* [PATCH] ext3_getblk should handle HOLE correctly
@ 2006-09-06 17:39 Badari Pulavarty
2006-09-06 17:42 ` Dave Kleikamp
2006-09-07 18:45 ` Andrew Morton
0 siblings, 2 replies; 5+ messages in thread
From: Badari Pulavarty @ 2006-09-06 17:39 UTC (permalink / raw)
To: akpm; +Cc: lkml, ext4, Will Simoneau, cmm
Hi Andrew,
Its been reported that ext3_getblk() is not doing the right thing
and triggering following WARN():
BUG: warning at fs/ext3/inode.c:1016/ext3_getblk()
<c01c5140> ext3_getblk+0x98/0x2a6 <c03b2806> md_wakeup_thread
+0x26/0x2a
<c01c536d> ext3_bread+0x1f/0x88 <c01cedf9> ext3_quota_read+0x136/0x1ae
<c018b683> v1_read_dqblk+0x61/0xac <c0188f32> dquot_acquire+0xf6/0x107
<c01ceaba> ext3_acquire_dquot+0x46/0x68 <c01897d4> dqget+0x155/0x1e7
<c018a97b> dquot_transfer+0x3e0/0x3e9 <c016fe52> dput+0x23/0x13e
<c01c7986> ext3_setattr+0xc3/0x240 <c0120f66> current_fs_time
+0x52/0x6a
<c017320e> notify_change+0x2bd/0x30d <c0159246> chown_common+0x9c/0xc5
<c02a222c> strncpy_from_user+0x3b/0x68 <c0167fe6> do_path_lookup
+0xdf/0x266
<c016841b> __user_walk_fd+0x44/0x5a <c01592b9> sys_chown+0x4a/0x55
<c015a43c> vfs_write+0xe7/0x13c <c01695d4> sys_mkdir+0x1f/0x23
<c0102a97> syscall_call+0x7/0xb
Looking at the code, it looks like its not handle HOLE correctly.
It ends up returning -EIO. Here is the patch to fix it.
If we really want to be paranoid, we can allow return values
0 (HOLE), 1 (we asked for one block) and return -EIO for
more than 1 block. But I really don't see a reason for
doing it - all we need is the block# here. (doesn't matter
how many blocks are mapped).
Thanks,
Badari
ext3_get_blocks_handle() returns number of blocks it mapped.
It returns 0 in case of HOLE. ext3_getblk() should handle
HOLE properly (currently its dumping warning stack and
returning -EIO).
Signed-off-by: Badari Pulavarty <pbadari@us.ibm.com>
Acked-by: Mingming Cao <cmm@us.ibm.com>
---
fs/ext3/inode.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
Index: linux-2.6.18-rc5/fs/ext3/inode.c
===================================================================
--- linux-2.6.18-rc5.orig/fs/ext3/inode.c 2006-08-27 20:41:48.000000000 -0700
+++ linux-2.6.18-rc5/fs/ext3/inode.c 2006-09-05 15:32:57.000000000 -0700
@@ -1009,11 +1009,12 @@ struct buffer_head *ext3_getblk(handle_t
buffer_trace_init(&dummy.b_history);
err = ext3_get_blocks_handle(handle, inode, block, 1,
&dummy, create, 1);
- if (err == 1) {
+ /*
+ * ext3_get_blocks_handle() returns number of blocks
+ * mapped. 0 in case of a HOLE.
