Date:   Thu, 18 Nov 2021 19:22:21 +0100
From:   Peter Zijlstra <peterz@infradead.org>
To:     Eric Dumazet <eric.dumazet@gmail.com>
Cc:     linux-kernel <linux-kernel@vger.kernel.org>,
        Eric Dumazet <edumazet@google.com>,
        Alexander Duyck <alexanderduyck@fb.com>,
        Johannes Berg <johannes.berg@intel.com>,
        kernel test robot <lkp@intel.com>
Subject: Re: [PATCH] x86/csum: fix compilation error for UM
Message-ID: <20211118182221.GI174703@worktop.programming.kicks-ass.net>
References: <20211118175239.1525650-1-eric.dumazet@gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20211118175239.1525650-1-eric.dumazet@gmail.com>
Precedence: bulk

On Thu, Nov 18, 2021 at 09:52:39AM -0800, Eric Dumazet wrote:
> From: Eric Dumazet <edumazet@google.com>
> 
> load_unaligned_zeropad() is not yet universal.
> 
> ARCH=um SUBARCH=x86_64 builds do not have it.
> 
> When CONFIG_DCACHE_WORD_ACCESS is not set, simply continue
> the bisection with 4, 2 and 1 byte steps.
> 
> Fixes: df4554cebdaa ("x86/csum: Rewrite/optimize csum_partial()")
> Signed-off-by: Eric Dumazet <edumazet@google.com>
> Cc: Peter Zijlstra (Intel) <peterz@infradead.org>
> Cc: Alexander Duyck <alexanderduyck@fb.com>
> Cc: Johannes Berg <johannes.berg@intel.com>
> Reported-by: kernel test robot <lkp@intel.com>
> ---

Yeah, that's much nicer. I'll go feed that to the robots I suppose :-)

>  arch/x86/lib/csum-partial_64.c | 26 ++++++++++++++++++++++++++
>  1 file changed, 26 insertions(+)
> 
> diff --git a/arch/x86/lib/csum-partial_64.c b/arch/x86/lib/csum-partial_64.c
> index 5ec35626945b6db2f7f41c6d46d5e422810eac46..1eb8f2d11f7c785be624eba315fe9ca7989fd56d 100644
> --- a/arch/x86/lib/csum-partial_64.c
> +++ b/arch/x86/lib/csum-partial_64.c
> @@ -92,6 +92,7 @@ __wsum csum_partial(const void *buff, int len, __wsum sum)
>  		buff += 8;
>  	}
>  	if (len & 7) {
> +#ifdef CONFIG_DCACHE_WORD_ACCESS
>  		unsigned int shift = (8 - (len & 7)) * 8;
>  		unsigned long trail;
>  
> @@ -101,6 +102,31 @@ __wsum csum_partial(const void *buff, int len, __wsum sum)
>  		    "adcq $0,%[res]"
>  			: [res] "+r" (temp64)
>  			: [trail] "r" (trail));
> +#else
> +		if (len & 4) {
> +			asm("addq %[val],%[res]\n\t"
> +			    "adcq $0,%[res]"
> +				: [res] "+r" (temp64)
> +				: [val] "r" ((u64)*(u32 *)buff)
> +				: "memory");
> +			buff += 4;
> +		}
> +		if (len & 2) {
> +			asm("addq %[val],%[res]\n\t"
> +			    "adcq $0,%[res]"
> +				: [res] "+r" (temp64)
> +				: [val] "r" ((u64)*(u16 *)buff)
> +				: "memory");
> +			buff += 2;
> +		}
> +		if (len & 1) {
> +			asm("addq %[val],%[res]\n\t"
> +			    "adcq $0,%[res]"
> +				: [res] "+r" (temp64)
> +				: [val] "r" ((u64)*(u8 *)buff)
> +				: "memory");
> +		}
> +#endif
>  	}
>  	result = add32_with_carry(temp64 >> 32, temp64 & 0xffffffff);
>  	if (unlikely(odd)) { 
> -- 
> 2.34.0.rc1.387.gb447b232ab-goog
>