[Pvfs2-developers] some testing questions

Sam Lang slang at mcs.anl.gov
Mon Apr 23 19:27:10 EDT 2007


On Apr 23, 2007, at 6:18 PM, David Brown wrote:

>> As I see it, if you need root for one step, then you might as well
>> use it for both.  The kernel module has to be loaded as root anyway,
>> so what does running the client as a non-root user buy you?
>
> Well technically you don't, with apps like dbus you can have modules
> auto-loaded by non-root applications on demand. Also with .service
> files dbus can auto-launch userspace apps as well, so the pvfs2-client
> can be shutdown when the pvfs2 file system isn't be accessed and
> started again when someone wants to access it. This could all be done
> with automount I would think... that'd be kinda cool :)

That does sound cool.

>
> This is of course a wish list sort of thing I'd understand if its not
> very high on the priority list ;).
>
>> I think its probably because its just easier to use mknod then figure
>> out how to use the syscall table in the kernel, but if you know of a
>> better way, we'd certainly consider any patches you send.  In
>> general, the pvfs model is keep the kernel module as lightweight as
>> possible, so anything that can go in the userspace client gets put
>> there.
>
> Yeah I'm not familiar with that aspect of the kernel either... but has
> their been any push to attempt to get the kernel part into the vanilla
> kernel?

Since the client has to be shipped separately, and the versions  
between kernel module and client have to match, I think it ends up  
being easier to just ship them together.  That being said, I've been  
smacked by the #ifdef checks in the kernel module in order to support  
all the different kernel versions still out there.  It can make the  
code hard to follow.  Someone with more history can answer this  
better though I'm sure.

-sam

>
> - David Brown
>



More information about the Pvfs2-developers mailing list