diff options
authorJouni Malinen <j@w1.fi>2003-07-09 17:48:15 (GMT)
committerJouni Malinen <j@w1.fi>2003-07-09 17:48:15 (GMT)
commitd27bebc92fcadaf8f2fdee5532fe98351ad87fb5 (patch)
parent2b5ef382e67670918d7c73ca50ccdec26531426f (diff)
1 files changed, 13 insertions, 10 deletions
diff --git a/todo.lst b/todo.lst
index d9a2c42..d0c3bdb 100644
--- a/todo.lst
+++ b/todo.lst
@@ -58,9 +58,7 @@
- check when port_reset could be omitted after WEP key change (this would
eliminate new authentication and would help at least .1x with Cisco AP)
- make sure that hostscan setting is re-set after card reset
-- reuse prism2_tx code with prism2_tx_80211, fix prism2_tx_80211 to support
- PS buffering (this may be a reason for some problems seen in tests with
- WinXP Supplicant)
+- reuse prism2_tx code with prism2_tx_80211
- new module code should use try_module_get() for crypto modules
- could actually integrate hostap_crypt.o module into hostap.o.. I don't think
it is that likely that only encryption stuff would be used in other drivers
@@ -89,13 +87,6 @@
to wireless media and Linux net stack
- RX: could possibly remove one tasklet and use dev_queue_xmit() directly
since it is now in softirq context
-- investigate WEP+WDS problems:
- Joe Parks: frequent deauth errors; only when both WEP and WDS are being used
- (turn off either one and things are fine); often tons of packet loss
- and "Invalid misc"; sometimes for brief periods everything seems to
- work fine; STA f/w 1.3.6 & 1.5.6; both host and fw encrypt, both
- compliant and non-compliant WDS frame; ARM processor
- Ricardo Galli: (2002-Dec)
- add option for making wlan#wds? interfaces point-to-point
* is '/sbin/ip ad ad peer dev eth0' enough?
# /sbin/ip ad
@@ -120,3 +111,15 @@
27 .. 154;
- #ifdef blocks for removing wlan#sta interface and default to removing it
(it does not work with new f/w versions)
+- problem with sta->users counter:
+ STA=00:90:4b:0d:fd:89
+ users=-1641
+ (this happened when testing iwconfig rate command on both STA and AP;
+ in addition, lot of Auth frames were not ACKed)
+- fix basic_rates when change supported rates (these rates must be from the
+ supported rate set and basic rate set must be non-empty)
+ (this seems to cause STA-Prism2 cards not sending ACKs at least in some
+ cases)
+- test Shared Key Authentication with kernel driver; it did not seem to succeed
+ (STA not found in auth_cb, or if STA was previously authenticated,
+ "authentication failure")