Ira Weiny wrote: > Dan Williams wrote: > > Ira Weiny wrote: > > [snip] > > > > > > > - guard(rwsem_read)(&cxl_cper_rw_sem); > > > - if (cper_callback) > > > - cper_callback(event_type, rec); > > > > Given a work function can be set atomically there is no need to create / > > manage a registration lock. Set a 'struct work' instance to a CXL > > provided routine on cxl_pci module load and restore it to a nop function > > + cancel_work_sync() on cxl_pci module exit. > > Ok I'll look into this. > > > > > > + wi = kmalloc(sizeof(*wi), GFP_ATOMIC); > > > > The system is already under distress trying to report an error it should > > not dip into emergency memory reserves to report errors. Use a kfifo() > > similar to how memory_failure_queue() avoids memory allocation in the > > error reporting path. > > I have a question on ghes_proc() [ghes_do_proc()]. Can they be called by > 2 threads at the same time? It seems like there could be multiple > platform devices which end up queueing into the single kfifo. Yes, that is already the case for memory_failure_queue() and aer_recover_queue(). > there needs to be a kfifo per device or synchronization with multiple > writers. Yes, follow the other _queue() examples. kfifo_in_spinlocked() looks useful for this purpose. I expect no lock needed on the read side since the reader is only the single workqueue context.
