Skip to contentSkip to conversation

x86/vdso: Remove remnants of the fake section table

PATCH Superseded Sean Christopherson 2 participants 9 messages Dec 4, 2018

Subsystems (3)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

cover x86/vdso: Remove remnants of the fake section table

Remove pieces of the vDSO's fake section table mechanism that were left behind when the vDSO build process reverted to using "objdump -S" to strip the userspace image.

One of the removed pieces is a 0x340 byte reservation (for 64-bit) in the .rodata section. Trimming that fat drops the current image's size to under 4k, i.e. reduces the effective size of the userspace vDSO mapping by a full page.

Sean Christopherson (2):
x86/vdso: Remove obsolete "fake section table" reservation x86/vdso: Remove a stale/misleading comment from the linker script

 arch/x86/entry/vdso/vdso-layout.lds.S | 27 ---------------------------
 arch/x86/entry/vdso/vdso2c.c          |  6 ------
 2 files changed, 33 deletions(-)
Signature · 1 line
2.19.2
#1 lore raw

PATCH 1/2 x86/vdso: Remove obsolete "fake section table" reservation

At one point the vDSO image was manually stripped down by vdso2c in an attempt to minimize the size of the image mapped into userspace. Part of that stripping process involved building a fake section table so as not to break userspace processes that parse the section table. Memory for the fake section table was reserved in the .rodata section so that vdso2c could simply copy the entire PT_LOAD segment into the userspace image after building the fake table.

Eventually, the entire fake section table approach was dropped in favor of stripping the vdso "the old fashioned way", i.e. via objdump -S. But, the reservation in .rodata for the fake table was left behind. Remove the reserveration along with a few other related defines and section entries.

Removing the fake section table placeholder zaps a whopping 0x340 bytes from the 64-bit vDSO image, which drops the current image's size to under 4k, i.e. reduces the effective size of the userspace vDSO mapping by a full page.

Fixes da861e18eccc ("x86, vdso: Get rid of the fake section mechanism")
Cc Andy Lutomirski <[email protected]>
---
 arch/x86/entry/vdso/vdso-layout.lds.S | 22 ----------------------
 arch/x86/entry/vdso/vdso2c.c          |  6 ------
 2 files changed, 28 deletions(-)
arch/x86/entry/vdso/vdso-layout.lds.S +0 -22
index acfd5ba7d943..0cedc905c8d6 100644
@@ -7,16 +7,6 @@
  * This script controls its layout.
  */
 
-#if defined(BUILD_VDSO64)
-# define SHDR_SIZE 64
-#elif defined(BUILD_VDSO32) || defined(BUILD_VDSOX32)
-# define SHDR_SIZE 40
-#else
-# error unknown VDSO target
-#endif
-
-#define NUM_FAKE_SHDRS 13
-
 SECTIONS
 {
 	/*
@@ -60,20 +50,8 @@ SECTIONS
 		*(.bss*)
 		*(.dynbss*)
 		*(.gnu.linkonce.b.*)
-
-		/*
-		 * Ideally this would live in a C file, but that won't
-		 * work cleanly for x32 until we start building the x32
-		 * C code using an x32 toolchain.
-		 */
-		VDSO_FAKE_SECTION_TABLE_START = .;
-		. = . + NUM_FAKE_SHDRS * SHDR_SIZE;
-		VDSO_FAKE_SECTION_TABLE_END = .;
 	}						:text
 
-	.fake_shstrtab	: { *(.fake_shstrtab) }		:text
-
-
 	.note		: { *(.note.*) }		:text	:note
 
 	.eh_frame_hdr	: { *(.eh_frame_hdr) }		:text	:eh_frame_hdr
arch/x86/entry/vdso/vdso2c.c +0 -6
index 4674f58581a1..2479a454b15c 100644
@@ -98,12 +98,6 @@ struct vdso_sym required_syms[] = {
 	[sym_hpet_page] = {"hpet_page", true},
 	[sym_pvclock_page] = {"pvclock_page", true},
 	[sym_hvclock_page] = {"hvclock_page", true},
-	[sym_VDSO_FAKE_SECTION_TABLE_START] = {
-		"VDSO_FAKE_SECTION_TABLE_START", false
-	},
-	[sym_VDSO_FAKE_SECTION_TABLE_END] = {
-		"VDSO_FAKE_SECTION_TABLE_END", false
-	},
 	{"VDSO32_NOTE_MASK", true},
 	{"__kernel_vsyscall", true},
 	{"__kernel_sigreturn", true},
Signature · 1 line
2.19.2
#2 lore raw

Re: [PATCH 1/2] x86/vdso: Remove obsolete "fake section table" reservation

Quoted 28 lines · Sean Christopherson

At one point the vDSO image was manually stripped down by vdso2c in an attempt to minimize the size of the image mapped into userspace. Part of that stripping process involved building a fake section table so as not to break userspace processes that parse the section table. Memory for the fake section table was reserved in the .rodata section so that vdso2c could simply copy the entire PT_LOAD segment into the userspace image after building the fake table.

Eventually, the entire fake section table approach was dropped in favor of stripping the vdso "the old fashioned way", i.e. via objdump -S. But, the reservation in .rodata for the fake table was left behind. Remove the reserveration along with a few other related defines and section entries.

Removing the fake section table placeholder zaps a whopping 0x340 bytes from the 64-bit vDSO image, which drops the current image's size to under 4k, i.e. reduces the effective size of the userspace vDSO mapping by a full page.

Fixes da861e18eccc ("x86, vdso: Get rid of the fake section mechanism")
Cc Andy Lutomirski <[email protected]>

---

arch/x86/entry/vdso/vdso2c.c +0 -0
index 4674f58581a1..2479a454b15c 100644

arch/x86/entry/vdso/vdso2c.c struct vdso_sym required_syms[] = { · line 106

Quoted 9 lines
arch/x86/entry/vdso/vdso2c.c +0 -6
@@ -98,12 +98,6 @@ struct vdso_sym required_syms[] = {
 	[sym_hpet_page] = {"hpet_page", true},
 	[sym_pvclock_page] = {"pvclock_page", true},
 	[sym_hvclock_page] = {"hvclock_page", true},
-	[sym_VDSO_FAKE_SECTION_TABLE_START] = {
-		"VDSO_FAKE_SECTION_TABLE_START", false
-	},
-	[sym_VDSO_FAKE_SECTION_TABLE_END] = {
-		"VDSO_FAKE_SECTION_TABLE_END", false
-	},

Doh, I missed removing the definitions for sym_VDSO_FAKE_SECTION_TABLE_*.

Quoted 6 lines
arch/x86/entry/vdso/vdso2c.c +0 -0
 	{"VDSO32_NOTE_MASK", true},
 	{"__kernel_vsyscall", true},
 	{"__kernel_sigreturn", true},
Signature · 1 line
2.19.2
#3 lore raw

Re: [PATCH 1/2] x86/vdso: Remove obsolete "fake section table" reservation

Quoted 32 lines · Sean Christopherson
Quoted 31 lines · Sean Christopherson

At one point the vDSO image was manually stripped down by vdso2c in an attempt to minimize the size of the image mapped into userspace. Part of that stripping process involved building a fake section table so as not to break userspace processes that parse the section table. Memory for the fake section table was reserved in the .rodata section so that vdso2c could simply copy the entire PT_LOAD segment into the userspace image after building the fake table.

Eventually, the entire fake section table approach was dropped in favor of stripping the vdso "the old fashioned way", i.e. via objdump -S. But, the reservation in .rodata for the fake table was left behind. Remove the reserveration along with a few other related defines and section entries.

Removing the fake section table placeholder zaps a whopping 0x340 bytes from the 64-bit vDSO image, which drops the current image's size to under 4k, i.e. reduces the effective size of the userspace vDSO mapping by a full page.

Fixes da861e18eccc ("x86, vdso: Get rid of the fake section mechanism")
Cc Andy Lutomirski <[email protected]>

---

arch/x86/entry/vdso/vdso2c.c +0 -0
index 4674f58581a1..2479a454b15c 100644
@@ -98,12 +98,6 @@ struct vdso_sym required_syms[] = {
 	[sym_hpet_page] = {"hpet_page", true},
 	[sym_pvclock_page] = {"pvclock_page", true},
 	[sym_hvclock_page] = {"hvclock_page", true},
Quoted 8 lines

arch/x86/entry/vdso/vdso2c.c struct vdso_sym required_syms[] = { · line 106

Quoted 6 lines
arch/x86/entry/vdso/vdso2c.c +0 -6
-	[sym_VDSO_FAKE_SECTION_TABLE_START] = {
-		"VDSO_FAKE_SECTION_TABLE_START", false
-	},
-	[sym_VDSO_FAKE_SECTION_TABLE_END] = {
-		"VDSO_FAKE_SECTION_TABLE_END", false
-	},

Doh, I missed removing the definitions for sym_VDSO_FAKE_SECTION_TABLE_*.

And with sym_VDSO_FAKE_SECTION_TABLE_* gone all symbols are exported, meaning required_syms can be a char* array and struct vdso_sym can be removed.

Quoted 6 lines
Quoted 6 lines
arch/x86/entry/vdso/vdso2c.c +0 -0
 	{"VDSO32_NOTE_MASK", true},
 	{"__kernel_vsyscall", true},
 	{"__kernel_sigreturn", true},
Signature · 1 line
2.19.2
maintainer
#4 lore raw

Re: [PATCH 1/2] x86/vdso: Remove obsolete "fake section table" reservation

Quoted 43 lines · Sean Christopherson
Quoted 32 lines · Sean Christopherson
Quoted 31 lines · Sean Christopherson

At one point the vDSO image was manually stripped down by vdso2c in an attempt to minimize the size of the image mapped into userspace. Part of that stripping process involved building a fake section table so as not to break userspace processes that parse the section table. Memory for the fake section table was reserved in the .rodata section so that vdso2c could simply copy the entire PT_LOAD segment into the userspace image after building the fake table.

Eventually, the entire fake section table approach was dropped in favor of stripping the vdso "the old fashioned way", i.e. via objdump -S. But, the reservation in .rodata for the fake table was left behind. Remove the reserveration along with a few other related defines and section entries.

Removing the fake section table placeholder zaps a whopping 0x340 bytes from the 64-bit vDSO image, which drops the current image's size to under 4k, i.e. reduces the effective size of the userspace vDSO mapping by a full page.

Fixes da861e18eccc ("x86, vdso: Get rid of the fake section mechanism")
Cc Andy Lutomirski <[email protected]>

---

arch/x86/entry/vdso/vdso2c.c +0 -0
index 4674f58581a1..2479a454b15c 100644
@@ -98,12 +98,6 @@ struct vdso_sym required_syms[] = {
    [sym_hpet_page] = {"hpet_page", true},
    [sym_pvclock_page] = {"pvclock_page", true},
    [sym_hvclock_page] = {"hvclock_page", true},
Quoted 8 lines

arch/x86/entry/vdso/vdso2c.c struct vdso_sym required_syms[] = { · line 106

Quoted 6 lines
arch/x86/entry/vdso/vdso2c.c +0 -6
-   [sym_VDSO_FAKE_SECTION_TABLE_START] = {
-           "VDSO_FAKE_SECTION_TABLE_START", false
-   },
-   [sym_VDSO_FAKE_SECTION_TABLE_END] = {
-           "VDSO_FAKE_SECTION_TABLE_END", false
-   },

Doh, I missed removing the definitions for sym_VDSO_FAKE_SECTION_TABLE_*.

Quoted 3 lines

And with sym_VDSO_FAKE_SECTION_TABLE_* gone all symbols are exported, meaning required_syms can be a char* array and struct vdso_sym can be removed.

I bet that we'll want that field to parse out the extable entries once all the dust settles, though.

