From: Sean Christopherson <seanjc@google.com>
To: Paolo Bonzini <pbonzini@redhat.com>
Cc: kvm@vger.kernel.org, Sean Christopherson <seanjc@google.com>
Subject: [kvm-unit-tests PATCH 4/4] x86/debug: Add actual test coverage for Bus Lock Detect #DBs
Date: Fri, 25 Sep 2026 16:00:03 -0700 [thread overview]
Message-ID: <20260925230003.2362261-5-seanjc@google.com> (raw)
In-Reply-To: <20260925230003.2362261-1-seanjc@google.com>
Extend the debug test to verify that bus locks to atomic accesses that
split cache lines actually trigger Bus Lock Detect #DBs, while still
allowing the access to go through. Note, as already commented in the
test, Bus Lock Detect only applies to CPL>0.
Signed-off-by: Sean Christopherson <seanjc@google.com>
---
x86/debug.c | 48 +++++++++++++++++++++++++++++++++++++++---------
1 file changed, 39 insertions(+), 9 deletions(-)
diff --git a/x86/debug.c b/x86/debug.c
index 5427797f..d359a281 100644
--- a/x86/debug.c
+++ b/x86/debug.c
@@ -429,30 +429,60 @@ static noinline uint64_t bus_lock(uint64_t magic)
return READ_ONCE(*(uint64_t *)val);
}
-static void bus_lock_test(void)
+static void bus_lock_test(u64 debugctl)
{
const uint64_t magic = 0xdeadbeefdeadbeefull;
- bool bus_lock_db = false;
uint64_t val;
+ bool ign;
+
+ if (debugctl & DEBUGCTLMSR_BUS_LOCK_DETECT)
+ report_fail("DEBUGCTL.BUS_LOCK_DETECT should be '0' at RESET");
/*
* Generate a bus lock (via a locked access that splits cache lines)
* in CPL0 and again in CPL3 (Bus Lock Detect only affects CPL3), and
* verify that no #AC or #DB is generated (the relevant features are
- * not enabled).
+ * not yet enabled).
*/
val = bus_lock(magic);
report(!got_ac && !n && val == magic,
"CPL0 Split Lock #AC = %u (#DB = %u), val = %lx (wanted %lx)",
got_ac, n, val, magic);
- val = run_in_user((usermode_func)bus_lock, DB_VECTOR, magic, 0, 0, 0, &bus_lock_db);
- report(!bus_lock_db && val == magic,
- "CPL3 Bus Lock #DB = %u, val = %lx (wanted %lx)",
- bus_lock_db, val, magic);
+ val = run_in_user((usermode_func)bus_lock, GP_VECTOR, magic, 0, 0, 0, &ign);
+ report(!got_ac && !n && val == magic,
+ "CPL3 Bus Lock #AC = %u, #DB = %u, val = %lx (wanted %lx)",
+ got_ac, n, val, magic);
- n = 0;
got_ac = false;
+
+ if (!this_cpu_has(X86_FEATURE_BUS_LOCK_DETECT))
+ return;
+
+ wrmsr(MSR_IA32_DEBUGCTLMSR, debugctl | DEBUGCTLMSR_BUS_LOCK_DETECT);
+
+ val = bus_lock(magic);
+ report(!got_ac && !n && val == magic,
+ "CPL0 Split Lock shouldn't trigger Bus Lock Detect: #DB = %u, #AC = %u, val = %lx (wanted %lx)",
+ n, got_ac, val, magic);
+
+ val = run_in_user((usermode_func)bus_lock, GP_VECTOR, magic, 0, 0, 0, &ign);
+ report(n == 1 && !(dr6[0] & DR6_BUS_LOCK) && val == magic && !got_ac,
+ "CPL3 Bus Lock Detect #DB (nr #DBs = %u, DR6 = %lx, #AC = %u, val = %lx (wanted %lx))",
+ n, dr6[0], got_ac, val, magic);
+
+ n = 0;
+ write_dr7(DR7_FIXED_1 | DR7_GD);
+ write_dr7(DR7_FIXED_1);
+
+ /* Verify DR6.BLD is preserved on an unrelated #DB. */
+ report(n == 1 && !(dr6[0] & DR6_BUS_LOCK) && (dr6[0] & DR6_BD),
+ "DR6.BLD preserved on General Detect #DB (nr #DBs = %u, DR6 = %lx)",
+ n, dr6[0]);
+
+ n = 0;
+ write_dr6(dr6_control_value);
+ wrmsr(MSR_IA32_DEBUGCTLMSR, debugctl);
}
static void run_tests(unsigned long __dr6_control_value)
@@ -473,7 +503,7 @@ static void run_tests(unsigned long __dr6_control_value)
if (!this_cpu_has(X86_FEATURE_BUS_LOCK_DETECT))
dr6_base_value |= DR6_BUS_LOCK;
- bus_lock_test();
+ bus_lock_test(debugctl);
/*
* Enable Bus Lock Detect to workaround an AMD ucode bug where DR6.BLD
--
2.56.0.rc1.315.gc6ed9934b7-goog
next prev parent reply other threads:[~2026-09-25 23:00 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-09-25 22:59 [kvm-unit-tests PATCH 0/4] x86/debug: Add Bus Lock Detect coverage Sean Christopherson
2026-09-25 23:00 ` [kvm-unit-tests PATCH 1/4] x86/debug: Set all active-low DR6 bits when resetting DR6 Sean Christopherson
2026-09-25 23:00 ` [kvm-unit-tests PATCH 2/4] x86/debug: Run the "debug" test with the host CPU model Sean Christopherson
2026-09-25 23:00 ` [kvm-unit-tests PATCH 3/4] x86/debug: Verify that KVM correctly emulates DR6 active-low and fixed-1 semantics Sean Christopherson
2026-09-25 23:00 ` Sean Christopherson [this message]
2026-10-02 21:07 ` [kvm-unit-tests PATCH 0/4] x86/debug: Add Bus Lock Detect coverage Sean Christopherson
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20260925230003.2362261-5-seanjc@google.com \
--to=seanjc@google.com \
--cc=kvm@vger.kernel.org \
--cc=pbonzini@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox