Implement automated binary re-relativization for AArch32 and AArch64 #4016
Assignees
Comments
This was referenced Jan 14, 2020
|
Just checking, but this issue is still only a concern when DR is used in standalone mode right? (in relation to #731) Or are there mangling issues which were recently discovered please? |
|
This does not impact core DR's basic operations, no. It impacts someone using decode_from_copy() or other decoding operations, which could be standalone or in a client. The workaround of forcing re-encode should ensure there are no problems today except when we have parts of an instr that are not recorded and not recreated on that re-encode: such as this case which we hit recently: #4017. |
derekbruening
added a commit
that referenced
this issue
Jan 15, 2020
Adds rip-relative information tracking at IR levels 1-3 for instruction references on x86, extending the existing tracking for data references. Adds decode_sizeof_ex() and instr_get_rel_data_or_instr_target() to support rip-relative displacement and target handling. Changes decode_from_copy() to preserve raw bits for all instructions, eliminating problems where untracked encoding features are lost such as in #4017. Documents the changes and describes how to avoid the new behavior. Does just that for intra-sequence cti's in the rseq native code copy. Adds test cases to api.ir. Implementing the same thing for AArchXX is left unimplemented, tracked by i#4016. Issue: #731, #3271, #3339, #4016, #4017 Fixes #731 Fixes #3271 Fixes #4017
derekbruening
added a commit
that referenced
this issue
Jan 16, 2020
Adds rip-relative information tracking at IR levels 1-3 for instruction references on x86, extending the existing tracking for data references. Adds decode_sizeof_ex() and instr_get_rel_data_or_instr_target() to support rip-relative displacement and target handling. Changes decode_from_copy() to preserve raw bits for all instructions, eliminating problems where untracked encoding features are lost such as in #4017. Documents the changes and describes how to avoid the new behavior. Does just that for intra-sequence cti's in the rseq native code copy and Windows syscall wrapper copies. Adds test cases to api.ir. Implementing the same thing for AArchXX is left unimplemented, tracked by i#4016. Issue: #731, #3271, #3339, #4016, #4017 Fixes #731 Fixes #3271 Fixes #4017
|
This is causing loss of encoding fidelity for cases like #7132. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Xref related issues: #731, #3271, #3339
This issue covers implementing automated binary re-relativization of AArch32 and AArch64 instructions. We will need a different strategy than the x86 approach of storing the location of a 32-bit immediate, due to the varying immediate encodings. For now, decode_from_copy() invalidates the raw bits to force a re-encoding as a workaround for not having this feature.
The text was updated successfully, but these errors were encountered: