[PATCH 0/2] Revert to 32-bit 3.15-style vvar and vdso stripping

[PATCH 0/2] Revert to 32-bit 3.15-style vvar and vdso stripping

[Andy Lutomirski]

The fancy ideas of aggressively stripping the vdso and placing the
vvar data after the vdso have been considerably more painful than I
hoped.  This reverts to using objcopy -S to strip the vdso (thus
eliminating the whole fake section mechanism and all of its
attendent build-time fragility, I hope) and places the vvar data
before the vdso.

The latter is a bit messy: ELF wasn't really designed to have
symbols at negative virtual addresses.  I think I got all the
bitness issues and cross-compilation stuff right, but I'm not
well-equipped to test that.  The kbuild robot hasn't complained yet,
though.

I don't know whether this is 3.16 material.  I'm not aware of an
actual problem at this point, but the current state of affairs is
quite fragile.

Applies on top of tip/x86/urgent as of a few minutes ago.


Andy Lutomirski (2):
  x86,vdso: Move the vvar area before the vdso text
  x86,vdso: Get rid of the fake section mechanism

 arch/x86/include/asm/vdso.h       |  18 +--
 arch/x86/vdso/Makefile            |  16 ++-
 arch/x86/vdso/vdso-fakesections.c |  21 ----
 arch/x86/vdso/vdso-layout.lds.S   |  44 ++++----
 arch/x86/vdso/vdso2c.c            | 128 ++++++++++++++++-----
 arch/x86/vdso/vdso2c.h            | 227 +++++++-------------------------------
 arch/x86/vdso/vma.c               |  20 ++--
 7 files changed, 188 insertions(+), 286 deletions(-)
 delete mode 100644 arch/x86/vdso/vdso-fakesections.c

-- 
1.9.3

[PATCH 1/2] x86,vdso: Move the vvar area before the vdso text

[Andy Lutomirski]

Putting the vvar area after the vdso text is rather complicated: it
only works of the total length of the vdso text mapping is known at
vdso link time, and the linker doesn't allow symbol addresses to
depend on the sizes of non-allocatable data after the PT_LOAD
segment.

Moving the vvar area before the vdso text will allow is to safely
map non-allocatable data after the vdso text, which is a nice
simplification.

Signed-off-by: Andy Lutomirski <luto@amacapital.net>
---
 arch/x86/include/asm/vdso.h     | 18 ++++++++---------
 arch/x86/vdso/vdso-layout.lds.S | 44 ++++++++++++++++++-----------------------
 arch/x86/vdso/vdso2c.c          | 12 ++++++-----
 arch/x86/vdso/vdso2c.h          | 25 ++++++++++++++---------
 arch/x86/vdso/vma.c             | 20 ++++++++++---------
 5 files changed, 62 insertions(+), 57 deletions(-)

diff --git a/arch/x86/include/asm/vdso.h b/arch/x86/include/asm/vdso.h
index 30be253..8021bd2 100644
--- a/arch/x86/include/asm/vdso.h
+++ b/arch/x86/include/asm/vdso.h
@@ -18,15 +18,15 @@ struct vdso_image {
 
 	unsigned long alt, alt_len;
 
-	unsigned long sym_end_mapping;  /* Total size of the mapping */
-
-	unsigned long sym_vvar_page;
-	unsigned long sym_hpet_page;
-	unsigned long sym_VDSO32_NOTE_MASK;
-	unsigned long sym___kernel_sigreturn;
-	unsigned long sym___kernel_rt_sigreturn;
-	unsigned long sym___kernel_vsyscall;
-	unsigned long sym_VDSO32_SYSENTER_RETURN;
+	long sym_vvar_start;  /* Negative offset to the vvar area */
+
+	long sym_vvar_page;
+	long sym_hpet_page;
+	long sym_VDSO32_NOTE_MASK;
+	long sym___kernel_sigreturn;
+	long sym___kernel_rt_sigreturn;
+	long sym___kernel_vsyscall;
+	long sym_VDSO32_SYSENTER_RETURN;
 };
 
 #ifdef CONFIG_X86_64
diff --git a/arch/x86/vdso/vdso-layout.lds.S b/arch/x86/vdso/vdso-layout.lds.S
index 9197544..de2c921 100644
--- a/arch/x86/vdso/vdso-layout.lds.S
+++ b/arch/x86/vdso/vdso-layout.lds.S
@@ -18,6 +18,25 @@
 
 SECTIONS
 {
+	/*
+	 * User/kernel shared data is before the vDSO.  This may be a little
+	 * uglier than putting it after the vDSO, but it avoids issues with
+	 * non-allocatable things that dangle past the end of the PT_LOAD
+	 * segment.
+	 */
+
+	vvar_start = . - 2 * PAGE_SIZE;
+	vvar_page = vvar_start;
+
+	/* Place all vvars at the offsets in asm/vvar.h. */
+#define EMIT_VVAR(name, offset) vvar_ ## name = vvar_page + offset;
+#define __VVAR_KERNEL_LDS
+#include <asm/vvar.h>
+#undef __VVAR_KERNEL_LDS
+#undef EMIT_VVAR
+
+	hpet_page = vvar_start + PAGE_SIZE;
+
 	. = SIZEOF_HEADERS;
 
 	.hash		: { *(.hash) }			:text
@@ -74,31 +93,6 @@ SECTIONS
 	.altinstructions	: { *(.altinstructions) }	:text
 	.altinstr_replacement	: { *(.altinstr_replacement) }	:text
 
-	/*
-	 * The remainder of the vDSO consists of special pages that are
-	 * shared between the kernel and userspace.  It needs to be at the
-	 * end so that it doesn't overlap the mapping of the actual
-	 * vDSO image.
-	 */
-
-	. = ALIGN(PAGE_SIZE);
-	vvar_page = .;
-
-	/* Place all vvars at the offsets in asm/vvar.h. */
-#define EMIT_VVAR(name, offset) vvar_ ## name = vvar_page + offset;
-#define __VVAR_KERNEL_LDS
-#include <asm/vvar.h>
-#undef __VVAR_KERNEL_LDS
-#undef EMIT_VVAR
-
-	. = vvar_page + PAGE_SIZE;
-
-	hpet_page = .;
-	. = . + PAGE_SIZE;
-
-	. = ALIGN(PAGE_SIZE);
-	end_mapping = .;
-
 	/DISCARD/ : {
 		*(.discard)
 		*(.discard.*)
diff --git a/arch/x86/vdso/vdso2c.c b/arch/x86/vdso/vdso2c.c
index 238dbe82..22c54d0 100644
--- a/arch/x86/vdso/vdso2c.c
+++ b/arch/x86/vdso/vdso2c.c
@@ -20,9 +20,9 @@ const char *outfilename;
 
 /* Symbols that we need in vdso2c. */
 enum {
+	sym_vvar_start,
 	sym_vvar_page,
 	sym_hpet_page,
-	sym_end_mapping,
 	sym_VDSO_FAKE_SECTION_TABLE_START,
 	sym_VDSO_FAKE_SECTION_TABLE_END,
 };
@@ -38,9 +38,9 @@ struct vdso_sym {
 };
 
 struct vdso_sym required_syms[] = {
+	[sym_vvar_start] = {"vvar_start", true},
 	[sym_vvar_page] = {"vvar_page", true},
 	[sym_hpet_page] = {"hpet_page", true},
-	[sym_end_mapping] = {"end_mapping", true},
 	[sym_VDSO_FAKE_SECTION_TABLE_START] = {
 		"VDSO_FAKE_SECTION_TABLE_START", false
 	},
@@ -96,9 +96,11 @@ extern void bad_put_le(void);
 
 #define NSYMS (sizeof(required_syms) / sizeof(required_syms[0]))
 
-#define BITSFUNC3(name, bits) name##bits
-#define BITSFUNC2(name, bits) BITSFUNC3(name, bits)
-#define BITSFUNC(name) BITSFUNC2(name, ELF_BITS)
+#define BITSFUNC3(name, bits, suffix) name##bits##suffix
+#define BITSFUNC2(name, bits, suffix) BITSFUNC3(name, bits, suffix)
+#define BITSFUNC(name) BITSFUNC2(name, ELF_BITS, )
+
+#define INT_BITS BITSFUNC2(int, ELF_BITS, _t)
 
 #define ELF_BITS_XFORM2(bits, x) Elf##bits##_##x
 #define ELF_BITS_XFORM(bits, x) ELF_BITS_XFORM2(bits, x)
diff --git a/arch/x86/vdso/vdso2c.h b/arch/x86/vdso/vdso2c.h
index 11b65d4..2da32fbc 100644
--- a/arch/x86/vdso/vdso2c.h
+++ b/arch/x86/vdso/vdso2c.h
@@ -132,7 +132,7 @@ static void BITSFUNC(go)(void *addr, size_t len,
 		*alt_sec = NULL;
 	ELF(Dyn) *dyn = 0, *dyn_end = 0;
 	const char *secstrings;
-	uint64_t syms[NSYMS] = {};
+	INT_BITS syms[NSYMS] = {};
 
 	struct BITSFUNC(fake_sections) fake_sections = {};
 
@@ -209,6 +209,13 @@ static void BITSFUNC(go)(void *addr, size_t len,
 					fail("duplicate symbol %s\n",
 					     required_syms[k].name);
 				}
+
+				/*
+				 * Careful: we use negative addresses, but
+				 * st_value is unsigned, so we rely
+				 * on syms[k] being a signed type of the
+				 * correct width.
+				 */
 				syms[k] = GET_LE(&sym->st_value);
 			}
 		}
@@ -263,15 +270,15 @@ static void BITSFUNC(go)(void *addr, size_t len,
 		if (syms[i] % 4096)
 			fail("%s must be a multiple of 4096\n",
 			     required_syms[i].name);
-		if (syms[i] < data_size)
-			fail("%s must be after the text mapping\n",
+		if (syms[sym_vvar_start] > syms[i] + 4096)
+			fail("%s underruns begin_vvar\n",
 			     required_syms[i].name);
-		if (syms[sym_end_mapping] < syms[i] + 4096)
-			fail("%s overruns end_mapping\n",
+		if (syms[i] + 4096 > 0)
+			fail("%s is on the wrong side of the vdso text\n",
 			     required_syms[i].name);
 	}
-	if (syms[sym_end_mapping] % 4096)
-		fail("end_mapping must be a multiple of 4096\n");
+	if (syms[sym_vvar_start] % 4096)
+		fail("vvar_begin must be a multiple of 4096\n");
 
 	if (!name) {
 		fwrite(addr, load_size, 1, outfile);
@@ -311,8 +318,8 @@ static void BITSFUNC(go)(void *addr, size_t len,
 	}
 	for (i = 0; i < NSYMS; i++) {
 		if (required_syms[i].export && syms[i])
-			fprintf(outfile, "\t.sym_%s = 0x%" PRIx64 ",\n",
-				required_syms[i].name, syms[i]);
+			fprintf(outfile, "\t.sym_%s = %" PRIi64 ",\n",
+				required_syms[i].name, (int64_t)syms[i]);
 	}
 	fprintf(outfile, "};\n");
 }
diff --git a/arch/x86/vdso/vma.c b/arch/x86/vdso/vma.c
index 5a5176d..dbef622 100644
--- a/arch/x86/vdso/vma.c
+++ b/arch/x86/vdso/vma.c
@@ -93,7 +93,7 @@ static int map_vdso(const struct vdso_image *image, bool calculate_addr)
 {
 	struct mm_struct *mm = current->mm;
 	struct vm_area_struct *vma;
-	unsigned long addr;
+	unsigned long addr, text_start;
 	int ret = 0;
 	static struct page *no_pages[] = {NULL};
 	static struct vm_special_mapping vvar_mapping = {
@@ -103,26 +103,28 @@ static int map_vdso(const struct vdso_image *image, bool calculate_addr)
 
 	if (calculate_addr) {
 		addr = vdso_addr(current->mm->start_stack,
-				 image->sym_end_mapping);
+				 image->size - image->sym_vvar_start);
 	} else {
 		addr = 0;
 	}
 
 	down_write(&mm->mmap_sem);
 
-	addr = get_unmapped_area(NULL, addr, image->sym_end_mapping, 0, 0);
+	addr = get_unmapped_area(NULL, addr,
+				 image->size - image->sym_vvar_start, 0, 0);
 	if (IS_ERR_VALUE(addr)) {
 		ret = addr;
 		goto up_fail;
 	}
 
-	current->mm->context.vdso = (void __user *)addr;
+	text_start = addr - image->sym_vvar_start;
+	current->mm->context.vdso = (void __user *)text_start;
 
 	/*
 	 * MAYWRITE to allow gdb to COW and set breakpoints
 	 */
 	vma = _install_special_mapping(mm,
-				       addr,
+				       text_start,
 				       image->size,
 				       VM_READ|VM_EXEC|
 				       VM_MAYREAD|VM_MAYWRITE|VM_MAYEXEC,
@@ -134,8 +136,8 @@ static int map_vdso(const struct vdso_image *image, bool calculate_addr)
 	}
 
 	vma = _install_special_mapping(mm,
-				       addr + image->size,
-				       image->sym_end_mapping - image->size,
+				       addr,
+				       -image->sym_vvar_start,
 				       VM_READ,
 				       &vvar_mapping);
 
@@ -146,7 +148,7 @@ static int map_vdso(const struct vdso_image *image, bool calculate_addr)
 
 	if (image->sym_vvar_page)
 		ret = remap_pfn_range(vma,
-				      addr + image->sym_vvar_page,
+				      text_start + image->sym_vvar_page,
 				      __pa_symbol(&__vvar_page) >> PAGE_SHIFT,
 				      PAGE_SIZE,
 				      PAGE_READONLY);
@@ -157,7 +159,7 @@ static int map_vdso(const struct vdso_image *image, bool calculate_addr)
 #ifdef CONFIG_HPET_TIMER
 	if (hpet_address && image->sym_hpet_page) {
 		ret = io_remap_pfn_range(vma,
-			addr + image->sym_hpet_page,
+			text_start + image->sym_hpet_page,
 			hpet_address >> PAGE_SHIFT,
 			PAGE_SIZE,
 			pgprot_noncached(PAGE_READONLY));
-- 
1.9.3

[tip:x86/vdso] x86, vdso: Move the vvar area before the vdso text

[tip-bot for Andy Lutomirski]

Commit-ID:  e6577a7ce99a506b587bcd1d2cd803cb45119557
Gitweb:     http://git.kernel.org/tip/e6577a7ce99a506b587bcd1d2cd803cb45119557
Author:     Andy Lutomirski <luto@amacapital.net>
AuthorDate: Thu, 10 Jul 2014 18:13:15 -0700
Committer:  H. Peter Anvin <hpa@linux.intel.com>
CommitDate: Fri, 11 Jul 2014 16:57:51 -0700

x86, vdso: Move the vvar area before the vdso text

Putting the vvar area after the vdso text is rather complicated: it
only works of the total length of the vdso text mapping is known at
vdso link time, and the linker doesn't allow symbol addresses to
depend on the sizes of non-allocatable data after the PT_LOAD
segment.

Moving the vvar area before the vdso text will allow is to safely
map non-allocatable data after the vdso text, which is a nice
simplification.

Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Link: http://lkml.kernel.org/r/156c78c0d93144ff1055a66493783b9e56813983.1405040914.git.luto@amacapital.net
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
---
 arch/x86/include/asm/vdso.h     | 18 ++++++++---------
 arch/x86/vdso/vdso-layout.lds.S | 44 ++++++++++++++++++-----------------------
 arch/x86/vdso/vdso2c.c          | 12 ++++++-----
 arch/x86/vdso/vdso2c.h          | 25 ++++++++++++++---------
 arch/x86/vdso/vma.c             | 20 ++++++++++---------
 5 files changed, 62 insertions(+), 57 deletions(-)

diff --git a/arch/x86/include/asm/vdso.h b/arch/x86/include/asm/vdso.h
index 30be253..8021bd2 100644
--- a/arch/x86/include/asm/vdso.h
+++ b/arch/x86/include/asm/vdso.h
@@ -18,15 +18,15 @@ struct vdso_image {
 
 	unsigned long alt, alt_len;
 
-	unsigned long sym_end_mapping;  /* Total size of the mapping */
-
-	unsigned long sym_vvar_page;
-	unsigned long sym_hpet_page;
-	unsigned long sym_VDSO32_NOTE_MASK;
-	unsigned long sym___kernel_sigreturn;
-	unsigned long sym___kernel_rt_sigreturn;
-	unsigned long sym___kernel_vsyscall;
-	unsigned long sym_VDSO32_SYSENTER_RETURN;
+	long sym_vvar_start;  /* Negative offset to the vvar area */
+
+	long sym_vvar_page;
+	long sym_hpet_page;
+	long sym_VDSO32_NOTE_MASK;
+	long sym___kernel_sigreturn;
+	long sym___kernel_rt_sigreturn;
+	long sym___kernel_vsyscall;
+	long sym_VDSO32_SYSENTER_RETURN;
 };
 
 #ifdef CONFIG_X86_64
diff --git a/arch/x86/vdso/vdso-layout.lds.S b/arch/x86/vdso/vdso-layout.lds.S
index 9197544..de2c921 100644
--- a/arch/x86/vdso/vdso-layout.lds.S
+++ b/arch/x86/vdso/vdso-layout.lds.S
@@ -18,6 +18,25 @@
 
 SECTIONS
 {
+	/*
+	 * User/kernel shared data is before the vDSO.  This may be a little
+	 * uglier than putting it after the vDSO, but it avoids issues with
+	 * non-allocatable things that dangle past the end of the PT_LOAD
+	 * segment.
+	 */
+
+	vvar_start = . - 2 * PAGE_SIZE;
+	vvar_page = vvar_start;
+
+	/* Place all vvars at the offsets in asm/vvar.h. */
+#define EMIT_VVAR(name, offset) vvar_ ## name = vvar_page + offset;
+#define __VVAR_KERNEL_LDS
+#include <asm/vvar.h>
+#undef __VVAR_KERNEL_LDS
+#undef EMIT_VVAR
+
+	hpet_page = vvar_start + PAGE_SIZE;
+
 	. = SIZEOF_HEADERS;
 
 	.hash		: { *(.hash) }			:text
@@ -74,31 +93,6 @@ SECTIONS
 	.altinstructions	: { *(.altinstructions) }	:text
 	.altinstr_replacement	: { *(.altinstr_replacement) }	:text
 
-	/*
-	 * The remainder of the vDSO consists of special pages that are
-	 * shared between the kernel and userspace.  It needs to be at the
-	 * end so that it doesn't overlap the mapping of the actual
-	 * vDSO image.
-	 */
-
-	. = ALIGN(PAGE_SIZE);
-	vvar_page = .;
-
-	/* Place all vvars at the offsets in asm/vvar.h. */
-#define EMIT_VVAR(name, offset) vvar_ ## name = vvar_page + offset;
-#define __VVAR_KERNEL_LDS
-#include <asm/vvar.h>
-#undef __VVAR_KERNEL_LDS
-#undef EMIT_VVAR
-
-	. = vvar_page + PAGE_SIZE;
-
-	hpet_page = .;
-	. = . + PAGE_SIZE;
-
-	. = ALIGN(PAGE_SIZE);
-	end_mapping = .;
-
 	/DISCARD/ : {
 		*(.discard)
 		*(.discard.*)
diff --git a/arch/x86/vdso/vdso2c.c b/arch/x86/vdso/vdso2c.c
index 238dbe82..22c54d0 100644
--- a/arch/x86/vdso/vdso2c.c
+++ b/arch/x86/vdso/vdso2c.c
@@ -20,9 +20,9 @@ const char *outfilename;
 
