Skip to contentSkip to conversation

DIAG 318 tests and fix

PATCH In review Collin Walling 3 participants 9 messages Oct 15, 2020 – Oct 29, 2020

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

v2 cover DIAG 318 tests and fix

Two patches: one that fixes a case where the DIAG 318 info was not actually being reset during load normal, and another patch to inroduce selftests for DIAG 318 -- which helped discover the aforementioned issue ;)

Collin Walling (2):
s390/kvm: fix diag318 reset
self_tests/kvm: sync_regs and reset tests for diag318

 arch/s390/kvm/kvm-s390.c                      |  2 +-
 tools/testing/selftests/kvm/Makefile          |  2 +-
 .../kvm/include/s390x/diag318_test_handler.h  | 13 +++
 .../kvm/lib/s390x/diag318_test_handler.c      | 82 +++++++++++++++++++
 tools/testing/selftests/kvm/s390x/resets.c    | 14 ++++
 .../selftests/kvm/s390x/sync_regs_test.c      | 16 +++-
 6 files changed, 126 insertions(+), 3 deletions(-)
 create mode 100644 tools/testing/selftests/kvm/include/s390x/diag318_test_handler.h
 create mode 100644 tools/testing/selftests/kvm/lib/s390x/diag318_test_handler.c
Signature · 1 line
2.26.2
#1 lore raw

PATCH v2 1/2 s390/kvm: fix diag318 reset

The DIAGNOSE 0x0318 instruction must be reset on a normal and clear reset. However, this was missed for the clear reset case.

Let's fix this by resetting the information during a normal reset. Since clear reset is a superset of normal reset, the info will still reset on a clear reset.

---
 arch/s390/kvm/kvm-s390.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/s390/kvm/kvm-s390.c +1 -1
index 6b74b92c1a58..b0cf8367e261 100644
@@ -3516,6 +3516,7 @@ static void kvm_arch_vcpu_ioctl_normal_reset(struct kvm_vcpu *vcpu)
 	vcpu->arch.sie_block->gpsw.mask &= ~PSW_MASK_RI;
 	vcpu->arch.pfault_token = KVM_S390_PFAULT_TOKEN_INVALID;
 	memset(vcpu->run->s.regs.riccb, 0, sizeof(vcpu->run->s.regs.riccb));
+	vcpu->run->s.regs.diag318 = 0;
 
 	kvm_clear_async_pf_completion_queue(vcpu);
 	if (!kvm_s390_user_cpu_state_ctrl(vcpu->kvm))
@@ -3582,7 +3583,6 @@ static void kvm_arch_vcpu_ioctl_clear_reset(struct kvm_vcpu *vcpu)
 
 	regs->etoken = 0;
 	regs->etoken_extension = 0;
-	regs->diag318 = 0;
 }
 
 int kvm_arch_vcpu_ioctl_set_regs(struct kvm_vcpu *vcpu, struct kvm_regs *regs)
Signature · 1 line
2.26.2
#8 lore raw

PATCH v2 2/2 self_tests/kvm: sync_regs and reset tests for diag318

In reply to Collin Walling

The DIAGNOSE 0x0318 instruction, unique to s390x, is a privileged call that must be intercepted via SIE, handled in userspace, and the information set by the instruction is communicated back to KVM.

This information relates to a Control Program Name & Version code that helps diagnose the environment a guest is running in.

To test the instruction interception, an ad-hoc handler is defined which simply has a VM execute the instruction and then userspace will extract the necessary info. The handler is defined such that the instruction invocation occurs only once. It is up the the caller to determine how the info returned by this handler should be used.

The diag318 info is communicated from userspace to KVM via a sync_regs call. This is tested during a sync_regs test, where the diag318 info is requested via the handler, then the info is stored in the appropriate register in KVM via a sync registers call.

The diag318 info is checked to be 0 after a normal and clear reset.

If KVM does not support diag318, then the tests will print a message stating that diag318 was skipped, and the asserts will simply test against a value of 0.

