* [GIT PULL 1/9] KVM: s390: log runtime instrumentation enablement
2017-04-07 10:47 [GIT PULL 0/9] KVM: s390: features for 4.12 (via kvm/next) Christian Borntraeger
@ 2017-04-07 10:47 ` Christian Borntraeger
2017-04-07 10:47 ` [GIT PULL 2/9] KVM: s390: Handle sthyi also for instruction intercept Christian Borntraeger
` (7 subsequent siblings)
8 siblings, 0 replies; 17+ messages in thread
From: Christian Borntraeger @ 2017-04-07 10:47 UTC (permalink / raw)
To: Paolo Bonzini, Radim Krčmář
Cc: KVM, Christian Borntraeger, Cornelia Huck, linux-s390
We handle runtime instrumentation enablement either lazy or via
sync_regs on migration. Make sure to add a debug log entry for that
per CPU on the first occurrence.
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
---
arch/s390/kvm/kvm-s390.c | 14 ++++++++------
arch/s390/kvm/priv.c | 1 +
2 files changed, 9 insertions(+), 6 deletions(-)
diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
index fd6cd05b..5fda468 100644
--- a/arch/s390/kvm/kvm-s390.c
+++ b/arch/s390/kvm/kvm-s390.c
@@ -2719,6 +2719,9 @@ static int __vcpu_run(struct kvm_vcpu *vcpu)
static void sync_regs(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
+ struct runtime_instr_cb *riccb;
+
+ riccb = (struct runtime_instr_cb *) &kvm_run->s.regs.riccb;
vcpu->arch.sie_block->gpsw.mask = kvm_run->psw_mask;
vcpu->arch.sie_block->gpsw.addr = kvm_run->psw_addr;
if (kvm_run->kvm_dirty_regs & KVM_SYNC_PREFIX)
@@ -2747,12 +2750,11 @@ static void sync_regs(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
* we should enable RI here instead of doing the lazy enablement.
*/
if ((kvm_run->kvm_dirty_regs & KVM_SYNC_RICCB) &&
- test_kvm_facility(vcpu->kvm, 64)) {
- struct runtime_instr_cb *riccb =
- (struct runtime_instr_cb *) &kvm_run->s.regs.riccb;
-
- if (riccb->valid)
- vcpu->arch.sie_block->ecb3 |= 0x01;
+ test_kvm_facility(vcpu->kvm, 64) &&
+ riccb->valid &&
+ !(vcpu->arch.sie_block->ecb3 & 0x01)) {
+ VCPU_EVENT(vcpu, 3, "%s", "ENABLE: RI (sync_regs)");
+ vcpu->arch.sie_block->ecb3 |= 0x01;
}
save_access_regs(vcpu->arch.host_acrs);
restore_access_regs(vcpu->run->s.regs.acrs);
diff --git a/arch/s390/kvm/priv.c b/arch/s390/kvm/priv.c
index 64b6a30..93d6cde 100644
--- a/arch/s390/kvm/priv.c
+++ b/arch/s390/kvm/priv.c
@@ -37,6 +37,7 @@
static int handle_ri(struct kvm_vcpu *vcpu)
{
if (test_kvm_facility(vcpu->kvm, 64)) {
+ VCPU_EVENT(vcpu, 3, "%s", "ENABLE: RI (lazy)");
vcpu->arch.sie_block->ecb3 |= 0x01;
kvm_s390_retry_instr(vcpu);
return 0;
--
2.7.4
^ permalink raw reply related [flat|nested] 17+ messages in thread* [GIT PULL 2/9] KVM: s390: Handle sthyi also for instruction intercept
2017-04-07 10:47 [GIT PULL 0/9] KVM: s390: features for 4.12 (via kvm/next) Christian Borntraeger
2017-04-07 10:47 ` [GIT PULL 1/9] KVM: s390: log runtime instrumentation enablement Christian Borntraeger
@ 2017-04-07 10:47 ` Christian Borntraeger
2017-04-07 10:47 ` [GIT PULL 3/9] KVM: s390: use defines for execution controls Christian Borntraeger
` (6 subsequent siblings)
8 siblings, 0 replies; 17+ messages in thread
From: Christian Borntraeger @ 2017-04-07 10:47 UTC (permalink / raw)
To: Paolo Bonzini, Radim Krčmář
Cc: KVM, Christian Borntraeger, Cornelia Huck, linux-s390
Right now we handle the STHYI only via the operation exception intercept
(illegal instruction). If hardware ever decides to provide an
instruction intercept for STHYI, we should handle that as well.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
---
arch/s390/kvm/intercept.c | 3 +--
arch/s390/kvm/priv.c | 1 +
arch/s390/kvm/sthyi.c | 3 +++
3 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/arch/s390/kvm/intercept.c b/arch/s390/kvm/intercept.c
index 59920f9..0a17646 100644
--- a/arch/s390/kvm/intercept.c
+++ b/arch/s390/kvm/intercept.c
@@ -368,8 +368,7 @@ static int handle_operexc(struct kvm_vcpu *vcpu)
trace_kvm_s390_handle_operexc(vcpu, vcpu->arch.sie_block->ipa,
vcpu->arch.sie_block->ipb);
- if (vcpu->arch.sie_block->ipa == 0xb256 &&
- test_kvm_facility(vcpu->kvm, 74))
+ if (vcpu->arch.sie_block->ipa == 0xb256)
return handle_sthyi(vcpu);
if (vcpu->arch.sie_block->ipa == 0 && vcpu->kvm->arch.user_instr0)
diff --git a/arch/s390/kvm/priv.c b/arch/s390/kvm/priv.c
index 93d6cde..26f30ba 100644
--- a/arch/s390/kvm/priv.c
+++ b/arch/s390/kvm/priv.c
@@ -760,6 +760,7 @@ static const intercept_handler_t b2_handlers[256] = {
[0x3b] = handle_io_inst,
[0x3c] = handle_io_inst,
[0x50] = handle_ipte_interlock,
+ [0x56] = handle_sthyi,
[0x5f] = handle_io_inst,
[0x74] = handle_io_inst,
[0x76] = handle_io_inst,
diff --git a/arch/s390/kvm/sthyi.c b/arch/s390/kvm/sthyi.c
index 05c98bb..926b524 100644
--- a/arch/s390/kvm/sthyi.c
+++ b/arch/s390/kvm/sthyi.c
@@ -404,6 +404,9 @@ int handle_sthyi(struct kvm_vcpu *vcpu)
u64 code, addr, cc = 0;
struct sthyi_sctns *sctns = NULL;
+ if (!test_kvm_facility(vcpu->kvm, 74))
+ return kvm_s390_inject_program_int(vcpu, PGM_OPERATION);
+
/*
* STHYI requires extensive locking in the higher hypervisors
* and is very computational/memory expensive. Therefore we
--
2.7.4
^ permalink raw reply related [flat|nested] 17+ messages in thread* [GIT PULL 3/9] KVM: s390: use defines for execution controls
2017-04-07 10:47 [GIT PULL 0/9] KVM: s390: features for 4.12 (via kvm/next) Christian Borntraeger
2017-04-07 10:47 ` [GIT PULL 1/9] KVM: s390: log runtime instrumentation enablement Christian Borntraeger
2017-04-07 10:47 ` [GIT PULL 2/9] KVM: s390: Handle sthyi also for instruction intercept Christian Borntraeger
@ 2017-04-07 10:47 ` Christian Borntraeger
2017-04-07 10:47 ` [GIT PULL 4/9] KVM: s390: Use defines for intercept code Christian Borntraeger
` (5 subsequent siblings)
8 siblings, 0 replies; 17+ messages in thread
From: Christian Borntraeger @ 2017-04-07 10:47 UTC (permalink / raw)
To: Paolo Bonzini, Radim Krčmář
Cc: KVM, Christian Borntraeger, Cornelia Huck, linux-s390, David Hildenbrand
From: David Hildenbrand <david@redhat.com>
Let's replace the bitmasks by defines. Reconstructed from code, comments
and commit messages.
Tried to keep the defines short and map them to feature names. In case
they don't completely map to features, keep them in the stye of ICTL
defines.
This effectively drops all "U" from the existing numbers. I think this
should be fine (as similarly done for e.g. ICTL defines).
I am not 100% sure about the ECA_MVPGI and ECA_PROTEXCI bits as they are
always used in pairs.
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20170313104828.13362-1-david@redhat.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
[some renames, add one missing place]
---
arch/s390/include/asm/kvm_host.h | 18 +++++++++++++++++-
arch/s390/kvm/gaccess.c | 6 +++---
arch/s390/kvm/kvm-s390.c | 36 ++++++++++++++++++------------------
arch/s390/kvm/kvm-s390.h | 2 +-
arch/s390/kvm/priv.c | 2 +-
arch/s390/kvm/vsie.c | 33 ++++++++++++++++-----------------
6 files changed, 56 insertions(+), 41 deletions(-)
diff --git a/arch/s390/include/asm/kvm_host.h b/arch/s390/include/asm/kvm_host.h
index a41faf3..18e9c9d 100644
--- a/arch/s390/include/asm/kvm_host.h
+++ b/arch/s390/include/asm/kvm_host.h
@@ -164,6 +164,13 @@ struct kvm_s390_sie_block {
#define ICTL_RRBE 0x00001000
#define ICTL_TPROT 0x00000200
__u32 ictl; /* 0x0048 */
+#define ECA_CEI 0x80000000
+#define ECA_IB 0x40000000
+#define ECA_SIGPI 0x10000000
+#define ECA_MVPGI 0x01000000
+#define ECA_VX 0x00020000
+#define ECA_PROTEXCI 0x00002000
+#define ECA_SII 0x00000001
__u32 eca; /* 0x004c */
#define ICPT_INST 0x04
#define ICPT_PROGI 0x08
@@ -182,10 +189,18 @@ struct kvm_s390_sie_block {
__u32 ipb; /* 0x0058 */
__u32 scaoh; /* 0x005c */
__u8 reserved60; /* 0x0060 */
+#define ECB_TE 0x10
+#define ECB_SRSI 0x04
+#define ECB_HOSTPROTINT 0x02
__u8 ecb; /* 0x0061 */
+#define ECB2_CMMA 0x80
+#define ECB2_IEP 0x20
+#define ECB2_PFMFI 0x08
+#define ECB2_ESCA 0x04
__u8 ecb2; /* 0x0062 */
-#define ECB3_AES 0x04
#define ECB3_DEA 0x08
+#define ECB3_AES 0x04
+#define ECB3_RI 0x01
__u8 ecb3; /* 0x0063 */
__u32 scaol; /* 0x0064 */
__u8 reserved68[4]; /* 0x0068 */
@@ -224,6 +239,7 @@ struct kvm_s390_sie_block {
__u8 reserved1a4[20]; /* 0x01a4 */
__u64 cbrlo; /* 0x01b8 */
__u8 reserved1c0[8]; /* 0x01c0 */
+#define ECD_HOSTREGMGMT 0x20000000
__u32 ecd; /* 0x01c8 */
__u8 reserved1cc[18]; /* 0x01cc */
__u64 pp; /* 0x01de */
diff --git a/arch/s390/kvm/gaccess.c b/arch/s390/kvm/gaccess.c
index d55c829..709aca9 100644
--- a/arch/s390/kvm/gaccess.c
+++ b/arch/s390/kvm/gaccess.c
@@ -262,7 +262,7 @@ struct aste {
int ipte_lock_held(struct kvm_vcpu *vcpu)
{
- if (vcpu->arch.sie_block->eca & 1) {
+ if (vcpu->arch.sie_block->eca & ECA_SII) {
