On Thu, 28 Mar 2019 10:24:20 -0700 (PDT), Dinosaur
<
vande...@gmail.com> declaimed the
following:
>Hi All
>
>Thanks for the prompt responses.
>
>The Libpruio had difficulties when booting from uSD with the MMCBLK1 in
>tact.
>I quote the Freebasic post relating to this.
>
>This is the problem, I guess. The kernel finds a valid drive system on eMMC
>> and binds it as /dev/mmcblk0 in this case.
>> Instead it should bind the uSD as /dev/mmcblk0, and completly boot from
>> uSD. You'll have to erase the partition table so that the kernel cannot
>> bind this drive! I recommend
>>
That hasn't been the situation for years as I recall... I think Wheezy
(Debian 7.x) was the last OS image that had the behavior of changing device
binding.
Booting my BBB with only the eMMC gives:
debian@beaglebone:~$ df
Filesystem 1K-blocks Used Available Use% Mounted on
udev 220164 0 220164 0% /dev
tmpfs 49476 5292 44184 11% /run
/dev/mmcblk1p1 3704040 3036352 459816 87% /
tmpfs 247376 0 247376 0% /dev/shm
tmpfs 5120 4 5116 1% /run/lock
tmpfs 247376 0 247376 0% /sys/fs/cgroup
tmpfs 49472 0 49472 0% /run/user/1000
debian@beaglebone:~$ ls -l /dev/m*
crw-r----- 1 root kmem 1, 1 Mar 25 14:13 /dev/mem
crw------- 1 root root 10, 59 Mar 25 14:13 /dev/memory_bandwidth
brw-rw---- 1 root disk 179, 0 Mar 25 14:12 /dev/mmcblk1
brw-rw---- 1 root disk 179, 8 Mar 25 14:12 /dev/mmcblk1boot0
brw-rw---- 1 root disk 179, 16 Mar 25 14:12 /dev/mmcblk1boot1
brw-rw---- 1 root disk 179, 1 Mar 25 14:13 /dev/mmcblk1p1
brw-rw---- 1 root disk 179, 24 Mar 25 14:13 /dev/mmcblk1rpmb
/dev/mapper:
total 0
crw------- 1 root root 10, 236 Nov 3 2016 control
/dev/mqueue:
total 0
eMMC is mmcblk1 device... Inserting an SD card and rebooting gives:
debian@beaglebone:~$ df
Filesystem 1K-blocks Used Available Use% Mounted on
udev 220168 0 220168 0% /dev
tmpfs 49476 5928 43548 12% /run
/dev/mmcblk0p1 7570776 2928044 4283728 41% /
tmpfs 247376 0 247376 0% /dev/shm
tmpfs 5120 4 5116 1% /run/lock
tmpfs 247376 0 247376 0% /sys/fs/cgroup
tmpfs 49472 0 49472 0% /run/user/1000
debian@beaglebone:~$ ls -l /dev/m*
crw-r----- 1 root kmem 1, 1 Mar 28 21:27 /dev/mem
crw------- 1 root root 10, 59 Mar 28 21:27 /dev/memory_bandwidth
brw-rw---- 1 root disk 179, 0 Mar 28 21:27 /dev/mmcblk0
brw-rw---- 1 root disk 179, 1 Mar 28 21:27 /dev/mmcblk0p1
brw-rw---- 1 root disk 179, 8 Mar 28 21:27 /dev/mmcblk1
brw-rw---- 1 root disk 179, 16 Mar 28 21:27 /dev/mmcblk1boot0
brw-rw---- 1 root disk 179, 24 Mar 28 21:27 /dev/mmcblk1boot1
brw-rw---- 1 root disk 179, 9 Mar 28 21:27 /dev/mmcblk1p1
brw-rw---- 1 root disk 179, 32 Mar 28 21:27 /dev/mmcblk1rpmb
/dev/mapper:
total 0
crw------- 1 root root 10, 236 Mar 28 21:26 control
/dev/mqueue:
total 0
debian@beaglebone:~$ uname -a
Linux beaglebone 4.14.49-ti-r54 #1 SMP PREEMPT Fri Jun 15 22:14:13 UTC 2018
armv7l GNU/Linux
SD card is mmcblk0, and with or without SD card, eMMC is mmcblk1. What
does change is the device number/node data... Without SD card, mmcblk1 is
179,0; with SD card mmcblk0 is 179,0 and mmcblk1 is 179,8
The u-Boot image that is used to start the boot may be coming from eMMC
{the full boot sequence is a mystery to me}, but when it detects the SD
card, it appears to transfer control to it.
Have you flashed the eMMC using one of the standard images (LXQT or
IoT) before going on to your SD image? My recommendation would be to ensure
the system works with a standard image in eMMC, then with a standard image
on SD card withOUT erasing eMMC blocks.