Correct a performance bug from Bill Paul's original FreeBSD bge(4) driver:

Each call to the FreeBSD bge_start() routine the transmit producer
pointer index from the chip mailbox register BGE_MBX_TX_HOST_PROD0_LO.
The local copy of that value is then updated by bge_encap() as
bge_encap() encapsulates packets in the Tx ring. If bge_encap()
succeds in encpuslating one or more packets, bge_start() tells the
chip to start sending the newly-encinitiates writes the new value back
to the chip mailbox register.

However, comparison of the Linux drivers (Broadcom-supplied and
open-source tg3.c) and to the OpenSolaris driver confirms that
register BGE_MBX_TX_HOST_PROD0_LO is write-only to software.
Thus, we can just keep a copy in the softc, and eliminate the
(expensive) PCI register write on each call to bge_start().

From jonathan NetBSD

tested by krw@, sturm@ and I on a few different bge NICs.

1 files changed, 2 insertions, 1 deletions

diff --git a/sys/dev/pci/if_bgereg.h b/sys/dev/pci/if_bgereg.h
index d0962d8a1a0..8ab2b8f0344 100644
--- a/sys/dev/pci/if_bgereg.h
+++ b/sys/dev/pci/if_bgereg.h
@@ -1,4 +1,4 @@
-/*	$OpenBSD: if_bgereg.h,v 1.30 2005/10/09 23:41:55 brad Exp $	*/
+/*	$OpenBSD: if_bgereg.h,v 1.31 2005/11/19 23:04:48 brad Exp $	*/
 
 /*
  * Copyright (c) 2001 Wind River Systems
@@ -2328,6 +2328,7 @@ struct bge_softc {
 	u_int16_t		bge_rx_saved_considx;
 	u_int16_t		bge_ev_saved_considx;
 	u_int16_t		bge_return_ring_cnt;
+	u_int32_t		bge_tx_prodidx;
 	u_int16_t		bge_std;	/* current std ring head */
 	u_int16_t		bge_jumbo;	/* current jumo ring head */
 	SLIST_HEAD(__bge_jfreehead, bge_jpool_entry)	bge_jfree_listhead;