linux-ext4.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jan Kara <jack@suse.cz>
To: David Rientjes <rientjes@google.com>
Cc: Andrew Morton <akpm@linux-foundation.org>,
	Steven Whitehouse <swhiteho@redhat.com>,
	Jens Axboe <jaxboe@fusionio.com>, Jan Kara <jack@suse.cz>,
	cluster-devel@redhat.com, linux-ext4@vger.kernel.org,
	linux-kernel@vger.kernel.org
Subject: Re: [patch 3/5] fs: add nofail variant of alloc_buffer_head
Date: Tue, 24 Aug 2010 14:17:03 +0200	[thread overview]
Message-ID: <20100824121702.GC3713@quack.suse.cz> (raw)
In-Reply-To: <alpine.DEB.2.00.1008240347070.19596@chino.kir.corp.google.com>

On Tue 24-08-10 03:50:30, David Rientjes wrote:
> Add alloc_buffer_head_nofail().  This function is equivalent to
> alloc_buffer_head(), except that it will never return NULL and instead
> loop forever trying to allocate memory.
> 
> If the first allocation attempt fails, a warning will be emitted,
> including a call trace.  Subsequent failures will suppress this warning.
> 
> This was added as a helper function for documentation and auditability.
> No future callers should be added.
  Acked-by: Jan Kara <jack@suse.cz>
for the JBD part here as well.

								Honza
> 
> Signed-off-by: David Rientjes <rientjes@google.com>
> ---
>  fs/buffer.c                 |   18 ++++++++++++++++++
>  fs/gfs2/log.c               |    2 +-
>  fs/jbd/journal.c            |    2 +-
>  include/linux/buffer_head.h |    1 +
>  4 files changed, 21 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/buffer.c b/fs/buffer.c
> --- a/fs/buffer.c
> +++ b/fs/buffer.c
> @@ -3238,6 +3238,24 @@ struct buffer_head *alloc_buffer_head(gfp_t gfp_flags)
>  }
>  EXPORT_SYMBOL(alloc_buffer_head);
>  
> +/*
> + * NOTE: no new callers of this function should be implemented!
> + * All memory allocations should be failable whenever possible.
> + */
> +struct buffer_head *alloc_buffer_head_nofail(gfp_t gfp_flags)
> +{
> +	struct buffer_head *ret;
> +
> +	for (;;) {
> +		ret = alloc_buffer_head(gfp_flags);
> +		if (ret)
> +			return ret;
> +		WARN_ONCE(1, "Out of memory; no fallback implemented "
> +				"(flags=0x%x)\n",
> +				gfp_flags);
> +	}
> +}
> +
>  void free_buffer_head(struct buffer_head *bh)
>  {
>  	BUG_ON(!list_empty(&bh->b_assoc_buffers));
> diff --git a/fs/gfs2/log.c b/fs/gfs2/log.c
> --- a/fs/gfs2/log.c
> +++ b/fs/gfs2/log.c
> @@ -523,7 +523,7 @@ struct buffer_head *gfs2_log_fake_buf(struct gfs2_sbd *sdp,
>  	u64 blkno = log_bmap(sdp, sdp->sd_log_flush_head);
>  	struct buffer_head *bh;
>  
> -	bh = alloc_buffer_head(GFP_NOFS | __GFP_NOFAIL);
> +	bh = alloc_buffer_head_nofail(GFP_NOFS);
>  	atomic_set(&bh->b_count, 1);
>  	bh->b_state = (1 << BH_Mapped) | (1 << BH_Uptodate) | (1 << BH_Lock);
>  	set_bh_page(bh, real->b_page, bh_offset(real));
> diff --git a/fs/jbd/journal.c b/fs/jbd/journal.c
> --- a/fs/jbd/journal.c
> +++ b/fs/jbd/journal.c
> @@ -301,7 +301,7 @@ int journal_write_metadata_buffer(transaction_t *transaction,
>  	 */
>  	J_ASSERT_BH(bh_in, buffer_jbddirty(bh_in));
>  
> -	new_bh = alloc_buffer_head(GFP_NOFS|__GFP_NOFAIL);
> +	new_bh = alloc_buffer_head_nofail(GFP_NOFS);
>  	/* keep subsequent assertions sane */
>  	new_bh->b_state = 0;
>  	init_buffer(new_bh, NULL, NULL);
> diff --git a/include/linux/buffer_head.h b/include/linux/buffer_head.h
> --- a/include/linux/buffer_head.h
> +++ b/include/linux/buffer_head.h
> @@ -176,6 +176,7 @@ void __breadahead(struct block_device *, sector_t block, unsigned int size);
>  struct buffer_head *__bread(struct block_device *, sector_t block, unsigned size);
>  void invalidate_bh_lrus(void);
>  struct buffer_head *alloc_buffer_head(gfp_t gfp_flags);
> +struct buffer_head *alloc_buffer_head_nofail(gfp_t gfp_flags);
>  void free_buffer_head(struct buffer_head * bh);
>  void unlock_buffer(struct buffer_head *bh);
>  void __lock_buffer(struct buffer_head *bh);
-- 
Jan Kara <jack@suse.cz>
SUSE Labs, CR

  reply	other threads:[~2010-08-24 12:17 UTC|newest]

