Hello.
I have just got this card.
On the 3 1/2 diskette there is code for the module. these two files:
dm9xs.c dmfe.c
but the linux kernel 2.4.17 has got dmfe.c only.
The files that came with the card are said to be for 2.4 and 2.2 respectively.
How does the dmfe.c from the latest kernel differs from the ones that come with the card?
and which should I stick with?
Thank you.
--
ICQ: 15605359 Bicho
=^..^=
First, they ignore you. Then they laugh at you. Then they fight you. Then you win. Mahatma Gandhi.
-------------------------------$B5$8!BN$N0lCW(B------------------------------------
$B=k$54($5$bH`4_$^$G!#(B
$BNx$K$7$7$g$&$J$7!#Nx$O$7$"$s$NB>!#(B
$B%"%s(B $B%"%s(B $B%"%s(B $B$H$C$F$bBg9%$-(B
> The files that came with the card are said to be for 2.4 and 2.2 respectively.
> How does the dmfe.c from the latest kernel differs from the ones that come with the card?
> and which should I stick with?
Unless your card is very old - use the tulip driver
On Sat, 23 Mar 2002, David Eduardo Gomez Noguera wrote:
> Hello.
> I have just got this card.
> On the 3 1/2 diskette there is code for the module. these two files:
> dm9xs.c dmfe.c
> but the linux kernel 2.4.17 has got dmfe.c only.
The version in the kernel is based on the latest dm9xs.c driver, but is
called dmfe.c. I have no idea why Davicom decided to rename the driver
for 2.4.
> The files that came with the card are said to be for 2.4 and 2.2
> respectively.
Correct.
> How does the dmfe.c from the latest kernel differs from the ones that
> come with the card? and which should I stick with?
The version in the kernel is based on the latest dm9xs.c driver, but
contains some fixes that are not present in Davicom's driver. I recommend
using the one included with the Linux kernel.
As Alan mentioned, you can also use the tulip driver with this card since
it is (yet) a(nother) tulip clone. With your card you will probably get
better tx performance using the tulip driver at the moment.
/Tobias
On Sun, 24 Mar 2002 00:52:05 +0100 (CET)
Tobias Ringstrom <[email protected]> wrote:
>
> As Alan mentioned, you can also use the tulip driver with this card since
> it is (yet) a(nother) tulip clone. With your card you will probably get
> better tx performance using the tulip driver at the moment.
>
> /Tobias
Thank you both.
Why would tulip perform better though?
--
ICQ: 15605359 Bicho
=^..^=
First, they ignore you. Then they laugh at you. Then they fight you. Then you win. Mahatma Gandhi.
-------------------------------$B5$8!BN$N0lCW(B------------------------------------
$B=k$54($5$bH`4_$^$G!#(B
$BNx$K$7$7$g$&$J$7!#Nx$O$7$"$s$NB>!#(B
$B%"%s(B $B%"%s(B $B%"%s(B $B$H$C$F$bBg9%$-(B
On Sat, 23 Mar 2002, David Eduardo Gomez Noguera wrote:
> Why would tulip perform better though?
The dmfe driver's packet transmission code is not very efficient. I have
a patch that fixes it, but I do not know if it will break with older
Davicom chips, althouhg I see no reason why it should. I have included
the patch at the end of this email.
It would be nice if you could try the patch and let me know if it works
for you. I'd be very surprised if it does not work with your card (since
I have one of those).
/Tobias
--- /usr/src/linux-2.4.18.crypto/drivers/net/dmfe.c Fri Mar 1 10:21:38 2002
+++ ./dmfe.c Sun Mar 24 09:40:52 2002
@@ -61,7 +61,7 @@
*/
#define DRV_NAME "dmfe"
-#define DRV_VERSION "1.36.4"
+#define DRV_VERSION "1.36.4tx"
#define DRV_RELDATE "2002-01-17"
#include <linux/module.h>
@@ -100,11 +100,9 @@
#define DM9102_IO_SIZE 0x80
#define DM9102A_IO_SIZE 0x100
-#define TX_MAX_SEND_CNT 0x1 /* Maximum tx packet per time */
#define TX_DESC_CNT 0x10 /* Allocated Tx descriptors */
