[PATCH] erofs-utils: mkfs: fix image reproducibility of `-E(all-)fragments`

[Hongbo Li]

On 2025/6/2 20:23, Gao Xiang wrote:
> Hi Hongbo,
> 
> On 2025/6/2 17:31, Hongbo Li wrote:
>> Hi, Xiang,
>>
>> On 2025/5/31 8:29, Gao Xiang wrote:
>>> The timestamp of the packed inode should be fixed to the build time.
>>>
>>> Fixes: 9fa9b017f773 ("erofs-utils: mkfs: support fragments")
>>> Signed-off-by: Gao Xiang <hsiangkao at linux.alibaba.com>
>>> ---
>>>   lib/inode.c | 14 ++++++++++----
>>>   1 file changed, 10 insertions(+), 4 deletions(-)
>>>
>>> diff --git a/lib/inode.c b/lib/inode.c
>>> index 7a10624..ca49a80 100644
>>> --- a/lib/inode.c
>>> +++ b/lib/inode.c
>>> @@ -910,7 +910,8 @@ out:
>>>       return 0;
>>>   }
>>> -static bool erofs_should_use_inode_extended(struct erofs_inode *inode)
>>> +static bool erofs_should_use_inode_extended(struct erofs_inode *inode,
>>> +                        const char *path)
>>>   {
>>>       if (cfg.c_force_inodeversion == FORCE_INODE_EXTENDED)
>>>           return true;
>>> @@ -924,7 +925,8 @@ static bool 
>>> erofs_should_use_inode_extended(struct erofs_inode *inode)
>>>           return true;
>>>       if (inode->i_nlink > USHRT_MAX)
>>>           return true;
>>> -    if ((inode->i_mtime != inode->sbi->build_time ||
>>> +    if (path != EROFS_PACKED_INODE &&
>>> +        (inode->i_mtime != inode->sbi->build_time ||
>>>            inode->i_mtime_nsec != inode->sbi->build_time_nsec) &&
>>>           !cfg.c_ignore_mtime)
>>>           return true;
>>> @@ -1016,6 +1018,10 @@ int __erofs_fill_inode(struct erofs_inode 
>>> *inode, struct stat *st,
>>>           erofs_err("gid overflow @ %s", path);
>>>       inode->i_gid += cfg.c_gid_offset;
>>> +    if (path == EROFS_PACKED_INODE) {
>>> +        inode->i_mtime = sbi->build_time;
>>> +        inode->i_mtime_nsec = 0;
>>> +    }
>>>       inode->i_mtime = st->st_mtime;
>>>       inode->i_mtime_nsec = ST_MTIM_NSEC(st);
>>
>> Should we put the condition in here? Because it will be reassigned if 
>> we do like this.
> 
> Oh, that is my fault, I will fix it soon.
> 
>>
>> And what about assigning sbi->build_time_nsec to inode->i_mtime_nsec 
>> like the FIXED case?
> 
> I just would like a quick fix for this because I have
> other features to work on.
> 
> If you want to improve that, could you submit a patch?
> 
Ok, I send the new one based on your v2.

Thanks,
Hongbo

> Thanks,
> Gao Xiang
> 
>>
>> Thanks,
>> Hongbo
>>