nexoniarz/SchOS
nexoniarz/SchOS
Watch
1
0
Fork
You've already forked SchOS
 0
Code Issues Pull requests Projects Releases 1 Packages Wiki Activity Actions

net: a network *class*, so the stack stops knowing what an e1000 is

Browse source 
The TCP/IP stack reached straight into the first driver anyone wrote.
net.c called e1000_present(), the lwIP glue was named netif_e1000.c and
called e1000_transmit(), and include/defs.h declared the driver's
internals to the whole kernel. The second NIC driver would therefore
have arrived with a near-identical copy of the glue file and a patch to
the protocol stack, which is the arrangement include/device.h was
written to prevent -- the e1000 even said so, with `.ops = 0, // no
class vtable yet; there is no net class`.

There is now. A driver publishes a struct netdev_ops -- transmit,
receive, mac, link_up -- and is finished: it never mentions lwIP, and
nothing above it learns its name. The display class was already this
shape; this is the same shape for networking, and the two now read the
same way in kernel/dev/device.c.

What that made possible immediately:

  - One netif per bound device instead of one per driver. lwIP's glue is
    netif_eth.c, it takes a struct device, and it holds up to NETIF_MAX
    interfaces. net.c attaches every network device it finds and starts
    DHCP on each, because which interface the network answers on is not
    ours to decide.
  - /proc/netif lists all of them, one block of fields each, with the
    interface name and the driver that claimed it.
  - The ARP self-test moved from the e1000 to the class. Nothing in it
    was chip-specific, and it is the first thing anyone will want to run
    against the next driver. Still `echo n > /proc/sysrq`.
  - Private headers live with their driver: e1000.h was in include/,
    where exactly one file could use it. The driver build now adds the
    driver's own directory to the include path.

The e1000 also drives more than one card. Its probe() refused the second
outright, because there was a single static instance -- so a machine
with two of them drove one and silently ignored the other. Instances are
an array now, and the instance is hung on dev->priv by a new optional
`attached` hook that the bus calls once the device exists. probe() runs
before device_add(), so there was previously nowhere to put it; that is
the whole reason the hook exists rather than a line inside probe.

QEMU gets an explicit NIC on every target (QEMU_NET). The default device
happens to be an e1000 on this machine type today, which is exactly the
sort of thing that stays true until a QEMU release changes it and the
network stops being exercised without anyone noticing.

Verified:
  - bootcheck -Net passes, 48 [ok], no failures -- including the moved
    self-test and two new checks that /proc/netif names both the
    interface and the driver behind it, which together say the stack
    went through the class rather than through a driver it was compiled
    to know.
  - Two NICs attached by hand (-device e1000 twice, second with its own
    MAC): eth0 and eth1 both attach, each reports its own MAC in
    /proc/netif, and each gets its own DHCP lease. Before this commit
    the same machine produced one interface.
  - make schfs-check still passes in full.

One trap worth recording: moving a header broke the build for anyone
with stale .d files -- make stopped with "No rule to make target
'include/e1000.h'". CFLAGS gains -MP, which emits a phony target per
header so a moved or deleted one can never wedge the build again.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This commit is contained in:
Nexoniarz 2026-08-15 16:42:34 +02:00
commit 20cc673a41
16 changed files with 729 additions and 367 deletions
Download patch file Download diff file Expand all files Collapse all files

