[PATCH v4 0/6] net/eth: Fix stack-buffer-overflow in _eth_get_rss_ex_dst_addr()

[PATCH v4 0/6] net/eth: Fix stack-buffer-overflow in _eth_get_rss_ex_dst_addr()

[Philippe Mathieu-Daudé]

I had a look at the patch from Miroslav trying to silence a
compiler warning which in fact is a nasty bug. Here is a fix.
https://www.mail-archive.com/qemu-devel@nongnu.org/msg772735.html

Since v3:
- reworked in multiple trivial patches (Stefano)
- reset R-b/A-b tags

Philippe Mathieu-Daudé (6):
  net/eth: Simplify _eth_get_rss_ex_dst_addr()
  net/eth: Better describe _eth_get_rss_ex_dst_addr's offset argument
  net/eth: Initialize input_size variable earlier
  net/eth: Check rt_hdr size before casting to ip6_ext_hdr
  net/eth: Remove now useless size check
  net/eth: Return earlier in _eth_get_rss_ex_dst_addr()

 net/eth.c                      | 42 +++++++++++++--------------
 tests/qtest/fuzz-e1000e-test.c | 53 ++++++++++++++++++++++++++++++++++
 MAINTAINERS                    |  1 +
 tests/qtest/meson.build        |  1 +
 4 files changed, 75 insertions(+), 22 deletions(-)
 create mode 100644 tests/qtest/fuzz-e1000e-test.c

-- 
2.26.2

[PATCH v4 1/6] net/eth: Simplify _eth_get_rss_ex_dst_addr()

[Philippe Mathieu-Daudé]

The length field is already contained in the ip6_ext_hdr structure.
Check it direcly in eth_parse_ipv6_hdr() before calling
_eth_get_rss_ex_dst_addr(), which gets a bit simplified.

Reviewed-by: Miroslav Rezanina <mrezanin@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 net/eth.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

diff --git a/net/eth.c b/net/eth.c
index 1e0821c5f81..7d4dd48c1ff 100644
--- a/net/eth.c
+++ b/net/eth.c
@@ -407,9 +407,7 @@ _eth_get_rss_ex_dst_addr(const struct iovec *pkt, int pkt_frags,
 {
     struct ip6_ext_hdr_routing *rthdr = (struct ip6_ext_hdr_routing *) ext_hdr;
 
-    if ((rthdr->rtype == 2) &&
-        (rthdr->len == sizeof(struct in6_address) / 8) &&
-        (rthdr->segleft == 1)) {
+    if ((rthdr->rtype == 2) && (rthdr->segleft == 1)) {
 
         size_t input_size = iov_size(pkt, pkt_frags);
         size_t bytes_read;
@@ -528,10 +526,12 @@ bool eth_parse_ipv6_hdr(const struct iovec *pkt, int pkt_frags,
         }
 
         if (curr_ext_hdr_type == IP6_ROUTING) {
-            info->rss_ex_dst_valid =
-                _eth_get_rss_ex_dst_addr(pkt, pkt_frags,
-                                         ip6hdr_off + info->full_hdr_len,
-                                         &ext_hdr, &info->rss_ex_dst);
+            if (ext_hdr.ip6r_len == sizeof(struct in6_address) / 8) {
+                info->rss_ex_dst_valid =
+                    _eth_get_rss_ex_dst_addr(pkt, pkt_frags,
+                                             ip6hdr_off + info->full_hdr_len,
+                                             &ext_hdr, &info->rss_ex_dst);
+            }
         } else if (curr_ext_hdr_type == IP6_DESTINATON) {
             info->rss_ex_src_valid =
                 _eth_get_rss_ex_src_addr(pkt, pkt_frags,
-- 
2.26.2

[PATCH v4 2/6] net/eth: Better describe _eth_get_rss_ex_dst_addr's offset argument

[Philippe Mathieu-Daudé]

The 'offset' argument represents the offset to the ip6_ext_hdr
header, rename it as 'ext_hdr_offset'.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 net/eth.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/net/eth.c b/net/eth.c
index 7d4dd48c1ff..6a5a1d04e0b 100644
--- a/net/eth.c
+++ b/net/eth.c
@@ -401,7 +401,7 @@ eth_is_ip6_extension_header_type(uint8_t hdr_type)
 
 static bool
 _eth_get_rss_ex_dst_addr(const struct iovec *pkt, int pkt_frags,
-                        size_t rthdr_offset,
+                        size_t ext_hdr_offset,
                         struct ip6_ext_hdr *ext_hdr,
                         struct in6_address *dst_addr)
 {
@@ -412,12 +412,12 @@ _eth_get_rss_ex_dst_addr(const struct iovec *pkt, int pkt_frags,
         size_t input_size = iov_size(pkt, pkt_frags);
         size_t bytes_read;
 
-        if (input_size < rthdr_offset + sizeof(*ext_hdr)) {
+        if (input_size < ext_hdr_offset + sizeof(*ext_hdr)) {
             return false;
         }
 
         bytes_read = iov_to_buf(pkt, pkt_frags,
-                                rthdr_offset + sizeof(*ext_hdr),
+                                ext_hdr_offset + sizeof(*ext_hdr),
                                 dst_addr, sizeof(*dst_addr));
 
         return bytes_read == sizeof(*dst_addr);
-- 
2.26.2

[PATCH v4 3/6] net/eth: Initialize input_size variable earlier

[Philippe Mathieu-Daudé]

Initialize the 'input_size' variable earlier to make the
next commit easier to review.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 net/eth.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/net/eth.c b/net/eth.c
index 6a5a1d04e0b..77af2b673bb 100644
--- a/net/eth.c
+++ b/net/eth.c
@@ -405,11 +405,10 @@ _eth_get_rss_ex_dst_addr(const struct iovec *pkt, int pkt_frags,
                         struct ip6_ext_hdr *ext_hdr,
                         struct in6_address *dst_addr)
 {
+    size_t input_size = iov_size(pkt, pkt_frags);
     struct ip6_ext_hdr_routing *rthdr = (struct ip6_ext_hdr_routing *) ext_hdr;
 
     if ((rthdr->rtype == 2) && (rthdr->segleft == 1)) {
-
-        size_t input_size = iov_size(pkt, pkt_frags);
         size_t bytes_read;
 
         if (input_size < ext_hdr_offset + sizeof(*ext_hdr)) {
-- 
2.26.2

[PATCH v4 4/6] net/eth: Check rt_hdr size before casting to ip6_ext_hdr

[Philippe Mathieu-Daudé]

Do not cast our ip6_ext_hdr pointer to ip6_ext_hdr_routing if there
isn't enough data in the buffer for a such structure.

This fix a 2 bytes buffer overrun in eth_parse_ipv6_hdr() reported
by QEMU fuzzer:

  $ cat << EOF | ./qemu-system-i386 -M pc-q35-5.0 \
    -accel qtest -monitor none \
    -serial none -nographic -qtest stdio
  outl 0xcf8 0x80001010
  outl 0xcfc 0xe1020000
  outl 0xcf8 0x80001004
  outw 0xcfc 0x7
  write 0x25 0x1 0x86
  write 0x26 0x1 0xdd
  write 0x4f 0x1 0x2b
  write 0xe1020030 0x4 0x190002e1
  write 0xe102003a 0x2 0x0807
  write 0xe1020048 0x4 0x12077cdd
  write 0xe1020400 0x4 0xba077cdd
  write 0xe1020420 0x4 0x190002e1
  write 0xe1020428 0x4 0x3509d807
  write 0xe1020438 0x1 0xe2
  EOF
  =================================================================
  ==2859770==ERROR: AddressSanitizer: stack-buffer-overflow on address 0x7ffdef904902 at pc 0x561ceefa78de bp 0x7ffdef904820 sp 0x7ffdef904818
  READ of size 1 at 0x7ffdef904902 thread T0
      #0 0x561ceefa78dd in _eth_get_rss_ex_dst_addr net/eth.c:410:17
      #1 0x561ceefa41fb in eth_parse_ipv6_hdr net/eth.c:532:17
      #2 0x561cef7de639 in net_tx_pkt_parse_headers hw/net/net_tx_pkt.c:228:14
      #3 0x561cef7dbef4 in net_tx_pkt_parse hw/net/net_tx_pkt.c:273:9
      #4 0x561ceec29f22 in e1000e_process_tx_desc hw/net/e1000e_core.c:730:29
      #5 0x561ceec28eac in e1000e_start_xmit hw/net/e1000e_core.c:927:9
      #6 0x561ceec1baab in e1000e_set_tdt hw/net/e1000e_core.c:2444:9
      #7 0x561ceebf300e in e1000e_core_write hw/net/e1000e_core.c:3256:9
      #8 0x561cef3cd4cd in e1000e_mmio_write hw/net/e1000e.c:110:5

