From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ryusuke Konishi Subject: [PATCH] nilfs2: fix hang problem after bio_alloc() failed Date: Fri, 19 Jun 2009 02:34:55 +0900 (JST) Message-ID: <20090619.023455.111674749.ryusuke@osrg.net> References: <20090614153256.GA4020@homero.springfield.home> <20090615.030245.104791670.konishi.ryusuke@gmail.com> <20090614181313.GA16597@homero.springfield.home> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20090614181313.GA16597@homero.springfield.home> Sender: linux-kernel-owner@vger.kernel.org List-ID: Content-Type: Text/Plain; charset="us-ascii" To: Leandro Lucarella Cc: konishi.ryusuke@lab.ntt.co.jp, linux-kernel@vger.kernel.org, albertito@blitiri.com.ar, users@nilfs.org Hi Leandro, On Sun, 14 Jun 2009 15:13:13 -0300, Leandro Lucarella wrote: > Ryusuke Konishi, el 15 de junio a las 03:02 me escribiste: > [snip] > > > Here is the complete trace: > > > http://pastebin.lugmen.org.ar/4931 > > > > Thank you for your help. > > > > According to your log, there seems to be a leakage in clear processing > > of the writeback flag on pages. I will review the error path of log > > writer to narrow down the cause. Here is a patch that hopefully fixes the hang problem after bio_alloc() failed. This is composed of three separate patches, but I now attach them together. If you still see problems with it, please let me know. > Oh! I forgot to tell you that the cleanerd process was not running. When > I mounted the NILFS2 filesystem the cleanerd daemon said: > > nilfs_cleanerd[29534]: start > nilfs_cleanerd[29534]: cannot create cleanerd on /dev/loop0 > nilfs_cleanerd[29534]: shutdown > > I thought I might have an old utilities version and that could be because > of that (since the nilfs website was down I couldn't check), but now > I checked in the backup nilfs website that latest version is 2.0.12 and > I have that (Debian package), so I don't know why the cleanerd failed to > start in the first place. The message indicates that initialization of the cleanerd failed. I don't know for sure, but another memory allocation failure might cause it. Thanks, Ryusuke Konishi --- Ryusuke Konishi (3): nilfs2: fix mis-conversion of error code by unlikely directive nilfs2: fix hang problem of log writer that follows on write failures nilfs2: remove incorrect warning in nilfs_dat_commit_start function fs/nilfs2/dat.c | 9 --------- fs/nilfs2/segment.c | 28 +++++++--------------------- 2 files changed, 7 insertions(+), 30 deletions(-) diff --git a/fs/nilfs2/dat.c b/fs/nilfs2/dat.c index bb8a581..e2646c3 100644 --- a/fs/nilfs2/dat.c +++ b/fs/nilfs2/dat.c @@ -149,15 +149,6 @@ void nilfs_dat_commit_start(struct inode *dat, struct nilfs_palloc_req *req, entry = nilfs_palloc_block_get_entry(dat, req->pr_entry_nr, req->pr_entry_bh, kaddr); entry->de_start = cpu_to_le64(nilfs_mdt_cno(dat)); - if (entry->de_blocknr != cpu_to_le64(0) || - entry->de_end != cpu_to_le64(NILFS_CNO_MAX)) { - printk(KERN_CRIT - "%s: vbn = %llu, start = %llu, end = %llu, pbn = %llu\n", - __func__, (unsigned long long)req->pr_entry_nr, - (unsigned long long)le64_to_cpu(entry->de_start), - (unsigned long long)le64_to_cpu(entry->de_end), - (unsigned long long)le64_to_cpu(entry->de_blocknr)); - } entry->de_blocknr = cpu_to_le64(blocknr); kunmap_atomic(kaddr, KM_USER0); diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c index 22c7f65..e8f188b 100644 --- a/fs/nilfs2/segment.c +++ b/fs/nilfs2/segment.c @@ -1846,26 +1846,13 @@ static int nilfs_segctor_write(struct nilfs_sc_info *sci, err = nilfs_segbuf_write(segbuf, &wi); res = nilfs_segbuf_wait(segbuf, &wi); - err = unlikely(err) ? : res; + err = unlikely(err) ? err : res; if (unlikely(err)) return err; } return 0; } -static int nilfs_page_has_uncleared_buffer(struct page *page) -{ - struct buffer_head *head, *bh; - - head = bh = page_buffers(page); - do { - if (buffer_dirty(bh) && !list_empty(&bh->b_assoc_buffers)) - return 1; - bh = bh->b_this_page; - } while (bh != head); - return 0; -} - static void __nilfs_end_page_io(struct page *page, int err) { if (!err) { @@ -1889,12 +1876,11 @@ static void nilfs_end_page_io(struct page *page, int err) if (!page) return; - if (buffer_nilfs_node(page_buffers(page)) && - nilfs_page_has_uncleared_buffer(page)) - /* For b-tree node pages, this function may be called twice - or more because they might be split in a segment. - This check assures that cleanup has been done for all - buffers in a split btnode page. */ + if (buffer_nilfs_node(page_buffers(page)) && !PageWriteback(page)) + /* + * For b-tree node pages, this function may be called twice + * or more because they might be split in a segment. + */ return; __nilfs_end_page_io(page, err); @@ -1957,7 +1943,7 @@ static void nilfs_segctor_abort_write(struct nilfs_sc_info *sci, } if (bh->b_page != fs_page) { nilfs_end_page_io(fs_page, err); - if (unlikely(fs_page == failed_page)) + if (fs_page && fs_page == failed_page) goto done; fs_page = bh->b_page; } -- 1.6.2