Received-SPF: pass (google.com: domain of linux-kernel+bounces-145864-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
Date: Mon, 15 Apr 2024 13:10:53 -0700
From: Andrew Morton <akpm@linux-foundation.org>
To: Long Li <leo.lilong@huawei.com>
Cc: <willy@infradead.org>, <linux-fsdevel@vger.kernel.org>,
 <linux-kernel@vger.kernel.org>, <yi.zhang@huawei.com>,
 <houtao1@huawei.com>, <yangerkun@huawei.com>
Subject: Re: [PATCH] xarray: inline xas_descend to improve performance
Message-Id: <20240415131053.051e60135eacf281df6921f6@linux-foundation.org>
In-Reply-To: <20240415012136.3636671-1-leo.lilong@huawei.com>
References: <20240415012136.3636671-1-leo.lilong@huawei.com>
Precedence: bulk
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

On Mon, 15 Apr 2024 09:21:36 +0800 Long Li <leo.lilong@huawei.com> wrote:

> The commit 63b1898fffcd ("XArray: Disallow sibling entries of nodes")
> modified the xas_descend function in such a way that it was no longer
> being compiled as an inline function, because it increased the size of
> xas_descend(), and the compiler no longer optimizes it as inline. This
> had a negative impact on performance, xas_descend is called frequently
> to traverse downwards in the xarray tree, making it a hot function.
> 
> Inlining xas_descend has been shown to significantly improve performance
> by approximately 4.95% in the iozone write test.
> 
>   Machine: Intel(R) Xeon(R) Gold 6240 CPU @ 2.60GHz
>   #iozone i 0 -i 1 -s 64g -r 16m -f /test/tmptest
> 
> Before this patch:
> 
>        kB    reclen    write   rewrite     read    reread
>  67108864     16384  2230080   3637689 6 315197   5496027
> 
> After this patch:
> 
>        kB    reclen    write   rewrite     read    reread
>  67108864     16384  2340360   3666175  6272401   5460782
> 
> Percentage change:
>                        4.95%     0.78%   -0.68%    -0.64%
> 
> This patch introduces inlining to the xas_descend function. While this
> change increases the size of lib/xarray.o, the performance gains in
> critical workloads make this an acceptable trade-off.
> 
> Size comparison before and after patch:
> .text		.data		.bss		file
> 0x3502		    0		   0		lib/xarray.o.before
> 0x3602		    0		   0		lib/xarray.o.after
> 
> ...
>
> --- a/lib/xarray.c
> +++ b/lib/xarray.c
> @@ -200,7 +200,7 @@ static void *xas_start(struct xa_state *xas)
>  	return entry;
>  }
>  
> -static void *xas_descend(struct xa_state *xas, struct xa_node *node)
> +static inline void *xas_descend(struct xa_state *xas, struct xa_node *node)
>  {
>  	unsigned int offset = get_offset(xas->xa_index, node);
>  	void *entry = xa_entry(xas->xa, node, offset);

I thought gcc nowadays treats `inline' as avisory and still makes up
its own mind?

Perhaps we should use __always_inline here?