  Address 0x7ffdef904902 is located in stack of thread T0 at offset 34 in frame
      #0 0x561ceefa320f in eth_parse_ipv6_hdr net/eth.c:486

    This frame has 1 object(s):
      [32, 34) 'ext_hdr' (line 487) <== Memory access at offset 34 overflows this variable
  HINT: this may be a false positive if your program uses some custom stack unwind mechanism, swapcontext or vfork
        (longjmp and C++ exceptions *are* supported)
  SUMMARY: AddressSanitizer: stack-buffer-overflow net/eth.c:410:17 in _eth_get_rss_ex_dst_addr
  Shadow bytes around the buggy address:
    0x10003df188d0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
    0x10003df188e0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
    0x10003df188f0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
    0x10003df18900: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
    0x10003df18910: 00 00 00 00 00 00 00 00 00 00 00 00 f1 f1 f1 f1
  =>0x10003df18920:[02]f3 f3 f3 00 00 00 00 00 00 00 00 00 00 00 00
    0x10003df18930: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
    0x10003df18940: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
    0x10003df18950: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
    0x10003df18960: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
    0x10003df18970: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
  Shadow byte legend (one shadow byte represents 8 application bytes):
    Addressable:           00
    Partially addressable: 01 02 03 04 05 06 07
    Stack left redzone:      f1
    Stack right redzone:     f3
  ==2859770==ABORTING

Add the corresponding qtest case with the fuzzer reproducer.

FWIW GCC 11 similarly reported:

  net/eth.c: In function 'eth_parse_ipv6_hdr':
  net/eth.c:410:15: error: array subscript 'struct ip6_ext_hdr_routing[0]' is partly outside array bounds of 'struct ip6_ext_hdr[1]' [-Werror=array-bounds]
    410 |     if ((rthdr->rtype == 2) && (rthdr->segleft == 1)) {
        |          ~~~~~^~~~~~~
  net/eth.c:485:24: note: while referencing 'ext_hdr'
    485 |     struct ip6_ext_hdr ext_hdr;
        |                        ^~~~~~~
  net/eth.c:410:38: error: array subscript 'struct ip6_ext_hdr_routing[0]' is partly outside array bounds of 'struct ip6_ext_hdr[1]' [-Werror=array-bounds]
    410 |     if ((rthdr->rtype == 2) && (rthdr->segleft == 1)) {
        |                                 ~~~~~^~~~~~~~~
  net/eth.c:485:24: note: while referencing 'ext_hdr'
    485 |     struct ip6_ext_hdr ext_hdr;
        |                        ^~~~~~~

