[PATCH 0/5] xfsprogs: 4.15 rollup pt. 2

[PATCH 0/5] xfsprogs: 4.15 rollup pt. 2

[Darrick J. Wong]

Hi all,

Here are some patches fixing up random odds and ends in 4.15.  We fix some
CoW staging extent misreporting in xfs_db, an xfs_db crash when address
selectors have a trailing slash, libxfs api violations in xfs_repair, and
a grep pattern error in the libxfs api violation checking tool.

Eric: You might already have the first few of these; they haven't changed.

--D

[PATCH 1/5] libhandle: zap fdhash in fshandle_destroy

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Set fdhash_head to zero once we've destroyed the handle list to avoid
dangling pointer problems.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 libhandle/handle.c |    1 +
 1 file changed, 1 insertion(+)


diff --git a/libhandle/handle.c b/libhandle/handle.c
index e6971dd..878d14d 100644
--- a/libhandle/handle.c
+++ b/libhandle/handle.c
@@ -73,6 +73,7 @@ fshandle_destroy(void)
 		free(h);
 		h = nexth;
 	}
+	fdhash_head = NULL;
 }
 
 int

[PATCH 2/5] xfs_db: check should deal with cow staging events correctly

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Teach xfs_check to record cow staging events correctly.  This means that
we strip off the high bit before using startblock.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 db/check.c |   19 +++++++++++++++----
 1 file changed, 15 insertions(+), 4 deletions(-)


diff --git a/db/check.c b/db/check.c
index 3afde02..2f8dee5 100644
--- a/db/check.c
+++ b/db/check.c
@@ -4730,13 +4730,24 @@ scanfunc_refcnt(
 		lastblock = 0;
 		for (i = 0; i < be16_to_cpu(block->bb_numrecs); i++) {
 			if (be32_to_cpu(rp[i].rc_refcount) == 1) {
-				dbprintf(_(
-		"leftover CoW extent (%u/%u) len %u\n"),
+				xfs_agblock_t	agbno;
+				char		*msg;
+
+				agbno = be32_to_cpu(rp[i].rc_startblock);
+				if (agbno >= XFS_REFC_COW_START) {
+					agbno -= XFS_REFC_COW_START;
+					msg = _(
+		"leftover CoW extent (%u/%u) len %u\n");
+				} else {
+					msg = _(
+		"leftover CoW extent at unexpected address (%u/%u) len %u\n");
+				}
+				dbprintf(msg,
 					seqno,
-					be32_to_cpu(rp[i].rc_startblock),
+					agbno,
 					be32_to_cpu(rp[i].rc_blockcount));
 				set_dbmap(seqno,
-					be32_to_cpu(rp[i].rc_startblock),
+					agbno,
 					be32_to_cpu(rp[i].rc_blockcount),
 					DBM_COWDATA, seqno, bno);
 			} else {

[PATCH 3/5] xfs_db: fix crash when field list selector string has trailing slash

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

If I run the following command:

xfs_db /dev/sdf -x -c 'agf 0' -c 'addr refcntroot' -c 'addr ptrs[1]\'

then ftok_free crashes on an invalid free() because picking up the
previous token (the closing bracket) xrealloc'd the token array to be 5
elements long but never set the last element's tok pointer.
Consequently the ftok_free tries to free whatever garbage pointer is in
that last element and kaboom.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 db/flist.c |    1 +
 1 file changed, 1 insertion(+)


diff --git a/db/flist.c b/db/flist.c
index e11acbf..b207354 100644
--- a/db/flist.c
+++ b/db/flist.c
@@ -400,6 +400,7 @@ flist_split(
 		strncpy(a, s, l);
 		a[l] = '\0';
 		v = xrealloc(v, (nv + 2) * sizeof(*v));
+		v[nv + 1].tok = NULL;
 		v[nv].tok = a;
 		v[nv].tokty = t;
 		nv++;

[PATCH 4/5] xfs_repair: fix libxfs namespace problems

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Call libxfs_ functions, not xfs_ functions.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 libxfs/libxfs_api_defs.h |    2 +-
 repair/dino_chunks.c     |    2 +-
 repair/phase6.c          |    2 +-
 3 files changed, 3 insertions(+), 3 deletions(-)


diff --git a/libxfs/libxfs_api_defs.h b/libxfs/libxfs_api_defs.h
index 84a4fc2..0a5c84d 100644
--- a/libxfs/libxfs_api_defs.h
+++ b/libxfs/libxfs_api_defs.h
@@ -115,7 +115,7 @@
 #define xfs_rmap_irec_offset_unpack	libxfs_rmap_irec_offset_unpack
 #define xfs_rmapbt_init_cursor		libxfs_rmapbt_init_cursor
 #define xfs_btree_del_cursor		libxfs_btree_del_cursor
-
+#define xfs_mode_to_ftype		libxfs_mode_to_ftype
 