--Andy

#5 lore raw

Re: [PATCH 1/2] x86/vdso: Remove obsolete "fake section table" reservation

Quoted 45 lines · Andy Lutomirski
Quoted 43 lines · Sean Christopherson
Quoted 40 lines · Sean Christopherson
Quoted 28 lines · Sean Christopherson

At one point the vDSO image was manually stripped down by vdso2c in an attempt to minimize the size of the image mapped into userspace. Part of that stripping process involved building a fake section table so as not to break userspace processes that parse the section table. Memory for the fake section table was reserved in the .rodata section so that vdso2c could simply copy the entire PT_LOAD segment into the userspace image after building the fake table.

Eventually, the entire fake section table approach was dropped in favor of stripping the vdso "the old fashioned way", i.e. via objdump -S. But, the reservation in .rodata for the fake table was left behind. Remove the reserveration along with a few other related defines and section entries.

Removing the fake section table placeholder zaps a whopping 0x340 bytes from the 64-bit vDSO image, which drops the current image's size to under 4k, i.e. reduces the effective size of the userspace vDSO mapping by a full page.

Fixes da861e18eccc ("x86, vdso: Get rid of the fake section mechanism")
Cc Andy Lutomirski <[email protected]>

---

arch/x86/entry/vdso/vdso2c.c +0 -0
index 4674f58581a1..2479a454b15c 100644

arch/x86/entry/vdso/vdso2c.c struct vdso_sym required_syms[] = { · line 106

Quoted 9 lines
arch/x86/entry/vdso/vdso2c.c +0 -6
@@ -98,12 +98,6 @@ struct vdso_sym required_syms[] = {
    [sym_hpet_page] = {"hpet_page", true},
    [sym_pvclock_page] = {"pvclock_page", true},
    [sym_hvclock_page] = {"hvclock_page", true},
-   [sym_VDSO_FAKE_SECTION_TABLE_START] = {
-           "VDSO_FAKE_SECTION_TABLE_START", false
-   },
-   [sym_VDSO_FAKE_SECTION_TABLE_END] = {
-           "VDSO_FAKE_SECTION_TABLE_END", false
-   },

Doh, I missed removing the definitions for sym_VDSO_FAKE_SECTION_TABLE_*.

Quoted 6 lines
Quoted 3 lines

And with sym_VDSO_FAKE_SECTION_TABLE_* gone all symbols are exported, meaning required_syms can be a char* array and struct vdso_sym can be removed.

I bet that we'll want that field to parse out the extable entries once all the dust settles, though.

Hmm, the extable stuff will be sections, a la .altinstructions, I don't think we'll need separate symbols.

What if I send out a RFC for the extable stuff in parallel to this series and to the SGX series? It'd also be nice to iterate on that code without having to spin a full SGX series.

maintainer
#6 lore raw

Re: [PATCH 1/2] x86/vdso: Remove obsolete "fake section table" reservation

Quoted 51 lines · Sean Christopherson
Quoted 49 lines · Andy Lutomirski
Quoted 46 lines · Sean Christopherson
Quoted 32 lines · Sean Christopherson
Quoted 31 lines · Sean Christopherson

At one point the vDSO image was manually stripped down by vdso2c in an attempt to minimize the size of the image mapped into userspace. Part of that stripping process involved building a fake section table so as not to break userspace processes that parse the section table. Memory for the fake section table was reserved in the .rodata section so that vdso2c could simply copy the entire PT_LOAD segment into the userspace image after building the fake table.

Eventually, the entire fake section table approach was dropped in favor of stripping the vdso "the old fashioned way", i.e. via objdump -S. But, the reservation in .rodata for the fake table was left behind. Remove the reserveration along with a few other related defines and section entries.

Removing the fake section table placeholder zaps a whopping 0x340 bytes from the 64-bit vDSO image, which drops the current image's size to under 4k, i.e. reduces the effective size of the userspace vDSO mapping by a full page.

