Created attachment 258827 [details] pstore Hello, I own a Dell XPS 13 9343 with an Intel Wireless 7265 chip. Everything works fine in all Wifis except the one in my university (PEAP authentication). After some time (can be a few minutes or a few hours) no data goes trough the network anymore. Nothing interesting shows up in dmesg and NetworkManager reports that I'm still connected. If I press disconnect I get immediately an OOPS. I attached the data which was stored in pstore. Please tell me if you need more information. Cheers David
I understand that there are two issues here: the lack of traffic and the Oops on disconnection, right? Regarding the Oops, we'd need the actual object file so that we can check exactly where it is failing. Thanks.
Created attachment 258835 [details] iwlmvm.ko
Created attachment 258837 [details] iwlmvm.o
Created attachment 258839 [details] tx.o
Created attachment 258841 [details] tdls.o
iwlmvm.ko is enough :) No need for all the .o :)
(In reply to Emmanuel Grumbach from comment #1) > I understand that there are two issues here: the lack of traffic and the > Oops on disconnection, right? They are related. Disconnection works usually fine on the network. The OOPs only happens when I disconnect after traffic stops. > > Regarding the Oops, we'd need the actual object file so that we can check > exactly where it is failing. I attached iwlmvm.o, tx.o, tdls.o and iwlmvm.ko. Tell me if you need anything else.
Actually, the Oops is in iwlwifi.ko. Please attach that one. Thanks.
Created attachment 258843 [details] iwlwifi.ko
There is no debug info in this object file. Can you please enable frame pointers in the kernel compilation? I can't see much w/o more debug from the compiler. The offset in iwl_trans_pcie_tx is quite big (0x8b9), so I can't even guess where the crash happened.
(In reply to Emmanuel Grumbach from comment #10) > There is no debug info in this object file. Can you please enable frame > pointers in the kernel compilation? CONFIG_FRAME_POINTER=y was already set but CONFIG_DEBUG_INFO=y was missing. I'll try to catch an OOPS with this option
Created attachment 258845 [details] iwlwifi.ko Here's the new iwlwifi.ko with debug symbols. I guess the offset changed because of the recompilation. I try to get a new OOPS
Created attachment 260245 [details] dmesg
Now the traffic stopped as usual but I could disconnect and connect again. I hope the stuff in dmesg is useful. This also happened at my university. At every other wifi everything works fine.
Nope - not the same problem. I do see a SYSASSERT that reminds me another FW bug. Can you please move to -31.ucode?
(In reply to Emmanuel Grumbach from comment #15) > Nope - not the same problem. > > I do see a SYSASSERT that reminds me another FW bug. > > Can you please move to -31.ucode? Where can I get this? The newest I can see is iwlwifi-7265D-29.ucode I checked https://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/linux-firmware.git/tree/ and https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/
iwlwifi-7265D-29.ucode is the latest version for 7265D. The -31 is not supported for this NIC. But you could try the newer one that I published last week from here: https://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/linux-firmware.git/plain/iwlwifi-7265D-29.ucode It includes some fixes compared to the version in the official linux-firmware.git tree. I already sent a pull-request but it will take a while for it to reach it.
Ok, I updated to this version(29.588277.0) and will see if the problem appears again
Created attachment 260293 [details] dmesg Again the traffic stopped but I could disconnect and connect again. Attached is the dmesg output.
This now really looks like https://bugzilla.kernel.org/show_bug.cgi?id=197279.
Yes, now we have the same BAD_COMMAND happening. Marking as duplicate. *** This bug has been marked as a duplicate of bug 197279 ***