| Commit message (Collapse) | Author | Age |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
cchecklink is now using program header information to figure out the
initial address space of the program, rather than the information in the
parent section of each symbol. This decouples the resolution of symbols
from inaccurate section information, reflecting more the actual program
loading. Additionally, a -relaxed option has been added to deal with
some strange ELFs, for instance when symbols data is dynamically
bootstrapped from another place by boot code different than the program
loader.
git-svn-id: https://yquem.inria.fr/compcert/svn/compcert/trunk@1893 fca1b0fc-160b-0410-b1d3-a4f43f01ea2e
|
| |
|
|
|
|
|
|
|
| |
Section mapping is now discovered on-the-fly, and linker script remappings
are reported as warnings at the end. Symbol data lookup is now able to
gracefully fail if the symbol's virtual address is not within the range of
its parent section's virtual address space.
git-svn-id: https://yquem.inria.fr/compcert/svn/compcert/trunk@1878 fca1b0fc-160b-0410-b1d3-a4f43f01ea2e
|
| |
|
|
|
|
|
|
|
| |
ndxes_of_sym_name used to have an O(s^2) complexity where s was the number
of symbols in the ELF file. It has now been reduced to an O(s*ln(s)) by
pre-computing the sets of symbols corresponding to each normalized symbol
name.
git-svn-id: https://yquem.inria.fr/compcert/svn/compcert/trunk@1875 fca1b0fc-160b-0410-b1d3-a4f43f01ea2e
|
| |
|
|
| |
git-svn-id: https://yquem.inria.fr/compcert/svn/compcert/trunk@1872 fca1b0fc-160b-0410-b1d3-a4f43f01ea2e
|
|
|
cparser: renamed Errors to Cerrors; removed packing into Cparser.
git-svn-id: https://yquem.inria.fr/compcert/svn/compcert/trunk@1856 fca1b0fc-160b-0410-b1d3-a4f43f01ea2e
|
varobert
xleroy