Von: "Mauro Carvalho Chehab" <mchehab+huawei@xxxxxxxxxx> > This patch seems too risky. While I don't know the specifics of this > specific chipset, usually the ROM is on a separate chip that may or > may not be present. It may even have an unusable or problematic > firmware, depending on when the firmware was flashed. Hi Mauro, thanks for your review. Could you help me understand what risk you mean? Before this change _all_ users of this driver had to rely on the ROM firmware, because the driver simply never downloaded any firmware patches into it. With my change, the following scenarios are possible: 1. The user has no si2157 firmware patch file in /lib/firmware. That will probably be the case for the vast majority of users, as this file is not found e.g. in linux-firmware.git. In this case the device will continue to work just as it did before, no regressions, no improvements. 2. The user has a valid si2157 firmware patch file in /lib/firmware, which is downloaded into the si2157. Should the user experience any issues with the updated firmware (which I think is rather unlikely, as I would expect SiLabs to have revoked it otherwise), simply deleting the firmware patch file from /lib/firmware will cure it and revert to scenario 1 above. 3. The user has an invalid si2157 firmware patch file in /lib/firmware, which looks "right" (to the file size check that's already been in the driver), but does not fit the si2157. I found that in this case, the si2157 will just operate with the original ROM firmware, i.e. also yield the same result as scenario 1 above. I have tested all 3 scenarios on my Hauppauge WinTV-QuadHD, and the result was a fully functional tuner in all cases. I haven't managed to produce a scenario where things broke. Could you sketch what risk you still see of things breaking/regressing with my patch? Best Regards, -Robert Schlabbach