---
 tools/testing/selftests/kvm/Makefile          |  2 +-
 .../kvm/include/s390x/diag318_test_handler.h  | 13 +++
 .../kvm/lib/s390x/diag318_test_handler.c      | 82 +++++++++++++++++++
 tools/testing/selftests/kvm/s390x/resets.c    | 14 ++++
 .../selftests/kvm/s390x/sync_regs_test.c      | 16 +++-
 5 files changed, 125 insertions(+), 2 deletions(-)
 create mode 100644 tools/testing/selftests/kvm/include/s390x/diag318_test_handler.h
 create mode 100644 tools/testing/selftests/kvm/lib/s390x/diag318_test_handler.c
tools/testing/selftests/kvm/Makefile +1 -1
index 4a166588d99f..ed172a0b83b6 100644
@@ -36,7 +36,7 @@ endif
 LIBKVM = lib/assert.c lib/elf.c lib/io.c lib/kvm_util.c lib/sparsebit.c lib/test_util.c
 LIBKVM_x86_64 = lib/x86_64/processor.c lib/x86_64/vmx.c lib/x86_64/svm.c lib/x86_64/ucall.c
 LIBKVM_aarch64 = lib/aarch64/processor.c lib/aarch64/ucall.c
-LIBKVM_s390x = lib/s390x/processor.c lib/s390x/ucall.c
+LIBKVM_s390x = lib/s390x/processor.c lib/s390x/ucall.c lib/s390x/diag318_test_handler.c
 
 TEST_GEN_PROGS_x86_64 = x86_64/cr4_cpuid_sync_test
 TEST_GEN_PROGS_x86_64 += x86_64/evmcs_test
tools/testing/selftests/kvm/include/s390x/diag318_test_handler.h new file +13 -0
new file mode 100644
index 000000000000..d8233ebf317b
@@ -0,0 +1,13 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Test handler for the s390x DIAGNOSE 0x0318 instruction.
+ *
+ * Copyright (C) 2020, IBM
+ */
+
+#ifndef SELFTEST_KVM_DIAG318_TEST_HANDLER
+#define SELFTEST_KVM_DIAG318_TEST_HANDLER
+
+uint64_t get_diag318_info(void);
+
+#endif
tools/testing/selftests/kvm/lib/s390x/diag318_test_handler.c new file +82 -0

82 lines not on this page · raw message

tools/testing/selftests/kvm/s390x/resets.c +14 -0
index b143db6d8693..d0853cebc402 100644
@@ -12,6 +12,7 @@
 
 #include "test_util.h"
 #include "kvm_util.h"
+#include "diag318_test_handler.h"
 
 #define VCPU_ID 3
 #define LOCAL_IRQS 32
@@ -110,6 +111,8 @@ static void assert_clear(void)
 
 	TEST_ASSERT(!memcmp(sync_regs->vrs, regs_null, sizeof(sync_regs->vrs)),
 		    "vrs0-15 == 0 (sync_regs)");
+
+	TEST_ASSERT(sync_regs->diag318 == 0, "diag318 == 0 (sync_regs)");
 }
 
 static void assert_initial_noclear(void)
@@ -182,6 +185,7 @@ static void assert_normal(void)
 	test_one_reg(KVM_REG_S390_PFTOKEN, KVM_S390_PFAULT_TOKEN_INVALID);
 	TEST_ASSERT(sync_regs->pft == KVM_S390_PFAULT_TOKEN_INVALID,
 			"pft == 0xff.....  (sync_regs)");
+	TEST_ASSERT(sync_regs->diag318 == 0, "diag318 == 0 (sync_regs)");
 	assert_noirq();
 }
 
@@ -208,6 +212,11 @@ static void test_normal(void)
 	run = vcpu_state(vm, VCPU_ID);
 	sync_regs = &run->s.regs;
 
+	if (get_diag318_info() > 0) {
+		run->s.regs.diag318 = get_diag318_info();
+		run->kvm_dirty_regs |= KVM_SYNC_DIAG318;
+	}
+
 	vcpu_run(vm, VCPU_ID);
 
 	inject_irq(VCPU_ID);
@@ -252,6 +261,11 @@ static void test_clear(void)
 	run = vcpu_state(vm, VCPU_ID);
 	sync_regs = &run->s.regs;
 
