Save/restore bound registers for _dl_runtime_profile

This patch saves and restores bound registers in x86-64 PLT for
ld.so profile and LD_AUDIT:

	* sysdeps/x86_64/bits/link.h (La_x86_64_regs): Add lr_bnd.
	(La_x86_64_retval): Add lrv_bnd0 and lrv_bnd1.
	* sysdeps/x86_64/dl-trampoline.S (_dl_runtime_profile): Save
	Intel MPX bound registers before _dl_profile_fixup.
	* sysdeps/x86_64/dl-trampoline.h: Restore Intel MPX bound
	registers after _dl_profile_fixup.  Save and restore bound
	registers bnd0/bnd1 when calling _dl_call_pltexit.
	* sysdeps/x86_64/link-defines.sym (BND_SIZE): New.
	(LR_BND_OFFSET): Likewise.
	(LRV_BND0_OFFSET): Likewise.
	(LRV_BND1_OFFSET): Likewise.

5 files changed, 77 insertions, 0 deletions

diff --git a/ChangeLog b/ChangeLog
index 722770992e..5efa3caee4 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,17 @@
+2014-04-16  Igor Zamyatin  <igor.zamyatin@intel.com>
+
+	* sysdeps/x86_64/bits/link.h (La_x86_64_regs): Add lr_bnd.
+	(La_x86_64_retval): Add lrv_bnd0 and lrv_bnd1.
+	* sysdeps/x86_64/dl-trampoline.S (_dl_runtime_profile): Save
+	Intel MPX bound registers before _dl_profile_fixup.
+	* sysdeps/x86_64/dl-trampoline.h: Restore Intel MPX bound
+	registers after _dl_profile_fixup.  Save and restore bound
+	registers bnd0/bnd1 when calling _dl_call_pltexit.
+	* sysdeps/x86_64/link-defines.sym (BND_SIZE): New.
+	(LR_BND_OFFSET): Likewise.
+	(LRV_BND0_OFFSET): Likewise.
+	(LRV_BND1_OFFSET): Likewise.
+
 2014-04-16  Samuel Thibault  <samuel.thibault@ens-lyon.org>
 
 	* sysdeps/mach/hurd/tls.h (dtv, dtv_t, tcbhead_t): Move declarations
diff --git a/sysdeps/x86/bits/link.h b/sysdeps/x86/bits/link.h
index 8673b2119b..82522e5cc9 100644
--- a/sysdeps/x86/bits/link.h
+++ b/sysdeps/x86/bits/link.h
@@ -93,6 +93,9 @@ typedef struct La_x86_64_regs
   uint64_t lr_rsp;
   La_x86_64_xmm lr_xmm[8];
   La_x86_64_vector lr_vector[8];
+#ifndef __ILP32__
+  __int128 lr_bnd[4];
+#endif
 } La_x86_64_regs;
 
 /* Return values for calls from PLT on x86-64.  */
@@ -106,6 +109,10 @@ typedef struct La_x86_64_retval
   long double lrv_st1;
   La_x86_64_vector lrv_vector0;
   La_x86_64_vector lrv_vector1;
+#ifndef __ILP32__
+  __int128 lrv_bnd0;
+  __int128 lrv_bnd1;
+#endif
 } La_x86_64_retval;
 
 #define La_x32_regs La_x86_64_regs
diff --git a/sysdeps/x86_64/dl-trampoline.S b/sysdeps/x86_64/dl-trampoline.S
index 3c0f54274a..bf3cff7b86 100644
--- a/sysdeps/x86_64/dl-trampoline.S
+++ b/sysdeps/x86_64/dl-trampoline.S
@@ -189,6 +189,20 @@ _dl_runtime_profile:
 	movaps %xmm6, (LR_XMM_OFFSET + XMM_SIZE*6)(%rsp)
 	movaps %xmm7, (LR_XMM_OFFSET + XMM_SIZE*7)(%rsp)
 
+# ifndef __ILP32__
+#  ifdef HAVE_MPX_SUPPORT
+	bndmov %bnd0, 		   (LR_BND_OFFSET)(%rsp)  # Preserve bound
+	bndmov %bnd1, (LR_BND_OFFSET +   BND_SIZE)(%rsp)  # registers. Nops if
+	bndmov %bnd2, (LR_BND_OFFSET + BND_SIZE*2)(%rsp)  # MPX not available
+	bndmov %bnd3, (LR_BND_OFFSET + BND_SIZE*3)(%rsp)  # or disabled.
+#  else
+	.byte 0x66,0x0f,0x1b,0x84,0x24;.long (LR_BND_OFFSET)
+	.byte 0x66,0x0f,0x1b,0x8c,0x24;.long (LR_BND_OFFSET + BND_SIZE)
+	.byte 0x66,0x0f,0x1b,0x84,0x24;.long (LR_BND_OFFSET + BND_SIZE*2)
+	.byte 0x66,0x0f,0x1b,0x8c,0x24;.long (LR_BND_OFFSET + BND_SIZE*3)
+#  endif
+# endif
+
 # if defined HAVE_AVX_SUPPORT || defined HAVE_AVX512_ASM_SUPPORT
 	.data
 L(have_avx):
diff --git a/sysdeps/x86_64/dl-trampoline.h b/sysdeps/x86_64/dl-trampoline.h
index 161af0ff57..416ed0e89c 100644
--- a/sysdeps/x86_64/dl-trampoline.h
+++ b/sysdeps/x86_64/dl-trampoline.h
@@ -63,6 +63,20 @@
 	movaps (LR_XMM_OFFSET + XMM_SIZE*6)(%rsp), %xmm6
 	movaps (LR_XMM_OFFSET + XMM_SIZE*7)(%rsp), %xmm7
 