 /* Symbols that we need in vdso2c. */
 enum {
+	sym_vvar_start,
 	sym_vvar_page,
 	sym_hpet_page,
-	sym_end_mapping,
 	sym_VDSO_FAKE_SECTION_TABLE_START,
 	sym_VDSO_FAKE_SECTION_TABLE_END,
 };
@@ -38,9 +38,9 @@ struct vdso_sym {
 };
 
 struct vdso_sym required_syms[] = {
+	[sym_vvar_start] = {"vvar_start", true},
 	[sym_vvar_page] = {"vvar_page", true},
 	[sym_hpet_page] = {"hpet_page", true},
-	[sym_end_mapping] = {"end_mapping", true},
 	[sym_VDSO_FAKE_SECTION_TABLE_START] = {
 		"VDSO_FAKE_SECTION_TABLE_START", false
 	},
@@ -96,9 +96,11 @@ extern void bad_put_le(void);
 
 #define NSYMS (sizeof(required_syms) / sizeof(required_syms[0]))
 
-#define BITSFUNC3(name, bits) name##bits
-#define BITSFUNC2(name, bits) BITSFUNC3(name, bits)
-#define BITSFUNC(name) BITSFUNC2(name, ELF_BITS)
+#define BITSFUNC3(name, bits, suffix) name##bits##suffix
+#define BITSFUNC2(name, bits, suffix) BITSFUNC3(name, bits, suffix)
+#define BITSFUNC(name) BITSFUNC2(name, ELF_BITS, )
+
+#define INT_BITS BITSFUNC2(int, ELF_BITS, _t)
 
 #define ELF_BITS_XFORM2(bits, x) Elf##bits##_##x
 #define ELF_BITS_XFORM(bits, x) ELF_BITS_XFORM2(bits, x)
diff --git a/arch/x86/vdso/vdso2c.h b/arch/x86/vdso/vdso2c.h
index 11b65d4..2da32fbc 100644
--- a/arch/x86/vdso/vdso2c.h
+++ b/arch/x86/vdso/vdso2c.h
@@ -132,7 +132,7 @@ static void BITSFUNC(go)(void *addr, size_t len,
 		*alt_sec = NULL;
 	ELF(Dyn) *dyn = 0, *dyn_end = 0;
 	const char *secstrings;
-	uint64_t syms[NSYMS] = {};
+	INT_BITS syms[NSYMS] = {};
 
 	struct BITSFUNC(fake_sections) fake_sections = {};
 
@@ -209,6 +209,13 @@ static void BITSFUNC(go)(void *addr, size_t len,
 					fail("duplicate symbol %s\n",
 					     required_syms[k].name);
 				}
+
+				/*
+				 * Careful: we use negative addresses, but
+				 * st_value is unsigned, so we rely
+				 * on syms[k] being a signed type of the
+				 * correct width.
+				 */
 				syms[k] = GET_LE(&sym->st_value);
 			}
 		}
@@ -263,15 +270,15 @@ static void BITSFUNC(go)(void *addr, size_t len,
 		if (syms[i] % 4096)
 			fail("%s must be a multiple of 4096\n",
 			     required_syms[i].name);
-		if (syms[i] < data_size)
-			fail("%s must be after the text mapping\n",
+		if (syms[sym_vvar_start] > syms[i] + 4096)
+			fail("%s underruns begin_vvar\n",
 			     required_syms[i].name);
-		if (syms[sym_end_mapping] < syms[i] + 4096)
-			fail("%s overruns end_mapping\n",
+		if (syms[i] + 4096 > 0)
+			fail("%s is on the wrong side of the vdso text\n",
 			     required_syms[i].name);
 	}
-	if (syms[sym_end_mapping] % 4096)
-		fail("end_mapping must be a multiple of 4096\n");
+	if (syms[sym_vvar_start] % 4096)
+		fail("vvar_begin must be a multiple of 4096\n");
 
 	if (!name) {
 		fwrite(addr, load_size, 1, outfile);
@@ -311,8 +318,8 @@ static void BITSFUNC(go)(void *addr, size_t len,
 	}
 	for (i = 0; i < NSYMS; i++) {
 		if (required_syms[i].export && syms[i])
-			fprintf(outfile, "\t.sym_%s = 0x%" PRIx64 ",\n",
-				required_syms[i].name, syms[i]);
+			fprintf(outfile, "\t.sym_%s = %" PRIi64 ",\n",
+				required_syms[i].name, (int64_t)syms[i]);
 	}
 	fprintf(outfile, "};\n");
 }
diff --git a/arch/x86/vdso/vma.c b/arch/x86/vdso/vma.c
index 5a5176d..dbef622 100644
--- a/arch/x86/vdso/vma.c
+++ b/arch/x86/vdso/vma.c
@@ -93,7 +93,7 @@ static int map_vdso(const struct vdso_image *image, bool calculate_addr)
 {
 	struct mm_struct *mm = current->mm;
 	struct vm_area_struct *vma;
-	unsigned long addr;
+	unsigned long addr, text_start;
 	int ret = 0;
 	static struct page *no_pages[] = {NULL};
 	static struct vm_special_mapping vvar_mapping = {
@@ -103,26 +103,28 @@ static int map_vdso(const struct vdso_image *image, bool calculate_addr)
 
 	if (calculate_addr) {
 		addr = vdso_addr(current->mm->start_stack,
-				 image->sym_end_mapping);
+				 image->size - image->sym_vvar_start);
 	} else {
 		addr = 0;
 	}
 
 	down_write(&mm->mmap_sem);
 
-	addr = get_unmapped_area(NULL, addr, image->sym_end_mapping, 0, 0);
+	addr = get_unmapped_area(NULL, addr,
+				 image->size - image->sym_vvar_start, 0, 0);
 	if (IS_ERR_VALUE(addr)) {
 		ret = addr;
 		goto up_fail;
 	}
 
-	current->mm->context.vdso = (void __user *)addr;
+	text_start = addr - image->sym_vvar_start;
+	current->mm->context.vdso = (void __user *)text_start;
 
 	/*
 	 * MAYWRITE to allow gdb to COW and set breakpoints
 	 */
 	vma = _install_special_mapping(mm,
-				       addr,
+				       text_start,
 				       image->size,
 				       VM_READ|VM_EXEC|
 				       VM_MAYREAD|VM_MAYWRITE|VM_MAYEXEC,
@@ -134,8 +136,8 @@ static int map_vdso(const struct vdso_image *image, bool calculate_addr)
 	}
 
 	vma = _install_special_mapping(mm,
-				       addr + image->size,
-				       image->sym_end_mapping - image->size,
+				       addr,
+				       -image->sym_vvar_start,
 				       VM_READ,
 				       &vvar_mapping);
 
@@ -146,7 +148,7 @@ static int map_vdso(const struct vdso_image *image, bool calculate_addr)
 
 	if (image->sym_vvar_page)
 		ret = remap_pfn_range(vma,
-				      addr + image->sym_vvar_page,
+				      text_start + image->sym_vvar_page,
 				      __pa_symbol(&__vvar_page) >> PAGE_SHIFT,
 				      PAGE_SIZE,
 				      PAGE_READONLY);
@@ -157,7 +159,7 @@ static int map_vdso(const struct vdso_image *image, bool calculate_addr)
 #ifdef CONFIG_HPET_TIMER
 	if (hpet_address && image->sym_hpet_page) {
 		ret = io_remap_pfn_range(vma,
-			addr + image->sym_hpet_page,
+			text_start + image->sym_hpet_page,
 			hpet_address >> PAGE_SHIFT,
 			PAGE_SIZE,
 			pgprot_noncached(PAGE_READONLY));

[PATCH 2/2] x86,vdso: Get rid of the fake section mechanism

[Andy Lutomirski]

Now that we can tolerate extra things dangling off the end of the
vdso image, we can strip the vdso the old fashioned way rather than
using an overcomplicated custom stripping algorithm.

This is a partial reversion of:
    6f121e5 x86, vdso: Reimplement vdso.so preparation in build-time C

Signed-off-by: Andy Lutomirski <luto@amacapital.net>
---
 arch/x86/vdso/Makefile            |  16 +--
 arch/x86/vdso/vdso-fakesections.c |  21 ----
 arch/x86/vdso/vdso2c.c            | 116 +++++++++++++++++-----
 arch/x86/vdso/vdso2c.h            | 202 +++++---------------------------------
 4 files changed, 126 insertions(+), 229 deletions(-)
 delete mode 100644 arch/x86/vdso/vdso-fakesections.c

diff --git a/arch/x86/vdso/Makefile b/arch/x86/vdso/Makefile
index 61b04fe..5a4affe 100644
--- a/arch/x86/vdso/Makefile
+++ b/arch/x86/vdso/Makefile
@@ -10,7 +10,7 @@ VDSO32-$(CONFIG_X86_32)		:= y
 VDSO32-$(CONFIG_COMPAT)		:= y
 
 # files to link into the vdso
-vobjs-y := vdso-note.o vclock_gettime.o vgetcpu.o vdso-fakesections.o
+vobjs-y := vdso-note.o vclock_gettime.o vgetcpu.o
 
 # files to link into kernel
 obj-y				+= vma.o
@@ -37,7 +37,8 @@ vdso_img_sodbg := $(vdso_img-y:%=vdso%.so.dbg)
 obj-y += $(vdso_img_objs)
 targets += $(vdso_img_cfiles)
 targets += $(vdso_img_sodbg)
-.SECONDARY: $(vdso_img-y:%=$(obj)/vdso-image-%.c)
+.SECONDARY: $(vdso_img-y:%=$(obj)/vdso-image-%.c) \
+	$(vdso_img-y:%=$(obj)/vdso%.so)
 
 export CPPFLAGS_vdso.lds += -P -C
 
@@ -54,10 +55,10 @@ hostprogs-y			+= vdso2c
 
 quiet_cmd_vdso2c = VDSO2C  $@
 define cmd_vdso2c
-	$(obj)/vdso2c $< $@
+	$(obj)/vdso2c $< $(<:%.dbg=%) $@
 endef
 
-$(obj)/vdso-image-%.c: $(obj)/vdso%.so.dbg $(obj)/vdso2c FORCE
+$(obj)/vdso-image-%.c: $(obj)/vdso%.so.dbg $(obj)/vdso%.so $(obj)/vdso2c FORCE
 	$(call if_changed,vdso2c)
 
 #
@@ -113,6 +114,10 @@ $(obj)/%-x32.o: $(obj)/%.o FORCE
 
 targets += vdsox32.lds $(vobjx32s-y)
 