152
kernel/net/lwip/netif_e1000.c
View file

@ -1,152 +0,0 @@
// SPDX-License-Identifier: Apache-2.0
//
// The lwIP network interface backed by the e1000 driver.
//
// Two directions and almost nothing else: a linkoutput that flattens a pbuf
// chain into a frame and hands it to the driver, and a poll that takes frames
// off the driver and hands them up. Everything about IP, ARP and TCP lives in
// lwIP; everything about descriptors and registers lives in the driver; this
// file is the seam.
#include "types.h"
#include "defs.h"
#include "param.h"
#include "lwip/opt.h"
#include "lwip/netif.h"
#include "lwip/pbuf.h"
#include "lwip/etharp.h"
#include "lwip/ethip6.h"
#include "netif/ethernet.h"
#define ETH_MTU 1500
#define ETH_FRAMESZ 1518 // 1500 payload + 14 header + 4 FCS allowance
static struct netif e1000_netif;
// Frames in and out, for /proc/netif. Counted here rather than in the driver
// because this is the seam lwIP actually uses: a frame the driver received and
// this file dropped never reached the stack, and a counter kept one layer down
// would say otherwise. Only the network thread touches them, so no lock.
static uint64 rx_frames, tx_frames;
void
net_e1000_counts(uint64 *rx, uint64 *tx)
{
*rx = rx_frames;
*tx = tx_frames;
}
// One staging buffer, not one per call: a frame is 1518 bytes and the kernel
// stack is 4096 in total, so a local of this size on a path that is several
// frames deep is exactly the overflow the kstack canary exists to catch.
// Safe as a single static because every caller holds the lwIP lock.
static char txframe[ETH_FRAMESZ];
static char rxframe[ETH_FRAMESZ];
// Send one packet. lwIP hands over a pbuf chain; the driver wants one flat
// buffer, so this walks the chain into txframe.
static err_t
e1000_linkoutput(struct netif *netif, struct pbuf *p)
{
(void)netif;
if (p->tot_len > sizeof(txframe))
return ERR_IF;
u16_t off = 0;
for (struct pbuf *q = p; q != NULL; q = q->next) {
if (off + q->len > sizeof(txframe))
return ERR_IF;
memmove(txframe + off, q->payload, q->len);
off += q->len;
}
// A full transmit ring is congestion, not failure. Reporting ERR_IF would
// make lwIP drop the segment and wait for a retransmission timer; ERR_WOULDBLOCK
// lets it try again, which for TCP is the difference between a stall of
// milliseconds and one of seconds.
if (e1000_transmit(txframe, off) != 0)
return ERR_WOULDBLOCK;
tx_frames++;
return ERR_OK;
}
// Bring the interface up. Called once, by the network thread.
static err_t
e1000_netif_init(struct netif *netif)
{
const unsigned char *mac = e1000_mac();
netif->name[0] = 'e';
netif->name[1] = 'n';
netif->output = etharp_output; // IP -> ARP -> linkoutput
netif->linkoutput = e1000_linkoutput;
netif->mtu = ETH_MTU;
netif->hwaddr_len = 6;
for (int i = 0; i < 6; i++)
netif->hwaddr[i] = mac[i];
netif->flags = NETIF_FLAG_BROADCAST | NETIF_FLAG_ETHARP | NETIF_FLAG_LINK_UP;
return ERR_OK;
}
struct netif *
net_e1000_netif(void)
{
return &e1000_netif;
}
// Attach the interface with no address yet -- DHCP fills it in.
int
net_e1000_attach(void)
{
ip4_addr_t any;
IP4_ADDR(&any, 0, 0, 0, 0);
if (netif_add(&e1000_netif, &any, &any, &any, NULL,
e1000_netif_init, netif_input) == NULL)
return -1;
netif_set_default(&e1000_netif);
netif_set_up(&e1000_netif);
netif_set_link_up(&e1000_netif);
return 0;
}
// Take whatever the driver has and push it into lwIP. Returns how many frames
// moved, so the caller can tell a busy poll from an idle one.
//
// Bounded rather than draining: a receive loop with no limit starves everything
// else on a busy network, and the caller comes straight back anyway.
int
net_e1000_poll(void)
{
int moved = 0;
for (int i = 0; i < 32; i++) {
int len = e1000_receive(rxframe, sizeof(rxframe));
if (len <= 0)
break;
// PBUF_POOL, so this comes from lwIP's own pool rather than a heap it
// does not have. A null return means the pool is exhausted, which is a
// dropped packet -- entirely normal under load, and what the protocol
// above is designed to survive.
struct pbuf *p = pbuf_alloc(PBUF_RAW, (u16_t)len, PBUF_POOL);
if (p == NULL)
break;
pbuf_take(p, rxframe, (u16_t)len);
// netif->input is netif_input, set at netif_add: it dispatches on the
// Ethernet type. Ownership of the pbuf passes with the call, including
// when it fails, so there is nothing to free here.
if (e1000_netif.input(p, &e1000_netif) != ERR_OK)
pbuf_free(p);
rx_frames++;
moved++;
}
return moved;
}

