Skip to contentSkip to conversation

KVM: x86: Rep string I/O WARN removal and test

PATCH Applied Sean Christopherson 2 participants 6 messages Oct 25, 2021 – Dec 9, 2021

Subsystems (4)
lore
Share & tools
lore

Sean Christopherson

author maintainer
#0 lore raw

cover KVM: x86: Rep string I/O WARN removal and test

Remove a WARN that was added as part of the recent I/O overhaul to play nice with SEV-ES string I/O.

For the record, my FIXME in lieu of a WARN was deliberate, as I suspected userspace could trigger a WARN ;-)

Based on kvm/master, commit 95e16b4792b0 ("KVM: SEV-ES: go over the sev_pio_data buffer in multiple passes if needed").

Sean Christopherson (2):
KVM: x86: Don't WARN if userspace mucks with RCX during string I/O
exit
KVM: selftests: Add test to verify KVM doesn't explode on "bad" I/O

 arch/x86/kvm/x86.c                            |   9 +-
 tools/testing/selftests/kvm/.gitignore        |   1 +
 tools/testing/selftests/kvm/Makefile          |   1 +
 .../selftests/kvm/x86_64/userspace_io_test.c  | 114 ++++++++++++++++++
 4 files changed, 123 insertions(+), 2 deletions(-)
 create mode 100644 tools/testing/selftests/kvm/x86_64/userspace_io_test.c
Signature · 1 line
2.33.0.1079.g6e70778dc9-goog
author maintainer
#1 lore raw

PATCH 1/2 KVM: x86: Don't WARN if userspace mucks with RCX during string I/O exit

Replace a WARN with a comment to call out that userspace can modify RCX during an exit to userspace to handle string I/O. KVM doesn't actually support changing the rep count during an exit, i.e. the scenario can be ignored, but the WARN needs to go as it's trivial to trigger from userspace.

Fixes 3b27de271839 ("KVM: x86: split the two parts of emulator_pio_in")
---
 arch/x86/kvm/x86.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)
arch/x86/kvm/x86.c +7 -2
index b26647a5ea22..1726812f31af 100644
@@ -6948,7 +6948,13 @@ static int emulator_pio_in(struct kvm_vcpu *vcpu, int size,
 			   unsigned short port, void *val, unsigned int count)
 {
 	if (vcpu->arch.pio.count) {
-		/* Complete previous iteration.  */
+		/*
+		 * Complete a previous iteration that required userspace I/O.
+		 * Note, @count isn't guaranteed to match pio.count as userspace
+		 * can modify ECX before rerunning the vCPU.  Ignore any such
+		 * shenanigans as KVM doesn't support modifying the rep count,
+		 * and the emulator ensures @count doesn't overflow the buffer.
+		 */
 	} else {
 		int r = __emulator_pio_in(vcpu, size, port, count);
 		if (!r)
@@ -6957,7 +6963,6 @@ static int emulator_pio_in(struct kvm_vcpu *vcpu, int size,
 		/* Results already available, fall through.  */
 	}
 
-	WARN_ON(count != vcpu->arch.pio.count);
 	complete_emulator_pio_in(vcpu, val);
 	return 1;
 }
Signature · 1 line
2.33.0.1079.g6e70778dc9-goog
author maintainer
#2 lore raw

PATCH 2/2 KVM: selftests: Add test to verify KVM doesn't explode on "bad" I/O

In reply to Sean Christopherson

Add an x86 selftest to verify that KVM doesn't WARN or otherwise explode if userspace modifies RCX during a userspace exit to handle string I/O. This is a regression test for a user-triggerable WARN introduced by commit 3b27de271839 ("KVM: x86: split the two parts of emulator_pio_in").

---
 tools/testing/selftests/kvm/.gitignore        |   1 +
 tools/testing/selftests/kvm/Makefile          |   1 +
 .../selftests/kvm/x86_64/userspace_io_test.c  | 114 ++++++++++++++++++
 3 files changed, 116 insertions(+)
 create mode 100644 tools/testing/selftests/kvm/x86_64/userspace_io_test.c
tools/testing/selftests/kvm/.gitignore +1 -0
index b8dbabe24ac2..691262c7b122 100644
@@ -27,6 +27,7 @@
 /x86_64/svm_int_ctl_test
 /x86_64/sync_regs_test
 /x86_64/tsc_msrs_test
+/x86_64/userspace_io_test
 /x86_64/userspace_msr_exit_test
 /x86_64/vmx_apic_access_test
 /x86_64/vmx_close_while_nested_test
tools/testing/selftests/kvm/Makefile +1 -0
index d1774f461393..1a611cb80c4e 100644
@@ -58,6 +58,7 @@ TEST_GEN_PROGS_x86_64 += x86_64/vmx_preemption_timer_test
 TEST_GEN_PROGS_x86_64 += x86_64/svm_vmcall_test
 TEST_GEN_PROGS_x86_64 += x86_64/svm_int_ctl_test
 TEST_GEN_PROGS_x86_64 += x86_64/sync_regs_test
+TEST_GEN_PROGS_x86_64 += x86_64/userspace_io_test
 TEST_GEN_PROGS_x86_64 += x86_64/userspace_msr_exit_test
 TEST_GEN_PROGS_x86_64 += x86_64/vmx_apic_access_test
 TEST_GEN_PROGS_x86_64 += x86_64/vmx_close_while_nested_test
tools/testing/selftests/kvm/x86_64/userspace_io_test.c new file +114 -0

114 lines not on this page · raw message

Signature · 1 line
2.33.0.1079.g6e70778dc9-goog
author maintainer
#3 lore raw
In reply to Sean Christopherson
Quoted 14 lines · Sean Christopherson

Remove a WARN that was added as part of the recent I/O overhaul to play nice with SEV-ES string I/O.

For the record, my FIXME in lieu of a WARN was deliberate, as I suspected userspace could trigger a WARN ;-)

Based on kvm/master, commit 95e16b4792b0 ("KVM: SEV-ES: go over the sev_pio_data buffer in multiple passes if needed").

Sean Christopherson (2):
KVM: x86: Don't WARN if userspace mucks with RCX during string I/O
exit
KVM: selftests: Add test to verify KVM doesn't explode on "bad" I/O

Quoted 6 lines
 arch/x86/kvm/x86.c                            |   9 +-
 tools/testing/selftests/kvm/.gitignore        |   1 +
 tools/testing/selftests/kvm/Makefile          |   1 +
 .../selftests/kvm/x86_64/userspace_io_test.c  | 114 ++++++++++++++++++
 4 files changed, 123 insertions(+), 2 deletions(-)
 create mode 100644 tools/testing/selftests/kvm/x86_64/userspace_io_test.c

Ping. I completely forgot about this too, until I unintentionally ran a userspace_io_test that was lying around.

maintainer Applied
#4 lore raw
Quoted 22 lines · Sean Christopherson
Quoted 14 lines · Sean Christopherson

Remove a WARN that was added as part of the recent I/O overhaul to play nice with SEV-ES string I/O.

For the record, my FIXME in lieu of a WARN was deliberate, as I suspected userspace could trigger a WARN ;-)

Based on kvm/master, commit 95e16b4792b0 ("KVM: SEV-ES: go over the sev_pio_data buffer in multiple passes if needed").

Sean Christopherson (2):
KVM: x86: Don't WARN if userspace mucks with RCX during string I/O
exit
KVM: selftests: Add test to verify KVM doesn't explode on "bad" I/O

Quoted 6 lines
  arch/x86/kvm/x86.c                            |   9 +-
  tools/testing/selftests/kvm/.gitignore        |   1 +
  tools/testing/selftests/kvm/Makefile          |   1 +
  .../selftests/kvm/x86_64/userspace_io_test.c  | 114 ++++++++++++++++++
  4 files changed, 123 insertions(+), 2 deletions(-)

create mode 100644 tools/testing/selftests/kvm/x86_64/userspace_io_test.c

Quoted 3 lines

Ping. I completely forgot about this too, until I unintentionally ran a userspace_io_test that was lying around.

Queued now, thanks. I don't know if I want the honor of having KVM singled out again on the -rc release message, but these are bugs nevertheless...

Paolo

author maintainer
#5 lore raw
Quoted 23 lines · Paolo Bonzini
Quoted 22 lines · Sean Christopherson
Quoted 20 lines · Sean Christopherson

Remove a WARN that was added as part of the recent I/O overhaul to play nice with SEV-ES string I/O.

For the record, my FIXME in lieu of a WARN was deliberate, as I suspected userspace could trigger a WARN ;-)

Based on kvm/master, commit 95e16b4792b0 ("KVM: SEV-ES: go over the sev_pio_data buffer in multiple passes if needed").

Sean Christopherson (2):
KVM: x86: Don't WARN if userspace mucks with RCX during string I/O
exit
KVM: selftests: Add test to verify KVM doesn't explode on "bad" I/O

  arch/x86/kvm/x86.c                            |   9 +-
  tools/testing/selftests/kvm/.gitignore        |   1 +
  tools/testing/selftests/kvm/Makefile          |   1 +
  .../selftests/kvm/x86_64/userspace_io_test.c  | 114 ++++++++++++++++++
  4 files changed, 123 insertions(+), 2 deletions(-)

create mode 100644 tools/testing/selftests/kvm/x86_64/userspace_io_test.c

Quoted 6 lines
Quoted 3 lines

Ping. I completely forgot about this too, until I unintentionally ran a userspace_io_test that was lying around.

Queued now, thanks. I don't know if I want the honor of having KVM singled out again on the -rc release message, but these are bugs nevertheless...

Heh, you could always wait an -rc to send the fixes and hope Linus has a short memory :-D