On Fri, 2011-09-16 at 11:42 +0200, Helmut Schaa wrote: > <Hi, > > On Thu, Sep 15, 2011 at 5:36 PM, Wey-Yi Guy <wey-yi.w.guy@xxxxxxxxx> wrote: > > From: Don Fry <donald.h.fry@xxxxxxxxx> > > > > This patch reverts commit 9b7688328422b88a7a15dc0dc123ad9ab1a6e22d which > > was introduced in 2.6.38-rc1. It works around a problem where the iwlagn > > driver stimulates a bug crashing (requiring power cycle to recover) some > > APs under heavy traffic. > > [...] > > > + /* > > + * Including the following line will crash some AP's. This > > + * workaround removes the stimulus which causes the crash until > > + * the AP software can be fixed. > > hw->max_tx_aggregation_subframes = LINK_QUAL_AGG_FRAME_LIMIT_DEF; > > + */ > > Would be interesting to know which AP models suffer from this issue since > other drivers might also be affected if max_tx_aggregation_subframes is set > to something != 64 or 0. > > Did this appear with the previous LINK_QUAL_AGG_FRAME_LIMIT_DEF > value of 31 too? > > If we can crash APs by sending values != 64 or 0 it might even make sense > to not allow drivers to overwrite the max AMPDU size ... It isn't that sending that directly crashes the APs right away. It still needs some traffic, so there's a possibility that we send too long AMPDUs or something else like that and just can't find that bug ... johannes -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