+$(obj)/%.so: OBJCOPYFLAGS := -S
+$(obj)/%.so: $(obj)/%.so.dbg
+	$(call if_changed,objcopy)
+
 $(obj)/vdsox32.so.dbg: $(src)/vdsox32.lds $(vobjx32s) FORCE
 	$(call if_changed,vdso)
 
@@ -134,7 +139,7 @@ override obj-dirs = $(dir $(obj)) $(obj)/vdso32/
 
 targets += vdso32/vdso32.lds
 targets += vdso32/note.o vdso32/vclock_gettime.o $(vdso32.so-y:%=vdso32/%.o)
-targets += vdso32/vclock_gettime.o vdso32/vdso-fakesections.o
+targets += vdso32/vclock_gettime.o
 
 $(obj)/vdso32.o: $(vdso32-images:%=$(obj)/%)
 
@@ -156,7 +161,6 @@ $(vdso32-images:%=$(obj)/%.dbg): KBUILD_CFLAGS = $(KBUILD_CFLAGS_32)
 $(vdso32-images:%=$(obj)/%.dbg): $(obj)/vdso32-%.so.dbg: FORCE \
 				 $(obj)/vdso32/vdso32.lds \
 				 $(obj)/vdso32/vclock_gettime.o \
-				 $(obj)/vdso32/vdso-fakesections.o \
 				 $(obj)/vdso32/note.o \
 				 $(obj)/vdso32/%.o
 	$(call if_changed,vdso)
diff --git a/arch/x86/vdso/vdso-fakesections.c b/arch/x86/vdso/vdso-fakesections.c
deleted file mode 100644
index aa5fbfa..0000000
--- a/arch/x86/vdso/vdso-fakesections.c
+++ /dev/null
@@ -1,21 +0,0 @@
-/*
- * Copyright 2014 Andy Lutomirski
- * Subject to the GNU Public License, v.2
- *
- * String table for loadable section headers.  See vdso2c.h for why
- * this exists.
- */
-
-const char fake_shstrtab[] __attribute__((section(".fake_shstrtab"))) =
-	".hash\0"
-	".dynsym\0"
-	".dynstr\0"
-	".gnu.version\0"
-	".gnu.version_d\0"
-	".dynamic\0"
-	".rodata\0"
-	".fake_shstrtab\0"  /* Yay, self-referential code. */
-	".note\0"
-	".eh_frame_hdr\0"
-	".eh_frame\0"
-	".text";
diff --git a/arch/x86/vdso/vdso2c.c b/arch/x86/vdso/vdso2c.c
index 22c54d0..8627db2 100644
--- a/arch/x86/vdso/vdso2c.c
+++ b/arch/x86/vdso/vdso2c.c
@@ -1,3 +1,53 @@
+/*
+ * vdso2c - A vdso image preparation tool
+ * Copyright (c) 2014 Andy Lutomirski and others
+ * Licensed under the GPL v2
+ *
+ * vdso2c requires stripped and unstripped input.  It would be trivial
+ * to fully strip the input in here, but, for reasons described below,
+ * we need to write a section table.  Doing this is more or less
+ * equivalent to dropping all non-allocatable sections, but it's
+ * easier to let objcopy handle that instead of doing it ourselves.
+ * If we ever need to do something fancier than what objcopy provides,
+ * it would be straightforward to add here.
+ *
+ * We're keep a section table for a few reasons:
+ *
+ * The Go runtime had a couple of bugs: it would read the section
+ * table to try to figure out how many dynamic symbols there were (it
+ * shouldn't have looked at the section table at all) and, if there
+ * were no SHT_SYNDYM section table entry, it would use an
+ * uninitialized value for the number of symbols.  An empty DYNSYM
+ * table would work, but I see no reason not to write a valid one (and
+ * keep full performance for old Go programs).  This hack is only
+ * needed on x86_64.
+ *
+ * The bug was introduced on 2012-08-31 by:
+ * https://code.google.com/p/go/source/detail?r=56ea40aac72b
+ * and was fixed on 2014-06-13 by:
+ * https://code.google.com/p/go/source/detail?r=fc1cd5e12595
+ *
+ * Binutils has issues debugging the vDSO: it reads the section table to
+ * find SHT_NOTE; it won't look at PT_NOTE for the in-memory vDSO, which
+ * would break build-id if we removed the section table.  Binutils
+ * also requires that shstrndx != 0.  See:
+ * https://sourceware.org/bugzilla/show_bug.cgi?id=17064
+ *
+ * elfutils might not look for PT_NOTE if there is a section table at
+ * all.  I don't know whether this matters for any practical purpose.
+ *
+ * For simplicity, rather than hacking up a partial section table, we
+ * just write a mostly complete one.  We omit non-dynamic symbols,
+ * though, since they're rather large.
+ *
+ * Once binutils gets fixed, we might be able to drop this for all but
+ * the 64-bit vdso, since build-id only works in kernel RPMs, and
+ * systems that update to new enough kernel RPMs will likely update
+ * binutils in sync.  build-id has never worked for home-built kernel
+ * RPMs without manual symlinking, and I suspect that no one ever does
+ * that.
+ */
+
 #include <inttypes.h>
 #include <stdint.h>
 #include <unistd.h>
@@ -61,7 +111,8 @@ static void fail(const char *format, ...)
 	va_start(ap, format);
 	fprintf(stderr, "Error: ");
 	vfprintf(stderr, format, ap);
-	unlink(outfilename);
+	if (outfilename)
+		unlink(outfilename);
 	exit(1);
 	va_end(ap);
 }
@@ -114,30 +165,53 @@ extern void bad_put_le(void);
 #include "vdso2c.h"
 #undef ELF_BITS
 
-static void go(void *addr, size_t len, FILE *outfile, const char *name)
+static void go(void *raw_addr, size_t raw_len,
+	       void *stripped_addr, size_t stripped_len,
+	       FILE *outfile, const char *name)
 {
-	Elf64_Ehdr *hdr = (Elf64_Ehdr *)addr;
+	Elf64_Ehdr *hdr = (Elf64_Ehdr *)raw_addr;
 
 	if (hdr->e_ident[EI_CLASS] == ELFCLASS64) {
-		go64(addr, len, outfile, name);
+		go64(raw_addr, raw_len, stripped_addr, stripped_len,
+		     outfile, name);
 	} else if (hdr->e_ident[EI_CLASS] == ELFCLASS32) {
-		go32(addr, len, outfile, name);
+		go32(raw_addr, raw_len, stripped_addr, stripped_len,
+		     outfile, name);
 	} else {
 		fail("unknown ELF class\n");
 	}
 }
 
+static void map_input(const char *name, void **addr, size_t *len, int prot)
+{
+	off_t tmp_len;
+
+	int fd = open(name, O_RDONLY);
+	if (fd == -1)
+		err(1, "%s", name);
+
+	tmp_len = lseek(fd, 0, SEEK_END);
+	if (tmp_len == (off_t)-1)
+		err(1, "lseek");
+	*len = (size_t)tmp_len;
+
+	*addr = mmap(NULL, tmp_len, prot, MAP_PRIVATE, fd, 0);
+	if (*addr == MAP_FAILED)
+		err(1, "mmap");
+
+	close(fd);
+}
+
 int main(int argc, char **argv)
 {
-	int fd;
-	off_t len;
-	void *addr;
+	size_t raw_len, stripped_len;
+	void *raw_addr, *stripped_addr;
 	FILE *outfile;
 	char *name, *tmp;
 	int namelen;
 
-	if (argc != 3) {
-		printf("Usage: vdso2c INPUT OUTPUT\n");
+	if (argc != 4) {
+		printf("Usage: vdso2c RAW_INPUT STRIPPED_INPUT OUTPUT\n");
 		return 1;
 	}
 
@@ -145,7 +219,7 @@ int main(int argc, char **argv)
 	 * Figure out the struct name.  If we're writing to a .so file,
 	 * generate raw output insted.
 	 */
-	name = strdup(argv[2]);
+	name = strdup(argv[3]);
 	namelen = strlen(name);
 	if (namelen >= 3 && !strcmp(name + namelen - 3, ".so")) {
 		name = NULL;
@@ -161,26 +235,18 @@ int main(int argc, char **argv)
 				*tmp = '_';
 	}
 
-	fd = open(argv[1], O_RDONLY);
-	if (fd == -1)
-		err(1, "%s", argv[1]);
-
-	len = lseek(fd, 0, SEEK_END);
-	if (len == (off_t)-1)
-		err(1, "lseek");
-
-	addr = mmap(NULL, len, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0);
-	if (addr == MAP_FAILED)
-		err(1, "mmap");
+	map_input(argv[1], &raw_addr, &raw_len, PROT_READ);
+	map_input(argv[2], &stripped_addr, &stripped_len, PROT_READ);
 
-	outfilename = argv[2];
+	outfilename = argv[3];
 	outfile = fopen(outfilename, "w");
 	if (!outfile)
 		err(1, "%s", argv[2]);
 
-	go(addr, (size_t)len, outfile, name);
+	go(raw_addr, raw_len, stripped_addr, stripped_len, outfile, name);
 
-	munmap(addr, len);
+	munmap(raw_addr, raw_len);
+	munmap(stripped_addr, stripped_len);
 	fclose(outfile);
 
 	return 0;
diff --git a/arch/x86/vdso/vdso2c.h b/arch/x86/vdso/vdso2c.h
index 2da32fbc..fd57829 100644
--- a/arch/x86/vdso/vdso2c.h
+++ b/arch/x86/vdso/vdso2c.h
@@ -4,128 +4,14 @@
  * are built for 32-bit userspace.
  */
 
