(Didn't receive the email in my mailbox so copying Samuel's reply from speakup mail archives web page) > Thinking again about it, you can not just put the kref on the tty, you > need to know whether the work happened (thus already put the kref) or > not (thus hasn't put the kref). > So instead of just setting tty to NULL, you should xchg it with NULL, > and if it wasn't NULL (the work didn't happen), put the kref. Yes that's good point, thanks. There is also another, abeit unlikely, race where we put kref inside speakup_cancel_selection() before speakup_set_selection() has had a chance to get kref but after compxchg succeeded. I'm not sure of the best way to address that. One way would be to use a global bool to indicate whether put kref is needed. But that will need to be atomic with respect to tty_kref_get() in speakup_set_selection() and to tty_kref_put() in __speakup_set_selection(), which means addition of another lock. I wonder if there is a better solution. Thanks, Okash _______________________________________________ Speakup mailing list Speakup@xxxxxxxxxxxxxxxxx http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
