It wasn't immediately apparent to me how to undo the Ubuntu backports option Luis suggested in his email, so I opted instead to download source code -- compat-wireless-2010-05-18.
I couldn't bring myself to run "make install" because it was apparent that "make uninstall" would not revert all the changes made by installing, so instead I did: rmmod ath5k rmmod ath rmmod mac80211 rmmod cfg80211 followed by insmod ./net/wireless/cfg80211.ko insmod ./net/mac80211/mac80211.ko insmod ./drivers/net/wireless/ath/ath.ko insmod ./drivers/net/wireless/ath/ath5k/ath5k.ko I believe this is sufficient to install the new ath5k driver. Let me know if it's not. After loading the new driver, I downloaded some big binary files with both wget and ftp. With both I was able to reproduce data corruption. In the first case a sequence of 1368 bytes was repeated in the destination file. In the second case I saw a repeated sequence of 1448 bytes. In both cases the incorrect data was the first instance of the repeated block, not the second instance that immediately followed it in the destination file. Whenever there was data corruption I also saw messages like those below, but I have also seen similar messages without data corruption. What other experiments can I do that will help figure out the cause of this problem? bob P.S. I have a couple of suggestions for compat-wireless .... First, my initial attempt to compile drivers failed because scripts/driver-select thought I was not in the right directory. I had to comment out the following code. # if [ ! -f compat-release ]; then # echo "Must run $0 from the compat-wireless top level directory" # exit # fi Perhaps the compat-release file is always present in official releases. Second, I think the Makefile should contain an uninstall-scripts target. Something like: uninstall-scripts: rm -f $(DESTDIR)/usr/lib/compat-wireless/scripts/modlib.sh rm -f $(DESTDIR)/usr/sbin/scripts/madwifi-unload rm -f $(DESTDIR)/usr/sbin/scripts/athenable rm -f $(DESTDIR)/usr/sbin/scripts/b43enable rm -f $(DESTDIR)/usr/sbin/scripts/iwl-enable rm -f $(DESTDIR)/usr/sbin/scripts/athload rm -f $(DESTDIR)/usr/sbin/scripts/b43load rm -f $(DESTDIR)/usr/sbin/scripts/iwl-load rmdir --ignore-fail-on-non-empty $(DESTDIR)/usr/lib/compat-wireless bob ==================== dmesg output during file transfers ==================== May 22 22:55:02 ugg kernel: [115957.336563] ACPI: EC: GPE storm detected, transactions will use polling mode May 22 23:01:26 ugg kernel: [116340.532470] cfg80211: Calling CRDA to update world regulatory domain May 22 23:01:26 ugg kernel: [116340.881306] cfg80211: World regulatory domain updated: May 22 23:01:26 ugg kernel: [116340.881320] (start_freq - end_freq @ bandwidth), (max_antenna_gain, max_eirp) May 22 23:01:26 ugg kernel: [116340.881337] (2402000 KHz - 2472000 KHz @ 40000 KHz), (300 mBi, 2000 mBm) May 22 23:01:26 ugg kernel: [116340.881350] (2457000 KHz - 2482000 KHz @ 20000 KHz), (300 mBi, 2000 mBm) May 22 23:01:26 ugg kernel: [116340.881363] (2474000 KHz - 2494000 KHz @ 20000 KHz), (300 mBi, 2000 mBm) May 22 23:01:26 ugg kernel: [116340.881376] (5170000 KHz - 5250000 KHz @ 40000 KHz), (300 mBi, 2000 mBm) May 22 23:01:26 ugg kernel: [116340.881390] (5735000 KHz - 5835000 KHz @ 40000 KHz), (300 mBi, 2000 mBm) ==================== May 22 16:55:06 ugg rsyslogd: [origin software="rsyslogd" swVersion="4.2.0" x-pid="901" x-info="http://www.rsyslog.com"] rsyslogd was HUPed, type 'lightweight'. May 22 16:56:19 ugg kernel: [94433.533234] cfg80211: Calling CRDA to update world regulatory domain May 22 16:56:20 ugg kernel: [94435.039856] cfg80211: World regulatory domain updated: May 22 16:56:20 ugg kernel: [94435.039864] (start_freq - end_freq @ bandwidth), (max_antenna_gain, max_eirp) May 22 16:56:20 ugg kernel: [94435.039873] (2402000 KHz - 2472000 KHz @ 40000 KHz), (300 mBi, 2000 mBm) May 22 16:56:20 ugg kernel: [94435.039879] (2457000 KHz - 2482000 KHz @ 20000 KHz), (300 mBi, 2000 mBm) May 22 16:56:20 ugg kernel: [94435.039886] (2474000 KHz - 2494000 KHz @ 20000 KHz), (300 mBi, 2000 mBm) May 22 16:56:20 ugg kernel: [94435.039892] (5170000 KHz - 5250000 KHz @ 40000 KHz), (300 mBi, 2000 mBm) May 22 16:56:20 ugg kernel: [94435.039898] (5735000 KHz - 5835000 KHz @ 40000 KHz), (300 mBi, 2000 mBm) ========================= On Mon, May 17, 2010 at 11:02 PM, Luis R. Rodriguez <mcg...@gmail.com> wrote: > On Mon, May 17, 2010 at 7:51 PM, Bruno Randolf <b...@einfach.org> wrote: >> On Tuesday 18 May 2010 11:29:27 Robert Brown wrote: >>> I'm running Ubuntu Linux 10.04 on an Acer Aspire ZG5 netbook that contains >>> Atheros wifi hardware. There's a sticker on the back that says "Atheros >>> AR5BXB63". The uname command reports the kernel version as >>> 2.6.32-22-generic. I'm using the ath5k driver. >> >> could you tell us the chipset version? the output of dmesg, just the line >> which looks similar to: >> "ath5k phy2: Atheros AR5414 chip found (MAC: 0xa5, PHY: 0x61)" >> >>> I often get corrupted data when downloading files using wget and often >>> experience broken ssh connections with errors that imply corrupted packets. >>> I see messages like the following in the system log: >>> >>> [24051.016838] ath5k phy0: unsupported jumbo >> >> this warning is nothing to worry about. it has been removed in newer kernels. >> >> the corrupted packets are something to worry about, however... ;( >> >>> I didn't have the data corruption problems 6 or 8 months ago with the >>> previous "Karmic" Ubuntu release and older kernels. They appeared at some >>> point with a system update. >> >> that's interresting. so we have to figure out which change caused this >> regression. >> >>> I'd like to help figure out what's wrong. What can I do to assist with >>> tracking down this problem? >> >> could you try a more recent version of ath5k? first thing to try is the >> recent >> compat-wireless. i'm not sure what's the ubuntu way of getting it, but it >> should be possible to compile it manually against your kernel. > > For ubuntu there are lbm packages but that is based on the stable > releases of compat-wireless. To avoid getting the jumbo frame messages > he'd have to use bleeding edge stuff: > > http://wireless.kernel.org/en/users/Download/#Getting_compat-wireless_on_Ubuntu > > Luis > _______________________________________________ > ath5k-devel mailing list > ath5k-devel@lists.ath5k.org > https://lists.ath5k.org/mailman/listinfo/ath5k-devel > _______________________________________________ ath5k-devel mailing list ath5k-devel@lists.ath5k.org https://lists.ath5k.org/mailman/listinfo/ath5k-devel