Two NICs, and unlike the RTL8111 last commit, QEMU emulates both -- so both are tested rather than reasoned about. **RTL8139** is its own driver and not a mode of the 8169, because it is its own chip. No descriptor rings: the receiver is one circular buffer the chip writes packets into back to back behind four-byte headers, and the transmitter is four fixed slots. Putting both programming models behind one set of function pointers is how a mode flag ends up threaded through every path. Three details are the whole driver. CAPR runs sixteen bytes behind the true read position -- a quirk of the part, not an off-by-one. WRAP means the chip runs a packet past the end of the buffer rather than splitting it, so the 8 KB buffer is allocated as 8 KB + 16 + one maximum frame; getting that wrong is a chip writing past its allocation at line rate. And a length field of 0xFFF0 means "still writing this one", so treating it as a length copies a partial packet and loses the alignment for everything after it. **The e1000 now matches the PCIe 8257x parts.** Its header refused them, with a note saying they "differ in NVM access and initialisation, and claiming a device this driver has not been shown to initialise is worse than not claiming it". That was right, and the second half is answerable: QEMU emulates the 82574L, so it can be *shown* to initialise. What differs turned out to be one register. EERD moved its address field from bits 8-15 to bits 2-15 and its done bit from 4 to 1; the rings, RCTL, TCTL and ICR are identical. So there is a per-instance flag for the NVM layout and seven more ids -- not a second copy of four hundred lines of ring code. The wrong pair of constants fails silently, reading the wrong word or waiting for a bit that never sets, and the MAC comes out as zeroes; that is why it is named in the header rather than papered over. Still refused: the ICH/PCH parts (82562, 82577-82579, I217-I219), which put the PHY behind a MAC-PHY interconnect and need a different bring-up nothing here can test. Tested, each booted alone: rtl8139 (10ec:8139) and 82574L (8086:10d3) both bind, read a real MAC, complete DHCP and ping 4 of 4 with no loss. The classic 82540EM does the same, which is the regression that mattered for the NVM change. Both supported display configurations: one boot each, no traps, no panics, socktest and memfdtest green. Also recorded: the Chromebox CN62 carries an RTL8111, not an Intel part, so last commit's driver covers it too. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
127 lines
4.9 KiB
C
127 lines
4.9 KiB
C
// The driver registry.
|
|
//
|
|
// One table, one line per driver. This is the same shape as
|
|
// kernel/sys/table.inc and kernel/dev/cdev.c's devsw: a single place that
|
|
// lists what exists, so a driver cannot be half-registered.
|
|
//
|
|
// It is a plain array rather than the linker-section trick a bigger kernel
|
|
// would use (a __attribute__((section(".drivers"))) entry per driver, walked
|
|
// between __start_ and __stop_ symbols). That approach lets a driver register
|
|
// itself by existing, which is genuinely nicer -- but it also means the list
|
|
// of what is in your kernel is not written down anywhere you can read, and
|
|
// a driver silently dropped by the linker looks exactly like one that was
|
|
// never written. At this size, being able to read the list wins.
|
|
//
|
|
// To add a driver: write kernel/drivers/<name>.c exporting a struct
|
|
// pci_driver, declare it here, and add it to the array and to the Makefile's
|
|
// KERN_DRIVERS.
|
|
|
|
#include "types.h"
|
|
#include "defs.h"
|
|
#include "param.h"
|
|
#include "spinlock.h"
|
|
#include "sleeplock.h"
|
|
#include "fs.h"
|
|
#include "file.h"
|
|
#include "pci.h"
|
|
|
|
#include "usb.h"
|
|
#include "device.h"
|
|
|
|
extern const struct pci_driver vga_pci_driver;
|
|
extern const struct pci_driver xhci_driver;
|
|
extern const struct pci_driver virtio_gpu_driver;
|
|
extern const struct pci_driver e1000_driver;
|
|
extern const struct pci_driver ac97_driver;
|
|
extern const struct pci_driver r8169_driver;
|
|
extern const struct pci_driver rtl8139_driver;
|
|
|
|
// Order matters here in one respect: virtio-gpu is offered the bus before the
|
|
// VBE framebuffer driver, so that on a machine with both, the device with real
|
|
// mode setting and an explicit flush wins the display class. vga_pci still
|
|
// claims the stdvga card underneath, which is what the boot splash and the
|
|
// text console are drawn on.
