On Mon, 21 Sep 2009, Sam Leffler wrote: > Warren Block wrote: >> >> ath0: <Atheros 5416> mem 0xf8200000-0xf820ffff irq 19 at device 0.0 on >> pci20 >> ath0: [ITHREAD] >> ath0: AR5418 mac 12.10 RF5133 phy 8.1 >> >> It sort-of works with a Linkys WRT54G b/g access point, but has some >> problems that keep it from working well. >> >> Transfers are usually stalled, then will suddenly jump to 2MB/sec for a >> little while, then stall again with the output of bb hang messages. >> Access point LED shows long pauses. >> >> kernel: ath0: bb hang detected (0x80), reseting >> kernel: ath0: bb hang detected (0x80), reseting >> kernel: ath0: bb hang detected (0x4), reseting >> kernel: ath0: bb hang detected (0x4), reseting > > This problem is a hardware bug in the 5416 that also appears in other > Atheros 11n cards (i've seen it w/ 9160- and 9280-based cards). Atheros > refused to help me resolve this, except to say the problem doesn't > hobble their linux driver. So to fix this problem you'll need to compare > how the linux code works to how the freebsd code works. If you can > collect register access traces from the linux driver then you can > compare them to what happens in freebsd (use the ATH_DEBUG_ALQ option). > Otherwise you can try to do this through code inspection. Beyond my abilities at present, unfortunately. Fortunately, the AR5007EX/AR5BXB63 cards are relatively easy to find, cheap, and work great. Thank you for all the work you've done to make that happen. -Warren Block * Rapid City, South Dakota USAReceived on Tue Sep 22 2009 - 12:30:51 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:55 UTC