Skip to contentSkip to conversation

cciss: use safer test on the result of find_first_zero_bit

PATCH [email protected] 3 participants 5 messages Jun 4, 2014

lore
Share & tools
lore

[email protected]

author Reviewed-by
#0 lore raw
Quoted 32 lines

Find_first_zero_bit considers BITS_PER_LONG bits at a time, and thus may return a larger number than the maximum position argument if that position is not a multiple of BITS_PER_LONG.

The semantic match that finds this problem is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
expression e1,e2,e3;
statement S1,S2;
@@

e1 = find_first_zero_bit(e2,e3)
...
if (e1
- ==
+ >=
e3)
S1 else S2
// </smpl>

---
 drivers/block/cciss.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff -u -p a/drivers/block/cciss.c b/drivers/block/cciss.c

drivers/block/cciss.c static CommandList_struct *cmd_alloc(ctl · line 986

Quoted 8 lines
drivers/block/cciss.c +1 -1
@@ -980,7 +980,7 @@ static CommandList_struct *cmd_alloc(ctl
 
 	do {
 		i = find_first_zero_bit(h->cmd_pool_bits, h->nr_cmds);
-		if (i == h->nr_cmds)
+		if (i >= h->nr_cmds)
 			return NULL;
 	} while (test_and_set_bit(i, h->cmd_pool_bits) != 0);
 	c = h->cmd_pool + i;

Thanks. Ack.

You can add

to this patch if you want.

You might consider adding "Cc: [email protected]" into the sign-off area as well.

-- steve

#1 lore raw
Quoted 44 lines · On Wed, 4 Jun 2014, [email protected] wrote:
Quoted 32 lines

Find_first_zero_bit considers BITS_PER_LONG bits at a time, and thus may return a larger number than the maximum position argument if that position is not a multiple of BITS_PER_LONG.

The semantic match that finds this problem is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
expression e1,e2,e3;
statement S1,S2;
@@

e1 = find_first_zero_bit(e2,e3)
...
if (e1
- ==
+ >=
e3)
S1 else S2
// </smpl>

---
 drivers/block/cciss.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff -u -p a/drivers/block/cciss.c b/drivers/block/cciss.c

drivers/block/cciss.c static CommandList_struct *cmd_alloc(ctl · line 986

Quoted 8 lines
drivers/block/cciss.c +1 -1
@@ -980,7 +980,7 @@ static CommandList_struct *cmd_alloc(ctl
 
 	do {
 		i = find_first_zero_bit(h->cmd_pool_bits, h->nr_cmds);
-		if (i == h->nr_cmds)
+		if (i >= h->nr_cmds)
 			return NULL;
 	} while (test_and_set_bit(i, h->cmd_pool_bits) != 0);
 	c = h->cmd_pool + i;

Thanks. Ack.

Quoted 8 lines

You can add

to this patch if you want.

You might consider adding "Cc: [email protected]" into the sign-off area as well.

Likewise here, the change is not needed.

julia

#2 lore raw
In reply to [email protected]
Quoted 44 lines · On 06/04/2014 08:51 AM, [email protected] wrote:
Quoted 32 lines

Find_first_zero_bit considers BITS_PER_LONG bits at a time, and thus may return a larger number than the maximum position argument if that position is not a multiple of BITS_PER_LONG.

The semantic match that finds this problem is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
expression e1,e2,e3;
statement S1,S2;
@@

e1 = find_first_zero_bit(e2,e3)
...
if (e1
- ==
+ >=
e3)
S1 else S2
// </smpl>

---
 drivers/block/cciss.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff -u -p a/drivers/block/cciss.c b/drivers/block/cciss.c

drivers/block/cciss.c static CommandList_struct *cmd_alloc(ctl · line 986

Quoted 8 lines
drivers/block/cciss.c +1 -1
@@ -980,7 +980,7 @@ static CommandList_struct *cmd_alloc(ctl
 
 	do {
 		i = find_first_zero_bit(h->cmd_pool_bits, h->nr_cmds);
-		if (i == h->nr_cmds)
+		if (i >= h->nr_cmds)
 			return NULL;
 	} while (test_and_set_bit(i, h->cmd_pool_bits) != 0);
 	c = h->cmd_pool + i;

Thanks. Ack.

Quoted 8 lines

You can add

to this patch if you want.

You might consider adding "Cc: [email protected]" into the sign-off area as well.

There are two such instances in cciss.c, btw.

Signature · 1 line
Jens Axboe
#3 lore raw
Quoted 47 lines · Jens Axboe
Quoted 46 lines · On 06/04/2014 08:51 AM, [email protected] wrote:
Quoted 32 lines

Find_first_zero_bit considers BITS_PER_LONG bits at a time, and thus may return a larger number than the maximum position argument if that position is not a multiple of BITS_PER_LONG.

The semantic match that finds this problem is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
expression e1,e2,e3;
statement S1,S2;
@@

e1 = find_first_zero_bit(e2,e3)
...
if (e1
- ==
+ >=
e3)
S1 else S2
// </smpl>

---
 drivers/block/cciss.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff -u -p a/drivers/block/cciss.c b/drivers/block/cciss.c

drivers/block/cciss.c static CommandList_struct *cmd_alloc(ctl · line 986

Quoted 8 lines
drivers/block/cciss.c +1 -1
@@ -980,7 +980,7 @@ static CommandList_struct *cmd_alloc(ctl
 
 	do {
 		i = find_first_zero_bit(h->cmd_pool_bits, h->nr_cmds);
-		if (i == h->nr_cmds)
+		if (i >= h->nr_cmds)
 			return NULL;
 	} while (test_and_set_bit(i, h->cmd_pool_bits) != 0);
 	c = h->cmd_pool + i;

Thanks. Ack.

You can add

Quoted 8 lines
Quoted 6 lines

to this patch if you want.

You might consider adding "Cc: [email protected]" into the sign-off area as well.

There are two such instances in cciss.c, btw.

Actually, there seem to be three, and I didn't find the other two because the assignment is inlined into the test. But the patch isn't needed anyway, because it turns out that the result never goes over the bound value.

thanks,
julia

#4 lore raw
Quoted 58 lines · Julia Lawall
Quoted 54 lines · Jens Axboe
Quoted 52 lines · On 06/04/2014 08:51 AM, [email protected] wrote:
Quoted 32 lines

Find_first_zero_bit considers BITS_PER_LONG bits at a time, and thus may return a larger number than the maximum position argument if that position is not a multiple of BITS_PER_LONG.

The semantic match that finds this problem is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
expression e1,e2,e3;
statement S1,S2;
@@

e1 = find_first_zero_bit(e2,e3)
...
if (e1
- ==
+ >=
e3)
S1 else S2
// </smpl>

---
 drivers/block/cciss.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff -u -p a/drivers/block/cciss.c b/drivers/block/cciss.c

drivers/block/cciss.c static CommandList_struct *cmd_alloc(ctl · line 986

Quoted 8 lines
drivers/block/cciss.c +1 -1
@@ -980,7 +980,7 @@ static CommandList_struct *cmd_alloc(ctl
 
 	do {
 		i = find_first_zero_bit(h->cmd_pool_bits, h->nr_cmds);
-		if (i == h->nr_cmds)
+		if (i >= h->nr_cmds)
 			return NULL;
 	} while (test_and_set_bit(i, h->cmd_pool_bits) != 0);
 	c = h->cmd_pool + i;

Thanks. Ack.

You can add

to this patch if you want.

You might consider adding "Cc: [email protected]" into the sign-off area as well.

Quoted 6 lines
Quoted 1 line

There are two such instances in cciss.c, btw.

Actually, there seem to be three, and I didn't find the other two because the assignment is inlined into the test. But the patch isn't needed anyway, because it turns out that the result never goes over the bound value.

I have always defensively programmed it, but it would make a shitty API if it did.

Signature · 1 line
Jens Axboe