[PATCH v3 6/8] bpf ppc64: Access only if addr is kernel address

[Christophe Leroy]

Le 21/09/2021 à 15:29, Hari Bathini a écrit :
> From: Ravi Bangoria <ravi.bangoria at linux.ibm.com>
> 
> On PPC64 with KUAP enabled, any kernel code which wants to
> access userspace needs to be surrounded by disable-enable KUAP.
> But that is not happening for BPF_PROBE_MEM load instruction.
> So, when BPF program tries to access invalid userspace address,
> page-fault handler considers it as bad KUAP fault:
> 
>    Kernel attempted to read user page (d0000000) - exploit attempt? (uid: 0)
> 
> Considering the fact that PTR_TO_BTF_ID (which uses BPF_PROBE_MEM
> mode) could either be a valid kernel pointer or NULL but should
> never be a pointer to userspace address, execute BPF_PROBE_MEM load
> only if addr is kernel address, otherwise set dst_reg=0 and move on.
> 
> This will catch NULL, valid or invalid userspace pointers. Only bad
> kernel pointer will be handled by BPF exception table.
> 
> [Alexei suggested for x86]
> Suggested-by: Alexei Starovoitov <ast at kernel.org>
> Signed-off-by: Ravi Bangoria <ravi.bangoria at linux.ibm.com>
> Signed-off-by: Hari Bathini <hbathini at linux.ibm.com>
> ---
> 
> Changes in v3:
> * Used is_kernel_addr() logic instead of using TASK_SIZE_MAX check
>    all the time.
> * Addressed other comments from Christophe.
> 
> 
>   arch/powerpc/net/bpf_jit_comp64.c | 29 +++++++++++++++++++++++++++++
>   1 file changed, 29 insertions(+)
> 
> diff --git a/arch/powerpc/net/bpf_jit_comp64.c b/arch/powerpc/net/bpf_jit_comp64.c
> index 506934c13ef7..06e1206a4266 100644
> --- a/arch/powerpc/net/bpf_jit_comp64.c
> +++ b/arch/powerpc/net/bpf_jit_comp64.c
> @@ -734,6 +734,35 @@ int bpf_jit_build_body(struct bpf_prog *fp, u32 *image, struct codegen_context *
>   		case BPF_LDX | BPF_MEM | BPF_DW:
>   			fallthrough;
>   		case BPF_LDX | BPF_PROBE_MEM | BPF_DW:
> +			/*
> +			 * As PTR_TO_BTF_ID that uses BPF_PROBE_MEM mode could either be a valid
> +			 * kernel pointer or NULL but not a userspace address, execute BPF_PROBE_MEM
> +			 * load only if addr is kernel address (see is_kernel_addr()), otherwise
> +			 * set dst_reg=0 and move on.
> +			 */
> +			if (BPF_MODE(code) == BPF_PROBE_MEM) {
> +				EMIT(PPC_RAW_ADDI(b2p[TMP_REG_1], src_reg, off));
> +#ifdef CONFIG_PPC_BOOK3E_64

It is better to use IS_ENABLED() whenever possible,

		if (IS_ENABLED((CONFIG_PPC_BOOK3E_64))
			PPC_LI64(b2p[TMP_REG_2], 0x8000000000000000ul);
		else
			PPC_LI64(b2p[TMP_REG_2], PAGE_OFFSET);


> +				PPC_LI64(b2p[TMP_REG_2], 0x8000000000000000ul);
> +#elif defined(CONFIG_PPC_BOOK3S_64)
> +				PPC_LI64(b2p[TMP_REG_2], PAGE_OFFSET);
> +#else
> +				PPC_LI64(b2p[TMP_REG_2], TASK_SIZE);
> +#endif

PPC64 is either CONFIG_PPC_BOOK3S_64 or CONFIG_PPC_BOOK3E_64. The else 
is PPC32.


> +				EMIT(PPC_RAW_CMPLD(b2p[TMP_REG_1], b2p[TMP_REG_2]));
> +				PPC_BCC(COND_GT, (ctx->idx + 4) * 4);
> +				EMIT(PPC_RAW_LI(dst_reg, 0));
> +				/*
> +				 * Check if 'off' is word aligned because PPC_BPF_LL()
> +				 * (BPF_DW case) generates two instructions if 'off' is not
> +				 * word-aligned and one instruction otherwise.
> +				 */
> +				if (BPF_SIZE(code) == BPF_DW && (off & 3))
> +					PPC_JMP((ctx->idx + 3) * 4);
> +				else
> +					PPC_JMP((ctx->idx + 2) * 4);
> +			}
> +
>   			switch (size) {
>   			case BPF_B:
>   				EMIT(PPC_RAW_LBZ(dst_reg, src_reg, off));
>