linux-kernel - Re: [PATCH] cifs: set length when cifs_copy_pages_to

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:   Sun, 29 May 2022 05:52:33 -0700
From:   Tom Rix <trix@...hat.com>
To:     Steve French <smfrench@...il.com>,
        Nick Desaulniers <ndesaulniers@...gle.com>
Cc:     Steve French <stfrench@...rosoft.com>,
        Nathan Chancellor <nathan@...nel.org>,
        David Howells <dhowells@...hat.com>,
        CIFS <linux-cifs@...r.kernel.org>,
        samba-technical <samba-technical@...ts.samba.org>,
        LKML <linux-kernel@...r.kernel.org>, llvm@...ts.linux.dev,
        Steve French <sfrench@...ba.org>
Subject: Re: [PATCH] cifs: set length when cifs_copy_pages_to_iter is
 successful


On 5/28/22 4:31 PM, Steve French wrote:
> Presumably this was in Dave Howell's patch set which we took out of
> for-next to restructure in some of Al's feedback and some things found
> during testing. So nothing to fix in current mainline or for-next ...
> right?

Yup. Nothing to fix.

So drop this patch.

Tom

>
> On Sat, May 28, 2022 at 3:40 PM Nick Desaulniers
> <ndesaulniers@...gle.com> wrote:
>> + Steve's @microsoft.com email addr.
>>
>> On Thu, May 26, 2022 at 7:02 AM Tom Rix <trix@...hat.com> wrote:
>>> clang build fails with
>>> fs/cifs/smb2ops.c:4984:7: error: variable 'length' is used uninitialized whenever 'if' condition is false [-Werror,-Wsometimes-uninitialized]
>>>    if (rdata->result != 0) {
>>>        ^~~~~~~~~~~~~~~~~~
>>>
>>> handle_read_data() returns the number of bytes handled by setting the length variable.
>>> This only happens in the copy_to_iter() branch, it needs to also happen in the
>>> cifs_copy_pages_to_iter() branch.  When cifs_copy_pages_to_iter() is successful,
>>> its parameter data_len is how many bytes were handled, so set length to data_len.
>>>
>>> Fixes: 67fd8cff2b0f ("cifs: Change the I/O paths to use an iterator rather than a page list")
>>> Signed-off-by: Tom Rix <trix@...hat.com>
>>> ---
>>>   fs/cifs/smb2ops.c | 2 +-
>>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/fs/cifs/smb2ops.c b/fs/cifs/smb2ops.c
>>> index 3630e132781f..bfad482ec186 100644
>>> --- a/fs/cifs/smb2ops.c
>>> +++ b/fs/cifs/smb2ops.c
>>> @@ -4988,7 +4988,7 @@ handle_read_data(struct TCP_Server_Info *server, struct mid_q_entry *mid,
>>>                                  dequeue_mid(mid, rdata->result);
>>>                          return 0;
>>>                  }
>>> -               rdata->got_bytes = pages_len;
>>> +               length = rdata->got_bytes = pages_len;
>>>
>>>          } else if (buf_len >= data_offset + data_len) {
>>>                  /* read response payload is in buf */
>>> --
>>> 2.27.0
>>>
>>
>> --
>> Thanks,
>> ~Nick Desaulniers
>
>