-/*
- * We're writing a section table for a few reasons:
- *
- * The Go runtime had a couple of bugs: it would read the section
- * table to try to figure out how many dynamic symbols there were (it
- * shouldn't have looked at the section table at all) and, if there
- * were no SHT_SYNDYM section table entry, it would use an
- * uninitialized value for the number of symbols.  An empty DYNSYM
- * table would work, but I see no reason not to write a valid one (and
- * keep full performance for old Go programs).  This hack is only
- * needed on x86_64.
- *
- * The bug was introduced on 2012-08-31 by:
- * https://code.google.com/p/go/source/detail?r=56ea40aac72b
- * and was fixed on 2014-06-13 by:
- * https://code.google.com/p/go/source/detail?r=fc1cd5e12595
- *
- * Binutils has issues debugging the vDSO: it reads the section table to
- * find SHT_NOTE; it won't look at PT_NOTE for the in-memory vDSO, which
- * would break build-id if we removed the section table.  Binutils
- * also requires that shstrndx != 0.  See:
- * https://sourceware.org/bugzilla/show_bug.cgi?id=17064
- *
- * elfutils might not look for PT_NOTE if there is a section table at
- * all.  I don't know whether this matters for any practical purpose.
- *
- * For simplicity, rather than hacking up a partial section table, we
- * just write a mostly complete one.  We omit non-dynamic symbols,
- * though, since they're rather large.
- *
- * Once binutils gets fixed, we might be able to drop this for all but
- * the 64-bit vdso, since build-id only works in kernel RPMs, and
- * systems that update to new enough kernel RPMs will likely update
- * binutils in sync.  build-id has never worked for home-built kernel
- * RPMs without manual symlinking, and I suspect that no one ever does
- * that.
- */
-struct BITSFUNC(fake_sections)
-{
-	ELF(Shdr) *table;
-	unsigned long table_offset;
-	int count, max_count;
-
-	int in_shstrndx;
-	unsigned long shstr_offset;
-	const char *shstrtab;
-	size_t shstrtab_len;
-
-	int out_shstrndx;
-};
-
-static unsigned int BITSFUNC(find_shname)(struct BITSFUNC(fake_sections) *out,
-					  const char *name)
-{
-	const char *outname = out->shstrtab;
-	while (outname - out->shstrtab < out->shstrtab_len) {
-		if (!strcmp(name, outname))
-			return (outname - out->shstrtab) + out->shstr_offset;
-		outname += strlen(outname) + 1;
-	}
-
-	if (*name)
-		printf("Warning: could not find output name \"%s\"\n", name);
-	return out->shstr_offset + out->shstrtab_len - 1;  /* Use a null. */
-}
-
-static void BITSFUNC(init_sections)(struct BITSFUNC(fake_sections) *out)
-{
-	if (!out->in_shstrndx)
-		fail("didn't find the fake shstrndx\n");
-
-	memset(out->table, 0, out->max_count * sizeof(ELF(Shdr)));
-
-	if (out->max_count < 1)
-		fail("we need at least two fake output sections\n");
-
-	PUT_LE(&out->table[0].sh_type, SHT_NULL);
-	PUT_LE(&out->table[0].sh_name, BITSFUNC(find_shname)(out, ""));
-
-	out->count = 1;
-}
-
-static void BITSFUNC(copy_section)(struct BITSFUNC(fake_sections) *out,
-				   int in_idx, const ELF(Shdr) *in,
-				   const char *name)
-{
-	uint64_t flags = GET_LE(&in->sh_flags);
-
-	bool copy = flags & SHF_ALLOC &&
-		(GET_LE(&in->sh_size) ||
-		 (GET_LE(&in->sh_type) != SHT_RELA &&
-		  GET_LE(&in->sh_type) != SHT_REL)) &&
-		strcmp(name, ".altinstructions") &&
-		strcmp(name, ".altinstr_replacement");
-
-	if (!copy)
-		return;
-
-	if (out->count >= out->max_count)
-		fail("too many copied sections (max = %d)\n", out->max_count);
-
-	if (in_idx == out->in_shstrndx)
-		out->out_shstrndx = out->count;
-
-	out->table[out->count] = *in;
-	PUT_LE(&out->table[out->count].sh_name,
-	       BITSFUNC(find_shname)(out, name));
-
-	/* elfutils requires that a strtab have the correct type. */
-	if (!strcmp(name, ".fake_shstrtab"))
-		PUT_LE(&out->table[out->count].sh_type, SHT_STRTAB);
-
-	out->count++;
-}
-
-static void BITSFUNC(go)(void *addr, size_t len,
+static void BITSFUNC(go)(void *raw_addr, size_t raw_len,
+			 void *stripped_addr, size_t stripped_len,
 			 FILE *outfile, const char *name)
 {
 	int found_load = 0;
 	unsigned long load_size = -1;  /* Work around bogus warning */
-	unsigned long data_size;
-	ELF(Ehdr) *hdr = (ELF(Ehdr) *)addr;
+	unsigned long mapping_size;
+	ELF(Ehdr) *hdr = (ELF(Ehdr) *)raw_addr;
 	int i;
 	unsigned long j;
 	ELF(Shdr) *symtab_hdr = NULL, *strtab_hdr, *secstrings_hdr,
@@ -134,9 +20,7 @@ static void BITSFUNC(go)(void *addr, size_t len,
 	const char *secstrings;
 	INT_BITS syms[NSYMS] = {};
 
-	struct BITSFUNC(fake_sections) fake_sections = {};
-
-	ELF(Phdr) *pt = (ELF(Phdr) *)(addr + GET_LE(&hdr->e_phoff));
+	ELF(Phdr) *pt = (ELF(Phdr) *)(raw_addr + GET_LE(&hdr->e_phoff));
 
 	/* Walk the segment table. */
 	for (i = 0; i < GET_LE(&hdr->e_phnum); i++) {
@@ -154,14 +38,16 @@ static void BITSFUNC(go)(void *addr, size_t len,
 			load_size = GET_LE(&pt[i].p_memsz);
 			found_load = 1;
 		} else if (GET_LE(&pt[i].p_type) == PT_DYNAMIC) {
-			dyn = addr + GET_LE(&pt[i].p_offset);
-			dyn_end = addr + GET_LE(&pt[i].p_offset) +
+			dyn = raw_addr + GET_LE(&pt[i].p_offset);
+			dyn_end = raw_addr + GET_LE(&pt[i].p_offset) +
 				GET_LE(&pt[i].p_memsz);
 		}
 	}
 	if (!found_load)
 		fail("no PT_LOAD seg\n");
-	data_size = (load_size + 4095) / 4096 * 4096;
+
+	if (stripped_len < load_size)
+		fail("stripped input is too short\n");
 
 	/* Walk the dynamic table */
 	for (i = 0; dyn + i < dyn_end &&
@@ -173,11 +59,11 @@ static void BITSFUNC(go)(void *addr, size_t len,
 	}
 
 	/* Walk the section table */
-	secstrings_hdr = addr + GET_LE(&hdr->e_shoff) +
+	secstrings_hdr = raw_addr + GET_LE(&hdr->e_shoff) +
 		GET_LE(&hdr->e_shentsize)*GET_LE(&hdr->e_shstrndx);
-	secstrings = addr + GET_LE(&secstrings_hdr->sh_offset);
+	secstrings = raw_addr + GET_LE(&secstrings_hdr->sh_offset);
 	for (i = 0; i < GET_LE(&hdr->e_shnum); i++) {
-		ELF(Shdr) *sh = addr + GET_LE(&hdr->e_shoff) +
+		ELF(Shdr) *sh = raw_addr + GET_LE(&hdr->e_shoff) +
 			GET_LE(&hdr->e_shentsize) * i;
 		if (GET_LE(&sh->sh_type) == SHT_SYMTAB)
 			symtab_hdr = sh;
@@ -190,7 +76,7 @@ static void BITSFUNC(go)(void *addr, size_t len,
 	if (!symtab_hdr)
 		fail("no symbol table\n");
 
-	strtab_hdr = addr + GET_LE(&hdr->e_shoff) +
+	strtab_hdr = raw_addr + GET_LE(&hdr->e_shoff) +
 		GET_LE(&hdr->e_shentsize) * GET_LE(&symtab_hdr->sh_link);
 
 	/* Walk the symbol table */
@@ -198,9 +84,9 @@ static void BITSFUNC(go)(void *addr, size_t len,
 	     i < GET_LE(&symtab_hdr->sh_size) / GET_LE(&symtab_hdr->sh_entsize);
 	     i++) {
 		int k;
-		ELF(Sym) *sym = addr + GET_LE(&symtab_hdr->sh_offset) +
+		ELF(Sym) *sym = raw_addr + GET_LE(&symtab_hdr->sh_offset) +
 			GET_LE(&symtab_hdr->sh_entsize) * i;
-		const char *name = addr + GET_LE(&strtab_hdr->sh_offset) +
+		const char *name = raw_addr + GET_LE(&strtab_hdr->sh_offset) +
 			GET_LE(&sym->st_name);
 
 		for (k = 0; k < NSYMS; k++) {
@@ -219,48 +105,7 @@ static void BITSFUNC(go)(void *addr, size_t len,
 				syms[k] = GET_LE(&sym->st_value);
 			}
 		}
-
-		if (!strcmp(name, "fake_shstrtab")) {
-			ELF(Shdr) *sh;
-
-			fake_sections.in_shstrndx = GET_LE(&sym->st_shndx);
-			fake_sections.shstrtab = addr + GET_LE(&sym->st_value);
-			fake_sections.shstrtab_len = GET_LE(&sym->st_size);
-			sh = addr + GET_LE(&hdr->e_shoff) +
-				GET_LE(&hdr->e_shentsize) *
-				fake_sections.in_shstrndx;
-			fake_sections.shstr_offset = GET_LE(&sym->st_value) -
-				GET_LE(&sh->sh_addr);
-		}
-	}
-
-	/* Build the output section table. */
-	if (!syms[sym_VDSO_FAKE_SECTION_TABLE_START] ||
-	    !syms[sym_VDSO_FAKE_SECTION_TABLE_END])
-		fail("couldn't find fake section table\n");
-	if ((syms[sym_VDSO_FAKE_SECTION_TABLE_END] -
-	     syms[sym_VDSO_FAKE_SECTION_TABLE_START]) % sizeof(ELF(Shdr)))
-		fail("fake section table size isn't a multiple of sizeof(Shdr)\n");
-	fake_sections.table = addr + syms[sym_VDSO_FAKE_SECTION_TABLE_START];
-	fake_sections.table_offset = syms[sym_VDSO_FAKE_SECTION_TABLE_START];
-	fake_sections.max_count = (syms[sym_VDSO_FAKE_SECTION_TABLE_END] -
-				   syms[sym_VDSO_FAKE_SECTION_TABLE_START]) /
-		sizeof(ELF(Shdr));
-
-	BITSFUNC(init_sections)(&fake_sections);
-	for (i = 0; i < GET_LE(&hdr->e_shnum); i++) {
-		ELF(Shdr) *sh = addr + GET_LE(&hdr->e_shoff) +
-			GET_LE(&hdr->e_shentsize) * i;
-		BITSFUNC(copy_section)(&fake_sections, i, sh,
-				       secstrings + GET_LE(&sh->sh_name));
 	}
-	if (!fake_sections.out_shstrndx)
-		fail("didn't generate shstrndx?!?\n");
-
-	PUT_LE(&hdr->e_shoff, fake_sections.table_offset);
-	PUT_LE(&hdr->e_shentsize, sizeof(ELF(Shdr)));
-	PUT_LE(&hdr->e_shnum, fake_sections.count);
-	PUT_LE(&hdr->e_shstrndx, fake_sections.out_shstrndx);
 