#define RX_DESC_CNT 0x20 /* Allocated Rx descriptors */
-#define TX_FREE_DESC_CNT (TX_DESC_CNT - 2) /* Max TX packet count */
-#define TX_WAKE_DESC_CNT (TX_DESC_CNT - 3) /* TX wakeup count */
+#define TX_IRQ_THR 12
#define DESC_ALL_CNT (TX_DESC_CNT + RX_DESC_CNT)
#define TX_BUF_ALLOC 0x600
#define RX_ALLOC_SIZE 0x620
@@ -211,8 +209,9 @@
struct rx_desc *first_rx_desc;
struct rx_desc *rx_insert_ptr;
struct rx_desc *rx_ready_ptr; /* packet come pointer */
- unsigned long tx_packet_cnt; /* transmitted packet count */
- unsigned long tx_queue_cnt; /* wait to send packet count */
+ int tx_int_pkt_num; /* number of packets to transmit until
+ * a transmit interrupt is requested */
+ u32 tx_live_cnt; /* number of used/live tx slots */
unsigned long rx_avail_cnt; /* available rx descriptor count */
unsigned long interval_rx_cnt; /* rx packet count a callback time */
@@ -541,8 +540,9 @@
db->buf_pool_ptr, db->buf_pool_dma_ptr);
unregister_netdev(dev);
pci_release_regions(pdev);
- kfree(dev); /* free board information */
+ pci_disable_device(pdev);
pci_set_drvdata(pdev, NULL);
+ kfree(dev); /* free board information */
}
DMFE_DBUG(0, "dmfe_remove_one() exit", 0);
@@ -567,8 +567,8 @@
/* system variable init */
db->cr6_data = CR6_DEFAULT | dmfe_cr6_user_set;
- db->tx_packet_cnt = 0;
- db->tx_queue_cnt = 0;
+ db->tx_int_pkt_num = TX_IRQ_THR;
+ db->tx_live_cnt = 0;
db->rx_avail_cnt = 0;
db->link_failed = 1;
db->wait_reset = 0;
@@ -687,9 +687,6 @@
DMFE_DBUG(0, "dmfe_start_xmit", 0);
- /* Resource flag check */
- netif_stop_queue(dev);
-
/* Too large packet check */
if (skb->len > MAX_PACKET_SIZE) {
printk(KERN_ERR DRV_NAME ": big packet = %d\n", (u16)skb->len);
@@ -699,38 +696,33 @@
spin_lock_irqsave(&db->lock, flags);
- /* No Tx resource check, it never happen nromally */
- if (db->tx_queue_cnt >= TX_FREE_DESC_CNT) {
- spin_unlock_irqrestore(&db->lock, flags);
- printk(KERN_ERR DRV_NAME ": No Tx resource %ld\n", db->tx_queue_cnt);
- return 1;
- }
-
/* Disable NIC interrupt */
outl(0, dev->base_addr + DCR7);
/* transmit this packet */
txptr = db->tx_insert_ptr;
memcpy(txptr->tx_buf_ptr, skb->data, skb->len);
- txptr->tdes1 = cpu_to_le32(0xe1000000 | skb->len);
+ if (--db->tx_int_pkt_num < 0)
+ {
+ txptr->tdes1 = cpu_to_le32(0xe1000000 | skb->len);
+ db->tx_int_pkt_num = TX_IRQ_THR;
+ }
+ else
+ txptr->tdes1 = cpu_to_le32(0x61000000 | skb->len);
+
+ /* Transmit Packet Process */
+ txptr->tdes0 = cpu_to_le32(0x80000000); /* set owner bit to DM910X */
+ outl(0x1, dev->base_addr + DCR1); /* Issue Tx polling command */
+ dev->trans_start = jiffies; /* saved the time stamp */
/* Point to next transmit free descriptor */
- db->tx_insert_ptr = txptr->next_tx_desc;
+ txptr = txptr->next_tx_desc;
- /* Transmit Packet Process */
- if ( (!db->tx_queue_cnt) && (db->tx_packet_cnt < TX_MAX_SEND_CNT) ) {
- txptr->tdes0 = cpu_to_le32(0x80000000); /* Set owner bit */
- db->tx_packet_cnt++; /* Ready to send */
- outl(0x1, dev->base_addr + DCR1); /* Issue Tx polling */
- dev->trans_start = jiffies; /* saved time stamp */
- } else {
- db->tx_queue_cnt++; /* queue TX packet */
- outl(0x1, dev->base_addr + DCR1); /* Issue Tx polling */
- }
+ if (txptr->tdes0 & cpu_to_le32(0x80000000))
+ netif_stop_queue(dev);
- /* Tx resource check */
- if ( db->tx_queue_cnt < TX_FREE_DESC_CNT )
- netif_wake_queue(dev);
+ db->tx_insert_ptr = txptr;
+ db->tx_live_cnt++;
/* free this SKB */
dev_kfree_skb(skb);
@@ -860,18 +852,18 @@
static void dmfe_free_tx_pkt(struct DEVICE *dev, struct dmfe_board_info * db)
{
struct tx_desc *txptr;
- unsigned long ioaddr = dev->base_addr;
u32 tdes0;
txptr = db->tx_remove_ptr;
- while(db->tx_packet_cnt) {
+ while (db->tx_live_cnt > 0)
+ {
tdes0 = le32_to_cpu(txptr->tdes0);
/* printk(DRV_NAME ": tdes0=%x\n", tdes0); */
if (tdes0 & 0x80000000)
break;
/* A packet sent completed */
- db->tx_packet_cnt--;
+ db->tx_live_cnt--;
db->stats.tx_packets++;
/* Transmit statistic counter */
@@ -908,17 +900,8 @@
/* Update TX remove pointer to next */
db->tx_remove_ptr = txptr;
- /* Send the Tx packet in queue */
- if ( (db->tx_packet_cnt < TX_MAX_SEND_CNT) && db->tx_queue_cnt ) {
- txptr->tdes0 = cpu_to_le32(0x80000000); /* Set owner bit */
- db->tx_packet_cnt++; /* Ready to send */
- db->tx_queue_cnt--;
- outl(0x1, ioaddr + DCR1); /* Issue Tx polling */
- dev->trans_start = jiffies; /* saved time stamp */
- }
-
/* Resource available check */
- if ( db->tx_queue_cnt < TX_WAKE_DESC_CNT )
+ if ((db->tx_insert_ptr->tdes0 & (cpu_to_le32(0x80000000))) == 0)
netif_wake_queue(dev); /* Active upper layer, send again */
}
@@ -1159,7 +1142,7 @@
db->interval_rx_cnt = 0;
/* TX polling kick monitor */
- if ( db->tx_packet_cnt &&
+ if ( db->tx_live_cnt > TX_IRQ_THR &&
time_after(jiffies, dev->trans_start + DMFE_TX_KICK) ) {
outl(0x1, dev->base_addr + DCR1); /* Tx polling again */
@@ -1167,13 +1150,13 @@
if ( time_after(jiffies, dev->trans_start + DMFE_TX_TIMEOUT) ) {
db->reset_TXtimeout++;
db->wait_reset = 1;
- printk(KERN_WARNING "%s: Tx timeout - resetting\n",
- dev->name);
+ printk(KERN_WARNING "%s: Tx timeout - resetting (%d)\n",
+ dev->name, db->tx_live_cnt);
}
}
if (db->wait_reset) {
- DMFE_DBUG(0, "Dynamic Reset device", db->tx_packet_cnt);
+ DMFE_DBUG(0, "Dynamic Reset device", db->tx_live_cnt);
db->reset_count++;
dmfe_dynamic_reset(dev);
db->first_in_callback = 0;
@@ -1271,8 +1254,8 @@
dmfe_free_rxbuffer(db);
/* system variable init */
- db->tx_packet_cnt = 0;
- db->tx_queue_cnt = 0;
+ db->tx_int_pkt_num = TX_IRQ_THR;
+ db->tx_live_cnt = 0;
db->rx_avail_cnt = 0;
db->link_failed = 1;
db->wait_reset = 0;
@@ -1464,6 +1447,12 @@
txptr = db->tx_insert_ptr;
suptr = (u32 *) txptr->tx_buf_ptr;
+ if (txptr->tdes0 & cpu_to_le32(0x80000000)) {
+ printk(KERN_WARNING "%s: Too busy to send filter frame\n",
+ dev->name);
+ return;
+ }
+
/* Node address */
addrptr = (u16 *) dev->dev_addr;
*suptr++ = addrptr[0];
@@ -1491,19 +1480,15 @@
/* prepare the setup frame */
db->tx_insert_ptr = txptr->next_tx_desc;
+ db->tx_live_cnt++;
txptr->tdes1 = cpu_to_le32(0x890000c0);
- /* Resource Check and Send the setup packet */
- if (!db->tx_packet_cnt) {
- /* Resource Empty */
- db->tx_packet_cnt++;
- txptr->tdes0 = cpu_to_le32(0x80000000);
- update_cr6(db->cr6_data | 0x2000, dev->base_addr);
- outl(0x1, dev->base_addr + DCR1); /* Issue Tx polling */
- update_cr6(db->cr6_data, dev->base_addr);
- dev->trans_start = jiffies;
- } else
- db->tx_queue_cnt++; /* Put in TX queue */
+ /* Send the setup frame */
+ dev->trans_start = jiffies; /* saved the time stamp */
+ txptr->tdes0 = cpu_to_le32(0x80000000);
+ update_cr6(db->cr6_data | 0x2000, dev->base_addr);
+ outl(0x1, dev->base_addr + DCR1); /* Issue Tx polling command */
+ update_cr6(db->cr6_data, dev->base_addr);
}
Hello.
How can i track down CRC Errors?
They are not very common, but happen some times with 2.4.x kernels.
Are they on the side of the boot loader or on the kernel?
I use LILO 21.4-4 to boot.
If they can be tracked, is there anything i can do to help to?
Thank you.
--
ICQ: 15605359 Bicho
=^..^=
First, they ignore you. Then they laugh at you. Then they fight you. Then
you win. Mahatma Gandhi.-------------------------------$B5$8!BN$N0lCW(B
------------------------------------ $B=k$54($5$bH`4_$^$G!#(B
$BNx$K$7$7$g$&$J$7!#Nx$O$7$"$s$NB>!#(B
$B%"%s(B $B%"%s(B $B%"%s(B $B$H$C$F$bBg9%$-(B