| Message ID | 20240529121251.1993135-1-ptosi@google.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | KVM: arm64: Add support for hypervisor kCFI | expand |
On Wed, May 29, 2024 at 01:12:06PM +0100, Pierre-Clément Tosi wrote: > CONFIG_CFI_CLANG ("kernel Control Flow Integrity") makes the compiler inject > runtime type checks before any indirect function call. On AArch64, it generates > a BRK instruction to be executed on type mismatch and encodes the indices of the > registers holding the branch target and expected type in the immediate of the > instruction. As a result, a synchronous exception gets triggered on kCFI failure > and the fault handler can retrieve the immediate (and indices) from ESR_ELx. > > This feature has been supported at EL1 ("host") since it was introduced by > b26e484b8bb3 ("arm64: Add CFI error handling"), where cfi_handler() decodes > ESR_EL1, giving informative panic messages such as > > [ 21.885179] CFI failure at lkdtm_indirect_call+0x2c/0x44 [lkdtm] > (target: lkdtm_increment_int+0x0/0x1c [lkdtm]; expected type: 0x7e0c52a) > [ 21.886593] Internal error: Oops - CFI: 0 [#1] PREEMPT SMP > > However, it is not or only partially supported at EL2: in nVHE (or pKVM), > CONFIG_CFI_CLANG gets filtered out at build time, preventing the compiler from > injecting the checks. In VHE, EL2 code gets compiled with the checks but the > handlers in VBAR_EL2 are not aware of kCFI and will produce a generic and > not-so-helpful panic message such as > > [ 36.456088][ T200] Kernel panic - not syncing: HYP panic: > [ 36.456088][ T200] PS:204003c9 PC:ffffffc080092310 ESR:f2008228 > [ 36.456088][ T200] FAR:0000000081a50000 HPFAR:000000000081a500 PAR:1de7ec7edbadc0de > [ 36.456088][ T200] VCPU:00000000e189c7cf > > To address this, > > - [01/13] fixes an existing bug where the ELR_EL2 was getting clobbered on > synchronous exceptions, causing the wrong "PC" to be reported by > nvhe_hyp_panic_handler() or __hyp_call_panic(). This is particularly limiting > for kCFI, as it would mask the location of the failed type check. > - [02/13] fixes a minor C/asm ABI mismatch which would trigger a kCFI failure > - [03/13] to [09/13] prepare nVHE for CONFIG_CFI_CLANG and [10/13] enables it > - [11/13] improves kCFI error messages by saving then parsing the CPU context > - [12/13] adds a kCFI test module for VHE and [13/13] extends it to nVHE & pKVM > > As a result, an informative kCFI panic message is printed by or on behalf of EL2 > giving the expected type and target address (possibly resolved to a symbol) for > VHE, nVHE, and pKVM (iff CONFIG_NVHE_EL2_DEBUG=y). > > Note that kCFI errors remain fatal at EL2, even when CONFIG_CFI_PERMISSIVE=y. > > Changes in v4: > - Addressed Will's comments on v3: nit: but please keep reviewers on CC when you post a new version. I missed this initially. Will
CONFIG_CFI_CLANG ("kernel Control Flow Integrity") makes the compiler inject runtime type checks before any indirect function call. On AArch64, it generates a BRK instruction to be executed on type mismatch and encodes the indices of the registers holding the branch target and expected type in the immediate of the instruction. As a result, a synchronous exception gets triggered on kCFI failure and the fault handler can retrieve the immediate (and indices) from ESR_ELx. This feature has been supported at EL1 ("host") since it was introduced by b26e484b8bb3 ("arm64: Add CFI error handling"), where cfi_handler() decodes ESR_EL1, giving informative panic messages such as [ 21.885179] CFI failure at lkdtm_indirect_call+0x2c/0x44 [lkdtm] (target: lkdtm_increment_int+0x0/0x1c [lkdtm]; expected type: 0x7e0c52a) [ 21.886593] Internal error: Oops - CFI: 0 [#1] PREEMPT SMP However, it is not or only partially supported at EL2: in nVHE (or pKVM), CONFIG_CFI_CLANG gets filtered out at build time, preventing the compiler from injecting the checks. In VHE, EL2 code gets compiled with the checks but the handlers in VBAR_EL2 are not aware of kCFI and will produce a generic and not-so-helpful panic message such as [ 36.456088][ T200] Kernel panic - not syncing: HYP panic: [ 36.456088][ T200] PS:204003c9 PC:ffffffc080092310 ESR:f2008228 [ 36.456088][ T200] FAR:0000000081a50000 HPFAR:000000000081a500 PAR:1de7ec7edbadc0de [ 36.456088][ T200] VCPU:00000000e189c7cf To address this, - [01/13] fixes an existing bug where the ELR_EL2 was getting clobbered on synchronous exceptions, causing the wrong "PC" to be reported by nvhe_hyp_panic_handler() or __hyp_call_panic(). This is particularly limiting for kCFI, as it would mask the location of the failed type check. - [02/13] fixes a minor C/asm ABI mismatch which would trigger a kCFI failure - [03/13] to [09/13] prepare nVHE for CONFIG_CFI_CLANG and [10/13] enables it - [11/13] improves kCFI error messages by saving then parsing the CPU context - [12/13] adds a kCFI test module for VHE and [13/13] extends it to nVHE & pKVM As a result, an informative kCFI panic message is printed by or on behalf of EL2 giving the expected type and target address (possibly resolved to a symbol) for VHE, nVHE, and pKVM (iff CONFIG_NVHE_EL2_DEBUG=y). Note that kCFI errors remain fatal at EL2, even when CONFIG_CFI_PERMISSIVE=y. Changes in v4: - Addressed Will's comments on v3: - Removed save/restore of x0-x1 & used __guest_exit_panic ABI for new routines - Reworked __pkvm_init_switch_pgd new API with separate args - Moved cosmetic changes (renaming to __hyp_panic) into dedicated commit - Further clarified the commit message regarding R_AARCH64_ABS32 - early_brk64() uses esr_is_cfi_brk() (now introduced along esr_brk_comment()) - Added helper to display nvHE panic banner - Moved the test module to the end of the series Pierre-Clément Tosi (13): KVM: arm64: Fix clobbered ELR in sync abort/SError KVM: arm64: Fix __pkvm_init_switch_pgd call ABI KVM: arm64: nVHE: Simplify __guest_exit_panic path KVM: arm64: nVHE: Add EL2h sync exception handler KVM: arm64: Rename __guest_exit_panic __hyp_panic KVM: arm64: nVHE: gen-hyprel: Skip R_AARCH64_ABS32 KVM: arm64: VHE: Mark __hyp_call_panic __noreturn arm64: Introduce esr_comment() & esr_is_cfi_brk() KVM: arm64: Introduce print_nvhe_hyp_panic helper KVM: arm64: nVHE: Support CONFIG_CFI_CLANG at EL2 KVM: arm64: Improve CONFIG_CFI_CLANG error message KVM: arm64: VHE: Add test module for hyp kCFI KVM: arm64: nVHE: Support test module for hyp kCFI arch/arm64/include/asm/esr.h | 11 ++ arch/arm64/include/asm/kvm_asm.h | 3 + arch/arm64/include/asm/kvm_cfi.h | 38 +++++++ arch/arm64/include/asm/kvm_hyp.h | 3 +- arch/arm64/kernel/asm-offsets.c | 1 + arch/arm64/kernel/debug-monitors.c | 4 +- arch/arm64/kernel/traps.c | 8 +- arch/arm64/kvm/Kconfig | 20 ++++ arch/arm64/kvm/Makefile | 3 + arch/arm64/kvm/handle_exit.c | 50 ++++++++- arch/arm64/kvm/hyp/cfi.c | 37 +++++++ arch/arm64/kvm/hyp/entry.S | 34 +++++- arch/arm64/kvm/hyp/hyp-entry.S | 4 +- arch/arm64/kvm/hyp/include/hyp/cfi.h | 47 +++++++++ arch/arm64/kvm/hyp/include/hyp/switch.h | 5 +- arch/arm64/kvm/hyp/nvhe/Makefile | 7 +- arch/arm64/kvm/hyp/nvhe/gen-hyprel.c | 6 ++ arch/arm64/kvm/hyp/nvhe/host.S | 21 ++-- arch/arm64/kvm/hyp/nvhe/hyp-init.S | 23 ++-- arch/arm64/kvm/hyp/nvhe/hyp-main.c | 19 ++++ arch/arm64/kvm/hyp/nvhe/setup.c | 4 +- arch/arm64/kvm/hyp/nvhe/switch.c | 7 ++ arch/arm64/kvm/hyp/vhe/Makefile | 1 + arch/arm64/kvm/hyp/vhe/switch.c | 34 +++++- arch/arm64/kvm/hyp_cfi_test.c | 75 +++++++++++++ arch/arm64/kvm/hyp_cfi_test_module.c | 135 ++++++++++++++++++++++++ 26 files changed, 553 insertions(+), 47 deletions(-) create mode 100644 arch/arm64/include/asm/kvm_cfi.h create mode 100644 arch/arm64/kvm/hyp/cfi.c create mode 100644 arch/arm64/kvm/hyp/include/hyp/cfi.h create mode 100644 arch/arm64/kvm/hyp_cfi_test.c create mode 100644 arch/arm64/kvm/hyp_cfi_test_module.c