Fixes da861e18eccc ("x86, vdso: Get rid of the fake section mechanism")
Cc Andy Lutomirski <[email protected]>

---

arch/x86/entry/vdso/vdso2c.c +0 -0
index 4674f58581a1..2479a454b15c 100644
@@ -98,12 +98,6 @@ struct vdso_sym required_syms[] = {
    [sym_hpet_page] = {"hpet_page", true},
    [sym_pvclock_page] = {"pvclock_page", true},
    [sym_hvclock_page] = {"hvclock_page", true},
Quoted 8 lines

arch/x86/entry/vdso/vdso2c.c struct vdso_sym required_syms[] = { · line 106

Quoted 6 lines
arch/x86/entry/vdso/vdso2c.c +0 -6
-   [sym_VDSO_FAKE_SECTION_TABLE_START] = {
-           "VDSO_FAKE_SECTION_TABLE_START", false
-   },
-   [sym_VDSO_FAKE_SECTION_TABLE_END] = {
-           "VDSO_FAKE_SECTION_TABLE_END", false
-   },

Doh, I missed removing the definitions for sym_VDSO_FAKE_SECTION_TABLE_*.

And with sym_VDSO_FAKE_SECTION_TABLE_* gone all symbols are exported, meaning required_syms can be a char* array and struct vdso_sym can be removed.

Quoted 5 lines
Quoted 2 lines

I bet that we'll want that field to parse out the extable entries once all the dust settles, though.

Hmm, the extable stuff will be sections, a la .altinstructions, I don't think we'll need separate symbols.

In my mind, the easiest nice implementation is to have a section that gets thrown away by objcopy --remove-section, and to have the actual table in that section be bracketed by a pair of symbols, since vdso2c already parses that. But vdso2c could certainly also just find the section direclty. Then vdso2c could expose an list of extable entries in the vdso_image, and that would be that. Does that seem reasonable?

Quoted 4 lines

What if I send out a RFC for the extable stuff in parallel to this series and to the SGX series? It'd also be nice to iterate on that code without having to spin a full SGX series.

Sounds great!

#7 lore raw

PATCH 2/2 x86/vdso: Remove a stale/misleading comment from the linker script

In reply to Sean Christopherson

Once upon a time, vdso2c aggressively stripped data from the vDSO image when generating the final userspace image. This included stripping the .altinstructions and .altinstr_replacement sections. Eventually, the stripping process reverted to "objdump -S" and no longer removed the aforementioned sections, but the comment remained.

Keeping the .alt* sections at the end of the PT_LOAD segment is no longer necessary, but there's no harm in doing so and it's a helpful reminder that they don't need to be included in the final vDSO image, i.e. someone may want to take another stab at zapping/stripping the unneeded sections.

Fixes da861e18eccc ("x86, vdso: Get rid of the fake section mechanism")
Cc Andy Lutomirski <[email protected]>
---
 arch/x86/entry/vdso/vdso-layout.lds.S | 5 -----
 1 file changed, 5 deletions(-)
arch/x86/entry/vdso/vdso-layout.lds.S +0 -5
index 0cedc905c8d6..93c6dc7812d0 100644
@@ -65,11 +65,6 @@ SECTIONS
 
 	.text		: { *(.text*) }			:text	=0x90909090,
 
-	/*
-	 * At the end so that eu-elflint stays happy when vdso2c strips
-	 * these.  A better implementation would avoid allocating space
-	 * for these.
-	 */
 	.altinstructions	: { *(.altinstructions) }	:text
 	.altinstr_replacement	: { *(.altinstr_replacement) }	:text
 
Signature · 1 line
2.19.2
maintainer Acked-by on v1 2/2
#8 lore raw

Re: [PATCH 2/2] x86/vdso: Remove a stale/misleading comment from the linker script

Quoted 7 lines · Sean Christopherson

Once upon a time, vdso2c aggressively stripped data from the vDSO image when generating the final userspace image. This included stripping the .altinstructions and .altinstr_replacement sections. Eventually, the stripping process reverted to "objdump -S" and no longer removed the aforementioned sections, but the comment remained.

Quoted 5 lines

Keeping the .alt* sections at the end of the PT_LOAD segment is no longer necessary, but there's no harm in doing so and it's a helpful reminder that they don't need to be included in the final vDSO image, i.e. someone may want to take another stab at zapping/stripping the unneeded sections.