Skip to contentSkip to conversation

xfsprogs: avoid to use NULL pointer

PATCH In review zhanchengbin 3 participants 6 messages Jun 6, 2022

lore
Share & tools
lore
#0 lore raw

cover xfsprogs: avoid to use NULL pointer

Change malloc to xmalloc for avoid to use NULL pointer.

zhanchengbin (3):
mkfs/proto.c: avoid to use NULL pointer
db: avoid to use NULL pointer
repair/slab.c: avoid to use NULL pointer

 db/block.c    | 2 +-
 db/check.c    | 4 ++--
 db/faddr.c    | 6 +++---
 db/namei.c    | 2 +-
 mkfs/proto.c  | 4 ++--
 repair/slab.c | 2 +-
 6 files changed, 10 insertions(+), 10 deletions(-)
Signature · 1 line
2.27.0
#1 lore raw

PATCH 1/3 mkfs/proto.c: avoid to use NULL pointer

Change malloc to xmalloc.

---
 mkfs/proto.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
mkfs/proto.c +2 -2
index 127d87dd..f3b8710c 100644
@@ -70,7 +70,7 @@ setup_proto(
 		goto out_fail;
 	}
 
-	buf = malloc(size + 1);
+	buf = xmalloc(size + 1);
 	if (read(fd, buf, size) < size) {
 		fprintf(stderr, _("%s: read failed on %s: %s\n"),
 			progname, fname, strerror(errno));
@@ -303,7 +303,7 @@ newregfile(
 		exit(1);
 	}
 	if ((*len = (int)size)) {
-		buf = malloc(size);
+		buf = xmalloc(size);
 		if (read(fd, buf, size) < size) {
 			fprintf(stderr, _("%s: read failed on %s: %s\n"),
 				progname, fname, strerror(errno));
Signature · 1 line
2.27.0
#2 lore raw

Re: [PATCH 1/3] mkfs/proto.c: avoid to use NULL pointer

Quoted 1 line · zhanchengbin

Change malloc to xmalloc.

The commit log and cover letter say nothing about this, but apparently "xmalloc" is often defined locally to abort() on a memory failure, so I guess you are trying to make (some of?) xfsprogs do that.

First, this doesn't seem to build....

Building mkfs
[CC] proto.o
proto.c: In function ‘setup_proto’:
proto.c:73:15: warning: implicit declaration of function ‘xmalloc’; did you mean ‘malloc’? [-Wimplicit-function-declaration]

   73 |         buf = xmalloc(size + 1);
      |               ^~~~~~~
      |               malloc

proto.c:73:13: warning: assignment to ‘char *’ from ‘int’ makes pointer from integer without a cast [-Wint-conversion]
73 | buf = xmalloc(size + 1);
| ^
proto.c: In function ‘newregfile’:
proto.c:306:21: warning: assignment to ‘char *’ from ‘int’ makes pointer from integer without a cast [-Wint-conversion]
306 | buf = xmalloc(size);
| ^
[LD] mkfs.xfs
/usr/bin/ld: proto.o: in function `setup_proto':
/src/git/xfsprogs-dev/mkfs/proto.c:73: undefined reference to `xmalloc'
/usr/bin/ld: proto.o: in function `newregfile':
/src/git/xfsprogs-dev/mkfs/proto.c:306: undefined reference to `xmalloc'
collect2: error: ld returned 1 exit status
gmake[2]: *** [../include/buildrules:65: mkfs.xfs] Error 1
gmake[1]: *** [include/buildrules:36: mkfs] Error 2
make: *** [Makefile:92: default] Error 2

Second, we have calls to malloc all over the codebase, including around 100 outside of the internal libraries in the various userspace utilities. Why change only mkfs/db/repair?

Third, what problem are you solving? Granted, we should be checking every malloc call, and it seems that we don't. But have you ever seen these failures in practice? Your commit log should at least explain why this makes the code better (and, I guess, where xmalloc is supposed to come from...)