Cc: qemu-stable@nongnu.org
Buglink: https://bugs.launchpad.net/qemu/+bug/1879531
Reported-by: Alexander Bulekov <alxndr@bu.edu>
Reported-by: Miroslav Rezanina <mrezanin@redhat.com>
Fixes: eb700029c78 ("net_pkt: Extend packet abstraction as required by e1000e functionality")
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 net/eth.c                      |  7 ++++-
 tests/qtest/fuzz-e1000e-test.c | 53 ++++++++++++++++++++++++++++++++++
 MAINTAINERS                    |  1 +
 tests/qtest/meson.build        |  1 +
 4 files changed, 61 insertions(+), 1 deletion(-)
 create mode 100644 tests/qtest/fuzz-e1000e-test.c

diff --git a/net/eth.c b/net/eth.c
index 77af2b673bb..f0c8dfe8df7 100644
--- a/net/eth.c
+++ b/net/eth.c
@@ -406,7 +406,12 @@ _eth_get_rss_ex_dst_addr(const struct iovec *pkt, int pkt_frags,
                         struct in6_address *dst_addr)
 {
     size_t input_size = iov_size(pkt, pkt_frags);
-    struct ip6_ext_hdr_routing *rthdr = (struct ip6_ext_hdr_routing *) ext_hdr;
+    struct ip6_ext_hdr_routing *rthdr;
+
+    if (input_size < ext_hdr_offset + sizeof(*rthdr)) {
+        return false;
+    }
+    rthdr = (struct ip6_ext_hdr_routing *) ext_hdr;
 
     if ((rthdr->rtype == 2) && (rthdr->segleft == 1)) {
         size_t bytes_read;
diff --git a/tests/qtest/fuzz-e1000e-test.c b/tests/qtest/fuzz-e1000e-test.c
new file mode 100644
index 00000000000..66229e60964
--- /dev/null
+++ b/tests/qtest/fuzz-e1000e-test.c
@@ -0,0 +1,53 @@
+/*
+ * QTest testcase for e1000e device generated by fuzzer
+ *
+ * Copyright (c) 2021 Red Hat, Inc.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#include "qemu/osdep.h"
+
+#include "libqos/libqtest.h"
+
+/*
+ * https://bugs.launchpad.net/qemu/+bug/1879531
+ */
+static void test_lp1879531_eth_get_rss_ex_dst_addr(void)
+{
+    QTestState *s;
+
+    s = qtest_init("-nographic -monitor none -serial none -M pc-q35-5.0");
+
+    qtest_outl(s, 0xcf8, 0x80001010);
+    qtest_outl(s, 0xcfc, 0xe1020000);
+    qtest_outl(s, 0xcf8, 0x80001004);
+    qtest_outw(s, 0xcfc, 0x7);
+    qtest_writeb(s, 0x25, 0x86);
+    qtest_writeb(s, 0x26, 0xdd);
+    qtest_writeb(s, 0x4f, 0x2b);
+
+    qtest_writel(s, 0xe1020030, 0x190002e1);
+    qtest_writew(s, 0xe102003a, 0x0807);
+    qtest_writel(s, 0xe1020048, 0x12077cdd);
+    qtest_writel(s, 0xe1020400, 0xba077cdd);
+    qtest_writel(s, 0xe1020420, 0x190002e1);
+    qtest_writel(s, 0xe1020428, 0x3509d807);
+    qtest_writeb(s, 0xe1020438, 0xe2);
+    qtest_writeb(s, 0x4f, 0x2b);
+    qtest_quit(s);
+}
+
+int main(int argc, char **argv)
+{
+    const char *arch = qtest_get_arch();
+
+    g_test_init(&argc, &argv, NULL);
+
+    if (strcmp(arch, "i386") == 0 || strcmp(arch, "x86_64") == 0) {
+        qtest_add_func("fuzz/test_lp1879531_eth_get_rss_ex_dst_addr",
+                       test_lp1879531_eth_get_rss_ex_dst_addr);
+    }
+
+    return g_test_run();
+}
diff --git a/MAINTAINERS b/MAINTAINERS
index f22d83c1782..4eb5784ff83 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -2003,6 +2003,7 @@ e1000e
 M: Dmitry Fleytman <dmitry.fleytman@gmail.com>
 S: Maintained
 F: hw/net/e1000e*
+F: tests/qtest/fuzz-e1000e-test.c
 
