Re: [ath5k-devel] [PATCH 0/5] Misc ath5k patches

2009/4/13 Bob Copeland : > Here are a few random patches I have sitting in my local tree, currently > based before the ath5k/9k merge.  The last one comes from a perl > script I wrote to check the initvals (also wrote one to check rfgain > values, but no errors there).  Comments welcome. > > Bob Co

[ath5k-devel] [PATCH 4/5] ath5k: use rx hw descriptor pointer for self-linked check

This patch simplifies the code used to detect when the self-linked DMA buffer is still in use by hardware, by checking the hardware's rxdp register instead of looking at the software buffer list. Signed-off-by: Bob Copeland --- drivers/net/wireless/ath5k/base.c | 24 d

[ath5k-devel] [PATCH 1/5] ath5k: use bool for modparams

Current code uses int types, but both modparams are boolean values. Changes-licensed-under: 3-Clause-BSD Signed-off-by: Bob Copeland --- drivers/net/wireless/ath5k/base.c |4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/wireless/ath5k/base.c b/drivers/net

[ath5k-devel] [PATCH 2/5] ath5k: manipulate rxlink and descriptor address under rxbuf lock

Grabbing an ath5k_buf then dropping the lock is racy because the referenced descriptor can be obtained in another thread and released before the buffer is handed to the hardware. Likewise, manipulating sc->rxlink without the lock can lead to having multiple self-linked hardware descriptors. Chang

[ath5k-devel] [PATCH 5/5] ath5k: fix initvals errors

This patch corrects a few errors in the initvals tables to match those in the HAL tables. Namely, remove a couple of repetitions, fix some turbo mode errors, and correct a register for the CCK rate power table. Changes-licensed-under: ISC Signed-off-by: Bob Copeland --- drivers/net/wireless/at

[ath5k-devel] [PATCH 0/5] Misc ath5k patches

Here are a few random patches I have sitting in my local tree, currently based before the ath5k/9k merge. The last one comes from a perl script I wrote to check the initvals (also wrote one to check rfgain values, but no errors there). Comments welcome. Bob Copeland (5): ath5k: use bool for m

[ath5k-devel] [PATCH 3/5] ath5k: use tasklet_hi_schedule for beacon queue

For embedded platforms, beacon transmission can be starved when flooded with data packets. Prioritize beacons by giving the beacon queue the first shot when the isr completes. Changes-licensed-under: 3-Clause-BSD Signed-off-by: Bob Copeland --- drivers/net/wireless/ath5k/base.c |2 +- 1 fi

Re: [ath5k-devel] beacon loss from AP - sending probe request

On Sun, Apr 12, 2009 at 02:14:55AM +0200, Andreas Seidler wrote: > > Yup, it's normal. A pending patch will remove the message. > > > but i dont want to get rid of this message, i dont want to loose my > connection to AP while scanning... Understood - see this thread for an explanation of what