| Submitter | Alexander Graf |
|---|---|
| Date | Oct. 5, 2012, 2:36 a.m. |
| Message ID | <1349404606-24057-1-git-send-email-agraf@suse.de> |
| Download | mbox | patch |
| Permalink | /patch/189404/ |
| State | New |
| Headers | show |
Comments
On Fri, Oct 05, 2012 at 04:36:46AM +0200, Alexander Graf wrote: > In the sregs API, upper and lower 32bit segments of the BAT registers > are swapped when doing a set. Since we need to support old kernels out > there, don't bother to fix it in the kernel, but instead work around > the problem in QEMU by swapping on put. > > Signed-off-by: Alexander Graf <agraf@suse.de> We should probably at least add warning comments to the kernel headers.
On 05.10.2012, at 05:29, David Gibson <david@gibson.dropbear.id.au> wrote: > On Fri, Oct 05, 2012 at 04:36:46AM +0200, Alexander Graf wrote: >> In the sregs API, upper and lower 32bit segments of the BAT registers >> are swapped when doing a set. Since we need to support old kernels out >> there, don't bother to fix it in the kernel, but instead work around >> the problem in QEMU by swapping on put. >> >> Signed-off-by: Alexander Graf <agraf@suse.de> > > We should probably at least add warning comments to the kernel > headers. I would prefer to switch all of the sregs mess to one_reg and declare it deprecated then :). Alex > > -- > David Gibson | I'll have my music baroque, and my code > david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_ > | _way_ _around_! > http://www.ozlabs.org/~dgibson
On Fri, Oct 05, 2012 at 04:36:46AM +0200, Alexander Graf wrote: > In the sregs API, upper and lower 32bit segments of the BAT registers > are swapped when doing a set. Since we need to support old kernels out > there, don't bother to fix it in the kernel, but instead work around > the problem in QEMU by swapping on put. > > Signed-off-by: Alexander Graf <agraf@suse.de> > --- > target-ppc/kvm.c | 9 +++++---- > 1 files changed, 5 insertions(+), 4 deletions(-) > > diff --git a/target-ppc/kvm.c b/target-ppc/kvm.c > index 1975323..93c5bb7 100644 > --- a/target-ppc/kvm.c > +++ b/target-ppc/kvm.c > @@ -493,10 +493,11 @@ int kvm_arch_put_registers(CPUPPCState *env, int level) > > /* Sync BATs */ > for (i = 0; i < 8; i++) { > - sregs.u.s.ppc32.dbat[i] = ((uint64_t)env->DBAT[1][i] << 32) > - | env->DBAT[0][i]; > - sregs.u.s.ppc32.ibat[i] = ((uint64_t)env->IBAT[1][i] << 32) > - | env->IBAT[0][i]; > + /* Beware. We have to swap upper and lower bits here */ > + sregs.u.s.ppc32.dbat[i] = ((uint64_t)env->DBAT[0][i] << 32) > + | env->DBAT[1][i]; > + sregs.u.s.ppc32.ibat[i] = ((uint64_t)env->IBAT[0][i] << 32) > + | env->IBAT[1][i]; > } > > ret = kvm_vcpu_ioctl(env, KVM_SET_SREGS, &sregs); Applied (actually as part of the PPC pull).
Patch
diff --git a/target-ppc/kvm.c b/target-ppc/kvm.c index 1975323..93c5bb7 100644 --- a/target-ppc/kvm.c +++ b/target-ppc/kvm.c @@ -493,10 +493,11 @@ int kvm_arch_put_registers(CPUPPCState *env, int level) /* Sync BATs */ for (i = 0; i < 8; i++) { - sregs.u.s.ppc32.dbat[i] = ((uint64_t)env->DBAT[1][i] << 32) - | env->DBAT[0][i]; - sregs.u.s.ppc32.ibat[i] = ((uint64_t)env->IBAT[1][i] << 32) - | env->IBAT[0][i]; + /* Beware. We have to swap upper and lower bits here */ + sregs.u.s.ppc32.dbat[i] = ((uint64_t)env->DBAT[0][i] << 32) + | env->DBAT[1][i]; + sregs.u.s.ppc32.ibat[i] = ((uint64_t)env->IBAT[0][i] << 32) + | env->IBAT[1][i]; } ret = kvm_vcpu_ioctl(env, KVM_SET_SREGS, &sregs);
In the sregs API, upper and lower 32bit segments of the BAT registers are swapped when doing a set. Since we need to support old kernels out there, don't bother to fix it in the kernel, but instead work around the problem in QEMU by swapping on put. Signed-off-by: Alexander Graf <agraf@suse.de> --- target-ppc/kvm.c | 9 +++++---- 1 files changed, 5 insertions(+), 4 deletions(-)