From: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>
To: Mingming Cao <cmm@us.ibm.com>
Cc: tytso <tytso@mit.edu>, Shehjar Tikoo <shehjart@cse.unsw.edu.au>,
linux-ext4@vger.kernel.org
Subject: Re: [RFC]Ext4: journal credits reservation fixes for DIO, fallocate and delalloc writepages
Date: Tue, 29 Jul 2008 11:54:27 +0530 [thread overview]
Message-ID: <20080729062427.GC4545@skywalker> (raw)
In-Reply-To: <1217272053.6379.7.camel@mingming-laptop>
On Mon, Jul 28, 2008 at 12:07:33PM -0700, Mingming Cao wrote:
.....
> > > > > *
> > > > > @@ -1142,13 +1133,13 @@ int ext4_get_block(struct inode *inode,
> > > > > handle_t *handle = ext4_journal_current_handle();
> > > > > int ret = 0, started = 0;
> > > > > unsigned max_blocks = bh_result->b_size >> inode->i_blkbits;
> > > > > + int dio_credits = EXT4_DATA_TRANS_BLOCKS(inode->i_sb);
> > > >
> > > > Again this should be
> > > > int dio_credits = ext4_writeblocks_trans(inode, DIO_MAX_BLOCKS);
> > > >
> > > No. DIO case is different than writepages(). When get_block() is called
> > > from DIO path, the get_block() is called in a loop, and the credit is
> > > reserved inside the loop. Each time get_block(),will return a single
> > > extent, so we should use EXT4_DATA_TRANS_BLOCKS(inode->i_sb) which is
> > > calculate a single chunk of allocation credits.
> >
> >
> > That is true only for extent format. With noextents we need something
> > like below
> >
>
> Not really, even with non extent format block allocation, ext4_get_block() will only allocate/map a chunk of contiguous blocks (i.e. an extent), so it will not hit the worse case.
>
> > if (!(EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL))
> > dio_credits = ext4_writeblocks_trans_credits_old(inode, max_blocks);
> > else {
> > /*
> > * For extent format get_block return only one extent
> > */
> > dio_credits = EXT4_DATA_TRANS_BLOCKS(inode->i_sb);
> > }
> >
> >
>
but dio_credits = EXT4_DATA_TRANS_BLOCKS(inode->i_sb); doesn't account
for updating the bitmap and group descriptor related to all indirect and
dindirect and tindirect blocks allocated.
We actually need
a) 1 bitmap for the blocks allocatted
b) 1 group desc for the blocks allocatted
c) 2 indirect
d) 2 dindirect
e) 1 tindirect
f) 5 bitmap for the indirect blocks allocated (2 + 2 + 1)
g) 5 group desc for the indirect blocks allocated
h) 1 inode
i) 1 super blocks
j) 2 * EXT4_QUOTA_TRANS_BLOCKS for quota
-aneesh
next prev parent reply other threads:[~2008-07-29 6:25 UTC|newest]
Thread overview: 61+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-07-21 4:28 Crash and stack trace for jbd2 under ext4 Shehjar Tikoo
2008-07-21 8:20 ` Aneesh Kumar K.V
2008-07-23 0:49 ` Mingming Cao
2008-07-23 0:51 ` [RFC]Ext4: journal credits reservation fixes for DIO, fallocate and delalloc writepages Mingming Cao
2008-07-23 1:18 ` Andreas Dilger
