> > On 4. Jan 2024, at 09:21, Johan Hovold <johan@xxxxxxxxxx> wrote: > > On Thu, Jan 04, 2024 at 08:50:19AM +0100, Sven Peter wrote: >>> On 4. Jan 2024, at 08:47, Aditya Garg <gargaditya08@xxxxxxxx> wrote: >>>> On 28-Dec-2023, at 5:41 PM, Johan Hovold <johan@xxxxxxxxxx> wrote: > >>>> Ok, good, then this patch and the one I posted are mostly equivalent >>>> assuming that the BCM4378/4387 return an invalid address during setup. >>>> >>>> This patch may be preferred as it does not need to rely on such >>>> assumptions, though. > >>> So what's the final take on this? Which one is gonna be merged upstream? >> >> I would’ve preferred this one (possibly with a better commit message) >> since it’s more explicit and doesn’t rely on additional assumptions >> but it looks like Johan’s version was already merged. > > Which addresses do BCM4378/4387 return before they are configured? > Should be easy enough to verify that the current check for invalid > addresses catches those or otherwise add them to the list. > > Johan I think the check used to work for BRCM4378 when I originally wrote the driver but I don’t have any BRCM4387 hardware so can’t test that myself. Sven
