x86-64: Add user_64bit_mode paravirt op

Three places in the kernel assume that the only long mode CPL 3
selector is __USER_CS.  This is not true on Xen -- Xen's sysretq
changes cs to the magic value 0xe033.

Two of the places are corner cases, but as of "x86-64: Improve
vsyscall emulation CS and RIP handling"
(c9712944b2a12373cb6ff8059afcfb7e826a6c54), vsyscalls will segfault
if called with Xen's extra CS selector.  This causes a panic when
older init builds die.

It seems impossible to make Xen use __USER_CS reliably without
taking a performance hit on every system call, so this fixes the
tests instead with a new paravirt op.  It's a little ugly because
ptrace.h can't include paravirt.h.

Signed-off-by: Andy Lutomirski <luto@mit.edu>
Link: http://lkml.kernel.org/r/f4fcb3947340d9e96ce1054a432f183f9da9db83.1312378163.git.luto@mit.edu
Reported-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>

3 files changed, 27 insertions, 2 deletions

diff --git a/arch/x86/include/asm/desc.h b/arch/x86/include/asm/desc.h
index 7b439d9aea2..41935fadfdf 100644
--- a/arch/x86/include/asm/desc.h
+++ b/arch/x86/include/asm/desc.h
@@ -27,8 +27,8 @@ static inline void fill_ldt(struct desc_struct *desc, const struct user_desc *in
 
         desc->base2             = (info->base_addr & 0xff000000) >> 24;
         /*
-         * Don't allow setting of the lm bit. It is useless anyway
-         * because 64bit system calls require __USER_CS:
+         * Don't allow setting of the lm bit. It would confuse
+         * user_64bit_mode and would get overridden by sysret anyway.
          */
         desc->l                 = 0;
 }
diff --git a/arch/x86/include/asm/paravirt_types.h b/arch/x86/include/asm/paravirt_types.h
index 82885099c86..96a0f80407b 100644
--- a/arch/x86/include/asm/paravirt_types.h
+++ b/arch/x86/include/asm/paravirt_types.h
@@ -41,6 +41,7 @@
 
 #include <asm/desc_defs.h>
 #include <asm/kmap_types.h>
+#include <asm/pgtable_types.h>
 
 struct page;
 struct thread_struct;
@@ -63,6 +64,11 @@ struct paravirt_callee_save {
 struct pv_info {
         unsigned int kernel_rpl;
         int shared_kernel_pmd;
+
+#ifdef CONFIG_X86_64
+        u16 extra_user_64bit_cs;  /* __USER_CS if none */
+#endif
+
         int paravirt_enabled;
         const char *name;
 };
diff --git a/arch/x86/include/asm/ptrace.h b/arch/x86/include/asm/ptrace.h
index 94e7618fcac..35664547125 100644
--- a/arch/x86/include/asm/ptrace.h
+++ b/arch/x86/include/asm/ptrace.h
@@ -131,6 +131,9 @@ struct pt_regs {
 #ifdef __KERNEL__
 
 #include <linux/init.h>
+#ifdef CONFIG_PARAVIRT
+#include <asm/paravirt_types.h>
+#endif
 
 struct cpuinfo_x86;
 struct task_struct;
@@ -187,6 +190,22 @@ static inline int v8086_mode(struct pt_regs *regs)
 #endif
 }
 
+#ifdef CONFIG_X86_64
+static inline bool user_64bit_mode(struct pt_regs *regs)
+{
+#ifndef CONFIG_PARAVIRT
+        /*
+         * On non-paravirt systems, this is the only long mode CPL 3
+         * selector.  We do not allow long mode selectors in the LDT.
+         */
+        return regs->cs == __USER_CS;
+#else
+        /* Headers are too twisted for this to go in paravirt.h. */
+        return regs->cs == __USER_CS || regs->cs == pv_info.extra_user_64bit_cs;
+#endif
+}
+#endif
+
 /*
  * X86_32 CPUs don't save ss and esp if the CPU is already in kernel mode
  * when it traps.  The previous stack will be directly underneath the saved