[Courgette] Fix AssemblyProgram parsing for ELF-ARM.

[Courgette] Fix AssemblyProgram parsing for ELF-ARM.

This fixes two problems:
(A) In DisassemblerElf32, RVAToFileOffset() used Program Segment Header,
  but FileOffsetToRVA() used Section Header. For consistency, both
  should use the same one. We choose the latter (for now).
(B) Even if a section has sh_type=SHT_PROGBITS, it can still have
  sh_addr=0. Extracting Rel32 address from these sections would add
  overlapping RVA chaos, and so should be avoided.

Also using elf-arm7 in a unit test. The test fails before the fix and
passes after.

BUG=579206
Committed: https://crrev.com/7ec152caaf55f64431dcb2c63cc052890c4a93f1
Cr-Commit-Position: refs/heads/master@{#373639}

[huangs, 2016-01-31 04:37:12]

Description was changed from

==========
q# Enter a description of the change.
[Courgette] Fix AssemblyProgram parsing for ELF and ARM.

BUG= 579206
==========

to

==========
[Courgette] Fix AssemblyProgram parsing for ELF and ARM.

BUG= 579206
==========

[huangs, 2016-01-31 05:21:53]

The CQ bit was checked by huangs@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-01-31 05:22:02]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1658463002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1658463002/1

[huangs, 2016-01-31 05:39:38]

Description was changed from

==========
[Courgette] Fix AssemblyProgram parsing for ELF and ARM.

BUG= 579206
==========

to

==========
[Courgette] Fix AssemblyProgram parsing for ELF-ARM.

This fixes two problems:
(A) In DisassemblerElf32, RVAToFileOffset() used Program Segment Header,
  but FileOffsetToRVA() used Section Header. Both should use the latter.
(B) Even if a section has sh_type=SHT_PROGBITS, it can still have
  sh_addr=0. Extract Rel32 address from these sections adds overlapping
  RVA chaos, and should be avoided.

Verified that with the fix, Courgette for ELF-x86 still works.

BUG= 579206
==========

[commit-bot: I haz the power, 2016-01-31 06:28:19]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-01-31 06:28:20]

Dry run: This issue passed the CQ dry run.

[huangs, 2016-02-01 13:19:32]

huangs@chromium.org changed reviewers:
+ chrisha@chromium.org, wfh@chromium.org

[huangs, 2016-02-01 13:19:32]

PTAL.  Thanks!

[Will Harris, 2016-02-01 19:05:10]

On 2016/02/01 13:19:32, huangs wrote:
> PTAL.  Thanks!

both look like good fixes but can you add a test ELF executable that did not
work previously and now works?

[huangs, 2016-02-02 17:41:16]

That's a good idea, though it looks like trivial ELF executables don't trigger
the bug.

Meanwhile I realized that there's another cause, and (A) merely masks it. Will
comment in the bug description.

[huangs, 2016-02-02 21:07:20]

Will need to rethink the fix. Meanwhile, test binary is added in
http://crrev.com/1662633002/.

[huangs, 2016-02-04 18:39:07]

Added the test, and verified that before fix the test would fail (error code
25).  Note that I only added 1 test, although elf-32-1 appears 5 times.  The 4
"missing" tests are not added because:
- bsdiff_memory_unittest.cc: Also uses elf-32-2, but right now we only have
single ARM test.
- ensemble_unittest.cc: Same.
- encoded_program_fuzz_unittest.cc: courgette_fuzz is currently broken, don't
want to complicate (also, why does it currently use nonexistent elf-32-1.exe?)
- disassembler_elf_32_x86_unittest.cc: Should create counterpart for ARM, and
refactor -- punting for now. Also the test seems broken: the check for abs32 &
rel32 overlap does not consider 4-byte width of pointers.

PTAL.  Thanks!

[huangs, 2016-02-04 18:39:24]

The CQ bit was checked by huangs@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-02-04 18:39:52]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1658463002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1658463002/40001

[Will Harris, 2016-02-04 18:54:21]

https://codereview.chromium.org/1658463002/diff/40001/courgette/disassembler_...
File courgette/disassembler_elf_32.cc (left):

https://codereview.chromium.org/1658463002/diff/40001/courgette/disassembler_...
courgette/disassembler_elf_32.cc:170: Elf32_Addr begin =
ProgramSegmentMemoryBegin(i);
are ProgramSegmentMemoryBegin and ProgramSegmentMemorySize needed any more now?

[commit-bot: I haz the power, 2016-02-04 19:56:41]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-02-04 19:56:41]

Dry run: This issue passed the CQ dry run.

[huangs, 2016-02-04 20:13:50]

Updated, PTAL.

https://codereview.chromium.org/1658463002/diff/40001/courgette/disassembler_...
File courgette/disassembler_elf_32.cc (left):

https://codereview.chromium.org/1658463002/diff/40001/courgette/disassembler_...
courgette/disassembler_elf_32.cc:170: Elf32_Addr begin =
ProgramSegmentMemoryBegin(i);
Ah good catch.  Also removing other unused accessors.

[Will Harris, 2016-02-04 20:37:26]

lgtm

[huangs, 2016-02-04 21:09:42]

Thanks! Committing.

