tech-kern archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: missing memory with uefi



On Thu, May 02, 2019 at 11:33:35AM +0700, Robert Elz wrote:
>     Date:        Thu, 2 May 2019 16:11:55 +1200
>     From:        Mark Davies <mark%ecs.vuw.ac.nz@localhost>
>     Message-ID:  <3a37d1df-aa49-a36c-15c5-61156cbcfe36%ecs.vuw.ac.nz@localhost>
> 
>   | If it helps someone this is what it reported:
> 
>   | 000000004c3d0000/000000005bd59fff available [ConventionalMemory]
> 
> This gets to about 1.5MB (and is the last ConventionalMemory reported
> until this one (much lower down)...
> 
>   | 0000000100000000/000000087fffffff available [ConventionalMemory]
> 
> Which fills in the rest of the 32GB.
> 
> At a guess, something is running out of space, and the examination
> of the memory map stops before it ever finds the last one.


The bootloader reports all entries from EFI to the kernel. The kernel
combines contigous entries of the same type, but uses only the
first VM_PHYSSEG_MAX, For amd64 that's 64.

There should be a console message

WARNING: too many memory segment (increase VM_PHYSSEG_MAX)

but it's probably not visible that early in bootstrap.


On the other hand, the types are mapped to a few BIOS memory types
before combination. Even this map should shrink that way to only
a few entries unless the stripped parts vary a lot between "available"
and "reserved" or contain lots of holes.



Greetings,
-- 
                                Michael van Elst
Internet: mlelstv%serpens.de@localhost
                                "A potential Snark may lurk in every tree."


Home | Main Index | Thread Index | Old Index