-Eric

Quoted 27 lines

---
 mkfs/proto.c | 4 ++--

 1 file changed, 2 insertions(+), 2 deletions(-)
mkfs/proto.c +2 -2
index 127d87dd..f3b8710c 100644
@@ -70,7 +70,7 @@ setup_proto(
          goto out_fail;
      }
 
-    buf = malloc(size + 1);
+    buf = xmalloc(size + 1);
      if (read(fd, buf, size) < size) {
          fprintf(stderr, _("%s: read failed on %s: %s\n"),
              progname, fname, strerror(errno));
@@ -303,7 +303,7 @@ newregfile(
          exit(1);
      }
      if ((*len = (int)size)) {
-        buf = malloc(size);
+        buf = xmalloc(size);
          if (read(fd, buf, size) < size) {
              fprintf(stderr, _("%s: read failed on %s: %s\n"),
                  progname, fname, strerror(errno));
#3 lore raw

Re: [PATCH 1/3] mkfs/proto.c: avoid to use NULL pointer

Quoted 33 lines · Eric Sandeen
Quoted 1 line · zhanchengbin

Change malloc to xmalloc.

The commit log and cover letter say nothing about this, but apparently "xmalloc" is often defined locally to abort() on a memory failure, so I guess you are trying to make (some of?) xfsprogs do that.

First, this doesn't seem to build....

Building mkfs
[CC] proto.o
proto.c: In function ‘setup_proto’:
proto.c:73:15: warning: implicit declaration of function ‘xmalloc’; did you mean ‘malloc’? [-Wimplicit-function-declaration]

   73 |         buf = xmalloc(size + 1);
      |               ^~~~~~~
      |               malloc

proto.c:73:13: warning: assignment to ‘char *’ from ‘int’ makes pointer from integer without a cast [-Wint-conversion]
73 | buf = xmalloc(size + 1);
| ^
proto.c: In function ‘newregfile’:
proto.c:306:21: warning: assignment to ‘char *’ from ‘int’ makes pointer from integer without a cast [-Wint-conversion]
306 | buf = xmalloc(size);
| ^
[LD] mkfs.xfs
/usr/bin/ld: proto.o: in function `setup_proto':
/src/git/xfsprogs-dev/mkfs/proto.c:73: undefined reference to `xmalloc'
/usr/bin/ld: proto.o: in function `newregfile':
/src/git/xfsprogs-dev/mkfs/proto.c:306: undefined reference to `xmalloc'
collect2: error: ld returned 1 exit status
gmake[2]: *** [../include/buildrules:65: mkfs.xfs] Error 1
gmake[1]: *** [include/buildrules:36: mkfs] Error 2
make: *** [Makefile:92: default] Error 2

Quoted 3 lines

Second, we have calls to malloc all over the codebase, including around 100 outside of the internal libraries in the various userspace utilities. Why change only mkfs/db/repair?

xmalloc is local to xfs_db (db/malloc.c), it's not part of libxfs. It's local to xfs_db because it is the tool that always runs out of memory on large filesystems. It's just a wrapper that prints "out of memory" and then exits immediately with no chance to recover. So, essentially, it's no different to immediately SEGV on a NULL pointer and dying.

Quoted 2 lines

Third, what problem are you solving? Granted, we should be checking every malloc call, and it seems that we don't.

In general, that's because we can't do anything useful to free memory when we fail memory allocation - we generally exit immediately on memory allocation failure, so checking for failure is almost redundant....

Cheers,

Dave.

Signature · 2 lines
Dave Chinner
[email protected]
#4 lore raw

PATCH 2/3 db: avoid to use NULL pointer

In reply to zhanchengbin

Changed malloc to xmalloc.

