Assembly instrumentation fails on Android-x86

[ramosian-glider]

Originally reported on Google Code with ID 353

http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux/builds/13369/steps/run%20asan%20tests%20%5BAndroid%2Fx86-K%5D/logs/stdio

We've got hardcoded shadow mapping in asm instrumentation code.
Android uses zero-based shadow, unlike Linux.

We should share and reuse AddressSanitizerPass's knowledge about shadow mappings on
various platforms.

Reported by eugenis@google.com on 2014-10-10 12:23:14

[ramosian-glider]

The url in the original message does not work anymore, here is a recent more recent
run of the same test: http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux/builds/14766/steps/run%20asan%20tests%20%5BAndroid%2Fx86-K%5D/logs/stdio

Reported by krasin@google.com on 2014-12-20 03:21:35

[ramosian-glider]

Adding Project:AddressSanitizer as part of GitHub migration.

Reported by ramosian.glider on 2015-07-30 09:06:34

• Labels added: ProjectAddressSanitizer

[eugenis]

More problems.

• Hardcoded shadow mapping.
• cmake in asan/tests checks for CAN_TARGET_I386, but android is I686
• instrumentation of rep movs uses EBX as a scratch register. On x86 EBX is a PIC register. On Android we link ASan runtime as a shared library, and crash when entering the PLT stub for __asan_report_load with bad EBX.

[eugenis]

And we are out of registers.

[morehouse]

@eugenis: This bug is over 3 years old. Do we want to keep it open?

[eugenis]

I think we want to deprecate the whole assembly instrumentation thing.
It never evolved into anything useful.