[Ocfs2-tools-devel] [PATCH 14/14] ocfs2-tools: mount not showing fully resolved path

[Joseph Qi]

On 2016/4/13 16:30, Junxiao Bi wrote:
> On 04/13/2016 04:12 PM, Joseph Qi wrote:
>> Hi Junxiao,
>>
>> On 2016/4/1 11:10, Junxiao Bi wrote:
>>> From: Srinivas Eeda <srinivas.eeda at oracle.com>
>>>
>>> In /etc/mtab, device path is not fully resolved because fgets return code
>>> checked incorrectly. Device path is getting set to /dev/dm-X instead of
>>> /dev/mapper/xxx.
>>>
>>> Signed-off-by: Srinivas Eeda <srinivas.eeda at oracle.com>
>>> Reviewed-by: Junxiao Bi <junxiao.bi at oracle.com>
>>> ---
>>>  mount.ocfs2/sundries.c |    6 +++---
>>>  1 file changed, 3 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/mount.ocfs2/sundries.c b/mount.ocfs2/sundries.c
>>> index 2e1b95a24214..c4e9ccd86e23 100644
>>> --- a/mount.ocfs2/sundries.c
>>> +++ b/mount.ocfs2/sundries.c
>>> @@ -235,7 +235,7 @@ canonicalize_dm_name(const char *ptname)
>>>  	FILE	*f;
>>>  	size_t	sz;
>>>  	char	path[256], name[256], *res = NULL;
>>> -	int	err;
>>> +	char	*ret;
>>>  
>>>  	snprintf(path, sizeof(path), "/sys/block/%s/dm/name", ptname);
>>>  	f = fopen(path, "r");
>>> @@ -243,9 +243,9 @@ canonicalize_dm_name(const char *ptname)
>>>  		return NULL;
>>>  
>>>  	/* read "<name>\n" from sysfs */
>>> -	err = fgets(name, sizeof(name), f);
>>> +	ret = fgets(name, sizeof(name), f);
>>>  	sz = strlen(name);
>>> -	if (!err && sz > 1) {
>>> +	if (ret && sz > 1) {
>> Should it be "!ret && sz > 1"?
> No, fgets() return the buffer for success or null for failure.

You are right.

Thanks,
Joseph
> 
> Thanks,
> Junxiao.
>> I have also fixed this issue in my patchset "fix build warnings".
>> https://oss.oracle.com/pipermail/ocfs2-tools-devel/2016-March/004337.html
>> And now it is queued in my github repo and I am planning to push it to
>> Mark's repo this week.
>>
>> Thanks,
>> Joseph
>>>  		name[sz - 1] = '\0';
>>>  		snprintf(path, sizeof(path), "/dev/mapper/%s", name);
>>>  		res = strdup(path);
>>>
>>
>>
> 
> 
> .
>