[rt2x00-users] rt2800pci: hang on module unload (SoC) and possible fix

[Gertjan van Wingerde]

On 04/21/10 18:02, Helmut Schaa wrote:
> Hi,
> 
> when using rt2800pci on the 305x SoC board it hangs sometimes when
> taking the interface
> down (everytime if at least one frame has been transmitted). It seems
> like the device gets
> stuck somehow and hangs the whole system. Yeah, that's the advantage of
> using a SoC ;)
> 
> Is anybody aware of similar issues on pci cards?
> 
> Nevertheless I've managed to find the cause of the hang  but I'm unsure
> on howto fix this
> in a nice and clean way. In order to disable the tx queues I had to
> disable TX DMA first, to
> be precise, before disabling the first queue already.
> 
> Using the patch below (disable TX DMA just before killing the tx queues)
> works just fine
> but is more like a hack.
> 
> What would be a clean way of fixing this issue?
> 1) Introduce a new callback disable_dma which would be called before calling
>    kill_tx_queue for every queue (from rt2x00 code)?
> 2) Just use the patch below as the code is only run from disable_radio
> which will
>    disable DMA anyway (later)?
> 3) just do nothing in rt2800_kill_tx_queue as I couldn't find any
> reference to similar
>    code in the ralink drivers?
> 4) Any other ideas?
> 
> Thanks,
> Helmut
> 
> 
> 
> Signed-off-by: Helmut Schaa <helmut.schaa at googlemail.com
> <mailto:helmut.schaa at googlemail.com>>
> ---
>  drivers/net/wireless/rt2x00/rt2800pci.c |    4 ++++
>  1 files changed, 4 insertions(+), 0 deletions(-)
> 
> diff --git a/drivers/net/wireless/rt2x00/rt2800pci.c
> b/drivers/net/wireless/rt2x00/rt2800pci.c
> index 0e52f17..f96fa8a 100644
> --- a/drivers/net/wireless/rt2x00/rt2800pci.c
> +++ b/drivers/net/wireless/rt2x00/rt2800pci.c
> @@ -806,6 +806,10 @@ static void rt2800pci_kill_tx_queue(struct
> rt2x00_dev *rt2x00dev,
>          rt2800_register_write(rt2x00dev, BCN_TIME_CFG, 0);
>          return;
>      }
> +
> +    rt2800_register_read(rt2x00dev, WPDMA_GLO_CFG, &reg);
> +    rt2x00_set_field32(&reg, WPDMA_GLO_CFG_ENABLE_TX_DMA, 0);
> +    rt2800_register_write(rt2x00dev, WPDMA_GLO_CFG, reg);
>  
>      rt2800_register_read(rt2x00dev, WPDMA_RST_IDX, &reg);
>      rt2x00_set_field32(&reg, WPDMA_RST_IDX_DTX_IDX0, (qid == QID_AC_BE));
> -- 
> 1.6.4.2
> 
> 

Hi, 

Sorry, I only just found this in my inbox.

It looks like my rt2800pci based PCI devices suffer from the same issue.
However, your patch doesn't seem to help. On my system it simply speeds up
the hang.

I don't have the time right now to fully investigate this, but your patch
doesn't seem to fix the issue.


---
Gertjan.