On 18 Dec, Don Lewis wrote: > Something that looks odd to me is that it appears that > aic_terminate_recovery_thread() can be called twice in at least one > case. In ahc_free(), ahc_terminate_recovery_thread() is unconditionaly > called, and then depending on the init level, ahc_shutdown() may be > called. When the recovery thread is killed, > aic->platform_data->recovery_thread does not get cleared, so if the > first call to ahc_terminate_recovery_thread() kills the recovery thread, > the second call to ahc_terminate_recovery_thread() (from ahc_shutdown()) > will hang because the recovery thread will not be around to wake it up > the second time. > > It looks like either ahc_terminate_recovery_thread() or > aic_recovery_thread() should clear aic->platform_data->recovery_thread > when the thread exits. I was just comparing the ahc code with the ahd code and noticed that ahc_shutdown() calls ahc_terminate_recovery_thread(), but ahd_shutdown() does not call ahd_terminate_recovery_thread() ...Received on Thu Dec 18 2003 - 23:06:32 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:34 UTC