Skip to contentSkip to conversation

x86, kaslr: Kernel base can be randomized at 0-1G offset.

PATCH In review WANG Chao 3 participants 7 messages Mar 14, 2014 – Mar 20, 2014

Subsystems (2)
lore
Share & tools
lore

WANG Chao

author
#0 lore raw

PATCH v2 1/1 x86, kaslr: Kernel base can be randomized at 0-1G offset.

With kASLR enabled (CONFIG_RANDOMIZED_BASE=y), kernel virtual address base is PAGE_OFFSET plus a randomized offset from 0 to 1G.

Current kexec-tools gets kernel vaddr and size from /proc/kcore. It assumes kernel vaddr start/end is within the range [0,512M). If kaslr enabled, kernel vaddr start/end will stay at [0+offset, 512M+offset).

To adapt kaslr, introduce a new macro X86_64_RANDOMIZED_BASE_MAX_OFFSET to address the max offset and use this macro to filter out the kernel text PT_LOAD from /proc/kcore.

---
 kexec/arch/i386/crashdump-x86.c | 3 ++-
 kexec/arch/i386/crashdump-x86.h | 3 +++
 2 files changed, 5 insertions(+), 1 deletion(-)
kexec/arch/i386/crashdump-x86.c +2 -1
index cb19e7d..1e6d3a3 100644
@@ -156,7 +156,8 @@ static int get_kernel_vaddr_and_size(struct kexec_info *UNUSED(info),
 
 			/* Look for kernel text mapping header. */
 			if ((saddr >= X86_64__START_KERNEL_map) &&
-			    (eaddr <= X86_64__START_KERNEL_map + X86_64_KERNEL_TEXT_SIZE)) {
+			    (saddr <= X86_64__START_KERNEL_map + X86_64_RANDOMIZE_BASE_MAX_OFFSET) &&
+			    (eaddr - saddr  < X86_64_KERNEL_TEXT_SIZE)) {
 				saddr = _ALIGN_DOWN(saddr, X86_64_KERN_VADDR_ALIGN);
 				elf_info->kern_vaddr_start = saddr;
 				size = eaddr - saddr;
kexec/arch/i386/crashdump-x86.h +3 -0
index e68b626..71a09f8 100644
@@ -15,6 +15,9 @@ int load_crashdump_segments(struct kexec_info *info, char *mod_cmdline,
 #define X86_64_PAGE_OFFSET_PRE_2_6_27	0xffff810000000000ULL
 #define X86_64_PAGE_OFFSET		0xffff880000000000ULL
 
+/* kASLR - Kernel base offset could be randomized up to 1G */
+#define X86_64_RANDOMIZE_BASE_MAX_OFFSET	0x40000000
+
 #define X86_64_MAXMEM        		0x3fffffffffffUL
 
 /* Kernel text size */
Signature · 7 lines
1.8.5.3


_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec
Acked-by on v2 1/1
#1 lore raw
Quoted 36 lines · WANG Chao

With kASLR enabled (CONFIG_RANDOMIZED_BASE=y), kernel virtual address base is PAGE_OFFSET plus a randomized offset from 0 to 1G.

Current kexec-tools gets kernel vaddr and size from /proc/kcore. It assumes kernel vaddr start/end is within the range [0,512M). If kaslr enabled, kernel vaddr start/end will stay at [0+offset, 512M+offset).

To adapt kaslr, introduce a new macro X86_64_RANDOMIZED_BASE_MAX_OFFSET to address the max offset and use this macro to filter out the kernel text PT_LOAD from /proc/kcore.

---
 kexec/arch/i386/crashdump-x86.c | 3 ++-
 kexec/arch/i386/crashdump-x86.h | 3 +++
 2 files changed, 5 insertions(+), 1 deletion(-)
kexec/arch/i386/crashdump-x86.c +2 -1
index cb19e7d..1e6d3a3 100644
@@ -156,7 +156,8 @@ static int get_kernel_vaddr_and_size(struct kexec_info *UNUSED(info),
 
 			/* Look for kernel text mapping header. */
 			if ((saddr >= X86_64__START_KERNEL_map) &&
-			    (eaddr <= X86_64__START_KERNEL_map + X86_64_KERNEL_TEXT_SIZE)) {
+			    (saddr <= X86_64__START_KERNEL_map + X86_64_RANDOMIZE_BASE_MAX_OFFSET) &&
+			    (eaddr - saddr  < X86_64_KERNEL_TEXT_SIZE)) {
 				saddr = _ALIGN_DOWN(saddr, X86_64_KERN_VADDR_ALIGN);
 				elf_info->kern_vaddr_start = saddr;
 				size = eaddr - saddr;
kexec/arch/i386/crashdump-x86.h +0 -0
index e68b626..71a09f8 100644

kexec/arch/i386/crashdump-x86.h int load_crashdump_segments(struct kexec_info *info, char *mod_cmdline, · line 23

Quoted 9 lines
kexec/arch/i386/crashdump-x86.h +3 -0
@@ -15,6 +15,9 @@ int load_crashdump_segments(struct kexec_info *info, char *mod_cmdline,
 #define X86_64_PAGE_OFFSET_PRE_2_6_27	0xffff810000000000ULL
 #define X86_64_PAGE_OFFSET		0xffff880000000000ULL
 
+/* kASLR - Kernel base offset could be randomized up to 1G */
+#define X86_64_RANDOMIZE_BASE_MAX_OFFSET	0x40000000
+
 #define X86_64_MAXMEM        		0x3fffffffffffUL
 
 /* Kernel text size */

Thanks
Dave

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

#2 lore raw
In reply to WANG Chao
Quoted 6 lines · WANG Chao

With kASLR enabled (CONFIG_RANDOMIZED_BASE=y), kernel virtual address base is PAGE_OFFSET plus a randomized offset from 0 to 1G.

Current kexec-tools gets kernel vaddr and size from /proc/kcore. It assumes kernel vaddr start/end is within the range [0,512M). If kaslr enabled, kernel vaddr start/end will stay at [0+offset, 512M+offset).

Hi Chao,

Documentation/x86/x86_64/mm.txt still says that kernel text mapping area is 512MB.

ffffffff80000000 - ffffffffa0000000 (=512 MB) kernel text mapping, from phys 0

So has that changed now due to kASLR.

Thanks
Vivek

Quoted 41 lines

To adapt kaslr, introduce a new macro X86_64_RANDOMIZED_BASE_MAX_OFFSET to address the max offset and use this macro to filter out the kernel text PT_LOAD from /proc/kcore.

---
 kexec/arch/i386/crashdump-x86.c | 3 ++-
 kexec/arch/i386/crashdump-x86.h | 3 +++
 2 files changed, 5 insertions(+), 1 deletion(-)
kexec/arch/i386/crashdump-x86.c +2 -1
index cb19e7d..1e6d3a3 100644
@@ -156,7 +156,8 @@ static int get_kernel_vaddr_and_size(struct kexec_info *UNUSED(info),
 
 			/* Look for kernel text mapping header. */
 			if ((saddr >= X86_64__START_KERNEL_map) &&
-			    (eaddr <= X86_64__START_KERNEL_map + X86_64_KERNEL_TEXT_SIZE)) {
+			    (saddr <= X86_64__START_KERNEL_map + X86_64_RANDOMIZE_BASE_MAX_OFFSET) &&
+			    (eaddr - saddr  < X86_64_KERNEL_TEXT_SIZE)) {
 				saddr = _ALIGN_DOWN(saddr, X86_64_KERN_VADDR_ALIGN);
 				elf_info->kern_vaddr_start = saddr;
 				size = eaddr - saddr;
kexec/arch/i386/crashdump-x86.h +3 -0
index e68b626..71a09f8 100644
@@ -15,6 +15,9 @@ int load_crashdump_segments(struct kexec_info *info, char *mod_cmdline,
 #define X86_64_PAGE_OFFSET_PRE_2_6_27	0xffff810000000000ULL
 #define X86_64_PAGE_OFFSET		0xffff880000000000ULL
 
+/* kASLR - Kernel base offset could be randomized up to 1G */
+#define X86_64_RANDOMIZE_BASE_MAX_OFFSET	0x40000000
+
 #define X86_64_MAXMEM        		0x3fffffffffffUL
 
 /* Kernel text size */
Signature · 1 line
1.8.5.3

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

author
#3 lore raw
Quoted 7 lines · Vivek Goyal
Quoted 6 lines · WANG Chao

With kASLR enabled (CONFIG_RANDOMIZED_BASE=y), kernel virtual address base is PAGE_OFFSET plus a randomized offset from 0 to 1G.

Current kexec-tools gets kernel vaddr and size from /proc/kcore. It assumes kernel vaddr start/end is within the range [0,512M). If kaslr enabled, kernel vaddr start/end will stay at [0+offset, 512M+offset).

NACK this patch myself.

There are several mistakes I made. I misunderstood some concepts. Then I realize this kASLR issue is not trivial to fix.

I think if kexec-tools needs to get kernel text mapping from kcore in kALSR case, the max base offset (CONFIG_RANDOMIZE_MAX_BASE_OFFSET) of the kernel text area must be exposed to userspace some where. We can use that value to determine which area is for kernel text mapping and which is for modules text mapping.

Quoted 10 lines

Hi Chao,

Documentation/x86/x86_64/mm.txt still says that kernel text mapping area is 512MB.

ffffffff80000000 - ffffffffa0000000 (=512 MB) kernel text mapping, from phys 0

So has that changed now due to kASLR.

Yes, with kASLR enabled, kernel text mapping is as following

ffffffff80000000 - (ffffffff80000000+CONFIG_RANDOMIZE_BASE_MAX_OFFSET)

That said, if using CONFIG_RANDOMIZE_BASE_MAX_OFFSET=0x40000000 by default, the kernel text mapping is as following:

ffffffff80000000 - ffffffffc0000000

Thanks
WANG Chao

Quoted 37 lines

Thanks
Vivek

Quoted 33 lines

To adapt kaslr, introduce a new macro X86_64_RANDOMIZED_BASE_MAX_OFFSET to address the max offset and use this macro to filter out the kernel text PT_LOAD from /proc/kcore.

---
 kexec/arch/i386/crashdump-x86.c | 3 ++-
 kexec/arch/i386/crashdump-x86.h | 3 +++
 2 files changed, 5 insertions(+), 1 deletion(-)
kexec/arch/i386/crashdump-x86.c +2 -1
index cb19e7d..1e6d3a3 100644
@@ -156,7 +156,8 @@ static int get_kernel_vaddr_and_size(struct kexec_info *UNUSED(info),
 
 			/* Look for kernel text mapping header. */
 			if ((saddr >= X86_64__START_KERNEL_map) &&
-			    (eaddr <= X86_64__START_KERNEL_map + X86_64_KERNEL_TEXT_SIZE)) {
+			    (saddr <= X86_64__START_KERNEL_map + X86_64_RANDOMIZE_BASE_MAX_OFFSET) &&
+			    (eaddr - saddr  < X86_64_KERNEL_TEXT_SIZE)) {
 				saddr = _ALIGN_DOWN(saddr, X86_64_KERN_VADDR_ALIGN);
 				elf_info->kern_vaddr_start = saddr;
 				size = eaddr - saddr;
kexec/arch/i386/crashdump-x86.h +0 -0
index e68b626..71a09f8 100644
@@ -15,6 +15,9 @@ int load_crashdump_segments(struct kexec_info *info, char *mod_cmdline,
 #define X86_64_PAGE_OFFSET_PRE_2_6_27	0xffff810000000000ULL
 #define X86_64_PAGE_OFFSET		0xffff880000000000ULL
 
Quoted 8 lines
Quoted 8 lines
kexec/arch/i386/crashdump-x86.h +3 -0
+/* kASLR - Kernel base offset could be randomized up to 1G */
+#define X86_64_RANDOMIZE_BASE_MAX_OFFSET	0x40000000
+
 #define X86_64_MAXMEM        		0x3fffffffffffUL
 
 /* Kernel text size */
Signature · 1 line
1.8.5.3

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

#4 lore raw
Quoted 11 lines · WANG Chao
Quoted 7 lines · Vivek Goyal
Quoted 6 lines · WANG Chao

With kASLR enabled (CONFIG_RANDOMIZED_BASE=y), kernel virtual address base is PAGE_OFFSET plus a randomized offset from 0 to 1G.

Current kexec-tools gets kernel vaddr and size from /proc/kcore. It assumes kernel vaddr start/end is within the range [0,512M). If kaslr enabled, kernel vaddr start/end will stay at [0+offset, 512M+offset).

NACK this patch myself.

Quoted 8 lines

There are several mistakes I made. I misunderstood some concepts. Then I realize this kASLR issue is not trivial to fix.

I think if kexec-tools needs to get kernel text mapping from kcore in kALSR case, the max base offset (CONFIG_RANDOMIZE_MAX_BASE_OFFSET) of the kernel text area must be exposed to userspace some where. We can use that value to determine which area is for kernel text mapping and which is for modules text mapping.

How about looking at /proc/kallsyms and look at address of one of the symbols say _text. And search for the ELF header in kcore which contains _text address and that's ELF header representing kernel text mapping.

That way you don't have to worry about the value of CONFIG_RANDOMIZE_MAX_BASE_OFFSET.

Quoted 12 lines
Quoted 11 lines

Hi Chao,

Documentation/x86/x86_64/mm.txt still says that kernel text mapping area is 512MB.

ffffffff80000000 - ffffffffa0000000 (=512 MB) kernel text mapping, from phys 0

So has that changed now due to kASLR.

Quoted 9 lines

Yes, with kASLR enabled, kernel text mapping is as following

ffffffff80000000 - (ffffffff80000000+CONFIG_RANDOMIZE_BASE_MAX_OFFSET)

That said, if using CONFIG_RANDOMIZE_BASE_MAX_OFFSET=0x40000000 by default, the kernel text mapping is as following:

ffffffff80000000 - ffffffffc0000000

Agreed. help text for CONFIG_RANDOMIZE_BASE_MAX_OFFSET says following.

On 64-bit this is limited by how the kernel fixmap page table is positioned, so this cannot be larger than 1GiB currently. Without RANDOMIZE_BASE, there is a 512MiB to 1.5GiB split between kernel and modules. When RANDOMIZE_BASE_MAX_OFFSET is above 512MiB, the modules area will shrink to compensate, up to the current maximum 1GiB to 1GiB split. The default is 1GiB.

Thanks
Vivek

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

author
#5 lore raw

[CC Simon]

Quoted 21 lines · Vivek Goyal
Quoted 11 lines · WANG Chao
Quoted 7 lines · Vivek Goyal
Quoted 6 lines · WANG Chao

With kASLR enabled (CONFIG_RANDOMIZED_BASE=y), kernel virtual address base is PAGE_OFFSET plus a randomized offset from 0 to 1G.

Current kexec-tools gets kernel vaddr and size from /proc/kcore. It assumes kernel vaddr start/end is within the range [0,512M). If kaslr enabled, kernel vaddr start/end will stay at [0+offset, 512M+offset).

NACK this patch myself.

Quoted 8 lines

There are several mistakes I made. I misunderstood some concepts. Then I realize this kASLR issue is not trivial to fix.

I think if kexec-tools needs to get kernel text mapping from kcore in kALSR case, the max base offset (CONFIG_RANDOMIZE_MAX_BASE_OFFSET) of the kernel text area must be exposed to userspace some where. We can use that value to determine which area is for kernel text mapping and which is for modules text mapping.

Quoted 3 lines

How about looking at /proc/kallsyms and look at address of one of the symbols say _text. And search for the ELF header in kcore which contains _text address and that's ELF header representing kernel text mapping.

Cool. I'd like to go with _stext. _text presents in /proc/kallsyms only when CONFIG_KALLSYMS_ALL=y. What do you think?

Quoted 35 lines

That way you don't have to worry about the value of CONFIG_RANDOMIZE_MAX_BASE_OFFSET.

Quoted 12 lines
Quoted 11 lines

Hi Chao,

Documentation/x86/x86_64/mm.txt still says that kernel text mapping area is 512MB.

ffffffff80000000 - ffffffffa0000000 (=512 MB) kernel text mapping, from phys 0

So has that changed now due to kASLR.

Quoted 9 lines

Yes, with kASLR enabled, kernel text mapping is as following

ffffffff80000000 - (ffffffff80000000+CONFIG_RANDOMIZE_BASE_MAX_OFFSET)

That said, if using CONFIG_RANDOMIZE_BASE_MAX_OFFSET=0x40000000 by default, the kernel text mapping is as following:

ffffffff80000000 - ffffffffc0000000

Agreed. help text for CONFIG_RANDOMIZE_BASE_MAX_OFFSET says following.

On 64-bit this is limited by how the kernel fixmap page table is positioned, so this cannot be larger than 1GiB currently. Without RANDOMIZE_BASE, there is a 512MiB to 1.5GiB split between kernel and modules. When RANDOMIZE_BASE_MAX_OFFSET is above 512MiB, the modules area will shrink to compensate, up to the current maximum 1GiB to 1GiB split. The default is 1GiB.

Quoted 2 lines

Thanks
Vivek

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

#6 lore raw
Quoted 24 lines · WANG Chao

[CC Simon]

Quoted 21 lines · Vivek Goyal
Quoted 19 lines · WANG Chao
Quoted 7 lines · Vivek Goyal
Quoted 6 lines · WANG Chao

With kASLR enabled (CONFIG_RANDOMIZED_BASE=y), kernel virtual address base is PAGE_OFFSET plus a randomized offset from 0 to 1G.

Current kexec-tools gets kernel vaddr and size from /proc/kcore. It assumes kernel vaddr start/end is within the range [0,512M). If kaslr enabled, kernel vaddr start/end will stay at [0+offset, 512M+offset).

NACK this patch myself.

There are several mistakes I made. I misunderstood some concepts. Then I realize this kASLR issue is not trivial to fix.

I think if kexec-tools needs to get kernel text mapping from kcore in kALSR case, the max base offset (CONFIG_RANDOMIZE_MAX_BASE_OFFSET) of the kernel text area must be exposed to userspace some where. We can use that value to determine which area is for kernel text mapping and which is for modules text mapping.

Quoted 6 lines
Quoted 3 lines

How about looking at /proc/kallsyms and look at address of one of the symbols say _text. And search for the ELF header in kcore which contains _text address and that's ELF header representing kernel text mapping.

Cool. I'd like to go with _stext. _text presents in /proc/kallsyms only when CONFIG_KALLSYMS_ALL=y. What do you think?

That's fine. Though looks like _stext will also work only if CONFIG_KALLSYMS=y. So that's not a perfect solution either for those who have built kernel using CONFIG_KALLSYMS=n.

Thanks
Vivek

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec