LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Dmitriy Monakhov <dmonakhov@sw.ru>
To: Nick Piggin <npiggin@suse.de>
Cc: Linux Memory Management <linux-mm@kvack.org>,
	Linux Kernel <linux-kernel@vger.kernel.org>,
	Linux Filesystems <linux-fsdevel@vger.kernel.org>,
	Andrew Morton <akpm@osdl.org>
Subject: Re: [patch 6/10] mm: be sure to trim blocks
Date: Sun, 14 Jan 2007 17:25:44 +0300	[thread overview]
Message-ID: <87bql1lm6v.fsf@sw.ru> (raw)
In-Reply-To: <20070113011255.9449.33228.sendpatchset@linux.site> (Nick Piggin's message of "Sat, 13 Jan 2007 04:25:11 +0100 (CET)")

Nick Piggin <npiggin@suse.de> writes:

> If prepare_write fails with AOP_TRUNCATED_PAGE, or if commit_write fails, then
> we may have failed the write operation despite prepare_write having
> instantiated blocks past i_size. Fix this, and consolidate the trimming into
> one place.
>
> Signed-off-by: Nick Piggin <npiggin@suse.de>
>
> Index: linux-2.6/mm/filemap.c
> ===================================================================
> --- linux-2.6.orig/mm/filemap.c
> +++ linux-2.6/mm/filemap.c
> @@ -1911,22 +1911,9 @@ generic_file_buffered_write(struct kiocb
>  		}
>  
>  		status = a_ops->prepare_write(file, page, offset, offset+bytes);
> -		if (unlikely(status)) {
> -			loff_t isize = i_size_read(inode);
> +		if (unlikely(status))
> +			goto fs_write_aop_error;
May be it's stupid question but still..
Why we treat non zero prepare_write() return code as error, it may be positive.
Positive error code may be used as fine grained 'bytes' limiter in case of 
blksize < pgsize as follows:

                status = a_ops->prepare_write(file, page, offset, offset+bytes);
		if (unlikely(status)) {
                        if (status > 0) {
                                bytes = min(bytes, status);
                                status = 0;
                        } else {
                	        goto fs_write_aop_error;
                        }
                }
---
This is useful because fs may want to reduce 'bytes' by number of reasons,
for example make it blksize bound. 
Example : filesystem has 1k blksize and only two free blocks. And we try 
write 4k bytes.
Currently  write(fd, buff, 4096) will return -ENOSPC
But after this fix write(fd, buff, 4096) will return as mutch as it can (2048).
>  
> -			if (status != AOP_TRUNCATED_PAGE)
> -				unlock_page(page);
> -			page_cache_release(page);
> -			if (status == AOP_TRUNCATED_PAGE)
> -				continue;
> -			/*
> -			 * prepare_write() may have instantiated a few blocks
> -			 * outside i_size.  Trim these off again.
> -			 */
> -			if (pos + bytes > isize)
> -				vmtruncate(inode, isize);
> -			break;
> -		}
>  		if (likely(nr_segs == 1))
>  			copied = filemap_copy_from_user(page, offset,
>  							buf, bytes);
> @@ -1935,10 +1922,9 @@ generic_file_buffered_write(struct kiocb
>  						cur_iov, iov_offset, bytes);
>  		flush_dcache_page(page);
>  		status = a_ops->commit_write(file, page, offset, offset+bytes);
> -		if (status == AOP_TRUNCATED_PAGE) {
> -			page_cache_release(page);
> -			continue;
> -		}
> +		if (unlikely(status))
> +			goto fs_write_aop_error;
> +
>  		if (likely(copied > 0)) {
>  			if (!status)
>  				status = copied;
> @@ -1969,6 +1955,25 @@ generic_file_buffered_write(struct kiocb
>  			break;
>  		balance_dirty_pages_ratelimited(mapping);
>  		cond_resched();
> +		continue;
> +
> +fs_write_aop_error:
> +		if (status != AOP_TRUNCATED_PAGE)
> +			unlock_page(page);
> +		page_cache_release(page);
> +
> +		/*
> +		 * prepare_write() may have instantiated a few blocks
> +		 * outside i_size.  Trim these off again. Don't need
> +		 * i_size_read because we hold i_mutex.
> +		 */
> +		if (pos + bytes > inode->i_size)
> +			vmtruncate(inode, inode->i_size);
> +		if (status == AOP_TRUNCATED_PAGE)
> +			continue;
> +		else
> +			break;
> +
>  	} while (count);
>  	*ppos = pos;
>  
> -
> 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/


  reply	other threads:[~2007-01-14 14:25 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-01-13  3:24 [patch 0/10] buffered write deadlock fix Nick Piggin
2007-01-13  3:24 ` [patch 1/10] fs: libfs buffered write leak fix Nick Piggin
2007-01-13  3:24 ` [patch 2/10] mm: revert "generic_file_buffered_write(): handle zero length iovec segments" Nick Piggin
2007-01-13  3:24 ` [patch 3/10] mm: revert "generic_file_buffered_write(): deadlock on vectored write" Nick Piggin
2007-01-13  3:24 ` [patch 4/10] mm: generic_file_buffered_write cleanup Nick Piggin
2007-01-13  3:25 ` [patch 5/10] mm: debug write deadlocks Nick Piggin
2007-01-13  3:25 ` [patch 6/10] mm: be sure to trim blocks Nick Piggin
2007-01-14 14:25   ` Dmitriy Monakhov [this message]
2007-01-20  3:50     ` Nick Piggin
2007-01-16 17:36   ` Peter Zijlstra
2007-01-16 19:14     ` Peter Zijlstra
2007-01-20  3:52       ` Nick Piggin
2007-01-13  3:25 ` [patch 7/10] mm: cleanup pagecache insertion operations Nick Piggin
2007-01-13  3:25 ` [patch 8/10] mm: generic_file_buffered_write cleanup more Nick Piggin
2007-01-13  3:25 ` [patch 9/10] mm: generic_file_buffered_write iovec cleanup Nick Piggin
2007-01-13  3:25 ` [patch 10/10] mm: fix pagecache write deadlocks Nick Piggin
2007-01-14  3:59   ` Nick Piggin

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=87bql1lm6v.fsf@sw.ru \
    --to=dmonakhov@sw.ru \
    --cc=akpm@osdl.org \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=npiggin@suse.de \
    --subject='Re: [patch 6/10] mm: be sure to trim blocks' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).