amd64 process sizes

[Kostik Belousov]

On Sat, Sep 08, 2007 at 10:53:06PM -0700, Peter Wemm wrote:
> Peter Wemm wrote:
> >
> >2) There is a bug in sys/kern/imgact_elf.c.  It assumes that the 
> >PT_LOAD sections are contiguous, which isn't a given.
> 
> I think I may have outsmarted myself here.  This should be coming from 
> rtld, not the kernel.  In any case, it is still our bug.

The extra mapping is mostly harmless. It only increases the VSZ.
Moreover, I think that it prevents kernel from selecting gaps for
mapping some chunks between segments of the elf object. This might cause
funny effects, and even break gdb more.

What seems to be a real bug is the fact that first mapping (for whole
region, inside which the actual segments are remmapped later) is
not aligned according to the elf object segments requirement. As
consequence, our rtld ignores alignment specification in the elf
segments that are greater that page size. It seems we were lucky that
gcc and most other compilers does not generate code that depends on
bigger alignment.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-amd64/attachments/20070909/6c39d596/attachment.pgp