https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273289
Mark Johnston <[email protected]> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |[email protected], | |[email protected] Status|New |Open --- Comment #2 from Mark Johnston <[email protected]> --- (In reply to John F. Carr from comment #1) Yes, there doesn't seem to be any particular reason to hold the lock across the free() calls. Though, I cannot see a reason to use a MTX_SPIN lock here at all. It looks like softs->devlist_lock should be a MTX_DEF mutex. Spin mutexes are only needed when synchronizing with interrupt handlers, and it doesn't look like this mutex has to deal with that. Though, fixing that would be more involved and would require some modification to this OS abstraction layer in the driver. -- You are receiving this mail because: You are the assignee for the bug.
