| Message ID | 20240817101541.1664-1-yuzenghui@huawei.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | KVM: arm64: vgic-debug: Don't put unmarked LPIs | expand |
On Sat, 17 Aug 2024 11:15:41 +0100, Zenghui Yu <yuzenghui@huawei.com> wrote: > > If there were LPIs being mapped behind our back (i.e., between .start() and > .stop()), we would put them at iter_unmark_lpis() without checking if they > were actually *marked*, which is obviously not good. > > Switch to use the xa_for_each_marked() iterator to fix it. > > Fixes: 85d3ccc8b75b ("KVM: arm64: vgic-debug: Use an xarray mark for debug iterator") > Signed-off-by: Zenghui Yu <yuzenghui@huawei.com> > --- > arch/arm64/kvm/vgic/vgic-debug.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/arch/arm64/kvm/vgic/vgic-debug.c b/arch/arm64/kvm/vgic/vgic-debug.c > index bc74d06398ef..e1397ab2072a 100644 > --- a/arch/arm64/kvm/vgic/vgic-debug.c > +++ b/arch/arm64/kvm/vgic/vgic-debug.c > @@ -85,7 +85,7 @@ static void iter_unmark_lpis(struct kvm *kvm) > struct vgic_irq *irq; > unsigned long intid; > > - xa_for_each(&dist->lpi_xa, intid, irq) { > + xa_for_each_marked(&dist->lpi_xa, intid, irq, LPI_XA_MARK_DEBUG_ITER) { > xa_clear_mark(&dist->lpi_xa, intid, LPI_XA_MARK_DEBUG_ITER); > vgic_put_irq(kvm, irq); > } Ouch. Nicely caught. I think this deserves a Cc: stable@vger.kernel.org # v6.10 With that, Reviewed-by: Marc Zyngier <maz@kernel.org> Thanks, M.
On 2024/8/17 18:25, Marc Zyngier wrote: > On Sat, 17 Aug 2024 11:15:41 +0100, > Zenghui Yu <yuzenghui@huawei.com> wrote: > > > > If there were LPIs being mapped behind our back (i.e., between .start() and > > .stop()), we would put them at iter_unmark_lpis() without checking if they > > were actually *marked*, which is obviously not good. > > > > Switch to use the xa_for_each_marked() iterator to fix it. > > > > Fixes: 85d3ccc8b75b ("KVM: arm64: vgic-debug: Use an xarray mark for debug iterator") > > Signed-off-by: Zenghui Yu <yuzenghui@huawei.com> > > --- > > arch/arm64/kvm/vgic/vgic-debug.c | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/arch/arm64/kvm/vgic/vgic-debug.c b/arch/arm64/kvm/vgic/vgic-debug.c > > index bc74d06398ef..e1397ab2072a 100644 > > --- a/arch/arm64/kvm/vgic/vgic-debug.c > > +++ b/arch/arm64/kvm/vgic/vgic-debug.c > > @@ -85,7 +85,7 @@ static void iter_unmark_lpis(struct kvm *kvm) > > struct vgic_irq *irq; > > unsigned long intid; > > > > - xa_for_each(&dist->lpi_xa, intid, irq) { > > + xa_for_each_marked(&dist->lpi_xa, intid, irq, LPI_XA_MARK_DEBUG_ITER) { > > xa_clear_mark(&dist->lpi_xa, intid, LPI_XA_MARK_DEBUG_ITER); > > vgic_put_irq(kvm, irq); > > } > > Ouch. Nicely caught. I think this deserves a > > Cc: stable@vger.kernel.org # v6.10 Yup. I guess Oliver will help to add it when applying ;-) > > With that, > > Reviewed-by: Marc Zyngier <maz@kernel.org> Thanks! Zenghui
On Sat, Aug 17, 2024 at 06:31:45PM +0800, Zenghui Yu wrote: > On 2024/8/17 18:25, Marc Zyngier wrote: > > On Sat, 17 Aug 2024 11:15:41 +0100, > > Zenghui Yu <yuzenghui@huawei.com> wrote: > > > > > > If there were LPIs being mapped behind our back (i.e., between .start() and > > > .stop()), we would put them at iter_unmark_lpis() without checking if they > > > were actually *marked*, which is obviously not good. > > > > > > Switch to use the xa_for_each_marked() iterator to fix it. Urgh, that's what I'd meant to do. Thanks for catching this Zenghui. > > > Fixes: 85d3ccc8b75b ("KVM: arm64: vgic-debug: Use an xarray mark for debug iterator") > > > Signed-off-by: Zenghui Yu <yuzenghui@huawei.com> > > > --- > > > arch/arm64/kvm/vgic/vgic-debug.c | 2 +- > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > diff --git a/arch/arm64/kvm/vgic/vgic-debug.c b/arch/arm64/kvm/vgic/vgic-debug.c > > > index bc74d06398ef..e1397ab2072a 100644 > > > --- a/arch/arm64/kvm/vgic/vgic-debug.c > > > +++ b/arch/arm64/kvm/vgic/vgic-debug.c > > > @@ -85,7 +85,7 @@ static void iter_unmark_lpis(struct kvm *kvm) > > > struct vgic_irq *irq; > > > unsigned long intid; > > > > > > - xa_for_each(&dist->lpi_xa, intid, irq) { > > > + xa_for_each_marked(&dist->lpi_xa, intid, irq, LPI_XA_MARK_DEBUG_ITER) { > > > xa_clear_mark(&dist->lpi_xa, intid, LPI_XA_MARK_DEBUG_ITER); > > > vgic_put_irq(kvm, irq); > > > } > > > > Ouch. Nicely caught. I think this deserves a > > > > Cc: stable@vger.kernel.org # v6.10 > > Yup. I guess Oliver will help to add it when applying ;-) Indeed, I'll grab it here in a moment. > > > > With that, > > > > Reviewed-by: Marc Zyngier <maz@kernel.org> > Thanks both!
On Sat, 17 Aug 2024 18:15:41 +0800, Zenghui Yu wrote: > If there were LPIs being mapped behind our back (i.e., between .start() and > .stop()), we would put them at iter_unmark_lpis() without checking if they > were actually *marked*, which is obviously not good. > > Switch to use the xa_for_each_marked() iterator to fix it. > > > [...] Applied to kvmarm/fixes, thanks! [1/1] KVM: arm64: vgic-debug: Don't put unmarked LPIs https://git.kernel.org/kvmarm/kvmarm/c/2240a50e6294 -- Best, Oliver
diff --git a/arch/arm64/kvm/vgic/vgic-debug.c b/arch/arm64/kvm/vgic/vgic-debug.c index bc74d06398ef..e1397ab2072a 100644 --- a/arch/arm64/kvm/vgic/vgic-debug.c +++ b/arch/arm64/kvm/vgic/vgic-debug.c @@ -85,7 +85,7 @@ static void iter_unmark_lpis(struct kvm *kvm) struct vgic_irq *irq; unsigned long intid; - xa_for_each(&dist->lpi_xa, intid, irq) { + xa_for_each_marked(&dist->lpi_xa, intid, irq, LPI_XA_MARK_DEBUG_ITER) { xa_clear_mark(&dist->lpi_xa, intid, LPI_XA_MARK_DEBUG_ITER); vgic_put_irq(kvm, irq); }
If there were LPIs being mapped behind our back (i.e., between .start() and .stop()), we would put them at iter_unmark_lpis() without checking if they were actually *marked*, which is obviously not good. Switch to use the xa_for_each_marked() iterator to fix it. Fixes: 85d3ccc8b75b ("KVM: arm64: vgic-debug: Use an xarray mark for debug iterator") Signed-off-by: Zenghui Yu <yuzenghui@huawei.com> --- arch/arm64/kvm/vgic/vgic-debug.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)