Thread overview: 56+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-08-24 10:50 [patch 1/5] mm: add nofail variants of kmalloc kcalloc and kzalloc David Rientjes
2010-08-24 10:50 ` [patch 3/5] fs: add nofail variant of alloc_buffer_head David Rientjes
2010-08-24 12:17   ` Jan Kara [this message]
2010-08-24 12:15 ` [patch 1/5] mm: add nofail variants of kmalloc kcalloc and kzalloc Jan Kara
2010-08-24 13:29 ` Peter Zijlstra
2010-08-24 13:33   ` Jens Axboe
2010-08-24 20:11     ` David Rientjes
2010-08-25 11:24       ` Ted Ts'o
2010-08-25 11:35         ` Peter Zijlstra
2010-08-25 11:57           ` Ted Ts'o
2010-08-25 12:48             ` Peter Zijlstra
2010-08-25 12:52               ` Peter Zijlstra
2010-08-25 13:20                 ` Theodore Tso
2010-08-25 13:31                   ` Peter Zijlstra
2010-08-25 20:43                     ` David Rientjes
2010-08-25 20:55                       ` Peter Zijlstra
2010-08-25 21:11                         ` David Rientjes
2010-08-25 21:27                           ` Peter Zijlstra
2010-08-25 23:11                             ` David Rientjes
2010-08-26  0:19                               ` Ted Ts'o
2010-08-26  0:30                                 ` David Rientjes
     [not found]                                 ` <alpine.DEB.2.00.1008251724360.25783@chino.kir.corp.google.com>
2010-08-26  1:48                                   ` Ted Ts'o
2010-08-26  3:09                                     ` David Rientjes
2010-08-26  7:06                                       ` Dave Chinner
2010-08-26  8:29                                       ` Peter Zijlstra
2010-08-26  6:38                                     ` Dave Chinner
2010-08-25 13:34                   ` Peter Zijlstra
2010-08-25 13:24               ` Dave Chinner
2010-08-25 13:35                 ` Peter Zijlstra
2010-08-25 20:53                   ` Ted Ts'o
2010-08-25 20:59                     ` David Rientjes
2010-08-25 21:35                     ` Peter Zijlstra
2010-08-25 20:58                   ` David Rientjes
2010-08-25 21:11                     ` Christoph Lameter
2010-08-25 21:21                       ` Peter Zijlstra
2010-08-25 21:23                       ` David Rientjes
2010-08-25 21:35                         ` Christoph Lameter
2010-08-25 23:05                           ` David Rientjes
2010-08-26  1:30                             ` Christoph Lameter
2010-08-26  3:12                               ` David Rientjes
2010-08-26 14:16                                 ` Christoph Lameter
2010-08-26 22:31                                   ` David Rientjes
2010-08-26  0:09                   ` Dave Chinner
2010-08-25 14:13         ` Peter Zijlstra
2010-08-24 13:55   ` Dave Chinner
2010-08-24 14:03     ` Peter Zijlstra
2010-08-24 20:12     ` David Rientjes
2010-08-24 20:08   ` David Rientjes
2010-09-02  1:02 ` [patch v2 " David Rientjes
2010-09-02  1:03   ` [patch v2 3/5] fs: add nofail variant of alloc_buffer_head David Rientjes
2010-09-02  7:59   ` [patch v2 1/5] mm: add nofail variants of kmalloc kcalloc and kzalloc Jiri Slaby
2010-09-02 14:51     ` Jan Kara
2010-09-02 21:15       ` Neil Brown
2010-09-05 23:03         ` David Rientjes
2010-09-05 23:01     ` David Rientjes
2010-09-06  9:05   ` David Rientjes

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=20100824121702.GC3713@quack.suse.cz \
    --to=jack@suse.cz \
    --cc=akpm@linux-foundation.org \
    --cc=cluster-devel@redhat.com \
    --cc=jaxboe@fusionio.com \
    --cc=linux-ext4@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=rientjes@google.com \
    --cc=swhiteho@redhat.com \
    /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
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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).