[PATCH -next] nbd: fix possible overflow on 'first_minor' in nbd_dev_add()

[PATCH -next] nbd: fix possible overflow on 'first_minor' in nbd_dev_add()

[Zhang Wensheng]

When 'index' is a big numbers, it may become negative which forced
to 'int'. then 'index << part_shift' might overflow to a positive
value that is not greater than '0xfffff', then sysfs might complains
about duplicate creation. Because of this, move the 'index' judgment
to the front will fix it and be better.

Fixes: b0d9111a2d53 ("nbd: use an idr to keep track of nbd devices")
Fixes: 940c264984fd ("nbd: fix possible overflow for 'first_minor' in nbd_dev_add()")
Signed-off-by: Zhang Wensheng <zhangwensheng5@huawei.com>
---
 drivers/block/nbd.c | 23 ++++++++++++-----------
 1 file changed, 12 insertions(+), 11 deletions(-)

diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c
index 5a1f98494ddd..9448aacbcf0f 100644
--- a/drivers/block/nbd.c
+++ b/drivers/block/nbd.c
@@ -1800,17 +1800,7 @@ static struct nbd_device *nbd_dev_add(int index, unsigned int refs)
 	refcount_set(&nbd->refs, 0);
 	INIT_LIST_HEAD(&nbd->list);
 	disk->major = NBD_MAJOR;
-
-	/* Too big first_minor can cause duplicate creation of
-	 * sysfs files/links, since index << part_shift might overflow, or
-	 * MKDEV() expect that the max bits of first_minor is 20.
-	 */
 	disk->first_minor = index << part_shift;
-	if (disk->first_minor < index || disk->first_minor > MINORMASK) {
-		err = -EINVAL;
-		goto out_free_work;
-	}
-
 	disk->minors = 1 << part_shift;
 	disk->fops = &nbd_fops;
 	disk->private_data = nbd;
@@ -1915,8 +1905,19 @@ static int nbd_genl_connect(struct sk_buff *skb, struct genl_info *info)
 	if (!netlink_capable(skb, CAP_SYS_ADMIN))
 		return -EPERM;
 
-	if (info->attrs[NBD_ATTR_INDEX])
+	if (info->attrs[NBD_ATTR_INDEX]) {
 		index = nla_get_u32(info->attrs[NBD_ATTR_INDEX]);
+
+		/*
+		 * Too big first_minor can cause duplicate creation of
+		 * sysfs files/links, since index << part_shift might overflow, or
+		 * MKDEV() expect that the max bits of first_minor is 20.
+		 */
+		if (index < 0 || index > MINORMASK >> part_shift) {
+			printk(KERN_ERR "nbd: illegal input index %d\n", index);
+			return -EINVAL;
+		}
+	}
 	if (!info->attrs[NBD_ATTR_SOCKETS]) {
 		printk(KERN_ERR "nbd: must specify at least one socket\n");
 		return -EINVAL;
-- 
2.31.1

Re: [PATCH -next] nbd: fix possible overflow on 'first_minor' in nbd_dev_add()

[Jens Axboe]

On 4/6/22 5:24 AM, Zhang Wensheng wrote:
> When 'index' is a big numbers, it may become negative which forced
> to 'int'. then 'index << part_shift' might overflow to a positive
> value that is not greater than '0xfffff', then sysfs might complains
> about duplicate creation. Because of this, move the 'index' judgment
> to the front will fix it and be better.

What's changed in this version? Always add a v2 to both the subject
line and below the '---' so that folks reviewing this will know
what changes you made since the last posting.

-- 
Jens Axboe

[PATCH -next] nbd: fix possible overflow on 'first_minor' in nbd_dev_add()

[Zhang Wensheng]

When 'index' is a big numbers, it may become negative which forced
to 'int'. then 'index << part_shift' might overflow to a positive
value that is not greater than '0xfffff', then sysfs might complains
about duplicate creation. Because of this, move the 'index' judgment
to the front will fix it and be better.

