llvm dev - Jun 2020 - Is it possible to get the exact PC address of an instruction?

I wish to perform some instrumentation at a particular PC
address/instruction using LLVM and I wished to know if getting the PC
address of the instruction is in fact possible given the nature of the llvm
compiler? For instance, I wish to inject a fault multiple times at the same
instruction address and analyze this behavior. I wish to get the PC address
of the instruction so as to be able to point to it as my injection target
for all my runs.

Regards
Abhay Potharaju
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20200609/6abb26d1/attachment.html>

Hello Abhay,

Are you talking about fault injection at runtime?  With an instruction generated
from C/C++ source?  I can imagine using a label in the source code, and finding
the address of that label afterward, which limits you to the first instruction
generated after the label.
But in the general case, if I had to do this myself, I’d probably end up
disassembling the object file to find the instruction of interest; or generating
asm source (with -S) and inserting a label before running it through the
assembler.  None of these tactics do anything clever within LLVM, it would work
with any toolchain.

Good luck,
--paulr

From: llvm-dev <llvm-dev-bounces at lists.llvm.org> On Behalf Of Abhay
Kamesh Potharaju via llvm-dev
Sent: Tuesday, June 9, 2020 11:16 AM
To: llvm-dev at lists.llvm.org
Subject: [llvm-dev] Is it possible to get the exact PC address of an
instruction?

I wish to perform some instrumentation at a particular PC address/instruction
using LLVM and I wished to know if getting the PC address of the instruction is
in fact possible given the nature of the llvm compiler? For instance, I wish to
inject a fault multiple times at the same instruction address and analyze this
behavior. I wish to get the PC address of the instruction so as to be able to
point to it as my injection target for all my runs.

Regards
Abhay Potharaju

-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20200609/11d4061a/attachment.html>