[Devel] Re: [RFC v14-rc3][PATCH 36/36] Stub implementation of IPC namespace c/r

[Oren Laadan]

Serge E. Hallyn wrote:
> Quoting Oren Laadan (orenl at cs.columbia.edu):
>> From: Dan Smith <danms at us.ibm.com>
>>
>> Changes:
>>  - Update to match UTS changes
>>
>> Signed-off-by: Dan Smith <danms at us.ibm.com>
>> Signed-off-by: Oren Laadan <orenl at cs.columbia.edu>
> 
> Acked-by: Serge Hallyn <serue at us.ibm.com>
> 
> However...
> 
>> +	if (!!ipc_ns ^ !(flags & CLONE_NEWIPC))
>> +		return -EINVAL;
> 
> Every time I see this I have to think about whether it is right or not.
> I'm not sure whether it's worth commenting (at each such meme) that
> CLONE_NEWIPC only needed to be set the first time we ran across that
> ipcns, or whether it's indicative that there is a simpler way the code
> could be done.  But if it just took me a twice-over to see that it's
> right, when I'd already confirmed that with the CLONE_NEWUTS version
> last week, then a fresh reviewer will be cursing your name...

This is a consistency check: ensure that the state of the objhash as
reflected in the value of ipc_ns is consistent with the flags.

I'll add a comment.

Oren.

> 
> -serge
> 
_______________________________________________
Containers mailing list
Containers at lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers