Fix two pointer arithmetic errors.

Fix two pointer arithmetic errors.

In the heuristic for detecting relative references, we had two non-fatal, but
sub-optimal mistakes with pointer arithmetic. This fixes them for both
ELF 32 and Win 32.

BUG=chromiumos:22677
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=109172

[dgarrett, 2011-11-08 21:05:04]

This fixes two issues pointed out in the previous code review. It DOES introduce
version incompatibility in some patch files.

[sra1, 2011-11-08 21:18:20]

On 2011/11/08 21:05:04, dgarrett wrote:
> This fixes two issues pointed out in the previous code review. It DOES
introduce
> version incompatibility in some patch files.

patches for files which have text sections ending with a jmp or call?

[sra1, 2011-11-08 21:18:45]

lgtm

http://codereview.chromium.org/8501023/diff/1/courgette/disassembler_elf_32_x...
File courgette/disassembler_elf_32_x86.cc (right):

http://codereview.chromium.org/8501023/diff/1/courgette/disassembler_elf_32_x...
courgette/disassembler_elf_32_x86.cc:575: if (p + 6 < end_pointer) {
<= here too

http://codereview.chromium.org/8501023/diff/1/courgette/disassembler_win32_x8...
File courgette/disassembler_win32_x86.cc (right):

http://codereview.chromium.org/8501023/diff/1/courgette/disassembler_win32_x8...
courgette/disassembler_win32_x86.cc:474: if (p + 6 < end_pointer) {
<= here too

[sra1, 2011-11-08 21:20:49]

On 2011/11/08 21:05:04, dgarrett wrote:
> This fixes two issues pointed out in the previous code review. It DOES
introduce
> version incompatibility in some patch files.

In current usage, the incompatibility should be controlled for in the 'three
stage rocket'.

[dgarrett, 2011-11-08 21:32:43]

Yes, Chrome should be safe against the compatibility problem.

We're debating something similar for ChromeOS, but haven't decided yet.

http://codereview.chromium.org/8501023/diff/1/courgette/disassembler_elf_32_x...
File courgette/disassembler_elf_32_x86.cc (right):

http://codereview.chromium.org/8501023/diff/1/courgette/disassembler_elf_32_x...
courgette/disassembler_elf_32_x86.cc:575: if (p + 6 < end_pointer) {
On 2011/11/08 21:18:45, sra1 wrote:
> <= here too

Sigh, you're right. And I thought I looked for more problems like this.

Done.

[commit-bot: I haz the power, 2011-11-08 22:15:43]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/dgarrett@chromium.org/8501023/2002

[commit-bot: I haz the power, 2011-11-09 00:39:26]

Try job failure for 8501023-2002 (retry) on win_rel for step
"chrome_frame_net_tests".
It's a second try, previously, step "chrome_frame_net_tests" failed.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win_rel&nu...

[commit-bot: I haz the power, 2011-11-09 00:44:39]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/dgarrett@chromium.org/8501023/2002

[commit-bot: I haz the power, 2011-11-09 02:15:49]

Change committed as 109172