On Thu, May 19, 2016 at 06:10:37PM -0400, nick wrote: > Here is the issue though it does not happen on v4.4 but on newer > kernels. I bisected it and it does work if the commit I stated is > reverted, and working at that commit the only line changed in any > function is in my patch. Here we can add a check like your asking:
This is not what I hinted at. I hinted at checking ->op_state in edac_mc_reset_delay_period() before doing edac_mod_work(). Here's another hint: ->op_state gets set to OP_RUNNING_POLL only when ->edac_check is not NULL. Better? -- Regards/Gruss, Boris. ECO tip #101: Trim your mails when you reply.