 	/* Validate mapping addresses. */
 	for (i = 0; i < sizeof(special_pages) / sizeof(special_pages[0]); i++) {
@@ -281,10 +126,12 @@ static void BITSFUNC(go)(void *addr, size_t len,
 		fail("vvar_begin must be a multiple of 4096\n");
 
 	if (!name) {
-		fwrite(addr, load_size, 1, outfile);
+		fwrite(stripped_addr, stripped_len, 1, outfile);
 		return;
 	}
 
+	mapping_size = (stripped_len + 4095) / 4096 * 4096;
+
 	fprintf(outfile, "/* AUTOMATICALLY GENERATED -- DO NOT EDIT */\n\n");
 	fprintf(outfile, "#include <linux/linkage.h>\n");
 	fprintf(outfile, "#include <asm/page_types.h>\n");
@@ -292,20 +139,21 @@ static void BITSFUNC(go)(void *addr, size_t len,
 	fprintf(outfile, "\n");
 	fprintf(outfile,
 		"static unsigned char raw_data[%lu] __page_aligned_data = {",
-		data_size);
-	for (j = 0; j < load_size; j++) {
+		mapping_size);
+	for (j = 0; j < stripped_len; j++) {
 		if (j % 10 == 0)
 			fprintf(outfile, "\n\t");
-		fprintf(outfile, "0x%02X, ", (int)((unsigned char *)addr)[j]);
+		fprintf(outfile, "0x%02X, ",
+			(int)((unsigned char *)stripped_addr)[j]);
 	}
 	fprintf(outfile, "\n};\n\n");
 
 	fprintf(outfile, "static struct page *pages[%lu];\n\n",
-		data_size / 4096);
+		mapping_size / 4096);
 
 	fprintf(outfile, "const struct vdso_image %s = {\n", name);
 	fprintf(outfile, "\t.data = raw_data,\n");
-	fprintf(outfile, "\t.size = %lu,\n", data_size);
+	fprintf(outfile, "\t.size = %lu,\n", mapping_size);
 	fprintf(outfile, "\t.text_mapping = {\n");
 	fprintf(outfile, "\t\t.name = \"[vdso]\",\n");
 	fprintf(outfile, "\t\t.pages = pages,\n");
-- 
1.9.3

[tip:x86/vdso] x86, vdso: Get rid of the fake section mechanism

[tip-bot for Andy Lutomirski]

Commit-ID:  da861e18ecccb5c126b9eb95ff720ce082a46286
Gitweb:     http://git.kernel.org/tip/da861e18ecccb5c126b9eb95ff720ce082a46286
Author:     Andy Lutomirski <luto@amacapital.net>
AuthorDate: Thu, 10 Jul 2014 18:13:16 -0700
Committer:  H. Peter Anvin <hpa@linux.intel.com>
CommitDate: Fri, 11 Jul 2014 16:58:07 -0700

x86, vdso: Get rid of the fake section mechanism

Now that we can tolerate extra things dangling off the end of the
vdso image, we can strip the vdso the old fashioned way rather than
using an overcomplicated custom stripping algorithm.

This is a partial reversion of:
    6f121e5 x86, vdso: Reimplement vdso.so preparation in build-time C

Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Link: http://lkml.kernel.org/r/50e01ed6dcc0575d20afd782f9fe98d5ee3e2d8a.1405040914.git.luto@amacapital.net
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
---
 arch/x86/vdso/Makefile            |  16 +--
 arch/x86/vdso/vdso-fakesections.c |  21 ----
 arch/x86/vdso/vdso2c.c            | 116 +++++++++++++++++-----
 arch/x86/vdso/vdso2c.h            | 202 +++++---------------------------------
 4 files changed, 126 insertions(+), 229 deletions(-)

diff --git a/arch/x86/vdso/Makefile b/arch/x86/vdso/Makefile
index 61b04fe..5a4affe 100644
--- a/arch/x86/vdso/Makefile
+++ b/arch/x86/vdso/Makefile
@@ -10,7 +10,7 @@ VDSO32-$(CONFIG_X86_32)		:= y
 VDSO32-$(CONFIG_COMPAT)		:= y
 
 # files to link into the vdso
-vobjs-y := vdso-note.o vclock_gettime.o vgetcpu.o vdso-fakesections.o
+vobjs-y := vdso-note.o vclock_gettime.o vgetcpu.o
 
 # files to link into kernel
 obj-y				+= vma.o
@@ -37,7 +37,8 @@ vdso_img_sodbg := $(vdso_img-y:%=vdso%.so.dbg)
 obj-y += $(vdso_img_objs)
 targets += $(vdso_img_cfiles)
 targets += $(vdso_img_sodbg)
-.SECONDARY: $(vdso_img-y:%=$(obj)/vdso-image-%.c)
+.SECONDARY: $(vdso_img-y:%=$(obj)/vdso-image-%.c) \
+	$(vdso_img-y:%=$(obj)/vdso%.so)
 
 export CPPFLAGS_vdso.lds += -P -C
 
@@ -54,10 +55,10 @@ hostprogs-y			+= vdso2c
 
 quiet_cmd_vdso2c = VDSO2C  $@
 define cmd_vdso2c
-	$(obj)/vdso2c $< $@
+	$(obj)/vdso2c $< $(<:%.dbg=%) $@
 endef
 
-$(obj)/vdso-image-%.c: $(obj)/vdso%.so.dbg $(obj)/vdso2c FORCE
+$(obj)/vdso-image-%.c: $(obj)/vdso%.so.dbg $(obj)/vdso%.so $(obj)/vdso2c FORCE
 	$(call if_changed,vdso2c)
 
 #
@@ -113,6 +114,10 @@ $(obj)/%-x32.o: $(obj)/%.o FORCE
 
 targets += vdsox32.lds $(vobjx32s-y)
 
+$(obj)/%.so: OBJCOPYFLAGS := -S
+$(obj)/%.so: $(obj)/%.so.dbg
+	$(call if_changed,objcopy)
+
 $(obj)/vdsox32.so.dbg: $(src)/vdsox32.lds $(vobjx32s) FORCE
 	$(call if_changed,vdso)
 
@@ -134,7 +139,7 @@ override obj-dirs = $(dir $(obj)) $(obj)/vdso32/
 
 targets += vdso32/vdso32.lds
 targets += vdso32/note.o vdso32/vclock_gettime.o $(vdso32.so-y:%=vdso32/%.o)
-targets += vdso32/vclock_gettime.o vdso32/vdso-fakesections.o
+targets += vdso32/vclock_gettime.o
 
 $(obj)/vdso32.o: $(vdso32-images:%=$(obj)/%)
 
@@ -156,7 +161,6 @@ $(vdso32-images:%=$(obj)/%.dbg): KBUILD_CFLAGS = $(KBUILD_CFLAGS_32)
 $(vdso32-images:%=$(obj)/%.dbg): $(obj)/vdso32-%.so.dbg: FORCE \
 				 $(obj)/vdso32/vdso32.lds \
 				 $(obj)/vdso32/vclock_gettime.o \
-				 $(obj)/vdso32/vdso-fakesections.o \
 				 $(obj)/vdso32/note.o \
 				 $(obj)/vdso32/%.o
 	$(call if_changed,vdso)
diff --git a/arch/x86/vdso/vdso-fakesections.c b/arch/x86/vdso/vdso-fakesections.c
deleted file mode 100644
index aa5fbfa..0000000
--- a/arch/x86/vdso/vdso-fakesections.c
+++ /dev/null
@@ -1,21 +0,0 @@
-/*
- * Copyright 2014 Andy Lutomirski
- * Subject to the GNU Public License, v.2
- *
- * String table for loadable section headers.  See vdso2c.h for why
- * this exists.
- */
-
-const char fake_shstrtab[] __attribute__((section(".fake_shstrtab"))) =
-	".hash\0"
-	".dynsym\0"
-	".dynstr\0"
-	".gnu.version\0"
-	".gnu.version_d\0"
-	".dynamic\0"
-	".rodata\0"
-	".fake_shstrtab\0"  /* Yay, self-referential code. */
-	".note\0"
-	".eh_frame_hdr\0"
-	".eh_frame\0"
-	".text";
diff --git a/arch/x86/vdso/vdso2c.c b/arch/x86/vdso/vdso2c.c
index 22c54d0..8627db2 100644
--- a/arch/x86/vdso/vdso2c.c
+++ b/arch/x86/vdso/vdso2c.c
@@ -1,3 +1,53 @@
+/*
+ * vdso2c - A vdso image preparation tool
+ * Copyright (c) 2014 Andy Lutomirski and others
+ * Licensed under the GPL v2
+ *
+ * vdso2c requires stripped and unstripped input.  It would be trivial
+ * to fully strip the input in here, but, for reasons described below,
+ * we need to write a section table.  Doing this is more or less
+ * equivalent to dropping all non-allocatable sections, but it's
+ * easier to let objcopy handle that instead of doing it ourselves.
+ * If we ever need to do something fancier than what objcopy provides,
+ * it would be straightforward to add here.
+ *
+ * We're keep a section table for a few reasons:
+ *
+ * The Go runtime had a couple of bugs: it would read the section
+ * table to try to figure out how many dynamic symbols there were (it
+ * shouldn't have looked at the section table at all) and, if there
+ * were no SHT_SYNDYM section table entry, it would use an
+ * uninitialized value for the number of symbols.  An empty DYNSYM
+ * table would work, but I see no reason not to write a valid one (and
+ * keep full performance for old Go programs).  This hack is only
+ * needed on x86_64.
+ *
+ * The bug was introduced on 2012-08-31 by:
+ * https://code.google.com/p/go/source/detail?r=56ea40aac72b
+ * and was fixed on 2014-06-13 by:
+ * https://code.google.com/p/go/source/detail?r=fc1cd5e12595
+ *
+ * Binutils has issues debugging the vDSO: it reads the section table to
+ * find SHT_NOTE; it won't look at PT_NOTE for the in-memory vDSO, which
+ * would break build-id if we removed the section table.  Binutils
+ * also requires that shstrndx != 0.  See:
+ * https://sourceware.org/bugzilla/show_bug.cgi?id=17064
+ *
+ * elfutils might not look for PT_NOTE if there is a section table at
+ * all.  I don't know whether this matters for any practical purpose.
+ *
+ * For simplicity, rather than hacking up a partial section table, we
+ * just write a mostly complete one.  We omit non-dynamic symbols,
+ * though, since they're rather large.
+ *
+ * Once binutils gets fixed, we might be able to drop this for all but
+ * the 64-bit vdso, since build-id only works in kernel RPMs, and
+ * systems that update to new enough kernel RPMs will likely update
+ * binutils in sync.  build-id has never worked for home-built kernel
+ * RPMs without manual symlinking, and I suspect that no one ever does
+ * that.
+ */
+
 #include <inttypes.h>
 #include <stdint.h>
 #include <unistd.h>
@@ -61,7 +111,8 @@ static void fail(const char *format, ...)
 	va_start(ap, format);
 	fprintf(stderr, "Error: ");
 	vfprintf(stderr, format, ap);
-	unlink(outfilename);
+	if (outfilename)
+		unlink(outfilename);
 	exit(1);
 	va_end(ap);
 }
@@ -114,30 +165,53 @@ extern void bad_put_le(void);
 #include "vdso2c.h"
 #undef ELF_BITS
 
-static void go(void *addr, size_t len, FILE *outfile, const char *name)
+static void go(void *raw_addr, size_t raw_len,
+	       void *stripped_addr, size_t stripped_len,
+	       FILE *outfile, const char *name)
 {
-	Elf64_Ehdr *hdr = (Elf64_Ehdr *)addr;
+	Elf64_Ehdr *hdr = (Elf64_Ehdr *)raw_addr;
 
 	if (hdr->e_ident[EI_CLASS] == ELFCLASS64) {
-		go64(addr, len, outfile, name);
+		go64(raw_addr, raw_len, stripped_addr, stripped_len,
+		     outfile, name);
 	} else if (hdr->e_ident[EI_CLASS] == ELFCLASS32) {
-		go32(addr, len, outfile, name);
+		go32(raw_addr, raw_len, stripped_addr, stripped_len,
+		     outfile, name);
 	} else {
 		fail("unknown ELF class\n");
 	}
 }
 
+static void map_input(const char *name, void **addr, size_t *len, int prot)
+{
+	off_t tmp_len;
+
+	int fd = open(name, O_RDONLY);
+	if (fd == -1)
+		err(1, "%s", name);
+
+	tmp_len = lseek(fd, 0, SEEK_END);
+	if (tmp_len == (off_t)-1)
+		err(1, "lseek");
+	*len = (size_t)tmp_len;
+
+	*addr = mmap(NULL, tmp_len, prot, MAP_PRIVATE, fd, 0);
+	if (*addr == MAP_FAILED)
+		err(1, "mmap");
+
+	close(fd);
+}
+
 int main(int argc, char **argv)
 {
-	int fd;
-	off_t len;
-	void *addr;
+	size_t raw_len, stripped_len;
+	void *raw_addr, *stripped_addr;
 	FILE *outfile;
 	char *name, *tmp;
 	int namelen;
 
-	if (argc != 3) {
-		printf("Usage: vdso2c INPUT OUTPUT\n");
+	if (argc != 4) {
+		printf("Usage: vdso2c RAW_INPUT STRIPPED_INPUT OUTPUT\n");
 		return 1;
 	}
 
@@ -145,7 +219,7 @@ int main(int argc, char **argv)
 	 * Figure out the struct name.  If we're writing to a .so file,
 	 * generate raw output insted.
 	 */
-	name = strdup(argv[2]);
+	name = strdup(argv[3]);
 	namelen = strlen(name);
 	if (namelen >= 3 && !strcmp(name + namelen - 3, ".so")) {
 		name = NULL;
@@ -161,26 +235,18 @@ int main(int argc, char **argv)
 				*tmp = '_';
 	}
 
-	fd = open(argv[1], O_RDONLY);
-	if (fd == -1)
-		err(1, "%s", argv[1]);
-
-	len = lseek(fd, 0, SEEK_END);
-	if (len == (off_t)-1)
-		err(1, "lseek");
-
-	addr = mmap(NULL, len, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0);
-	if (addr == MAP_FAILED)
-		err(1, "mmap");
+	map_input(argv[1], &raw_addr, &raw_len, PROT_READ);
+	map_input(argv[2], &stripped_addr, &stripped_len, PROT_READ);
 
-	outfilename = argv[2];
+	outfilename = argv[3];
 	outfile = fopen(outfilename, "w");
 	if (!outfile)
 		err(1, "%s", argv[2]);
 
-	go(addr, (size_t)len, outfile, name);
+	go(raw_addr, raw_len, stripped_addr, stripped_len, outfile, name);
 
-	munmap(addr, len);
+	munmap(raw_addr, raw_len);
+	munmap(stripped_addr, stripped_len);
 	fclose(outfile);
 
 	return 0;
diff --git a/arch/x86/vdso/vdso2c.h b/arch/x86/vdso/vdso2c.h
index 2da32fbc..fd57829 100644
--- a/arch/x86/vdso/vdso2c.h
+++ b/arch/x86/vdso/vdso2c.h
@@ -4,128 +4,14 @@
  * are built for 32-bit userspace.
  */
 
-/*
- * We're writing a section table for a few reasons:
- *
- * The Go runtime had a couple of bugs: it would read the section
- * table to try to figure out how many dynamic symbols there were (it
- * shouldn't have looked at the section table at all) and, if there
- * were no SHT_SYNDYM section table entry, it would use an
- * uninitialized value for the number of symbols.  An empty DYNSYM
- * table would work, but I see no reason not to write a valid one (and
- * keep full performance for old Go programs).  This hack is only
- * needed on x86_64.
- *
- * The bug was introduced on 2012-08-31 by:
- * https://code.google.com/p/go/source/detail?r=56ea40aac72b
- * and was fixed on 2014-06-13 by:
- * https://code.google.com/p/go/source/detail?r=fc1cd5e12595
- *
- * Binutils has issues debugging the vDSO: it reads the section table to
- * find SHT_NOTE; it won't look at PT_NOTE for the in-memory vDSO, which
- * would break build-id if we removed the section table.  Binutils
- * also requires that shstrndx != 0.  See:
- * https://sourceware.org/bugzilla/show_bug.cgi?id=17064
- *
- * elfutils might not look for PT_NOTE if there is a section table at
- * all.  I don't know whether this matters for any practical purpose.
- *
- * For simplicity, rather than hacking up a partial section table, we
- * just write a mostly complete one.  We omit non-dynamic symbols,
- * though, since they're rather large.
- *
- * Once binutils gets fixed, we might be able to drop this for all but
- * the 64-bit vdso, since build-id only works in kernel RPMs, and
- * systems that update to new enough kernel RPMs will likely update
- * binutils in sync.  build-id has never worked for home-built kernel
- * RPMs without manual symlinking, and I suspect that no one ever does
- * that.
- */
-struct BITSFUNC(fake_sections)
-{
-	ELF(Shdr) *table;
-	unsigned long table_offset;
-	int count, max_count;
-
-	int in_shstrndx;
-	unsigned long shstr_offset;
-	const char *shstrtab;
-	size_t shstrtab_len;
-
-	int out_shstrndx;
-};
-
-static unsigned int BITSFUNC(find_shname)(struct BITSFUNC(fake_sections) *out,
-					  const char *name)
-{
-	const char *outname = out->shstrtab;
-	while (outname - out->shstrtab < out->shstrtab_len) {
-		if (!strcmp(name, outname))
-			return (outname - out->shstrtab) + out->shstr_offset;
-		outname += strlen(outname) + 1;
-	}
-
-	if (*name)
-		printf("Warning: could not find output name \"%s\"\n", name);
-	return out->shstr_offset + out->shstrtab_len - 1;  /* Use a null. */
-}
-
-static void BITSFUNC(init_sections)(struct BITSFUNC(fake_sections) *out)
-{
-	if (!out->in_shstrndx)
-		fail("didn't find the fake shstrndx\n");
-
-	memset(out->table, 0, out->max_count * sizeof(ELF(Shdr)));
-
-	if (out->max_count < 1)
-		fail("we need at least two fake output sections\n");
-
-	PUT_LE(&out->table[0].sh_type, SHT_NULL);
-	PUT_LE(&out->table[0].sh_name, BITSFUNC(find_shname)(out, ""));
-
-	out->count = 1;
-}
-
-static void BITSFUNC(copy_section)(struct BITSFUNC(fake_sections) *out,
-				   int in_idx, const ELF(Shdr) *in,
-				   const char *name)
-{
-	uint64_t flags = GET_LE(&in->sh_flags);
-
-	bool copy = flags & SHF_ALLOC &&
-		(GET_LE(&in->sh_size) ||
-		 (GET_LE(&in->sh_type) != SHT_RELA &&
-		  GET_LE(&in->sh_type) != SHT_REL)) &&
-		strcmp(name, ".altinstructions") &&
-		strcmp(name, ".altinstr_replacement");
-
-	if (!copy)
-		return;
-
-	if (out->count >= out->max_count)
-		fail("too many copied sections (max = %d)\n", out->max_count);
-
-	if (in_idx == out->in_shstrndx)
-		out->out_shstrndx = out->count;
-
-	out->table[out->count] = *in;
-	PUT_LE(&out->table[out->count].sh_name,
-	       BITSFUNC(find_shname)(out, name));
-
-	/* elfutils requires that a strtab have the correct type. */
-	if (!strcmp(name, ".fake_shstrtab"))
-		PUT_LE(&out->table[out->count].sh_type, SHT_STRTAB);
-
-	out->count++;
-}
-
-static void BITSFUNC(go)(void *addr, size_t len,
+static void BITSFUNC(go)(void *raw_addr, size_t raw_len,
+			 void *stripped_addr, size_t stripped_len,
 			 FILE *outfile, const char *name)
 {
 	int found_load = 0;
 	unsigned long load_size = -1;  /* Work around bogus warning */
-	unsigned long data_size;
-	ELF(Ehdr) *hdr = (ELF(Ehdr) *)addr;
+	unsigned long mapping_size;
+	ELF(Ehdr) *hdr = (ELF(Ehdr) *)raw_addr;
 	int i;
 	unsigned long j;
 	ELF(Shdr) *symtab_hdr = NULL, *strtab_hdr, *secstrings_hdr,
@@ -134,9 +20,7 @@ static void BITSFUNC(go)(void *addr, size_t len,
 	const char *secstrings;
 	INT_BITS syms[NSYMS] = {};
 
