Re: [kvm-unit-tests PATCH v2] s390x: Fix uv_call() exception behavior

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 18/01/2021 16.09, Janosch Frank wrote:
On a program exception we usually skip the instruction that caused the
exception and continue. That won't work for UV calls since a "brc
3,0b" will retry the instruction if the CC is > 1. Let's forgo the brc
when checking for privilege exceptions and use a uv_call_once().

Signed-off-by: Janosch Frank <frankja@xxxxxxxxxxxxx>
Suggested-by: Claudio Imbrenda <imbrenda@xxxxxxxxxxxxx>

---
  lib/s390x/asm/uv.h | 24 ++++++++++++++++--------
  s390x/uv-guest.c   |  6 +++---
  2 files changed, 19 insertions(+), 11 deletions(-)

diff --git a/lib/s390x/asm/uv.h b/lib/s390x/asm/uv.h
index 4c2fc48..39d2dc0 100644
--- a/lib/s390x/asm/uv.h
+++ b/lib/s390x/asm/uv.h
@@ -50,19 +50,12 @@ struct uv_cb_share {
  	u64 reserved28;
  } __attribute__((packed))  __attribute__((aligned(8)));
-static inline int uv_call(unsigned long r1, unsigned long r2)
+static inline int uv_call_once(unsigned long r1, unsigned long r2)
  {
  	int cc;
- /*
-	 * The brc instruction will take care of the cc 2/3 case where
-	 * we need to continue the execution because we were
-	 * interrupted. The inline assembly will only return on
-	 * success/error i.e. cc 0/1.
-	*/
  	asm volatile(
  		"0:	.insn rrf,0xB9A40000,%[r1],%[r2],0,0\n"
-		"		brc	3,0b\n"
  		"		ipm	%[cc]\n"
  		"		srl	%[cc],28\n"
  		: [cc] "=d" (cc)
@@ -71,4 +64,19 @@ static inline int uv_call(unsigned long r1, unsigned long r2)
  	return cc;
  }
+static inline int uv_call(unsigned long r1, unsigned long r2)
+{
+	int cc;
+
+	/*
+	 * CC 2 and 3 tell us to re-execute because the instruction
+	 * hasn't yet finished.
+	 */
+	do {
+		cc = uv_call_once(r1, r2);
+	} while (cc > 1);
+
+	return cc;
+}
+
  #endif
diff --git a/s390x/uv-guest.c b/s390x/uv-guest.c
index d47333e..091a19b 100644
--- a/s390x/uv-guest.c
+++ b/s390x/uv-guest.c
@@ -31,7 +31,7 @@ static void test_priv(void)
  	uvcb.len = sizeof(struct uv_cb_qui);
  	expect_pgm_int();
  	enter_pstate();
-	uv_call(0, (u64)&uvcb);
+	uv_call_once(0, (u64)&uvcb);
  	check_pgm_int_code(PGM_INT_CODE_PRIVILEGED_OPERATION);
  	report_prefix_pop();
@@ -40,7 +40,7 @@ static void test_priv(void)
  	uvcb.len = sizeof(struct uv_cb_share);
  	expect_pgm_int();
  	enter_pstate();
-	uv_call(0, (u64)&uvcb);
+	uv_call_once(0, (u64)&uvcb);
  	check_pgm_int_code(PGM_INT_CODE_PRIVILEGED_OPERATION);
  	report_prefix_pop();
@@ -49,7 +49,7 @@ static void test_priv(void)
  	uvcb.len = sizeof(struct uv_cb_share);
  	expect_pgm_int();
  	enter_pstate();
-	uv_call(0, (u64)&uvcb);
+	uv_call_once(0, (u64)&uvcb);
  	check_pgm_int_code(PGM_INT_CODE_PRIVILEGED_OPERATION);
  	report_prefix_pop();

That looks nicer, indeed.

Reviewed-by: Thomas Huth <thuth@xxxxxxxxxx>




[Index of Archives]     [KVM ARM]     [KVM ia64]     [KVM ppc]     [Virtualization Tools]     [Spice Development]     [Libvirt]     [Libvirt Users]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite Questions]     [Linux Kernel]     [Linux SCSI]     [XFree86]

  Powered by Linux