+	if (get_diag318_info() > 0) {
+		run->s.regs.diag318 = get_diag318_info();
+		run->kvm_dirty_regs |= KVM_SYNC_DIAG318;
+	}
+
 	vcpu_run(vm, VCPU_ID);
 
 	inject_irq(VCPU_ID);
tools/testing/selftests/kvm/s390x/sync_regs_test.c +15 -1
index 5731ccf34917..caf7b8859a94 100644
@@ -20,6 +20,7 @@
 
 #include "test_util.h"
 #include "kvm_util.h"
+#include "diag318_test_handler.h"
 
 #define VCPU_ID 5
 
@@ -70,7 +71,7 @@ static void compare_sregs(struct kvm_sregs *left, struct kvm_sync_regs *right)
 
 #undef REG_COMPARE
 
-#define TEST_SYNC_FIELDS   (KVM_SYNC_GPRS|KVM_SYNC_ACRS|KVM_SYNC_CRS)
+#define TEST_SYNC_FIELDS   (KVM_SYNC_GPRS|KVM_SYNC_ACRS|KVM_SYNC_CRS|KVM_SYNC_DIAG318)
 #define INVALID_SYNC_FIELD 0x80000000
 
 int main(int argc, char *argv[])
@@ -152,6 +153,12 @@ int main(int argc, char *argv[])
 
 	run->kvm_valid_regs = TEST_SYNC_FIELDS;
 	run->kvm_dirty_regs = KVM_SYNC_GPRS | KVM_SYNC_ACRS;
+
+	if (get_diag318_info() > 0) {
+		run->s.regs.diag318 = get_diag318_info();
+		run->kvm_dirty_regs |= KVM_SYNC_DIAG318;
+	}
+
 	rv = _vcpu_run(vm, VCPU_ID);
 	TEST_ASSERT(rv == 0, "vcpu_run failed: %d\n", rv);
 	TEST_ASSERT(run->exit_reason == KVM_EXIT_S390_SIEIC,
@@ -164,6 +171,9 @@ int main(int argc, char *argv[])
 	TEST_ASSERT(run->s.regs.acrs[0]  == 1 << 11,
 		    "acr0 sync regs value incorrect 0x%x.",
 		    run->s.regs.acrs[0]);
+	TEST_ASSERT(run->s.regs.diag318 == get_diag318_info(),
+		    "diag318 sync regs value incorrect 0x%llx.",
+		    run->s.regs.diag318);
 
 	vcpu_regs_get(vm, VCPU_ID, &regs);
 	compare_regs(&regs, &run->s.regs);
@@ -177,6 +187,7 @@ int main(int argc, char *argv[])
 	run->kvm_valid_regs = TEST_SYNC_FIELDS;
 	run->kvm_dirty_regs = 0;
 	run->s.regs.gprs[11] = 0xDEADBEEF;
+	run->s.regs.diag318 = 0x4B1D;
 	rv = _vcpu_run(vm, VCPU_ID);
 	TEST_ASSERT(rv == 0, "vcpu_run failed: %d\n", rv);
 	TEST_ASSERT(run->exit_reason == KVM_EXIT_S390_SIEIC,
@@ -186,6 +197,9 @@ int main(int argc, char *argv[])
 	TEST_ASSERT(run->s.regs.gprs[11] != 0xDEADBEEF,
 		    "r11 sync regs value incorrect 0x%llx.",
 		    run->s.regs.gprs[11]);
+	TEST_ASSERT(run->s.regs.diag318 != 0x4B1D,
+		    "diag318 sync regs value incorrect 0x%llx.",
+		    run->s.regs.diag318);
 
 	kvm_vm_free(vm);
 
Signature · 1 line
2.26.2
maintainer
#2 lore raw

Re: [PATCH v2 1/2] s390/kvm: fix diag318 reset

In reply to Collin Walling
Quoted 6 lines · Collin Walling

The DIAGNOSE 0x0318 instruction must be reset on a normal and clear reset. However, this was missed for the clear reset case.

Let's fix this by resetting the information during a normal reset. Since clear reset is a superset of normal reset, the info will still reset on a clear reset.

