Bug 432267
| Summary: | rt2500pci kernel panics when interface activated | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Vaclav "sHINOBI" Misek <misek> | ||||||||
| Component: | kernel | Assignee: | John W. Linville <linville> | ||||||||
| Status: | CLOSED RAWHIDE | QA Contact: | Fedora Extras Quality Assurance <extras-qa> | ||||||||
| Severity: | high | Docs Contact: | |||||||||
| Priority: | low | ||||||||||
| Version: | rawhide | CC: | cebbert, davej, ivdoorn, madko, vedran | ||||||||
| Target Milestone: | --- | Keywords: | Reopened | ||||||||
| Target Release: | --- | ||||||||||
| Hardware: | x86_64 | ||||||||||
| OS: | Linux | ||||||||||
| Whiteboard: | |||||||||||
| Fixed In Version: | Doc Type: | Bug Fix | |||||||||
| Doc Text: | Story Points: | --- | |||||||||
| Clone Of: | Environment: | ||||||||||
| Last Closed: | 2008-04-03 22:00:23 UTC | Type: | --- | ||||||||
| Regression: | --- | Mount Type: | --- | ||||||||
| Documentation: | --- | CRM: | |||||||||
| Verified Versions: | Category: | --- | |||||||||
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||||||
| Cloudforms Team: | --- | Target Upstream Version: | |||||||||
| Embargoed: | |||||||||||
| Attachments: |
|
||||||||||
|
Description
Vaclav "sHINOBI" Misek
2008-02-10 18:55:58 UTC
Created attachment 294504 [details]
2 listed bugs from /var/log/message
stack trace does not imply rt2x00 to be the cause, but I have a suspicious feeling that this trace was made with rt2x00 2.1.0. What is the result of "modinfo rt2x00lib" Version 2.1.0 is known to be broken and causes memory corruption. 2.1.1 fixes most of those issues but will not bring RX/TX activity back to any of the rt2x00 drivers yet. #modinfo rt2x00lib filename: /lib/modules/2.6.24.1-26.fc9/kernel/drivers/net/wireless/rt2x00/rt2x00lib.ko license: GPL description: rt2x00 library version: 2.0.14 author: http://rt2x00.serialmonkey.com srcversion: 96EDA04F07CF1E49EA178D4 depends: mac80211,input-polldev,crc-itu-t,rfkill,crc-ccitt vermagic: 2.6.24.1-26.fc9 SMP mod_unload Hmm thats odd. Well for rt2x00 2.0.14 this kind of errors are unknown so far. It doesn't look like a complete random crash, since it seems to be consistent about a list error. It is hard to see in which module this bug is occuring, but since rt2x00 doesn't use lists... I'm sorry, but I am not really convinced this is a bug in rt2500pci... Are you sure this bug doesn't appear when rt2500pci is not loaded, or the device is not present? Not completely sure, because the device is build in and I can't remove it. But when the system is started without NetworkManager there is no error. But after ifconfig wlan0 up - bang. There were many patches regarding rt2x00 merged in the latest kernels. - rt2x00: don't write past the end when writing short descriptors on rt61 - rt2x00: Update copyright notice - rt2x00: Add new USB ID to rt2500usb - rt2x00: Fix tx parameter initialization - rt2x00: Enable master and adhoc mode again - rt2x00: Driver requiring firmware should select crc algo - rt2x00: Add per-interface structure - rt2x00: Remove TX_MGMT queue usage - rt2x00: Initialize QID from queue->qid - rt2x00: Move beacon and atim queue defines into rt2x00 - rt2x00: Fix queue index handling - rt2x00: Queue handling overhaul All kernels after these patches doesn't work for me, but it can be certainly related to the general wireless changes. Any idea how to debug it? Ah yes, that list correspond to the rt2x00 2.1.0 patch series, John probably merged all patches for 2.1.0 except the version bump itself. This means that: yes this bug and the other one you reported are caused by rt2x00, and those issues have been resolved in the 2.1.1 release. But unfortunately even in rt2x00 2.1.1 the TX and RX remain broken so it isn't a useful version yet. Kernel panics are gone in the kernel 2.6.25-0.64.rc2.git5.fc9, but the other bug 9(432266) remains. Unfortunately the device is still not working, maybe in another update ;-). Ivo, for reverence fedora kernel version 2.6.25-0.64.rc2.git5.fc9 corresponds to wireless-testing commit e040078e1d5700aaf2a554612db41c9384bc864a. At the moment rt2x00 2.1.3 (released minutes ago) should address the issue of the TX/RX not working (Current score: 3 people reported success, 1 reported failure). The kernel-2.6.25-0.87.rc3.git4.fc9 which should include 2.1.3 is still not working for my hw, but no kernel panic. *** Bug 433201 has been marked as a duplicate of this bug. *** Closing on the basis of "no kernel panic". I believe bug 435591 covers the "still not working" bit... I doubt. All those bugs cover stuff that didn't work even before 2.6.24, while in my case it stopped working with 2.6.24-something, and only in fedora (it works just fine with debian's 2.6.24). Let's hope it will magically get fixed somehow in final 2.6.25... Created attachment 297235 [details]
/var/log/messages calltrace
Hmm I just noticed with the latest rawhide, that the crash is unfortunately
back. kernel-2.6.25-0.95.rc4.fc9
Ok, that panic sounds pretty straightforward. :) I'll look into this. I have a fix for this in the rt2x00 2.1.4 release which was released a few minutes ago. Wonderful! The latest koji kernel (with included 2.1.4) 2.6.25-0.113.rc5.git2.fc9 did the trick. No crash and working wireless. Thanks a lot. Created attachment 299971 [details]
/var/log/messages
Unfortunately the bug is back. The latest working kernel is
kernel-2.6.25-0.155.rc6.git8.fc9.x86_64, the latest tested not working
kernel-2.6.25-0.177.rc7.git6.fc9.x86_64.
A bit of luck, I released a patch for this issue a few days ago. :) John, this bug can be fixed with the patch: [PATCH] rt2x00: Invert scheduled packet_filter check which I send to linux-wireless recently. That patch is available in the kernels here: http://koji.fedoraproject.org/koji/buildinfo?buildID=44780 Do those kernels change the situation for you? Great with the above mentioned kernel there is no crash and the wireless is working. Thanks a lot. |