On Sat, Mar 29, 2025 at 02:02:10AM -0400, Valdis Klētnieks wrote: > On Sat, 29 Mar 2025 10:06:39 +0800, Ming Lei said: > > On Fri, Mar 28, 2025 at 02:10:53PM -0400, Valdis Kl??tnieks wrote: > > > Saw this during boot on a Dell Inspiron 5559 laptop. > > > > > > In addition, the external USB ports all gave up, rendering a USB mouse and a > > > USB external drive totally dead in the water. May or may not be related, I didn't > > > dig too far into it. > > > > It shouldn't be related to the warning. > > > For this lockdep warning, feel free to try patch in the following link: > > > > https://lore.kernel.org/linux-block/Z-dUCLvf06SfTOHy@fedora/ > > After applying that patch, USB *didn't* die during boot. So apparently > *something* changed. That is surprising, and maybe the USB die isn't 100% thing. > > Also, the patch merely caused a *different* lockdep warning. > Rather than &q->q_usage_counter(io) and &q->elevator_lock, the > new one is &q->elevator_lock versus pcpu_alloc_mutex... > > Looks like it's a bit more convoluted than first looked? That is another story wrt. freeze lock, fs_reclaim & percpu allocator lock, looks one real risk, I will try to work one patch and see if all can be addressed. Thanks, Ming