 eepro100
 M: Stefan Weil <sw@weilnetz.de>
diff --git a/tests/qtest/meson.build b/tests/qtest/meson.build
index 58efc46144e..7997d895449 100644
--- a/tests/qtest/meson.build
+++ b/tests/qtest/meson.build
@@ -60,6 +60,7 @@
   (config_all_devices.has_key('CONFIG_TPM_TIS_ISA') ? ['tpm-tis-test'] : []) +              \
   (config_all_devices.has_key('CONFIG_TPM_TIS_ISA') ? ['tpm-tis-swtpm-test'] : []) +        \
   (config_all_devices.has_key('CONFIG_RTL8139_PCI') ? ['rtl8139-test'] : []) +              \
+  (config_all_devices.has_key('CONFIG_E1000E_PCI_EXPRESS') ? ['fuzz-e1000e-test'] : []) +   \
   qtests_pci +                                                                              \
   ['fdc-test',
    'ide-test',
-- 
2.26.2

[PATCH v4 5/6] net/eth: Remove now useless size check

[Philippe Mathieu-Daudé]

In the previous commit we checked there is enough room
for a ip6_ext_hdr_routing in the buffer.

The ip6_ext_hdr_routing structure starts with a ip6_ext_hdr,
so we already checked that input_size is big enough for a
ip6_ext_hdr :) Remove that now useless check.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 net/eth.c | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/net/eth.c b/net/eth.c
index f0c8dfe8df7..e984edcfb0b 100644
--- a/net/eth.c
+++ b/net/eth.c
@@ -416,10 +416,6 @@ _eth_get_rss_ex_dst_addr(const struct iovec *pkt, int pkt_frags,
     if ((rthdr->rtype == 2) && (rthdr->segleft == 1)) {
         size_t bytes_read;
 
-        if (input_size < ext_hdr_offset + sizeof(*ext_hdr)) {
-            return false;
-        }
-
         bytes_read = iov_to_buf(pkt, pkt_frags,
                                 ext_hdr_offset + sizeof(*ext_hdr),
                                 dst_addr, sizeof(*dst_addr));
-- 
2.26.2

[PATCH v4 6/6] net/eth: Return earlier in _eth_get_rss_ex_dst_addr()

[Philippe Mathieu-Daudé]

Slightly simplify _eth_get_rss_ex_dst_addr() by returning earlier.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 net/eth.c | 16 +++++++---------
 1 file changed, 7 insertions(+), 9 deletions(-)

diff --git a/net/eth.c b/net/eth.c
index e984edcfb0b..b44439d31c5 100644
--- a/net/eth.c
+++ b/net/eth.c
@@ -407,23 +407,21 @@ _eth_get_rss_ex_dst_addr(const struct iovec *pkt, int pkt_frags,
 {
     size_t input_size = iov_size(pkt, pkt_frags);
     struct ip6_ext_hdr_routing *rthdr;
+    size_t bytes_read;
 
     if (input_size < ext_hdr_offset + sizeof(*rthdr)) {
         return false;
     }
     rthdr = (struct ip6_ext_hdr_routing *) ext_hdr;
 
-    if ((rthdr->rtype == 2) && (rthdr->segleft == 1)) {
-        size_t bytes_read;
-
-        bytes_read = iov_to_buf(pkt, pkt_frags,
-                                ext_hdr_offset + sizeof(*ext_hdr),
-                                dst_addr, sizeof(*dst_addr));
-
-        return bytes_read == sizeof(*dst_addr);
+    if ((rthdr->rtype != 2) || (rthdr->segleft != 1)) {
+        return false;
     }
 
-    return false;
+    bytes_read = iov_to_buf(pkt, pkt_frags, ext_hdr_offset + sizeof(*ext_hdr),
+                            dst_addr, sizeof(*dst_addr));
+
+    return bytes_read == sizeof(*dst_addr);
 }
 
 static bool
-- 
2.26.2

Re: [PATCH v4 2/6] net/eth: Better describe _eth_get_rss_ex_dst_addr's offset argument

