[uClinux-dev] Re: Tiny memory

John Williams jwilliams at itee.uq.edu.au
Tue Mar 25 16:35:47 EST 2003

Greg Ungerer wrote:
> Non-beleivers, all of you :-)
> I guess I'll just have to prove it for you :-)

There is a reason you are called the Chief Software Wizard! )

> Below is the output log of a uClinx/ColdFire kernel
> and filesystem limited to 1MB on a Motorola M5272C3 board.
> Kernel is 2.4.20, only app configured is a shell. I left
> the proc fs enabled just for fun, networking is disabled.
> Kernel and filesystem are in RAM, and this setup
> actually loses the first 128k of the 1MB of RAM (due to
> being compitible with the dBUG firmware of the M5272C3
> board, so it is really 896k :-)

[snip depressing and seemingly far away uClinux bootup log!]

Ok Ok so I believe you now!  Well I believed you all along actually.

Now here's the big question, did you enable the page_alloc2 stuff, or 
just use the standard settings?  That's the key to my question, I need 
to determine whether it's my small memory size that's causing my slab.c 
BUG() (in which case I could use page_alloc2), or if something else is 
rotten, and it just manifests itself in slab.c.  The more I see, the 
more I think it's the latter ... back to the debugger :(



More information about the uClinux-dev mailing list