diff mbox series

KVM: x86: move guest_pv_has out of user_access section

Message ID 20211112075422.3821671-1-pbonzini@redhat.com (mailing list archive)
State New, archived
Headers show
Series KVM: x86: move guest_pv_has out of user_access section | expand

Commit Message

Paolo Bonzini Nov. 12, 2021, 7:54 a.m. UTC
When UBSAN is enabled, the code emitted for the call to guest_pv_has
includes a call to __ubsan_handle_load_invalid_value.  objtool
complains that this call happens with UACCESS enabled; to avoid
the warning, pull the calls to user_access_begin into both arms
of the "if" statement, after the check for guest_pv_has.

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 arch/x86/kvm/x86.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

Comments

David Woodhouse Nov. 12, 2021, 8:11 a.m. UTC | #1
On Fri, 2021-11-12 at 02:54 -0500, Paolo Bonzini wrote:
> When UBSAN is enabled, the code emitted for the call to guest_pv_has
> includes a call to __ubsan_handle_load_invalid_value.  objtool
> complains that this call happens with UACCESS enabled; to avoid
> the warning, pull the calls to user_access_begin into both arms
> of the "if" statement, after the check for guest_pv_has.
> 
> Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
> Cc: David Woodhouse <dwmw2@infradead.org>
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Acked-by: David Woodhouse <dwmw@amazon.co.uk>

Thanks. I found a bunch of those objtool warnings during my original
experimentation and testing, but clearly missed that one. I'll turn on
CONFIG_UBSAN.
diff mbox series

Patch

diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index dd7b8b465675..dc7eb5fddfd3 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -3292,9 +3292,6 @@  static void record_steal_time(struct kvm_vcpu *vcpu)
 	}
 
 	st = (struct kvm_steal_time __user *)ghc->hva;
-	if (!user_access_begin(st, sizeof(*st)))
-		return;
-
 	/*
 	 * Doing a TLB flush here, on the guest's behalf, can avoid
 	 * expensive IPIs.
@@ -3303,6 +3300,9 @@  static void record_steal_time(struct kvm_vcpu *vcpu)
 		u8 st_preempted = 0;
 		int err = -EFAULT;
 
+		if (!user_access_begin(st, sizeof(*st)))
+			return;
+
 		asm volatile("1: xchgb %0, %2\n"
 			     "xor %1, %1\n"
 			     "2:\n"
@@ -3325,6 +3325,9 @@  static void record_steal_time(struct kvm_vcpu *vcpu)
 		if (!user_access_begin(st, sizeof(*st)))
 			goto dirty;
 	} else {
+		if (!user_access_begin(st, sizeof(*st)))
+			return;
+
 		unsafe_put_user(0, &st->preempted, out);
 		vcpu->arch.st.preempted = 0;
 	}