Skip to contentSkip to conversation

tmpfs: fix mounts when size is less than the page size

PATCH Superseded Michael Marineau 2 participants 5 messages Nov 3, 2007 – Nov 4, 2007

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 tmpfs: fix mounts when size is less than the page size

When tmpfs is mounted with a size less than one page the number of blocks is set to 0 which makes the tmpfs mount unlimited. This can lead to a quick and surprising death is someone typos a tmpfs mount command and writes to much.

tmpfs can still be mounted as unlimited if size or nr_blocks is exactly 0.

---
 mm/shmem.c |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)
mm/shmem.c +2 -0
index 253d205..66b07f2 100644
@@ -2138,6 +2138,8 @@ static int shmem_parse_options(char *options, int *mode, uid_t *uid,
 			if (*rest)
 				goto bad_val;
 			*blocks = size >> PAGE_CACHE_SHIFT;
+			if (size && blocks == 0)
+				blocks = 1;
 		} else if (!strcmp(this_char,"nr_blocks")) {
 			*blocks = memparse(value,&rest);
 			if (*rest)
Signature · 1 line
1.5.1.6
#1 lore raw

PATCH RESEND 1/1 tmpfs: fix mounts when size is less than the page size

When tmpfs is mounted with a size less than one page the number of blocks is set to 0 which makes the tmpfs mount unlimited. This can lead to a quick and surprising death is someone typos a tmpfs mount command and writes to much.

tmpfs can still be mounted as unlimited if size or nr_blocks is exactly 0.

---
Somehow I was a moron and didn't notice the compiler shouting insults at me, here is the correct version of the patch. Sorry for the noise :-(

 mm/shmem.c |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)
mm/shmem.c +2 -0
index 253d205..86b47d8 100644
@@ -2138,6 +2138,8 @@ static int shmem_parse_options(char *options, int *mode, uid_t *uid,
 			if (*rest)
 				goto bad_val;
 			*blocks = size >> PAGE_CACHE_SHIFT;
+			if (size && *blocks == 0)
+				*blocks = 1;
 		} else if (!strcmp(this_char,"nr_blocks")) {
 			*blocks = memparse(value,&rest);
 			if (*rest)
Signature · 1 line
1.5.1.6
#2 lore raw

Fair enough, that does match the documentation of "size=0" better. Though either way, someone who typos is going to get one kind of surprise or another. Do you mind if we do it slightly differently, achieving the same effect without a special case, by rounding up instead of down?

From: Michael Marineau <[email protected]>

When tmpfs is mounted with a size less than one page, the number of blocks is set to 0 which makes the tmpfs mount unlimited. This can lead to a quick and surprising death if someone typos a tmpfs mount command and writes too much.

tmpfs can still be mounted as unlimited if size or nr_blocks is exactly 0, as Documentation/filesystems/tmpfs.txt says.

Hugh: do this by rounding size up instead of down in all cases: which slightly expands other odd-sized tmpfs mounts, but in a consistent way.

---

 mm/shmem.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
mm/shmem.c +1 -1
@@ -2137,7 +2137,7 @@ static int shmem_parse_options(char *opt
 			}
 			if (*rest)
 				goto bad_val;
-			*blocks = size >> PAGE_CACHE_SHIFT;
+			*blocks = DIV_ROUND_UP(size, PAGE_CACHE_SIZE);
 		} else if (!strcmp(this_char,"nr_blocks")) {
 			*blocks = memparse(value,&rest);
 			if (*rest)
#3 lore raw
Quoted 4 lines · Hugh Dickins

Fair enough, that does match the documentation of "size=0" better. Though either way, someone who typos is going to get one kind of surprise or another. Do you mind if we do it slightly differently, achieving the same effect without a special case, by rounding up instead of down?

Sounds reasonable to me. I didn't know about DIV_ROUND_UP :-)

Quoted 34 lines

From: Michael Marineau <[email protected]>

When tmpfs is mounted with a size less than one page, the number of blocks is set to 0 which makes the tmpfs mount unlimited. This can lead to a quick and surprising death if someone typos a tmpfs mount command and writes too much.

tmpfs can still be mounted as unlimited if size or nr_blocks is exactly 0, as Documentation/filesystems/tmpfs.txt says.

Hugh: do this by rounding size up instead of down in all cases: which slightly expands other odd-sized tmpfs mounts, but in a consistent way.

---

 mm/shmem.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
mm/shmem.c 2007-11-04 11:46:09.000000000 +0000 renamed from mm/shmem.c 2007-10-31 06:18:05.000000000 +0000 +1 -1
@@ -2137,7 +2137,7 @@ static int shmem_parse_options(char *opt
                        }
                        if (*rest)
                                goto bad_val;
-                       *blocks = size >> PAGE_CACHE_SHIFT;
+                       *blocks = DIV_ROUND_UP(size, PAGE_CACHE_SIZE);
                } else if (!strcmp(this_char,"nr_blocks")) {
                        *blocks = memparse(value,&rest);
                        if (*rest)
Signature · 3 lines
Michael Marineau
Oregon State University
[email protected]
#4 lore raw
Quoted 7 lines · Michael Marineau
Quoted 4 lines · Hugh Dickins

Fair enough, that does match the documentation of "size=0" better. Though either way, someone who typos is going to get one kind of surprise or another. Do you mind if we do it slightly differently, achieving the same effect without a special case, by rounding up instead of down?

Sounds reasonable to me. I didn't know about DIV_ROUND_UP :-)

I'm not familiar with it either (had to look it up to check it's the right thing) - in mm/ we're accustomed to writing things like "(size + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT"; but that took the line over 80 columns, and I dreaded someone somewhere mailing in to ask "Why didn't you use DIV_ROUND_UP?" ;-)

Hugh