> >> ' Error wrong command queue 43 command id ___0x6B___' 6b looks more > >> like slub poison -- accessing already freed skb > >> > > > > > hmm, 0x6B indeed is not a documented command ID... > > Only triggering under load must point to some overflow or race i guess. > > > Yep. Though the queue number actually keeps varying, it wasn't always 43, but you're right, 0x6b sounds like poison. Have you enabled poison on your test machines? > > I should get myself a new laptop to be able to play with this... > > The best i can do now, is wonder if this patch > > "[PATCH 08/10] iwlwifi: decrement rx skb counter in scan abort handler" > > might be responsible, but that's just fuzzy string matching "recent patches" > > with > > "freed skb" ;-) > > No, that's a good patch.. Johannes failure looks like he got it right > in the begining before scanning. > We need open more logs and check what slub allocator is in use Indeed, I'm using slub on a 6.5G 64-bit powerpc box with IOMMU so there's lots of room for differences... johannes
Attachment:
signature.asc
Description: This is a digitally signed message part
