[PULL 0/6] KVM: s390: patches for kvm-next

[PULL 0/6] KVM: s390: patches for kvm-next

[Cornelia Huck]

Gleb, Paolo,

here are some more fixes and improvements that have been sitting
in our tree. Please consider pulling.

The following changes since commit d6d63b51fe3bfea0cf596993afa480b0b3b02c32:

  Merge tag 'kvm-s390-20131128' of git://git.kernel.org/pub/scm/linux/kernel/git/kvms390/linux into kvm-next (2013-11-28 15:30:16 +0100)

are available in the git repository at:


  git://git.kernel.org/pub/scm/linux/kernel/git/kvms390/linux.git tags/kvm-s390-20131211

for you to fetch changes up to ff1f3cb4b3ac5d039f02679f34cb1498d110d241:

  KVM: s390: ioeventfd: ignore leftmost bits (2013-12-11 19:05:41 +0100)

----------------------------------------------------------------
Some further s390 patches for kvm-next.

Various improvements and bugfixes in the signal processor handling.
Document kvm support for diagnose (s390 hypercalls). And last but
not least, fix a bug in the s390 ioeventfd backend that was causing
us grief in scenarios with 4G+ memory.

----------------------------------------------------------------
Cornelia Huck (1):
  KVM: s390: diagnose call documentation

Dominik Dingel (1):
  KVM: s390: ioeventfd: ignore leftmost bits

Thomas Huth (4):
  KVM: s390: Use helper function to set CC in SIGP handler
  KVM: s390: Add the SIGP order CONDITIONAL EMERGENCY SIGNAL
  KVM: s390: Reworked SIGP RESTART order
  KVM: s390: SIGP START has to report BUSY while stopping a CPU

 Documentation/virtual/kvm/hypercalls.txt |    3 ++
 Documentation/virtual/kvm/s390-diag.txt  |   80 ++++++++++++++++++++++++++++++
 arch/s390/include/asm/sigp.h             |    2 +
 arch/s390/kvm/diag.c                     |    2 +-
 arch/s390/kvm/sigp.c                     |   64 ++++++++++++++++++++----
 5 files changed, 139 insertions(+), 12 deletions(-)
 create mode 100644 Documentation/virtual/kvm/s390-diag.txt

-- 
1.7.9.5

[PULL 1/6] KVM: s390: diagnose call documentation

[Cornelia Huck]

Add some further documentation on the DIAGNOSE calls we support.

Reviewed-by: Thomas Huth <thuth@linux.vnet.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
---
 Documentation/virtual/kvm/hypercalls.txt |    3 ++
 Documentation/virtual/kvm/s390-diag.txt  |   80 ++++++++++++++++++++++++++++++
 2 files changed, 83 insertions(+)
 create mode 100644 Documentation/virtual/kvm/s390-diag.txt

diff --git a/Documentation/virtual/kvm/hypercalls.txt b/Documentation/virtual/kvm/hypercalls.txt
index 022198e..d922d73 100644
--- a/Documentation/virtual/kvm/hypercalls.txt
+++ b/Documentation/virtual/kvm/hypercalls.txt
@@ -17,6 +17,9 @@ S390:
   S390 uses diagnose instruction as hypercall (0x500) along with hypercall
   number in R1.
 
+  For further information on the S390 diagnose call as supported by KVM,
+  refer to Documentation/virtual/kvm/s390-diag.txt.
+
  PowerPC:
   It uses R3-R10 and hypercall number in R11. R4-R11 are used as output registers.
   Return value is placed in R3.
diff --git a/Documentation/virtual/kvm/s390-diag.txt b/Documentation/virtual/kvm/s390-diag.txt
new file mode 100644
index 0000000..f1de4fb
--- /dev/null
+++ b/Documentation/virtual/kvm/s390-diag.txt
@@ -0,0 +1,80 @@
+The s390 DIAGNOSE call on KVM
+=============================
+
+KVM on s390 supports the DIAGNOSE call for making hypercalls, both for
+native hypercalls and for selected hypercalls found on other s390
+hypervisors.
+
+Note that bits are numbered as by the usual s390 convention (most significant
+bit on the left).
+
+
+General remarks
+---------------
+
+DIAGNOSE calls by the guest cause a mandatory intercept. This implies
+all supported DIAGNOSE calls need to be handled by either KVM or its
+userspace.
+
+All DIAGNOSE calls supported by KVM use the RS-a format:
+
+--------------------------------------
+|  '83'  | R1 | R3 | B2 |     D2     |
+--------------------------------------
+0        8    12   16   20           31
+
+The second-operand address (obtained by the base/displacement calculation)
+is not used to address data. Instead, bits 48-63 of this address specify
+the function code, and bits 0-47 are ignored.
+
+The supported DIAGNOSE function codes vary by the userspace used. For
+DIAGNOSE function codes not specific to KVM, please refer to the
+documentation for the s390 hypervisors defining them.
+
+
+DIAGNOSE function code 'X'500' - KVM virtio functions
+-----------------------------------------------------
+
+If the function code specifies 0x500, various virtio-related functions
+are performed.
+
+General register 1 contains the virtio subfunction code. Supported
+virtio subfunctions depend on KVM's userspace. Generally, userspace
+provides either s390-virtio (subcodes 0-2) or virtio-ccw (subcode 3).
+
+Upon completion of the DIAGNOSE instruction, general register 2 contains
+the function's return code, which is either a return code or a subcode
+specific value.
+
+Subcode 0 - s390-virtio notification and early console printk
+    Handled by userspace.
+
+Subcode 1 - s390-virtio reset
+    Handled by userspace.
+
+Subcode 2 - s390-virtio set status
+    Handled by userspace.
+
+Subcode 3 - virtio-ccw notification
+    Handled by either userspace or KVM (ioeventfd case).
+
+    General register 2 contains a subchannel-identification word denoting
+    the subchannel of the virtio-ccw proxy device to be notified.
+
+    General register 3 contains the number of the virtqueue to be notified.
+
+    General register 4 contains a 64bit identifier for KVM usage (the
+    kvm_io_bus cookie). If general register 4 does not contain a valid
+    identifier, it is ignored.
+
+    After completion of the DIAGNOSE call, general register 2 may contain
+    a 64bit identifier (in the kvm_io_bus cookie case).
+
+    See also the virtio standard for a discussion of this hypercall.
+
+
+DIAGNOSE function code 'X'501 - KVM breakpoint
+----------------------------------------------
+
+If the function code specifies 0x501, breakpoint functions may be performed.
+This function code is handled by userspace.
-- 
1.7.9.5

[PULL 2/6] KVM: s390: Use helper function to set CC in SIGP handler

[Cornelia Huck]

From: Thomas Huth <thuth@linux.vnet.ibm.com>

We've got a helper function for setting the condition code now,
so let's use it in the SIGP handler, too.

Signed-off-by: Thomas Huth <thuth@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
---
 arch/s390/kvm/sigp.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/arch/s390/kvm/sigp.c b/arch/s390/kvm/sigp.c
index c370058..bc0d85a 100644
--- a/arch/s390/kvm/sigp.c
+++ b/arch/s390/kvm/sigp.c
@@ -435,7 +435,6 @@ int kvm_s390_handle_sigp(struct kvm_vcpu *vcpu)
 	if (rc < 0)
 		return rc;
 
-	vcpu->arch.sie_block->gpsw.mask &= ~(3ul << 44);
-	vcpu->arch.sie_block->gpsw.mask |= (rc & 3ul) << 44;
+	kvm_s390_set_psw_cc(vcpu, rc);
 	return 0;
 }
-- 
1.7.9.5

[PULL 3/6] KVM: s390: Add the SIGP order CONDITIONAL EMERGENCY SIGNAL

[Cornelia Huck]

From: Thomas Huth <thuth@linux.vnet.ibm.com>

This patch adds the missing SIGP order "conditional emergency
signal" by calling the "emergency signal" SIGP handler if the
required conditions are met.

Signed-off-by: Thomas Huth <thuth@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
---
 arch/s390/include/asm/sigp.h |    1 +
 arch/s390/kvm/sigp.c         |   37 ++++++++++++++++++++++++++++++++++++-
 2 files changed, 37 insertions(+), 1 deletion(-)

diff --git a/arch/s390/include/asm/sigp.h b/arch/s390/include/asm/sigp.h
index 5a87d16..c002cd5 100644
--- a/arch/s390/include/asm/sigp.h
+++ b/arch/s390/include/asm/sigp.h
@@ -12,6 +12,7 @@
 #define SIGP_SET_PREFIX		     13
 #define SIGP_STORE_STATUS_AT_ADDRESS 14
 #define SIGP_SET_ARCHITECTURE	     18
+#define SIGP_COND_EMERGENCY_SIGNAL   19
 #define SIGP_SENSE_RUNNING	     21
 
 /* SIGP condition codes */
diff --git a/arch/s390/kvm/sigp.c b/arch/s390/kvm/sigp.c
index bc0d85a..eee1402 100644
--- a/arch/s390/kvm/sigp.c
+++ b/arch/s390/kvm/sigp.c
@@ -1,7 +1,7 @@
 /*
  * handling interprocessor communication
  *
- * Copyright IBM Corp. 2008, 2009
+ * Copyright IBM Corp. 2008, 2013
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License (version 2 only)
@@ -89,6 +89,37 @@ unlock:
 	return rc;
 }
 
+static int __sigp_conditional_emergency(struct kvm_vcpu *vcpu, u16 cpu_addr,
+					u16 asn, u64 *reg)
+{
+	struct kvm_vcpu *dst_vcpu = NULL;
+	const u64 psw_int_mask = PSW_MASK_IO | PSW_MASK_EXT;
+	u16 p_asn, s_asn;
+	psw_t *psw;
+	u32 flags;
+
+	if (cpu_addr < KVM_MAX_VCPUS)
+		dst_vcpu = kvm_get_vcpu(vcpu->kvm, cpu_addr);
+	if (!dst_vcpu)
+		return SIGP_CC_NOT_OPERATIONAL;
+	flags = atomic_read(&dst_vcpu->arch.sie_block->cpuflags);
+	psw = &dst_vcpu->arch.sie_block->gpsw;
+	p_asn = dst_vcpu->arch.sie_block->gcr[4] & 0xffff;  /* Primary ASN */
+	s_asn = dst_vcpu->arch.sie_block->gcr[3] & 0xffff;  /* Secondary ASN */
+
+	/* Deliver the emergency signal? */
+	if (!(flags & CPUSTAT_STOPPED)
+	    || (psw->mask & psw_int_mask) != psw_int_mask
+	    || ((flags & CPUSTAT_WAIT) && psw->addr != 0)
+	    || (!(flags & CPUSTAT_WAIT) && (asn == p_asn || asn == s_asn))) {
+		return __sigp_emergency(vcpu, cpu_addr);
+	} else {
+		*reg &= 0xffffffff00000000UL;
+		*reg |= SIGP_STATUS_INCORRECT_STATE;
+		return SIGP_CC_STATUS_STORED;
+	}
+}
+
 static int __sigp_external_call(struct kvm_vcpu *vcpu, u16 cpu_addr)
 {
 	struct kvm_s390_float_interrupt *fi = &vcpu->kvm->arch.float_int;
@@ -417,6 +448,10 @@ int kvm_s390_handle_sigp(struct kvm_vcpu *vcpu)
 		rc = __sigp_set_prefix(vcpu, cpu_addr, parameter,
 				       &vcpu->run->s.regs.gprs[r1]);
 		break;
+	case SIGP_COND_EMERGENCY_SIGNAL:
+		rc = __sigp_conditional_emergency(vcpu, cpu_addr, parameter,
+						  &vcpu->run->s.regs.gprs[r1]);
+		break;
 	case SIGP_SENSE_RUNNING:
 		vcpu->stat.instruction_sigp_sense_running++;
 		rc = __sigp_sense_running(vcpu, cpu_addr,
-- 
1.7.9.5

[PULL 4/6] KVM: s390: Reworked SIGP RESTART order

[Cornelia Huck]

From: Thomas Huth <thuth@linux.vnet.ibm.com>

When SIGP RESTART detected an illegal CPU address, there is no need to
drop to userspace, we can return CC3 to the guest directly instead.
Also renamed __sigp_restart() to sigp_check_callable() (since this
is a better description of what the function is really doing) and
moved a string specific to RESTART to the calling place instead, so
that this function gets usable by other SIGP orders, too.

Signed-off-by: Thomas Huth <thuth@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
---
 arch/s390/kvm/sigp.c |   19 +++++++++++--------
 1 file changed, 11 insertions(+), 8 deletions(-)

diff --git a/arch/s390/kvm/sigp.c b/arch/s390/kvm/sigp.c
index eee1402..509547d 100644
--- a/arch/s390/kvm/sigp.c
+++ b/arch/s390/kvm/sigp.c
@@ -363,7 +363,8 @@ static int __sigp_sense_running(struct kvm_vcpu *vcpu, u16 cpu_addr,
 	return rc;
 }
 
