Hello,

Here's the mqueue namespace patchset. 

If you're OK with it, I'll rebase on the next -mm which should contain
the new CONFIG_NAMESPACES and send to andrew.

Thanks !

C.

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

Cedric Le Goater <clg@fr.ibm.com> writes:

> Hello,
>
> Here's the mqueue namespace patchset. 
>
> If you're OK with it, I'll rebase on the next -mm which should contain
> the new CONFIG_NAMESPACES and send to andrew.

Well the CLONE flag is a conflict with the network namespace.

I was hoping someone had managed to make message queues work
through the filesystem API and then this would just be a matter
of multiple mounts of a filesystem.

However given the syscalls and the rest I guess this really does
count as a separate namespace :(

Eric
_______________________________________________
Containers mailing list
Containers@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers

Cedric Le Goater <clg@fr.ibm.com> writes:

> Hello,
>
> Here's the mqueue namespace patchset. 
>
> If you're OK with it, I'll rebase on the next -mm which should contain
> the new CONFIG_NAMESPACES and send to andrew.

And before I forget.

At least until someone has done an audit and we are certain we
have gotten all of the user space interfaces.

We need to Kconfig the whole thing so support creating one of these
things depends on CONFIG_EXPERIMENTAL.

Once we have the code merged and have audited the merge result
we can remove the Kconfig and the CONFIG_EXPERIMENTAL bit.

Eric
_______________________________________________
Containers mailing list
Containers@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers

Eric W. Biederman wrote:
> Cedric Le Goater <clg@fr.ibm.com> writes:
> 
>> Hello,
>>
>> Here's the mqueue namespace patchset. 
>>
>> If you're OK with it, I'll rebase on the next -mm which should contain
>> the new CONFIG_NAMESPACES and send to andrew.
> 
> Well the CLONE flag is a conflict with the network namespace.

Indeed. For the moment, this is for review because I'm waiting for a few 
namespace patches to go in -mm. When this time comes, I'll rebase.  

> I was hoping someone had managed to make message queues work
> through the filesystem API and then this would just be a matter
> of multiple mounts of a filesystem.

I've added a small helper routine to handle the case of a single
super_block per namespace. I called it get_sb_single_per_data()

It could be used in pid namespace also and probably a few other fs.

> However given the syscalls and the rest I guess this really does
> count as a separate namespace :(

yep ...

C.
_______________________________________________
Containers mailing list
Containers@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers

Eric W. Biederman wrote:
> Cedric Le Goater <clg@fr.ibm.com> writes:
> 
>> Hello,
>>
>> Here's the mqueue namespace patchset. 
>>
>> If you're OK with it, I'll rebase on the next -mm which should contain
>> the new CONFIG_NAMESPACES and send to andrew.
> 
> And before I forget.
> 
> At least until someone has done an audit and we are certain we
> have gotten all of the user space interfaces.
> 
> We need to Kconfig the whole thing so support creating one of these
> things depends on CONFIG_EXPERIMENTAL.
> 
> Once we have the code merged and have audited the merge result
> we can remove the Kconfig and the CONFIG_EXPERIMENTAL bit.

sure. I'm just waiting for pavel's new CONFIG_NAMESPACES* options 
to go in -mm. 

Thanks !

C.
_______________________________________________
Containers mailing list
Containers@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers