On Mon, Feb 15, 2016 at 02:59:13PM +0100, Maarten Brock wrote: > > The uartlite driver suffers from missing/duplicate/corrupted character > > data when the interrupt handler runs concurrently with access to the > > device from another cpu. Take the port spinlock to exclude concurrent > > access. > > Though I am very glad this is finally fixed in this driver, I don't think it > is the best solution. I (a.k.a. vanmierlo) already had an extensive discussion > about it here: > https://github.com/Xilinx/linux-xlnx/issues/24 > > I think the spinlock should be taken and released inside the while loop. > > I'll try to find some time to send in a patch for this and also for some other > issues from that thread. I hope this doesn't end up being a case of perfect being the enemy of good. If you or someone else has a better fix ready to commit, I'd be happy to see that replace my patch, but if not, then in the mean time the driver is completely broken and unsafe on SMP and I think it makes sense for my patch to move forward. Rich -- To unsubscribe from this list: send the line "unsubscribe linux-serial" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
