[PATCH 3.19.y-ckt 220/251] ath10k: add extra check for frame tracing

[Kamal Mostafa]

3.19.8-ckt4 -stable review patch.  If anyone has any objections, please let me 
know.

------------------

From: Michal Kazior <michal.kaz...@tieto.com>

commit 36d8230b7d4fcd6f1a0c98b579d229bc919fa1fb upstream.

Frames are logged via tracing in two slices:
header and payload, separately. This is done for
performance reasons when one wants to, e.g.
analyse metadata only of frames only.

If for some reason device delivered a frame buffer
which was sized below what 802.11 header implied
tracing logic would blow doing an invalid memory
accesses.

I've hit this problem when running IBSS on QCA988X
with 999.999.0.636 and tracing at the same time.

Fixes: 5ce8e7fdcc7a ("ath10k: handle ieee80211 header and payload tracing 
separately")
Signed-off-by: Michal Kazior <michal.kaz...@tieto.com>
Signed-off-by: Kalle Valo <kv...@qca.qualcomm.com>
Signed-off-by: Kamal Mostafa <ka...@canonical.com>
---
 drivers/net/wireless/ath/ath10k/trace.h | 20 +++++++++++++-------
 1 file changed, 13 insertions(+), 7 deletions(-)

diff --git a/drivers/net/wireless/ath/ath10k/trace.h 
b/drivers/net/wireless/ath/ath10k/trace.h
index b289378..571d7cb 100644
--- a/drivers/net/wireless/ath/ath10k/trace.h
+++ b/drivers/net/wireless/ath/ath10k/trace.h
@@ -21,11 +21,16 @@
 #include "core.h"
 
 #if !defined(_TRACE_H_)
-static inline u32 ath10k_frm_hdr_len(const void *buf)
+static inline u32 ath10k_frm_hdr_len(const void *buf, size_t len)
 {
        const struct ieee80211_hdr *hdr = buf;
 
-       return ieee80211_hdrlen(hdr->frame_control);
+       /* In some rare cases (e.g. fcs error) device reports frame buffer
+        * shorter than what frame header implies (e.g. len = 0). The buffer
+        * can still be accessed so do a simple min() to guarantee caller
+        * doesn't get value greater than len.
+        */
+       return min_t(u32, len, ieee80211_hdrlen(hdr->frame_control));
 }
 #endif
 
@@ -360,13 +365,13 @@ DECLARE_EVENT_CLASS(ath10k_hdr_event,
                __string(device, dev_name(ar->dev))
                __string(driver, dev_driver_string(ar->dev))
                __field(size_t, len)
-               __dynamic_array(u8, data, ath10k_frm_hdr_len(data))
+               __dynamic_array(u8, data, ath10k_frm_hdr_len(data, len))
        ),
 
        TP_fast_assign(
                __assign_str(device, dev_name(ar->dev));
                __assign_str(driver, dev_driver_string(ar->dev));
-               __entry->len = ath10k_frm_hdr_len(data);
+               __entry->len = ath10k_frm_hdr_len(data, len);
                memcpy(__get_dynamic_array(data), data, __entry->len);
        ),
 
@@ -387,15 +392,16 @@ DECLARE_EVENT_CLASS(ath10k_payload_event,
                __string(device, dev_name(ar->dev))
                __string(driver, dev_driver_string(ar->dev))
                __field(size_t, len)
-               __dynamic_array(u8, payload, (len - ath10k_frm_hdr_len(data)))
+               __dynamic_array(u8, payload, (len -
+                                             ath10k_frm_hdr_len(data, len)))
        ),
 
        TP_fast_assign(
                __assign_str(device, dev_name(ar->dev));
                __assign_str(driver, dev_driver_string(ar->dev));
-               __entry->len = len - ath10k_frm_hdr_len(data);
+               __entry->len = len - ath10k_frm_hdr_len(data, len);
                memcpy(__get_dynamic_array(payload),
-                      data + ath10k_frm_hdr_len(data), __entry->len);
+                      data + ath10k_frm_hdr_len(data, len), __entry->len);
        ),
 
        TP_printk(
-- 
1.9.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/