[+Daniel and Jean-Philippe]
On Thu, May 23, 2019 at 05:12:00PM +0900, Yoshihiro Shimoda wrote:
> The following build warning happens on gcc 8.1.0.
>
> linux/arch/arm64/include/asm/insn.h: In function 'aarch64_insn_is_ldadd':
> linux/arch/arm64/include/asm/insn.h:280:257: warning: bitwise comparison always evaluates to false [-Wtautological-compare]
> __AARCH64_INSN_FUNCS(ldadd, 0x3F20FC00, 0xB8200000)
>
> Since the second argument is mask value and compare with the third
> argument value, the bit 31 is always masked and then this macro is
> always false. So, this patch fixes the issue.
>
> Reported-by: Kuninori Morimoto <kuninori.morimoto.gx@xxxxxxxxxxx>
> Fixes: 34b8ab091f9ef57a ("bpf, arm64: use more scalable stadd over ldxr / stxr loop in xadd")
> Tested-by: Kuninori Morimoto <kuninori.morimoto.gx@xxxxxxxxxxx>
> Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@xxxxxxxxxxx>
> ---
> I'm not sure the second argument "0xBF20FC00" is OK or not (we can set
> to 0xFF20FC00 instead). So, I marked RFC on this patch.
>
> arch/arm64/include/asm/insn.h | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/arch/arm64/include/asm/insn.h b/arch/arm64/include/asm/insn.h
> index ec894de..c9e3cdc 100644
> --- a/arch/arm64/include/asm/insn.h
> +++ b/arch/arm64/include/asm/insn.h
> @@ -277,7 +277,7 @@ __AARCH64_INSN_FUNCS(adrp, 0x9F000000, 0x90000000)
> __AARCH64_INSN_FUNCS(prfm, 0x3FC00000, 0x39800000)
> __AARCH64_INSN_FUNCS(prfm_lit, 0xFF000000, 0xD8000000)
> __AARCH64_INSN_FUNCS(str_reg, 0x3FE0EC00, 0x38206800)
> -__AARCH64_INSN_FUNCS(ldadd, 0x3F20FC00, 0xB8200000)
> +__AARCH64_INSN_FUNCS(ldadd, 0xBF20FC00, 0xB8200000)
Looking at the ISA encoding, I think that top digit should indeed be 'B',
but I haven't checked the rest of the instruction.
However, I'm fairly sure we tested this so now I'm a bit worried that I'm
missing something :/
Will
