rtl8723au
rtl8723au copied to clipboard
lwfinger
BOGUS urb xfer, pipe 3 != type 1
I'm compiling and using linux 3.13.0 from the git repo and I get this every so often:
http://pastebin.com/VcFPd4Yt
"stupid crap" indeed. I have no way to check this out. What type is endpoint 3?
Sorry for the late reply. Apparently it is type 1 (iso) according to lsusb -s 1:4 -v...I'm at a loss.
FULL DISCLOSURE: I'm not a kernel hacker or a developer, I learned about usb endpoints/pipes ten minutes ago from a wiki article. Any tips on how to proceed and be more helpful would be appreciated.
http://pastebin.com/PBH1NvNa
Now we know that the end point is usually of type 1, which is why it only fails intermittently. Either your hardware is triggering a bug in the USB layer, or the hardware is failing intermittently.
As my knowledge of the USB layer is quite limited, let me post an inquiry on the USB mailing list. Perhaps the experts will know something to try. I'll get back to you when I learn anything.
I should say something too. As I said, I'm using a clone of the git linux-stable for my kernel. I was however using Ubuntu a few months ago and I don't recall this issue.
Forgive the close/open. That damn button is next to the comment button, I clicked it thinking it was clear. Damn gui-learned reflexes.
If it's helpful, I was using Ubuntu previously and I don't recall this particular issue. It's now that I'm using my own clone of linux-stable that I've picked this up. I might be missing some _CONFIG that hid this issue, perhaps that was part of the debian/ubuntu build of the kernel.
What kernel was Ubuntu using? The chance of a kernel regression is much higher than the likelihood of a configuration problem.
The latest version I used on ubuntu was 3.12.0. I actually have a dmesg output from the old install. I recalled booting Gentoo (current install) with an earlier version of linux, and I found that I still had 3.10.17 lying around (with its modules in /lib/modules) and tried booting with that a few minutes, same issue.
I'm not sure this is a "gentoo issue" because I am not currently using gentoo's version of the kernel (there is such a thing with a special set of patches, but as I said, I'm just using a clone of the actual git tree now).
Actually, after looking around, I eerily see similar crashes for other gentoo users... Here[1] is a crash that seems the same. You don't have to read the whole thread, the OP basically complains about the wireless not working. I had similar issues and I basically had to set a number of missing CONFIGs that weren't set. If that's relevant, I've linked the thread[2] where I discuss getting it working.
[1] http://forums.gentoo.org/viewtopic-t-959596-start-0.html [2] http://forums.gentoo.org/viewtopic-t-977664-start-0.html
The issue of the wireless device being considered as an ethernet one was fixed in my repo on October 17, 2013. If anyone still has that problem, they need to use my repo, or make the fix in commit e3e9c3a7.
As to the issue of needing to get rid of the Realtek wireless card support configuration, I still do not see how that can have any possible effect. I would need proof.
The crash in reference [1] is indeed the same as you get, but I still do not know why.