All of lore.kernel.org
 help / color / mirror / Atom feed
From: Evgeniy Dushistov <dushistov@mail.ru>
To: "Vladimir 'φ-coder/phcoder' Serbinenko" <phcoder@gmail.com>
Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org
Subject: Re: [PATCH] update i_blocks on UFS
Date: Mon, 14 May 2012 01:59:26 +0400	[thread overview]
Message-ID: <20120513215925.GA1998@maclin> (raw)
In-Reply-To: <4FAFBCAE.1050401@gmail.com>

On Sun, May 13, 2012 at 03:52:46PM +0200, Vladimir 'φ-coder/phcoder' Serbinenko wrote:
> UFS code dosn't update i_blocks which leads to the problems ranging from wrong
> stat/du to long symlinks (as opposed to inline ones) being damaged. This is
> a fix.
> 

Yes, when quota support was removed, code that update statistic
about usage of blocks also was removed. But why so big change to fix this?
I found this problem some time ago and fixed it with code bellow,
but may be I missed something?

From 3bba2473d2d09c0ebde5e287d694346ba0e4b4c0 Mon Sep 17 00:00:00 2001
From: Evgeniy Dushistov <dushistov@mail.ru>
Date: Fri, 18 Mar 2011 06:42:57 +0300
Subject: [PATCH] Looks like with quote support code was removed code
 that update statistic about usage blocks per inode.
 So fix this.

---
 fs/ufs/balloc.c |    9 ++++++---
 1 files changed, 6 insertions(+), 3 deletions(-)

diff --git a/fs/ufs/balloc.c b/fs/ufs/balloc.c
index 100d2ee..e7bf2fd 100644
--- a/fs/ufs/balloc.c
+++ b/fs/ufs/balloc.c
@@ -83,7 +83,7 @@ void ufs_free_fragments(struct inode *inode, u64 fragment, unsigned count)
 			ufs_error (sb, "ufs_free_fragments",
 				   "bit already cleared for fragment %u", i);
 	}
-	
+	inode_sub_bytes(inode, count << uspi->s_fshift);
 	fs32_add(sb, &ucg->cg_cs.cs_nffree, count);
 	uspi->cs_total.cs_nffree += count;
 	fs32_add(sb, &UFS_SB(sb)->fs_cs(cgno).cs_nffree, count);
@@ -189,6 +189,7 @@ do_more:
 		ubh_setblock(UCPI_UBH(ucpi), ucpi->c_freeoff, blkno);
 		if ((UFS_SB(sb)->s_flags & UFS_CG_MASK) == UFS_CG_44BSD)
 			ufs_clusteracct (sb, ucpi, blkno, 1);
+		inode_sub_bytes(inode, uspi->s_bsize);
 
 		fs32_add(sb, &ucg->cg_cs.cs_nbfree, 1);
 		uspi->cs_total.cs_nbfree++;
@@ -547,7 +548,7 @@ static u64 ufs_add_fragments(struct inode *inode, u64 fragment,
 		fs32_add(sb, &ucg->cg_frsum[fragsize - count], 1);
 	for (i = oldcount; i < newcount; i++)
 		ubh_clrbit (UCPI_UBH(ucpi), ucpi->c_freeoff, fragno + i);
-
+	inode_add_bytes(inode, count << uspi->s_fshift);
 	fs32_sub(sb, &ucg->cg_cs.cs_nffree, count);
 	fs32_sub(sb, &UFS_SB(sb)->fs_cs(cgno).cs_nffree, count);
 	uspi->cs_total.cs_nffree -= count;
@@ -649,6 +650,7 @@ cg_found:
 		for (i = count; i < uspi->s_fpb; i++)
 			ubh_setbit (UCPI_UBH(ucpi), ucpi->c_freeoff, goal + i);
 		i = uspi->s_fpb - count;
+		inode_sub_bytes(inode, i << uspi->s_fshift);
 
 		fs32_add(sb, &ucg->cg_cs.cs_nffree, i);
 		uspi->cs_total.cs_nffree += i;
@@ -660,6 +662,7 @@ cg_found:
 	result = ufs_bitmap_search (sb, ucpi, goal, allocsize);
 	if (result == INVBLOCK)
 		return 0;
+	inode_add_bytes(inode, count << uspi->s_fshift);
 	for (i = 0; i < count; i++)
 		ubh_clrbit (UCPI_UBH(ucpi), ucpi->c_freeoff, result + i);
 	
@@ -725,7 +728,7 @@ gotit:
 	ubh_clrblock (UCPI_UBH(ucpi), ucpi->c_freeoff, blkno);
 	if ((UFS_SB(sb)->s_flags & UFS_CG_MASK) == UFS_CG_44BSD)
 		ufs_clusteracct (sb, ucpi, blkno, -1);
-
+	inode_add_bytes(inode, uspi->s_bsize);
 	fs32_sub(sb, &ucg->cg_cs.cs_nbfree, 1);
 	uspi->cs_total.cs_nbfree--;
 	fs32_sub(sb, &UFS_SB(sb)->fs_cs(ucpi->c_cgx).cs_nbfree, 1);
-- 
1.7.3.4

-- 
/Evgeniy

WARNING: multiple messages have this Message-ID (diff)
From: Evgeniy Dushistov <dushistov@mail.ru>
To: "Vladimir 'φ-coder/phcoder' Serbinenko" <phcoder@gmail.com>
Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org
Subject: Re: [PATCH] update i_blocks on UFS
Date: Mon, 14 May 2012 01:59:26 +0400	[thread overview]
Message-ID: <20120513215925.GA1998@maclin> (raw)
In-Reply-To: <4FAFBCAE.1050401@gmail.com>

On Sun, May 13, 2012 at 03:52:46PM +0200, Vladimir 'φ-coder/phcoder' Serbinenko wrote:
> UFS code dosn't update i_blocks which leads to the problems ranging from wrong
> stat/du to long symlinks (as opposed to inline ones) being damaged. This is
> a fix.
> 

Yes, when quota support was removed, code that update statistic
about usage of blocks also was removed. But why so big change to fix this?
I found this problem some time ago and fixed it with code bellow,
but may be I missed something?

