Any ideas on how to fix rssi?

Any ideas on how to fix rssi?

[Ben Greear]

I'm seeing incorrect signal strength when running a rate-vs-range
test through an attenuator.  Maybe the FIXME below is the reason?

Any suggestions on what needs doing here?

Thanks,
Ben


static void ath10k_htt_rx_frag_handler(struct ath10k_htt *htt,
				struct htt_rx_fragment_indication *frag)
{
	struct sk_buff *msdu_head, *msdu_tail;
	enum htt_rx_mpdu_encrypt_type enctype;
	struct htt_rx_desc *rxd;
	enum rx_msdu_decap_format fmt;
	struct ieee80211_rx_status *rx_status = &htt->rx_status;
	struct ieee80211_hdr *hdr;
	int ret;
	bool tkip_mic_err;
	bool decrypt_err;
	u8 *fw_desc;
	int fw_desc_len, hdrlen, paramlen;
	int trim;

	fw_desc_len = __le16_to_cpu(frag->fw_rx_desc_bytes);
	fw_desc = (u8 *)frag->fw_msdu_rx_desc;

	msdu_head = NULL;
	msdu_tail = NULL;

	spin_lock_bh(&htt->rx_ring.lock);
	ret = ath10k_htt_rx_amsdu_pop(htt, &fw_desc, &fw_desc_len,
				      &msdu_head, &msdu_tail);
	spin_unlock_bh(&htt->rx_ring.lock);

	ath10k_dbg(ATH10K_DBG_HTT_DUMP, "htt rx frag ahead\n");

	if (ret) {
		ath10k_warn("failed to pop amsdu from httr rx ring for fragmented rx %d\n",
			    ret);
		ath10k_htt_rx_free_msdu_chain(msdu_head);
		return;
	}

	/* FIXME: implement signal strength */


-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com


_______________________________________________
ath10k mailing list
ath10k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath10k

Re: Any ideas on how to fix rssi?

[Ben Greear]

On 05/14/2014 04:22 PM, Ben Greear wrote:
> I'm seeing incorrect signal strength when running a rate-vs-range
> test through an attenuator.  Maybe the FIXME below is the reason?
> 
> Any suggestions on what needs doing here?

Actually, it wasn't (just?) this...the patch below makes the numbers look
good, but obviously there could be a better fix:

diff --git a/drivers/net/wireless/ath/ath10k/htt_rx.c b/drivers/net/wireless/ath/ath10k/htt_rx.c
index ccbc363..f26f296 100644
--- a/drivers/net/wireless/ath/ath10k/htt_rx.c
+++ b/drivers/net/wireless/ath/ath10k/htt_rx.c
@@ -1218,8 +1218,13 @@ static void ath10k_htt_rx_handler(struct ath10k_htt *htt,
        mpdu_ranges = htt_rx_ind_get_mpdu_ranges(rx);

        /* Fill this once, while this is per-ppdu */
-       rx_status->signal  = ATH10K_DEFAULT_NOISE_FLOOR;
-       rx_status->signal += rx->ppdu.combined_rssi;
+       if (rx->ppdu.combined_rssi) {
+               rx_status->signal  = ATH10K_DEFAULT_NOISE_FLOOR;
+               rx_status->signal += rx->ppdu.combined_rssi;
+       }
+       else {
+               rx_status->flag |= RX_FLAG_NO_SIGNAL_VAL;
+       }

        if (rx->ppdu.info0 & HTT_RX_INDICATION_INFO0_END_VALID) {
                /* TSF available only in 32-bit */
@@ -1342,7 +1347,8 @@ static void ath10k_htt_rx_frag_handler(struct ath10k_htt *htt,
        }

        /* FIXME: implement signal strength */
-
+       rx_status->flag |= RX_FLAG_NO_SIGNAL_VAL;
+
        hdr = (struct ieee80211_hdr *)msdu_head->data;
        rxd = (void *)msdu_head->data - sizeof(*rxd);
        tkip_mic_err = !!(__le32_to_cpu(rxd->attention.flags) &



