x86: Make x32 syscall support conditional on a kernel parameter

Enabling x32 in the standard amd64 kernel would increase its attack
surface while provide no benefit to the vast majority of its users.
No-one seems interested in regularly checking for vulnerabilities
specific to x32 (at least no-one with a white hat).

Still, adding another flavour just to turn on x32 seems wasteful.  And
the only differences on syscall entry are two instructions (mask out
the x32 flag and compare the syscall number).

So pad the standard comparison with a nop and add a kernel parameter
"syscall.x32" which controls whether this is replaced with the x32
version at boot time.  Add a Kconfig parameter to set the default.

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Gbp-Pq: Topic features/x86
Gbp-Pq: Name x86-make-x32-syscall-support-conditional.patch

diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt
index 7737ab5d04b2941621df23b7fc83a4cdba36e5ba..4399d13a4579a12c4e03889358220310d9cf6ec7 100644 (file)
--- a/Documentation/admin-guide/kernel-parameters.txt
+++ b/Documentation/admin-guide/kernel-parameters.txt
@@ -3906,6 +3906,10 @@
 
        switches=       [HW,M68k]
 
+       syscall.x32=    [KNL,x86_64] Enable/disable use of x32 syscalls on
+                       an x86_64 kernel where CONFIG_X86_X32 is enabled.
+                       Default depends on CONFIG_X86_X32_DISABLED.
+
        sysfs.deprecated=0|1 [KNL]
                        Enable/disable old style sysfs layout for old udev
                        on older distributions. When this option is enabled

diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 0efb4c9497bce7e491665688840c4ced8e2c66c6..6e043a136a1eb245df877cca8397f54fcadc8f12 100644 (file)
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -2765,6 +2765,14 @@ config COMPAT_32
        select HAVE_UID16
        select OLD_SIGSUSPEND3
 
+config X86_X32_DISABLED
+       bool "x32 ABI disabled by default"
+       depends on X86_X32
+       default n
+       help
+         Disable the x32 ABI unless explicitly enabled using the
+         kernel paramter "syscall.x32=y".
+
 config COMPAT
        def_bool y
        depends on IA32_EMULATION || X86_X32

diff --git a/arch/x86/entry/common.c b/arch/x86/entry/common.c
index cdefcfdd9e63effa6a06660adf013c4fe96ea853..18616cb068b5344e7c82a7d4017b2a657493bc8f 100644 (file)
--- a/arch/x86/entry/common.c
+++ b/arch/x86/entry/common.c
@@ -268,6 +268,7 @@ __visible void do_syscall_64(struct pt_regs *regs)
 {
        struct thread_info *ti = current_thread_info();
        unsigned long nr = regs->orig_ax;
+       unsigned int syscall_mask, nr_syscalls_enabled;
 
        enter_from_user_mode();
        local_irq_enable();
@@ -280,8 +281,19 @@ __visible void do_syscall_64(struct pt_regs *regs)
         * table.  The only functional difference is the x32 bit in
         * regs->orig_ax, which changes the behavior of some syscalls.
         */
-       if (likely((nr & __SYSCALL_MASK) < NR_syscalls)) {
-               regs->ax = sys_call_table[nr & __SYSCALL_MASK](
+       if (__SYSCALL_MASK == ~0U || x32_enabled) {
+               syscall_mask = __SYSCALL_MASK;
+               nr_syscalls_enabled = NR_syscalls;
+       } else {
+               /*
+                * x32 syscalls present but not enabled.  Don't mask out
+                * the x32 flag and don't enable any x32-specific calls.
+                */
+               syscall_mask = ~0U;
+               nr_syscalls_enabled = 512;
+       }
+       if (likely((nr & syscall_mask) < nr_syscalls_enabled)) {
+               regs->ax = sys_call_table[nr & syscall_mask](
                        regs->di, regs->si, regs->dx,
                        regs->r10, regs->r8, regs->r9);
        }

diff --git a/arch/x86/entry/entry_64.S b/arch/x86/entry/entry_64.S
index 4a4c0834f9659bd9b4855e23801d47103e644f09..b612a6ea4f683b91f89ccd8003d1b9839548eb41 100644 (file)
--- a/arch/x86/entry/entry_64.S
+++ b/arch/x86/entry/entry_64.S
@@ -189,8 +189,12 @@ entry_SYSCALL_64_fastpath:
 #if __SYSCALL_MASK == ~0
        cmpq    $__NR_syscall_max, %rax
 #else
-       andl    $__SYSCALL_MASK, %eax
-       cmpl    $__NR_syscall_max, %eax
+.global system_call_fast_compare
+.global system_call_fast_compare_end
+system_call_fast_compare:
+       cmpq    $511, %rax                      /* x32 syscalls start at 512 */
+       .byte   P6_NOP4
+system_call_fast_compare_end:
 #endif
        ja      1f                              /* return -ENOSYS (already in pt_regs->ax) */
        movq    %r10, %rcx
@@ -322,6 +326,16 @@ opportunistic_sysret_failed:
        jmp     restore_c_regs_and_iret
 END(entry_SYSCALL_64)
 
