Date: Tue, 15 Dec 2015 08:45:20 -0500 (EST)
From: Bob Peterson <rpeterso@redhat.com>
To: Abhi Das <adas@redhat.com>
Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org,
        swhiteho@redhat.com
Message-ID: <827318549.33702096.1450187120615.JavaMail.zimbra@redhat.com>
In-Reply-To: <1450143229-40537-1-git-send-email-adas@redhat.com>
References: <1450143229-40537-1-git-send-email-adas@redhat.com>
Subject: Re: [RFC PATCH] fs: __generic_file_splice_read retry lookup on
 AOP_TRUNCATED_PAGE
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit
Thread-Topic: __generic_file_splice_read retry lookup on AOP_TRUNCATED_PAGE
Thread-Index: dq+62dDQM1qR2UOwoUHNM8kz+T6g+A==
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 2748
Lines: 76

Hi Abhi,

Comment below.

----- Original Message -----
> During testing, I discovered that __generic_file_splice_read() returns
> 0 (EOF) when aops->readpage fails with AOP_TRUNCATED_PAGE on the first
> page of a single/multi-page splice read operation. This EOF return code
> causes the userspace test to (correctly) report a zero-length read error
> when it was expecting otherwise.
> 
> The current strategy of returning a partial non-zero read when ->readpage
> returns AOP_TRUNCATED_PAGE works only when the failed page is not the
> first of the lot being processed.
> 
> This patch attempts to retry lookup and call ->readpage again on pages
> that had previously failed with AOP_TRUNCATED_PAGE. With this patch, my
> tests pass and I haven't noticed any unwanted side effects.
> 
> Signed-off-by: Abhi Das <adas@redhat.com>
> ---
>  fs/splice.c | 12 +++++++-----
>  1 file changed, 7 insertions(+), 5 deletions(-)
> 
> diff --git a/fs/splice.c b/fs/splice.c
> index 801c21c..c424d44 100644
> --- a/fs/splice.c
> +++ b/fs/splice.c
> @@ -387,6 +387,7 @@ __generic_file_splice_read(struct file *in, loff_t *ppos,
>  	spd.nr_pages = 0;
>  	for (page_nr = 0; page_nr < nr_pages; page_nr++) {
>  		unsigned int this_len;
> +		int retries = 0;
>  
>  		if (!len)
>  			break;
> @@ -415,6 +416,7 @@ __generic_file_splice_read(struct file *in, loff_t *ppos,
>  			 */
>  			if (!page->mapping) {
>  				unlock_page(page);
> +retry_lookup:
>  				page = find_or_create_page(mapping, index,
>  						mapping_gfp_mask(mapping));
>  
> @@ -439,13 +441,13 @@ __generic_file_splice_read(struct file *in, loff_t
> *ppos,
>  			error = mapping->a_ops->readpage(in, page);
>  			if (unlikely(error)) {
>  				/*
> -				 * We really should re-lookup the page here,
> -				 * but it complicates things a lot. Instead
> -				 * lets just do what we already stored, and
> -				 * we'll get it the next time we are called.
> +				 * Re-lookup the page
>  				 */
> -				if (error == AOP_TRUNCATED_PAGE)
> +				if (error == AOP_TRUNCATED_PAGE) {
> +					if (retries++ < 3)
> +						goto retry_lookup;
>  					error = 0;

I have no problem with doing the retry. However, I think we should probably set
error = 0 before we decide whether to retry here. Otherwise, if error is
AOP_TRUNCATED_PAGE, it might jump to retry_lookup, but then discover that the
page is PageUptodate(), goto fill_it with error still set to AOP_TRUNCATED_PAGE,
then end up returning that to the caller.

Regards,

Bob Peterson
Red Hat File Systems
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/