+ */
+ if (err > 0) {
err = 0;
- } else if (err >= 0) {
- WARN_ON(1);
- err = -EIO;
}
*errp = err;
if (!err && buffer_mapped(&dummy)) {
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] ext3_getblk should handle HOLE correctly
2006-09-06 17:39 [PATCH] ext3_getblk should handle HOLE correctly Badari Pulavarty
@ 2006-09-06 17:42 ` Dave Kleikamp
2006-09-07 18:45 ` Andrew Morton
1 sibling, 0 replies; 5+ messages in thread
From: Dave Kleikamp @ 2006-09-06 17:42 UTC (permalink / raw)
To: Badari Pulavarty; +Cc: akpm, lkml, ext4, Will Simoneau, cmm
On Wed, 2006-09-06 at 10:39 -0700, Badari Pulavarty wrote:
> Index: linux-2.6.18-rc5/fs/ext3/inode.c
> ===================================================================
> --- linux-2.6.18-rc5.orig/fs/ext3/inode.c 2006-08-27 20:41:48.000000000 -0700
> +++ linux-2.6.18-rc5/fs/ext3/inode.c 2006-09-05 15:32:57.000000000 -0700
> @@ -1009,11 +1009,12 @@ struct buffer_head *ext3_getblk(handle_t
> buffer_trace_init(&dummy.b_history);
> err = ext3_get_blocks_handle(handle, inode, block, 1,
> &dummy, create, 1);
> - if (err == 1) {
> + /*
> + * ext3_get_blocks_handle() returns number of blocks
> + * mapped. 0 in case of a HOLE.
> + */
> + if (err > 0) {
> err = 0;
> - } else if (err >= 0) {
> - WARN_ON(1);
> - err = -EIO;
> }
I'd get rid of the {} too.
> *errp = err;
> if (!err && buffer_mapped(&dummy)) {
--
David Kleikamp
IBM Linux Technology Center
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] ext3_getblk should handle HOLE correctly
2006-09-06 17:39 [PATCH] ext3_getblk should handle HOLE correctly Badari Pulavarty
2006-09-06 17:42 ` Dave Kleikamp
@ 2006-09-07 18:45 ` Andrew Morton
2006-09-07 19:22 ` Badari Pulavarty
1 sibling, 1 reply; 5+ messages in thread
From: Andrew Morton @ 2006-09-07 18:45 UTC (permalink / raw)
To: Badari Pulavarty; +Cc: lkml, ext4, Will Simoneau, cmm
On Wed, 06 Sep 2006 10:39:06 -0700
Badari Pulavarty <pbadari@us.ibm.com> wrote:
> Hi Andrew,
>
> Its been reported that ext3_getblk() is not doing the right thing
> and triggering following WARN():
>
> BUG: warning at fs/ext3/inode.c:1016/ext3_getblk()
> <c01c5140> ext3_getblk+0x98/0x2a6 <c03b2806> md_wakeup_thread
> +0x26/0x2a
> <c01c536d> ext3_bread+0x1f/0x88 <c01cedf9> ext3_quota_read+0x136/0x1ae
> <c018b683> v1_read_dqblk+0x61/0xac <c0188f32> dquot_acquire+0xf6/0x107
> <c01ceaba> ext3_acquire_dquot+0x46/0x68 <c01897d4> dqget+0x155/0x1e7
> <c018a97b> dquot_transfer+0x3e0/0x3e9 <c016fe52> dput+0x23/0x13e
> <c01c7986> ext3_setattr+0xc3/0x240 <c0120f66> current_fs_time
> +0x52/0x6a
> <c017320e> notify_change+0x2bd/0x30d <c0159246> chown_common+0x9c/0xc5
> <c02a222c> strncpy_from_user+0x3b/0x68 <c0167fe6> do_path_lookup
> +0xdf/0x266
> <c016841b> __user_walk_fd+0x44/0x5a <c01592b9> sys_chown+0x4a/0x55
> <c015a43c> vfs_write+0xe7/0x13c <c01695d4> sys_mkdir+0x1f/0x23
> <c0102a97> syscall_call+0x7/0xb
>
> Looking at the code, it looks like its not handle HOLE correctly.
> It ends up returning -EIO.
Strange. The fs should be spewing these warnings all over the place. For
some reason this code is hard to trigger. Why??
> - if (err == 1) {
> + /*
> + * ext3_get_blocks_handle() returns number of blocks
> + * mapped. 0 in case of a HOLE.
> + */
> + if (err > 0) {
> err = 0;
> - } else if (err >= 0) {
> - WARN_ON(1);
> - err = -EIO;
> }
That removes the warning if ext3_get_blocks_handle() returned a positive
number greater than one. And it looks like we still need debugging support
in this area.