---
 db/block.c | 2 +-
 db/check.c | 4 ++--
 db/faddr.c | 6 +++---
 db/namei.c | 2 +-
 4 files changed, 7 insertions(+), 7 deletions(-)
db/block.c +1 -1
index f31f9544..349cf6ec 100644
@@ -180,7 +180,7 @@ dblock_f(
 		return 0;
 	}
 	nex = nb = type == TYP_DIR2 ? mp->m_dir_geo->fsbcount : 1;
-	bmp = malloc(nb * sizeof(*bmp));
+	bmp = xmalloc(nb * sizeof(*bmp));
 	bmap(bno, nb, XFS_DATA_FORK, &nex, bmp);
 	if (nex == 0) {
 		dbprintf(_("file data block is unmapped\n"));
db/check.c +2 -2
index a9c6e632..cc3c7832 100644
@@ -1782,7 +1782,7 @@ dir_hash_add(
 	dirhash_t		*p;
 
 	i = DIR_HASH_FUNC(hash, addr);
-	p = malloc(sizeof(*p));
+	p = xmalloc(sizeof(*p));
 	p->next = dirhash[i];
 	dirhash[i] = p;
 	p->hashval = hash;
@@ -3094,7 +3094,7 @@ process_leaf_node_dir_v2(
 	v2 = verbose || id->ilist;
 	v = parent = 0;
 	dbno = NULLFILEOFF;
-	freetab = malloc(FREETAB_SIZE(dirsize / mp->m_dir_geo->blksize));
+	freetab = xmalloc(FREETAB_SIZE(dirsize / mp->m_dir_geo->blksize));
 	freetab->naents = (int)(dirsize / mp->m_dir_geo->blksize);
 	freetab->nents = 0;
 	for (i = 0; i < freetab->naents; i++)
db/faddr.c +3 -3
index 0127c5d1..cc21faaf 100644
@@ -137,7 +137,7 @@ fa_cfileoffd(
 		return;
 	}
 	nex = nb = next == TYP_DIR2 ? mp->m_dir_geo->fsbcount : 1;
-	bmp = malloc(nb * sizeof(*bmp));
+	bmp = xmalloc(nb * sizeof(*bmp));
 	bmap(bno, nb, XFS_DATA_FORK, &nex, bmp);
 	if (nex == 0) {
 		dbprintf(_("file block is unmapped\n"));
@@ -221,7 +221,7 @@ fa_dfiloffd(
 		return;
 	}
 	nex = nb = next == TYP_DIR2 ? mp->m_dir_geo->fsbcount : 1;
-	bmp = malloc(nb * sizeof(*bmp));
+	bmp = xmalloc(nb * sizeof(*bmp));
 	bmap(bno, nb, XFS_DATA_FORK, &nex, bmp);
 	if (nex == 0) {
 		dbprintf(_("file block is unmapped\n"));
@@ -274,7 +274,7 @@ fa_dirblock(
 		return;
 	}
 	nex = mp->m_dir_geo->fsbcount;
-	bmp = malloc(nex * sizeof(*bmp));
+	bmp = xmalloc(nex * sizeof(*bmp));
 	bmap(bno, mp->m_dir_geo->fsbcount, XFS_DATA_FORK, &nex, bmp);
 	if (nex == 0) {
 		dbprintf(_("directory block is unmapped\n"));
db/namei.c +1 -1
index e44667a9..d1a9904b 100644
@@ -266,7 +266,7 @@ dir_emit(
 		 * copy the string to a buffer so that we can add the null
 		 * terminator.
 		 */
-		display_name = malloc(namelen + 1);
+		display_name = xmalloc(namelen + 1);
 		memcpy(display_name, name, namelen);
 		display_name[namelen] = 0;
 		xname.len = namelen;
Signature · 1 line
2.27.0
#5 lore raw

PATCH 3/3 repair/slab.c: avoid to use NULL pointer

In reply to zhanchengbin

Changed malloc to xmalloc.

---
 repair/slab.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
repair/slab.c +1 -1
index 165f97ef..13fc8203 100644
@@ -237,7 +237,7 @@ qsort_slab(
 	create_work_queue(&wq, NULL, platform_nproc());
 	hdr = slab->s_first;
 	while (hdr) {
-		qs = malloc(sizeof(struct qsort_slab));
+		qs = xmalloc(sizeof(struct qsort_slab));
 		qs->slab = slab;
 		qs->hdr = hdr;
 		qs->compare_fn = compare_fn;
Signature · 1 line
2.27.0