x86: implement x86_32 stack protector

Impact: stack protector for x86_32

Implement stack protector for x86_32.  GDT entry 28 is used for it.
It's set to point to stack_canary-20 and have the length of 24 bytes.
CONFIG_CC_STACKPROTECTOR turns off CONFIG_X86_32_LAZY_GS and sets %gs
to the stack canary segment on entry.  As %gs is otherwise unused by
the kernel, the canary can be anywhere.  It's defined as a percpu
variable.

x86_32 exception handlers take register frame on stack directly as
struct pt_regs.  With -fstack-protector turned on, gcc copies the
whole structure after the stack canary and (of course) doesn't copy
back on return thus losing all changed.  For now, -fno-stack-protector
is added to all files which contain those functions.  We definitely
need something better.

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

6 files changed, 52 insertions, 8 deletions

diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile
index 37fa30bada17..b1f8be33300d 100644
--- a/arch/x86/kernel/Makefile
+++ b/arch/x86/kernel/Makefile
@@ -24,6 +24,24 @@ CFLAGS_vsyscall_64.o	:= $(PROFILING) -g0 $(nostackp)
 CFLAGS_hpet.o           := $(nostackp)
 CFLAGS_tsc.o            := $(nostackp)
 CFLAGS_paravirt.o       := $(nostackp)
+#
+# On x86_32, register frame is passed verbatim on stack as struct
+# pt_regs.  gcc considers the parameter to belong to the callee and
+# with -fstack-protector it copies pt_regs to the callee's stack frame
+# to put the structure after the stack canary causing changes made by
+# the exception handlers to be lost.  Turn off stack protector for all
+# files containing functions which take struct pt_regs from register
+# frame.
+#
+# The proper way to fix this is to teach gcc that the argument belongs
+# to the caller for these functions, oh well...
+#
+ifdef CONFIG_X86_32
+CFLAGS_process_32.o     := $(nostackp)
+CFLAGS_vm86_32.o        := $(nostackp)
+CFLAGS_signal.o         := $(nostackp)
+CFLAGS_traps.o          := $(nostackp)
+endif
 
 obj-y                   := process_$(BITS).o signal.o entry_$(BITS).o
 obj-y                   += traps.o irq.o irq_$(BITS).o dumpstack_$(BITS).o
diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
index 41b0de6df873..260fe4cb2c82 100644
--- a/arch/x86/kernel/cpu/common.c
+++ b/arch/x86/kernel/cpu/common.c
@@ -39,6 +39,7 @@
 #include <asm/sections.h>
 #include <asm/setup.h>
 #include <asm/hypervisor.h>
+#include <asm/stackprotector.h>
 
 #include "cpu.h"
 
@@ -122,6 +123,7 @@ DEFINE_PER_CPU_PAGE_ALIGNED(struct gdt_page, gdt_page) = { .gdt = {
 
         [GDT_ENTRY_ESPFIX_SS] = { { { 0x00000000, 0x00c09200 } } },
         [GDT_ENTRY_PERCPU] = { { { 0x0000ffff, 0x00cf9200 } } },
+        GDT_STACK_CANARY_INIT
 #endif
 } };
 EXPORT_PER_CPU_SYMBOL_GPL(gdt_page);
@@ -261,6 +263,7 @@ void load_percpu_segment(int cpu)
         loadsegment(gs, 0);
         wrmsrl(MSR_GS_BASE, (unsigned long)per_cpu(irq_stack_union.gs_base, cpu));
 #endif
+        load_stack_canary_segment();
 }
 
 /* Current gdt points %fs at the "master" per-cpu area: after this,
@@ -946,16 +949,21 @@ unsigned long kernel_eflags;
  */
 DEFINE_PER_CPU(struct orig_ist, orig_ist);
 
-#else
+#else   /* x86_64 */
+
+#ifdef CONFIG_CC_STACKPROTECTOR
+DEFINE_PER_CPU(unsigned long, stack_canary);
+#endif
 
-/* Make sure %fs is initialized properly in idle threads */
+/* Make sure %fs and %gs are initialized properly in idle threads */
 struct pt_regs * __cpuinit idle_regs(struct pt_regs *regs)
 {
         memset(regs, 0, sizeof(struct pt_regs));
         regs->fs = __KERNEL_PERCPU;
+        regs->gs = __KERNEL_STACK_CANARY;
         return regs;
 }
-#endif
+#endif  /* x86_64 */
 