|
|
static const struct pci_driver *drivers[] = {
|
|
&virtio_gpu_driver,
|
|
&vga_pci_driver,
|
|
&e1000_driver,
|
|
&r8169_driver,
|
|
&rtl8139_driver,
|
|
&xhci_driver,
|
|
&ac97_driver,
|
|
};
|
|
|
|
void
|
|
pci_bind_all(void)
|
|
{
|
|
for (unsigned i = 0; i < sizeof(drivers) / sizeof(drivers[0]); i++)
|
|
pci_bind(drivers[i]);
|
|
}
|
|
|
|
// USB class drivers, same arrangement and for the same reason. These are not
|
|
// found on the PCI bus -- they are offered a device by the USB core once its
|
|
// interface descriptor has been read -- but a driver you cannot find in a
|
|
// list is a driver nobody knows is there, whichever bus it sits on.
|
|
//
|
|
// To add one: write kernel/drivers/usb_<name>.c exporting a struct
|
|
// usb_driver, declare it here, add it to the array and to KERN_DRIVERS.
|
|
|
|
extern const struct usb_driver usb_hid_driver;
|
|
extern const struct usb_driver usb_storage_driver;
|
|
extern const struct usb_driver usb_hub_driver;
|
|
|
|
static const struct usb_driver *usb_drivers[] = {
|
|
&usb_hid_driver,
|
|
&usb_storage_driver,
|
|
&usb_hub_driver,
|
|
};
|
|
|
|
int usb_offer(const struct usb_driver *, struct usb_dev *);
|
|
|
|
void
|
|
usb_bind_all(struct usb_dev *dev)
|
|
{
|
|
for (unsigned i = 0; i < sizeof(usb_drivers) / sizeof(usb_drivers[0]); i++)
|
|
if (usb_offer(usb_drivers[i], dev)) {
|
|
// Same as the PCI path: publish it in the bus-agnostic list so that
|
|
// "what is attached" has one answer covering every bus. Class comes from
|
|
// the driver that claimed it rather than the USB class code, because a
|
|
// driver knows what it made of the device and the descriptor only says
|
|
// what the device claims to be.
|
|
const char *n = usb_drivers[i]->name;
|
|
struct device *d = device_add(n, DEVBUS_USB, usb_drivers[i]->devclass, n);
|
|
if (d)
|
|
device_set_desc(d, usb_drivers[i]->name);
|
|
return;
|
|
}
|
|
}
|
|
|
|
// Devices that are not enumerated by any bus -- they are simply known to be
|
|
// there on a PC, and the driver pokes fixed I/O ports to find out whether it
|
|
// is right. Listing them matters because a device manager that shows the
|
|
// graphics card but not the keyboard is not showing the machine.
|
|
//
|
|
// Registered here rather than inside each driver's init so that the whole set
|
|
// is visible in one place, which is the same reasoning as the driver arrays
|
|
// above.
|
|
void
|
|
platform_register_devices(uint64 fb_addr)
|
|
{
|
|
struct device *d;
|
|
|
|
// The framebuffer first, before anything that might want to draw. It is a
|
|
// driver call, and this file is where driver names belong -- main() used to
|
|
// call vga_init() directly, which put one display driver's name in the
|
|
// kernel's startup path and would have needed editing to boot on a machine
|
|
// whose framebuffer came from anywhere else.
|
|
vga_fb_init(fb_addr);
|
|
|
|
if ((d = device_add("kbd", DEVBUS_PLATFORM, DEVCLASS_INPUT, "ps2kbd")))
|
|
device_set_desc(d, "PS/2 keyboard");
|
|
if ((d = device_add("mouse", DEVBUS_PLATFORM, DEVCLASS_INPUT, "ps2mouse")))
|
|
device_set_desc(d, "PS/2 mouse");
|
|
if ((d = device_add("uart", DEVBUS_PLATFORM, DEVCLASS_OTHER, "uart8250")))
|
|
device_set_desc(d, "16550 serial port");
|
|
if ((d = device_add("ide", DEVBUS_PLATFORM, DEVCLASS_STORAGE, "ide")))
|
|
device_set_desc(d, "ATA disk controller");
|
|
}
|