Re: wireless-testing, p54 and sinus 154 data no longer works

[Artur Skawina <art.08.09@gmail.com>]

Christian Lamparter wrote:
> On Thursday 15 January 2009 18:55:19 Artur Skawina wrote:
>> Artur Skawina wrote:
>>> Artur Skawina wrote:
>>>>>>> reboots, but locks up after only a few packets go over the hostap driven
>>>>>>> p54usb device. I need the box to be up, that limits the number of tests i can
>>>> after switching from SLUB to SLAB and enabling some debugging i finally caught this:
>>> and slub debugging gave something a bit more useful:
>>>
>>> =============================================================================
>>> BUG kmalloc-4096: Poison overwritten
>> tried reproducing (using the old w-t/pending pull) on different machine,
>> but no oops or crash there...
>> (similar connectivity issues though, such as connecting only works
>> exactly once after starting hostapd, to reconnect i have to restart
>> hostapd)
> hmm, maybe Windows Mobile uses another PS mechanism, e.g WMM-PS.

hmm, i'll have to set up another device to sniff the traffic i guess.

>> Upgraded to current wireless-testing/pending on the problematic box
>> and almost immediately got [1]. No slab corruption this time (at least
>> not yet). Will switch to GFP_ATOMIC and retry w/ the new fw.
>>
>> artur
> yes that's a bug, p54_set_tim must use GFP_ATOMIC, do you want to post the patch?

nah, too trivial :) fwiw i used this:

diff --git a/drivers/net/wireless/p54/p54common.c b/drivers/net/wireless/p54/p54common.c
index 635b423..579bc6c 100644
--- a/drivers/net/wireless/p54/p54common.c
+++ b/drivers/net/wireless/p54/p54common.c
@@ -1295,7 +1295,7 @@ static int p54_set_tim(struct ieee80211_hw *dev, struct ieee80211_sta *sta,
 	struct p54_tim *tim;
 
 	skb = p54_alloc_skb(dev, P54_HDR_FLAG_CONTROL_OPSET, sizeof(*tim),
-			    P54_CONTROL_TYPE_TIM, GFP_KERNEL);
+			    P54_CONTROL_TYPE_TIM, GFP_ATOMIC);
 	if (!skb)
 		return -ENOMEM;


Will try to do some more tests in the next few days.

Thanks,

artur