Hi,
The driver 8139too configures interframe gap with the shortest possible
interval. This behaviour was commited by Jeff on version 0.9.4.1. From
changelog:
* Do not set Interfame Gap (IFG) bits in TxConfig
The chip spec, though, says that these bits must be set in order to
conform to 802.3. There is a comment on the code that says exactly the
opposite. I would like to know why the configuration is set this way.
I have a problem where an external device inserts VLAN tags in frames,
thus reducing the gap more yet. After that, too close frames (except the
first one) are not recognized by a switch.
I have set these bits and my problem disappears. My patch follows:
--- linux-2.6.10/drivers/net/8139too.c 2004-12-24 19:33:47.000000000 -0200
+++ linux-2.6.10-mod/drivers/net/8139too.c 2005-02-03 13:47:15.000000000 -0200
@@ -391,7 +391,7 @@
/* Bits in TxConfig. */
enum tx_config_bits {
TxIFG1 = (1 << 25), /* Interframe Gap Time */
- TxIFG0 = (1 << 24), /* Enabling these bits violates IEEE 802.3 */
+ TxIFG0 = (1 << 24), /* Disabling any of these bits violates IEEE 802.3 */
TxLoopBack = (1 << 18) | (1 << 17), /* enable loopback test mode */
TxCRC = (1 << 16), /* DISABLE appending CRC to end of Tx packets */
TxClearAbt = (1 << 0), /* Clear abort (WO) */
@@ -723,7 +723,7 @@
#endif
static const unsigned int rtl8139_tx_config =
- (TX_DMA_BURST << TxDMAShift) | (TX_RETRY << TxRetryShift);
+ TxIFG1 | TxIFG0 | (TX_DMA_BURST << TxDMAShift) | (TX_RETRY << TxRetryShift);
static void __rtl8139_cleanup_dev (struct net_device *dev)
{
Thanks.
Marcelo Abreu
-
: send the line "unsubscribe linux-net" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
