[Ocfs2-tools-devel] [PATCH 1/2] fsck: supporting fixing inode alloc group desc

piaojun piaojun at huawei.com
Wed Jan 31 00:37:48 PST 2018


Eric,

On 2018/1/31 16:18, Eric Ren wrote:
> Jun,
> 
>> This won't cause crash, as '*pp = NULL' is just 'ptr = NULL'.
> 
> You're right! I overlooked it, sorry.
> 
>>> And, during iterating all files, if one file doesn't belong to any block group, it is a sign of problem, can we prompt
>>> a warning for user to make a decision?
>> We only iterate the gds in first level of gd chains, so probably there
> 
> This is a bad thing, but a thing we cannot think up better solution, so we
> have to compromise, right?
Yes, we could only try our best to solve this.

> 
> By iterating files, I think you assume all files that cannot travels are valid files.
> If there is one that doesn't fits into any bg, it indicates the unusual "bad thing" did
> happen, why not give a warning in that case?
What I concern is that too many files not in first level of chains may
cause too many warnings. And this may seem strange for users.

thanks,
Jun

> 
>> will be some files not belong to any of the gds. That is to say, we
>> can't distinguish them from the bad files. And I think this patch
>> should not be responsible for the bad files.
> Right, but my point is we'd better show the fact to user as we can.
Sure, it will be better to notice user about that.
> 
> Eric
> 
> 
> 
> .
> 



More information about the Ocfs2-tools-devel mailing list