Em Mon, 27 Oct 2014 13:38:38 -0400 Michael Ira Krufky <mkrufky@xxxxxxxxxxx> escreveu: > On Mon, Oct 27, 2014 at 12:22 PM, Johannes Stezenbach <js@xxxxxxxxxxx> wrote: > > On Mon, Oct 27, 2014 at 01:57:27PM -0200, Mauro Carvalho Chehab wrote: > >> Em Mon, 27 Oct 2014 10:25:48 -0400 > >> Michael Ira Krufky <mkrufky@xxxxxxxxxxx> escreveu: > >> > >> > I like the idea of supporting older firmware revisions if the new one > >> > is not present, but, the established president for this sort of thing > >> > has always been to replace older firmware with newer firmware without > >> > backward compatibility support for older binaries. > >> > >> No, we're actually adding backward support. There are some drivers > >> already with it. See for example xc4000 (changeset da7bfa2c5df). > >> > >> > Although the current driver can work with both old and new firmware > >> > versions, this hasn't been the case in the past, and won't always be > >> > the case with future firmware revisions. > >> > >> Yeah, we did a very crap job breaking backward firmware compat in > >> the past. We're not doing it anymore ;) > >> > >> > Hauppauge has provided links to the new firmware for both the XC5000 > >> > and XC5000C chips along with licensing. Maybe instead, we can just > >> > upstream those into the linux-firmware packages for distribution. > >> > >> Upstreaming to linux-firmware was done already for the previous firmwares. > >> The firmwares at linux-firmware for xc5000 and xc5000c were merged back > >> there for 3.17 a few weeks ago. > >> > >> Feel free to submit them a new version. > >> > >> > I don't think supporting two different firmware versions is a good > >> > idea for the case of the xc5000 driver. > >> > >> Why not? It should work as-is with either version. We can always add > >> some backward compat code if needed. > > > > FWIW, Linus recently addressed the topic wrt wireless firmware: > > http://article.gmane.org/gmane.linux.kernel.wireless.general/126794 > > > > > > HTH, > > Johannes > > OK, I read Linus' email. I am willing to add an additional patch that > will look for the new firmware image and fall back to the older one if > the new one is not present, but I strongly believe that we should only > support both firmware revisions for a finite period of time -- this > can give people (and distros) time to update to the newer firmware, > and will help to eliminate future bug reports and quality issues that > would otherwise have been resolved by moving to the new firmware. > The new firmware image itself is a bug-fix and improves tuning > performance. If users complain of quality issues using the old > firmware, it will not be very likely to gain developer interest, as > only the new firmware is considered to be truly "supported" now. Well, perhaps you could add a printk message warning the user that the driver is not using the latest firmware and performance/quality could be badly affected. > > Is this acceptable? > > -Mike Krufky Regards, Mauro -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
