Re: ELF bugfixes

[phcoder]

Robert Millan wrote:
> On Sat, Mar 21, 2009 at 06:58:58PM +0100, phcoder wrote:
> > Robert Millan wrote:
> > > On Wed, Mar 18, 2009 at 02:26:40PM +0100, phcoder wrote:
> > > > Robert Millan wrote:
> > > > > On Fri, Mar 13, 2009 at 09:52:39PM +0100, phcoder wrote:
> > > > > > -  grub_multiboot_payload_entry_offset = ehdr->e_entry - 
> > > > > > phdr(lowest_segment)->p_vaddr;
> > > > > > +  for (i = 0; i < ehdr->e_phnum; i++)
> > > > > > +    if (phdr(i)->p_vaddr <= ehdr->e_entry +	&& phdr(i)->p_vaddr 
> > > > > > + phdr(i)->p_memsz > ehdr->e_entry)
> > > > > > +      grub_multiboot_payload_entry_offset = (ehdr->e_entry - phdr(i)->p_vaddr)
> > > > > > +       + (phdr(i)->p_paddr  - phdr(lowest_segment)->p_paddr);
> > > > > You need to handle the case in which grub_multiboot_payload_entry_offset is left
> > > > > uninitialized (it needs to be initialized each time the multiboot command is
> > > > > run, not just when the module is loaded).
> > > > module? actually it's when loading image. Perhaps you mean that   
> > > > additional error check is necessary
> > > I meant GRUB's multiboot.mod, not the payload's module.  Sorry I wasn't clear.
> > With this error check if grub_multiboot_payload_entry_offset it can  
> > happen only if no image is loaded. And actually  
> > grub_multiboot_payload_entry_offset is set to 0 at multiboot.mod load
> > So I don't really understand the problem
>
> You can't rely on grub_multiboot_payload_entry_offset being set to 0, because
> any subsequent call of "multiboot /something" has the potential to override
> this.  You must not assume the multiboot command is only going to be run once.
No but it always corresponds to the current image. It's set either in 
multiboot.c or in grub_multiboot_load_elf

--

Regards
Vladimir 'phcoder' Serbinenko