[RFC PATCH 1/3] ath10k: ignore diagnose pipes in ath10k_pci_ce_recv_data()/_send_done()

[Kalle Valo]

Michal Kazior <michal.kazior at tieto.com> writes:

> On 11 September 2014 22:31, Kalle Valo <kvalo at qca.qualcomm.com> wrote:
>> As ath10k_pci_diag_read_mem() uses polling to receive data from CE, the CE callbacks
>> have to ignore the pipe used for diagnose reads. Otherwise ath10k crashes due
>> to NULL dereference and polling reads timeout.
>>
>> Signed-off-by: Kalle Valo <kvalo at qca.qualcomm.com>
>> ---
>>  drivers/net/wireless/ath/ath10k/pci.c |    6 ++++++
>>  1 file changed, 6 insertions(+)
>>
>> diff --git a/drivers/net/wireless/ath/ath10k/pci.c b/drivers/net/wireless/ath/ath10k/pci.c
>> index baeb98e78b1f..154451ab3e3c 100644
>> --- a/drivers/net/wireless/ath/ath10k/pci.c
>> +++ b/drivers/net/wireless/ath/ath10k/pci.c
>> @@ -819,6 +819,9 @@ static void ath10k_pci_ce_send_done(struct ath10k_ce_pipe *ce_state)
>>         unsigned int nbytes;
>>         unsigned int transfer_id;
>>
>> +       if (ce_state->id == 7)
>> +               return;
>
> I think you can check it in a saner way: (ce_state == ar_pci->ce_diag).
>
> Another approach is to prevent this from happening in the first place
> by never unmasking the diagnostic window copy engine pipe in
> ath10k_ce_enable_interrupts() and preventing
> ath10k_ce_per_engine_service() from being called in
> ath10k_pci_hif_send_complete_check() for the diagnostic ce pipe.

I like the latter approach and I'll try to implement that first. Thanks!

-- 
Kalle Valo