diff mbox series

[v1] cfg80211: Fix bitrates overflow issue

Message ID 1665493468-24774-1-git-send-email-quic_paulz@quicinc.com
State New
Headers show
Series [v1] cfg80211: Fix bitrates overflow issue | expand

Commit Message

Paul Zhang Oct. 11, 2022, 1:04 p.m. UTC
When invoking function cfg80211_calculate_bitrate_eht about
(320 MHz, EHT-MCS 13, EHT-NSS 2, EHT-GI 0), which means the
parameters as flags: 0x80, bw: 7, mcs: 13, eht_gi: 0, nss: 2,
this formula (result * rate->nss) will overflow and causes
the returned bitrate to be 3959 when it should be 57646.

Here is the explanation:
 u64 tmp;
 u32 result;
 …
 /* tmp = result = 4 * rates_996[0]
  *     = 4 * 480388888 = 0x72889c60
  */
 tmp = result;

 /* tmp = 0x72889c60 * 6144 = 0xabccea90000 */
 tmp *= SCALE;

 /* tmp = 0xabccea90000 / mcs_divisors[13]
  *     = 0xabccea90000 / 5120 = 0x8970bba6
  */
 do_div(tmp, mcs_divisors[rate->mcs]);

 /* result = 0x8970bba6 */
 result = tmp;

 /* normally (result * rate->nss) = 0x8970bba6 * 2 = 0x112e1774c,
  * but since result is u32, (result * rate->nss) = 0x12e1774c,
  * overflow happens and it loses the highest bit.
  * Then result =  0x12e1774c / 8 = 39595753,
  */
 result = (result * rate->nss) / 8;

Signed-off-by: Paul Zhang <quic_paulz@quicinc.com>
---
 net/wireless/util.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
diff mbox series

Patch

diff --git a/net/wireless/util.c b/net/wireless/util.c
index f09d528..59d4947 100644
--- a/net/wireless/util.c
+++ b/net/wireless/util.c
@@ -1557,10 +1557,12 @@  static u32 cfg80211_calculate_bitrate_eht(struct rate_info *rate)
 	tmp = result;
 	tmp *= SCALE;
 	do_div(tmp, mcs_divisors[rate->mcs]);
-	result = tmp;
 
 	/* and take NSS */
-	result = (result * rate->nss) / 8;
+	tmp *= rate->nss;
+	do_div(tmp, 8);
+
+	result = tmp;
 
 	return result / 10000;
 }