215
kernel/net/lwip/netif_eth.c Normal file
View file

@ -0,0 +1,215 @@
// SPDX-License-Identifier: Apache-2.0
//
// lwIP network interfaces over the netdev class.
//
// The seam, and it is deliberately driver-agnostic: everything about IP, ARP
// and TCP lives in lwIP, everything about descriptors and registers lives in a
// driver, and this file knows only that a network device can transmit a frame,
// receive a frame and state its MAC address. It was `netif_e1000.c` and called
// `e1000_transmit` directly, which meant the second NIC driver would have
// arrived with a second near-identical copy of this file. See include/netdev.h.
//
// One netif per bound interface, up to NETIF_MAX. Two NICs is an ordinary
// machine -- a laptop with Ethernet and the virtual one QEMU adds is already
// two -- and the limit exists only because a fixed array is simpler than
// allocation for something enumerated once at boot.
#include "types.h"
#include "defs.h"
#include "param.h"
#include "device.h"
#include "netdev.h"
#include "lwip/opt.h"
#include "lwip/netif.h"
#include "lwip/pbuf.h"
#include "lwip/etharp.h"
#include "lwip/ethip6.h"
#include "netif/ethernet.h"
#define ETH_MTU 1500
#define ETH_FRAMESZ 1518 // 1500 payload + 14 header + 4 FCS allowance
struct eth_if {
struct netif netif;
struct device *dev;
// Frames in and out, for /proc/netif. Counted at this seam rather than in
// the driver on purpose: a frame the driver received and this file dropped
// never reached the stack, and a counter one layer down would say it did.
uint64 rx_frames, tx_frames;
int used;
};
static struct eth_if ifs[NETIF_MAX];
static int nifs;
// One staging buffer each, not one per call: a frame is 1518 bytes and the
// kernel stack is 4096 in total, so a local of this size on a path that is
// several frames deep is exactly the overflow the kstack canary exists to
// catch. Safe as a single static because every caller holds the lwIP lock --
// which is also why one pair serves every interface.
static char txframe[ETH_FRAMESZ];
static char rxframe[ETH_FRAMESZ];
// Send one packet. lwIP hands over a pbuf chain; a driver wants one flat
// buffer, so this walks the chain into txframe.
static err_t
eth_linkoutput(struct netif *netif, struct pbuf *p)
{
struct eth_if *e = netif->state;
if (p->tot_len > sizeof(txframe))
return ERR_IF;
u16_t off = 0;
for (struct pbuf *q = p; q != NULL; q = q->next) {
if (off + q->len > sizeof(txframe))
return ERR_IF;
memmove(txframe + off, q->payload, q->len);
off += q->len;
}
// A full transmit ring is congestion, not failure. Reporting ERR_IF would
// make lwIP drop the segment and wait for a retransmission timer;
// ERR_WOULDBLOCK lets it try again, which for TCP is the difference between
// a stall of milliseconds and one of seconds.
if (netdev_transmit(e->dev, txframe, off) != 0)
return ERR_WOULDBLOCK;
e->tx_frames++;
return ERR_OK;
}
// Bring one interface up. Called by netif_add, once per device.
static err_t
eth_netif_init(struct netif *netif)
{