-	struct BITSFUNC(fake_sections) fake_sections = {};
-
-	ELF(Phdr) *pt = (ELF(Phdr) *)(addr + GET_LE(&hdr->e_phoff));
+	ELF(Phdr) *pt = (ELF(Phdr) *)(raw_addr + GET_LE(&hdr->e_phoff));
 
 	/* Walk the segment table. */
 	for (i = 0; i < GET_LE(&hdr->e_phnum); i++) {
@@ -154,14 +38,16 @@ static void BITSFUNC(go)(void *addr, size_t len,
 			load_size = GET_LE(&pt[i].p_memsz);
 			found_load = 1;
 		} else if (GET_LE(&pt[i].p_type) == PT_DYNAMIC) {
-			dyn = addr + GET_LE(&pt[i].p_offset);
-			dyn_end = addr + GET_LE(&pt[i].p_offset) +
+			dyn = raw_addr + GET_LE(&pt[i].p_offset);
+			dyn_end = raw_addr + GET_LE(&pt[i].p_offset) +
 				GET_LE(&pt[i].p_memsz);
 		}
 	}
 	if (!found_load)
 		fail("no PT_LOAD seg\n");
-	data_size = (load_size + 4095) / 4096 * 4096;
+
+	if (stripped_len < load_size)
+		fail("stripped input is too short\n");
 
 	/* Walk the dynamic table */
 	for (i = 0; dyn + i < dyn_end &&
@@ -173,11 +59,11 @@ static void BITSFUNC(go)(void *addr, size_t len,
 	}
 
 	/* Walk the section table */
-	secstrings_hdr = addr + GET_LE(&hdr->e_shoff) +
+	secstrings_hdr = raw_addr + GET_LE(&hdr->e_shoff) +
 		GET_LE(&hdr->e_shentsize)*GET_LE(&hdr->e_shstrndx);
-	secstrings = addr + GET_LE(&secstrings_hdr->sh_offset);
+	secstrings = raw_addr + GET_LE(&secstrings_hdr->sh_offset);
 	for (i = 0; i < GET_LE(&hdr->e_shnum); i++) {
-		ELF(Shdr) *sh = addr + GET_LE(&hdr->e_shoff) +
+		ELF(Shdr) *sh = raw_addr + GET_LE(&hdr->e_shoff) +
 			GET_LE(&hdr->e_shentsize) * i;
 		if (GET_LE(&sh->sh_type) == SHT_SYMTAB)
 			symtab_hdr = sh;
@@ -190,7 +76,7 @@ static void BITSFUNC(go)(void *addr, size_t len,
 	if (!symtab_hdr)
 		fail("no symbol table\n");
 
-	strtab_hdr = addr + GET_LE(&hdr->e_shoff) +
+	strtab_hdr = raw_addr + GET_LE(&hdr->e_shoff) +
 		GET_LE(&hdr->e_shentsize) * GET_LE(&symtab_hdr->sh_link);
 
 	/* Walk the symbol table */
@@ -198,9 +84,9 @@ static void BITSFUNC(go)(void *addr, size_t len,
 	     i < GET_LE(&symtab_hdr->sh_size) / GET_LE(&symtab_hdr->sh_entsize);
 	     i++) {
 		int k;
-		ELF(Sym) *sym = addr + GET_LE(&symtab_hdr->sh_offset) +
+		ELF(Sym) *sym = raw_addr + GET_LE(&symtab_hdr->sh_offset) +
 			GET_LE(&symtab_hdr->sh_entsize) * i;
-		const char *name = addr + GET_LE(&strtab_hdr->sh_offset) +
+		const char *name = raw_addr + GET_LE(&strtab_hdr->sh_offset) +
 			GET_LE(&sym->st_name);
 
 		for (k = 0; k < NSYMS; k++) {
@@ -219,48 +105,7 @@ static void BITSFUNC(go)(void *addr, size_t len,
 				syms[k] = GET_LE(&sym->st_value);
 			}
 		}
-
-		if (!strcmp(name, "fake_shstrtab")) {
-			ELF(Shdr) *sh;
-
-			fake_sections.in_shstrndx = GET_LE(&sym->st_shndx);
-			fake_sections.shstrtab = addr + GET_LE(&sym->st_value);
-			fake_sections.shstrtab_len = GET_LE(&sym->st_size);
-			sh = addr + GET_LE(&hdr->e_shoff) +
-				GET_LE(&hdr->e_shentsize) *
-				fake_sections.in_shstrndx;
-			fake_sections.shstr_offset = GET_LE(&sym->st_value) -
-				GET_LE(&sh->sh_addr);
-		}
-	}
-
-	/* Build the output section table. */
-	if (!syms[sym_VDSO_FAKE_SECTION_TABLE_START] ||
-	    !syms[sym_VDSO_FAKE_SECTION_TABLE_END])
-		fail("couldn't find fake section table\n");
-	if ((syms[sym_VDSO_FAKE_SECTION_TABLE_END] -
-	     syms[sym_VDSO_FAKE_SECTION_TABLE_START]) % sizeof(ELF(Shdr)))
-		fail("fake section table size isn't a multiple of sizeof(Shdr)\n");
-	fake_sections.table = addr + syms[sym_VDSO_FAKE_SECTION_TABLE_START];
-	fake_sections.table_offset = syms[sym_VDSO_FAKE_SECTION_TABLE_START];
-	fake_sections.max_count = (syms[sym_VDSO_FAKE_SECTION_TABLE_END] -
-				   syms[sym_VDSO_FAKE_SECTION_TABLE_START]) /
-		sizeof(ELF(Shdr));
-
-	BITSFUNC(init_sections)(&fake_sections);
-	for (i = 0; i < GET_LE(&hdr->e_shnum); i++) {
-		ELF(Shdr) *sh = addr + GET_LE(&hdr->e_shoff) +
-			GET_LE(&hdr->e_shentsize) * i;
-		BITSFUNC(copy_section)(&fake_sections, i, sh,
-				       secstrings + GET_LE(&sh->sh_name));
 	}
-	if (!fake_sections.out_shstrndx)
-		fail("didn't generate shstrndx?!?\n");
-
-	PUT_LE(&hdr->e_shoff, fake_sections.table_offset);
-	PUT_LE(&hdr->e_shentsize, sizeof(ELF(Shdr)));
-	PUT_LE(&hdr->e_shnum, fake_sections.count);
-	PUT_LE(&hdr->e_shstrndx, fake_sections.out_shstrndx);
 
 	/* Validate mapping addresses. */
 	for (i = 0; i < sizeof(special_pages) / sizeof(special_pages[0]); i++) {
@@ -281,10 +126,12 @@ static void BITSFUNC(go)(void *addr, size_t len,
 		fail("vvar_begin must be a multiple of 4096\n");
 
 	if (!name) {
-		fwrite(addr, load_size, 1, outfile);
+		fwrite(stripped_addr, stripped_len, 1, outfile);
 		return;
 	}
 
+	mapping_size = (stripped_len + 4095) / 4096 * 4096;
+
 	fprintf(outfile, "/* AUTOMATICALLY GENERATED -- DO NOT EDIT */\n\n");
 	fprintf(outfile, "#include <linux/linkage.h>\n");
 	fprintf(outfile, "#include <asm/page_types.h>\n");
@@ -292,20 +139,21 @@ static void BITSFUNC(go)(void *addr, size_t len,
 	fprintf(outfile, "\n");
 	fprintf(outfile,
 		"static unsigned char raw_data[%lu] __page_aligned_data = {",
-		data_size);
-	for (j = 0; j < load_size; j++) {
+		mapping_size);
+	for (j = 0; j < stripped_len; j++) {
 		if (j % 10 == 0)
 			fprintf(outfile, "\n\t");
-		fprintf(outfile, "0x%02X, ", (int)((unsigned char *)addr)[j]);
+		fprintf(outfile, "0x%02X, ",
+			(int)((unsigned char *)stripped_addr)[j]);
 	}
 	fprintf(outfile, "\n};\n\n");
 
 	fprintf(outfile, "static struct page *pages[%lu];\n\n",
-		data_size / 4096);
+		mapping_size / 4096);
 
 	fprintf(outfile, "const struct vdso_image %s = {\n", name);
 	fprintf(outfile, "\t.data = raw_data,\n");
-	fprintf(outfile, "\t.size = %lu,\n", data_size);
+	fprintf(outfile, "\t.size = %lu,\n", mapping_size);
 	fprintf(outfile, "\t.text_mapping = {\n");
 	fprintf(outfile, "\t\t.name = \"[vdso]\",\n");
 	fprintf(outfile, "\t\t.pages = pages,\n");