+#if __SYSCALL_MASK != ~0
+       /* This replaces the usual comparisons if syscall.x32 is set */
+.global system_call_mask_compare
+.global system_call_mask_compare_end
+system_call_mask_compare:
+       andl    $__SYSCALL_MASK, %eax
+       cmpl    $__NR_syscall_max, %eax
+system_call_mask_compare_end:
+#endif
+
 ENTRY(stub_ptregs_64)
        /*
         * Syscalls marked as needing ptregs land here.

diff --git a/arch/x86/entry/syscall_64.c b/arch/x86/entry/syscall_64.c
index 9dbc5abb6162fa20581069499667a8c49b254868..753eeae0a72e0fac8fca10cc20112a686ad6edc0 100644 (file)
--- a/arch/x86/entry/syscall_64.c
+++ b/arch/x86/entry/syscall_64.c
@@ -3,8 +3,14 @@
 #include <linux/linkage.h>
 #include <linux/sys.h>
 #include <linux/cache.h>
+#include <linux/moduleparam.h>
+#undef MODULE_PARAM_PREFIX
+#define MODULE_PARAM_PREFIX "syscall."
+#include <linux/bug.h>
+#include <linux/init.h>
 #include <asm/asm-offsets.h>
 #include <asm/syscall.h>
+#include <asm/text-patching.h>
 
 #define __SYSCALL_64_QUAL_(sym) sym
 #define __SYSCALL_64_QUAL_ptregs(sym) ptregs_##sym
@@ -25,3 +31,36 @@ asmlinkage const sys_call_ptr_t sys_call_table[__NR_syscall_max+1] = {
        [0 ... __NR_syscall_max] = &sys_ni_syscall,
 #include <asm/syscalls_64.h>
 };
+
+#ifdef CONFIG_X86_X32_ABI
+
+/* Maybe enable x32 syscalls */
+
+bool x32_enabled = !IS_ENABLED(CONFIG_X86_X32_DISABLED);
+module_param_named(x32, x32_enabled, bool, 0444);
+
+extern char system_call_fast_compare_end[], system_call_fast_compare[],
+       system_call_mask_compare_end[], system_call_mask_compare[];
+
+static int __init x32_enable(void)
+{
+       BUG_ON(system_call_fast_compare_end - system_call_fast_compare != 10);
+       BUG_ON(system_call_mask_compare_end - system_call_mask_compare != 10);
+
+       if (x32_enabled) {
+               text_poke_early(system_call_fast_compare,
+                               system_call_mask_compare, 10);
+#ifdef CONFIG_X86_X32_DISABLED
+               pr_info("Enabled x32 syscalls\n");
+#endif
+       }
+#ifndef CONFIG_X86_X32_DISABLED
+       else
+               pr_info("Disabled x32 syscalls\n");
+#endif
+
+       return 0;
+}
+late_initcall(x32_enable);
+
+#endif

diff --git a/arch/x86/include/asm/elf.h b/arch/x86/include/asm/elf.h
index 9aeb91935ce02387d8dae5e2f51bf1750f420a43..0145d28b9177def9ff4993b5c83633053bfef55e 100644 (file)
--- a/arch/x86/include/asm/elf.h
+++ b/arch/x86/include/asm/elf.h
@@ -9,6 +9,7 @@
 #include <asm/ptrace.h>
 #include <asm/user.h>
 #include <asm/auxvec.h>
+#include <asm/syscall.h>
 
 typedef unsigned long elf_greg_t;
 
@@ -162,7 +163,7 @@ do {                                                \
 
 #define compat_elf_check_arch(x)                                       \
        (elf_check_arch_ia32(x) ||                                      \
-        (IS_ENABLED(CONFIG_X86_X32_ABI) && (x)->e_machine == EM_X86_64))
+        (x32_enabled && (x)->e_machine == EM_X86_64))
 
 #if __USER32_DS != __USER_DS
 # error "The following code assumes __USER32_DS == __USER_DS"

diff --git a/arch/x86/include/asm/syscall.h b/arch/x86/include/asm/syscall.h
index e3c95e8e61c5fca9808524ba47d539bd0377ece2..4cf8e121d874c52a5e37558255d24ee54c7c264e 100644 (file)
--- a/arch/x86/include/asm/syscall.h
+++ b/arch/x86/include/asm/syscall.h
@@ -35,6 +35,12 @@ extern const sys_call_ptr_t sys_call_table[];
 extern const sys_call_ptr_t ia32_sys_call_table[];
 #endif
 
+#if defined(CONFIG_X86_X32_ABI)
+extern bool x32_enabled;
+#else
+#define x32_enabled 0
+#endif
+
 /*
  * Only the low 32 bits of orig_ax are meaningful, so we return int.
  * This importantly ignores the high bits on 64-bit, so comparisons