This is an alternate solution to the EEM 'sentinel' CRC valiation issue.
CDC EEM allows using a 'sentinel' ethernet frame CRC of 0xdeadbeef in
place of a real CRC. The 'sentinel' value is transmitted in big-endian
order whereas the normal CRC is little-endian. This patch handles both
cases appropriately.
Signed-off-by: Brian Niebuhr <bniebuhr@xxxxxxxxxxxxx>
---
drivers/net/usb/cdc_eem.c | 17 ++++++++++-------
1 files changed, 10 insertions(+), 7 deletions(-)
diff --git a/drivers/net/usb/cdc_eem.c b/drivers/net/usb/cdc_eem.c
index 45cebfb..2330065 100644
--- a/drivers/net/usb/cdc_eem.c
+++ b/drivers/net/usb/cdc_eem.c
@@ -300,20 +300,23 @@ static int eem_rx_fixup(struct usbnet *dev, struct sk_buff *skb)
return 0;
}
- crc = get_unaligned_le32(skb2->data
- + len - ETH_FCS_LEN);
- skb_trim(skb2, len - ETH_FCS_LEN);
-
/*
* The bmCRC helps to denote when the CRC field in
* the Ethernet frame contains a calculated CRC:
* bmCRC = 1 : CRC is calculated
* bmCRC = 0 : CRC = 0xDEADBEEF
*/
- if (header & BIT(14))
- crc2 = ~crc32_le(~0, skb2->data, skb2->len);
- else
+ if (header & BIT(14)) {
+ crc = get_unaligned_le32(skb2->data
+ + len - ETH_FCS_LEN);
+ crc2 = ~crc32_le(~0, skb2->data, skb2->len
+ - ETH_FCS_LEN);
+ } else {
+ crc = get_unaligned_be32(skb2->data
+ + len - ETH_FCS_LEN);
crc2 = 0xdeadbeef;
+ }
+ skb_trim(skb2, len - ETH_FCS_LEN);
if (is_last)
return crc == crc2;
--
1.6.0.4
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