> 
> Thanks,
> Ben
> 
> 
> static void ath10k_htt_rx_frag_handler(struct ath10k_htt *htt,
> 				struct htt_rx_fragment_indication *frag)
> {
> 	struct sk_buff *msdu_head, *msdu_tail;
> 	enum htt_rx_mpdu_encrypt_type enctype;
> 	struct htt_rx_desc *rxd;
> 	enum rx_msdu_decap_format fmt;
> 	struct ieee80211_rx_status *rx_status = &htt->rx_status;
> 	struct ieee80211_hdr *hdr;
> 	int ret;
> 	bool tkip_mic_err;
> 	bool decrypt_err;
> 	u8 *fw_desc;
> 	int fw_desc_len, hdrlen, paramlen;
> 	int trim;
> 
> 	fw_desc_len = __le16_to_cpu(frag->fw_rx_desc_bytes);
> 	fw_desc = (u8 *)frag->fw_msdu_rx_desc;
> 
> 	msdu_head = NULL;
> 	msdu_tail = NULL;
> 
> 	spin_lock_bh(&htt->rx_ring.lock);
> 	ret = ath10k_htt_rx_amsdu_pop(htt, &fw_desc, &fw_desc_len,
> 				      &msdu_head, &msdu_tail);
> 	spin_unlock_bh(&htt->rx_ring.lock);
> 
> 	ath10k_dbg(ATH10K_DBG_HTT_DUMP, "htt rx frag ahead\n");
> 
> 	if (ret) {
> 		ath10k_warn("failed to pop amsdu from httr rx ring for fragmented rx %d\n",
> 			    ret);
> 		ath10k_htt_rx_free_msdu_chain(msdu_head);
> 		return;
> 	}
> 
> 	/* FIXME: implement signal strength */
> 
> 


-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com


_______________________________________________
ath10k mailing list
ath10k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath10k

Re: Any ideas on how to fix rssi?

[Janusz Dziedzic]

On 15 May 2014 02:07, Ben Greear <greearb@candelatech.com> wrote:
> On 05/14/2014 04:22 PM, Ben Greear wrote:
>> I'm seeing incorrect signal strength when running a rate-vs-range
>> test through an attenuator.  Maybe the FIXME below is the reason?
>>
>> Any suggestions on what needs doing here?
>
> Actually, it wasn't (just?) this...the patch below makes the numbers look
> good, but obviously there could be a better fix:
>
> diff --git a/drivers/net/wireless/ath/ath10k/htt_rx.c b/drivers/net/wireless/ath/ath10k/htt_rx.c
> index ccbc363..f26f296 100644
> --- a/drivers/net/wireless/ath/ath10k/htt_rx.c
> +++ b/drivers/net/wireless/ath/ath10k/htt_rx.c
> @@ -1218,8 +1218,13 @@ static void ath10k_htt_rx_handler(struct ath10k_htt *htt,
>         mpdu_ranges = htt_rx_ind_get_mpdu_ranges(rx);
>
>         /* Fill this once, while this is per-ppdu */
> -       rx_status->signal  = ATH10K_DEFAULT_NOISE_FLOOR;
> -       rx_status->signal += rx->ppdu.combined_rssi;

Seems you have old code without ath10k: fix rssi and rate reporting patch.

BR
Janusz
> +       if (rx->ppdu.combined_rssi) {
> +               rx_status->signal  = ATH10K_DEFAULT_NOISE_FLOOR;
> +               rx_status->signal += rx->ppdu.combined_rssi;
> +       }
> +       else {
> +               rx_status->flag |= RX_FLAG_NO_SIGNAL_VAL;
> +       }
>
>         if (rx->ppdu.info0 & HTT_RX_INDICATION_INFO0_END_VALID) {
>                 /* TSF available only in 32-bit */
> @@ -1342,7 +1347,8 @@ static void ath10k_htt_rx_frag_handler(struct ath10k_htt *htt,
>         }
>
>         /* FIXME: implement signal strength */
> -
> +       rx_status->flag |= RX_FLAG_NO_SIGNAL_VAL;
> +
>         hdr = (struct ieee80211_hdr *)msdu_head->data;
>         rxd = (void *)msdu_head->data - sizeof(*rxd);
>         tkip_mic_err = !!(__le32_to_cpu(rxd->attention.flags) &
>
>
>
>>
>> Thanks,
>> Ben
>>
>>
>> static void ath10k_htt_rx_frag_handler(struct ath10k_htt *htt,
>>                               struct htt_rx_fragment_indication *frag)
>> {
>>       struct sk_buff *msdu_head, *msdu_tail;
>>       enum htt_rx_mpdu_encrypt_type enctype;
>>       struct htt_rx_desc *rxd;
>>       enum rx_msdu_decap_format fmt;
>>       struct ieee80211_rx_status *rx_status = &htt->rx_status;
>>       struct ieee80211_hdr *hdr;
>>       int ret;
>>       bool tkip_mic_err;
>>       bool decrypt_err;
>>       u8 *fw_desc;
>>       int fw_desc_len, hdrlen, paramlen;
>>       int trim;
>>
>>       fw_desc_len = __le16_to_cpu(frag->fw_rx_desc_bytes);
>>       fw_desc = (u8 *)frag->fw_msdu_rx_desc;
>>
>>       msdu_head = NULL;
>>       msdu_tail = NULL;
>>
>>       spin_lock_bh(&htt->rx_ring.lock);
>>       ret = ath10k_htt_rx_amsdu_pop(htt, &fw_desc, &fw_desc_len,
>>                                     &msdu_head, &msdu_tail);
>>       spin_unlock_bh(&htt->rx_ring.lock);
>>
>>       ath10k_dbg(ATH10K_DBG_HTT_DUMP, "htt rx frag ahead\n");
>>
>>       if (ret) {
>>               ath10k_warn("failed to pop amsdu from httr rx ring for fragmented rx %d\n",
>>                           ret);
>>               ath10k_htt_rx_free_msdu_chain(msdu_head);
>>               return;
>>       }
>>
>>       /* FIXME: implement signal strength */
>>
>>
>
>
> --
> Ben Greear <greearb@candelatech.com>
> Candela Technologies Inc  http://www.candelatech.com
>
>
> _______________________________________________
> ath10k mailing list
> ath10k@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/ath10k

_______________________________________________
ath10k mailing list
ath10k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath10k

Re: Any ideas on how to fix rssi?

[Ben Greear]

On 05/14/2014 10:07 PM, Janusz Dziedzic wrote:
> On 15 May 2014 02:07, Ben Greear <greearb@candelatech.com> wrote:
>> On 05/14/2014 04:22 PM, Ben Greear wrote:
>>> I'm seeing incorrect signal strength when running a rate-vs-range
>>> test through an attenuator.  Maybe the FIXME below is the reason?
>>>
>>> Any suggestions on what needs doing here?
>>
>> Actually, it wasn't (just?) this...the patch below makes the numbers look
>> good, but obviously there could be a better fix:
>>
>> diff --git a/drivers/net/wireless/ath/ath10k/htt_rx.c b/drivers/net/wireless/ath/ath10k/htt_rx.c
>> index ccbc363..f26f296 100644
>> --- a/drivers/net/wireless/ath/ath10k/htt_rx.c
>> +++ b/drivers/net/wireless/ath/ath10k/htt_rx.c
>> @@ -1218,8 +1218,13 @@ static void ath10k_htt_rx_handler(struct ath10k_htt *htt,
>>         mpdu_ranges = htt_rx_ind_get_mpdu_ranges(rx);
>>
>>         /* Fill this once, while this is per-ppdu */
>> -       rx_status->signal  = ATH10K_DEFAULT_NOISE_FLOOR;
>> -       rx_status->signal += rx->ppdu.combined_rssi;
> 
> Seems you have old code without ath10k: fix rssi and rate reporting patch.

Yes, thanks for the suggestion.  I had patches on each side of that one,
but somehow I had missed it.  Will test with it today.

Thanks,
Ben

-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com


_______________________________________________
ath10k mailing list
ath10k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath10k