[PATCH] afs: use BIT_ULL for shifting to fix integer overflow

[PATCH] afs: use BIT_ULL for shifting to fix integer overflow

[Colin King]

From: Colin Ian King <colin.king@canonical.com>

The expression 1 << nr_slots is evaluated with 32 bit integer arithmetic
and can overflow before it is widened. Instead, use BIT_ULL to avoid
overflow.

Addresses-Coverity: ("Unintentional integer overflow")
Fixes: 63a4681ff39c ("afs: Locally edit directory data for mkdir/create/unlink/...")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
---
 fs/afs/dir_edit.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/fs/afs/dir_edit.c b/fs/afs/dir_edit.c
index d4fbe5f85f1b..f360119923aa 100644
--- a/fs/afs/dir_edit.c
+++ b/fs/afs/dir_edit.c
@@ -36,7 +36,7 @@ static int afs_find_contig_bits(union afs_xdr_dir_block *block, unsigned int nr_
 	bitmap |= (u64)block->hdr.bitmap[7] << 7 * 8;
 	bitmap >>= 1; /* The first entry is metadata */
 	bit = 1;
-	mask = (1 << nr_slots) - 1;
+	mask = BIT_ULL(nr_slots) - 1;
 
 	do {
 		if (sizeof(unsigned long) == 8)
@@ -70,7 +70,7 @@ static void afs_set_contig_bits(union afs_xdr_dir_block *block,
 {
 	u64 mask, before, after;
 
-	mask = (1 << nr_slots) - 1;
+	mask = BIT_ULL(nr_slots) - 1;
 	mask <<= bit;
 
 	before = *(u64 *)block->hdr.bitmap;
@@ -95,7 +95,7 @@ static void afs_clear_contig_bits(union afs_xdr_dir_block *block,
 {
 	u64 mask, before, after;
 
-	mask = (1 << nr_slots) - 1;
+	mask = BIT_ULL(nr_slots) - 1;
 	mask <<= bit;
 
 	before = *(u64 *)block->hdr.bitmap;
-- 
2.20.1

Re: [PATCH] afs: use BIT_ULL for shifting to fix integer overflow

[David Howells]

Colin King <colin.king@canonical.com> wrote:

> The expression 1 << nr_slots is evaluated with 32 bit integer arithmetic
> and can overflow before it is widened.

If it does, it's an error on the part of the caller.  See the banner comment:
1 <= nr_slots <= 9.

And, in any case, if nr_slots >= 64, using BIT_ULL wouldn't help...

David

Re: [PATCH] afs: use BIT_ULL for shifting to fix integer overflow

[Colin Ian King]

On 30/08/2019 15:01, David Howells wrote:
> Colin King <colin.king@canonical.com> wrote:
> 
>> The expression 1 << nr_slots is evaluated with 32 bit integer arithmetic
>> and can overflow before it is widened.
> 
> If it does, it's an error on the part of the caller.  See the banner comment:
> 1 <= nr_slots <= 9.

Doh, totally missed that. My bad.

> 
> And, in any case, if nr_slots >= 64, using BIT_ULL wouldn't help...
> 
> David
>