Fixes: b0d9111a2d53 ("nbd: use an idr to keep track of nbd devices")
Fixes: 940c264984fd ("nbd: fix possible overflow for 'first_minor' in nbd_dev_add()")
Signed-off-by: Zhang Wensheng <zhangwensheng5@huawei.com>
---
 drivers/block/nbd.c | 24 ++++++++++++------------
 1 file changed, 12 insertions(+), 12 deletions(-)

diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c
index 5a1f98494ddd..b3cdfc0ffb98 100644
--- a/drivers/block/nbd.c
+++ b/drivers/block/nbd.c
@@ -1800,17 +1800,6 @@ static struct nbd_device *nbd_dev_add(int index, unsigned int refs)
 	refcount_set(&nbd->refs, 0);
 	INIT_LIST_HEAD(&nbd->list);
 	disk->major = NBD_MAJOR;
-
-	/* Too big first_minor can cause duplicate creation of
-	 * sysfs files/links, since index << part_shift might overflow, or
-	 * MKDEV() expect that the max bits of first_minor is 20.
-	 */
-	disk->first_minor = index << part_shift;
-	if (disk->first_minor < index || disk->first_minor > MINORMASK) {
-		err = -EINVAL;
-		goto out_free_work;
-	}
-
 	disk->minors = 1 << part_shift;
 	disk->fops = &nbd_fops;
 	disk->private_data = nbd;
@@ -1915,8 +1904,19 @@ static int nbd_genl_connect(struct sk_buff *skb, struct genl_info *info)
 	if (!netlink_capable(skb, CAP_SYS_ADMIN))
 		return -EPERM;
 
-	if (info->attrs[NBD_ATTR_INDEX])
+	if (info->attrs[NBD_ATTR_INDEX]) {
 		index = nla_get_u32(info->attrs[NBD_ATTR_INDEX]);
+
+		/*
+		 * Too big first_minor can cause duplicate creation of
+		 * sysfs files/links, since index << part_shift might overflow, or
+		 * MKDEV() expect that the max bits of first_minor is 20.
+		 */
+		if (index < 0 || index > MINORMASK >> part_shift) {
+			printk(KERN_ERR "nbd: illegal input index %d\n", index);
+			return -EINVAL;
+		}
+	}
 	if (!info->attrs[NBD_ATTR_SOCKETS]) {
 		printk(KERN_ERR "nbd: must specify at least one socket\n");
 		return -EINVAL;
-- 
2.31.1

Re: [PATCH -next] nbd: fix possible overflow on 'first_minor' in nbd_dev_add()

[zhangwensheng (E)]

friendly ping...

在 2022/3/10 17:32, Zhang Wensheng 写道:
> When 'index' is a big numbers, it may become negative which forced
> to 'int'. then 'index << part_shift' might overflow to a positive
> value that is not greater than '0xfffff', then sysfs might complains
> about duplicate creation. Because of this, move the 'index' judgment
> to the front will fix it and be better.
>
> Fixes: b0d9111a2d53 ("nbd: use an idr to keep track of nbd devices")
> Fixes: 940c264984fd ("nbd: fix possible overflow for 'first_minor' in nbd_dev_add()")
> Signed-off-by: Zhang Wensheng <zhangwensheng5@huawei.com>
> ---
>   drivers/block/nbd.c | 24 ++++++++++++------------
>   1 file changed, 12 insertions(+), 12 deletions(-)
>
> diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c
> index 5a1f98494ddd..b3cdfc0ffb98 100644
> --- a/drivers/block/nbd.c
> +++ b/drivers/block/nbd.c
> @@ -1800,17 +1800,6 @@ static struct nbd_device *nbd_dev_add(int index, unsigned int refs)
>   	refcount_set(&nbd->refs, 0);
>   	INIT_LIST_HEAD(&nbd->list);
>   	disk->major = NBD_MAJOR;
> -
> -	/* Too big first_minor can cause duplicate creation of
> -	 * sysfs files/links, since index << part_shift might overflow, or
> -	 * MKDEV() expect that the max bits of first_minor is 20.
> -	 */
> -	disk->first_minor = index << part_shift;
> -	if (disk->first_minor < index || disk->first_minor > MINORMASK) {
> -		err = -EINVAL;
> -		goto out_free_work;
> -	}
> -
>   	disk->minors = 1 << part_shift;
>   	disk->fops = &nbd_fops;
>   	disk->private_data = nbd;
> @@ -1915,8 +1904,19 @@ static int nbd_genl_connect(struct sk_buff *skb, struct genl_info *info)
>   	if (!netlink_capable(skb, CAP_SYS_ADMIN))
>   		return -EPERM;
>   
> -	if (info->attrs[NBD_ATTR_INDEX])
> +	if (info->attrs[NBD_ATTR_INDEX]) {
>   		index = nla_get_u32(info->attrs[NBD_ATTR_INDEX]);
> +
> +		/*
> +		 * Too big first_minor can cause duplicate creation of
> +		 * sysfs files/links, since index << part_shift might overflow, or
> +		 * MKDEV() expect that the max bits of first_minor is 20.
> +		 */
> +		if (index < 0 || index > MINORMASK >> part_shift) {
> +			printk(KERN_ERR "nbd: illegal input index %d\n", index);
> +			return -EINVAL;
> +		}
> +	}
>   	if (!info->attrs[NBD_ATTR_SOCKETS]) {
>   		printk(KERN_ERR "nbd: must specify at least one socket\n");
>   		return -EINVAL;

Re: [PATCH -next] nbd: fix possible overflow on 'first_minor' in nbd_dev_add()

[zhangwensheng (E)]

friendly ping...

在 2022/3/11 10:43, zhangwensheng (E) 写道:
> friendly ping...
>
> 在 2022/3/10 17:32, Zhang Wensheng 写道:
>> When 'index' is a big numbers, it may become negative which forced
>> to 'int'. then 'index << part_shift' might overflow to a positive
>> value that is not greater than '0xfffff', then sysfs might complains
>> about duplicate creation. Because of this, move the 'index' judgment
>> to the front will fix it and be better.
>>
>> Fixes: b0d9111a2d53 ("nbd: use an idr to keep track of nbd devices")
>> Fixes: 940c264984fd ("nbd: fix possible overflow for 'first_minor' in 
>> nbd_dev_add()")
>> Signed-off-by: Zhang Wensheng <zhangwensheng5@huawei.com>
>> ---
>>   drivers/block/nbd.c | 24 ++++++++++++------------
>>   1 file changed, 12 insertions(+), 12 deletions(-)
>>
>> diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c
>> index 5a1f98494ddd..b3cdfc0ffb98 100644
>> --- a/drivers/block/nbd.c
>> +++ b/drivers/block/nbd.c
>> @@ -1800,17 +1800,6 @@ static struct nbd_device *nbd_dev_add(int 
>> index, unsigned int refs)
>>       refcount_set(&nbd->refs, 0);
>>       INIT_LIST_HEAD(&nbd->list);
>>       disk->major = NBD_MAJOR;
>> -
>> -    /* Too big first_minor can cause duplicate creation of
>> -     * sysfs files/links, since index << part_shift might overflow, or
>> -     * MKDEV() expect that the max bits of first_minor is 20.
>> -     */
>> -    disk->first_minor = index << part_shift;
>> -    if (disk->first_minor < index || disk->first_minor > MINORMASK) {
>> -        err = -EINVAL;
>> -        goto out_free_work;
>> -    }
>> -
>>       disk->minors = 1 << part_shift;
>>       disk->fops = &nbd_fops;
>>       disk->private_data = nbd;
>> @@ -1915,8 +1904,19 @@ static int nbd_genl_connect(struct sk_buff 
>> *skb, struct genl_info *info)
>>       if (!netlink_capable(skb, CAP_SYS_ADMIN))
>>           return -EPERM;
>>   -    if (info->attrs[NBD_ATTR_INDEX])
>> +    if (info->attrs[NBD_ATTR_INDEX]) {
>>           index = nla_get_u32(info->attrs[NBD_ATTR_INDEX]);
>> +
>> +        /*
>> +         * Too big first_minor can cause duplicate creation of
>> +         * sysfs files/links, since index << part_shift might 
>> overflow, or
>> +         * MKDEV() expect that the max bits of first_minor is 20.
>> +         */
>> +        if (index < 0 || index > MINORMASK >> part_shift) {
>> +            printk(KERN_ERR "nbd: illegal input index %d\n", index);
>> +            return -EINVAL;
>> +        }
>> +    }
>>       if (!info->attrs[NBD_ATTR_SOCKETS]) {
>>           printk(KERN_ERR "nbd: must specify at least one socket\n");
>>           return -EINVAL;
> .

Re: [PATCH -next] nbd: fix possible overflow on 'first_minor' in nbd_dev_add()

[zhangwensheng (E)]

friendly ping...