>From 3bba2473d2d09c0ebde5e287d694346ba0e4b4c0 Mon Sep 17 00:00:00 2001
From: Evgeniy Dushistov <dushistov@mail.ru>
Date: Fri, 18 Mar 2011 06:42:57 +0300
Subject: [PATCH] Looks like with quote support code was removed code
 that update statistic about usage blocks per inode.
 So fix this.

---
 fs/ufs/balloc.c |    9 ++++++---
 1 files changed, 6 insertions(+), 3 deletions(-)

diff --git a/fs/ufs/balloc.c b/fs/ufs/balloc.c
index 100d2ee..e7bf2fd 100644
--- a/fs/ufs/balloc.c
+++ b/fs/ufs/balloc.c
@@ -83,7 +83,7 @@ void ufs_free_fragments(struct inode *inode, u64 fragment, unsigned count)
 			ufs_error (sb, "ufs_free_fragments",
 				   "bit already cleared for fragment %u", i);
 	}
-	
+	inode_sub_bytes(inode, count << uspi->s_fshift);
 	fs32_add(sb, &ucg->cg_cs.cs_nffree, count);
 	uspi->cs_total.cs_nffree += count;
 	fs32_add(sb, &UFS_SB(sb)->fs_cs(cgno).cs_nffree, count);
@@ -189,6 +189,7 @@ do_more:
 		ubh_setblock(UCPI_UBH(ucpi), ucpi->c_freeoff, blkno);
 		if ((UFS_SB(sb)->s_flags & UFS_CG_MASK) == UFS_CG_44BSD)
 			ufs_clusteracct (sb, ucpi, blkno, 1);
+		inode_sub_bytes(inode, uspi->s_bsize);
 
 		fs32_add(sb, &ucg->cg_cs.cs_nbfree, 1);
 		uspi->cs_total.cs_nbfree++;
@@ -547,7 +548,7 @@ static u64 ufs_add_fragments(struct inode *inode, u64 fragment,
 		fs32_add(sb, &ucg->cg_frsum[fragsize - count], 1);
 	for (i = oldcount; i < newcount; i++)
 		ubh_clrbit (UCPI_UBH(ucpi), ucpi->c_freeoff, fragno + i);
-
+	inode_add_bytes(inode, count << uspi->s_fshift);
 	fs32_sub(sb, &ucg->cg_cs.cs_nffree, count);
 	fs32_sub(sb, &UFS_SB(sb)->fs_cs(cgno).cs_nffree, count);
 	uspi->cs_total.cs_nffree -= count;
@@ -649,6 +650,7 @@ cg_found:
 		for (i = count; i < uspi->s_fpb; i++)
 			ubh_setbit (UCPI_UBH(ucpi), ucpi->c_freeoff, goal + i);
 		i = uspi->s_fpb - count;
+		inode_sub_bytes(inode, i << uspi->s_fshift);
 
 		fs32_add(sb, &ucg->cg_cs.cs_nffree, i);
 		uspi->cs_total.cs_nffree += i;
@@ -660,6 +662,7 @@ cg_found:
 	result = ufs_bitmap_search (sb, ucpi, goal, allocsize);
 	if (result == INVBLOCK)
 		return 0;
+	inode_add_bytes(inode, count << uspi->s_fshift);
 	for (i = 0; i < count; i++)
 		ubh_clrbit (UCPI_UBH(ucpi), ucpi->c_freeoff, result + i);
 	
@@ -725,7 +728,7 @@ gotit:
 	ubh_clrblock (UCPI_UBH(ucpi), ucpi->c_freeoff, blkno);
 	if ((UFS_SB(sb)->s_flags & UFS_CG_MASK) == UFS_CG_44BSD)
 		ufs_clusteracct (sb, ucpi, blkno, -1);
-
+	inode_add_bytes(inode, uspi->s_bsize);
 	fs32_sub(sb, &ucg->cg_cs.cs_nbfree, 1);
 	uspi->cs_total.cs_nbfree--;
 	fs32_sub(sb, &UFS_SB(sb)->fs_cs(ucpi->c_cgx).cs_nbfree, 1);
-- 
1.7.3.4

-- 
/Evgeniy


  reply	other threads:[~2012-05-13 21:59 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-05-13 13:52 [PATCH] update i_blocks on UFS Vladimir 'φ-coder/phcoder' Serbinenko
2012-05-13 21:59 ` Evgeniy Dushistov [this message]
2012-05-13 21:59   ` Evgeniy Dushistov
2012-05-13 22:16   ` Vladimir 'φ-coder/phcoder' Serbinenko
2012-05-14 11:57     ` Evgeniy Dushistov
2012-05-14 11:57       ` Evgeniy Dushistov

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=20120513215925.GA1998@maclin \
    --to=dushistov@mail.ru \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=phcoder@gmail.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.