x86/shstk: Activate Supervisor Shadow Stacks

With all other plumbing in place, activate shadow stacks when possible.

Note that CET shares the similar problems to SMEP/SMAP with Ring1 being
supervisor to the processor, and that the layout of the shadow stack differs
between an IRET to Ring 1 and Ring 3.  Therefore, we disable PV32 when CET is
enabled.  Compatibility can be maintained if necessary via PV-Shim.

The BSP needs to wait until alternatives have run (to avoid interaction with
CR0.WP), and after the first reset_stack_and_jump() to avoid having a pristine
shadow stack interact in problematic ways with an in-use regular stack.
Activate shadow stack in reinit_bsp_stack().

APs have all infrastructure set up by the booting CPU, so enable shadow stacks
before entering C.  Adjust the logic to call start_secondary rather than jump
to it, so stack traces make more sense.

The crash path needs to turn CET off to avoid interfering with the crash
kernel's environment.

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

diff --git a/docs/misc/xen-command-line.pandoc b/docs/misc/xen-command-line.pandoc
index ec1c12cbe02d471da6e52a287c862ec214e7cd35..2d4d9639d61abf9b9add9562ed76244f4e2075bb 100644 (file)
--- a/docs/misc/xen-command-line.pandoc
+++ b/docs/misc/xen-command-line.pandoc
@@ -287,6 +287,10 @@ call/jmp COP/JOP) attacks.
     `cet=no-shstk` will cause Xen not to use Shadow Stacks even when support
     is available in hardware.
 
+    Shadow Stacks are incompatible with 32bit PV guests.  This option will
+    override the `pv=32` boolean to false.  Backwards compatibility can be
+    maintained with the `pv-shim` mechanism.
+
 ### clocksource (x86)
 > `= pit | hpet | acpi | tsc`
 
@@ -1726,6 +1730,10 @@ Controls for aspects of PV guest support.
 *   The `32` boolean controls whether 32bit PV guests can be created.  It
     defaults to `true`, and is ignored when `CONFIG_PV32` is compiled out.
 
+    32bit PV guests are incompatible with CET Shadow Stacks.  If Xen is using
+    shadow stacks, this option will be overridden to `false`.  Backwards
+    compatibility can be maintained with the `pv-shim` mechanism.
+
 ### pv-linear-pt (x86)
 > `= <boolean>`
 

diff --git a/xen/arch/x86/boot/x86_64.S b/xen/arch/x86/boot/x86_64.S
index 314a32a19ff46e78d518f908b42f11aa2943c2d4..551acd9e9447f56ceb0285c1a3ca7c354acf9de4 100644 (file)
--- a/xen/arch/x86/boot/x86_64.S
+++ b/xen/arch/x86/boot/x86_64.S
@@ -28,8 +28,39 @@ ENTRY(__high_start)
         lretq
 1:
         test    %ebx,%ebx
-        jnz     start_secondary
-
+        jz      .L_bsp
+
+        /* APs.  Set up shadow stacks before entering C. */
+
+        testl   $cpufeat_mask(X86_FEATURE_XEN_SHSTK), \
+                CPUINFO_FEATURE_OFFSET(X86_FEATURE_XEN_SHSTK) + boot_cpu_data(%rip)
+        je      .L_ap_shstk_done
+
+        /* Set up MSR_S_CET. */
+        mov     $MSR_S_CET, %ecx
+        xor     %edx, %edx
+        mov     $CET_SHSTK_EN | CET_WRSS_EN, %eax
+        wrmsr
+
+        /* Derive MSR_PL0_SSP from %rsp (token written when stack is allocated). */
+        mov     $MSR_PL0_SSP, %ecx
+        mov     %rsp, %rdx
+        shr     $32, %rdx
+        mov     %esp, %eax
+        and     $~(STACK_SIZE - 1), %eax
+        or      $(PRIMARY_SHSTK_SLOT + 1) * PAGE_SIZE - 8, %eax
+        wrmsr
+
+        /* Enable CET.  MSR_INTERRUPT_SSP_TABLE is set up later in load_system_tables(). */
+        mov     $XEN_MINIMAL_CR4 | X86_CR4_CET, %ecx
+        mov     %rcx, %cr4
+        setssbsy
+
+.L_ap_shstk_done:
+        call    start_secondary
+        BUG     /* start_secondary() shouldn't return. */
+
+.L_bsp:
         /* Pass off the Multiboot info structure to C land (if applicable). */
         mov     multiboot_ptr(%rip),%edi
         call    __start_xen

diff --git a/xen/arch/x86/cpu/common.c b/xen/arch/x86/cpu/common.c
index f483c81b7c238bd4f126a82e38599c78d5417fbb..58f0876180f8d2302778f9ad56227d849afd923a 100644 (file)
--- a/xen/arch/x86/cpu/common.c
+++ b/xen/arch/x86/cpu/common.c
@@ -329,6 +329,11 @@ void __init early_cpu_init(void)
               x86_cpuid_vendor_to_str(c->x86_vendor), c->x86, c->x86,
               c->x86_model, c->x86_model, c->x86_mask, eax);
 
