On 19/02/2020 13:44, Jan Beulich wrote: > On 19.02.2020 14:22, Roger Pau Monné wrote: >> On Wed, Feb 19, 2020 at 01:59:51PM +0100, Jan Beulich wrote: >>> On 13.02.2020 12:32, Roger Pau Monne wrote: >>>> Don't allow cpu_hotplug_begin to fail by converting the trylock into a >>>> blocking lock acquisition. Write users of the cpu_add_remove_lock are >>>> limited to CPU plug/unplug operations, and cannot deadlock between >>>> themselves or other users taking the lock in read mode as >>>> cpu_add_remove_lock is always locked with interrupts enabled. There >>>> are also no other locks taken during the plug/unplug operations. >>> I don't think the goal was deadlock avoidance, but rather limiting >>> of the time spent spinning while trying to acquire the lock, in >>> favor of having the caller retry. >> Now that the contention between read-only users is reduced as those >> can take the lock in parallel I think it's safe to switch writers to >> blocking mode. > I'd agree if writers couldn't be starved by (many) readers.
XSA-114. We fixed that vulnerability ages ago. A writer wanting the lock will prevent further readers from taking it before the writer drops it. ~Andrew _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel
