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

[Xen-changelog] [xen master] xen/bitops: Provide helpers to safely modify guest memory atomically



commit 6a5f01a57a662565e6aa63fc9f3081fa69e54465
Author:     Julien Grall <julien.grall@xxxxxxx>
AuthorDate: Mon Apr 29 15:05:26 2019 +0100
Commit:     Julien Grall <julien.grall@xxxxxxx>
CommitDate: Fri Jun 14 14:27:32 2019 +0100

    xen/bitops: Provide helpers to safely modify guest memory atomically
    
    On Arm, exclusive load-store atomics should only be used between trusted
    thread. As not all the guests are trusted, it may be possible to DoS Xen
    when updating shared memory with guest atomically.
    
    This patch adds a new set of helper that will update the guest memory
    safely. For x86, it is already possible to use the current helpers
    safely. So just wrap them.
    
    For Arm, we will first attempt to update the guest memory with the loop
    bounded by a maximum number of iterations. If it fails, we will pause the
    domain and try again.
    
    Note that this heuristics assumes that a page can only be shared between
    Xen and one domain. Not Xen and multiple domain.
    
    The maximum number of iterations is based on how many times a simple
    load-store atomic operation can be executed in 1uS. The maximum value is
    per-CPU to cater big.LITTLE and calculated when the CPU is booting. The
    heuristic was randomly chosen and can be modified if impact too much
    good-behaving guest.
    
    Note, while test_bit does not requires to use atomic operation, a
    wrapper for test_bit was added for completeness. In this case, the
    domain stays constified to avoid major rework in the caller for the
    time-being.
    
    This is part of XSA-295.
    
    Signed-of-by: Julien Grall <julien.grall@xxxxxxx>
    Acked-by: Jan Beulich <jbeulich@xxxxxxxx>
    Reviewed-by: Stefano Stabellini <sstabellini@xxxxxxxxxx>
---
 xen/arch/arm/Makefile               |  1 +
 xen/arch/arm/guest_atomics.c        | 91 +++++++++++++++++++++++++++++++++++++
 xen/include/asm-arm/guest_atomics.h | 76 +++++++++++++++++++++++++++++++
 xen/include/asm-x86/guest_atomics.h | 30 ++++++++++++
 4 files changed, 198 insertions(+)

diff --git a/xen/arch/arm/Makefile b/xen/arch/arm/Makefile
index cb902cb6fe..872a155b60 100644
--- a/xen/arch/arm/Makefile
+++ b/xen/arch/arm/Makefile
@@ -22,6 +22,7 @@ obj-$(CONFIG_GICV3) += gic-v3.o
 obj-$(CONFIG_HAS_ITS) += gic-v3-its.o
 obj-$(CONFIG_HAS_ITS) += gic-v3-lpi.o
 obj-y += guestcopy.o
+obj-y += guest_atomics.o
 obj-y += guest_walk.o
 obj-y += hvm.o
 obj-y += io.o
diff --git a/xen/arch/arm/guest_atomics.c b/xen/arch/arm/guest_atomics.c
new file mode 100644
index 0000000000..1b78a062f0
--- /dev/null
+++ b/xen/arch/arm/guest_atomics.c
@@ -0,0 +1,91 @@
+/*
+ * arch/arm/guest_atomics.c
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program; If not, see <http://www.gnu.org/licenses/>.
+ */
+#include <xen/cpu.h>
+
+#include <asm/guest_atomics.h>
+
+DEFINE_PER_CPU_READ_MOSTLY(unsigned int, guest_safe_atomic_max);
+
+/*
+ * Heuristic to find a safe upper-limit for load-store exclusive
+ * operations on memory shared with guest.
+ *
+ * At the moment, we calculate the number of iterations of a simple
+ * load-store atomic loop in 1uS.
+ */
+static void calibrate_safe_atomic(void)
+{
+    s_time_t deadline = NOW() + MICROSECS(1);
+    unsigned int counter = 0;
+    unsigned long mem = 0;
+
+    do
+    {
+        unsigned long res, tmp;
+
+#ifdef CONFIG_ARM_32
+        asm volatile (" ldrex   %2, %1\n"
+                      " add     %2, %2, #1\n"
+                      " strex   %0, %2, %1\n"
+                      : "=&r" (res), "+Q" (mem), "=&r" (tmp));
+#else
+        asm volatile (" ldxr    %w2, %1\n"
+                      " add     %w2, %w2, #1\n"
+                      " stxr    %w0, %w2, %1\n"
+                      : "=&r" (res), "+Q" (mem), "=&r" (tmp));
+#endif
+        counter++;
+    } while (NOW() < deadline);
+
+    this_cpu(guest_safe_atomic_max) = counter;
+
+    printk(XENLOG_DEBUG
+           "CPU%u: Guest atomics will try %u times before pausing the 
domain\n",
+           smp_processor_id(), counter);
+}
+
+static int cpu_guest_safe_atomic_callback(struct notifier_block *nfb,
+                                          unsigned long action,
+                                          void *hcpu)
+{
+    if ( action == CPU_STARTING )
+        calibrate_safe_atomic();
+
+    return NOTIFY_DONE;
+}
+
+static struct notifier_block cpu_guest_safe_atomic_nfb = {
+    .notifier_call = cpu_guest_safe_atomic_callback,
+};
+
+static int __init guest_safe_atomic_init(void)
+{
+    register_cpu_notifier(&cpu_guest_safe_atomic_nfb);
+
+    calibrate_safe_atomic();
+
+    return 0;
+}
+presmp_initcall(guest_safe_atomic_init);
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff --git a/xen/include/asm-arm/guest_atomics.h 
b/xen/include/asm-arm/guest_atomics.h
new file mode 100644
index 0000000000..4f127fda41
--- /dev/null
+++ b/xen/include/asm-arm/guest_atomics.h
@@ -0,0 +1,76 @@
+#ifndef _ARM_GUEST_ATOMICS_H
+#define _ARM_GUEST_ATOMICS_H
+
+#include <xen/bitops.h>
+#include <xen/sched.h>
+
+/*
+ * The guest atomics helpers shares the same logic. We first try to use
+ * the *_timeout version of the operation. If it didn't timeout, then we
+ * successfully updated the memory. Nothing else to do.
+ *
+ * If it did timeout, then it means we didn't manage to update the
+ * memory. This is possibly because the guest is misbehaving (i.e tight
+ * store loop) but can also happen for other reasons (i.e nested Xen).
+ * In that case pause the domain and retry the operation, this time
+ * without a timeout.
+ *
+ * Note, those helpers rely on other part of the code to prevent sharing
+ * a page between Xen and multiple domain.
+ */
+
+DECLARE_PER_CPU(unsigned int, guest_safe_atomic_max);
+
+#define guest_bitop(name)                                                   \
+static inline void guest_##name(struct domain *d, int nr, volatile void *p) \
+{                                                                           \
+    if ( name##_timeout(nr, p, this_cpu(guest_safe_atomic_max)) )           \
+        return;                                                             \
+                                                                            \
+    domain_pause_nosync(d);                                                 \
+    name(nr, p);                                                            \
+    domain_unpause(d);                                                      \
+}
+
+#define guest_testop(name)                                                  \
+static inline int guest_##name(struct domain *d, int nr, volatile void *p)  \
+{                                                                           \
+    bool succeed;                                                           \
+    int oldbit;                                                             \
+                                                                            \
+    succeed = name##_timeout(nr, p, &oldbit,                                \
+                             this_cpu(guest_safe_atomic_max));              \
+    if ( succeed )                                                          \
+        return oldbit;                                                      \
+                                                                            \
+    domain_pause_nosync(d);                                                 \
+    oldbit = name(nr, p);                                                   \
+    domain_unpause(d);                                                      \
+                                                                            \
+    return oldbit;                                                          \
+}
+
+guest_bitop(set_bit)
+guest_bitop(clear_bit)
+guest_bitop(change_bit)
+
+#undef guest_bitop
+
+/* test_bit does not use load-store atomic operations */
+#define guest_test_bit(d, nr, p) ((void)(d), test_bit(nr, p))
+
+guest_testop(test_and_set_bit)
+guest_testop(test_and_clear_bit)
+guest_testop(test_and_change_bit)
+
+#undef guest_testop
+
+#endif /* _ARM_GUEST_ATOMICS_H */
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff --git a/xen/include/asm-x86/guest_atomics.h 
b/xen/include/asm-x86/guest_atomics.h
new file mode 100644
index 0000000000..0c71d2d278
--- /dev/null
+++ b/xen/include/asm-x86/guest_atomics.h
@@ -0,0 +1,30 @@
+#ifndef _X86_GUEST_ATOMICS_H
+#define _X86_GUEST_ATOMICS_H
+
+#include <xen/bitops.h>
+
+/*
+ * It is safe to use the atomics helpers on x86 on memory shared with
+ * the guests.
+ */
+#define guest_set_bit(d, nr, p)     ((void)(d), set_bit(nr, p))
+#define guest_clear_bit(d, nr, p)   ((void)(d), clear_bit(nr, p))
+#define guest_change_bit(d, nr, p)  ((void)(d), change_bit(nr, p))
+#define guest_test_bit(d, nr, p)    ((void)(d), test_bit(nr, p))
+
+#define guest_test_and_set_bit(d, nr, p)    \
+    ((void)(d), test_and_set_bit(nr, p))
+#define guest_test_and_clear_bit(d, nr, p)  \
+    ((void)(d), test_and_clear_bit(nr, p))
+#define guest_test_and_change_bit(d, nr, p) \
+    ((void)(d), test_and_change_bit(nr, p))
+
+#endif /* _X86_GUEST_ATOMICS_H */
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
--
generated by git-patchbot for /home/xen/git/xen.git#master

_______________________________________________
Xen-changelog mailing list
Xen-changelog@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/xen-changelog

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.