Re: [Qemu-devel] [Qemu-ppc] [PATCH] ppc: Yet another fix for the huge page support detection mechanism

[Greg Kurz]

On Mon, 18 Jul 2016 11:33:16 +0200
Thomas Huth <address@hidden> wrote:

> On 18.07.2016 11:26, Greg Kurz wrote:
> > On Mon, 18 Jul 2016 11:04:39 +0200
> > Thomas Huth <address@hidden> wrote:
> >   
> >> On 18.07.2016 10:59, Greg Kurz wrote:  
> >>> On Mon, 18 Jul 2016 10:52:36 +1000
> >>> David Gibson <address@hidden> wrote:
> >>>     
> >>>> On Fri, Jul 15, 2016 at 10:10:25AM +0200, Thomas Huth wrote:    
> >>>>> Commit 86b50f2e1bef ("Disable huge page support if it is not available
> >>>>> for main RAM") already made sure that huge page support is not announced
> >>>>> to the guest if the normal RAM of non-NUMA configurations is not backed
> >>>>> by a huge page filesystem. However, there is one more case that can go
> >>>>> wrong: NUMA is enabled, but the RAM of the NUMA nodes are not configured
> >>>>> with huge page support (and only the memory of a DIMM is configured with
> >>>>> it). When QEMU is started with the following command line for example,
> >>>>> the Linux guest currently crashes because it is trying to use huge pages
> >>>>> on a memory region that does not support huge pages:
> >>>>>
> >>>>>  qemu-system-ppc64 -enable-kvm ... -m 1G,slots=4,maxmem=32G -object \
> >>>>>    
> >>>>> memory-backend-file,policy=default,mem-path=/hugepages,size=1G,id=mem-mem1
> >>>>>  \
> >>>>>    -device pc-dimm,id=dimm-mem1,memdev=mem-mem1 -smp 2 \
> >>>>>    -numa node,nodeid=0 -numa node,nodeid=1
> >>>>>
> >>>>> To fix this issue, we've got to make sure to disable huge page support,
> >>>>> too, when there is a NUMA node that is not using a memory backend with
> >>>>> huge page support.
> >>>>>
> >>>>> Fixes: 86b50f2e1befc33407bdfeb6f45f7b0d2439a740
> >>>>> Signed-off-by: Thomas Huth <address@hidden>
> >>>>> ---
> >>>>>  target-ppc/kvm.c | 10 +++++++---
> >>>>>  1 file changed, 7 insertions(+), 3 deletions(-)      
> >>>>
> >>>> Applied to ppc-for-2.7, thanks.
> >>>>    
> >>>
> >>> It looks like my replies to this patch were ignored... no big deal though 
> >>> :)    
> >>
> >> I'll try to come up with an additional patch that fixes the remaining
> >> problem that you've found... Meanwhile, did you find out why you get
> >> that assertion that I was not able to recreate? Could you maybe post the
> >> exact command line to trigger that assertion?
> >>  
> > 
> > I hit the assertion when I specify pc-dimm devices on the command line:
> > 
> > qemu-system-ppc64 -enable-kvm ... -m 2G,slots=4,maxmem=32G \
> >  -object 
> > memory-backend-file,policy=default,mem-path=/mnt/kvm_hugepage,size=1G,id=mem-mem1
> >  \
> >  -device pc-dimm,id=dimm-mem1,memdev=mem-mem1 \
> >  -object 
> > memory-backend-file,policy=default,mem-path=/mnt,size=1G,id=mem-mem2 \
> >  -device pc-dimm,id=dimm-mem2,memdev=mem-mem2 \
> >  -smp 2 -numa node,nodeid=0,memdev=mem-mem1 \
> >  -numa node,nodeid=1,memdev=mem-mem2  
> 
> FWIW, with that command line, I still don't get an assertion but a
> normal error message:
> 
> qemu-system-ppc64: -device pc-dimm,id=dimm-mem1,memdev=mem-mem1: can't
> use already busy memdev: mem-mem1
> 
>  Thomas
> 

I hit the assertion with this exact command line:

qemu-system-ppc64 -machine pseries,accel=kvm \
-m 1G,slots=4,maxmem=32G \
-object 
memory-backend-file,policy=default,mem-path=/dev/hugepages,size=1G,id=mem-mem1 \
-device pc-dimm,id=dimm-mem1,memdev=mem-mem1 \
-numa node,nodeid=0,memdev=mem-mem1 \
-S

QEMU was built against David's ppc-for-2.7 branch (commit 159d2e39a).

But I get the very same error as you with QEMU 2.6... regression ?

And BTW, I'm not sure to understand why it is wrong to specify both pc-dimm and
numa pointing to the same memory backend.

Cheers.

--
Greg

pgpCw3xa2SWqM.pgp
Description: OpenPGP digital signature