On Tue, May 03, 2022 at 11:41:17AM +0200, Daniels Umanovskis wrote: > On 5/3/22 11:21 AM, Richard Leitner - SKIDATA wrote: > > What's the maximum timeout you've observed? > > > > I guess it would be the simpler and "better" approach to just increase > > the timeout in the driver (if it's not too much above the 500µs). > > I saw 800-820 us at most, and my initial fix internally was just to increase > the sleep duration in the driver. But it's an increase of over 50% and I > don't feel it makes sense to change the driver's behavior for thousands of > users with properly working chips, hence the configurable timeout for > out-of-spec batches like the one we had here. I expect more users to run > across such batches in the coming months. > > In an ideal world, we'd just trash these hubs that should have surely failed > factory QA, but with today's component shortage that's an unimaginable > luxury... The only solution that works here is increase the timeout in the driver. Are you going to tweak the dtb based on what batch the chip is from? No, that's not possible. Having worked in a chip company, I can tell you how they would fix it. Better testing? No, they'd change the documentation. Rob