 /*
  * cpu_init() initializes state that is per-CPU. Some data is already
@@ -1120,9 +1128,6 @@ void __cpuinit cpu_init(void)
         __set_tss_desc(cpu, GDT_ENTRY_DOUBLEFAULT_TSS, &doublefault_tss);
 #endif
 
-        /* Clear %gs. */
-        asm volatile ("mov %0, %%gs" : : "r" (0));
-
         /* Clear all 6 debug registers: */
         set_debugreg(0, 0);
         set_debugreg(0, 1);
diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S
index 82e6868bee47..5f5bd22adcd4 100644
--- a/arch/x86/kernel/entry_32.S
+++ b/arch/x86/kernel/entry_32.S
@@ -186,7 +186,7 @@
         /*CFI_REL_OFFSET gs, PT_GS*/
 .endm
 .macro SET_KERNEL_GS reg
-        xorl \reg, \reg
+        movl $(__KERNEL_STACK_CANARY), \reg
         movl \reg, %gs
 .endm
 
diff --git a/arch/x86/kernel/head_32.S b/arch/x86/kernel/head_32.S
index 24c0e5cd71e3..924e31615fb6 100644
--- a/arch/x86/kernel/head_32.S
+++ b/arch/x86/kernel/head_32.S
@@ -19,6 +19,7 @@
 #include <asm/asm-offsets.h>
 #include <asm/setup.h>
 #include <asm/processor-flags.h>
+#include <asm/percpu.h>
 
 /* Physical address */
 #define pa(X) ((X) - __PAGE_OFFSET)
@@ -437,8 +438,25 @@ is386:	movl $2,%ecx		# set MP
         movl $(__KERNEL_PERCPU), %eax
         movl %eax,%fs                   # set this cpu's percpu
 
-        xorl %eax,%eax                  # Clear GS and LDT
+#ifdef CONFIG_CC_STACKPROTECTOR
+        /*
+         * The linker can't handle this by relocation.  Manually set
+         * base address in stack canary segment descriptor.
+         */
+        cmpb $0,ready
+        jne 1f
+        movl $per_cpu__gdt_page,%eax
+        movl $per_cpu__stack_canary,%ecx
+        movw %cx, 8 * GDT_ENTRY_STACK_CANARY + 2(%eax)
+        shrl $16, %ecx
+        movb %cl, 8 * GDT_ENTRY_STACK_CANARY + 4(%eax)
+        movb %ch, 8 * GDT_ENTRY_STACK_CANARY + 7(%eax)
+1:
+#endif
+        movl $(__KERNEL_STACK_CANARY),%eax
         movl %eax,%gs
+
+        xorl %eax,%eax                  # Clear LDT
         lldt %ax
 
         cld                     # gcc2 wants the direction flag cleared at all times
diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c
index 86122fa2a1ba..9a62383e7c3c 100644
--- a/arch/x86/kernel/process_32.c
+++ b/arch/x86/kernel/process_32.c
@@ -212,6 +212,7 @@ int kernel_thread(int (*fn)(void *), void *arg, unsigned long flags)
         regs.ds = __USER_DS;
         regs.es = __USER_DS;
         regs.fs = __KERNEL_PERCPU;
+        regs.gs = __KERNEL_STACK_CANARY;
         regs.orig_ax = -1;
         regs.ip = (unsigned long) kernel_thread_helper;
         regs.cs = __KERNEL_CS | get_kernel_rpl();
diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c
index ef91747bbed5..d992e6cff730 100644
--- a/arch/x86/kernel/setup_percpu.c
+++ b/arch/x86/kernel/setup_percpu.c
@@ -16,6 +16,7 @@
 #include <asm/proto.h>
 #include <asm/cpumask.h>
 #include <asm/cpu.h>
+#include <asm/stackprotector.h>
 
 #ifdef CONFIG_DEBUG_PER_CPU_MAPS
 # define DBG(x...) printk(KERN_DEBUG x)
@@ -95,6 +96,7 @@ void __init setup_per_cpu_areas(void)
                 per_cpu(this_cpu_off, cpu) = per_cpu_offset(cpu);
                 per_cpu(cpu_number, cpu) = cpu;
                 setup_percpu_segment(cpu);
+                setup_stack_canary_segment(cpu);
                 /*
                  * Copy data used in early init routines from the
                  * initial arrays to the per cpu data areas.  These