Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1512928 > unrolled thread
| Started by | Jann Horn <jann@thejh.net> |
|---|---|
| First post | 2016-10-31 22:40 +0100 |
| Last post | 2016-11-04 16:00 +0100 |
| Articles | 4 — 4 participants |
Back to article view | Back to linux.kernel
[PATCH] swapfile: fix memory corruption via malformed swapfile Jann Horn <jann@thejh.net> - 2016-10-31 22:40 +0100
Re: [PATCH] swapfile: fix memory corruption via malformed swapfile Kees Cook <keescook@chromium.org> - 2016-10-31 23:40 +0100
Re: [PATCH] swapfile: fix memory corruption via malformed swapfile Jerome Marchand <jmarchan@redhat.com> - 2016-11-01 10:20 +0100
Re: [PATCH] swapfile: fix memory corruption via malformed swapfile Johannes Weiner <hannes@cmpxchg.org> - 2016-11-04 16:00 +0100
| From | Jann Horn <jann@thejh.net> |
|---|---|
| Date | 2016-10-31 22:40 +0100 |
| Subject | [PATCH] swapfile: fix memory corruption via malformed swapfile |
| Message-ID | <systc-7Sy-21@gated-at.bofh.it> |
When root activates a swap partition whose header has the wrong endianness,
nr_badpages elements of badpages are swabbed before nr_badpages has been
checked, leading to a buffer overrun of up to 8GB.
This normally is not a security issue because it can only be exploited by
root (more specifically, a process with CAP_SYS_ADMIN or the ability to
modify a swap file/partition), and such a process can already e.g. modify
swapped-out memory of any other userspace process on the system.
Testcase for reproducing the bug (must be run as root, should crash your
kernel):
=================
#include <stdlib.h>
#include <unistd.h>
#include <sys/swap.h>
#include <limits.h>
#include <err.h>
#include <string.h>
#include <stdio.h>
#define PAGE_SIZE 4096
#define __u32 unsigned int
// from include/linux/swap.h
union swap_header {
struct {
char reserved[PAGE_SIZE - 10];
char magic[10]; /* SWAP-SPACE or SWAPSPACE2 */
} magic;
struct {
char bootbits[1024]; /* Space for disklabel etc. */
__u32 version;
__u32 last_page;
__u32 nr_badpages;
unsigned char sws_uuid[16];
unsigned char sws_volume[16];
__u32 padding[117];
__u32 badpages[1];
} info;
};
int main(void) {
char file[] = "/tmp/swapfile.XXXXXX";
int file_fd = mkstemp(file);
if (file_fd == -1)
err(1, "mkstemp");
if (ftruncate(file_fd, PAGE_SIZE))
err(1, "ftruncate");
union swap_header swap_header = {
.info = {
.version = __builtin_bswap32(1),
.nr_badpages = __builtin_bswap32(INT_MAX)
}
};
memcpy(swap_header.magic.magic, "SWAPSPACE2", 10);
if (write(file_fd, &swap_header, sizeof(swap_header)) !=
sizeof(swap_header))
err(1, "write");
// not because the attack needs it, just in case you forgot to
// sync yourself before crashing your machine
sync();
// now die
if (swapon(file, 0))
err(1, "swapon");
puts("huh, we survived");
if (swapoff(file))
err(1, "swapoff");
unlink(file);
}
=================
Cc: stable@vger.kernel.org
Signed-off-by: Jann Horn <jann@thejh.net>
---
mm/swapfile.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/mm/swapfile.c b/mm/swapfile.c
index 2210de290b54..f30438970cd1 100644
--- a/mm/swapfile.c
+++ b/mm/swapfile.c
@@ -2224,6 +2224,8 @@ static unsigned long read_swap_header(struct swap_info_struct *p,
swab32s(&swap_header->info.version);
swab32s(&swap_header->info.last_page);
swab32s(&swap_header->info.nr_badpages);
+ if (swap_header->info.nr_badpages > MAX_SWAP_BADPAGES)
+ return 0;
for (i = 0; i < swap_header->info.nr_badpages; i++)
swab32s(&swap_header->info.badpages[i]);
}
--
2.1.4
[toc] | [next] | [standalone]
| From | Kees Cook <keescook@chromium.org> |
|---|---|
| Date | 2016-10-31 23:40 +0100 |
| Message-ID | <sytpf-8ws-1@gated-at.bofh.it> |
| In reply to | #1512928 |
On Mon, Oct 31, 2016 at 2:32 PM, Jann Horn <jann@thejh.net> wrote:
> When root activates a swap partition whose header has the wrong endianness,
> nr_badpages elements of badpages are swabbed before nr_badpages has been
> checked, leading to a buffer overrun of up to 8GB.
