LWN: Comments on "Device dependencies and deferred probing" https://lwn.net/Articles/662820/ This is a special feed containing comments posted to the individual LWN article titled "Device dependencies and deferred probing". en-us Wed, 05 Nov 2025 01:32:21 +0000 Wed, 05 Nov 2025 01:32:21 +0000 https://www.rssboard.org/rss-specification lwn@lwn.net Device dependencies and deferred probing https://lwn.net/Articles/663455/ https://lwn.net/Articles/663455/ broonie <div class="FormattedComment"> Yes, exactly.<br> </div> Fri, 06 Nov 2015 12:30:35 +0000 Device dependencies and deferred probing https://lwn.net/Articles/663009/ https://lwn.net/Articles/663009/ tpetazzoni <i>Rafael Wysocki responded that the real issue is dependencies between drivers</i> <p>It is not exactly what Rafael said I believe. What Rafael said is that dependencies are not just between devices, but between devices that are actually *bound* to a driver to manage them. Indeed if a device A needs another device B, it doesn't just need this device B to have its platform_device registered. It also needs device B to be bound to a device driver, of which the -&gt;probe() method has been called to actually do something with device B, so that it can be used by device A.</p> <p>So device A -&gt;probe() should ideally only be triggered when device B -&gt;probe() has completed. Having device B "registered" is not sufficient.</p> Wed, 04 Nov 2015 13:14:35 +0000