linux-kernel - Re: [PATCH v1] qnx4_match: do not over run the buffer

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Sat, 21 Nov 2020 20:27:37 -0500
From:   Tong Zhang <ztong0001@...il.com>
To:     Anders Larsen <al@...rsen.net>
Cc:     linux-kernel@...r.kernel.org
Subject: Re: [PATCH v1] qnx4_match: do not over run the buffer

Thanks for the clarification! This sounds good to me.
I will send a revised patch.
Best,
- Tong

> On Nov 21, 2020, at 4:57 PM, Anders Larsen <al@...rsen.net> wrote:
> 
> On Saturday, 2020-11-21 22:47 Tong Zhang wrote:
>> 
>>> On Nov 21, 2020, at 4:40 PM, Anders Larsen <al@...rsen.net> wrote:
>>> 
>>> On Friday, 2020-11-20 22:21 Tong Zhang wrote:
>>>> the di_fname may not terminated by '\0', use strnlen to prevent buffer
>>>> overrun
>>>> 
>>>> ---
>>>> fs/qnx4/namei.c | 2 +-
>>>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>>> 
>>>> diff --git a/fs/qnx4/namei.c b/fs/qnx4/namei.c
>>>> index 8d72221735d7..c0e79094f578 100644
>>>> --- a/fs/qnx4/namei.c
>>>> +++ b/fs/qnx4/namei.c
>>>> @@ -40,7 +40,7 @@ static int qnx4_match(int len, const char *name,
>>>> 	} else {
>>>> 		namelen = QNX4_SHORT_NAME_MAX;
>>>> 	}
>>>> -	thislen = strlen( de->di_fname );
>>>> +	thislen = strnlen( de->di_fname, QNX4_SHORT_NAME_MAX );
>>> 
>>> that should be
>>> +	thislen = strnlen( de->di_fname, namelen );
>>> otherwise the length of a filename would always be limited to QNX4_SHORT_NAME_MAX (16) characters.
>>> 
>> Why should we put something bigger here if the size of qnx4_inode_entry->di_fname is QNX4_SHORT_NAME_MAX.
>> Won’t that be a problem?
> 
> If QNX4_FILE_LINK is set in de->di_status (see line 38), 'de' actually points to a struct qnx4_link_info which can hold a longer name.
> That's the reason for the namelen massage.
> (Please don't ask why it is not a union)
> 
> Cheers
> Anders
> 
>