+       if (c->cpuid_level >= 7) {
+               cpuid_count(7, 0, &eax, &ebx, &ecx, &edx);
+               c->x86_capability[cpufeat_word(X86_FEATURE_CET_SS)] = ecx;
+       }
+
        eax = cpuid_eax(0x80000000);
        if ((eax >> 16) == 0x8000 && eax >= 0x80000008) {
                eax = cpuid_eax(0x80000008);

diff --git a/xen/arch/x86/crash.c b/xen/arch/x86/crash.c
index 450eecd46baf4c1777d949933b57e6724458cb51..0611b4fb9b09a5692b5c7f45681f168714845a86 100644 (file)
--- a/xen/arch/x86/crash.c
+++ b/xen/arch/x86/crash.c
@@ -200,6 +200,13 @@ void machine_crash_shutdown(void)
     /* Reset CPUID masking and faulting to the host's default. */
     ctxt_switch_levelling(NULL);
 
+    /* Disable shadow stacks. */
+    if ( cpu_has_xen_shstk )
+    {
+        wrmsrl(MSR_S_CET, 0);
+        write_cr4(read_cr4() & ~X86_CR4_CET);
+    }
+
     info = kexec_crash_save_info();
     info->xen_phys_start = xen_phys_start;
     info->dom0_pfn_to_mfn_frame_list_list =

diff --git a/xen/arch/x86/setup.c b/xen/arch/x86/setup.c
index 584589baffad9e47bbac9a4a5a18be8b7993eff0..665274674dc302da1bd39168f2d91bfdf8e8f4a8 100644 (file)
--- a/xen/arch/x86/setup.c
+++ b/xen/arch/x86/setup.c
@@ -664,6 +664,14 @@ static void __init noreturn reinit_bsp_stack(void)
     stack_base[0] = stack;
     memguard_guard_stack(stack);
 
+    if ( cpu_has_xen_shstk )
+    {
+        wrmsrl(MSR_PL0_SSP,
+               (unsigned long)stack + (PRIMARY_SHSTK_SLOT + 1) * PAGE_SIZE - 8);
+        wrmsrl(MSR_S_CET, CET_SHSTK_EN | CET_WRSS_EN);
+        asm volatile ("setssbsy" ::: "memory");
+    }
+
     reset_stack_and_jump_nolp(init_done);
 }
 
@@ -1065,6 +1073,21 @@ void __init noreturn __start_xen(unsigned long mbi_p)
     /* This must come before e820 code because it sets paddr_bits. */
     early_cpu_init();
 
+    /* Choose shadow stack early, to set infrastructure up appropriately. */
+    if ( opt_xen_shstk && boot_cpu_has(X86_FEATURE_CET_SS) )
+    {
+        printk("Enabling Supervisor Shadow Stacks\n");
+
+        setup_force_cpu_cap(X86_FEATURE_XEN_SHSTK);
+#ifdef CONFIG_PV32
+        if ( opt_pv32 )
+        {
+            opt_pv32 = 0;
+            printk("  - Disabling PV32 due to Shadow Stacks\n");
+        }
+#endif
+    }
+
     /* Sanitise the raw E820 map to produce a final clean version. */
     max_page = raw_max_page = init_e820(memmap_type, &e820_raw);
 
@@ -1801,6 +1824,10 @@ void __init noreturn __start_xen(unsigned long mbi_p)
 
     alternative_branches();
 
+    /* Defer CR4.CET until alternatives have finished playing with CR0.WP */
+    if ( cpu_has_xen_shstk )
+        set_in_cr4(X86_CR4_CET);
+
     /*
      * NB: when running as a PV shim VCPUOP_up/down is wired to the shim
      * physical cpu_add/remove functions, so launch the guest with only

diff --git a/xen/arch/x86/spec_ctrl.c b/xen/arch/x86/spec_ctrl.c
index c5d8e587a805b7d695d1fef0d81dccac95b8cfb5..a94be2d59496e97dc13ba9401e65083de3bbc576 100644 (file)
--- a/xen/arch/x86/spec_ctrl.c
+++ b/xen/arch/x86/spec_ctrl.c
@@ -881,6 +881,14 @@ void __init init_speculation_mitigations(void)
 
     hw_smt_enabled = check_smt_enabled();
 
+    /*
+     * First, disable the use of retpolines if Xen is using shadow stacks, as
+     * they are incompatible.
+     */
+    if ( cpu_has_xen_shstk &&
+         (opt_thunk == THUNK_DEFAULT || opt_thunk == THUNK_RETPOLINE) )
+        thunk = THUNK_JMP;
+
     /*
      * Has the user specified any custom BTI mitigations?  If so, follow their
      * instructions exactly and disable all heuristics.