Commit 28588d31 fixes this pull-request

Why are you resetting poked here ?

Simo,

Thinking here was that if a config update sets kernel_nfsd = no, SIGHUP, and then sets kernel_nfsd = yes, and again SIGHUP, we should track these state changes and rewrite the proc file with 1 again. In the case where the proc file becomes out of sync with poked, it would allow poked to be reset via two SIGHUPs. Perhaps it would be best to check if we had previously written to the file (poked == 1) and then write it to zero if kernel_nfsd becomes no (in addition to the current behavior of writing it to 1 if it does switch back to yes).

Cipherboy, no that reading is incorrect, once set this setting can never be changed kernel side.
It should not cause harm to poke the kernel again, but it is not useful either, it should be ignored by the kernel IIRC. Although it might cause the kernel to try to reconnect immediately I guess.

Commit 938af94b fixes this pull-request

Commit 28588d31 fixes this pull-request