[Qemu-devel] Commit d8d95814609e89e5438a3318a647ec322fc4ff16 seems to break ioh3420

[Qemu-devel] Commit d8d95814609e89e5438a3318a647ec322fc4ff16 seems to break ioh3420

[Knut Omang]

I am struggling with the latest master and 

-device ioh3420,slot=0,id=pcie_port.0

as argument. I bisected it down to:

d8d95814609e89e5438a3318a647ec322fc4ff16 is the first bad commit
commit d8d95814609e89e5438a3318a647ec322fc4ff16
Author: Paolo Bonzini <pbonzini@redhat.com>
Date:   Wed Jun 11 12:42:01 2014 +0200

    memory: convert memory_region_destroy to object_unparent
    
    Explicitly call object_unparent in the few places where we
    will re-create the memory region.  If the memory region is
    simply being destroyed as part of device teardown, let QOM
    handle it.
    
    Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

:040000 040000 b38d10817e8297b9c1ef8827b3429362e47578f1 554905847501ec8e22669a6225cd979e42866905 M      docs
:040000 040000 a1a340ae1271b019975a5007673fef1490981a7b 57c10cdc7bb0ee0535fdde4c80edaef174270c62 M      hw
:100644 100644 dce94a313044e35bdd841a9499c2f218f848ee72 783a3ae6756ff4375267555607f46606a055b628 M      ioport.c
:100644 100644 64d7176193d829d9289c466bf4bd82b560b0c40a 23c5269b8ef077e3c2114df2cbd79ff9e206bf98 M      memory.c


With this commit, qemu crashes at boot if I supply the ioh3420 argument,
without it, it seems to work - I am not immediately able to pinpoint
where the problem is,...

Knut

Re: [Qemu-devel] Commit d8d95814609e89e5438a3318a647ec322fc4ff16 seems to break ioh3420

[Knut Omang]

Seems very similar to what Gonglei just reported...

Knut

On Wed, 2014-08-20 at 16:10 +0200, Knut Omang wrote:
> I am struggling with the latest master and 
> 
> -device ioh3420,slot=0,id=pcie_port.0
> 
> as argument. I bisected it down to:
> 
> d8d95814609e89e5438a3318a647ec322fc4ff16 is the first bad commit
> commit d8d95814609e89e5438a3318a647ec322fc4ff16
> Author: Paolo Bonzini <pbonzini@redhat.com>
> Date:   Wed Jun 11 12:42:01 2014 +0200
> 
>     memory: convert memory_region_destroy to object_unparent
>     
>     Explicitly call object_unparent in the few places where we
>     will re-create the memory region.  If the memory region is
>     simply being destroyed as part of device teardown, let QOM
>     handle it.
>     
>     Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> 
> :040000 040000 b38d10817e8297b9c1ef8827b3429362e47578f1 554905847501ec8e22669a6225cd979e42866905 M      docs
> :040000 040000 a1a340ae1271b019975a5007673fef1490981a7b 57c10cdc7bb0ee0535fdde4c80edaef174270c62 M      hw
> :100644 100644 dce94a313044e35bdd841a9499c2f218f848ee72 783a3ae6756ff4375267555607f46606a055b628 M      ioport.c
> :100644 100644 64d7176193d829d9289c466bf4bd82b560b0c40a 23c5269b8ef077e3c2114df2cbd79ff9e206bf98 M      memory.c
> 
> 
> With this commit, qemu crashes at boot if I supply the ioh3420 argument,
> without it, it seems to work - I am not immediately able to pinpoint
> where the problem is,...
> 
> Knut

Re: [Qemu-devel] Commit d8d95814609e89e5438a3318a647ec322fc4ff16 seems to break ioh3420

[Paolo Bonzini]

Il 20/08/2014 16:10, Knut Omang ha scritto:
> I am struggling with the latest master and 
> 
> -device ioh3420,slot=0,id=pcie_port.0
> 
> as argument. I bisected it down to:

My fault indeed.  I just sent a fix and re-reviewed my patch to look for
similar breakage (but didn't find any).

Paolo