Description of problem: iwlagn stops transferring all data after few MB data transfer at 5GHz 11n mode. For example, while viewing some web pages, all data transferring are stopped. Version-Release number of selected component (if applicable): kernel-2.6.32.9-67.fc12 How reproducible: Steps to Reproduce: 1. connect AP with 5GHz 11n mode. 2. transfer some MBytes data. (ex. viewing some web pages) 3. all data transfer are stopped. ping fails sending data with message "ping: sendmsg: No buffer space available" Actual results: All data transferring are stopped. Expected results: All data transferring are continued. Additional info: I cannot find any logs about this in message file. I haven't seen this trouble in other mode. kernel-2.6.31.12-174.2.22.fc12 works fine except 2.4GHz 11n mode.
Sorry, this is about wireless lan driver problem. I forgot to mention about it. I use 5300AGN on board chip.
kernel-2.6.32.9-70.fc12 has same problem in 5GHz 11n mode. But, there are two case. 1. Data transfer was stopped soon. 2. Data transfer is continued for a while.
I'm seeing something similar even in 2.4GHz 11n mode with the 2.6.32 kernels. It happens after a variable amount of time, and there doesn't seem to be much of interest in the logs. In my case, ping just reports the network as unreachable. I've also got a 5300 on-board chip (ThinkPad T400s). In the meantime, I'm using kernel 2.6.31.12-174.2.22.fc12.x86_64, which works just fine in 2.4GHz 11n mode but not in 5GHz (bug 536988).
Hi. Can You try a new kernel 2.6.32.10 from koji ( search here http://koji.fedoraproject.org/koji/packageinfo?packageID=8 ), which contains a N-only network fix, perhaps this fix helps also with that issue. If bug is not fixed, please provide logs when module is loaded with debug=0x47833fff parameter. This will generate a lots of debugging messages. When logging, please assure you have "kern.*" log to messages in /etc/rsyslog.conf like in example bellow: *.info;kern.*;mail.none;authpriv.none;cron.none /var/log/messages If you will modify /etc/rsyslog.conf please do not forgot to restart syslog daemon. Test of rawhide 2.6.34 kernel, will be also very much appreciated, as we would know if the bug is fixed upstream or not.
Created attachment 400732 [details] A message file compressed by bzip2.
Hi! This is a report of some tests. 2.6.32.10-78.fc12.i686.PAE: Both of 2.4GHz and 5GHz 11n modes work fine! But, message file has many logs like this: Mar 17 17:29:39 localhost kernel: iwlagn 0000:06:00.0: free more than tfds_in_queue (0:1) Hole message file are attached above. 2.6.34-0.10.rc1.git0.fc14.i686.PAE: Both of 2.4GHz and 5GHz 11n modes work fine! Thank you very much!
Hello Akira, I'm glad that driver works for you. We need to get rid of that "tdfs_in_queue" messages, however. This is some firmware/driver problem. Could you please provide logs with debug=0x47833fff module option. This will generate lots of debugging messages. Logs from iwlagn module load to first few occurrences of "tfds_in_queue" are fine. Thanks.
Hi! I re-generated message file. The message file attached before uses modprobe file. But this time I typed "rmmod iwlagn" and "modprobe iwlagn debug=0x47833fff". No error/warning messages were displayed. I cannot find so many differences between this and that attached before. So I attach new message file "as is". I used kernel-PAE-2.6.32.10-78.fc12.i686.rpm, but if other kernel is required, please inform me which one is correct. Best Regards, Anezaki
Created attachment 400971 [details] A message file compressed by bzip2. Old message file is generated with /etc/modprobe.d/iwlagn.conf file including a line: debug=0x47833fff This message file is generated by typing these lines: > /sbin/rmmod iwlagn > /sbin/modprobe iwlagn debug=0x47833fff > /etc/rc.d/init.d/NetworkManager start And connect 5GHz 11n AP by GUI operation.
I'm sorry, option should be debug50=0x47833fff.
Hi! I just finished it. I did like this: Boot and logged in. > rmmod iwlagn > modprobe iwlagn debug50=0x47833fff Connect 5GHz 11n AP by GUI operation. Viewing many Amazon pages. Switch 2.4GHz 11n AP by GUI operation. Viewing many Amazon pages. Shutdown. This message file contains many logs. It will be attached next. I hope these information help you. Best Regards, Anezaki
Created attachment 400978 [details] A message file compressed by bzip2. (3rd) A message file generated with modprobe iwlagn debug50=0x4783fff 1st block is 5GHz. 2nd block is 2.4GHz.
Akira, Could you please test below kernel (when it finish to built, it is compiling right now). Thanks. http://koji.fedoraproject.org/koji/taskinfo?taskID=2061297
Hi! I tested the kernel. It works fine and "tfds_in_queue" disappears! I'm sorry that I have i686 environment only and I can't test other kernels. Thank you very much! Best Regards, Akira Anezaki
*** Bug 574885 has been marked as a duplicate of this bug. ***
Hi! I have continued test more some days. It works fine and no "tdfs_in_queue" message. By the way, kernel-PAE-2.6.32.10-83.sg_test.fc12.i686 fails resuming from sleep/hibernate. Do you think that this problem has no relation to iwlagn driver/firmware, don't you? Regards, Anezaki
For what it's worth I'm running 2.6.32.10-83.sg_test.fc12.x86_64 and I have not see it fail to resume from sleep (I don't use hibernate). I did, however, see it fail to come back from S3 (screen blank), but only once.
Hi! I'm testing new kernel-PAE-2.6.32.10-90.fc12.i686. It looks quite good! 11n mode works in both 2.4GHz and 5GHz. No "tdfs_in_queue" appears. Resume is succeeded from both of sleep and hibernate. Stanislaw, thank you !
If still everything looks fine, time to close this bug :)
For what it's worth, I recently updated to 2.6.32.19-163.fc12.x86_64 and I'm seeing this issue again! I turned off my 'n' network configuration in order to get work done.
Hi Derek, This is other bug, please open new report for it and assign it to me. Provide dmesg output, and info about version last known working kernel.