> From: "Richard B. Johnson" <root@analogic.com>
>
> > Do both (i) and (ii):
> > (i) apply my patch to free_initmem() (or just comment it out)
> > (ii) get recent binutils
>
> Well I have the latest binutils installed!! binutils-2.8.0.3 with many of
> the source and 'build` file dates in April of this year! Do these things
> get changed weekly?
>
> Hmm - my fault was an old ld - perhaps your fault is running SMP?
>
> I am not near a kernel source tree, but you can easily test things,
> now that you know where to look.
> (Look in System.map at the addresses for __init_begin and __init_end.
> They should be multiples of 4096, i.e., end in 000. If not, your ld is bad.
> Comment out the body of free_initmem(). Test whether all is fine. Etc.)
>
> Andries
>
I "fixed" the problem, as you suggested by adding a "return" as the first
instruction in free_initmem(). The System.map file shows that the labels
that you specify truly start and end on a page boundary.... But, I think
you are correct, free_initmem() seems to free one page too many.
c01fc000 A __init_begin
c0203000 A __init_end
Cheers,
Dick Johnson
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Richard B Johnson
Project Engineer
Analogic Corporation
Voice : (508) 977-3000 ext. 3754
Fax : (508) 532-6097
Modem : (508) 977-6870
FTP : ftp@boneserver.analogic.com
Email : rjohnson@analogic.com, johnson@analogic.com
Penguin : Linux version 2.1.40 on an i586 machine (66.36 BogoMips).
Warning : I read unsolicited mail for $350.00 per hour. Supply billing address.
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-