+#ifndef __ILP32__
+# ifdef HAVE_MPX_SUPPORT
+	bndmov 		    (LR_BND_OFFSET)(%rsp), %bnd0  # Restore bound
+	bndmov (LR_BND_OFFSET +   BND_SIZE)(%rsp), %bnd1  # registers.
+	bndmov (LR_BND_OFFSET + BND_SIZE*2)(%rsp), %bnd2
+	bndmov (LR_BND_OFFSET + BND_SIZE*3)(%rsp), %bnd3
+# else
+	.byte 0x66,0x0f,0x1a,0x84,0x24;.long (LR_BND_OFFSET)
+	.byte 0x66,0x0f,0x1a,0x8c,0x24;.long (LR_BND_OFFSET + BND_SIZE)
+	.byte 0x66,0x0f,0x1a,0x94,0x24;.long (LR_BND_OFFSET + BND_SIZE*2)
+	.byte 0x66,0x0f,0x1a,0x9c,0x24;.long (LR_BND_OFFSET + BND_SIZE*3)
+# endif
+#endif
+
 #ifdef RESTORE_AVX
 	/* Check if any xmm0-xmm7 registers are changed by audit
 	   module.  */
@@ -222,6 +236,16 @@
 	vmovdqa %xmm1, (LRV_SIZE + XMM_SIZE)(%rcx)
 #endif
 
+#ifndef __ILP32__
+# ifdef HAVE_MPX_SUPPORT
+	bndmov %bnd0, LRV_BND0_OFFSET(%rcx)  # Preserve returned bounds.
+	bndmov %bnd1, LRV_BND1_OFFSET(%rcx)
+# else
+	.byte  0x66,0x0f,0x1b,0x81;.long (LRV_BND0_OFFSET)
+	.byte  0x66,0x0f,0x1b,0x89;.long (LRV_BND1_OFFSET)
+# endif
+#endif
+
 	fstpt LRV_ST0_OFFSET(%rcx)
 	fstpt LRV_ST1_OFFSET(%rcx)
 
@@ -254,6 +278,16 @@
 1:
 #endif
 
+#ifndef __ILP32__
+# ifdef HAVE_MPX_SUPPORT
+	bndmov LRV_BND0_OFFSET(%rcx), %bnd0  # Restore bound registers.
+	bndmov LRV_BND1_OFFSET(%rcx), %bnd1
+# else
+	.byte  0x66,0x0f,0x1a,0x81;.long (LRV_BND0_OFFSET)
+	.byte  0x66,0x0f,0x1a,0x89;.long (LRV_BND1_OFFSET)
+# endif
+#endif
+
 	fldt LRV_ST1_OFFSET(%rsp)
 	fldt LRV_ST0_OFFSET(%rsp)
 
diff --git a/sysdeps/x86_64/link-defines.sym b/sysdeps/x86_64/link-defines.sym
index 85d35ad034..862fa22024 100644
--- a/sysdeps/x86_64/link-defines.sym
+++ b/sysdeps/x86_64/link-defines.sym
@@ -6,6 +6,7 @@ VECTOR_SIZE		sizeof (La_x86_64_vector)
 XMM_SIZE		sizeof (La_x86_64_xmm)
 YMM_SIZE		sizeof (La_x86_64_ymm)
 ZMM_SIZE		sizeof (La_x86_64_zmm)
+BND_SIZE		sizeof (__int128)
 
 LR_SIZE			sizeof (struct La_x86_64_regs)
 LR_RDX_OFFSET		offsetof (struct La_x86_64_regs, lr_rdx)
@@ -18,6 +19,9 @@ LR_RBP_OFFSET		offsetof (struct La_x86_64_regs, lr_rbp)
 LR_RSP_OFFSET		offsetof (struct La_x86_64_regs, lr_rsp)
 LR_XMM_OFFSET		offsetof (struct La_x86_64_regs, lr_xmm)
 LR_VECTOR_OFFSET	offsetof (struct La_x86_64_regs, lr_vector)
+#ifndef __ILP32__
+LR_BND_OFFSET		offsetof (struct La_x86_64_regs, lr_bnd)
+#endif
 
 LRV_SIZE		sizeof (struct La_x86_64_retval)
 LRV_RAX_OFFSET		offsetof (struct La_x86_64_retval, lrv_rax)
@@ -28,3 +32,7 @@ LRV_ST0_OFFSET		offsetof (struct La_x86_64_retval, lrv_st0)
 LRV_ST1_OFFSET		offsetof (struct La_x86_64_retval, lrv_st1)
 LRV_VECTOR0_OFFSET	offsetof (struct La_x86_64_retval, lrv_vector0)
 LRV_VECTOR1_OFFSET	offsetof (struct La_x86_64_retval, lrv_vector1)
+#ifndef __ILP32__
+LRV_BND0_OFFSET		offsetof (struct La_x86_64_retval, lrv_bnd0)
+LRV_BND1_OFFSET		offsetof (struct La_x86_64_retval, lrv_bnd1)
+#endif