The architecture really confuses me here but I think we don't want this in the kernel VCPU reset handlers at all.

This needs to be reset per VM *NOT* per VCPU. Hence the resets are bound to diag308 and not SIGP.

I.e. we need to clear it in QEMU's VM reset handler.
It's still early and I have yet to consume my first coffee, am I missing something?

Quoted 27 lines
---
 arch/s390/kvm/kvm-s390.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/s390/kvm/kvm-s390.c +1 -1
index 6b74b92c1a58..b0cf8367e261 100644
@@ -3516,6 +3516,7 @@ static void kvm_arch_vcpu_ioctl_normal_reset(struct kvm_vcpu *vcpu)
 	vcpu->arch.sie_block->gpsw.mask &= ~PSW_MASK_RI;
 	vcpu->arch.pfault_token = KVM_S390_PFAULT_TOKEN_INVALID;
 	memset(vcpu->run->s.regs.riccb, 0, sizeof(vcpu->run->s.regs.riccb));
+	vcpu->run->s.regs.diag318 = 0;
 
 	kvm_clear_async_pf_completion_queue(vcpu);
 	if (!kvm_s390_user_cpu_state_ctrl(vcpu->kvm))
@@ -3582,7 +3583,6 @@ static void kvm_arch_vcpu_ioctl_clear_reset(struct kvm_vcpu *vcpu)
 
 	regs->etoken = 0;
 	regs->etoken_extension = 0;
-	regs->diag318 = 0;
 }
 
 int kvm_arch_vcpu_ioctl_set_regs(struct kvm_vcpu *vcpu, struct kvm_regs *regs)
#3 lore raw

Re: [PATCH v2 1/2] s390/kvm: fix diag318 reset

Quoted 11 lines · Janosch Frank
Quoted 6 lines · Collin Walling

The DIAGNOSE 0x0318 instruction must be reset on a normal and clear reset. However, this was missed for the clear reset case.

Let's fix this by resetting the information during a normal reset. Since clear reset is a superset of normal reset, the info will still reset on a clear reset.

The architecture really confuses me here but I think we don't want this in the kernel VCPU reset handlers at all.

Quoted 6 lines

This needs to be reset per VM *NOT* per VCPU. Hence the resets are bound to diag308 and not SIGP.

I.e. we need to clear it in QEMU's VM reset handler.
It's still early and I have yet to consume my first coffee, am I missing something?

I agree with Janosch. architecture indicates that this should only be reset for VM-wide resets, e.g. sigp orders 11 and 12 are explicitly mentioned to NOT reset the value.

Quoted 30 lines
Quoted 28 lines
---
 arch/s390/kvm/kvm-s390.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/s390/kvm/kvm-s390.c +1 -1
index 6b74b92c1a58..b0cf8367e261 100644
@@ -3516,6 +3516,7 @@ static void kvm_arch_vcpu_ioctl_normal_reset(struct kvm_vcpu *vcpu)
 	vcpu->arch.sie_block->gpsw.mask &= ~PSW_MASK_RI;
 	vcpu->arch.pfault_token = KVM_S390_PFAULT_TOKEN_INVALID;
 	memset(vcpu->run->s.regs.riccb, 0, sizeof(vcpu->run->s.regs.riccb));
+	vcpu->run->s.regs.diag318 = 0;
 
 	kvm_clear_async_pf_completion_queue(vcpu);
 	if (!kvm_s390_user_cpu_state_ctrl(vcpu->kvm))
@@ -3582,7 +3583,6 @@ static void kvm_arch_vcpu_ioctl_clear_reset(struct kvm_vcpu *vcpu)
 
 	regs->etoken = 0;
 	regs->etoken_extension = 0;
-	regs->diag318 = 0;
 }
 
 int kvm_arch_vcpu_ioctl_set_regs(struct kvm_vcpu *vcpu, struct kvm_regs *regs)
#4 lore raw

Re: [PATCH v2 1/2] s390/kvm: fix diag318 reset

Quoted 17 lines · Christian Borntraeger
Quoted 14 lines · Janosch Frank
Quoted 6 lines · Collin Walling

The DIAGNOSE 0x0318 instruction must be reset on a normal and clear reset. However, this was missed for the clear reset case.

