When using GCC on x86-64 to compile an usdt prog with -O1 or higher optimization, the compiler will generate SIB addressing mode for global array and PC-relative addressing mode for global variable, e.g. "1@-96(%rbp,%rax,8)" and "-1@4+t1(%rip)".
The current USDT implementation in libbpf cannot parse these two formats, causing `bpf_program__attach_usdt()` to fail with -ENOENT (unrecognized register).
This patch series adds support for SIB addressing mode in USDT probes. The main changes include: - add correct handling logic for SIB-addressed arguments in `parse_usdt_arg`. - force -O2 optimization for usdt.test.o to generate SIB addressing usdt argument spec. - change the global variable t1 to a local variable, to avoid compiler generating PC-relative addressing mode for it.
Testing shows that the SIB probe correctly generates 8@(%rcx,%rax,8) argument spec and passes all validation checks.
The modification history of this patch series: Change since v1: - refactor the code to make it more readable - modify the commit message to explain why and how
Change since v2: - fix the `scale` uninitialized error
Change since v3: - force -O2 optimization for usdt.test.o to generate SIB addressing usdt and pass all test cases.
Do we need to add support for PC-relative USDT argument spec handling in libbpf? I have some interest in this question, but currently have no ideas. Getting offsets based on symbols requires dependency on the symbol table. However, once the binary file is stripped, the symtab will also be removed, which will cause this approach to fail. Does anyone have any thoughts on this?
Jiawei Zhao (1): libbpf: fix USDT SIB argument handling causing unrecognized register error
tools/lib/bpf/usdt.bpf.h | 33 +++++++++++++- tools/lib/bpf/usdt.c | 43 ++++++++++++++++--- tools/testing/selftests/bpf/Makefile | 5 +++ tools/testing/selftests/bpf/prog_tests/usdt.c | 18 +++++--- 4 files changed, 86 insertions(+), 13 deletions(-)