Thanks for the sanity check. This is an AMD based setup using a combined CPU+GPU processor that shares system memory for both functions. (i.e. AMD APU on A88X chipset)
I dug around a bit just now and it looks to me like this is normal for a shared memory resource(?). I ran this command at the prompt to corroborate what I found:
which returned this:
[ 1.633246] radeon 0000:00:01.0: VRAM: 1024M 0x0000000000000000 - 0x000000003FFFFFFF (1024M used)
[ 1.633249] [drm] Detected VRAM RAM=1024M, BAR=256M
[ 1.633389] [drm] radeon: 1024M of VRAM memory ready
Then
returned this:
[ 0.000000] Memory: 6849856K/7276616K available (8211K kernel code, 1416K rwdata, 3340K rodata, 1604K init, 1520K bss, 426760K reserved, 0K cma-reserved)
[ 0.331252] x86/mm: Memory block size: 128MB
The numbers look like they mostly add up, and I'm guessing that with only 4gb before, the system borrowed less system memory.
Assuming that to be the case, I think I'm used to how Windows reports memory in that it will report all 8gb and I guess Linux is only telling me the "total" availability
after video allocation?
Is that sound logic?