At LSF we decided that if we truncate up from isize we shouldn't trim fallocated
blocks that were fallocated with KEEP_SIZE and are past the new i_size. This
patch fixes ext4 to do this. Thanks,
Signed-off-by: Josef Bacik <[email protected]>
---
fs/ext4/inode.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 0554b0b..92174dd 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -4640,6 +4640,7 @@ static void ext4_wait_for_tail_page_commit(struct inode *inode)
int ext4_setattr(struct dentry *dentry, struct iattr *attr)
{
struct inode *inode = d_inode(dentry);
+ loff_t oldsize = inode->i_size;
int error, rc = 0;
int orphan = 0;
const unsigned int ia_valid = attr->ia_valid;
@@ -4727,8 +4728,6 @@ int ext4_setattr(struct dentry *dentry, struct iattr *attr)
goto err_out;
}
} else {
- loff_t oldsize = inode->i_size;
On Tue, 19 May 2015, Josef Bacik wrote:
> Date: Tue, 19 May 2015 11:28:57 -0400
> From: Josef Bacik <[email protected]>
> To: [email protected], [email protected]
> Subject: [PATCH] ext4: only call ext4_truncate when size <= isize
>
> At LSF we decided that if we truncate up from isize we shouldn't trim fallocated
> blocks that were fallocated with KEEP_SIZE and are past the new i_size. This
> patch fixes ext4 to do this. Thanks,
Looks good, thanks!
Reviewed-by: Lukas Czerner <[email protected]>
>
> Signed-off-by: Josef Bacik <[email protected]>
> ---
> fs/ext4/inode.c | 5 ++---
> 1 file changed, 2 insertions(+), 3 deletions(-)
>
> diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
> index 0554b0b..92174dd 100644
> --- a/fs/ext4/inode.c
> +++ b/fs/ext4/inode.c
> @@ -4640,6 +4640,7 @@ static void ext4_wait_for_tail_page_commit(struct inode *inode)
> int ext4_setattr(struct dentry *dentry, struct iattr *attr)
> {
> struct inode *inode = d_inode(dentry);
> + loff_t oldsize = inode->i_size;
> int error, rc = 0;
> int orphan = 0;
> const unsigned int ia_valid = attr->ia_valid;
> @@ -4727,8 +4728,6 @@ int ext4_setattr(struct dentry *dentry, struct iattr *attr)
> goto err_out;
> }
> } else {
> - loff_t oldsize = inode->i_size;
> -
> i_size_write(inode, attr->ia_size);
> pagecache_isize_extended(inode, oldsize, inode->i_size);
> }
> @@ -4756,7 +4755,7 @@ int ext4_setattr(struct dentry *dentry, struct iattr *attr)
> * We want to call ext4_truncate() even if attr->ia_size ==
> * inode->i_size for cases like truncation of fallocated space
> */
> - if (attr->ia_valid & ATTR_SIZE)
> + if (attr->ia_valid & ATTR_SIZE && attr->ia_size <= oldsize)
> ext4_truncate(inode);
>
> if (!rc) {
>
On Tue, Jun 02, 2015 at 10:08:18AM +0200, Lukáš Czerner wrote:
> On Tue, 19 May 2015, Josef Bacik wrote:
>
> > Date: Tue, 19 May 2015 11:28:57 -0400
> > From: Josef Bacik <[email protected]>
> > To: [email protected], [email protected]
> > Subject: [PATCH] ext4: only call ext4_truncate when size <= isize
> >
> > At LSF we decided that if we truncate up from isize we shouldn't trim fallocated
> > blocks that were fallocated with KEEP_SIZE and are past the new i_size. This
> > patch fixes ext4 to do this. Thanks,
>
> Looks good, thanks!
>
> Reviewed-by: Lukas Czerner <[email protected]>
Applied, thanks.
- Ted
On Sat, Jun 06, 2015 at 04:20:11PM -0400, Theodore Ts'o wrote:
> On Tue, Jun 02, 2015 at 10:08:18AM +0200, Lukáš Czerner wrote:
> > On Tue, 19 May 2015, Josef Bacik wrote:
> >
> > > Date: Tue, 19 May 2015 11:28:57 -0400
> > > From: Josef Bacik <[email protected]>
> > > To: [email protected], [email protected]
> > > Subject: [PATCH] ext4: only call ext4_truncate when size <= isize
> > >
> > > At LSF we decided that if we truncate up from isize we shouldn't trim fallocated
> > > blocks that were fallocated with KEEP_SIZE and are past the new i_size. This
> > > patch fixes ext4 to do this. Thanks,
> >
> > Looks good, thanks!
> >
> > Reviewed-by: Lukas Czerner <[email protected]>
>
> Applied, thanks.
Unfortunately this patch is causing generic/239 to fail when run with
1k blocksizes.
I'm dropping this for now; could you take a look?
Thanks!
