg_cnt = ARRAY_SIZE(fec_enet_register_offset);- } else {
reg_list = fec_enet_register_offset_6ul;reg_cnt = ARRAY_SIZE(fec_enet_register_offset_6ul);- }
+#else
- /* coldfire */
- static u32 *reg_list = fec_enet_register_offset;
- static const u32 reg_cnt = ARRAY_SIZE(fec_enet_register_offset);
+#endif
First, COMPILE_TEST now changes behaviour of the code, which is unexpected. I guess coldfire machine with of_machine_is_compatible("fsl,imx6ul") == true is not probable, but still.
Second reg_list is static, which is quite confusing. I guess that was attempt to enable compiler optimalizations, but I don't think compiler needs this much help.
Best regards, Pavel