struct eth_if *e = netif->state;
const unsigned char *mac = netdev_mac(e->dev);
netif->name[0] = 'e';
netif->name[1] = 'n';
netif->output = etharp_output; // IP -> ARP -> linkoutput
netif->linkoutput = eth_linkoutput;
netif->mtu = ETH_MTU;
netif->hwaddr_len = 6;
for (int i = 0; i < 6; i++)
netif->hwaddr[i] = mac ? mac[i] : 0;
netif->flags = NETIF_FLAG_BROADCAST | NETIF_FLAG_ETHARP | NETIF_FLAG_LINK_UP;
return ERR_OK;
}
// Attach one network device. Returns its interface index, or -1.
//
// No address: DHCP fills that in, and an interface that came up with a
// hardcoded one would be wrong on every network but the one it was written on.
int
net_eth_attach(struct device *dev)
{
ip4_addr_t any;
if (nifs >= NETIF_MAX || dev == 0)
return -1;
struct eth_if *e = &ifs[nifs];
e->dev = dev;
e->used = 1;
IP4_ADDR(&any, 0, 0, 0, 0);
if (netif_add(&e->netif, &any, &any, &any, e,
eth_netif_init, netif_input) == NULL) {
e->used = 0;
return -1;
}
// The first one is the default route. With one interface that is the only
// possible answer; with two it is a policy decision that belongs to a
// routing table this kernel does not have yet, and picking the first is at
// least a rule rather than an accident of enumeration order.
if (nifs == 0)
netif_set_default(&e->netif);
netif_set_up(&e->netif);
if (netdev_link_up(dev))
netif_set_link_up(&e->netif);
else
netif_set_link_down(&e->netif);
return nifs++;
}
int
net_eth_count(void)
{
return nifs;
}
struct netif *
net_eth_netif(int i)
{
if (i < 0 || i >= nifs)
return 0;
return &ifs[i].netif;
}
struct device *
net_eth_device(int i)
{
if (i < 0 || i >= nifs)
return 0;
return ifs[i].dev;
}
void
net_eth_counts(int i, uint64 *rx, uint64 *tx)
{
if (i < 0 || i >= nifs) {
*rx = *tx = 0;
return;
}
*rx = ifs[i].rx_frames;
*tx = ifs[i].tx_frames;
}
// Take whatever the drivers have and push it into lwIP. Returns how many
// frames moved across all interfaces, so the caller can tell a busy poll from
// an idle one.
//
// Bounded per interface rather than draining: a receive loop with no limit
// starves everything else on a busy network, and the caller comes straight
// back anyway.
int
net_eth_poll(void)
{
int moved = 0;
for (int k = 0; k < nifs; k++) {
struct eth_if *e = &ifs[k];
for (int i = 0; i < 32; i++) {
int len = netdev_receive(e->dev, rxframe, sizeof(rxframe));
if (len <= 0)
break;
// PBUF_POOL, so this comes from lwIP's own pool rather than a heap
// it does not have. A null return means the pool is exhausted,
// which is a dropped packet -- entirely normal under load, and what
// the protocol above is designed to survive.
struct pbuf *p = pbuf_alloc(PBUF_RAW, (u16_t)len, PBUF_POOL);
if (p == NULL)
break;
pbuf_take(p, rxframe, (u16_t)len);
// netif->input is netif_input, set at netif_add: it dispatches on
// the Ethernet type. Ownership of the pbuf passes with the call,
// including when it fails, so there is nothing to free here.
if (e->netif.input(p, &e->netif) != ERR_OK)
pbuf_free(p);
e->rx_frames++;
moved++;
}
}
return moved;
}