- Ted
generic/239 22s ... [11:46:25] [11:46:54] 29s
_check_generic_filesystem: filesystem on /dev/vdd is inconsistent (see /results/results-1k/generic/239.full)
Ran: generic/239
Failures: generic/239
Failed 1 of 1 tests
e2fsck 1.43-WIP (18-May-2015)
Pass 1: Checking inodes, blocks, and sizes
Inode 12, i_size is 0, should be 4096. Fix? yes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information
/dev/vdd: ***** FILE SYSTEM WAS MODIFIED *****
/dev/vdd: 12/327680 files (0.0% non-contiguous), 120145/5242880 blocks
On Tue, Jun 09, 2015 at 12:07:39PM -0400, Theodore Ts'o wrote:
>
> Unfortunately this patch is causing generic/239 to fail when run with
> 1k blocksizes.
So it turns out the original patch was completely busted; the on-disk
i_size wasn't getting updated when we were truncating up. Part of the
problem was the original way ATTR_SIZE was handled in ext4_setattr was
a bit of a mess, and it made it hard to understand. So I've reworked
patch completely to clean up ext4_setattr while I was it.
- Ted
>From 0d209259066ca93548189668d1eb4c9497913932 Mon Sep 17 00:00:00 2001
From: Josef Bacik <[email protected]>
Date: Mon, 22 Jun 2015 00:31:26 -0400
Subject: [PATCH] ext4: only call ext4_truncate when size <= isize
At LSF we decided that if we truncate up from isize we shouldn't trim
fallocated blocks that were fallocated with KEEP_SIZE and are past the
new i_size. This patch fixes ext4 to do this.
[ Completely reworked patch so that i_disksize would actually get set
when truncating up. Also reworked the code for handling truncate so
that it's easier to handle. -- tytso ]
Signed-off-by: Josef Bacik <[email protected]>
Signed-off-by: Theodore Ts'o <[email protected]>
Reviewed-by: Lukas Czerner <[email protected]>
---
fs/ext4/inode.c | 38 ++++++++++++++++++--------------------
1 file changed, 18 insertions(+), 20 deletions(-)
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index ae93f0b..e057c6f 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -4681,8 +4681,10 @@ int ext4_setattr(struct dentry *dentry, struct iattr *attr)
ext4_journal_stop(handle);
}
- if (attr->ia_valid & ATTR_SIZE && attr->ia_size != inode->i_size) {
+ if (attr->ia_valid & ATTR_SIZE) {
handle_t *handle;
+ loff_t oldsize = inode->i_size;
+ int shrink = (attr->ia_size <= inode->i_size);
if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))) {
struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
@@ -4690,24 +4692,26 @@ int ext4_setattr(struct dentry *dentry, struct iattr *attr)
if (attr->ia_size > sbi->s_bitmap_maxbytes)
return -EFBIG;
}
+ if (!S_ISREG(inode->i_mode))
+ return -EINVAL;
if (IS_I_VERSION(inode) && attr->ia_size != inode->i_size)
inode_inc_iversion(inode);
- if (S_ISREG(inode->i_mode) &&
+ if (ext4_should_order_data(inode) &&
(attr->ia_size < inode->i_size)) {
- if (ext4_should_order_data(inode)) {
- error = ext4_begin_ordered_truncate(inode,
+ error = ext4_begin_ordered_truncate(inode,
attr->ia_size);
- if (error)
- goto err_out;
- }
+ if (error)
+ goto err_out;
+ }
+ if (attr->ia_size != inode->i_size) {
handle = ext4_journal_start(inode, EXT4_HT_INODE, 3);
if (IS_ERR(handle)) {
error = PTR_ERR(handle);
goto err_out;
}
- if (ext4_handle_valid(handle)) {
+ if (ext4_handle_valid(handle) && shrink) {
error = ext4_orphan_add(handle, inode);
orphan = 1;
}
@@ -4726,15 +4730,13 @@ int ext4_setattr(struct dentry *dentry, struct iattr *attr)
up_write(&EXT4_I(inode)->i_data_sem);
ext4_journal_stop(handle);
if (error) {
- ext4_orphan_del(NULL, inode);
+ if (orphan)
+ ext4_orphan_del(NULL, inode);
goto err_out;
}
- } else {
- loff_t oldsize = inode->i_size;
-
- i_size_write(inode, attr->ia_size);
- pagecache_isize_extended(inode, oldsize, inode->i_size);
}
+ if (!shrink)
+ pagecache_isize_extended(inode, oldsize, inode->i_size);
/*
* Blocks are going to be removed from the inode. Wait
@@ -4754,13 +4756,9 @@ int ext4_setattr(struct dentry *dentry, struct iattr *attr)
* in data=journal mode to make pages freeable.
*/
truncate_pagecache(inode, inode->i_size);
+ if (shrink)
+ ext4_truncate(inode);
}
- /*
- * We want to call ext4_truncate() even if attr->ia_size ==
- * inode->i_size for cases like truncation of fallocated space
- */
- if (attr->ia_valid & ATTR_SIZE)
- ext4_truncate(inode);
if (!rc) {
setattr_copy(inode, attr);
--
2.3.0