int rc;
read_lock(&vcpu->kvm->arch.sca_lock);
@@ -361,7 +361,7 @@ static void ipte_unlock_siif(struct kvm_vcpu *vcpu)
void ipte_lock(struct kvm_vcpu *vcpu)
{
- if (vcpu->arch.sie_block->eca & 1)
+ if (vcpu->arch.sie_block->eca & ECA_SII)
ipte_lock_siif(vcpu);
else
ipte_lock_simple(vcpu);
@@ -369,7 +369,7 @@ void ipte_lock(struct kvm_vcpu *vcpu)
void ipte_unlock(struct kvm_vcpu *vcpu)
{
- if (vcpu->arch.sie_block->eca & 1)
+ if (vcpu->arch.sie_block->eca & ECA_SII)
ipte_unlock_siif(vcpu);
else
ipte_unlock_simple(vcpu);
diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
index 5fda468..976373c 100644
--- a/arch/s390/kvm/kvm-s390.c
+++ b/arch/s390/kvm/kvm-s390.c
@@ -1646,7 +1646,7 @@ static void sca_add_vcpu(struct kvm_vcpu *vcpu)
sca->cpu[vcpu->vcpu_id].sda = (__u64) vcpu->arch.sie_block;
vcpu->arch.sie_block->scaoh = (__u32)(((__u64)sca) >> 32);
vcpu->arch.sie_block->scaol = (__u32)(__u64)sca & ~0x3fU;
- vcpu->arch.sie_block->ecb2 |= 0x04U;
+ vcpu->arch.sie_block->ecb2 |= ECB2_ESCA;
set_bit_inv(vcpu->vcpu_id, (unsigned long *) sca->mcn);
} else {
struct bsca_block *sca = vcpu->kvm->arch.sca;
@@ -1700,7 +1700,7 @@ static int sca_switch_to_extended(struct kvm *kvm)
kvm_for_each_vcpu(vcpu_idx, vcpu, kvm) {
vcpu->arch.sie_block->scaoh = scaoh;
vcpu->arch.sie_block->scaol = scaol;
- vcpu->arch.sie_block->ecb2 |= 0x04U;
+ vcpu->arch.sie_block->ecb2 |= ECB2_ESCA;
}
kvm->arch.sca = new_sca;
kvm->arch.use_esca = 1;
@@ -1939,8 +1939,8 @@ int kvm_s390_vcpu_setup_cmma(struct kvm_vcpu *vcpu)
if (!vcpu->arch.sie_block->cbrlo)
return -ENOMEM;
- vcpu->arch.sie_block->ecb2 |= 0x80;
- vcpu->arch.sie_block->ecb2 &= ~0x08;
+ vcpu->arch.sie_block->ecb2 |= ECB2_CMMA;
+ vcpu->arch.sie_block->ecb2 &= ~ECB2_PFMFI;
return 0;
}
@@ -1970,28 +1970,28 @@ int kvm_arch_vcpu_setup(struct kvm_vcpu *vcpu)
/* pgste_set_pte has special handling for !MACHINE_HAS_ESOP */
if (MACHINE_HAS_ESOP)
- vcpu->arch.sie_block->ecb |= 0x02;
+ vcpu->arch.sie_block->ecb |= ECB_HOSTPROTINT;
if (test_kvm_facility(vcpu->kvm, 9))
- vcpu->arch.sie_block->ecb |= 0x04;
+ vcpu->arch.sie_block->ecb |= ECB_SRSI;
if (test_kvm_facility(vcpu->kvm, 73))
- vcpu->arch.sie_block->ecb |= 0x10;
+ vcpu->arch.sie_block->ecb |= ECB_TE;
if (test_kvm_facility(vcpu->kvm, 8) && sclp.has_pfmfi)
- vcpu->arch.sie_block->ecb2 |= 0x08;
+ vcpu->arch.sie_block->ecb2 |= ECB2_PFMFI;
if (test_kvm_facility(vcpu->kvm, 130))
- vcpu->arch.sie_block->ecb2 |= 0x20;
- vcpu->arch.sie_block->eca = 0x1002000U;
+ vcpu->arch.sie_block->ecb2 |= ECB2_IEP;
+ vcpu->arch.sie_block->eca = ECA_MVPGI | ECA_PROTEXCI;
if (sclp.has_cei)
- vcpu->arch.sie_block->eca |= 0x80000000U;
+ vcpu->arch.sie_block->eca |= ECA_CEI;
if (sclp.has_ib)
- vcpu->arch.sie_block->eca |= 0x40000000U;
+ vcpu->arch.sie_block->eca |= ECA_IB;
if (sclp.has_siif)
- vcpu->arch.sie_block->eca |= 1;
+ vcpu->arch.sie_block->eca |= ECA_SII;
if (sclp.has_sigpif)
- vcpu->arch.sie_block->eca |= 0x10000000U;
+ vcpu->arch.sie_block->eca |= ECA_SIGPI;
if (test_kvm_facility(vcpu->kvm, 129)) {
- vcpu->arch.sie_block->eca |= 0x00020000;
- vcpu->arch.sie_block->ecd |= 0x20000000;
+ vcpu->arch.sie_block->eca |= ECA_VX;
+ vcpu->arch.sie_block->ecd |= ECD_HOSTREGMGMT;
}
vcpu->arch.sie_block->riccbd = (unsigned long) &vcpu->run->s.regs.riccb;
vcpu->arch.sie_block->ictl |= ICTL_ISKE | ICTL_SSKE | ICTL_RRBE;
@@ -2752,9 +2752,9 @@ static void sync_regs(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
if ((kvm_run->kvm_dirty_regs & KVM_SYNC_RICCB) &&
test_kvm_facility(vcpu->kvm, 64) &&
riccb->valid &&
- !(vcpu->arch.sie_block->ecb3 & 0x01)) {
+ !(vcpu->arch.sie_block->ecb3 & ECB3_RI)) {
VCPU_EVENT(vcpu, 3, "%s", "ENABLE: RI (sync_regs)");
- vcpu->arch.sie_block->ecb3 |= 0x01;
+ vcpu->arch.sie_block->ecb3 |= ECB3_RI;
}
save_access_regs(vcpu->arch.host_acrs);
restore_access_regs(vcpu->run->s.regs.acrs);
diff --git a/arch/s390/kvm/kvm-s390.h b/arch/s390/kvm/kvm-s390.h
index af9fa91..dfdcde1 100644
--- a/arch/s390/kvm/kvm-s390.h
+++ b/arch/s390/kvm/kvm-s390.h
@@ -25,7 +25,7 @@
typedef int (*intercept_handler_t)(struct kvm_vcpu *vcpu);
/* Transactional Memory Execution related macros */
-#define IS_TE_ENABLED(vcpu) ((vcpu->arch.sie_block->ecb & 0x10))
+#define IS_TE_ENABLED(vcpu) ((vcpu->arch.sie_block->ecb & ECB_TE))
#define TDB_FORMAT1 1
#define IS_ITDB_VALID(vcpu) ((*(char *)vcpu->arch.sie_block->itdba == TDB_FORMAT1))
diff --git a/arch/s390/kvm/priv.c b/arch/s390/kvm/priv.c
index 26f30ba..d42bb2d 100644
--- a/arch/s390/kvm/priv.c
+++ b/arch/s390/kvm/priv.c
@@ -38,7 +38,7 @@ static int handle_ri(struct kvm_vcpu *vcpu)
{
if (test_kvm_facility(vcpu->kvm, 64)) {
VCPU_EVENT(vcpu, 3, "%s", "ENABLE: RI (lazy)");
- vcpu->arch.sie_block->ecb3 |= 0x01;
+ vcpu->arch.sie_block->ecb3 |= ECB3_RI;
kvm_s390_retry_instr(vcpu);
return 0;
} else
diff --git a/arch/s390/kvm/vsie.c b/arch/s390/kvm/vsie.c
index 5491be3..d91f1df 100644
--- a/arch/s390/kvm/vsie.c
+++ b/arch/s390/kvm/vsie.c
@@ -249,7 +249,7 @@ static int shadow_scb(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page)
{
struct kvm_s390_sie_block *scb_o = vsie_page->scb_o;
struct kvm_s390_sie_block *scb_s = &vsie_page->scb_s;
- bool had_tx = scb_s->ecb & 0x10U;
+ bool had_tx = scb_s->ecb & ECB_TE;
unsigned long new_mso = 0;
int rc;
@@ -307,34 +307,34 @@ static int shadow_scb(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page)
scb_s->ihcpu = scb_o->ihcpu;
/* MVPG and Protection Exception Interpretation are always available */
- scb_s->eca |= scb_o->eca & 0x01002000U;
+ scb_s->eca |= scb_o->eca & (ECA_MVPGI | ECA_PROTEXCI);
/* Host-protection-interruption introduced with ESOP */
if (test_kvm_cpu_feat(vcpu->kvm, KVM_S390_VM_CPU_FEAT_ESOP))
- scb_s->ecb |= scb_o->ecb & 0x02U;
+ scb_s->ecb |= scb_o->ecb & ECB_HOSTPROTINT;
/* transactional execution */
if (test_kvm_facility(vcpu->kvm, 73)) {
/* remap the prefix is tx is toggled on */
- if ((scb_o->ecb & 0x10U) && !had_tx)
+ if ((scb_o->ecb & ECB_TE) && !had_tx)
prefix_unmapped(vsie_page);
- scb_s->ecb |= scb_o->ecb & 0x10U;
+ scb_s->ecb |= scb_o->ecb & ECB_TE;
}
/* SIMD */
if (test_kvm_facility(vcpu->kvm, 129)) {
- scb_s->eca |= scb_o->eca & 0x00020000U;
- scb_s->ecd |= scb_o->ecd & 0x20000000U;
+ scb_s->eca |= scb_o->eca & ECA_VX;
+ scb_s->ecd |= scb_o->ecd & ECD_HOSTREGMGMT;
}
/* Run-time-Instrumentation */
if (test_kvm_facility(vcpu->kvm, 64))
- scb_s->ecb3 |= scb_o->ecb3 & 0x01U;
+ scb_s->ecb3 |= scb_o->ecb3 & ECB3_RI;
/* Instruction Execution Prevention */
if (test_kvm_facility(vcpu->kvm, 130))
- scb_s->ecb2 |= scb_o->ecb2 & 0x20U;
+ scb_s->ecb2 |= scb_o->ecb2 & ECB2_IEP;
if (test_kvm_cpu_feat(vcpu->kvm, KVM_S390_VM_CPU_FEAT_SIIF))
- scb_s->eca |= scb_o->eca & 0x00000001U;
+ scb_s->eca |= scb_o->eca & ECA_SII;
if (test_kvm_cpu_feat(vcpu->kvm, KVM_S390_VM_CPU_FEAT_IB))
- scb_s->eca |= scb_o->eca & 0x40000000U;
+ scb_s->eca |= scb_o->eca & ECA_IB;
if (test_kvm_cpu_feat(vcpu->kvm, KVM_S390_VM_CPU_FEAT_CEI))
- scb_s->eca |= scb_o->eca & 0x80000000U;
+ scb_s->eca |= scb_o->eca & ECA_CEI;
prepare_ibc(vcpu, vsie_page);
rc = shadow_crycb(vcpu, vsie_page);
@@ -406,7 +406,7 @@ static int map_prefix(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page)
prefix += scb_s->mso;
rc = kvm_s390_shadow_fault(vcpu, vsie_page->gmap, prefix);
- if (!rc && (scb_s->ecb & 0x10U))
+ if (!rc && (scb_s->ecb & ECB_TE))
rc = kvm_s390_shadow_fault(vcpu, vsie_page->gmap,
prefix + PAGE_SIZE);
/*
@@ -543,7 +543,7 @@ static int pin_blocks(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page)
}
gpa = scb_o->itdba & ~0xffUL;
- if (gpa && (scb_s->ecb & 0x10U)) {
+ if (gpa && (scb_s->ecb & ECB_TE)) {
if (!(gpa & ~0x1fffU)) {
rc = set_validity_icpt(scb_s, 0x0080U);
goto unpin;
@@ -558,8 +558,7 @@ static int pin_blocks(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page)
}
gpa = scb_o->gvrd & ~0x1ffUL;
- if (gpa && (scb_s->eca & 0x00020000U) &&
- !(scb_s->ecd & 0x20000000U)) {
+ if (gpa && (scb_s->eca & ECA_VX) && !(scb_s->ecd & ECD_HOSTREGMGMT)) {
if (!(gpa & ~0x1fffUL)) {
rc = set_validity_icpt(scb_s, 0x1310U);
goto unpin;
@@ -577,7 +576,7 @@ static int pin_blocks(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page)
}
gpa = scb_o->riccbd & ~0x3fUL;
- if (gpa && (scb_s->ecb3 & 0x01U)) {
+ if (gpa && (scb_s->ecb3 & ECB3_RI)) {
if (!(gpa & ~0x1fffUL)) {
rc = set_validity_icpt(scb_s, 0x0043U);
goto unpin;
--
2.7.4
^ permalink raw reply related [flat|nested] 17+ messages in thread* [GIT PULL 4/9] KVM: s390: Use defines for intercept code
2017-04-07 10:47 [GIT PULL 0/9] KVM: s390: features for 4.12 (via kvm/next) Christian Borntraeger
` (2 preceding siblings ...)
2017-04-07 10:47 ` [GIT PULL 3/9] KVM: s390: use defines for execution controls Christian Borntraeger
@ 2017-04-07 10:47 ` Christian Borntraeger
2017-04-07 10:47 ` [GIT PULL 5/9] KVM: s390: gs support for kvm guests Christian Borntraeger
` (4 subsequent siblings)
8 siblings, 0 replies; 17+ messages in thread
From: Christian Borntraeger @ 2017-04-07 10:47 UTC (permalink / raw)
To: Paolo Bonzini, Radim Krčmář
Cc: KVM, Christian Borntraeger, Cornelia Huck, linux-s390, Farhan Ali
From: Farhan Ali <alifm@linux.vnet.ibm.com>
Let's use #define values for better readability.
Signed-off-by: Farhan Ali <alifm@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
---
arch/s390/include/asm/kvm_host.h | 3 +++
arch/s390/kvm/intercept.c | 20 ++++++++++----------
2 files changed, 13 insertions(+), 10 deletions(-)
diff --git a/arch/s390/include/asm/kvm_host.h b/arch/s390/include/asm/kvm_host.h
index 18e9c9d..54e36e7 100644
--- a/arch/s390/include/asm/kvm_host.h
+++ b/arch/s390/include/asm/kvm_host.h
@@ -175,7 +175,10 @@ struct kvm_s390_sie_block {
#define ICPT_INST 0x04
#define ICPT_PROGI 0x08
#define ICPT_INSTPROGI 0x0C
+#define ICPT_EXTREQ 0x10
#define ICPT_EXTINT 0x14
+#define ICPT_IOREQ 0x18
+#define ICPT_WAIT 0x1c
#define ICPT_VALIDITY 0x20
#define ICPT_STOP 0x28
#define ICPT_OPEREXC 0x2C
diff --git a/arch/s390/kvm/intercept.c b/arch/s390/kvm/intercept.c
index 0a17646..e831f4b 100644
--- a/arch/s390/kvm/intercept.c
+++ b/arch/s390/kvm/intercept.c
@@ -403,26 +403,26 @@ int kvm_handle_sie_intercept(struct kvm_vcpu *vcpu)
return -EOPNOTSUPP;
switch (vcpu->arch.sie_block->icptcode) {
- case 0x10:
- case 0x18:
+ case ICPT_EXTREQ:
+ case ICPT_IOREQ:
return handle_noop(vcpu);
- case 0x04:
+ case ICPT_INST:
rc = handle_instruction(vcpu);
break;
- case 0x08:
+ case ICPT_PROGI:
return handle_prog(vcpu);
- case 0x14:
+ case ICPT_EXTINT:
return handle_external_interrupt(vcpu);
- case 0x1c:
+ case ICPT_WAIT:
return kvm_s390_handle_wait(vcpu);
- case 0x20:
+ case ICPT_VALIDITY:
return handle_validity(vcpu);
- case 0x28:
+ case ICPT_STOP:
return handle_stop(vcpu);
- case 0x2c:
+ case ICPT_OPEREXC:
rc = handle_operexc(vcpu);
break;
- case 0x38:
+ case ICPT_PARTEXEC:
rc = handle_partial_execution(vcpu);
break;
default:
--
2.7.4
^ permalink raw reply related [flat|nested] 17+ messages in thread* [GIT PULL 5/9] KVM: s390: gs support for kvm guests
2017-04-07 10:47 [GIT PULL 0/9] KVM: s390: features for 4.12 (via kvm/next) Christian Borntraeger
` (3 preceding siblings ...)
2017-04-07 10:47 ` [GIT PULL 4/9] KVM: s390: Use defines for intercept code Christian Borntraeger
@ 2017-04-07 10:47 ` Christian Borntraeger
2017-04-07 11:42 ` David Hildenbrand
2017-04-07 10:47 ` [GIT PULL 6/9] KVM: s390: interface for suppressible I/O adapters Christian Borntraeger
` (3 subsequent siblings)
8 siblings, 1 reply; 17+ messages in thread
From: Christian Borntraeger @ 2017-04-07 10:47 UTC (permalink / raw)
To: Paolo Bonzini, Radim Krčmář
Cc: KVM, Christian Borntraeger, Cornelia Huck, linux-s390, Fan Zhang
From: Fan Zhang <zhangfan@linux.vnet.ibm.com>
This patch adds guarded storage support for KVM guest. We need to
setup the necessary control blocks, the kvm_run structure for the
new registers, the necessary wrappers for VSIE, as well as the
machine check save areas.
GS is enabled lazily and the register saving and reloading is done in
KVM code. As this feature adds new content for migration, we provide
a new capability for enablement (KVM_CAP_S390_GS).
Signed-off-by: Fan Zhang <zhangfan@linux.vnet.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Janosch Frank <frankja@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
---
Documentation/virtual/kvm/api.txt | 9 ++++++
arch/s390/include/asm/kvm_host.h | 8 ++++-
arch/s390/include/uapi/asm/kvm.h | 14 ++++++++-
arch/s390/kvm/intercept.c | 1 +
arch/s390/kvm/interrupt.c | 33 +++++++++++++++++++--
arch/s390/kvm/kvm-s390.c | 62 +++++++++++++++++++++++++++++++++++++++
arch/s390/kvm/kvm-s390.h | 1 +
arch/s390/kvm/priv.c | 27 +++++++++++++++++
arch/s390/kvm/vsie.c | 39 ++++++++++++++++++++++++
include/uapi/linux/kvm.h | 1 +
10 files changed, 191 insertions(+), 4 deletions(-)
diff --git a/Documentation/virtual/kvm/api.txt b/Documentation/virtual/kvm/api.txt
index 3c248f7..7252508 100644
--- a/Documentation/virtual/kvm/api.txt
+++ b/Documentation/virtual/kvm/api.txt
@@ -4101,6 +4101,15 @@ to take care of that.
This capability can be enabled dynamically even if VCPUs were already
created and are running.
+7.9 KVM_CAP_S390_GS
+
+Architectures: s390
+Parameters: none
+Returns: 0 on success; -EINVAL if the machine does not support
+ guarded storage; -EBUSY if a VCPU has already been created.
+
+Allows use of guarded storage for the KVM guest.
+
8. Other capabilities.
----------------------
diff --git a/arch/s390/include/asm/kvm_host.h b/arch/s390/include/asm/kvm_host.h
index 54e36e7..1af090d 100644
--- a/arch/s390/include/asm/kvm_host.h
+++ b/arch/s390/include/asm/kvm_host.h
@@ -25,6 +25,7 @@
#include <asm/cpu.h>
#include <asm/fpu/api.h>
#include <asm/isc.h>
+#include <asm/guarded_storage.h>
#define KVM_S390_BSCA_CPU_SLOTS 64
#define KVM_S390_ESCA_CPU_SLOTS 248
@@ -192,6 +193,7 @@ struct kvm_s390_sie_block {
__u32 ipb; /* 0x0058 */
__u32 scaoh; /* 0x005c */
__u8 reserved60; /* 0x0060 */
+#define ECB_GS 0x40
#define ECB_TE 0x10
#define ECB_SRSI 0x04
#define ECB_HOSTPROTINT 0x02
@@ -237,7 +239,9 @@ struct kvm_s390_sie_block {
__u32 crycbd; /* 0x00fc */
__u64 gcr[16]; /* 0x0100 */
__u64 gbea; /* 0x0180 */
- __u8 reserved188[24]; /* 0x0188 */
+ __u8 reserved188[8]; /* 0x0188 */
+ __u64 sdnxo; /* 0x0190 */
+ __u8 reserved198[8]; /* 0x0198 */
__u32 fac; /* 0x01a0 */
__u8 reserved1a4[20]; /* 0x01a4 */
__u64 cbrlo; /* 0x01b8 */
@@ -573,6 +577,7 @@ struct kvm_vcpu_arch {
/* if vsie is active, currently executed shadow sie control block */
struct kvm_s390_sie_block *vsie_block;
unsigned int host_acrs[NUM_ACRS];
+ struct gs_cb *host_gscb;
struct fpu host_fpregs;
struct kvm_s390_local_interrupt local_int;
struct hrtimer ckc_timer;
@@ -593,6 +598,7 @@ struct kvm_vcpu_arch {
*/
seqcount_t cputm_seqcount;
__u64 cputm_start;
+ bool gs_enabled;
};
struct kvm_vm_stat {
diff --git a/arch/s390/include/uapi/asm/kvm.h b/arch/s390/include/uapi/asm/kvm.h
index a2ffec4..5bd23cf 100644
--- a/arch/s390/include/uapi/asm/kvm.h
+++ b/arch/s390/include/uapi/asm/kvm.h
@@ -197,6 +197,10 @@ struct kvm_guest_debug_arch {
#define KVM_SYNC_VRS (1UL << 6)
#define KVM_SYNC_RICCB (1UL << 7)
#define KVM_SYNC_FPRS (1UL << 8)
+#define KVM_SYNC_GSCB (1UL << 9)
+/* length and alignment of the sdnx as a power of two */
+#define SDNXC 8
+#define SDNXL (1UL << SDNXC)
/* definition of registers in kvm_run */
struct kvm_sync_regs {
__u64 prefix; /* prefix register */
@@ -217,8 +221,16 @@ struct kvm_sync_regs {
};
__u8 reserved[512]; /* for future vector expansion */
__u32 fpc; /* valid on KVM_SYNC_VRS or KVM_SYNC_FPRS */
- __u8 padding[52]; /* riccb needs to be 64byte aligned */
+ __u8 padding1[52]; /* riccb needs to be 64byte aligned */
__u8 riccb[64]; /* runtime instrumentation controls block */
+ __u8 padding2[192]; /* sdnx needs to be 256byte aligned */
+ union {
+ __u8 sdnx[SDNXL]; /* state description annex */
+ struct {
+ __u64 reserved1[2];
+ __u64 gscb[4];
+ };
+ };
};
#define KVM_REG_S390_TODPR (KVM_REG_S390 | KVM_REG_SIZE_U32 | 0x1)
diff --git a/arch/s390/kvm/intercept.c b/arch/s390/kvm/intercept.c
index e831f4b..f5378f3 100644
--- a/arch/s390/kvm/intercept.c
+++ b/arch/s390/kvm/intercept.c
@@ -35,6 +35,7 @@ static const intercept_handler_t instruction_handlers[256] = {
[0xb6] = kvm_s390_handle_stctl,
[0xb7] = kvm_s390_handle_lctl,
[0xb9] = kvm_s390_handle_b9,
+ [0xe3] = kvm_s390_handle_e3,
[0xe5] = kvm_s390_handle_e5,
[0xeb] = kvm_s390_handle_eb,
};
diff --git a/arch/s390/kvm/interrupt.c b/arch/s390/kvm/interrupt.c
index 169558d..311eef0 100644
--- a/arch/s390/kvm/interrupt.c
+++ b/arch/s390/kvm/interrupt.c
@@ -410,6 +410,7 @@ static int __write_machine_check(struct kvm_vcpu *vcpu,
struct kvm_s390_mchk_info *mchk)
{
unsigned long ext_sa_addr;
+ unsigned long lc;
freg_t fprs[NUM_FPRS];
union mci mci;
int rc;
@@ -422,8 +423,28 @@ static int __write_machine_check(struct kvm_vcpu *vcpu,
/* Extended save area */
rc = read_guest_lc(vcpu, __LC_MCESAD, &ext_sa_addr,
sizeof(unsigned long));
- /* Only bits 0-53 are used for address formation */
- ext_sa_addr &= ~0x3ffUL;
+ /* Only bits 0 through 63-LC are used for address formation */
+ lc = ext_sa_addr & MCESA_LC_MASK;
+ if (test_kvm_facility(vcpu->kvm, 133)) {
+ switch (lc) {
+ case 0:
+ case 10:
+ ext_sa_addr &= ~0x3ffUL;
+ break;
+ case 11:
+ ext_sa_addr &= ~0x7ffUL;
+ break;
+ case 12:
+ ext_sa_addr &= ~0xfffUL;
+ break;
+ default:
+ ext_sa_addr = 0;
+ break;
+ }
+ } else {
+ ext_sa_addr &= ~0x3ffUL;
+ }
+
if (!rc && mci.vr && ext_sa_addr && test_kvm_facility(vcpu->kvm, 129)) {
if (write_guest_abs(vcpu, ext_sa_addr, vcpu->run->s.regs.vrs,
512))
@@ -431,6 +452,14 @@ static int __write_machine_check(struct kvm_vcpu *vcpu,
} else {
mci.vr = 0;
}
+ if (!rc && mci.gs && ext_sa_addr && test_kvm_facility(vcpu->kvm, 133)
+ && (lc == 11 || lc == 12)) {
+ if (write_guest_abs(vcpu, ext_sa_addr + 1024,
+ &vcpu->run->s.regs.gscb, 32))
+ mci.gs = 0;
+ } else {
+ mci.gs = 0;
+ }
/* General interruption information */
rc |= put_guest_lc(vcpu, 1, (u8 __user *) __LC_AR_MODE_ID);
diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
index 976373c..f83f18b 100644
--- a/arch/s390/kvm/kvm-s390.c
+++ b/arch/s390/kvm/kvm-s390.c
@@ -405,6 +405,9 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
case KVM_CAP_S390_RI:
r = test_facility(64);
break;
+ case KVM_CAP_S390_GS:
+ r = test_facility(133);
+ break;
default:
r = 0;
}
@@ -541,6 +544,20 @@ static int kvm_vm_ioctl_enable_cap(struct kvm *kvm, struct kvm_enable_cap *cap)
VM_EVENT(kvm, 3, "ENABLE: CAP_S390_RI %s",
r ? "(not available)" : "(success)");
break;
+ case KVM_CAP_S390_GS:
+ r = -EINVAL;
+ mutex_lock(&kvm->lock);
+ if (atomic_read(&kvm->online_vcpus)) {
+ r = -EBUSY;
+ } else if (test_facility(133)) {
+ set_kvm_facility(kvm->arch.model.fac_mask, 133);
+ set_kvm_facility(kvm->arch.model.fac_list, 133);
+ r = 0;
+ }
+ mutex_unlock(&kvm->lock);
+ VM_EVENT(kvm, 3, "ENABLE: CAP_S390_GS %s",
+ r ? "(not available)" : "(success)");
+ break;
case KVM_CAP_S390_USER_STSI:
VM_EVENT(kvm, 3, "%s", "ENABLE: CAP_S390_USER_STSI");
kvm->arch.user_stsi = 1;
@@ -1749,6 +1766,8 @@ int kvm_arch_vcpu_init(struct kvm_vcpu *vcpu)
kvm_s390_set_prefix(vcpu, 0);
if (test_kvm_facility(vcpu->kvm, 64))
vcpu->run->kvm_valid_regs |= KVM_SYNC_RICCB;
+ if (test_kvm_facility(vcpu->kvm, 133))
+ vcpu->run->kvm_valid_regs |= KVM_SYNC_GSCB;
/* fprs can be synchronized via vrs, even if the guest has no vx. With
* MACHINE_HAS_VX, (load|store)_fpu_regs() will work with vrs format.
*/
@@ -1993,6 +2012,8 @@ int kvm_arch_vcpu_setup(struct kvm_vcpu *vcpu)
vcpu->arch.sie_block->eca |= ECA_VX;
vcpu->arch.sie_block->ecd |= ECD_HOSTREGMGMT;
}
+ vcpu->arch.sie_block->sdnxo = ((unsigned long) &vcpu->run->s.regs.sdnx)
+ | SDNXC;
vcpu->arch.sie_block->riccbd = (unsigned long) &vcpu->run->s.regs.riccb;
vcpu->arch.sie_block->ictl |= ICTL_ISKE | ICTL_SSKE | ICTL_RRBE;
@@ -2720,8 +2741,10 @@ static int __vcpu_run(struct kvm_vcpu *vcpu)
static void sync_regs(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
{
struct runtime_instr_cb *riccb;
+ struct gs_cb *gscb;
riccb = (struct runtime_instr_cb *) &kvm_run->s.regs.riccb;
+ gscb = (struct gs_cb *) &kvm_run->s.regs.gscb;
vcpu->arch.sie_block->gpsw.mask = kvm_run->psw_mask;
vcpu->arch.sie_block->gpsw.addr = kvm_run->psw_addr;
if (kvm_run->kvm_dirty_regs & KVM_SYNC_PREFIX)
@@ -2756,6 +2779,19 @@ static void sync_regs(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
VCPU_EVENT(vcpu, 3, "%s", "ENABLE: RI (sync_regs)");
vcpu->arch.sie_block->ecb3 |= ECB3_RI;
}
+ /*
+ * If userspace sets the gscb (e.g. after migration) to non-zero,
+ * we should enable GS here instead of doing the lazy enablement.
+ */
+ if ((kvm_run->kvm_dirty_regs & KVM_SYNC_GSCB) &&
+ test_kvm_facility(vcpu->kvm, 133) &&
+ gscb->gssm &&
+ !vcpu->arch.gs_enabled) {
+ VCPU_EVENT(vcpu, 3, "%s", "ENABLE: GS (sync_regs)");
+ vcpu->arch.sie_block->ecb |= ECB_GS;
+ vcpu->arch.sie_block->ecd |= ECD_HOSTREGMGMT;
+ vcpu->arch.gs_enabled = 1;
+ }
save_access_regs(vcpu->arch.host_acrs);
restore_access_regs(vcpu->run->s.regs.acrs);
/* save host (userspace) fprs/vrs */
@@ -2770,6 +2806,20 @@ static void sync_regs(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
if (test_fp_ctl(current->thread.fpu.fpc))
/* User space provided an invalid FPC, let's clear it */
current->thread.fpu.fpc = 0;
+ if (MACHINE_HAS_GS) {
+ preempt_disable();
+ __ctl_set_bit(2, 4);
+ if (current->thread.gs_cb) {
+ vcpu->arch.host_gscb = current->thread.gs_cb;
+ save_gs_cb(vcpu->arch.host_gscb);
+ }
+ if (vcpu->arch.gs_enabled) {
+ current->thread.gs_cb = (struct gs_cb *)
+ &vcpu->run->s.regs.gscb;
+ restore_gs_cb(current->thread.gs_cb);
+ }
+ preempt_enable();
+ }
kvm_run->kvm_dirty_regs = 0;
}
@@ -2796,6 +2846,18 @@ static void store_regs(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
/* Restore will be done lazily at return */
current->thread.fpu.fpc = vcpu->arch.host_fpregs.fpc;
current->thread.fpu.regs = vcpu->arch.host_fpregs.regs;
+ if (MACHINE_HAS_GS) {
+ __ctl_set_bit(2, 4);
+ if (vcpu->arch.gs_enabled)
+ save_gs_cb(current->thread.gs_cb);
+ preempt_disable();
+ current->thread.gs_cb = vcpu->arch.host_gscb;
+ restore_gs_cb(vcpu->arch.host_gscb);
+ preempt_enable();
+ if (!vcpu->arch.host_gscb)
+ __ctl_clear_bit(2, 4);
+ vcpu->arch.host_gscb = NULL;
+ }
}
diff --git a/arch/s390/kvm/kvm-s390.h b/arch/s390/kvm/kvm-s390.h
index dfdcde1..455124f 100644
--- a/arch/s390/kvm/kvm-s390.h
+++ b/arch/s390/kvm/kvm-s390.h
@@ -246,6 +246,7 @@ static inline void kvm_s390_retry_instr(struct kvm_vcpu *vcpu)
int is_valid_psw(psw_t *psw);
int kvm_s390_handle_aa(struct kvm_vcpu *vcpu);
int kvm_s390_handle_b2(struct kvm_vcpu *vcpu);
+int kvm_s390_handle_e3(struct kvm_vcpu *vcpu);
int kvm_s390_handle_e5(struct kvm_vcpu *vcpu);
int kvm_s390_handle_01(struct kvm_vcpu *vcpu);
int kvm_s390_handle_b9(struct kvm_vcpu *vcpu);
diff --git a/arch/s390/kvm/priv.c b/arch/s390/kvm/priv.c
index d42bb2d..0ffe973 100644
--- a/arch/s390/kvm/priv.c
+++ b/arch/s390/kvm/priv.c
@@ -53,6 +53,33 @@ int kvm_s390_handle_aa(struct kvm_vcpu *vcpu)
return -EOPNOTSUPP;
}
+static int handle_gs(struct kvm_vcpu *vcpu)
+{
+ if (test_kvm_facility(vcpu->kvm, 133)) {
+ VCPU_EVENT(vcpu, 3, "%s", "ENABLE: GS (lazy)");
+ preempt_disable();
+ __ctl_set_bit(2, 4);
+ current->thread.gs_cb = (struct gs_cb *)&vcpu->run->s.regs.gscb;
+ restore_gs_cb(current->thread.gs_cb);
+ preempt_enable();
+ vcpu->arch.sie_block->ecb |= ECB_GS;
+ vcpu->arch.sie_block->ecd |= ECD_HOSTREGMGMT;
+ vcpu->arch.gs_enabled = 1;
+ kvm_s390_retry_instr(vcpu);
+ return 0;
+ } else
+ return kvm_s390_inject_program_int(vcpu, PGM_OPERATION);
+}
+
+int kvm_s390_handle_e3(struct kvm_vcpu *vcpu)
+{
+ int code = vcpu->arch.sie_block->ipb & 0xff;
+
+ if (code == 0x49 || code == 0x4d)
+ return handle_gs(vcpu);
+ else
+ return -EOPNOTSUPP;
+}
/* Handle SCK (SET CLOCK) interception */
static int handle_set_clock(struct kvm_vcpu *vcpu)
{
diff --git a/arch/s390/kvm/vsie.c b/arch/s390/kvm/vsie.c
index d91f1df..2fafc2b 100644
--- a/arch/s390/kvm/vsie.c
+++ b/arch/s390/kvm/vsie.c
@@ -329,6 +329,11 @@ static int shadow_scb(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page)
/* Instruction Execution Prevention */
if (test_kvm_facility(vcpu->kvm, 130))
scb_s->ecb2 |= scb_o->ecb2 & ECB2_IEP;
+ /* Guarded Storage */
+ if (test_kvm_facility(vcpu->kvm, 133)) {
+ scb_s->ecb |= scb_o->ecb & ECB_GS;
+ scb_s->ecd |= scb_o->ecd & ECD_HOSTREGMGMT;
+ }
if (test_kvm_cpu_feat(vcpu->kvm, KVM_S390_VM_CPU_FEAT_SIIF))
scb_s->eca |= scb_o->eca & ECA_SII;
if (test_kvm_cpu_feat(vcpu->kvm, KVM_S390_VM_CPU_FEAT_IB))
@@ -496,6 +501,13 @@ static void unpin_blocks(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page)
unpin_guest_page(vcpu->kvm, gpa, hpa);
scb_s->riccbd = 0;
}
+
+ hpa = scb_s->sdnxo;
+ if (hpa) {
+ gpa = scb_o->sdnxo;
+ unpin_guest_page(vcpu->kvm, gpa, hpa);
+ scb_s->sdnxo = 0;
+ }
}
/*
@@ -590,6 +602,33 @@ static int pin_blocks(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page)
goto unpin;
scb_s->riccbd = hpa;
}
+ if ((scb_s->ecb & ECB_GS) && !(scb_s->ecd & ECD_HOSTREGMGMT)) {
+ unsigned long sdnxc;
+
+ gpa = scb_o->sdnxo & ~0xfUL;
+ sdnxc = scb_o->sdnxo & 0xfUL;
+ if (!gpa || !(gpa & ~0x1fffUL)) {
+ rc = set_validity_icpt(scb_s, 0x10b0U);
+ goto unpin;
+ }
+ if (sdnxc < 6 || sdnxc > 12) {
+ rc = set_validity_icpt(scb_s, 0x10b1U);
+ goto unpin;
+ }
+ if (gpa & ((1 << sdnxc) - 1)) {
+ rc = set_validity_icpt(scb_s, 0x10b2U);
+ goto unpin;
+ }
+ /* Due to alignment rules (checked above) this cannot
+ * cross page boundaries
+ */
+ rc = pin_guest_page(vcpu->kvm, gpa, &hpa);
+ if (rc == -EINVAL)
+ rc = set_validity_icpt(scb_s, 0x10b0U);
+ if (rc)
+ goto unpin;
+ scb_s->sdnxo = hpa;
+ }
return 0;
unpin:
unpin_blocks(vcpu, vsie_page);
diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h
index f51d508..c9d5227 100644
--- a/include/uapi/linux/kvm.h
+++ b/include/uapi/linux/kvm.h
@@ -883,6 +883,7 @@ struct kvm_ppc_resize_hpt {
#define KVM_CAP_PPC_MMU_RADIX 134
#define KVM_CAP_PPC_MMU_HASH_V3 135
#define KVM_CAP_IMMEDIATE_EXIT 136
+#define KVM_CAP_S390_GS 137
#ifdef KVM_CAP_IRQ_ROUTING
--
2.7.4
^ permalink raw reply related [flat|nested] 17+ messages in thread* Re: [GIT PULL 5/9] KVM: s390: gs support for kvm guests
2017-04-07 10:47 ` [GIT PULL 5/9] KVM: s390: gs support for kvm guests Christian Borntraeger
@ 2017-04-07 11:42 ` David Hildenbrand
2017-04-07 12:01 ` Christian Borntraeger
2017-04-07 12:27 ` [PATCH] KVM: s390: Fix sdnxo setting for nested guests Christian Borntraeger
0 siblings, 2 replies; 17+ messages in thread
From: David Hildenbrand @ 2017-04-07 11:42 UTC (permalink / raw)
To: Christian Borntraeger, Paolo Bonzini, Radim Krčmář
Cc: KVM, Cornelia Huck, linux-s390, Fan Zhang
> +
> + hpa = scb_s->sdnxo;
> + if (hpa) {
> + gpa = scb_o->sdnxo;
Other blocks in this function explicitly extract the gpa portion
gpa = scb_o->sdnxo & ~0xfUL;
But this shouldn't matter as it will be dropped either way when
converting to pfn. At least for now this should work.
> + unpin_guest_page(vcpu->kvm, gpa, hpa);
> + scb_s->sdnxo = 0;
> + }
> }
>
> /*
> @@ -590,6 +602,33 @@ static int pin_blocks(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page)
> goto unpin;
> scb_s->riccbd = hpa;
> }
> + if ((scb_s->ecb & ECB_GS) && !(scb_s->ecd & ECD_HOSTREGMGMT)) {
> + unsigned long sdnxc;
> +
> + gpa = scb_o->sdnxo & ~0xfUL;
> + sdnxc = scb_o->sdnxo & 0xfUL;
> + if (!gpa || !(gpa & ~0x1fffUL)) {
> + rc = set_validity_icpt(scb_s, 0x10b0U);
> + goto unpin;
> + }
> + if (sdnxc < 6 || sdnxc > 12) {
> + rc = set_validity_icpt(scb_s, 0x10b1U);
> + goto unpin;
> + }
> + if (gpa & ((1 << sdnxc) - 1)) {
> + rc = set_validity_icpt(scb_s, 0x10b2U);
> + goto unpin;
> + }
> + /* Due to alignment rules (checked above) this cannot
> + * cross page boundaries
> + */
> + rc = pin_guest_page(vcpu->kvm, gpa, &hpa);
> + if (rc == -EINVAL)
> + rc = set_validity_icpt(scb_s, 0x10b0U);
> + if (rc)
> + goto unpin;
> + scb_s->sdnxo = hpa;
I'm curious, this will result in no sdnxc getting set. Shouldn't there
be something like the following, cause otherwise there would be a
validity icpt 0x10b1U (guessing from the code above).
scb_s->sdnxo = hpa | sdnxc;
> + }
> return 0;
> unpin:
> unpin_blocks(vcpu, vsie_page);
> diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h
> index f51d508..c9d5227 100644
> --- a/include/uapi/linux/kvm.h
> +++ b/include/uapi/linux/kvm.h
> @@ -883,6 +883,7 @@ struct kvm_ppc_resize_hpt {
> #define KVM_CAP_PPC_MMU_RADIX 134
> #define KVM_CAP_PPC_MMU_HASH_V3 135
> #define KVM_CAP_IMMEDIATE_EXIT 136
> +#define KVM_CAP_S390_GS 137
>
> #ifdef KVM_CAP_IRQ_ROUTING
>
>
--
Thanks,
David
^ permalink raw reply [flat|nested] 17+ messages in thread* Re: [GIT PULL 5/9] KVM: s390: gs support for kvm guests
2017-04-07 11:42 ` David Hildenbrand
@ 2017-04-07 12:01 ` Christian Borntraeger
2017-04-11 12:42 ` Christian Borntraeger
2017-04-07 12:27 ` [PATCH] KVM: s390: Fix sdnxo setting for nested guests Christian Borntraeger
1 sibling, 1 reply; 17+ messages in thread
From: Christian Borntraeger @ 2017-04-07 12:01 UTC (permalink / raw)
To: David Hildenbrand, Paolo Bonzini, Radim Krčmář
Cc: KVM, Cornelia Huck, linux-s390, Fan Zhang
On 04/07/2017 01:42 PM, David Hildenbrand wrote:
>
>> +
>> + hpa = scb_s->sdnxo;
>> + if (hpa) {
>> + gpa = scb_o->sdnxo;
>
> Other blocks in this function explicitly extract the gpa portion
>
> gpa = scb_o->sdnxo & ~0xfUL;
>
> But this shouldn't matter as it will be dropped either way when
> converting to pfn. At least for now this should work.
Yes, this should not matter as we unpin it on a page base anyway.
>
>> + unpin_guest_page(vcpu->kvm, gpa, hpa);
>> + scb_s->sdnxo = 0;
>> + }
>> }
>>
>> /*
>> @@ -590,6 +602,33 @@ static int pin_blocks(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page)
>> goto unpin;
>> scb_s->riccbd = hpa;
>> }
>> + if ((scb_s->ecb & ECB_GS) && !(scb_s->ecd & ECD_HOSTREGMGMT)) {
>> + unsigned long sdnxc;
>> +
>> + gpa = scb_o->sdnxo & ~0xfUL;
>> + sdnxc = scb_o->sdnxo & 0xfUL;
>> + if (!gpa || !(gpa & ~0x1fffUL)) {
>> + rc = set_validity_icpt(scb_s, 0x10b0U);
>> + goto unpin;
>> + }
>> + if (sdnxc < 6 || sdnxc > 12) {
>> + rc = set_validity_icpt(scb_s, 0x10b1U);
>> + goto unpin;
>> + }
>> + if (gpa & ((1 << sdnxc) - 1)) {
>> + rc = set_validity_icpt(scb_s, 0x10b2U);
>> + goto unpin;
>> + }
>> + /* Due to alignment rules (checked above) this cannot
>> + * cross page boundaries
>> + */
>> + rc = pin_guest_page(vcpu->kvm, gpa, &hpa);
>> + if (rc == -EINVAL)
>> + rc = set_validity_icpt(scb_s, 0x10b0U);
>> + if (rc)
>> + goto unpin;
>> + scb_s->sdnxo = hpa;
>
> I'm curious, this will result in no sdnxc getting set. Shouldn't there
> be something like the following, cause otherwise there would be a
> validity icpt 0x10b1U (guessing from the code above).
>
> scb_s->sdnxo = hpa | sdnxc;
Indeed this looks like a bug (and the right fix).
This was not noticed during test since KVM enables the host management of
registers and we do not use the sdnx. So for the KVM under KVM case this
does not matter. Still, we should fix it.
Radim, Paolo, do you want a respin of the whole series or can I simply send
a fixup patch with the next round of patches?
>
>> + }
>> return 0;
>> unpin:
>> unpin_blocks(vcpu, vsie_page);
>> diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h
>> index f51d508..c9d5227 100644
>> --- a/include/uapi/linux/kvm.h
>> +++ b/include/uapi/linux/kvm.h
>> @@ -883,6 +883,7 @@ struct kvm_ppc_resize_hpt {
>> #define KVM_CAP_PPC_MMU_RADIX 134
>> #define KVM_CAP_PPC_MMU_HASH_V3 135
>> #define KVM_CAP_IMMEDIATE_EXIT 136
>> +#define KVM_CAP_S390_GS 137
>>
>> #ifdef KVM_CAP_IRQ_ROUTING
>>
>>
>
>
^ permalink raw reply [flat|nested] 17+ messages in thread* Re: [GIT PULL 5/9] KVM: s390: gs support for kvm guests
2017-04-07 12:01 ` Christian Borntraeger
@ 2017-04-11 12:42 ` Christian Borntraeger
2017-04-11 14:04 ` David Hildenbrand
2017-04-11 19:23 ` Radim Krčmář
0 siblings, 2 replies; 17+ messages in thread
From: Christian Borntraeger @ 2017-04-11 12:42 UTC (permalink / raw)
To: David Hildenbrand, Paolo Bonzini, Radim Krčmář
Cc: KVM, Cornelia Huck, linux-s390, Fan Zhang
On 04/07/2017 02:01 PM, Christian Borntraeger wrote:
>
> Indeed this looks like a bug (and the right fix).
>
> This was not noticed during test since KVM enables the host management of
> registers and we do not use the sdnx. So for the KVM under KVM case this
> does not matter. Still, we should fix it.
>
> Radim, Paolo, do you want a respin of the whole series or can I simply send
> a fixup patch with the next round of patches?
Radim, Paolo,
is it ok to schedule the fixup patch for the next round of patches(I plan to
have some more for 4.12 anyway) and apply this series as is?
^ permalink raw reply [flat|nested] 17+ messages in thread
* Re: [GIT PULL 5/9] KVM: s390: gs support for kvm guests
2017-04-11 12:42 ` Christian Borntraeger
@ 2017-04-11 14:04 ` David Hildenbrand
2017-04-11 19:23 ` Radim Krčmář
1 sibling, 0 replies; 17+ messages in thread
From: David Hildenbrand @ 2017-04-11 14:04 UTC (permalink / raw)
To: Christian Borntraeger, Paolo Bonzini, Radim Krčmář
Cc: KVM, Cornelia Huck, linux-s390, Fan Zhang
On 11.04.2017 14:42, Christian Borntraeger wrote:
> On 04/07/2017 02:01 PM, Christian Borntraeger wrote:
>>
>> Indeed this looks like a bug (and the right fix).
>>
>> This was not noticed during test since KVM enables the host management of
>> registers and we do not use the sdnx. So for the KVM under KVM case this
>> does not matter. Still, we should fix it.
>>
>> Radim, Paolo, do you want a respin of the whole series or can I simply send
>> a fixup patch with the next round of patches?
>
> Radim, Paolo,
>
> is it ok to schedule the fixup patch for the next round of patches(I plan to
> have some more for 4.12 anyway) and apply this series as is?
This fix really has low priority (hw not available, KVM not using this
feature), so I think this fixup can wait.
--
Thanks,
David
^ permalink raw reply [flat|nested] 17+ messages in thread
* Re: [GIT PULL 5/9] KVM: s390: gs support for kvm guests
2017-04-11 12:42 ` Christian Borntraeger
2017-04-11 14:04 ` David Hildenbrand
@ 2017-04-11 19:23 ` Radim Krčmář
1 sibling, 0 replies; 17+ messages in thread
From: Radim Krčmář @ 2017-04-11 19:23 UTC (permalink / raw)
To: Christian Borntraeger
Cc: David Hildenbrand, Paolo Bonzini, KVM, Cornelia Huck, linux-s390,
Fan Zhang
2017-04-11 14:42+0200, Christian Borntraeger:
> On 04/07/2017 02:01 PM, Christian Borntraeger wrote:
> >
> > Indeed this looks like a bug (and the right fix).
> >
> > This was not noticed during test since KVM enables the host management of
> > registers and we do not use the sdnx. So for the KVM under KVM case this
> > does not matter. Still, we should fix it.
> >
> > Radim, Paolo, do you want a respin of the whole series or can I simply send
> > a fixup patch with the next round of patches?
>
> Radim, Paolo,
>
> is it ok to schedule the fixup patch for the next round of patches(I plan to
> have some more for 4.12 anyway) and apply this series as is?
I have not noticed the bottom hunk on Friday and waited before asking,
sorry. Pulled now, the new capability numbers are
#define KVM_CAP_S390_GS 140
#define KVM_CAP_S390_AIS 141
Thanks.
^ permalink raw reply [flat|nested] 17+ messages in thread
* [PATCH] KVM: s390: Fix sdnxo setting for nested guests
2017-04-07 11:42 ` David Hildenbrand
2017-04-07 12:01 ` Christian Borntraeger
@ 2017-04-07 12:27 ` Christian Borntraeger
2017-04-11 11:55 ` Cornelia Huck
1 sibling, 1 reply; 17+ messages in thread
From: Christian Borntraeger @ 2017-04-07 12:27 UTC (permalink / raw)
To: Paolo Bonzini, Radim Krčmář
Cc: KVM, Christian Borntraeger, Cornelia Huck, linux-s390, David Hildenbrand
If the guest does not use the host register management, but it uses
the sdnx area, we must fill in a proper sdnxo value (address of sdnx
and the sdnxc).
Reported-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
---
arch/s390/kvm/vsie.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/s390/kvm/vsie.c b/arch/s390/kvm/vsie.c
index 2fafc2b..025b1f2 100644
--- a/arch/s390/kvm/vsie.c
+++ b/arch/s390/kvm/vsie.c
@@ -627,7 +627,7 @@ static int pin_blocks(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page)
rc = set_validity_icpt(scb_s, 0x10b0U);
if (rc)
goto unpin;
- scb_s->sdnxo = hpa;
+ scb_s->sdnxo = hpa | sdnxc;
}
return 0;
unpin:
--
2.7.4
^ permalink raw reply related [flat|nested] 17+ messages in thread
* Re: [PATCH] KVM: s390: Fix sdnxo setting for nested guests
2017-04-07 12:27 ` [PATCH] KVM: s390: Fix sdnxo setting for nested guests Christian Borntraeger
@ 2017-04-11 11:55 ` Cornelia Huck
0 siblings, 0 replies; 17+ messages in thread
From: Cornelia Huck @ 2017-04-11 11:55 UTC (permalink / raw)
To: Christian Borntraeger
Cc: Paolo Bonzini, Radim Krčmář,
KVM, linux-s390, David Hildenbrand
On Fri, 7 Apr 2017 14:27:45 +0200
Christian Borntraeger <borntraeger@de.ibm.com> wrote:
> If the guest does not use the host register management, but it uses
> the sdnx area, we must fill in a proper sdnxo value (address of sdnx
> and the sdnxc).
>
> Reported-by: David Hildenbrand <david@redhat.com>
> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
> ---
> arch/s390/kvm/vsie.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/arch/s390/kvm/vsie.c b/arch/s390/kvm/vsie.c
> index 2fafc2b..025b1f2 100644
> --- a/arch/s390/kvm/vsie.c
> +++ b/arch/s390/kvm/vsie.c
> @@ -627,7 +627,7 @@ static int pin_blocks(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page)
> rc = set_validity_icpt(scb_s, 0x10b0U);
> if (rc)
> goto unpin;
> - scb_s->sdnxo = hpa;
> + scb_s->sdnxo = hpa | sdnxc;
> }
> return 0;
> unpin:
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
^ permalink raw reply [flat|nested] 17+ messages in thread
* [GIT PULL 6/9] KVM: s390: interface for suppressible I/O adapters
2017-04-07 10:47 [GIT PULL 0/9] KVM: s390: features for 4.12 (via kvm/next) Christian Borntraeger
` (4 preceding siblings ...)
2017-04-07 10:47 ` [GIT PULL 5/9] KVM: s390: gs support for kvm guests Christian Borntraeger
@ 2017-04-07 10:47 ` Christian Borntraeger
2017-04-07 10:48 ` [GIT PULL 7/9] KVM: s390: introduce ais mode modify function Christian Borntraeger
` (2 subsequent siblings)
8 siblings, 0 replies; 17+ messages in thread
From: Christian Borntraeger @ 2017-04-07 10:47 UTC (permalink / raw)
To: Paolo Bonzini, Radim Krčmář
Cc: KVM, Christian Borntraeger, Cornelia Huck, linux-s390, Fei Li
From: Fei Li <sherrylf@linux.vnet.ibm.com>
In order to properly implement adapter-interruption suppression, we
need a way for userspace to specify which adapters are subject to
suppression. Let's convert the existing (and unused) 'pad' field into
a 'flags' field and define a flag value for suppressible adapters.
Besides, add documentation for the interface.
Signed-off-by: Fei Li <sherrylf@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
---
Documentation/virtual/kvm/devices/s390_flic.txt | 12 +++++++++---
arch/s390/include/asm/kvm_host.h | 1 +
arch/s390/include/uapi/asm/kvm.h | 4 +++-
arch/s390/kvm/interrupt.c | 2 ++
4 files changed, 15 insertions(+), 4 deletions(-)
diff --git a/Documentation/virtual/kvm/devices/s390_flic.txt b/Documentation/virtual/kvm/devices/s390_flic.txt
index 6b0e115..e8ee3b6 100644
--- a/Documentation/virtual/kvm/devices/s390_flic.txt
+++ b/Documentation/virtual/kvm/devices/s390_flic.txt
@@ -64,12 +64,18 @@ struct kvm_s390_io_adapter {
__u8 isc;
__u8 maskable;
__u8 swap;
- __u8 pad;
+ __u8 flags;
};
id contains the unique id for the adapter, isc the I/O interruption subclass
- to use, maskable whether this adapter may be masked (interrupts turned off)
- and swap whether the indicators need to be byte swapped.
+ to use, maskable whether this adapter may be masked (interrupts turned off),
+ swap whether the indicators need to be byte swapped, and flags contains
+ further characteristics of the adapter.
+ Currently defined values for 'flags' are:
+ - KVM_S390_ADAPTER_SUPPRESSIBLE: adapter is subject to AIS
+ (adapter-interrupt-suppression) facility. This flag only has an effect if
+ the AIS capability is enabled.
+ Unknown flag values are ignored.
KVM_DEV_FLIC_ADAPTER_MODIFY
diff --git a/arch/s390/include/asm/kvm_host.h b/arch/s390/include/asm/kvm_host.h
index 1af090d..499c72c 100644
--- a/arch/s390/include/asm/kvm_host.h
+++ b/arch/s390/include/asm/kvm_host.h
@@ -621,6 +621,7 @@ struct s390_io_adapter {
bool maskable;
bool masked;
bool swap;
+ bool suppressible;
struct rw_semaphore maps_lock;
struct list_head maps;
atomic_t nr_maps;
diff --git a/arch/s390/include/uapi/asm/kvm.h b/arch/s390/include/uapi/asm/kvm.h
index 5bd23cf..5fa144d 100644
--- a/arch/s390/include/uapi/asm/kvm.h
+++ b/arch/s390/include/uapi/asm/kvm.h
@@ -41,9 +41,11 @@ struct kvm_s390_io_adapter {
__u8 isc;
__u8 maskable;
__u8 swap;
- __u8 pad;
+ __u8 flags;
};
+#define KVM_S390_ADAPTER_SUPPRESSIBLE 0x01
+
#define KVM_S390_IO_ADAPTER_MASK 1
#define KVM_S390_IO_ADAPTER_MAP 2
#define KVM_S390_IO_ADAPTER_UNMAP 3
diff --git a/arch/s390/kvm/interrupt.c b/arch/s390/kvm/interrupt.c
index 311eef0..dba51ad 100644
--- a/arch/s390/kvm/interrupt.c
+++ b/arch/s390/kvm/interrupt.c
@@ -1997,6 +1997,8 @@ static int register_io_adapter(struct kvm_device *dev,
adapter->maskable = adapter_info.maskable;
adapter->masked = false;
adapter->swap = adapter_info.swap;
+ adapter->suppressible = (adapter_info.flags) &
+ KVM_S390_ADAPTER_SUPPRESSIBLE;
dev->kvm->arch.adapters[adapter->id] = adapter;
return 0;
--
2.7.4
^ permalink raw reply related [flat|nested] 17+ messages in thread* [GIT PULL 7/9] KVM: s390: introduce ais mode modify function
2017-04-07 10:47 [GIT PULL 0/9] KVM: s390: features for 4.12 (via kvm/next) Christian Borntraeger
` (5 preceding siblings ...)
2017-04-07 10:47 ` [GIT PULL 6/9] KVM: s390: interface for suppressible I/O adapters Christian Borntraeger
@ 2017-04-07 10:48 ` Christian Borntraeger
2017-04-07 10:48 ` [GIT PULL 8/9] KVM: s390: introduce adapter interrupt inject function Christian Borntraeger
2017-04-07 10:48 ` [GIT PULL 9/9] KVM: s390: introduce AIS capability Christian Borntraeger
8 siblings, 0 replies; 17+ messages in thread
From: Christian Borntraeger @ 2017-04-07 10:48 UTC (permalink / raw)
To: Paolo Bonzini, Radim Krčmář
Cc: KVM, Christian Borntraeger, Cornelia Huck, linux-s390, Fei Li,
Yi Min Zhao
From: Fei Li <sherrylf@linux.vnet.ibm.com>
Provide an interface for userspace to modify AIS
(adapter-interruption-suppression) mode state, and add documentation
for the interface. Allowed target modes are ALL-Interruptions mode
and SINGLE-Interruption mode.
We introduce the 'simm' and 'nimm' fields in kvm_s390_float_interrupt
to store interruption modes for each ISC. Each bit in 'simm' and
'nimm' targets to one ISC, and collaboratively indicate three modes:
ALL-Interruptions, SINGLE-Interruption and NO-Interruptions. This
interface can initiate most transitions between the states; transition
from SINGLE-Interruption to NO-Interruptions via adapter interrupt
injection will be introduced in a following patch. The meaningful
combinations are as follows:
interruption mode | simm bit | nimm bit
------------------|----------|----------
ALL | 0 | 0
SINGLE | 1 | 0
NO | 1 | 1
Besides, add tracepoint to track AIS mode transitions.
