Quoting from kernel.spec: > Patch690: linux-2.6-at76.patch [...] > # Add misc wireless bits from upstream wireless tree > ApplyPatch linux-2.6-at76.patch [...] >* Fri Aug 01 2008 John W. Linville <linville> >- Revert at76_usb to version from before attempted mac80211 port This patch is unneeded, as at76 is upstream these days -- it's part of the staging tree/area that was merged into 2.6.28. The sources are apart from two small changes even identical with the ones we use Version-Release number of selected component (if applicable): kernel-2.6.28-1.fc11.x86_64 Suggested solution: Remove patch and use the driver from the staging tree; that way the kernel will also get tainted as soon as the driver gets loaded (that's what upstream does and it's likely the best to follow that example)
Seems reasonable, except that there is virtually no guarantee that the staging version will remain working or that anyone will fix it when it breaks. C'est la vie? Also, I don't know what (if any) policy we have or need w.r.t. enabling staging drivers?
(In reply to comment #1) > Seems reasonable, except that there is virtually no guarantee that the staging > version will remain working or that anyone will fix it when it breaks. C'est > la vie? Well, the same afaics applies to the driver we include as patch right now, as it nearly the same code base ;-) So maybe the real solution to for the at76-usb driver is to drop it completely from Fedora... > Also, I don't know what (if any) policy we have or need w.r.t. enabling staging > drivers? That was one of the reasons why I opened this bug. But the decisions is likely up to Davej, Cebbert and Kylem.
One difference is that the staging driver will eventually go back to the (broken) mac80211 port of the driver, whereas the version in fedora will remain where it is until we have a working alternative. As for dropping it, that might be a fine alternative provided you don't mind hurting those users.
*sigh* Seems http://thread.gmane.org/gmane.linux.kernel.wireless.general/26045/focus=26487 was to late. http://git.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=02227c28391b5059a7710d6039c52912b0ee2c1d > Staging: at76_usb: update drivers/staging/at76_usb w/ mac80211 port > > This updates drivers/staging/at76_usb to correspond with the latest > version previously available in wireless-testing. Which makes this bug mostly irrelevant afaics :-/
Crap, confusion...that commit reflects what I was hinting at in comment 3. Of course, I thought the thread you quoted meant that we all agreed not to commit that -- Greg must have not connected the two points... :-(
at76 change was reverted afaics: http://git.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=89cb7e7fd6c0917bb9236ea48bf538d4668ed009 So we could again do now what was suggested in the initial comment of this bug: drop our private at76_usb copy that we patch in and instead use the one from the staging tree
*** Bug 484850 has been marked as a duplicate of this bug. ***
Change checked-in to rawhide.