149
kernel/net/net.c
View file

@ -4,11 +4,15 @@
//
// lwIP with NO_SYS=1 has no threads of its own, so something has to feed it
// received frames and run its timers. That something is a kernel thread rather
// than the e1000 interrupt handler, and the reason is the lock: a call into
// than a driver's interrupt handler, and the reason is the lock: a call into
// lwIP can reach the driver, the driver can sleep, and an interrupt handler
// cannot. So the interrupt keeps doing what it always did -- clear the cause
// register -- and this thread polls.
//
// Nothing here names a driver. Interfaces come from the netdev class
// (include/netdev.h): every bound network device is attached, and a machine
// with two NICs gets two interfaces without a line changing in this file.
//
// Polling costs a wakeup every few milliseconds when the link is idle. The
// alternative, waking this thread from the interrupt, is a worthwhile change
// later; it is deliberately not the first version, because a receive path that
@ -20,6 +24,8 @@
#include "spinlock.h"
#include "sleeplock.h"
#include "proc.h"
#include "device.h"
#include "netdev.h"
#include "lwip/opt.h"
#include "lwip/init.h"
@ -66,42 +72,63 @@ net_thread(void)
lwip_sys_lock();
lwip_init();
if (net_e1000_attach() != 0) {
// Every bound network device becomes an interface. DHCP is started on each
// of them: which one answers is the network's business, not ours, and an
// interface left unconfigured because it happened to be second would be a
// rule nobody could see.
int attached = 0;
for (int i = 0; i < netdev_count(); i++) {
struct device *d = netdev_at(i);
int idx = net_eth_attach(d);
if (idx < 0)
continue;
dhcp_start(net_eth_netif(idx));
attached++;
cprintf("net: %s attached as interface %d\n", d->name, idx);
}
if (attached == 0) {
lwip_sys_unlock();
cprintf("net: no e1000; the network thread has nothing to drive\n");
cprintf("net: no network device; the thread has nothing to drive\n");
// Not a panic and not an exit: a kernel thread must never return, and
// a machine with no NIC is an ordinary machine.
for (;;)
wait_tick();
}
struct netif *nif = net_e1000_netif();
dhcp_start(nif);
lwip_sys_unlock();
cprintf("net: lwIP %d.%d.%d up, requesting an address\n",
LWIP_VERSION_MAJOR, LWIP_VERSION_MINOR, LWIP_VERSION_REVISION);
cprintf("net: lwIP %d.%d.%d up on %d interface%s, requesting an address\n",
LWIP_VERSION_MAJOR, LWIP_VERSION_MINOR, LWIP_VERSION_REVISION,
attached, attached == 1 ? "" : "s");
int announced = 0;
// One bit per interface, so a second one binding later is announced too.
int announced[NETIF_MAX] = { 0 };
for (;;) {
int moved;
lwip_sys_lock();
moved = net_e1000_poll();
moved = net_eth_poll();
sys_check_timeouts();
// Report the address once, when it first arrives. Anyone who has spent
// Report each address once, when it first arrives. Anyone who has spent
// an evening wondering whether the guest has an address at all wants
// this on the console rather than in a file they have to know about.
if (!announced && !ip4_addr_isany_val(*netif_ip4_addr(nif))) {
for (int i = 0; i < attached; i++) {
struct netif *nif = net_eth_netif(i);
if (announced[i] || ip4_addr_isany_val(*netif_ip4_addr(nif)))
continue;
const ip4_addr_t *a = netif_ip4_addr(nif);
const ip4_addr_t *m = netif_ip4_netmask(nif);
const ip4_addr_t *gw = netif_ip4_gw(nif);
cprintf("net: %d.%d.%d.%d/%d.%d.%d.%d via %d.%d.%d.%d\n",
cprintf("net: %s %d.%d.%d.%d/%d.%d.%d.%d via %d.%d.%d.%d\n",
net_eth_device(i)->name,
ip4_addr1(a), ip4_addr2(a), ip4_addr3(a), ip4_addr4(a),
ip4_addr1(m), ip4_addr2(m), ip4_addr3(m), ip4_addr4(m),
ip4_addr1(gw), ip4_addr2(gw), ip4_addr3(gw), ip4_addr4(gw));
announced = 1;
announced[i] = 1;
net_bound = 1;
}
lwip_sys_unlock();
@ -183,65 +210,83 @@ put_field(char *out, const char *name)
return n;
}
// One block of fields per interface, separated by a blank line. A reader that
// can parse one can parse any number, which is the whole reason it is not a
// single flat list of values: with two NICs a flat list has no way to say which
// address belongs to which.
int
net_info(char *buf, int cap)
{
int len = 0;
if (cap < 512)
// Room for the largest block, so the loop below never needs a bounds check
// in the middle of writing one.
if (cap < 256 * (NETIF_MAX + 1))
return 0;
if (!net_started) {
if (!net_started || net_eth_count() == 0) {
len += put_field(buf + len, "state");
len += put_str(buf + len, e1000_present() ? "down\n" : "no interface\n");
len += put_str(buf + len, netdev_count() ? "down\n" : "no interface\n");
return len;
}
struct netif *nif = net_e1000_netif();
for (int i = 0; i < net_eth_count(); i++) {
struct netif *nif = net_eth_netif(i);
struct device *dev = net_eth_device(i);
lwip_sys_lock();
if (i)
buf[len++] = '\n';
len += put_field(buf + len, "interface");
buf[len++] = nif->name[0];
buf[len++] = nif->name[1];
len += put_u(buf + len, nif->num);
buf[len++] = '\n';
lwip_sys_lock();
len += put_field(buf + len, "mac");
len += put_mac(buf + len, nif->hwaddr);
buf[len++] = '\n';
len += put_field(buf + len, "interface");
len += put_str(buf + len, dev->name);
buf[len++] = '\n';
len += put_field(buf + len, "state");
len += put_str(buf + len, net_bound ? "up\n" : "no address\n");
len += put_field(buf + len, "driver");
len += put_str(buf + len, dev->driver ? dev->driver : "-");
buf[len++] = '\n';
len += put_field(buf + len, "address");
len += put_ip4(buf + len, netif_ip4_addr(nif));
buf[len++] = '\n';
len += put_field(buf + len, "mac");
len += put_mac(buf + len, nif->hwaddr);
buf[len++] = '\n';
len += put_field(buf + len, "netmask");
len += put_ip4(buf + len, netif_ip4_netmask(nif));
buf[len++] = '\n';
// Per interface, not per system: with two NICs one can be configured
// while the other is still asking.
len += put_field(buf + len, "state");
len += put_str(buf + len,
ip4_addr_isany_val(*netif_ip4_addr(nif)) ? "no address\n"
: "up\n");
len += put_field(buf + len, "gateway");
len += put_ip4(buf + len, netif_ip4_gw(nif));
buf[len++] = '\n';
len += put_field(buf + len, "address");
len += put_ip4(buf + len, netif_ip4_addr(nif));
buf[len++] = '\n';
len += put_field(buf + len, "mtu");
len += put_u(buf + len, nif->mtu);
buf[len++] = '\n';
len += put_field(buf + len, "netmask");
len += put_ip4(buf + len, netif_ip4_netmask(nif));
buf[len++] = '\n';
lwip_sys_unlock();
len += put_field(buf + len, "gateway");
len += put_ip4(buf + len, netif_ip4_gw(nif));
buf[len++] = '\n';
uint64 rx, tx;
net_e1000_counts(&rx, &tx);
len += put_field(buf + len, "mtu");
len += put_u(buf + len, nif->mtu);
buf[len++] = '\n';
len += put_field(buf + len, "rx_frames");
len += put_u(buf + len, rx);
buf[len++] = '\n';
lwip_sys_unlock();
len += put_field(buf + len, "tx_frames");
len += put_u(buf + len, tx);
buf[len++] = '\n';
uint64 rx, tx;
net_eth_counts(i, &rx, &tx);
len += put_field(buf + len, "rx_frames");
len += put_u(buf + len, rx);
buf[len++] = '\n';
len += put_field(buf + len, "tx_frames");
len += put_u(buf + len, tx);
buf[len++] = '\n';
}
return len;
}
@ -258,9 +303,9 @@ netinit(void)
// depends on that reasoning holding forever is not worth the two lines
// saved.
inet_init();
if (!e1000_present()) {
// Silent: a machine without an e1000 is not misconfigured, and this
// runs on every boot.
if (netdev_count() == 0) {
// Silent: a machine with no network device is not misconfigured, and
// this runs on every boot.
return;
}
net_started = 1;