[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [Patch 1 of 1] [xen-4.1.2] libxl: Seal tapdisk minor leak.



On Aug 7,  4:45pm, Ian Jackson wrote:
} Subject: Re: [Xen-devel] [Patch 1 of 1] [xen-4.1.2] libxl: Seal tapdisk mi

Hi, hope the day is going well for everyone.

> Ian Campbell writes ("Re: [Xen-devel] [Patch 1 of 1] [xen-4.1.2] libxl: Seal 
> tapdisk minor leak."):
> > On Tue, 2012-08-07 at 16:14 +0100, Ian Jackson wrote:
> > > Dr. Greg Wettstein writes ("[Patch 1 of 1] [xen-4.1.2] libxl: Seal 
> > > tapdisk minor leak."):
> > > > libxl: Seal tapdisk minor leak.
> > > ...
> > > > To implement correct cleanup of blktap devices in Xen 4.1.2.
> > > 
> > > Is this patch is supposed to be against xen-4.1-testing.hg ?
> > > 
> > > > diff -r b2b7a7a49af5 tools/libxl/libxl_blktap2.c
> > > > --- a/tools/libxl/libxl_blktap2.c       Sat Aug 04 16:17:08 2012 -0500
> > > > +++ b/tools/libxl/libxl_blktap2.c       Sun Aug 05 09:22:35 2012 -0500
> > > > @@ -59,6 +59,7 @@ void libxl__device_destroy_tapdisk(libxl
> > > 
> > > This function doesn't seem to exist in my copy of xen-4.1-testing.hg
> > > tip nor in 23172:3eca5bf65e6c.
> > 
> > It comes from <201208051440.q75Een7F008501@xxxxxxxxxxxxxxxxx> which is a
> > backport request that introduces this function.
> 
> Oh I see, these are two interdependent patches.  Let me read them
> again.

Sorry for the confusion, I should have explicitly called out the
inter-dependency.

The first patch is a backport of one which Ian did for xen-unstable,
corrected primarily, for the change in calling convention for the
context arguement to libxl__device_destroy_tapdisk().

By itself this patch corrects the problem with the tapdisk2 userspace
process not being kill when the guest exists.

The second builds on top of that patch and fixes an ordering problem
which causes a deadlock between xen-blkback and blktap2 when the user
space process attempts to unmap its ring buffer.  The major effect is
orphaning of the tapdisk minor.

The more subtle effect is the deadlock which stalls shutdown of the
guest and actually livelocks the kernel for the duration of the
timeout on the select() call used for the IPC socket communications.

> Ian.

Have a good day.

}-- End of excerpt from Ian Jackson

As always,
Dr. G.W. Wettstein, Ph.D.   Enjellic Systems Development, LLC.
4206 N. 19th Ave.           Specializing in information infra-structure
Fargo, ND  58102            development.
PH: 701-281-1686
FAX: 701-281-3949           EMAIL: greg@xxxxxxxxxxxx
------------------------------------------------------------------------------
"Some of them are.  A surprising number aren't.  A personal favorite of
 mine was the log from a cracker who couldn't figure out how to untar
 and install the trojan package he'd ftped onto the machine.  He tried a
 few times, and then eventually gave up and logged out."
                                -- Nat Lanza

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.