 #define xfs_log_sb			libxfs_log_sb
 #define xfs_sb_from_disk		libxfs_sb_from_disk
diff --git a/repair/dino_chunks.c b/repair/dino_chunks.c
index a3909ac..17de95f 100644
--- a/repair/dino_chunks.c
+++ b/repair/dino_chunks.c
@@ -845,7 +845,7 @@ process_inode_chunk(
 			 * phase 6.
 			 */
 			set_inode_ftype(ino_rec, irec_offset,
-				xfs_mode_to_ftype(be16_to_cpu(dino->di_mode)));
+				libxfs_mode_to_ftype(be16_to_cpu(dino->di_mode)));
 
 			/*
 			 * store on-disk nlink count for comparing in phase 7
diff --git a/repair/phase6.c b/repair/phase6.c
index b326929..1a398aa 100644
--- a/repair/phase6.c
+++ b/repair/phase6.c
@@ -1081,7 +1081,7 @@ mv_orphanage(
 	if ((err = -libxfs_iget(mp, NULL, ino, 0, &ino_p)))
 		do_error(_("%d - couldn't iget disconnected inode\n"), err);
 
-	xname.type = xfs_mode_to_ftype(VFS_I(ino_p)->i_mode);
+	xname.type = libxfs_mode_to_ftype(VFS_I(ino_p)->i_mode);
 
 	if (isa_dir)  {
 		irec = find_inode_rec(mp, XFS_INO_TO_AGNO(mp, orphanage_ino),

[PATCH 5/5] find-api-violations: fix a broken grep pattern

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

One of the grep patterns in find-api-violations is mistaken for a
(broken) range specifier when LC_ALL=C, so fix it to work properly.
This was found by wiring up the script to xfstests.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 tools/find-api-violations.sh |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)


diff --git a/tools/find-api-violations.sh b/tools/find-api-violations.sh
index c2be4c5..3b976d3 100755
--- a/tools/find-api-violations.sh
+++ b/tools/find-api-violations.sh
@@ -10,7 +10,7 @@ tool_dirs="copy db estimate fsck fsr growfs io logprint mdrestore mkfs quota rep
 
 # Calls to xfs_* functions in libxfs/*.c without the libxfs_ prefix
 find_possible_api_calls() {
-	grep -rn '[[:space:],-(]xfs_[a-z_]*(' $tool_dirs | sed -e 's/^.*\(xfs_[a-z_]*\)(.*$/\1/g' | sort | uniq
+	grep -rn '[-[:space:],(]xfs_[a-z_]*(' $tool_dirs | sed -e 's/^.*\(xfs_[a-z_]*\)(.*$/\1/g' | sort | uniq
 }
 
 check_if_api_calls() {

Re: [PATCH 2/5] xfs_db: check should deal with cow staging events correctly

[Darrick J. Wong]

On Wed, Jan 03, 2018 at 10:40:19AM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
> Subject: [PATCH 2/5] xfs_db: check should deal with cow staging events correctly

s/events/extents/

> Teach xfs_check to record cow staging events correctly.  This means that

s/events/extents/

> we strip off the high bit before using startblock.

high, indeed. :P

--D

> 
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> ---
>  db/check.c |   19 +++++++++++++++----
>  1 file changed, 15 insertions(+), 4 deletions(-)
> 
> 
> diff --git a/db/check.c b/db/check.c
> index 3afde02..2f8dee5 100644
> --- a/db/check.c
> +++ b/db/check.c
> @@ -4730,13 +4730,24 @@ scanfunc_refcnt(
>  		lastblock = 0;
>  		for (i = 0; i < be16_to_cpu(block->bb_numrecs); i++) {
>  			if (be32_to_cpu(rp[i].rc_refcount) == 1) {
> -				dbprintf(_(
> -		"leftover CoW extent (%u/%u) len %u\n"),
> +				xfs_agblock_t	agbno;
> +				char		*msg;
> +
> +				agbno = be32_to_cpu(rp[i].rc_startblock);
> +				if (agbno >= XFS_REFC_COW_START) {
> +					agbno -= XFS_REFC_COW_START;
> +					msg = _(
> +		"leftover CoW extent (%u/%u) len %u\n");
> +				} else {
> +					msg = _(
> +		"leftover CoW extent at unexpected address (%u/%u) len %u\n");
> +				}
> +				dbprintf(msg,
>  					seqno,
> -					be32_to_cpu(rp[i].rc_startblock),
> +					agbno,
>  					be32_to_cpu(rp[i].rc_blockcount));
>  				set_dbmap(seqno,
> -					be32_to_cpu(rp[i].rc_startblock),
> +					agbno,
>  					be32_to_cpu(rp[i].rc_blockcount),
>  					DBM_COWDATA, seqno, bno);
>  			} else {
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 2/5] xfs_db: check should deal with cow staging events correctly

[Eric Sandeen]

On 1/3/18 12:40 PM, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
> 
> Teach xfs_check to record cow staging events correctly.  This means that
> we strip off the high bit before using startblock.
> 
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>

Reviewed-by: Eric Sandeen <sandeen@redhat.com>

> ---
>  db/check.c |   19 +++++++++++++++----
>  1 file changed, 15 insertions(+), 4 deletions(-)
> 
> 
> diff --git a/db/check.c b/db/check.c
> index 3afde02..2f8dee5 100644
> --- a/db/check.c
> +++ b/db/check.c
> @@ -4730,13 +4730,24 @@ scanfunc_refcnt(
>  		lastblock = 0;
>  		for (i = 0; i < be16_to_cpu(block->bb_numrecs); i++) {
>  			if (be32_to_cpu(rp[i].rc_refcount) == 1) {
> -				dbprintf(_(
> -		"leftover CoW extent (%u/%u) len %u\n"),
> +				xfs_agblock_t	agbno;
> +				char		*msg;
> +
> +				agbno = be32_to_cpu(rp[i].rc_startblock);
> +				if (agbno >= XFS_REFC_COW_START) {
> +					agbno -= XFS_REFC_COW_START;
> +					msg = _(
> +		"leftover CoW extent (%u/%u) len %u\n");
> +				} else {
> +					msg = _(
> +		"leftover CoW extent at unexpected address (%u/%u) len %u\n");
> +				}
> +				dbprintf(msg,
>  					seqno,
> -					be32_to_cpu(rp[i].rc_startblock),
> +					agbno,
>  					be32_to_cpu(rp[i].rc_blockcount));
>  				set_dbmap(seqno,
> -					be32_to_cpu(rp[i].rc_startblock),
> +					agbno,
>  					be32_to_cpu(rp[i].rc_blockcount),
>  					DBM_COWDATA, seqno, bno);
>  			} else {
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 

Re: [PATCH 4/5] xfs_repair: fix libxfs namespace problems

[Eric Sandeen]

On 1/3/18 12:40 PM, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
> 
> Call libxfs_ functions, not xfs_ functions.
> 
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>

Oops, thanks.

Reviewed-by: Eric Sandeen <sandeen@redhat.com>

Re: [PATCH 5/5] find-api-violations: fix a broken grep pattern

[Eric Sandeen]

On 1/3/18 12:40 PM, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
> 
> One of the grep patterns in find-api-violations is mistaken for a
> (broken) range specifier when LC_ALL=C, so fix it to work properly.
> This was found by wiring up the script to xfstests.
> 
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>

I'm gonna have to take your word for it that your cat walked correctly
on the keyboard this time.

Reviewed-by: Eric Sandeen <sandeen@redhat.com>

> ---
>  tools/find-api-violations.sh |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> 
> diff --git a/tools/find-api-violations.sh b/tools/find-api-violations.sh
> index c2be4c5..3b976d3 100755
> --- a/tools/find-api-violations.sh
> +++ b/tools/find-api-violations.sh
> @@ -10,7 +10,7 @@ tool_dirs="copy db estimate fsck fsr growfs io logprint mdrestore mkfs quota rep
>  
>  # Calls to xfs_* functions in libxfs/*.c without the libxfs_ prefix
>  find_possible_api_calls() {
> -	grep -rn '[[:space:],-(]xfs_[a-z_]*(' $tool_dirs | sed -e 's/^.*\(xfs_[a-z_]*\)(.*$/\1/g' | sort | uniq
> +	grep -rn '[-[:space:],(]xfs_[a-z_]*(' $tool_dirs | sed -e 's/^.*\(xfs_[a-z_]*\)(.*$/\1/g' | sort | uniq
>  }
>  
>  check_if_api_calls() {
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>