Let's fix this by resetting the information during a normal reset. Since clear reset is a superset of normal reset, the info will still reset on a clear reset.

The architecture really confuses me here but I think we don't want this in the kernel VCPU reset handlers at all.

This needs to be reset per VM *NOT* per VCPU. Hence the resets are bound to diag308 and not SIGP.

Quoted 8 lines
Quoted 3 lines

I.e. we need to clear it in QEMU's VM reset handler.
It's still early and I have yet to consume my first coffee, am I missing something?

I agree with Janosch. architecture indicates that this should only be reset for VM-wide resets, e.g. sigp orders 11 and 12 are explicitly mentioned to NOT reset the value.

A few questions regarding how resets for diag318 should work here:

The AR states that any copies retained by the diag318 should be set to 0 on a clear reset and load normal -- I thought both of those resets were implicitly called by diag308 as well?

Should the register used to store diag318 info not be set to 0 *by KVM* then? Should the values be set *by QEMU* and a subsequent sync_regs will ensure things are sane on the KVM side?

Quoted 30 lines
Quoted 30 lines
Quoted 28 lines
---
 arch/s390/kvm/kvm-s390.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/s390/kvm/kvm-s390.c +1 -1
index 6b74b92c1a58..b0cf8367e261 100644
@@ -3516,6 +3516,7 @@ static void kvm_arch_vcpu_ioctl_normal_reset(struct kvm_vcpu *vcpu)
 	vcpu->arch.sie_block->gpsw.mask &= ~PSW_MASK_RI;
 	vcpu->arch.pfault_token = KVM_S390_PFAULT_TOKEN_INVALID;
 	memset(vcpu->run->s.regs.riccb, 0, sizeof(vcpu->run->s.regs.riccb));
+	vcpu->run->s.regs.diag318 = 0;
 
 	kvm_clear_async_pf_completion_queue(vcpu);
 	if (!kvm_s390_user_cpu_state_ctrl(vcpu->kvm))
@@ -3582,7 +3583,6 @@ static void kvm_arch_vcpu_ioctl_clear_reset(struct kvm_vcpu *vcpu)
 
 	regs->etoken = 0;
 	regs->etoken_extension = 0;
-	regs->diag318 = 0;
 }
 
 int kvm_arch_vcpu_ioctl_set_regs(struct kvm_vcpu *vcpu, struct kvm_regs *regs)
Signature · 4 lines
Regards,
Collin

Stay safe and stay healthy
maintainer
#5 lore raw

Re: [PATCH v2 1/2] s390/kvm: fix diag318 reset

Quoted 29 lines · Collin Walling
Quoted 17 lines · Christian Borntraeger
Quoted 14 lines · Janosch Frank
Quoted 6 lines · Collin Walling

The DIAGNOSE 0x0318 instruction must be reset on a normal and clear reset. However, this was missed for the clear reset case.

Let's fix this by resetting the information during a normal reset. Since clear reset is a superset of normal reset, the info will still reset on a clear reset.

The architecture really confuses me here but I think we don't want this in the kernel VCPU reset handlers at all.

This needs to be reset per VM *NOT* per VCPU. Hence the resets are bound to diag308 and not SIGP.

Quoted 8 lines
Quoted 3 lines

I.e. we need to clear it in QEMU's VM reset handler.
It's still early and I have yet to consume my first coffee, am I missing something?

I agree with Janosch. architecture indicates that this should only be reset for VM-wide resets, e.g. sigp orders 11 and 12 are explicitly mentioned to NOT reset the value.

A few questions regarding how resets for diag318 should work here:

Quoted 7 lines

The AR states that any copies retained by the diag318 should be set to 0 on a clear reset and load normal -- I thought both of those resets were implicitly called by diag308 as well?

Should the register used to store diag318 info not be set to 0 *by KVM* then? Should the values be set *by QEMU* and a subsequent sync_regs will ensure things are sane on the KVM side?

Just a FYI for the non-IBMers reading in:

I have spoken to the author of the architecture and cleared up our way forward.

