On Wed, Feb 8, 2012 at 8:14 PM, Liu Bo [off-list ref] wrote=
:

On 02/09/2012 05:01 AM, Mitch Harder wrote:

quoted

On Wed, Jan 18, 2012 at 10:13 AM, Mitch Harder
[off-list ref] wrote:

quoted

I have a Btrfs partition that is reliably reproducing premature ENO=

SPC

quoted

quoted

when restoring the disk from a tar file, but it is only happening w=

ith

quoted

quoted

zlib compression (lzo or no compression proceeds normally).

I've had the same issue at least back through the 3.1 kernel series=

,

quoted

quoted

and I've been having intermittent issues even further back.

I am currently using a 3.2.1 kernel merged with Chris' latest
integration branch.

I've performed about 12 trials trying to explore various combinatio=

ns

quoted

quoted

of compress, compress-force, compress[-force]=3D[zlib,lzo] and
autodefrag.

If I use no compression, or if I explicitly declare lzo compression=

, I

quoted

quoted

don't receive the premature ENOSPC when untarring my restoration
archive to the empty partition.

If I don't specify compression (zlib is the default) or specify zli=

b,

quoted

quoted

I get consistent premature ENOSPC errors regardless of other
combinations.

I apologize if this is already general knowledge, but I couldn't se=

e

quoted

quoted

where this has been posted to the list before.

As time allows, I will try to capture exactly where this ENOSPC is
being issued in btrfs by inserting WARN_ON's in my local version
where-ever ENOSPC is set.

Some follow-up...

I've injected some debugging code to isolate when the ENOSPC is bein=

g

quoted

generated when using zlib compression.

When using zlib, I'm getting intermittent ENOSPC in the
may_commit_transaction() function in extent-tree.c at this point:

=A0 =A0 =A0 if (delayed_rsv->size < bytes) {
=A0 =A0 =A0 =A0 =A0 =A0 =A0 spin_unlock(&delayed_rsv->lock);
=A0 =A0 =A0 =A0 =A0 =A0 =A0 return -ENOSPC;
=A0 =A0 =A0 }

The typical values for (delayed_rsv->size < bytes) have been:
delayed_rsv->size ( =3D 0x60000) < bytes ( =3D 0x78000)

This typically occurs when unzipping a section of my backup that
contains lots of small files that are probably being mostly in-lined=

=2E

s_root *root,

=A0 =A0 =A0 =A0if (force)
=A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0goto commit;

- =A0 =A0 =A0 /* See if there is enough pinned space to make this res=

ervation */

- =A0 =A0 =A0 spin_lock(&space_info->lock);
- =A0 =A0 =A0 if (space_info->bytes_pinned >=3D bytes) {
+ =A0 =A0 =A0 if (space_info !=3D delayed_rsv->space_info) {
+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 /*
+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0* For DATA:
+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0* See if there is enough pinned spac=

e to make this reservation

+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0*/
+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 spin_lock(&space_info->lock);
+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 if (space_info->bytes_pinned < bytes) {
+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 spin_unlock(&space_info=

->lock);

+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 return -ENOSPC;
+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 }
=A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0spin_unlock(&space_info->lock);
- =A0 =A0 =A0 =A0 =A0 =A0 =A0 goto commit;
- =A0 =A0 =A0 }
- =A0 =A0 =A0 spin_unlock(&space_info->lock);
-
- =A0 =A0 =A0 /*
- =A0 =A0 =A0 =A0* See if there is some space in the delayed insertio=

n reservation for

- =A0 =A0 =A0 =A0* this reservation.
- =A0 =A0 =A0 =A0*/
- =A0 =A0 =A0 if (space_info !=3D delayed_rsv->space_info)
- =A0 =A0 =A0 =A0 =A0 =A0 =A0 return -ENOSPC;
-
- =A0 =A0 =A0 spin_lock(&delayed_rsv->lock);
- =A0 =A0 =A0 if (delayed_rsv->size < bytes) {
+ =A0 =A0 =A0 } else {
+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 /*
+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0* For METADATA:
+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0* See if there is enough space(pinne=

d and delayed insertion)

+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0* to make this reservation
+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0*/
+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 spin_lock(&space_info->lock);
+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 spin_lock(&delayed_rsv->lock);
+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 if (space_info->bytes_pinned + delayed_=

rsv->size < bytes) {

+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 spin_unlock(&delayed_rs=

v->lock);

+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 spin_unlock(&space_info=

->lock);

+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 return -ENOSPC;
+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 }
=A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0spin_unlock(&delayed_rsv->lock);
- =A0 =A0 =A0 =A0 =A0 =A0 =A0 return -ENOSPC;
- =A0 =A0 =A0 }
- =A0 =A0 =A0 spin_unlock(&delayed_rsv->lock);
+ =A0 =A0 =A0 =A0 =A0 =A0 =A0 spin_unlock(&space_info->lock);

+ =A0 =A0 =A0 }
=A0commit:
=A0 =A0 =A0 =A0trans =3D btrfs_join_transaction(root);
=A0 =A0 =A0 =A0if (IS_ERR(trans))
--
1.6.5.2

Thanks for looking at this, but I'm still getting ENOSPC errors at the
same point after applying this patch.
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" =
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html