[190] in linux-net channel archive
Patch for Lance driver: multicasts, graceful restart, etc.
daemon@ATHENA.MIT.EDU (Dave Platt)
Mon Apr 10 21:59:46 1995
From: dplatt@3do.com (Dave Platt)
Date: Mon, 10 Apr 1995 17:22:53 PDT
To: linux-net@vger.rutgers.edu
Cc: becker@cesdis1.gsfc.nasa.gov, lau@eniac.seas.upenn.edu
Enclosed herein is a patch for the Lance Ethernet driver in Linux 1.2.4.
This patch can probably be applied to most recent versions of this
driver without difficulty.
The primary intent of this patch is to address a number of conditions
which can cause modern (79cXXX) versions of the AMD chip family to hang.
Most of these hangs arise from one crucial difference between the
original 7990 LANCE chip and the newer PCnet chips:
- When the 7990 chip is stopped, and restarted without
reinitialization, it retains the current transmit/receive ring
pointer settings.
- When PCnet chips are stopped and restarted, they always reset/reload
the ring pointer settings.
The lance.c driver as released, assumes the former behavior. If a PCnet
chip is stopped and restarted (due to a Tx FIFO error, a busmaster
arbitration error, or to set or clear multicast or promiscuous modes)
the PCnet chip and the lance.c driver would often have their ring
pointers out of sync - this usually leads to a complete paralysis of the
interface.
The enclosed patch addresses this problem, and also cleans up the code
in a couple of minor ways.
[1] The various quirks of different LANCE/PCnet chips are now described
as individual flags in the chip_table. All code which tests for
chip-specific behavior does so by checking bits, rather than by
testing the chip version number.
[2] A new routine has been added to purge the transmit ring during a
chip restart. This code walks the ring, locates all descriptors
which have skbuffs associated, and frees the skbuffs (in effect,
dropping these packets on the floor without transmitting them).
This keeps the higher-level protocol code from hanging, and allows
it to time out and retransmit the packets after the interface is
back up.
[3] A new lance_restart() routine has been added. This routine checks
to see if the chip will reset its ring buffers when restarted - if
so, it purges the transmit ring and reinitializes both rings.
[4] All code which had previously contained "outw" statements to restart
the chip, has been modified to use lance_restart().
The net effect of these patches is that the chip/card/driver recover
gracefully from busmaster arbitration errors, Tx FIFO underruns, and
set_multicast_address calls.
I've tested these changes on my current configuration (ASUS SP3G
motherboard, HP J2405A card with 79c960 PCInet chip). Without the
patches, I can wedge the interface within a minute or so by doing
tar cvf - . | rsh remotehost dd of=/dev/null
It looks as if the NCR 53c810's PCI busmastering is locking out the ISA
DMA channel busmaster for excessive periods (> 50 uS). With the patch
in place, my system keeps running and the commands do not hang, even
though the error log is spitting several "Bus master arbitrarion
failure" and "Tx FIFO error" messages per minute.
If you have a LANCE/PCnet-based Ethernet card (AT1500, NE2100, Boca
BEN1PI, etc.) I'd appreciate it very much if you could test this patch
and let me know whether it helps or hurts your system's operation. If
nobody complains within the next few days I'll send it to Linus.
--- linux-1.2.4/drivers/net/lance.c Fri Jan 20 01:34:39 1995
+++ linux/drivers/net/lance.c Sat Apr 8 16:54:08 1995
@@ -15,7 +15,7 @@
Code 930.5, Goddard Space Flight Center, Greenbelt MD 20771
*/
-static char *version = "lance.c:v1.07 1/18/95 becker@cesdis.gsfc.nasa.gov\n";
+static char *version = "lance.c:v1.08 4/10/95 dplatt@3do.com\n";
#include <linux/config.h>
#include <linux/kernel.h>
@@ -203,12 +203,18 @@
int dirty_rx, dirty_tx; /* The ring entries to be free()ed. */
int dma;
struct enet_statistics stats;
- char chip_version; /* See lance_chip_type. */
+ unsigned char chip_version; /* See lance_chip_type. */
char tx_full;
char lock;
int pad0, pad1; /* Used for 8-byte alignment */
};
+#define LANCE_MUST_PAD 0x00000001
+#define LANCE_ENABLE_AUTOSELECT 0x00000002
+#define LANCE_MUST_REINIT_RING 0x00000004
+#define LANCE_MUST_UNRESET 0x00000008
+#define LANCE_HAS_MISSED_FRAME 0x00000010
+
/* A mapping from the chip ID number to the part number and features.
These are from the datasheets -- in real life the '970 version
reportedly has the same ID as the '965. */
@@ -217,14 +223,25 @@
char *name;
int flags;
} chip_table[] = {
- {0x0000, "LANCE 7990", 0}, /* Ancient lance chip. */
- {0x0003, "PCnet/ISA 79C960", 0}, /* 79C960 PCnet/ISA. */
- {0x2260, "PCnet/ISA+ 79C961", 0}, /* 79C961 PCnet/ISA+, Plug-n-Play. */
- {0x2420, "PCnet/PCI 79C970", 0}, /* 79C970 or 79C974 PCnet-SCSI, PCI. */
+ {0x0000, "LANCE 7990", /* Ancient lance chip. */
+ LANCE_MUST_PAD + LANCE_MUST_UNRESET},
+ {0x0003, "PCnet/ISA 79C960", /* 79C960 PCnet/ISA. */
+ LANCE_ENABLE_AUTOSELECT + LANCE_MUST_REINIT_RING +
+ LANCE_HAS_MISSED_FRAME},
+ {0x2260, "PCnet/ISA+ 79C961", /* 79C961 PCnet/ISA+, Plug-n-Play. */
+ LANCE_ENABLE_AUTOSELECT + LANCE_MUST_REINIT_RING +
+ LANCE_HAS_MISSED_FRAME},
+ {0x2420, "PCnet/PCI 79C970", /* 79C970 or 79C974 PCnet-SCSI, PCI. */
+ LANCE_ENABLE_AUTOSELECT + LANCE_MUST_REINIT_RING +
+ LANCE_HAS_MISSED_FRAME},
/* Bug: the PCnet/PCI actually uses the PCnet/VLB ID number, so just call
it the PCnet32. */
- {0x2430, "PCnet32", 0}, /* 79C965 PCnet for VL bus. */
- {0x0, "PCnet (unknown)", 0},
+ {0x2430, "PCnet32", /* 79C965 PCnet for VL bus. */
+ LANCE_ENABLE_AUTOSELECT + LANCE_MUST_REINIT_RING +
+ LANCE_HAS_MISSED_FRAME},
+ {0x0, "PCnet (unknown)",
+ LANCE_ENABLE_AUTOSELECT + LANCE_MUST_REINIT_RING +
+ LANCE_HAS_MISSED_FRAME},
};
enum {OLD_LANCE = 0, PCNET_ISA=1, PCNET_ISAP=2, PCNET_PCI=3, PCNET_VLB=4, LANCE_UNKNOWN=5};
@@ -256,8 +273,6 @@
int *port;
#if defined(CONFIG_PCI)
-#define AMD_VENDOR_ID 0x1022
-#define AMD_DEVICE_ID 0x2000
if (pcibios_present()) {
int pci_index;
printk("lance.c: PCI bios is present, checking for devices...\n");
@@ -266,7 +281,8 @@
unsigned long pci_ioaddr;
unsigned short pci_command;
- if (pcibios_find_device (AMD_VENDOR_ID, AMD_DEVICE_ID, pci_index,
+ if (pcibios_find_device (PCI_VENDOR_ID_AMD,
+ PCI_DEVICE_ID_AMD_LANCE, pci_index,
&pci_bus, &pci_device_fn) != 0)
break;
pcibios_read_config_byte(pci_bus, pci_device_fn,
@@ -522,7 +538,7 @@
}
}
- if (lp->chip_version != OLD_LANCE) {
+ if (chip_table[lp->chip_version].flags & LANCE_ENABLE_AUTOSELECT) {
/* Turn on auto-select of media (10baseT or BNC) so that the user
can watch the LEDs even if the board isn't opened. */
outw(0x0002, ioaddr+LANCE_ADDR);
@@ -570,10 +586,10 @@
}
/* Un-Reset the LANCE, needed only for the NE2100. */
- if (lp->chip_version == OLD_LANCE)
+ if (chip_table[lp->chip_version].flags & LANCE_MUST_UNRESET)
outw(0, ioaddr+LANCE_RESET);
- if (lp->chip_version != OLD_LANCE) {
+ if (chip_table[lp->chip_version].flags & LANCE_ENABLE_AUTOSELECT) {
/* This is 79C960-specific: Turn on auto-select of media (AUI, BNC). */
outw(0x0002, ioaddr+LANCE_ADDR);
outw(0x0002, ioaddr+LANCE_BUS_IF);
@@ -617,6 +633,33 @@
return 0; /* Always succeed */
}
+/* The LANCE has been halted for one reason or another (busmaster memory
+ arbitration error, Tx FIFO underflow, driver stopped it to reconfigure,
+ etc.). Modern LANCE variants always reload their ring-buffer
+ configuration when restarted, so we must reinitialize our ring
+ context before restarting. As part of this reinitialization,
+ find all packets still on the Tx ring and pretend that they had been
+ sent (in effect, drop the packets on the floor) - the higher-level
+ protocols will time out and retransmit. It'd be better to shuffle
+ these skbs to a temp list and then actually re-Tx them after
+ restarting the chip, but I'm too lazy to do so right now. dplatt@3do.com
+*/
+
+static void
+lance_purge_tx_ring(struct device *dev)
+{
+ struct lance_private *lp = (struct lance_private *)dev->priv;
+ int i;
+
+ for (i = 0; i < TX_RING_SIZE; i++) {
+ if (lp->tx_skbuff[i]) {
+ dev_kfree_skb(lp->tx_skbuff[i],FREE_WRITE);
+ lp->tx_skbuff[i] = NULL;
+ }
+ }
+}
+
+
/* Initialize the LANCE Rx and Tx rings. */
static void
lance_init_ring(struct device *dev)
@@ -647,12 +690,27 @@
lp->init_block.tx_ring = (int)lp->tx_ring | TX_RING_LEN_BITS;
}
+static void
+lance_restart(struct device *dev, unsigned int csr0_bits, int must_reinit)
+{
+ struct lance_private *lp = (struct lance_private *)dev->priv;
+
+ if (must_reinit ||
+ (chip_table[lp->chip_version].flags & LANCE_MUST_REINIT_RING)) {
+ lance_purge_tx_ring(dev);
+ lance_init_ring(dev);
+ }
+ outw(0x0000, dev->base_addr + LANCE_ADDR);
+ outw(csr0_bits, dev->base_addr + LANCE_DATA);
+}
+
static int
lance_start_xmit(struct sk_buff *skb, struct device *dev)
{
struct lance_private *lp = (struct lance_private *)dev->priv;
int ioaddr = dev->base_addr;
int entry;
+ unsigned long flags;
/* Transmitter timeout, serious problems. */
if (dev->tbusy) {
@@ -681,8 +739,7 @@
printk("\n");
}
#endif
- lance_init_ring(dev);
- outw(0x0043, ioaddr+LANCE_DATA);
+ lance_restart(dev, 0x0043, 1);
dev->tbusy=0;
dev->trans_start = jiffies;
@@ -728,7 +785,7 @@
with the "ownership" bits last. */
/* The old LANCE chips doesn't automatically pad buffers to min. size. */
- if (lp->chip_version == OLD_LANCE) {
+ if (chip_table[lp->chip_version].flags & LANCE_MUST_PAD) {
lp->tx_ring[entry].length =
-(ETH_ZLEN < skb->len ? skb->len : ETH_ZLEN);
} else
@@ -758,13 +815,14 @@
dev->trans_start = jiffies;
+ save_flags(flags);
cli();
lp->lock = 0;
if (lp->tx_ring[(entry+1) & TX_RING_MOD_MASK].base == 0)
dev->tbusy=0;
else
lp->tx_full = 1;
- sti();
+ restore_flags(flags);
return 0;
}
@@ -776,6 +834,7 @@
struct device *dev = (struct device *)(irq2dev_map[irq]);
struct lance_private *lp;
int csr0, ioaddr, boguscnt=10;
+ int must_restart;
if (dev == NULL) {
printk ("lance_interrupt(): irq %d for unknown device.\n", irq);
@@ -795,6 +854,8 @@
/* Acknowledge all of the current interrupt sources ASAP. */
outw(csr0 & ~0x004f, dev->base_addr + LANCE_DATA);
+ must_restart = 0;
+
if (lance_debug > 5)
printk("%s: interrupt csr0=%#2.2x new csr=%#2.2x.\n",
dev->name, csr0, inw(dev->base_addr + LANCE_DATA));
@@ -828,7 +889,7 @@
printk("%s: Tx FIFO error! Status %4.4x.\n",
dev->name, csr0);
/* Restart the chip. */
- outw(0x0002, dev->base_addr + LANCE_DATA);
+ must_restart = 1;
}
} else {
if (status & 0x18000000)
@@ -871,7 +932,14 @@
printk("%s: Bus master arbitration failure, status %4.4x.\n",
dev->name, csr0);
/* Restart the chip. */
- outw(0x0002, dev->base_addr + LANCE_DATA);
+ must_restart = 1;
+ }
+
+ if (must_restart) {
+ /* stop the chip to clear the error condition, then restart */
+ outw(0x0000, dev->base_addr + LANCE_ADDR);
+ outw(0x0004, dev->base_addr + LANCE_DATA);
+ lance_restart(dev, 0x0002, 0);
}
}
@@ -961,7 +1029,7 @@
dev->start = 0;
dev->tbusy = 1;
- if (lp->chip_version != OLD_LANCE) {
+ if (chip_table[lp->chip_version].flags & LANCE_HAS_MISSED_FRAME) {
outw(112, ioaddr+LANCE_ADDR);
lp->stats.rx_missed_errors = inw(ioaddr+LANCE_DATA);
}
@@ -991,14 +1059,16 @@
struct lance_private *lp = (struct lance_private *)dev->priv;
short ioaddr = dev->base_addr;
short saved_addr;
+ unsigned long flags;
- if (lp->chip_version != OLD_LANCE) {
+ if (chip_table[lp->chip_version].flags & LANCE_HAS_MISSED_FRAME) {
+ save_flags(flags);
cli();
saved_addr = inw(ioaddr+LANCE_ADDR);
outw(112, ioaddr+LANCE_ADDR);
lp->stats.rx_missed_errors = inw(ioaddr+LANCE_DATA);
outw(saved_addr, ioaddr+LANCE_ADDR);
- sti();
+ restore_flags(flags);
}
return &lp->stats;
@@ -1015,11 +1085,9 @@
{
short ioaddr = dev->base_addr;
- /* We take the simple way out and always enable promiscuous mode. */
outw(0, ioaddr+LANCE_ADDR);
outw(0x0004, ioaddr+LANCE_DATA); /* Temporarily stop the lance. */
- outw(15, ioaddr+LANCE_ADDR);
if (num_addrs >= 0) {
short multicast_table[4];
int i;
@@ -1029,15 +1097,17 @@
outw(8 + i, ioaddr+LANCE_ADDR);
outw(multicast_table[i], ioaddr+LANCE_DATA);
}
+ outw(15, ioaddr+LANCE_ADDR);
outw(0x0000, ioaddr+LANCE_DATA); /* Unset promiscuous mode */
} else {
/* Log any net taps. */
printk("%s: Promiscuous mode enabled.\n", dev->name);
+ outw(15, ioaddr+LANCE_ADDR);
outw(0x8000, ioaddr+LANCE_DATA); /* Set promiscuous mode */
}
- outw(0, ioaddr+LANCE_ADDR);
- outw(0x0142, ioaddr+LANCE_DATA); /* Resume normal operation. */
+ lance_restart(dev, 0x0142, 0); /* Resume normal operation */
+
}