在 2022/3/17 21:05, zhangwensheng (E) 写道:
> friendly ping...
>
> 在 2022/3/11 10:43, zhangwensheng (E) 写道:
>> friendly ping...
>>
>> 在 2022/3/10 17:32, Zhang Wensheng 写道:
>>> When 'index' is a big numbers, it may become negative which forced
>>> to 'int'. then 'index << part_shift' might overflow to a positive
>>> value that is not greater than '0xfffff', then sysfs might complains
>>> about duplicate creation. Because of this, move the 'index' judgment
>>> to the front will fix it and be better.
>>>
>>> Fixes: b0d9111a2d53 ("nbd: use an idr to keep track of nbd devices")
>>> Fixes: 940c264984fd ("nbd: fix possible overflow for 'first_minor' 
>>> in nbd_dev_add()")
>>> Signed-off-by: Zhang Wensheng <zhangwensheng5@huawei.com>
>>> ---
>>>   drivers/block/nbd.c | 24 ++++++++++++------------
>>>   1 file changed, 12 insertions(+), 12 deletions(-)
>>>
>>> diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c
>>> index 5a1f98494ddd..b3cdfc0ffb98 100644
>>> --- a/drivers/block/nbd.c
>>> +++ b/drivers/block/nbd.c
>>> @@ -1800,17 +1800,6 @@ static struct nbd_device *nbd_dev_add(int 
>>> index, unsigned int refs)
>>>       refcount_set(&nbd->refs, 0);
>>>       INIT_LIST_HEAD(&nbd->list);
>>>       disk->major = NBD_MAJOR;
>>> -
>>> -    /* Too big first_minor can cause duplicate creation of
>>> -     * sysfs files/links, since index << part_shift might overflow, or
>>> -     * MKDEV() expect that the max bits of first_minor is 20.
>>> -     */
>>> -    disk->first_minor = index << part_shift;
>>> -    if (disk->first_minor < index || disk->first_minor > MINORMASK) {
>>> -        err = -EINVAL;
>>> -        goto out_free_work;
>>> -    }
>>> -
>>>       disk->minors = 1 << part_shift;
>>>       disk->fops = &nbd_fops;
>>>       disk->private_data = nbd;
>>> @@ -1915,8 +1904,19 @@ static int nbd_genl_connect(struct sk_buff 
>>> *skb, struct genl_info *info)
>>>       if (!netlink_capable(skb, CAP_SYS_ADMIN))
>>>           return -EPERM;
>>>   -    if (info->attrs[NBD_ATTR_INDEX])
>>> +    if (info->attrs[NBD_ATTR_INDEX]) {
>>>           index = nla_get_u32(info->attrs[NBD_ATTR_INDEX]);
>>> +
>>> +        /*
>>> +         * Too big first_minor can cause duplicate creation of
>>> +         * sysfs files/links, since index << part_shift might 
>>> overflow, or
>>> +         * MKDEV() expect that the max bits of first_minor is 20.
>>> +         */
>>> +        if (index < 0 || index > MINORMASK >> part_shift) {
>>> +            printk(KERN_ERR "nbd: illegal input index %d\n", index);
>>> +            return -EINVAL;
>>> +        }
>>> +    }
>>>       if (!info->attrs[NBD_ATTR_SOCKETS]) {
>>>           printk(KERN_ERR "nbd: must specify at least one socket\n");
>>>           return -EINVAL;
>> .
> .

Re: [PATCH -next] nbd: fix possible overflow on 'first_minor' in nbd_dev_add()

[Josef Bacik]

On Thu, Mar 10, 2022 at 05:32:24PM +0800, Zhang Wensheng wrote:
> When 'index' is a big numbers, it may become negative which forced
> to 'int'. then 'index << part_shift' might overflow to a positive
> value that is not greater than '0xfffff', then sysfs might complains
> about duplicate creation. Because of this, move the 'index' judgment
> to the front will fix it and be better.
> 
> Fixes: b0d9111a2d53 ("nbd: use an idr to keep track of nbd devices")
> Fixes: 940c264984fd ("nbd: fix possible overflow for 'first_minor' in nbd_dev_add()")
> Signed-off-by: Zhang Wensheng <zhangwensheng5@huawei.com>

Reviewed-by: Josef Bacik <josef@toxicpanda.com>

Thanks,

Josef

Re: [PATCH -next] nbd: fix possible overflow on 'first_minor' in nbd_dev_add()

[Jens Axboe]

On Thu, 10 Mar 2022 17:32:24 +0800, Zhang Wensheng wrote:
> When 'index' is a big numbers, it may become negative which forced
> to 'int'. then 'index << part_shift' might overflow to a positive
> value that is not greater than '0xfffff', then sysfs might complains
> about duplicate creation. Because of this, move the 'index' judgment
> to the front will fix it and be better.
> 
> 
> [...]

Applied, thanks!

[1/1] nbd: fix possible overflow on 'first_minor' in nbd_dev_add()
      commit: 6d35d04a9e18990040e87d2bbf72689252669d54

Best regards,
-- 
Jens Axboe