Aarch64 Pointer Authentication & Branch Target Enablement
Arm Pointer Authentication (PAC) is a method of hardening code from Return Oriented Programming (ROP) attacks. It uses a tag in a pointer to sign and verify pointers. Branch Target Identification (BTI) is another code hardening method, where the branch/jump target is identified with a special landing pad instruction. Outside of some system support in glibc+kernel, packages gain the additional hardening by compiling with the -mbranch-protection= flag available in recent versions of GCC. In particular -mbranch-protection=standard enables both BTI and PAC, with backwards compatible to armv8.0 code sequences that activate on v8.3 (PAC) & v8.5 (BTI) enabled Arm machines.
- Name: Jeremy Linton & ARM SIG
- Email: email@example.com
- Targeted release: Fedora 33
- Last updated: 2020-07-20
- FESCo issue: #2403
- Tracker bug: #1847148
- Release notes tracker: #512
Benefit to Fedora
PAC & BTI are code hardening features, they should serve to make fedora more resistant to a couple further classes of runtime attacks. By enabling this early, fedora is once again proven to be at the leading edge of security and linux development. If everything works as planned, this change will be invisible to the end user, except in cases where the applications will trap behaviour that appears to be caused by exploit attempts.
- Proposal owners:
Work with individual package maintainers in the case of build failures or runtime exceptions. In the latter case there are two possibilities. First on v8.0 hardware, which is currently the most common, the additional instruction sequences are treated as NOP's and should be completely ignored by the hardware. It may be possible on v8.3/8.5 hardware that PAC or BTI may need additional tweaks for hand written assembly which interacts with PAC/BTI enabled code.
- Other developers:
Assure their packages continue to compile and pass unit/integration/etc tests on v8.0 hardware. Continue to monitor runtime problems on v8.3+ for bugs, vs exploit attempts.
- Release engineering: 9505 (a check of an impact with Release Engineering is needed)
- Policies and guidelines:
At the moment, nothing needs to be changed as this should propagate as the default set of RPM build flags.
- Trademark approval: N/A (not needed for this Change)
If everything works as planned, this should be transparent to the end user.
How To Test
Testing falls into two categories. Assuring that the packages continue to work on existing arm v8.0 hardware without PAC, and testing on PAC+BTI enabled hardware. For the most part the expectation from the fedora community is that package maintainers assure their packages continue to work on existing systems. PAC+ hardware will be in limited supply during the F33 development cycle, so the expectation is that owners of that hardware will perform more complete systemwide testing and report any defects found against the packages in question along with fixes or hardware access.
There are various gcc and kernel related changes which have already landed, but there continue to be a few cleanup patches trickling into the toolchain/compiler as problems are discovered.
- Contingency mechanism: (What to do? Who will do it?)
Build affected packages with explicit compiler flags disabling the feature. Worse case the top level rpm macros reversion, and a rebuild of effected packages. Contingency deadline: Beta target.
- Blocks release? No, except for major functionality loss due to core package bug.
- Blocks product? No
Arm pointer authentication is a technology designed to make software more robust by providing hardware assistance for code hardening. It protects pointers by cryptographically signing them and verifying their signatures when used, thereby mitigating certain attack vectors. Core support is provided to applications and libraries transparently via kernel and toolchain changes to generate hardended code. Branch Target identification, similarly provides landing pads, to harden code paths by restricting the processor from jumping into unexpected parts of a function.