[huangs, 2016-02-04 21:09:47]

The CQ bit was checked by huangs@chromium.org

[commit-bot: I haz the power, 2016-02-04 21:11:49]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1658463002/60001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1658463002/60001

[huangs, 2016-02-04 21:40:37]

The CQ bit was unchecked by huangs@chromium.org

[huangs, 2016-02-04 21:42:19]

The CQ bit was checked by huangs@chromium.org

[commit-bot: I haz the power, 2016-02-04 21:43:04]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1658463002/60001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1658463002/60001

[huangs, 2016-02-04 21:43:07]

Hmm I said I would rethink this bug, but I think it's good to get this CL in
first anyway.  Will comment in the bug.

[huangs, 2016-02-04 22:09:01]

The CQ bit was unchecked by huangs@chromium.org

[huangs, 2016-02-04 22:12:54]

Description was changed from

==========
[Courgette] Fix AssemblyProgram parsing for ELF-ARM.

This fixes two problems:
(A) In DisassemblerElf32, RVAToFileOffset() used Program Segment Header,
  but FileOffsetToRVA() used Section Header. Both should use the latter.
(B) Even if a section has sh_type=SHT_PROGBITS, it can still have
  sh_addr=0. Extract Rel32 address from these sections adds overlapping
  RVA chaos, and should be avoided.

Verified that with the fix, Courgette for ELF-x86 still works.

BUG= 579206
==========

to

==========
[Courgette] Fix AssemblyProgram parsing for ELF-ARM.

This fixes two problems:
(A) In DisassemblerElf32, RVAToFileOffset() used Program Segment Header,
  but FileOffsetToRVA() used Section Header. For consistency, both
  should use the same one. We choose the latter (for now).
(B) Even if a section has sh_type=SHT_PROGBITS, it can still have
  sh_addr=0. Extracting Rel32 address from these sections would add
  overlapping RVA chaos, and so should be avoided.

Also using elf-arm7 in a unit test. The test fails before the fix and
passes after.

BUG= 579206
==========

[huangs, 2016-02-04 22:13:04]

The CQ bit was checked by huangs@chromium.org

[commit-bot: I haz the power, 2016-02-04 22:14:43]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1658463002/60001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1658463002/60001

[commit-bot: I haz the power, 2016-02-04 22:26:49]

Description was changed from

==========
[Courgette] Fix AssemblyProgram parsing for ELF-ARM.

This fixes two problems:
(A) In DisassemblerElf32, RVAToFileOffset() used Program Segment Header,
  but FileOffsetToRVA() used Section Header. For consistency, both
  should use the same one. We choose the latter (for now).
(B) Even if a section has sh_type=SHT_PROGBITS, it can still have
  sh_addr=0. Extracting Rel32 address from these sections would add
  overlapping RVA chaos, and so should be avoided.

Also using elf-arm7 in a unit test. The test fails before the fix and
passes after.

BUG= 579206
==========

to

==========
[Courgette] Fix AssemblyProgram parsing for ELF-ARM.

This fixes two problems:
(A) In DisassemblerElf32, RVAToFileOffset() used Program Segment Header,
  but FileOffsetToRVA() used Section Header. For consistency, both
  should use the same one. We choose the latter (for now).
(B) Even if a section has sh_type=SHT_PROGBITS, it can still have
  sh_addr=0. Extracting Rel32 address from these sections would add
  overlapping RVA chaos, and so should be avoided.

Also using elf-arm7 in a unit test. The test fails before the fix and
passes after.

BUG= 579206
==========

[commit-bot: I haz the power, 2016-02-04 22:26:50]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2016-02-04 22:27:48]

Description was changed from

==========
[Courgette] Fix AssemblyProgram parsing for ELF-ARM.

This fixes two problems:
(A) In DisassemblerElf32, RVAToFileOffset() used Program Segment Header,
  but FileOffsetToRVA() used Section Header. For consistency, both
  should use the same one. We choose the latter (for now).
(B) Even if a section has sh_type=SHT_PROGBITS, it can still have
  sh_addr=0. Extracting Rel32 address from these sections would add
  overlapping RVA chaos, and so should be avoided.

Also using elf-arm7 in a unit test. The test fails before the fix and
passes after.

BUG= 579206
==========

to

==========
[Courgette] Fix AssemblyProgram parsing for ELF-ARM.

This fixes two problems:
(A) In DisassemblerElf32, RVAToFileOffset() used Program Segment Header,
  but FileOffsetToRVA() used Section Header. For consistency, both
  should use the same one. We choose the latter (for now).
(B) Even if a section has sh_type=SHT_PROGBITS, it can still have
  sh_addr=0. Extracting Rel32 address from these sections would add
  overlapping RVA chaos, and so should be avoided.

Also using elf-arm7 in a unit test. The test fails before the fix and
passes after.

BUG= 579206

Committed: https://crrev.com/7ec152caaf55f64431dcb2c63cc052890c4a93f1
Cr-Commit-Position: refs/heads/master@{#373639}
==========

[commit-bot: I haz the power, 2016-02-04 22:27:48]

Patchset 4 (id:??) landed as
https://crrev.com/7ec152caaf55f64431dcb2c63cc052890c4a93f1
Cr-Commit-Position: refs/heads/master@{#373639}