[PATCH v13 8/8] arm64: expose FAR_EL1 tag bits in siginfo

[Catalin Marinas]

On Mon, Nov 02, 2020 at 08:09:44PM -0800, Peter Collingbourne wrote:
> The kernel currently clears the tag bits (i.e. bits 56-63) in the fault
> address exposed via siginfo.si_addr and sigcontext.fault_address. However,
> the tag bits may be needed by tools in order to accurately diagnose
> memory errors, such as HWASan [1] or future tools based on the Memory
> Tagging Extension (MTE).
> 
> We should not stop clearing these bits in the existing fault address
> fields, because there may be existing userspace applications that are
> expecting the tag bits to be cleared. Instead, create a new pair of
> union fields in siginfo._sigfault, and store the tag bits of FAR_EL1
> there, together with a mask specifying which bits are valid.

This comment is slightly confusing as they are not union fields in
_sigfault.

> diff --git a/include/uapi/asm-generic/siginfo.h b/include/uapi/asm-generic/siginfo.h
> index f43778355b77..2b2ed0394457 100644
> --- a/include/uapi/asm-generic/siginfo.h
> +++ b/include/uapi/asm-generic/siginfo.h
> @@ -94,6 +94,8 @@ union __sifields {
>  			void *_pad[6];
>  		};
>  		__u32 _faultflags;
> +		unsigned long _addr_tag_bits;
> +		unsigned long _addr_tag_bits_mask;
>  	} _sigfault;

Can we not actually add these as a struct to the union? Do we expect the
other fields to be valid?

Also there's a 32-bit gap I mentioned on the previous patch between
_faultflags and _addr_tag_bits.

That said, I wonder whether we could solve this for MTE without new
fields by always setting the tag in si_addr when si_code is SEGV_MTE*.
Alternatively, we could add a prctl() bit to require tagged si_addr.

Well, I don't mind the _addr_tag_bits* fields if they are part of the
union and keep si_addr intact.

-- 
Catalin