We sent Kevin two patches, originating from the TIGCC fork. These patches are a small subset of Patrick Pelissier (PpHd) 's patch, mentioned at http://tichessteamhq.yuku.com/topic/403?page=2 . Kevin seems to want the patches to be posted on this message board rather than in private mails (the latter being the means of communication multiple members of the fork have used previously for most of their previous contributions to TIGCC), so here are they:
* "ld-tigcc: add code+doc for __ld_bss_even_end" : http://lionel.debroux.fre...ion-of-special-__ld.patch .
Nothing special to say about this patch.
* "ld-tigcc: add code+doc for __ld_first_sector_after_os" : http://lionel.debroux.fre...ion-of-special-__ld.patch .
In Patrick's patch, this is called "__ld_archive_start". I think this name is not neutral enough: the data/code written at the next sector after the end of the OS may very well not be "archive memory" in the AMS sense (user-accessible storage). This is especially likely true on Punix OS, but maybe also on PedroM if someone ever implements FlashApps and stores them like in AMS, between the OS and "archive memory".
Alternative names for this builtin could be "first_sector_after_end_of_os", "next_sector_after_os" or something else. Anyway, whatever choice a s/// away.
There's a third patch, but Kevin says this is not the right solution to fix the problem, so we're not going to post the patch along with the two others. Maybe later.