-static int __sigp_restart(struct kvm_vcpu *vcpu, u16 cpu_addr)
+/* Test whether the destination CPU is available and not busy */
+static int sigp_check_callable(struct kvm_vcpu *vcpu, u16 cpu_addr)
 {
 	struct kvm_s390_float_interrupt *fi = &vcpu->kvm->arch.float_int;
 	struct kvm_s390_local_interrupt *li;
@@ -382,9 +383,6 @@ static int __sigp_restart(struct kvm_vcpu *vcpu, u16 cpu_addr)
 	spin_lock_bh(&li->lock);
 	if (li->action_bits & ACTION_STOP_ON_STOP)
 		rc = SIGP_CC_BUSY;
-	else
-		VCPU_EVENT(vcpu, 4, "sigp restart %x to handle userspace",
-			cpu_addr);
 	spin_unlock_bh(&li->lock);
 out:
 	spin_unlock(&fi->lock);
@@ -459,10 +457,15 @@ int kvm_s390_handle_sigp(struct kvm_vcpu *vcpu)
 		break;
 	case SIGP_RESTART:
 		vcpu->stat.instruction_sigp_restart++;
-		rc = __sigp_restart(vcpu, cpu_addr);
-		if (rc == SIGP_CC_BUSY)
-			break;
-		/* user space must know about restart */
+		rc = sigp_check_callable(vcpu, cpu_addr);
+		if (rc == SIGP_CC_ORDER_CODE_ACCEPTED) {
+			VCPU_EVENT(vcpu, 4,
+				   "sigp restart %x to handle userspace",
+				   cpu_addr);
+			/* user space must know about restart */
+			rc = -EOPNOTSUPP;
+		}
+		break;
 	default:
 		return -EOPNOTSUPP;
 	}
-- 
1.7.9.5

[PULL 5/6] KVM: s390: SIGP START has to report BUSY while stopping a CPU

[Cornelia Huck]

From: Thomas Huth <thuth@linux.vnet.ibm.com>

Just like the RESTART order, the START order also has to report BUSY
while a STOP request is pending, to avoid that the START might be
ignored due to a race condition between the STOP and the START order.

Signed-off-by: Thomas Huth <thuth@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
---
 arch/s390/include/asm/sigp.h |    1 +
 arch/s390/kvm/sigp.c         |    5 +++++
 2 files changed, 6 insertions(+)

diff --git a/arch/s390/include/asm/sigp.h b/arch/s390/include/asm/sigp.h
index c002cd5..d091aa1 100644
--- a/arch/s390/include/asm/sigp.h
+++ b/arch/s390/include/asm/sigp.h
@@ -5,6 +5,7 @@
 #define SIGP_SENSE		      1
 #define SIGP_EXTERNAL_CALL	      2
 #define SIGP_EMERGENCY_SIGNAL	      3
+#define SIGP_START		      4
 #define SIGP_STOP		      5
 #define SIGP_RESTART		      6
 #define SIGP_STOP_AND_STORE_STATUS    9
diff --git a/arch/s390/kvm/sigp.c b/arch/s390/kvm/sigp.c
index 509547d..87c2b3a 100644
--- a/arch/s390/kvm/sigp.c
+++ b/arch/s390/kvm/sigp.c
@@ -455,6 +455,11 @@ int kvm_s390_handle_sigp(struct kvm_vcpu *vcpu)
 		rc = __sigp_sense_running(vcpu, cpu_addr,
 					  &vcpu->run->s.regs.gprs[r1]);
 		break;
+	case SIGP_START:
+		rc = sigp_check_callable(vcpu, cpu_addr);
+		if (rc == SIGP_CC_ORDER_CODE_ACCEPTED)
+			rc = -EOPNOTSUPP;    /* Handle START in user space */
+		break;
 	case SIGP_RESTART:
 		vcpu->stat.instruction_sigp_restart++;
 		rc = sigp_check_callable(vcpu, cpu_addr);
-- 
1.7.9.5

[PULL 6/6] KVM: s390: ioeventfd: ignore leftmost bits

[Cornelia Huck]

From: Dominik Dingel <dingel@linux.vnet.ibm.com>

The diagnose 500 subcode 3 contains the 32 bit subchannel id in bits 32-63
(counting from the left). As for other I/O instructions, bits 0-31 should be
ignored and thus not be passed to kvm_io_bus_write_cookie().

This fixes a bug where the guest passed non-zero bits 0-31 which the
host tried to interpret, leading to ioeventfd notification failures.

Cc: stable@vger.kernel.org
Signed-off-by: Dominik Dingel <dingel@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
---
 arch/s390/kvm/diag.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/s390/kvm/diag.c b/arch/s390/kvm/diag.c
index 5ff29be..8216c0e 100644
--- a/arch/s390/kvm/diag.c
+++ b/arch/s390/kvm/diag.c
@@ -121,7 +121,7 @@ static int __diag_virtio_hypercall(struct kvm_vcpu *vcpu)
 	 * - gpr 4 contains the index on the bus (optionally)
 	 */
 	ret = kvm_io_bus_write_cookie(vcpu->kvm, KVM_VIRTIO_CCW_NOTIFY_BUS,
-				      vcpu->run->s.regs.gprs[2],
+				      vcpu->run->s.regs.gprs[2] & 0xffffffff,
 				      8, &vcpu->run->s.regs.gprs[3],
 				      vcpu->run->s.regs.gprs[4]);
 
-- 
1.7.9.5