This is the mail archive of the binutils@sourceware.org mailing list for the binutils project.
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |
Other format: | [Raw text] |
On Thursday 16 August 2007, Mike Frysinger wrote: > On Wednesday 15 August 2007, H.J. Lu wrote: > > On Wed, Aug 15, 2007 at 05:07:37PM +0100, Nick Clifton wrote: > > > Hi H.J. > > > > > > >We keep GNU attributes section in executable and shared library. Also > > > >There is a PT_GNU_STACK segment. But only its segment type is used. > > > >This patch maps GNU attributes section to PT_GNU_STACK segment and > > > >makes PT_GNU_ATTRIBUTES an alias of PT_GNU_STACK so that GNU > > > > attributes section is available to ELF loader. > > > > > > Why do you want to do this ? In particular why do you want to overload > > > the PT_GNU_STACK segment ? Why not create a new PT_GNU_ATTR segment ? > > > Is space in the header the only reason ? > > > > > > It seems to me that overloading the PT_GNU_STACK segment in the way you > > > propose is prone to confusing the user, and I think that we ought to > > > try to avoid this. > > > > The current usage of the PT_GNU_STACK segment is the flags field > > only. All other fields are ignored. The new segment ignores the > > flags field. Overloading the PT_GNU_STACK segment seems a good > > idea to me since it is totally backward and forward compatible. > > this is not correct ... on Blackfin, we use the other attributes of > PT_GNU_STACK (like the memsize field) to control the stack size (since > Blackfin is a no-mmu) to be a little more precise ... all Linux ports which support FDPIC ELF utilize the memsize field to set the stack size of the application, so this isnt Blackfin-specific -mike
Attachment:
signature.asc
Description: This is a digitally signed message part.
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |