| Message ID | 156925342310.974393.12235498904930019791.stgit@bahia.lan |
|---|---|
| State | Changes Requested |
| Headers | show |
| Series | KVM: PPC: Book3S: HV: XIVE: Allocate less VPs in OPAL | expand |
On 23/09/2019 17:43, Greg Kurz wrote: > If we cannot allocate the XIVE VPs in OPAL, the creation of a XIVE or > XICS-on-XIVE device is aborted as expected, but we leave kvm->arch.xive > set forever since the relase method isn't called in this case. Any release > subsequent tentative to create a XIVE or XICS-on-XIVE for this VM will > thus always fail. This is a problem for QEMU since it destroys and > re-creates these devices when the VM is reset: the VM would be > restricted to using the emulated XIVE or XICS forever. > > As an alternative to adding rollback, do not assign kvm->arch.xive before > making sure the XIVE VPs are allocated in OPAL. > > Fixes: 5422e95103cf ("KVM: PPC: Book3S HV: XIVE: Replace the 'destroy' method by a 'release' method") > Signed-off-by: Greg Kurz <groug@kaod.org> Reviewed-by: Cédric Le Goater <clg@kaod.org> C. > --- > arch/powerpc/kvm/book3s_xive.c | 11 +++++------ > arch/powerpc/kvm/book3s_xive_native.c | 2 +- > 2 files changed, 6 insertions(+), 7 deletions(-) > > diff --git a/arch/powerpc/kvm/book3s_xive.c b/arch/powerpc/kvm/book3s_xive.c > index cd2006bfcd3e..2ef43d037a4f 100644 > --- a/arch/powerpc/kvm/book3s_xive.c > +++ b/arch/powerpc/kvm/book3s_xive.c > @@ -2006,6 +2006,10 @@ static int kvmppc_xive_create(struct kvm_device *dev, u32 type) > > pr_devel("Creating xive for partition\n"); > > + /* Already there ? */ > + if (kvm->arch.xive) > + return -EEXIST; > + > xive = kvmppc_xive_get_device(kvm, type); > if (!xive) > return -ENOMEM; > @@ -2014,12 +2018,6 @@ static int kvmppc_xive_create(struct kvm_device *dev, u32 type) > xive->kvm = kvm; > mutex_init(&xive->lock); > > - /* Already there ? */ > - if (kvm->arch.xive) > - ret = -EEXIST; > - else > - kvm->arch.xive = xive; > - > /* We use the default queue size set by the host */ > xive->q_order = xive_native_default_eq_shift(); > if (xive->q_order < PAGE_SHIFT) > @@ -2040,6 +2038,7 @@ static int kvmppc_xive_create(struct kvm_device *dev, u32 type) > return ret; > > dev->private = xive; > + kvm->arch.xive = xive; > return 0; > } > > diff --git a/arch/powerpc/kvm/book3s_xive_native.c b/arch/powerpc/kvm/book3s_xive_native.c > index e9cbb42de424..84a354b90f60 100644 > --- a/arch/powerpc/kvm/book3s_xive_native.c > +++ b/arch/powerpc/kvm/book3s_xive_native.c > @@ -1087,7 +1087,6 @@ static int kvmppc_xive_native_create(struct kvm_device *dev, u32 type) > > xive->dev = dev; > xive->kvm = kvm; > - kvm->arch.xive = xive; > mutex_init(&xive->mapping_lock); > mutex_init(&xive->lock); > > @@ -1109,6 +1108,7 @@ static int kvmppc_xive_native_create(struct kvm_device *dev, u32 type) > return ret; > > dev->private = xive; > + kvm->arch.xive = xive; > return 0; > } > >
diff --git a/arch/powerpc/kvm/book3s_xive.c b/arch/powerpc/kvm/book3s_xive.c index cd2006bfcd3e..2ef43d037a4f 100644 --- a/arch/powerpc/kvm/book3s_xive.c +++ b/arch/powerpc/kvm/book3s_xive.c @@ -2006,6 +2006,10 @@ static int kvmppc_xive_create(struct kvm_device *dev, u32 type) pr_devel("Creating xive for partition\n"); + /* Already there ? */ + if (kvm->arch.xive) + return -EEXIST; + xive = kvmppc_xive_get_device(kvm, type); if (!xive) return -ENOMEM; @@ -2014,12 +2018,6 @@ static int kvmppc_xive_create(struct kvm_device *dev, u32 type) xive->kvm = kvm; mutex_init(&xive->lock); - /* Already there ? */ - if (kvm->arch.xive) - ret = -EEXIST; - else - kvm->arch.xive = xive; - /* We use the default queue size set by the host */ xive->q_order = xive_native_default_eq_shift(); if (xive->q_order < PAGE_SHIFT) @@ -2040,6 +2038,7 @@ static int kvmppc_xive_create(struct kvm_device *dev, u32 type) return ret; dev->private = xive; + kvm->arch.xive = xive; return 0; } diff --git a/arch/powerpc/kvm/book3s_xive_native.c b/arch/powerpc/kvm/book3s_xive_native.c index e9cbb42de424..84a354b90f60 100644 --- a/arch/powerpc/kvm/book3s_xive_native.c +++ b/arch/powerpc/kvm/book3s_xive_native.c @@ -1087,7 +1087,6 @@ static int kvmppc_xive_native_create(struct kvm_device *dev, u32 type) xive->dev = dev; xive->kvm = kvm; - kvm->arch.xive = xive; mutex_init(&xive->mapping_lock); mutex_init(&xive->lock); @@ -1109,6 +1108,7 @@ static int kvmppc_xive_native_create(struct kvm_device *dev, u32 type) return ret; dev->private = xive; + kvm->arch.xive = xive; return 0; }
If we cannot allocate the XIVE VPs in OPAL, the creation of a XIVE or XICS-on-XIVE device is aborted as expected, but we leave kvm->arch.xive set forever since the relase method isn't called in this case. Any subsequent tentative to create a XIVE or XICS-on-XIVE for this VM will thus always fail. This is a problem for QEMU since it destroys and re-creates these devices when the VM is reset: the VM would be restricted to using the emulated XIVE or XICS forever. As an alternative to adding rollback, do not assign kvm->arch.xive before making sure the XIVE VPs are allocated in OPAL. Fixes: 5422e95103cf ("KVM: PPC: Book3S HV: XIVE: Replace the 'destroy' method by a 'release' method") Signed-off-by: Greg Kurz <groug@kaod.org> --- arch/powerpc/kvm/book3s_xive.c | 11 +++++------ arch/powerpc/kvm/book3s_xive_native.c | 2 +- 2 files changed, 6 insertions(+), 7 deletions(-)