linux-kernel - Re: [PATCH] erofs: [PATCH v2] Prevent entering an infinite loop when i is 0

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]

Message-ID: <76c8e7c1-82c0-46bd-9e46-96c8eaac83b1@linux.alibaba.com>
Date: Thu, 29 Aug 2024 19:15:42 +0800
From: Gao Xiang <hsiangkao@...ux.alibaba.com>
To: liujinbao1 <jinbaoliu365@...il.com>
Cc: chao@...nel.org, linux-erofs@...ts.ozlabs.org,
 linux-kernel@...r.kernel.org, mazhenhua@...omi.com,
 liujinbao1 <liujinbao1@...omi.com>, Gao Xiang <xiang@...nel.org>,
 Hongzhen Luo <hongzhen@...ux.alibaba.com>
Subject: Re: [PATCH] erofs: [PATCH v2] Prevent entering an infinite loop when
 i is 0

Hi liujinbao1,

Since it's a trivial bug which can be simply fixed, I just
would like to confirm if you still have time and interest to
send a workable patch (after a week).

If you don't have time to follow up, I will ask Hongzhen to
fix this with your "Reported-by:" tomorrow.

Thanks,
Gao Xiang

On 2024/8/23 20:03, Gao Xiang wrote:
> 
> 
> On 2024/8/23 11:05, liujinbao1 wrote:
>> From: liujinbao1 <liujinbao1@...omi.com>
>>
>> When i=0 and err is not equal to 0,
>> the while(-1) loop will enter into an
>> infinite loop. This patch avoids this issue.
>>
>> Signed-off-by: liujinbao1 <liujinbao1@...omi.com>
>> ---
>>   fs/erofs/decompressor.c | 8 +++-----
>>   1 file changed, 3 insertions(+), 5 deletions(-)
>>
>>> Hi,
> 
> The patch is corrupted and the patch subject
> line is also broken.
> 
> I think it should be:
> [PATCH v2] erofs: prevent entering an infinite loop when i is 0
> 
>>>
>>> On 2024/8/22 14:27, liujinbao1 wrote:
>>>> From: liujinbao1 <liujinbao1@...omi.com>
>>>>
>>>> When i=0 and err is not equal to 0,
>>>> the while(-1) loop will enter into an
>>>> infinite loop. This patch avoids this issue.
>>>
>>> Missing your Signed-off-by here.
>>>
>>>> ---
>>>>   fs/erofs/decompressor.c | 2 ++
>>>>   1 file changed, 2 insertions(+)
>>>>
>>>> diff --git a/fs/erofs/decompressor.c b/fs/erofs/decompressor.c index
>>>> c2253b6a5416..1b2b8cc7911c 100644
>>>> --- a/fs/erofs/decompressor.c
>>>> +++ b/fs/erofs/decompressor.c
>>>> @@ -539,6 +539,8 @@ int __init z_erofs_init_decompressor(void)
>>>>       for (i = 0; i < Z_EROFS_COMPRESSION_MAX; ++i) {
>>>>               err = z_erofs_decomp[i] ? z_erofs_decomp[i]->init() : 0;
>>>>               if (err) {
>>>> +                    if (!i)
>>>> +                            return err;
>>>>                       while (--i)
>>>>                               if (z_erofs_decomp[i])
>>>>                                       z_erofs_decomp[i]->exit();
>>>
>>>
>>> Thanks for catching this, how about the following diff (space-demaged).
>>>
>>> If it looks good to you, could you please send another version?
>>
>>> diff --git a/fs/erofs/decompressor.c b/fs/erofs/decompressor.c index c2253b6a5416..c9b2bc1309d2 100644
>>> --- a/fs/erofs/decompressor.c
>>> +++ b/fs/erofs/decompressor.c
>>> @@ -534,18 +534,16 @@ int z_erofs_parse_cfgs(struct super_block *sb, struct erofs_super_block *dsb)
>>>
>>> int __init z_erofs_init_decompressor(void)
>>> {
>>> -      int i, err;
>>> +      int i, err = 0;
>>>
>>>         for (i = 0; i < Z_EROFS_COMPRESSION_MAX; ++i) {
>>>                 err = z_erofs_decomp[i] ? z_erofs_decomp[i]->init() : 0;
>>> -              if (err) {
>>> +              if (err && i)
>>>                         while (--i)
>>>                                 if (z_erofs_decomp[i])
>>>                                         z_erofs_decomp[i]->exit();
>>> -                      return err;
>> +                        break;
>>> -              }
>>>         }
>>> -      return 0;
>>> +      return err;
>>> }
>>>
>> missing break?
> 
> Why needing a break?
> 
> Thanks,
> Gao Xiang