Co-Authored-By: Yi Min Zhao <zyimin@linux.vnet.ibm.com>
Signed-off-by: Yi Min Zhao <zyimin@linux.vnet.ibm.com>
Signed-off-by: Fei Li <sherrylf@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
---
Documentation/virtual/kvm/devices/s390_flic.txt | 20 ++++++++++++
arch/s390/include/asm/kvm_host.h | 10 ++++++
arch/s390/include/uapi/asm/kvm.h | 6 ++++
arch/s390/kvm/interrupt.c | 43 +++++++++++++++++++++++++
arch/s390/kvm/kvm-s390.c | 4 +++
arch/s390/kvm/trace-s390.h | 31 ++++++++++++++++++
6 files changed, 114 insertions(+)
diff --git a/Documentation/virtual/kvm/devices/s390_flic.txt b/Documentation/virtual/kvm/devices/s390_flic.txt
index e8ee3b6..dfd42fd 100644
--- a/Documentation/virtual/kvm/devices/s390_flic.txt
+++ b/Documentation/virtual/kvm/devices/s390_flic.txt
@@ -14,6 +14,7 @@ FLIC provides support to
- purge one pending floating I/O interrupt (KVM_DEV_FLIC_CLEAR_IO_IRQ)
- enable/disable for the guest transparent async page faults
- register and modify adapter interrupt sources (KVM_DEV_FLIC_ADAPTER_*)
+- modify AIS (adapter-interruption-suppression) mode state (KVM_DEV_FLIC_AISM)
Groups:
KVM_DEV_FLIC_ENQUEUE
@@ -107,6 +108,25 @@ struct kvm_s390_io_adapter_req {
release a userspace page for the translated address specified in addr
from the list of mappings
+ KVM_DEV_FLIC_AISM
+ modify the adapter-interruption-suppression mode for a given isc if the
+ AIS capability is enabled. Takes a kvm_s390_ais_req describing:
+
+struct kvm_s390_ais_req {
+ __u8 isc;
+ __u16 mode;
+};
+
+ isc contains the target I/O interruption subclass, mode the target
+ adapter-interruption-suppression mode. The following modes are
+ currently supported:
+ - KVM_S390_AIS_MODE_ALL: ALL-Interruptions Mode, i.e. airq injection
+ is always allowed;
+ - KVM_S390_AIS_MODE_SINGLE: SINGLE-Interruption Mode, i.e. airq
+ injection is only allowed once and the following adapter interrupts
+ will be suppressed until the mode is set again to ALL-Interruptions
+ or SINGLE-Interruption mode.
+
Note: The KVM_SET_DEVICE_ATTR/KVM_GET_DEVICE_ATTR device ioctls executed on
FLIC with an unknown group or attribute gives the error code EINVAL (instead of
ENXIO, as specified in the API documentation). It is not possible to conclude
diff --git a/arch/s390/include/asm/kvm_host.h b/arch/s390/include/asm/kvm_host.h
index 499c72c..552c319 100644
--- a/arch/s390/include/asm/kvm_host.h
+++ b/arch/s390/include/asm/kvm_host.h
@@ -521,6 +521,12 @@ struct kvm_s390_local_interrupt {
#define FIRQ_CNTR_PFAULT 3
#define FIRQ_MAX_COUNT 4
+/* mask the AIS mode for a given ISC */
+#define AIS_MODE_MASK(isc) (0x80 >> isc)
+
+#define KVM_S390_AIS_MODE_ALL 0
+#define KVM_S390_AIS_MODE_SINGLE 1
+
struct kvm_s390_float_interrupt {
unsigned long pending_irqs;
spinlock_t lock;
@@ -530,6 +536,10 @@ struct kvm_s390_float_interrupt {
struct kvm_s390_ext_info srv_signal;
int next_rr_cpu;
unsigned long idle_mask[BITS_TO_LONGS(KVM_MAX_VCPUS)];
+ struct mutex ais_lock;
+ u8 simm;
+ u8 nimm;
+ int ais_enabled;
};
struct kvm_hw_wp_info_arch {
diff --git a/arch/s390/include/uapi/asm/kvm.h b/arch/s390/include/uapi/asm/kvm.h
index 5fa144d..50d2a92 100644
--- a/arch/s390/include/uapi/asm/kvm.h
+++ b/arch/s390/include/uapi/asm/kvm.h
@@ -26,6 +26,7 @@
#define KVM_DEV_FLIC_ADAPTER_REGISTER 6
#define KVM_DEV_FLIC_ADAPTER_MODIFY 7
#define KVM_DEV_FLIC_CLEAR_IO_IRQ 8
+#define KVM_DEV_FLIC_AISM 9
/*
* We can have up to 4*64k pending subchannels + 8 adapter interrupts,
* as well as up to ASYNC_PF_PER_VCPU*KVM_MAX_VCPUS pfault done interrupts.
@@ -46,6 +47,11 @@ struct kvm_s390_io_adapter {
#define KVM_S390_ADAPTER_SUPPRESSIBLE 0x01
+struct kvm_s390_ais_req {
+ __u8 isc;
+ __u16 mode;
+};
+
#define KVM_S390_IO_ADAPTER_MASK 1
#define KVM_S390_IO_ADAPTER_MAP 2
#define KVM_S390_IO_ADAPTER_UNMAP 3
diff --git a/arch/s390/kvm/interrupt.c b/arch/s390/kvm/interrupt.c
index dba51ad..96b689e 100644
--- a/arch/s390/kvm/interrupt.c
+++ b/arch/s390/kvm/interrupt.c
@@ -2152,6 +2152,45 @@ static int clear_io_irq(struct kvm *kvm, struct kvm_device_attr *attr)
return 0;
}
+static int modify_ais_mode(struct kvm *kvm, struct kvm_device_attr *attr)
+{
+ struct kvm_s390_float_interrupt *fi = &kvm->arch.float_int;
+ struct kvm_s390_ais_req req;
+ int ret = 0;
+
+ if (!fi->ais_enabled)
+ return -ENOTSUPP;
+
+ if (copy_from_user(&req, (void __user *)attr->addr, sizeof(req)))
+ return -EFAULT;
+
+ if (req.isc > MAX_ISC)
+ return -EINVAL;
+
+ trace_kvm_s390_modify_ais_mode(req.isc,
+ (fi->simm & AIS_MODE_MASK(req.isc)) ?
+ (fi->nimm & AIS_MODE_MASK(req.isc)) ?
+ 2 : KVM_S390_AIS_MODE_SINGLE :
+ KVM_S390_AIS_MODE_ALL, req.mode);
+
+ mutex_lock(&fi->ais_lock);
+ switch (req.mode) {
+ case KVM_S390_AIS_MODE_ALL:
+ fi->simm &= ~AIS_MODE_MASK(req.isc);
+ fi->nimm &= ~AIS_MODE_MASK(req.isc);
+ break;
+ case KVM_S390_AIS_MODE_SINGLE:
+ fi->simm |= AIS_MODE_MASK(req.isc);
+ fi->nimm &= ~AIS_MODE_MASK(req.isc);
+ break;
+ default:
+ ret = -EINVAL;
+ }
+ mutex_unlock(&fi->ais_lock);
+
+ return ret;
+}
+
static int flic_set_attr(struct kvm_device *dev, struct kvm_device_attr *attr)
{
int r = 0;
@@ -2188,6 +2227,9 @@ static int flic_set_attr(struct kvm_device *dev, struct kvm_device_attr *attr)
case KVM_DEV_FLIC_CLEAR_IO_IRQ:
r = clear_io_irq(dev->kvm, attr);
break;
+ case KVM_DEV_FLIC_AISM:
+ r = modify_ais_mode(dev->kvm, attr);
+ break;
default:
r = -EINVAL;
}
@@ -2207,6 +2249,7 @@ static int flic_has_attr(struct kvm_device *dev,
case KVM_DEV_FLIC_ADAPTER_REGISTER:
case KVM_DEV_FLIC_ADAPTER_MODIFY:
case KVM_DEV_FLIC_CLEAR_IO_IRQ:
+ case KVM_DEV_FLIC_AISM:
return 0;
}
return -ENXIO;
diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
index f83f18b..977cc16 100644
--- a/arch/s390/kvm/kvm-s390.c
+++ b/arch/s390/kvm/kvm-s390.c
@@ -1515,6 +1515,10 @@ int kvm_arch_init_vm(struct kvm *kvm, unsigned long type)
kvm_s390_crypto_init(kvm);
+ mutex_init(&kvm->arch.float_int.ais_lock);
+ kvm->arch.float_int.simm = 0;
+ kvm->arch.float_int.nimm = 0;
+ kvm->arch.float_int.ais_enabled = 0;
spin_lock_init(&kvm->arch.float_int.lock);
for (i = 0; i < FIRQ_LIST_COUNT; i++)
INIT_LIST_HEAD(&kvm->arch.float_int.lists[i]);
diff --git a/arch/s390/kvm/trace-s390.h b/arch/s390/kvm/trace-s390.h
index 396485b..b32994d 100644
--- a/arch/s390/kvm/trace-s390.h
+++ b/arch/s390/kvm/trace-s390.h
@@ -280,6 +280,37 @@ TRACE_EVENT(kvm_s390_enable_disable_ibs,
__entry->state ? "enabling" : "disabling", __entry->id)
);
+/*
+ * Trace point for modifying ais mode for a given isc.
+ */
+TRACE_EVENT(kvm_s390_modify_ais_mode,
+ TP_PROTO(__u8 isc, __u16 from, __u16 to),
+ TP_ARGS(isc, from, to),
+
+ TP_STRUCT__entry(
+ __field(__u8, isc)
+ __field(__u16, from)
+ __field(__u16, to)
+ ),
+
+ TP_fast_assign(
+ __entry->isc = isc;
+ __entry->from = from;
+ __entry->to = to;
+ ),
+
+ TP_printk("for isc %x, modifying interruption mode from %s to %s",
+ __entry->isc,
+ (__entry->from == KVM_S390_AIS_MODE_ALL) ?
+ "ALL-Interruptions Mode" :
+ (__entry->from == KVM_S390_AIS_MODE_SINGLE) ?
+ "Single-Interruption Mode" : "No-Interruptions Mode",
+ (__entry->to == KVM_S390_AIS_MODE_ALL) ?
+ "ALL-Interruptions Mode" :
+ (__entry->to == KVM_S390_AIS_MODE_SINGLE) ?
+ "Single-Interruption Mode" : "No-Interruptions Mode")
+ );
+
#endif /* _TRACE_KVMS390_H */
--
2.7.4
^ permalink raw reply related [flat|nested] 17+ messages in thread* [GIT PULL 8/9] KVM: s390: introduce adapter interrupt inject function
2017-04-07 10:47 [GIT PULL 0/9] KVM: s390: features for 4.12 (via kvm/next) Christian Borntraeger
` (6 preceding siblings ...)
2017-04-07 10:48 ` [GIT PULL 7/9] KVM: s390: introduce ais mode modify function Christian Borntraeger
@ 2017-04-07 10:48 ` Christian Borntraeger
2017-04-07 10:48 ` [GIT PULL 9/9] KVM: s390: introduce AIS capability Christian Borntraeger
8 siblings, 0 replies; 17+ messages in thread
From: Christian Borntraeger @ 2017-04-07 10:48 UTC (permalink / raw)
To: Paolo Bonzini, Radim Krčmář
Cc: KVM, Christian Borntraeger, Cornelia Huck, linux-s390,
Yi Min Zhao, Fei Li
From: Yi Min Zhao <zyimin@linux.vnet.ibm.com>
Inject adapter interrupts on a specified adapter which allows to
retrieve the adapter flags, e.g. if the adapter is subject to AIS
facility or not. And add documentation for this interface.
For adapters subject to AIS, handle the airq injection suppression
for a given ISC according to the interruption mode:
- before injection, if NO-Interruptions Mode, just return 0 and
suppress, otherwise, allow the injection.
- after injection, if SINGLE-Interruption Mode, change it to
NO-Interruptions Mode to suppress the following interrupts.
Besides, add tracepoint for suppressed airq and AIS mode transitions.
Signed-off-by: Yi Min Zhao <zyimin@linux.vnet.ibm.com>
Signed-off-by: Fei Li <sherrylf@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
---
Documentation/virtual/kvm/devices/s390_flic.txt | 9 +++++
arch/s390/include/uapi/asm/kvm.h | 1 +
arch/s390/kvm/interrupt.c | 53 ++++++++++++++++++++++---
arch/s390/kvm/trace-s390.h | 21 ++++++++++
4 files changed, 78 insertions(+), 6 deletions(-)
diff --git a/Documentation/virtual/kvm/devices/s390_flic.txt b/Documentation/virtual/kvm/devices/s390_flic.txt
index dfd42fd..c2518ce 100644
--- a/Documentation/virtual/kvm/devices/s390_flic.txt
+++ b/Documentation/virtual/kvm/devices/s390_flic.txt
@@ -15,6 +15,7 @@ FLIC provides support to
- enable/disable for the guest transparent async page faults
- register and modify adapter interrupt sources (KVM_DEV_FLIC_ADAPTER_*)
- modify AIS (adapter-interruption-suppression) mode state (KVM_DEV_FLIC_AISM)
+- inject adapter interrupts on a specified adapter (KVM_DEV_FLIC_AIRQ_INJECT)
Groups:
KVM_DEV_FLIC_ENQUEUE
@@ -127,6 +128,14 @@ struct kvm_s390_ais_req {
will be suppressed until the mode is set again to ALL-Interruptions
or SINGLE-Interruption mode.
+ KVM_DEV_FLIC_AIRQ_INJECT
+ Inject adapter interrupts on a specified adapter.
+ attr->attr contains the unique id for the adapter, which allows for
+ adapter-specific checks and actions.
+ For adapters subject to AIS, handle the airq injection suppression for
+ an isc according to the adapter-interruption-suppression mode on condition
+ that the AIS capability is enabled.
+
Note: The KVM_SET_DEVICE_ATTR/KVM_GET_DEVICE_ATTR device ioctls executed on
FLIC with an unknown group or attribute gives the error code EINVAL (instead of
ENXIO, as specified in the API documentation). It is not possible to conclude
diff --git a/arch/s390/include/uapi/asm/kvm.h b/arch/s390/include/uapi/asm/kvm.h
index 50d2a92..2c9ad25 100644
--- a/arch/s390/include/uapi/asm/kvm.h
+++ b/arch/s390/include/uapi/asm/kvm.h
@@ -27,6 +27,7 @@
#define KVM_DEV_FLIC_ADAPTER_MODIFY 7
#define KVM_DEV_FLIC_CLEAR_IO_IRQ 8
#define KVM_DEV_FLIC_AISM 9
+#define KVM_DEV_FLIC_AIRQ_INJECT 10
/*
* We can have up to 4*64k pending subchannels + 8 adapter interrupts,
* as well as up to ASYNC_PF_PER_VCPU*KVM_MAX_VCPUS pfault done interrupts.
diff --git a/arch/s390/kvm/interrupt.c b/arch/s390/kvm/interrupt.c
index 96b689e..482673e 100644
--- a/arch/s390/kvm/interrupt.c
+++ b/arch/s390/kvm/interrupt.c
@@ -2191,6 +2191,48 @@ static int modify_ais_mode(struct kvm *kvm, struct kvm_device_attr *attr)
return ret;
}
+static int kvm_s390_inject_airq(struct kvm *kvm,
+ struct s390_io_adapter *adapter)
+{
+ struct kvm_s390_float_interrupt *fi = &kvm->arch.float_int;
+ struct kvm_s390_interrupt s390int = {
+ .type = KVM_S390_INT_IO(1, 0, 0, 0),
+ .parm = 0,
+ .parm64 = (adapter->isc << 27) | 0x80000000,
+ };
+ int ret = 0;
+
+ if (!fi->ais_enabled || !adapter->suppressible)
+ return kvm_s390_inject_vm(kvm, &s390int);
+
+ mutex_lock(&fi->ais_lock);
+ if (fi->nimm & AIS_MODE_MASK(adapter->isc)) {
+ trace_kvm_s390_airq_suppressed(adapter->id, adapter->isc);
+ goto out;
+ }
+
+ ret = kvm_s390_inject_vm(kvm, &s390int);
+ if (!ret && (fi->simm & AIS_MODE_MASK(adapter->isc))) {
+ fi->nimm |= AIS_MODE_MASK(adapter->isc);
+ trace_kvm_s390_modify_ais_mode(adapter->isc,
+ KVM_S390_AIS_MODE_SINGLE, 2);
+ }
+out:
+ mutex_unlock(&fi->ais_lock);
+ return ret;
+}
+
+static int flic_inject_airq(struct kvm *kvm, struct kvm_device_attr *attr)
+{
+ unsigned int id = attr->attr;
+ struct s390_io_adapter *adapter = get_io_adapter(kvm, id);
+
+ if (!adapter)
+ return -EINVAL;
+
+ return kvm_s390_inject_airq(kvm, adapter);
+}
+
static int flic_set_attr(struct kvm_device *dev, struct kvm_device_attr *attr)
{
int r = 0;
@@ -2230,6 +2272,9 @@ static int flic_set_attr(struct kvm_device *dev, struct kvm_device_attr *attr)
case KVM_DEV_FLIC_AISM:
r = modify_ais_mode(dev->kvm, attr);
break;
+ case KVM_DEV_FLIC_AIRQ_INJECT:
+ r = flic_inject_airq(dev->kvm, attr);
+ break;
default:
r = -EINVAL;
}
@@ -2250,6 +2295,7 @@ static int flic_has_attr(struct kvm_device *dev,
case KVM_DEV_FLIC_ADAPTER_MODIFY:
case KVM_DEV_FLIC_CLEAR_IO_IRQ:
case KVM_DEV_FLIC_AISM:
+ case KVM_DEV_FLIC_AIRQ_INJECT:
return 0;
}
return -ENXIO;
@@ -2360,12 +2406,7 @@ static int set_adapter_int(struct kvm_kernel_irq_routing_entry *e,
ret = adapter_indicators_set(kvm, adapter, &e->adapter);
up_read(&adapter->maps_lock);
if ((ret > 0) && !adapter->masked) {
- struct kvm_s390_interrupt s390int = {
- .type = KVM_S390_INT_IO(1, 0, 0, 0),
- .parm = 0,
- .parm64 = (adapter->isc << 27) | 0x80000000,
- };
- ret = kvm_s390_inject_vm(kvm, &s390int);
+ ret = kvm_s390_inject_airq(kvm, adapter);
if (ret == 0)
ret = 1;
}
diff --git a/arch/s390/kvm/trace-s390.h b/arch/s390/kvm/trace-s390.h
index b32994d..78b7e84 100644
--- a/arch/s390/kvm/trace-s390.h
+++ b/arch/s390/kvm/trace-s390.h
@@ -311,6 +311,27 @@ TRACE_EVENT(kvm_s390_modify_ais_mode,
"Single-Interruption Mode" : "No-Interruptions Mode")
);
+/*
+ * Trace point for suppressed adapter I/O interrupt.
+ */
+TRACE_EVENT(kvm_s390_airq_suppressed,
+ TP_PROTO(__u32 id, __u8 isc),
+ TP_ARGS(id, isc),
+
+ TP_STRUCT__entry(
+ __field(__u32, id)
+ __field(__u8, isc)
+ ),
+
+ TP_fast_assign(
+ __entry->id = id;
+ __entry->isc = isc;
+ ),
+
+ TP_printk("adapter I/O interrupt suppressed (id:%x isc:%x)",
+ __entry->id, __entry->isc)
+ );
+
#endif /* _TRACE_KVMS390_H */
--
2.7.4
^ permalink raw reply related [flat|nested] 17+ messages in thread* [GIT PULL 9/9] KVM: s390: introduce AIS capability
2017-04-07 10:47 [GIT PULL 0/9] KVM: s390: features for 4.12 (via kvm/next) Christian Borntraeger
` (7 preceding siblings ...)
2017-04-07 10:48 ` [GIT PULL 8/9] KVM: s390: introduce adapter interrupt inject function Christian Borntraeger
@ 2017-04-07 10:48 ` Christian Borntraeger
8 siblings, 0 replies; 17+ messages in thread
From: Christian Borntraeger @ 2017-04-07 10:48 UTC (permalink / raw)
To: Paolo Bonzini, Radim Krčmář
Cc: KVM, Christian Borntraeger, Cornelia Huck, linux-s390,
Yi Min Zhao, Fei Li
From: Yi Min Zhao <zyimin@linux.vnet.ibm.com>
Introduce a cap to enable AIS facility bit, and add documentation
for this capability.
Signed-off-by: Yi Min Zhao <zyimin@linux.vnet.ibm.com>
Signed-off-by: Fei Li <sherrylf@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
---
Documentation/virtual/kvm/api.txt | 8 ++++++++
arch/s390/kvm/kvm-s390.c | 15 +++++++++++++++
include/uapi/linux/kvm.h | 1 +
3 files changed, 24 insertions(+)
diff --git a/Documentation/virtual/kvm/api.txt b/Documentation/virtual/kvm/api.txt
index 7252508..598278c 100644
--- a/Documentation/virtual/kvm/api.txt
+++ b/Documentation/virtual/kvm/api.txt
@@ -4110,6 +4110,14 @@ Returns: 0 on success; -EINVAL if the machine does not support
Allows use of guarded storage for the KVM guest.
+7.10 KVM_CAP_S390_AIS
+
+Architectures: s390
+Parameters: none
+
+Allow use of adapter-interruption suppression.
+Returns: 0 on success; -EBUSY if a VCPU has already been created.
+
8. Other capabilities.
----------------------
diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
index 977cc16..11b7d66 100644
--- a/arch/s390/kvm/kvm-s390.c
+++ b/arch/s390/kvm/kvm-s390.c
@@ -380,6 +380,7 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
case KVM_CAP_S390_SKEYS:
case KVM_CAP_S390_IRQ_STATE:
case KVM_CAP_S390_USER_INSTR0:
+ case KVM_CAP_S390_AIS:
r = 1;
break;
case KVM_CAP_S390_MEM_OP:
@@ -544,6 +545,20 @@ static int kvm_vm_ioctl_enable_cap(struct kvm *kvm, struct kvm_enable_cap *cap)
VM_EVENT(kvm, 3, "ENABLE: CAP_S390_RI %s",
r ? "(not available)" : "(success)");
break;
+ case KVM_CAP_S390_AIS:
+ mutex_lock(&kvm->lock);
+ if (kvm->created_vcpus) {
+ r = -EBUSY;
+ } else {
+ set_kvm_facility(kvm->arch.model.fac_mask, 72);
+ set_kvm_facility(kvm->arch.model.fac_list, 72);
+ kvm->arch.float_int.ais_enabled = 1;
+ r = 0;
+ }
+ mutex_unlock(&kvm->lock);
+ VM_EVENT(kvm, 3, "ENABLE: AIS %s",
+ r ? "(not available)" : "(success)");
+ break;
case KVM_CAP_S390_GS:
r = -EINVAL;
mutex_lock(&kvm->lock);
diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h
index c9d5227..33dd2a4 100644
--- a/include/uapi/linux/kvm.h
+++ b/include/uapi/linux/kvm.h
@@ -884,6 +884,7 @@ struct kvm_ppc_resize_hpt {
#define KVM_CAP_PPC_MMU_HASH_V3 135
#define KVM_CAP_IMMEDIATE_EXIT 136
#define KVM_CAP_S390_GS 137
+#define KVM_CAP_S390_AIS 138
#ifdef KVM_CAP_IRQ_ROUTING
--
2.7.4
^ permalink raw reply related [flat|nested] 17+ messages in thread