>
> This normally is not a security issue because it can only be exploited by
> root (more specifically, a process with CAP_SYS_ADMIN or the ability to
> modify a swap file/partition), and such a process can already e.g. modify
> swapped-out memory of any other userspace process on the system.
>
> Testcase for reproducing the bug (must be run as root, should crash your
> kernel):
> =================
> #include <stdlib.h>
> #include <unistd.h>
> #include <sys/swap.h>
> #include <limits.h>
> #include <err.h>
> #include <string.h>
> #include <stdio.h>
>
> #define PAGE_SIZE 4096
> #define __u32 unsigned int
>
>
> // from include/linux/swap.h
> union swap_header {
> struct {
> char reserved[PAGE_SIZE - 10];
> char magic[10]; /* SWAP-SPACE or SWAPSPACE2 */
> } magic;
> struct {
> char bootbits[1024]; /* Space for disklabel etc. */
> __u32 version;
> __u32 last_page;
> __u32 nr_badpages;
> unsigned char sws_uuid[16];
> unsigned char sws_volume[16];
> __u32 padding[117];
> __u32 badpages[1];
> } info;
> };
>
> int main(void) {
> char file[] = "/tmp/swapfile.XXXXXX";
> int file_fd = mkstemp(file);
> if (file_fd == -1)
> err(1, "mkstemp");
> if (ftruncate(file_fd, PAGE_SIZE))
> err(1, "ftruncate");
> union swap_header swap_header = {
> .info = {
> .version = __builtin_bswap32(1),
> .nr_badpages = __builtin_bswap32(INT_MAX)
> }
> };
> memcpy(swap_header.magic.magic, "SWAPSPACE2", 10);
> if (write(file_fd, &swap_header, sizeof(swap_header)) !=
> sizeof(swap_header))
> err(1, "write");
>
> // not because the attack needs it, just in case you forgot to
> // sync yourself before crashing your machine
> sync();
>
> // now die
> if (swapon(file, 0))
> err(1, "swapon");
> puts("huh, we survived");
> if (swapoff(file))
> err(1, "swapoff");
> unlink(file);
> }
> =================
>
> Cc: stable@vger.kernel.org
> Signed-off-by: Jann Horn <jann@thejh.net>
> ---
> mm/swapfile.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/mm/swapfile.c b/mm/swapfile.c
> index 2210de290b54..f30438970cd1 100644
> --- a/mm/swapfile.c
> +++ b/mm/swapfile.c
> @@ -2224,6 +2224,8 @@ static unsigned long read_swap_header(struct swap_info_struct *p,
> swab32s(&swap_header->info.version);
> swab32s(&swap_header->info.last_page);
> swab32s(&swap_header->info.nr_badpages);
> + if (swap_header->info.nr_badpages > MAX_SWAP_BADPAGES)
> + return 0;
> for (i = 0; i < swap_header->info.nr_badpages; i++)
> swab32s(&swap_header->info.badpages[i]);
> }
> --
> 2.1.4
>
Eww. Nice find. :) At least it's only init_ns CAP_SYS_ADMIN. :P
Acked-by: Kees Cook <keescook@chromium.org>
-Kees
--
Kees Cook
Nexus Security
[toc] | [prev] | [next] | [standalone]
| From | Jerome Marchand <jmarchan@redhat.com> |
|---|---|
| Date | 2016-11-01 10:20 +0100 |
| Message-ID | <syDoC-6JE-7@gated-at.bofh.it> |
| In reply to | #1512928 |
[Multipart message — attachments visible in raw view] — view raw
On 10/31/2016 10:32 PM, Jann Horn wrote:
> When root activates a swap partition whose header has the wrong endianness,
> nr_badpages elements of badpages are swabbed before nr_badpages has been
> checked, leading to a buffer overrun of up to 8GB.
>
> This normally is not a security issue because it can only be exploited by
> root (more specifically, a process with CAP_SYS_ADMIN or the ability to
> modify a swap file/partition), and such a process can already e.g. modify
> swapped-out memory of any other userspace process on the system.
>
> Testcase for reproducing the bug (must be run as root, should crash your
> kernel):
> =================
> #include <stdlib.h>
> #include <unistd.h>
> #include <sys/swap.h>
> #include <limits.h>
> #include <err.h>
> #include <string.h>
> #include <stdio.h>
>
> #define PAGE_SIZE 4096
> #define __u32 unsigned int
>
>
> // from include/linux/swap.h
> union swap_header {
> struct {
> char reserved[PAGE_SIZE - 10];
> char magic[10]; /* SWAP-SPACE or SWAPSPACE2 */
> } magic;
> struct {
> char bootbits[1024]; /* Space for disklabel etc. */
> __u32 version;
> __u32 last_page;
> __u32 nr_badpages;
> unsigned char sws_uuid[16];
> unsigned char sws_volume[16];
> __u32 padding[117];
> __u32 badpages[1];
> } info;
> };
>
> int main(void) {
> char file[] = "/tmp/swapfile.XXXXXX";
> int file_fd = mkstemp(file);
> if (file_fd == -1)
> err(1, "mkstemp");
> if (ftruncate(file_fd, PAGE_SIZE))
> err(1, "ftruncate");
> union swap_header swap_header = {
> .info = {
> .version = __builtin_bswap32(1),
> .nr_badpages = __builtin_bswap32(INT_MAX)
> }
> };
> memcpy(swap_header.magic.magic, "SWAPSPACE2", 10);
> if (write(file_fd, &swap_header, sizeof(swap_header)) !=
> sizeof(swap_header))
> err(1, "write");
>
> // not because the attack needs it, just in case you forgot to
> // sync yourself before crashing your machine
> sync();
>
> // now die
> if (swapon(file, 0))
> err(1, "swapon");
> puts("huh, we survived");
> if (swapoff(file))
> err(1, "swapoff");
> unlink(file);
> }
> =================
>
> Cc: stable@vger.kernel.org
> Signed-off-by: Jann Horn <jann@thejh.net>
> ---
> mm/swapfile.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/mm/swapfile.c b/mm/swapfile.c
> index 2210de290b54..f30438970cd1 100644
> --- a/mm/swapfile.c
> +++ b/mm/swapfile.c
> @@ -2224,6 +2224,8 @@ static unsigned long read_swap_header(struct swap_info_struct *p,
> swab32s(&swap_header->info.version);
> swab32s(&swap_header->info.last_page);
> swab32s(&swap_header->info.nr_badpages);
> + if (swap_header->info.nr_badpages > MAX_SWAP_BADPAGES)
> + return 0;
> for (i = 0; i < swap_header->info.nr_badpages; i++)
> swab32s(&swap_header->info.badpages[i]);
> }
>
Nice catch!
Acked-by: Jerome Marchand <jmarchan@redhat.com>
[toc] | [prev] | [next] | [standalone]
| From | Johannes Weiner <hannes@cmpxchg.org> |
|---|---|
| Date | 2016-11-04 16:00 +0100 |
| Message-ID | <szO8i-37p-25@gated-at.bofh.it> |
| In reply to | #1512928 |
On Mon, Oct 31, 2016 at 10:32:13PM +0100, Jann Horn wrote: > When root activates a swap partition whose header has the wrong endianness, > nr_badpages elements of badpages are swabbed before nr_badpages has been > checked, leading to a buffer overrun of up to 8GB. > > This normally is not a security issue because it can only be exploited by > root (more specifically, a process with CAP_SYS_ADMIN or the ability to > modify a swap file/partition), and such a process can already e.g. modify > swapped-out memory of any other userspace process on the system. > > Testcase for reproducing the bug (must be run as root, should crash your > kernel): [...] > Cc: stable@vger.kernel.org > Signed-off-by: Jann Horn <jann@thejh.net> Acked-by: Johannes Weiner <hannes@cmpxchg.org>
[toc] | [prev] | [standalone]
Back to top | Article view | linux.kernel
csiph-web