Re: [PATCH] ath10k: handle ieee80211 header and payload tracing separately

2014-10-30 Thread Rajkumar Manoharan 
On Wed, Oct 29, 2014 at 04:05:44PM +0100, Johannes Berg wrote:
 On Wed, 2014-10-29 at 19:39 +0530, Rajkumar Manoharan wrote:
 
  +   u32 hdr_len = ieee80211_hdrlen(hdr-frame_control);
 
  +   trace_ath10k_rx_hdr(ar, hdr, hdr_len);
  +   trace_ath10k_rx_payload(ar, skb-data + hdr_len, skb-len - hdr_len);
 
 Another note, since you're taking my suggestion :-)
 
 It would be worth hiding that inside the tracepoint's assign function,
 so instead of passing data/len here you'd pass the full skb, or the full
 skb data/skb len, like this:
 
 ar, skb-data, skb-len
 
 to both tracers. Then inside the tracer you can do the hdrlen check, and
 that way move the code into the tracing so it's not hit when tracing is
 disabled.
 
Nice. Let me send v2. Thanks once again.

-Rajkumar
--
To unsubscribe from this list: send the line unsubscribe linux-wireless in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH] ath10k: handle ieee80211 header and payload tracing separately

2014-10-29 Thread Rajkumar Manoharan 
For packet log, the transmitted frame 802.11 header alone is sufficient.
Recording entire packet is also consuming lot of disk space. To optimize
this, tx and rx data tracepoints are splitted into header and payload
tracepoints.

To record tx ieee80211 headers

 trace-cmd record -e ath10k_tx_hdr

To record complete packets

 trace-cmd record -e ath10k_tx_hdr -e ath10k_tx_payload

Cc: Michal Kazior michal.kaz...@tieto.com
Signed-off-by: Rajkumar Manoharan rmano...@qti.qualcomm.com
---
 drivers/net/wireless/ath/ath10k/htt_rx.c |  5 +++--
 drivers/net/wireless/ath/ath10k/htt_tx.c |  4 +++-
 drivers/net/wireless/ath/ath10k/trace.h  |  9 +
 drivers/net/wireless/ath/ath10k/wmi.c| 15 ---
 4 files changed, 23 insertions(+), 10 deletions(-)

diff --git a/drivers/net/wireless/ath/ath10k/htt_rx.c 
b/drivers/net/wireless/ath/ath10k/htt_rx.c
index a691fdf..462f808 100644
--- a/drivers/net/wireless/ath/ath10k/htt_rx.c
+++ b/drivers/net/wireless/ath/ath10k/htt_rx.c
@@ -297,8 +297,6 @@ static inline struct sk_buff 
*ath10k_htt_rx_netbuf_pop(struct ath10k_htt *htt)
 DMA_FROM_DEVICE);
ath10k_dbg_dump(ar, ATH10K_DBG_HTT_DUMP, NULL, htt rx netbuf pop: ,
msdu-data, msdu-len + skb_tailroom(msdu));
-   trace_ath10k_htt_rx_pop_msdu(ar, msdu-data, msdu-len +
-skb_tailroom(msdu));
 
return msdu;
 }
@@ -875,6 +873,7 @@ static void ath10k_process_rx(struct ath10k *ar,
struct ieee80211_rx_status *status;
struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)skb-data;
char tid[32];
+   u32 hdr_len = ieee80211_hdrlen(hdr-frame_control);
 
status = IEEE80211_SKB_RXCB(skb);
*status = *rx_status;
@@ -903,6 +902,8 @@ static void ath10k_process_rx(struct ath10k *ar,
   !!(status-flag  RX_FLAG_AMSDU_MORE));
ath10k_dbg_dump(ar, ATH10K_DBG_HTT_DUMP, NULL, rx skb: ,
skb-data, skb-len);
+   trace_ath10k_rx_hdr(ar, hdr, hdr_len);
+   trace_ath10k_rx_payload(ar, skb-data + hdr_len, skb-len - hdr_len);
 
ieee80211_rx(ar-hw, skb);
 }
diff --git a/drivers/net/wireless/ath/ath10k/htt_tx.c 
b/drivers/net/wireless/ath/ath10k/htt_tx.c
index b0df470..b19d43b 100644
--- a/drivers/net/wireless/ath/ath10k/htt_tx.c
+++ b/drivers/net/wireless/ath/ath10k/htt_tx.c
@@ -455,6 +455,7 @@ int ath10k_htt_tx(struct ath10k_htt *htt, struct sk_buff 
*msdu)
dma_addr_t paddr;
u32 frags_paddr;
bool use_frags;
+   u32 hdr_len = ieee80211_hdrlen(hdr-frame_control);
 
res = ath10k_htt_tx_inc_pending(htt);
if (res)
@@ -564,7 +565,8 @@ int ath10k_htt_tx(struct ath10k_htt *htt, struct sk_buff 
*msdu)
   (u32)skb_cb-paddr, vdev_id, tid);
ath10k_dbg_dump(ar, ATH10K_DBG_HTT_DUMP, NULL, htt tx msdu: ,
msdu-data, msdu-len);
-   trace_ath10k_htt_tx_msdu(ar, msdu-data, msdu-len);
+   trace_ath10k_tx_hdr(ar, hdr, hdr_len);
+   trace_ath10k_tx_payload(ar, msdu-data + hdr_len, msdu-len - hdr_len);
 
sg_items[0].transfer_id = 0;
sg_items[0].transfer_context = NULL;
diff --git a/drivers/net/wireless/ath/ath10k/trace.h 
b/drivers/net/wireless/ath/ath10k/trace.h
index b9a2ba6..fd9e6bc 100644
--- a/drivers/net/wireless/ath/ath10k/trace.h
+++ b/drivers/net/wireless/ath/ath10k/trace.h
@@ -368,22 +368,22 @@ DECLARE_EVENT_CLASS(ath10k_data_event,
)
 );
 
-DEFINE_EVENT(ath10k_data_event, ath10k_htt_tx_msdu,
+DEFINE_EVENT(ath10k_data_event, ath10k_tx_hdr,
 TP_PROTO(struct ath10k *ar, void *data, size_t len),
 TP_ARGS(ar, data, len)
 );
 
-DEFINE_EVENT(ath10k_data_event, ath10k_htt_rx_pop_msdu,
+DEFINE_EVENT(ath10k_data_event, ath10k_tx_payload,
 TP_PROTO(struct ath10k *ar, void *data, size_t len),
 TP_ARGS(ar, data, len)
 );
 
-DEFINE_EVENT(ath10k_data_event, ath10k_wmi_mgmt_tx,
+DEFINE_EVENT(ath10k_data_event, ath10k_rx_hdr,
 TP_PROTO(struct ath10k *ar, void *data, size_t len),
 TP_ARGS(ar, data, len)
 );
 
-DEFINE_EVENT(ath10k_data_event, ath10k_wmi_bcn_tx,
+DEFINE_EVENT(ath10k_data_event, ath10k_rx_payload,
 TP_PROTO(struct ath10k *ar, void *data, size_t len),
 TP_ARGS(ar, data, len)
 );
@@ -392,6 +392,7 @@ DEFINE_EVENT(ath10k_data_event, ath10k_htt_rx_desc,
 TP_PROTO(struct ath10k *ar, void *data, size_t len),
 TP_ARGS(ar, data, len)
 );
+
 #endif /* _TRACE_H_ || TRACE_HEADER_MULTI_READ*/
 
 /* we don't want to use include/trace/events */
diff --git a/drivers/net/wireless/ath/ath10k/wmi.c 
b/drivers/net/wireless/ath/ath10k/wmi.c
index ae746ce..c241446 100644
--- a/drivers/net/wireless/ath/ath10k/wmi.c
+++ b/drivers/net/wireless/ath/ath10k/wmi.c
@@ -798,9 +798,11 @@ int ath10k_wmi_mgmt_tx(struct ath10k *ar, struct sk_buff 
*skb)
int len;
u32 buf_len = skb-len;
u16 fc;
+   u32 hdr_len;

Re: [PATCH] ath10k: handle ieee80211 header and payload tracing separately

2014-10-29 Thread Johannes Berg 
On Wed, 2014-10-29 at 19:39 +0530, Rajkumar Manoharan wrote:

 + u32 hdr_len = ieee80211_hdrlen(hdr-frame_control);

 + trace_ath10k_rx_hdr(ar, hdr, hdr_len);
 + trace_ath10k_rx_payload(ar, skb-data + hdr_len, skb-len - hdr_len);

Another note, since you're taking my suggestion :-)

It would be worth hiding that inside the tracepoint's assign function,
so instead of passing data/len here you'd pass the full skb, or the full
skb data/skb len, like this:

ar, skb-data, skb-len

to both tracers. Then inside the tracer you can do the hdrlen check, and
that way move the code into the tracing so it's not hit when tracing is
disabled.

johannes

--
To unsubscribe from this list: send the line unsubscribe linux-wireless in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html