Responding to my query about CONFIG_MODULE_SIG in your .config, you wrote:
Why did I ask?CONFIG_MODULE_SIG=y
# CONFIG_MODULE_SIG_FORCE is not set
CONFIG_MODULE_SIG_ALL=y
CONFIG_MODULE_SIG_SHA1=y
# CONFIG_MODULE_SIG_SHA224 is not set
# CONFIG_MODULE_SIG_SHA256 is not set
# CONFIG_MODULE_SIG_SHA384 is not set
# CONFIG_MODULE_SIG_SHA512 is not set
CONFIG_MODULE_SIG_HASH="sha1"
Back in May 2016 you were helping me with ndiswrapper and xenial. I asked you to look at a quote that provided a possible reason for failure. You didn't answer. Perhaps it is appropriate in this instance too.
About half-way in my post, the source states that if kernel module signature verification is enabled and a module isn't properly signed, then the kernel will refuse to load it. In which case, the developer should disable CONFIG_MODULE_SIG_ALL and CONFIG_MODULE_SIG in .config before compiling.
My guess with your .config showing "CONFIG_MODULE_SIG=y" and "CONFIG_MODULE_SIG_ALL=y" is that it is enabled. I know from our conversations in the thread back then that you did some patching, which would most likely make it "improperly signed" and therefore doomed to fail.
What do you think? And even if you think it's a crock, please humour me and include these changes in your next compile. I'll gladly test and if it fails, I won't bother you anymore. Deal?