KVM: s390: interrupt: fix virtual-physical confusion for next alert GISA
v2 cover KVM: s390: interrupt: fix virtual-physical confusion for next alert GISA
v2:
---
* improve commit message (thanks Janosch)
Nico Boehr (1):
KVM: s390: interrupt: fix virtual-physical confusion for next alert
GISA
arch/s390/kvm/interrupt.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
Signature · 1 line
2.39.1
PATCH v2 1/1 KVM: s390: interrupt: fix virtual-physical confusion for next alert GISA
The gisa next alert address is defined as a host absolute address so let's use virt_to_phys() to make sure we always write an absolute address to this hardware structure.
This is not a bug and currently works, because virtual and physical addresses are the same.
--- arch/s390/kvm/interrupt.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
arch/s390/kvm/interrupt.c +2 -2
@@ -305,7 +305,7 @@ static inline u8 gisa_get_ipm_or_restore_iam(struct kvm_s390_gisa_interrupt *gi)static inline int gisa_in_alert_list (struct kvm_s390_gisa *gisa) { -return READ_ONCE (gisa->next_alert) != (u32 )(u64 )gisa; +return READ_ONCE (gisa->next_alert) != (u32 )virt_to_phys (gisa); }static inline void gisa_set_ipm_gisc (struct kvm_s390_gisa *gisa,u32 gisc)
@@ -3167,7 +3167,7 @@ void kvm_s390_gisa_init(struct kvm *kvm)hrtimer_init (&gi->timer,CLOCK_MONOTONIC ,HRTIMER_MODE_REL ); gi->timer.function = gisa_vcpu_kicker;memset (gi->origin,0 ,sizeof (struct kvm_s390_gisa )); - gi->origin->next_alert = (u32 )(u64 )gi->origin; + gi->origin->next_alert = (u32 )virt_to_phys (gi->origin);VM_EVENT (kvm,3 ,"gisa 0x%pK initialized" , gi->origin); }
Signature · 1 line
2.39.1
Claudio Imbrenda
maintainerQuoted 18 lines · On Fri, 24 Feb 2023 15:09:08 +0100 Nico Boehr <[email protected]> wrote:
The gisa next alert address is defined as a host absolute address so let's use
virt_to_phys()to make sure we always write an absolute address to this hardware structure.This is not a bug and currently works, because virtual and physical addresses are the same.
Signed-off-by Nico Boehr <[email protected]>Reviewed-by Janosch Frank <[email protected]>--- arch/s390/kvm/interrupt.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)arch/s390/kvm/interrupt.c +0
-0
arch/s390/kvm/interrupt.c static inline u8 gisa_get_ipm_or_restore_iam(struct kvm_s390_gisa_interrupt *gi) · line 308
Quoted 5 lines
arch/s390/kvm/interrupt.c +1
-1@@ -305,7 +305,7 @@ static inline u8 gisa_get_ipm_or_restore_iam(struct kvm_s390_gisa_interrupt *gi)static inline int gisa_in_alert_list (struct kvm_s390_gisa *gisa) { -return READ_ONCE (gisa->next_alert) != (u32 )(u64 )gisa; +return READ_ONCE (gisa->next_alert) != (u32 )virt_to_phys (gisa);
is gisa always allocated below 4G? (I assume 2G actually) should we check if things are proper?
the cast to (u32) might hide bugs if gisa is above 4G (which it shouldn't be, obviously)
or do we not care?
arch/s390/kvm/interrupt.c void kvm_s390_gisa_init(struct kvm *kvm) · line 3170
Quoted 9 lines
arch/s390/kvm/interrupt.c +1
-1}static inline void gisa_set_ipm_gisc (struct kvm_s390_gisa *gisa,u32 gisc)@@ -3167,7 +3167,7 @@ void kvm_s390_gisa_init(struct kvm *kvm)hrtimer_init (&gi->timer,CLOCK_MONOTONIC ,HRTIMER_MODE_REL ); gi->timer.function = gisa_vcpu_kicker;memset (gi->origin,0 ,sizeof (struct kvm_s390_gisa )); - gi->origin->next_alert = (u32 )(u64 )gi->origin; + gi->origin->next_alert = (u32 )virt_to_phys (gi->origin);
same here
Quoted 3 lines
arch/s390/kvm/interrupt.c +0
-0VM_EVENT (kvm,3 ,"gisa 0x%pK initialized" , gi->origin); }
Quoting Claudio Imbrenda (2023-02-28 18:16:33)
Quoted 29 lines
Quoted 18 lines · On Fri, 24 Feb 2023 15:09:08 +0100 Nico Boehr <[email protected]> wrote:
The gisa next alert address is defined as a host absolute address so let's use
virt_to_phys()to make sure we always write an absolute address to this hardware structure.This is not a bug and currently works, because virtual and physical addresses are the same.
Signed-off-by Nico Boehr <[email protected]>Reviewed-by Janosch Frank <[email protected]>--- arch/s390/kvm/interrupt.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)arch/s390/kvm/interrupt.c +0
-0arch/s390/kvm/interrupt.c static inline u8 gisa_get_ipm_or_restore_iam(struct kvm_s390_gisa_interrupt *gi) · line 308
Quoted 5 lines
arch/s390/kvm/interrupt.c +1
-1@@ -305,7 +305,7 @@ static inline u8 gisa_get_ipm_or_restore_iam(struct kvm_s390_gisa_interrupt *gi)static inline int gisa_in_alert_list (struct kvm_s390_gisa *gisa) { -return READ_ONCE (gisa->next_alert) != (u32 )(u64 )gisa; +return READ_ONCE (gisa->next_alert) != (u32 )virt_to_phys (gisa);is gisa always allocated below 4G? (I assume 2G actually)
Quoted 5 lines
should we check if things are proper?
the cast to (u32) might hide bugs if gisa is above 4G (which it shouldn't be, obviously)or do we not care?
Yes, the gisa is always below 2 GB since it's part of the sie_page2.
I don't mind getting rid of the u32 cast really, but if it is allocated above 2 GB, it already is broken as it is and I didn't want to introduce unrelated changes. Also note that there is a few other places where we currently don't verify things really are below 2 GB, so you already need to be careful when allocating.
Also not sure if this is the right place to do this check, since we've already given the address to firmware/hardware anyways in the CHSC SGIB call, in the sie_block etc... so if we want to check this we should maybe look for a better place to check this...
Claudio Imbrenda
maintainerQuoted 34 lines · On Mon, 06 Mar 2023 11:50:32 +0100 Nico Boehr <[email protected]> wrote:
Quoting Claudio Imbrenda (2023-02-28 18:16:33)
Quoted 33 lines
Quoted 18 lines · On Fri, 24 Feb 2023 15:09:08 +0100 Nico Boehr <[email protected]> wrote:
The gisa next alert address is defined as a host absolute address so let's use
virt_to_phys()to make sure we always write an absolute address to this hardware structure.This is not a bug and currently works, because virtual and physical addresses are the same.
Signed-off-by Nico Boehr <[email protected]>Reviewed-by Janosch Frank <[email protected]>--- arch/s390/kvm/interrupt.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)arch/s390/kvm/interrupt.c +0
-0arch/s390/kvm/interrupt.c static inline u8 gisa_get_ipm_or_restore_iam(struct kvm_s390_gisa_interrupt *gi) · line 308
Quoted 5 lines
arch/s390/kvm/interrupt.c +1
-1@@ -305,7 +305,7 @@ static inline u8 gisa_get_ipm_or_restore_iam(struct kvm_s390_gisa_interrupt *gi)static inline int gisa_in_alert_list (struct kvm_s390_gisa *gisa) { -return READ_ONCE (gisa->next_alert) != (u32 )(u64 )gisa; +return READ_ONCE (gisa->next_alert) != (u32 )virt_to_phys (gisa);is gisa always allocated below 4G? (I assume 2G actually)
should we check if things are proper?
the cast to (u32) might hide bugs if gisa is above 4G (which it shouldn't be, obviously)
Quoted 7 lines
Quoted 1 line
or do we not care?
Yes, the gisa is always below 2 GB since it's part of the
sie_page2.I don't mind getting rid of the u32 cast really, but if it is allocated above 2 GB, it already is broken as it is and I didn't want to introduce unrelated changes. Also note that there is a few other places where we currently don't verify things really are below 2 GB, so you already need to be careful when allocating.
Also not sure if this is the right place to do this check, since we've already given the address to
firmware/hardwareanyways in the CHSC SGIB call, in thesie_blocketc... so if we want to check this we should maybe look for a better place to check this...
fair enough