File xsa402-4.patch of Package xen.25148

Overview Repositories Revisions Requests Users Attributes Meta

File xsa402-4.patch of Package xen.25148

From: Andrew Cooper <andrew.cooper3@citrix.com>
Subject: x86/amd: Work around CLFLUSH ordering on older parts

On pre-CLFLUSHOPT AMD CPUs, CLFLUSH is weakely ordered with everything,
including reads and writes to the address, and LFENCE/SFENCE instructions.

This creates a multitude of problematic corner cases, laid out in the manual.
Arrange to use MFENCE on both sides of the CLFLUSH to force proper ordering.

This is part of XSA-402.

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>

--- a/xen/arch/x86/cpu/amd.c
+++ b/xen/arch/x86/cpu/amd.c
@@ -624,6 +624,14 @@ static void init_amd(struct cpuinfo_x86
 	if (!cpu_has_lfence_dispatch)
 		__set_bit(X86_FEATURE_MFENCE_RDTSC, c->x86_capability);
 
+	/*
+	 * On pre-CLFLUSHOPT AMD CPUs, CLFLUSH is weakly ordered with
+	 * everything, including reads and writes to address, and
+	 * LFENCE/SFENCE instructions.
+	 */
+	if (!boot_cpu_has(X86_FEATURE_CLFLUSHOPT))
+		setup_force_cpu_cap(X86_BUG_CLFLUSH_MFENCE);
+
 	switch(c->x86)
 	{
 	case 0xf ... 0x11:
--- a/xen/arch/x86/flushtlb.c
+++ b/xen/arch/x86/flushtlb.c
@@ -256,6 +256,13 @@ unsigned int flush_area_local(const void
     return flags;
 }
 
+/*
+ * On pre-CLFLUSHOPT AMD CPUs, CLFLUSH is weakly ordered with everything,
+ * including reads and writes to address, and LFENCE/SFENCE instructions.
+ *
+ * This function only works safely after alternatives have run.  Luckily, at
+ * the time of writing, we don't flush the caches that early.
+ */
 void cache_flush(const void *addr, unsigned int size)
 {
     /*
@@ -265,6 +272,8 @@ void cache_flush(const void *addr, unsig
     unsigned int clflush_size = current_cpu_data.x86_clflush_size ?: 16;
     const void *end = addr + size;
 
+    alternative(ASM_NOP3, "mfence", X86_BUG_CLFLUSH_MFENCE);
+
     addr -= (unsigned long)addr & (clflush_size - 1);
     for ( ; addr < end; addr += clflush_size )
     {
@@ -280,7 +289,9 @@ void cache_flush(const void *addr, unsig
                        [p] "m" (*(const char *)(addr)));
     }
 
-    alternative(ASM_NOP3, "sfence", X86_FEATURE_CLFLUSHOPT);
+    alternative_2(ASM_NOP3,
+                  "sfence", X86_FEATURE_CLFLUSHOPT,
+                  "mfence", X86_BUG_CLFLUSH_MFENCE);
 }
 
 void cache_writeback(const void *addr, unsigned int size)
--- a/xen/include/asm-x86/cpufeatures.h
+++ b/xen/include/asm-x86/cpufeatures.h
@@ -6,7 +6,8 @@
 
 #define FSCAPINTS FEATURESET_NR_ENTRIES
 
-#define NCAPINTS (FSCAPINTS + 1) /* N 32-bit words worth of info */
+/* Synthetic words follow the featureset words. */
+#define X86_NR_SYNTH 1
 
 /* Other features, Xen-defined mapping. */
 /* This range is used for feature bits which conflict or are synthesized */
@@ -37,3 +38,13 @@ XEN_CPUFEATURE(SC_VERW_PV,      (FSCAPIN
 XEN_CPUFEATURE(SC_VERW_HVM,     (FSCAPINTS+0)*32+24) /* VERW used by Xen for HVM */
 XEN_CPUFEATURE(SC_VERW_IDLE,    (FSCAPINTS+0)*32+25) /* VERW used by Xen for idle */
 XEN_CPUFEATURE(XEN_SELFSNOOP,   (FSCAPINTS+0)*32+26) /* SELFSNOOP gets used by Xen itself */
+
+/* Bug words follow the synthetic words. */
+#define X86_NR_BUG 1
+
+#define X86_BUG(x) ((FSCAPINTS + X86_NR_SYNTH) * 32 + (x))
+
+#define X86_BUG_CLFLUSH_MFENCE    X86_BUG( 2) /* MFENCE needed to serialise CLFLUSH */
+
+/* Total number of capability words, inc synth and bug words. */
+#define NCAPINTS (FSCAPINTS + X86_NR_SYNTH + X86_NR_BUG) /* N 32-bit words worth of info */

Places

File xsa402-4.patch of Package xen.25148

Places