2008-07-23 18:19 ` Theodore Tso
2008-07-25 19:38 ` Mingming Cao
2008-07-23 7:42 ` Aneesh Kumar K.V
2008-07-24 2:07 ` Andreas Dilger
2008-07-25 19:26 ` Mingming Cao
2008-07-28 16:11 ` Aneesh Kumar K.V
2008-07-28 19:07 ` Mingming Cao
2008-07-29 6:24 ` Aneesh Kumar K.V [this message]
2008-07-26 0:42 ` [PATCH v2]Ext4: " Mingming Cao
2008-07-30 1:58 ` [PATCH v3]Ext4: " Mingming Cao
2008-07-30 11:29 ` Frédéric Bohé
2008-07-31 18:07 ` Mingming Cao
2008-08-01 5:49 ` Theodore Tso
2008-08-01 10:51 ` Frédéric Bohé
2008-08-01 18:08 ` Mingming Cao
2008-08-01 18:03 ` Mingming Cao
2008-08-01 19:10 ` Theodore Tso
2008-08-02 0:03 ` Theodore Tso
2008-08-04 11:23 ` Frédéric Bohé
2008-08-04 13:20 ` Theodore Tso
2008-07-30 11:36 ` Andreas Dilger
2008-07-30 12:16 ` Aneesh Kumar K.V
2008-08-01 19:29 ` Theodore Tso
2008-08-02 0:22 ` Theodore Tso
2008-08-12 16:23 ` [PATCH 0/6 ]Ext4 journal credits reservation fixes Mingming Cao
2008-08-12 16:25 ` [PATCH 1/6 ]Ext4 credits caclulation cleanup and fix that for nonextent writepage Mingming Cao
2008-08-13 8:31 ` Aneesh Kumar K.V
2008-08-14 0:30 ` Mingming Cao
2008-08-13 10:19 ` Aneesh Kumar K.V
2008-08-14 1:02 ` Mingming Cao
2008-08-16 0:37 ` [PATCH 1/6 V2 " Mingming Cao
2008-08-12 16:27 ` [PATCH 2/6 ]Ext4: journal credits reservation fixes for extent file writepage Mingming Cao
2008-08-13 8:37 ` Aneesh Kumar K.V
2008-08-14 0:26 ` Mingming Cao
2008-08-14 8:28 ` Aneesh Kumar K.V
2008-08-16 0:38 ` [PATCH 2/6 V2]Ext4: " Mingming Cao
2008-08-16 4:25 ` Aneesh Kumar K.V
2008-08-12 16:29 ` [PATCH 3/6 ]Ext4: journal credits reservation fixes for DIO, fallocate Mingming Cao
2008-08-13 8:53 ` Aneesh Kumar K.V
2008-08-13 10:14 ` Aneesh Kumar K.V
2008-08-14 0:50 ` Mingming Cao
2008-08-16 0:39 ` [PATCH 3/6 V2 " Mingming Cao
2008-08-12 16:32 ` [PATCH 4/6 ]ext4: Rework the ext4_da_writepages Mingming Cao
2008-08-16 0:43 ` [PATCH 4/6 V2]ext4: " Mingming Cao
2008-08-12 16:35 ` [PATCH 5/6 ]Ext4 journal credits reservation fixes Mingming Cao
2008-08-13 9:46 ` Aneesh Kumar K.V
2008-08-14 1:01 ` Mingming Cao
2008-08-14 8:40 ` Aneesh Kumar K.V
2008-08-16 0:40 ` [PATCH 5/6 V2]Ext4 journal credits fixes for delalloc writepages Mingming Cao
2008-08-16 4:23 ` Aneesh Kumar K.V
2008-08-12 16:37 ` [PATCH 6/6 ]Ext4 journal credits reservation fixes for defrag Mingming Cao
2008-08-16 0:45 ` [PATCH 6/6 V2]Ext4 " Mingming Cao
2008-08-16 15:55 ` Theodore Tso
2008-08-15 17:33 ` [PATCH 0/6 ]Ext4 journal credits reservation fixes Aneesh Kumar K.V
2008-08-15 19:02 ` Mingming Cao
2008-08-16 0:34 ` Mingming Cao
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=20080729062427.GC4545@skywalker \
--to=aneesh.kumar@linux.vnet.ibm.com \
--cc=cmm@us.ibm.com \
--cc=linux-ext4@vger.kernel.org \
--cc=shehjart@cse.unsw.edu.au \
--cc=tytso@mit.edu \
/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