I reworked it like this:
--- a/fs/ext3/inode.c~ext3_getblk-should-handle-hole-correctly
+++ a/fs/ext3/inode.c
@@ -1010,11 +1010,14 @@ struct buffer_head *ext3_getblk(handle_t
buffer_trace_init(&dummy.b_history);
err = ext3_get_blocks_handle(handle, inode, block, 1,
&dummy, create, 1);
- if (err == 1) {
+ /*
+ * ext3_get_blocks_handle() returns number of blocks
+ * mapped. 0 in case of a HOLE.
+ */
+ if (err > 0) {
+ if (err > 1)
+ WARN_ON(1);
err = 0;
- } else if (err >= 0) {
- WARN_ON(1);
- err = -EIO;
}
*errp = err;
if (!err && buffer_mapped(&dummy)) {
_
ie:
/*
* ext3_get_blocks_handle() returns number of blocks
* mapped. 0 in case of a HOLE.
*/
if (err > 0) {
if (err > 1)
WARN_ON(1);
err = 0;
}
*errp = err;
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] ext3_getblk should handle HOLE correctly
2006-09-07 18:45 ` Andrew Morton
@ 2006-09-07 19:22 ` Badari Pulavarty
2006-09-07 19:55 ` Andrew Morton
0 siblings, 1 reply; 5+ messages in thread
From: Badari Pulavarty @ 2006-09-07 19:22 UTC (permalink / raw)
To: Andrew Morton; +Cc: lkml, ext4, Will Simoneau, cmm
On Thu, 2006-09-07 at 11:45 -0700, Andrew Morton wrote:
> On Wed, 06 Sep 2006 10:39:06 -0700
> Badari Pulavarty <pbadari@us.ibm.com> wrote:
>
> > Hi Andrew,
> >
> > Its been reported that ext3_getblk() is not doing the right thing
> > and triggering following WARN():
> >
> > BUG: warning at fs/ext3/inode.c:1016/ext3_getblk()
> > <c01c5140> ext3_getblk+0x98/0x2a6 <c03b2806> md_wakeup_thread
> > +0x26/0x2a
> > <c01c536d> ext3_bread+0x1f/0x88 <c01cedf9> ext3_quota_read+0x136/0x1ae
> > <c018b683> v1_read_dqblk+0x61/0xac <c0188f32> dquot_acquire+0xf6/0x107
> > <c01ceaba> ext3_acquire_dquot+0x46/0x68 <c01897d4> dqget+0x155/0x1e7
> > <c018a97b> dquot_transfer+0x3e0/0x3e9 <c016fe52> dput+0x23/0x13e
> > <c01c7986> ext3_setattr+0xc3/0x240 <c0120f66> current_fs_time
> > +0x52/0x6a
> > <c017320e> notify_change+0x2bd/0x30d <c0159246> chown_common+0x9c/0xc5
> > <c02a222c> strncpy_from_user+0x3b/0x68 <c0167fe6> do_path_lookup
> > +0xdf/0x266
> > <c016841b> __user_walk_fd+0x44/0x5a <c01592b9> sys_chown+0x4a/0x55
> > <c015a43c> vfs_write+0xe7/0x13c <c01695d4> sys_mkdir+0x1f/0x23
> > <c0102a97> syscall_call+0x7/0xb
> >
> > Looking at the code, it looks like its not handle HOLE correctly.
> > It ends up returning -EIO.
>
> Strange. The fs should be spewing these warnings all over the place. For
> some reason this code is hard to trigger. Why??
I guess - ext3_getblk() mostly used by ext3_bread() and most callers
to it would be reading already allocated block.
>
> > - if (err == 1) {
> > + /*
> > + * ext3_get_blocks_handle() returns number of blocks
> > + * mapped. 0 in case of a HOLE.
> > + */
> > + if (err > 0) {
> > err = 0;
> > - } else if (err >= 0) {
> > - WARN_ON(1);
> > - err = -EIO;
> > }
>
> That removes the warning if ext3_get_blocks_handle() returned a positive
> number greater than one. And it looks like we still need debugging support
> in this area.
I am not sure why we need it ? All we care about is one block. If
ext3_get_blocks_handle() returns more than one (which it shouldn't) -
it still be okay. Whats wrong with that ? Just curious ..
May be we should add a WARN() in ext3_get_blocks_handle() when it
returns more than asked for.
Thanks,
Badari
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] ext3_getblk should handle HOLE correctly
2006-09-07 19:22 ` Badari Pulavarty
@ 2006-09-07 19:55 ` Andrew Morton
0 siblings, 0 replies; 5+ messages in thread
From: Andrew Morton @ 2006-09-07 19:55 UTC (permalink / raw)
To: Badari Pulavarty; +Cc: lkml, ext4, Will Simoneau, cmm
On Thu, 07 Sep 2006 12:22:27 -0700
Badari Pulavarty <pbadari@us.ibm.com> wrote:
> On Thu, 2006-09-07 at 11:45 -0700, Andrew Morton wrote:
> > On Wed, 06 Sep 2006 10:39:06 -0700
> > Badari Pulavarty <pbadari@us.ibm.com> wrote:
> >
> > > Hi Andrew,
> > >
> > > Its been reported that ext3_getblk() is not doing the right thing
> > > and triggering following WARN():
> > >
> > > BUG: warning at fs/ext3/inode.c:1016/ext3_getblk()
> > > <c01c5140> ext3_getblk+0x98/0x2a6 <c03b2806> md_wakeup_thread
> > > +0x26/0x2a
> > > <c01c536d> ext3_bread+0x1f/0x88 <c01cedf9> ext3_quota_read+0x136/0x1ae
> > > <c018b683> v1_read_dqblk+0x61/0xac <c0188f32> dquot_acquire+0xf6/0x107
> > > <c01ceaba> ext3_acquire_dquot+0x46/0x68 <c01897d4> dqget+0x155/0x1e7
> > > <c018a97b> dquot_transfer+0x3e0/0x3e9 <c016fe52> dput+0x23/0x13e
> > > <c01c7986> ext3_setattr+0xc3/0x240 <c0120f66> current_fs_time
> > > +0x52/0x6a
> > > <c017320e> notify_change+0x2bd/0x30d <c0159246> chown_common+0x9c/0xc5
> > > <c02a222c> strncpy_from_user+0x3b/0x68 <c0167fe6> do_path_lookup
> > > +0xdf/0x266
> > > <c016841b> __user_walk_fd+0x44/0x5a <c01592b9> sys_chown+0x4a/0x55
> > > <c015a43c> vfs_write+0xe7/0x13c <c01695d4> sys_mkdir+0x1f/0x23
> > > <c0102a97> syscall_call+0x7/0xb
> > >
> > > Looking at the code, it looks like its not handle HOLE correctly.
> > > It ends up returning -EIO.
> >
> > Strange. The fs should be spewing these warnings all over the place. For
> > some reason this code is hard to trigger. Why??
>
> I guess - ext3_getblk() mostly used by ext3_bread() and most callers
> to it would be reading already allocated block.
OK.
> >
> > > - if (err == 1) {
> > > + /*
> > > + * ext3_get_blocks_handle() returns number of blocks
> > > + * mapped. 0 in case of a HOLE.
> > > + */
> > > + if (err > 0) {
> > > err = 0;
> > > - } else if (err >= 0) {
> > > - WARN_ON(1);
> > > - err = -EIO;
> > > }
> >
> > That removes the warning if ext3_get_blocks_handle() returned a positive
> > number greater than one. And it looks like we still need debugging support
> > in this area.
>
> I am not sure why we need it ? All we care about is one block. If
> ext3_get_blocks_handle() returns more than one (which it shouldn't) -
The operative part is "which it shouldn't". This code is fairly fresh, and
ext3 is paranoid. Once it's all settled down then after a year or so we
might decide that the debugging code is no longer needed.
Then again, we have plenty of five-year-old assertions in there..
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2006-09-07 19:56 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2006-09-06 17:39 [PATCH] ext3_getblk should handle HOLE correctly Badari Pulavarty
2006-09-06 17:42 ` Dave Kleikamp
2006-09-07 18:45 ` Andrew Morton
2006-09-07 19:22 ` Badari Pulavarty
2006-09-07 19:55 ` Andrew Morton
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox