Re: [Qemu-devel] [PATCH] check if we have space left for hotplugged memory

[Michael S. Tsirkin]

On Wed, Jun 25, 2014 at 05:33:31PM +0800, Hu Tao wrote:
> On Wed, Jun 25, 2014 at 12:17:57PM +0300, Michael S. Tsirkin wrote:
> > On Wed, Jun 25, 2014 at 05:01:04PM +0800, Hu Tao wrote:
> > > If pc-dimm is specified on qemu command line, but only with
> > > -m size (aka not -m size,maxmem,slots) then qemu will core dump.
> > > 
> > > This patch fixes the problem.
> > > 
> > > Signed-off-by: Hu Tao <address@hidden>
> > > ---
> > >  hw/i386/pc.c     | 5 +++++
> > >  hw/mem/pc-dimm.c | 5 +++++
> > >  2 files changed, 10 insertions(+)
> > > 
> > > diff --git a/hw/i386/pc.c b/hw/i386/pc.c
> > > index 2cf22b1..1a0d96e 100644
> > > --- a/hw/i386/pc.c
> > > +++ b/hw/i386/pc.c
> > > @@ -1566,6 +1566,11 @@ static void pc_dimm_plug(HotplugHandler 
> > > *hotplug_dev,
> > >          goto out;
> > >      }
> > >  
> > > +    if (memory_region_size(&pcms->hotplug_memory) == 0) {
> > > +        error_setg(&local_err, "no space left for hotplugged memory.");
> > > +        goto out;
> > > +    }
> > > +
> > >      addr = pc_dimm_get_free_addr(pcms->hotplug_memory_base,
> > >                                   
> > > memory_region_size(&pcms->hotplug_memory),
> > >                                   !addr ? NULL : &addr,
> > 
> > 
> > Why two errors in two places?
> 
> This is what Igor suggested(IIUC) which I think is reasonable. The check
> here prevents further operation even if there is no call to
> pc_dimm_get_free_addr.
> 
> > 
> > > diff --git a/hw/mem/pc-dimm.c b/hw/mem/pc-dimm.c
> > > index ad176b7..59ed9c0 100644
> > > --- a/hw/mem/pc-dimm.c
> > > +++ b/hw/mem/pc-dimm.c
> > > @@ -146,6 +146,11 @@ uint64_t pc_dimm_get_free_addr(uint64_t 
> > > address_space_start,
> > >      uint64_t new_addr, ret = 0;
> > >      uint64_t address_space_end = address_space_start + 
> > > address_space_size;
> > >  
> > > +    if (address_space_size == 0) {
> > > +        error_setg(errp, "no space left for hotplugged memory.");
> > > +        goto out;
> > > +    }
> > > +
> > >      assert(address_space_end > address_space_size);
> > >      object_child_foreach(qdev_get_machine(), pc_dimm_built_list, &list);
> > >  
> > 
> > 
> > Didn't we want a more specific message?
> 
> What about:
> 
>   no space left for hotplugged memory, please check maxmem option
> 
> ?

Maybe "please increase the maxmem value"?

> 
> Igor pointed out that explicitly state "-m maxmem=maxmem" here may not
> sync with future -m change.
> 
> 
> Hu