* We need to clear on diag 308 subcodes 0,1,3,4
* SIGP does not alter diag318 data in any way
* We need to set the cpnc on all VCPUs of the VM

Quoted 33 lines
Quoted 31 lines
Quoted 31 lines
Quoted 29 lines
---
 arch/s390/kvm/kvm-s390.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/s390/kvm/kvm-s390.c +1 -1
index 6b74b92c1a58..b0cf8367e261 100644
@@ -3516,6 +3516,7 @@ static void kvm_arch_vcpu_ioctl_normal_reset(struct kvm_vcpu *vcpu)
 	vcpu->arch.sie_block->gpsw.mask &= ~PSW_MASK_RI;
 	vcpu->arch.pfault_token = KVM_S390_PFAULT_TOKEN_INVALID;
 	memset(vcpu->run->s.regs.riccb, 0, sizeof(vcpu->run->s.regs.riccb));
+	vcpu->run->s.regs.diag318 = 0;
 
 	kvm_clear_async_pf_completion_queue(vcpu);
 	if (!kvm_s390_user_cpu_state_ctrl(vcpu->kvm))
@@ -3582,7 +3583,6 @@ static void kvm_arch_vcpu_ioctl_clear_reset(struct kvm_vcpu *vcpu)
 
 	regs->etoken = 0;
 	regs->etoken_extension = 0;
-	regs->diag318 = 0;
 }
 
 int kvm_arch_vcpu_ioctl_set_regs(struct kvm_vcpu *vcpu, struct kvm_regs *regs)
#6 lore raw

Re: [PATCH v2 1/2] s390/kvm: fix diag318 reset

Quoted 38 lines · Janosch Frank
Quoted 29 lines · Collin Walling
Quoted 17 lines · Christian Borntraeger
Quoted 14 lines · Janosch Frank
Quoted 6 lines · Collin Walling

The DIAGNOSE 0x0318 instruction must be reset on a normal and clear reset. However, this was missed for the clear reset case.

Let's fix this by resetting the information during a normal reset. Since clear reset is a superset of normal reset, the info will still reset on a clear reset.

The architecture really confuses me here but I think we don't want this in the kernel VCPU reset handlers at all.

This needs to be reset per VM *NOT* per VCPU. Hence the resets are bound to diag308 and not SIGP.

Quoted 8 lines
Quoted 3 lines

I.e. we need to clear it in QEMU's VM reset handler.
It's still early and I have yet to consume my first coffee, am I missing something?

I agree with Janosch. architecture indicates that this should only be reset for VM-wide resets, e.g. sigp orders 11 and 12 are explicitly mentioned to NOT reset the value.

A few questions regarding how resets for diag318 should work here:

Quoted 7 lines

The AR states that any copies retained by the diag318 should be set to 0 on a clear reset and load normal -- I thought both of those resets were implicitly called by diag308 as well?

Should the register used to store diag318 info not be set to 0 *by KVM* then? Should the values be set *by QEMU* and a subsequent sync_regs will ensure things are sane on the KVM side?

Quoted 7 lines

Just a FYI for the non-IBMers reading in:

I have spoken to the author of the architecture and cleared up our way forward.

* We need to clear on diag 308 subcodes 0,1,3,4
* SIGP does not alter diag318 data in any way

Just to make sure I fully understand the changes required here (since the wording in the documentation is a bit tricky) -- this means I need to remove the kvm_arch_vcpu_ioctl_*_reset code for the diag318 data and instead implement a way for this to be reset via 308 only?

If true, then resetting this data becomes tricky with the current implementation...

Quoted 3 lines

* We need to set the cpnc on all VCPUs of the VM

This begs a few questions about the current design of the diag318 code.

Since the diag318 info (CPNC and CPVC) are *attributes of the configuration*, then is the current implementation of the diag318 handling incorrect?

Right now, the diag318 data is synced per-vcpu via a sync_regs call, and only the CPU involved in the instruction interception will have its diag318 register synced. However, if we shouldn't use the kvm_vcpu_ioctl functions to reset the diag318 data, then does it make sense to use the sync_regs interface?

There are two approaches I can think of:

1. Modify the current implementation by "dirtying all the VCPU's registers" -- this will ensure the CPNC is set in all SIE blocks.

2. Fallback to the old implementation of this item and use an ioctl (only for setting the data)

This new ioctl can be utilized to reset the diag318 data (including the copy retained by QEMU for migration) during the respective 308 resets.

Any insights before I fall down the wrong rabbit hole?

[...]

Signature · 4 lines
Regards,
Collin

Stay safe and stay healthy
maintainer
#7 lore raw

Re: [PATCH v2 1/2] s390/kvm: fix diag318 reset

Quoted 44 lines · Collin Walling
Quoted 43 lines · Janosch Frank
Quoted 29 lines · Collin Walling
Quoted 17 lines · Christian Borntraeger
Quoted 14 lines · Janosch Frank
Quoted 6 lines · Collin Walling

The DIAGNOSE 0x0318 instruction must be reset on a normal and clear reset. However, this was missed for the clear reset case.

Let's fix this by resetting the information during a normal reset. Since clear reset is a superset of normal reset, the info will still reset on a clear reset.

The architecture really confuses me here but I think we don't want this in the kernel VCPU reset handlers at all.

This needs to be reset per VM *NOT* per VCPU. Hence the resets are bound to diag308 and not SIGP.

Quoted 8 lines
Quoted 3 lines

I.e. we need to clear it in QEMU's VM reset handler.
It's still early and I have yet to consume my first coffee, am I missing something?

I agree with Janosch. architecture indicates that this should only be reset for VM-wide resets, e.g. sigp orders 11 and 12 are explicitly mentioned to NOT reset the value.

A few questions regarding how resets for diag318 should work here:

Quoted 7 lines

The AR states that any copies retained by the diag318 should be set to 0 on a clear reset and load normal -- I thought both of those resets were implicitly called by diag308 as well?

Should the register used to store diag318 info not be set to 0 *by KVM* then? Should the values be set *by QEMU* and a subsequent sync_regs will ensure things are sane on the KVM side?

Just a FYI for the non-IBMers reading in:

I have spoken to the author of the architecture and cleared up our way forward.

Quoted 7 lines
Quoted 2 lines

* We need to clear on diag 308 subcodes 0,1,3,4
* SIGP does not alter diag318 data in any way

Just to make sure I fully understand the changes required here (since the wording in the documentation is a bit tricky) -- this means I need to remove the kvm_arch_vcpu_ioctl_*_reset code for the diag318 data and instead implement a way for this to be reset via 308 only?

Yes

Quoted 3 lines

If true, then resetting this data becomes tricky with the current implementation...

Not really, it's just a bit of a pain to iterate through all VCPUs

Quoted 22 lines
Quoted 4 lines

* We need to set the cpnc on all VCPUs of the VM

This begs a few questions about the current design of the diag318 code.

Since the diag318 info (CPNC and CPVC) are *attributes of the configuration*, then is the current implementation of the diag318 handling incorrect?

Right now, the diag318 data is synced per-vcpu via a sync_regs call, and only the CPU involved in the instruction interception will have its diag318 register synced. However, if we shouldn't use the kvm_vcpu_ioctl functions to reset the diag318 data, then does it make sense to use the sync_regs interface?

There are two approaches I can think of:

1. Modify the current implementation by "dirtying all the VCPU's registers" -- this will ensure the CPNC is set in all SIE blocks.

Quoted 6 lines

2. Fallback to the old implementation of this item and use an ioctl (only for setting the data)

This new ioctl can be utilized to reset the diag318 data (including the copy retained by QEMU for migration) during the respective 308 resets.

I don't think that we can tear down old IOCTLs that easily, we have to keep the API consistent.

Luckily a diag 308 reset also means VCPU resets, so all VCPUs are out of SIE and QEMU will change a lot of the sync reg contents anyway.

The tricky part is the instruction itself as synchronizing the CPNC change to all VCPUs will not be easy.

I'd suggest you start with the d308 changes and I'll help you next week after the KVM Forum

@Christian: Any other thoughts on this?

At the end of the year I'll take a day to find out how we went so wrong with this item...

Quoted 5 lines

Any insights before I fall down the wrong rabbit hole?

[...]