* [PATCH] rx checksum support for gige nForce ethernet
@ 2004-08-07 11:22 Manfred Spraul
0 siblings, 0 replies; only message in thread
From: Manfred Spraul @ 2004-08-07 11:22 UTC (permalink / raw)
To: Netdev
[-- Attachment #1: Type: text/plain, Size: 296 bytes --]
Hi,
With help from nVidia I have added rx checksumming to the forcedeth
driver. It's only supported by the gigabit version of the chipset.
Please test it - it works for me. Depends on the previous patches to
2.6.8-rc3-mm1, but rejects against 2.6.8-rc3-mm1 should be trivial.
--
Manfred
[-- Attachment #2: patch-forcedeth-rxcheck --]
[-- Type: text/plain, Size: 2587 bytes --]
// $Header$
// Kernel Version:
// VERSION = 2
// PATCHLEVEL = 6
// SUBLEVEL = 8
// EXTRAVERSION =-rc3-mm1
--- 2.6/drivers/net/forcedeth.c 2004-08-07 13:09:21.512115128 +0200
+++ build-2.6/drivers/net/forcedeth.c 2004-08-07 13:15:37.308985320 +0200
@@ -78,6 +78,7 @@
* 0.29: 28 Jul 2004: Add jumbo frame support. Add reset into nv_close,
* previous code clobbered kfree'd memory.
* 0.30: 07 Aug 2004: Add backup timer for link change notification.
+ * 0.31: 07 Aug 2004: rx checksum support
*
* Known bugs:
* We suspect that on some hardware no TX done interrupts are generated.
@@ -89,7 +90,7 @@
* DEV_NEED_TIMERIRQ will not harm you on sane hardware, only generating a few
* superfluous timer interrupts from the nic.
*/
-#define FORCEDETH_VERSION "0.30"
+#define FORCEDETH_VERSION "0.31"
#define DRV_NAME "forcedeth"
#include <linux/module.h>
@@ -219,6 +220,7 @@ enum {
#define NVREG_TXRXCTL_BIT2 0x0004
#define NVREG_TXRXCTL_IDLE 0x0008
#define NVREG_TXRXCTL_RESET 0x0010
+#define NVREG_TXRXCTL_RXCHECK 0x0400
NvRegMIIStatus = 0x180,
#define NVREG_MIISTAT_ERROR 0x0001
#define NVREG_MIISTAT_LINKCHANGE 0x0008
@@ -315,6 +317,10 @@ struct ring_desc {
#define NV_RX_ERROR (1<<30)
#define NV_RX_AVAIL (1<<31)
+#define NV_RX2_CHECKSUMMASK (0x1C000000)
+#define NV_RX2_CHECKSUMOK1 (0x10000000)
+#define NV_RX2_CHECKSUMOK2 (0x14000000)
+#define NV_RX2_CHECKSUMOK3 (0x18000000)
#define NV_RX2_DESCRIPTORVALID (1<<29)
#define NV_RX2_SUBSTRACT1 (1<<25)
#define NV_RX2_ERROR1 (1<<18)
@@ -377,8 +383,15 @@ struct ring_desc {
#define POLL_WAIT (1+HZ/100)
#define LINK_TIMEOUT (3*HZ)
+/*
+ * desc_ver values:
+ * This field has two purposes:
+ * - Newer nics uses a different ring layout. The layout is selected by
+ * comparing np->desc_ver with DESC_VER_xy.
+ * - It contains bits that are forced on when writing to NvRegTxRxControl.
+ */
#define DESC_VER_1 0x0
-#define DESC_VER_2 0x02100
+#define DESC_VER_2 (0x02100|NVREG_TXRXCTL_RXCHECK)
/* PHY defines */
#define PHY_OUI_MARVELL 0x5043
@@ -1200,6 +1213,15 @@ static void nv_rx_process(struct net_dev
goto next_pkt;
}
}
+ Flags &= NV_RX2_CHECKSUMMASK;
+ if (Flags == NV_RX2_CHECKSUMOK1 ||
+ Flags == NV_RX2_CHECKSUMOK2 ||
+ Flags == NV_RX2_CHECKSUMOK3) {
+ dprintk(KERN_DEBUG "%s: hw checksum hit!.\n", dev->name);
+ np->rx_skbuff[i]->ip_summed = CHECKSUM_UNNECESSARY;
+ } else {
+ dprintk(KERN_DEBUG "%s: hwchecksum miss!.\n", dev->name);
+ }
}
/* got a valid packet - forward it to the network core */
skb = np->rx_skbuff[i];
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2004-08-07 11:22 UTC | newest]
Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2004-08-07 11:22 [PATCH] rx checksum support for gige nForce ethernet Manfred Spraul
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).