Re: RPI 4/5 u-boot new port, is there any interest? [16 GiByte RPi5B (D0) basics operational via the EDK2 variant for such, NMVe booting too]

On Jun 6, 2025, at 08:39, Bjoern A. Zeeb <[email protected]> wrote:

> On Thu, 5 Jun 2025, Mark Millard wrote:
> 
> Can you share a
> 
> pciconf -lv
> 
> from when the RPi5 is booted?

Sure:

# pciconf -lv
pcib1@pci1:0:0:0:	class=0x060400 rev=0x30 hdr=0x01 vendor=0x14e4 device=0x2712 subvendor=0x0000 subdevice=0x0000
    vendor     = 'Broadcom Inc. and subsidiaries'
    device     = 'BCM2712 PCIe Bridge'
    class      = bridge
    subclass   = PCI-PCI
nvme0@pci1:1:0:0:	class=0x010802 rev=0x01 hdr=0x00 vendor=0x10ec device=0x5765 subvendor=0x10ec subdevice=0x5765
    vendor     = 'Realtek Semiconductor Co., Ltd.'
    device     = 'RTS5765DL NVMe SSD Controller (DRAM-less)'
    class      = mass storage
    subclass   = NVM


FYI for folks that might read for more general interest:

The system is normally used headless and has been building
ports into packages via poudriere-devel for a little over
12 hours so far. The configuration is a high load average
style of building with lots of SWAP space (62 GiBytes, so
RAM+SWAP==78 GiBytes) and USE_TMPFS=all is in use with a
large TMPFS_BLACKLIST. So far the largest observed SWAP
usage so far: 11252Mi "MaxObsUsed".

For this activity, the system seems to be stable. I did
take the top off the case to help with its cooling:
plastic, so more of an insulator with the insides fairly
packed. I made sure the other side's grid of holes had
air space. It does have the official heat sink/fan unit.

Relative to my normally-headless activity, oddities are:

) No login prompt at the serial console (debug port).

) Not all types of messages are being logged to the
  serial console (debug port). It seems to just be
  kernel messages.

) It seems that the keyboard (with a mouse attached)
  on USB2 sometimes can be long delayed for showing
  up (judging by when the kernel messages show up).
  (It appears that noticing them may sometimes be
  tied to my doing something unrelated via ssh.
  My normal use is via ssh.)

I've no clue how to track down what:

# ls -lodT /dev/tty*
crw-------  1 root wheel - 0x2f May  8 13:02:40 2025 /dev/ttyu0
crw-------  1 root wheel - 0x30 Jun  5 20:56:51 2025 /dev/ttyu0.init
crw-------  1 root wheel - 0x31 Jun  5 20:56:51 2025 /dev/ttyu0.lock

is actually tracing back to in the RPi5 board's hardware (in
case it is not the debug port).

# grep ttyu0 /etc/ttys
ttyu0	"/usr/libexec/getty 3wire"	vt100	onifconsole secure

As stands Ethernet is via a USB3 dongle.

===
Mark Millard
marklmi at yahoo.com