Re: rpi3b+ ue0 too late to be configured on boot?
- In reply to: Mark Millard : "Re: rpi3b+ ue0 too late to be configured on boot?"
- Go to: [ bottom of page ] [ top of archives ] [ this month ]
Date: Thu, 29 Dec 2022 21:01:15 UTC
On Dec 29, 2022, at 12:38, Mark Millard <marklmi@yahoo.com> wrote: > On Dec 29, 2022, at 11:39, Mike Karels <mike@karels.net> wrote: > >> On 28 Dec 2022, at 16:42, Bjoern A. Zeeb wrote: >> >>> Hi, >>> >>> does anyone else have the problem on latest main that the ue0 on an >>> rpi3b+ doesn't show up in time for the rc netif to configure it (or at >>> least it looks like)? >> >> I don’t see that problem. I tried the 12/24 snapshot of main on an >> RPi3B+, and ue0 configured via DHCP on the initial boot, and on a reboot. > > Are you also using EDK2's UEFI via the boot materials from: > > https://github.com/pftf/RPi3/releases/tag/v1.38 > > instead of the using the normal FreeBSD ports used in official > builds (the RPi firmware and a U-Boot)? The "pfpt" RPi3 materials > include both RPi* firmware and the EDK2 based material. > > Bjoern did not report enough information for the configuration > of that EDK2 based context to replicate his boot context in a > test. > > I'll note that the intructions for that EDK2 variant indicate > that linux should not use ACPI but instead should use the > Device Tree. As I remember, the ACPI is documented somewhere > to follow some non-standard Microsoft specifics. (The RPi4 > "pfpt" does not have the same issue.) Some related quotes > from https://github.com/pftf/RPi3 are: > > QUOTE > With recent Linux installs, please assure that the > firmware is running in DT mode, either via > "Device Manager"->"Raspberry Pi Configuration" > ->"Advanced Configuration"->"System Table Selection" > or the Linux/Grub command line with "acpi=off". > END QUOTE > I found the text about not being ACPI compliant: https://github.com/tianocore/edk2-platforms/tree/master/Platform/RaspberryPi/RPi3 reports: QUOTE ACPI . . . Note that the ACPI tables were derived or copied from the MS-IoT one. This means that they are not truly ACPI compliant, especially when it comes to their descriptors, and therefore not suitable for Linux environments. If you want to use a Linux HLOS, you are encouraged tp install a kernel that relies on Device Tree rather than ACPI. END QUOTE >> I did have problems with the HDMI console, some of which are new; I will >> follow up on that separately. Specifically, USB keyboard input didn’t >> work to the loader, and I didn’t see console output from user-level >> startup until I disabled boot_multicons and boot_serial. The second >> problem is relatively new. >> >> Mike >> >>> While during boot USB seems to need a wait-delay now, which it hadn't in >>> the last years for me? >>> >>> ------------------------------------------------------------------------ >>> ... >>> Feeding entropy: . >>> ELF ldconfig path: /lib /usr/lib >>> ugen1.2: <vendor 0x0424 product 0x9514> at usbus1 >>> uhub1 on uhub0 >>> uhub1: <vendor 0x0424 product 0x9514, class 9/0, rev 2.00/2.00, addr 2> on usbus1 >>> uhub1: MTT enabled >>> lo0: link state changed to UP >>> uhub1: 5 ports with 4 removable, self powered >>> ugen1.3: <vendor 0x0424 product 0xec00> at usbus1 >>> smsc0 on uhub1 >>> smsc0: <vendor 0x0424 product 0xec00, rev 2.00/2.00, addr 3> on usbus1 >>> smsc0: chip 0xec00, rev. 0002 >>> miibus0: <MII bus> on smsc0 >>> smscphy0: <SMC LAN8700 10/100 interface> PHY 1 on miibus0 >>> smscphy0: OUI 0x00800f, model 0x000c, rev. 3 >>> smscphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto >>> ue0: <USB Ethernet> on smsc0 >>> ue0: bpf attached >>> ue0: Ethernet address: xx:xx:xx:xx:xx:xx >>> Starting Network: lo0. >>> lo0: flags=8049<UP,LOOPBACK,RUNNING,MULTICAST> metric 0 mtu 16384 >>> ... >>> ------------------------------------------------------------------------ >>> >>> If running manually after boot (again) it all works fine? >>> >>> # sh /etc/rc.d/netif start ue0 >>> smsc0: chip 0xec00, rev. 0002 >>> Starting Network: ue0. >>> ue0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500 >>> ... >>> >>> >>> >>> Not that it should matter - this is with https://github.com/pftf/RPi3/releases/tag/v1.38 >>> >>> /bz >>> >>> -- >>> Bjoern A. Zeeb r15:7 >> > === Mark Millard marklmi at yahoo.com