Manuel Bouyer <bouyer@antioche.eu.org> writes:

> This happened at one time when the disk didn't have time to flush its cache.
> But for some time now a cache flush is issued at shutdown time, and
> there should be a 2s delay bewteen the last kernel action an poweroff.

It looks far faster than 2s to me, even with today's software.  I see
delay(500000) in machdep.c:cpu_reboot.  I wonder if that doesn't work
on my machine for some reason - I'll try bumping it up.

>> The builtin wm(4) sort of works, but PHY programming is clearly wrong.
>> I have to set media manually to 10baseT and then it will work; even
>> then after setting to 100baseTX my switch reports it's still at 10.
>
> What happens if you set it back to auto ?
> Can you send the dmesg ?

At auto, I get 100/FD on the switch (Linksys EZXS88W 10/100, never had
trouble with anything else), and no connectivity.  (I'll denote FD vs
- (no FD light)).  I ran tcpblast and am pretty sure the hub 10/100
light is accurate.  Note that this is a sequence; a given media
command doesn't always result in the particular behavior.

media 10baseT: 10/- works
media 10baseT mediaopt full-duplex: 10/- works
media 100baseTX: 10/- works
media 100baseTX mediaopt full-duplex: 10/- works
media autoselect: 100/FD FAILURE
media 10baseT: 10/- works

I have used the interface with auto with a Thinkpad R51 wm(4) and that
worked at 1000baseT.

wm0 at pci2 dev 0 function 0: Intel i82573L Gigabit Ethernet, rev. 0
wm0: interrupting at ioapic0 pin 16 (irq 11)
wm0: PCI-Express bus
wm0: 256 word (8 address bits) SPI EEPROM
wm0: Ethernet address 00:15:58:[redacted]
makphy0 at wm0 phy 1: Marvell 88E1111 Gigabit PHY, rev. 2
makphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-FDX, auto

pcictl reports (-n/regular):

002:00:0: 0x109a8086 (0x2000000)
002:00:0: Intel i82573L Gigabit Ethernet (ethernet network)

The machine is a 2613-ECU (T60 with T2600 Core Duo, 1400x1050 screen)
and 2 GB of RAM.