[Stefano Garzarella]

On Tue, Mar 09, 2021 at 07:27:05PM +0100, Philippe Mathieu-Daudé wrote:
>The 'offset' argument represents the offset to the ip6_ext_hdr
>header, rename it as 'ext_hdr_offset'.
>
>Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
>---
> net/eth.c | 6 +++---
> 1 file changed, 3 insertions(+), 3 deletions(-)

Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>

Re: [PATCH v4 3/6] net/eth: Initialize input_size variable earlier

[Stefano Garzarella]

On Tue, Mar 09, 2021 at 07:27:06PM +0100, Philippe Mathieu-Daudé wrote:
>Initialize the 'input_size' variable earlier to make the
>next commit easier to review.
>
>Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
>---
> net/eth.c | 3 +--
> 1 file changed, 1 insertion(+), 2 deletions(-)

Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>

Re: [PATCH v4 4/6] net/eth: Check rt_hdr size before casting to ip6_ext_hdr

[Stefano Garzarella]

On Tue, Mar 09, 2021 at 07:27:07PM +0100, Philippe Mathieu-Daudé wrote:
>Do not cast our ip6_ext_hdr pointer to ip6_ext_hdr_routing if there
>isn't enough data in the buffer for a such structure.
>
>This fix a 2 bytes buffer overrun in eth_parse_ipv6_hdr() reported
>by QEMU fuzzer:
>
>  $ cat << EOF | ./qemu-system-i386 -M pc-q35-5.0 \
>    -accel qtest -monitor none \
>    -serial none -nographic -qtest stdio
>  outl 0xcf8 0x80001010
>  outl 0xcfc 0xe1020000
>  outl 0xcf8 0x80001004
>  outw 0xcfc 0x7
>  write 0x25 0x1 0x86
>  write 0x26 0x1 0xdd
>  write 0x4f 0x1 0x2b
>  write 0xe1020030 0x4 0x190002e1
>  write 0xe102003a 0x2 0x0807
>  write 0xe1020048 0x4 0x12077cdd
>  write 0xe1020400 0x4 0xba077cdd
>  write 0xe1020420 0x4 0x190002e1
>  write 0xe1020428 0x4 0x3509d807
>  write 0xe1020438 0x1 0xe2
>  EOF
>  =================================================================
>  ==2859770==ERROR: AddressSanitizer: stack-buffer-overflow on address 0x7ffdef904902 at pc 0x561ceefa78de bp 0x7ffdef904820 sp 0x7ffdef904818
>  READ of size 1 at 0x7ffdef904902 thread T0
>      #0 0x561ceefa78dd in _eth_get_rss_ex_dst_addr net/eth.c:410:17
>      #1 0x561ceefa41fb in eth_parse_ipv6_hdr net/eth.c:532:17
>      #2 0x561cef7de639 in net_tx_pkt_parse_headers hw/net/net_tx_pkt.c:228:14
>      #3 0x561cef7dbef4 in net_tx_pkt_parse hw/net/net_tx_pkt.c:273:9
>      #4 0x561ceec29f22 in e1000e_process_tx_desc hw/net/e1000e_core.c:730:29
>      #5 0x561ceec28eac in e1000e_start_xmit hw/net/e1000e_core.c:927:9
>      #6 0x561ceec1baab in e1000e_set_tdt hw/net/e1000e_core.c:2444:9
>      #7 0x561ceebf300e in e1000e_core_write hw/net/e1000e_core.c:3256:9
>      #8 0x561cef3cd4cd in e1000e_mmio_write hw/net/e1000e.c:110:5
>
>  Address 0x7ffdef904902 is located in stack of thread T0 at offset 34 in frame
>      #0 0x561ceefa320f in eth_parse_ipv6_hdr net/eth.c:486
>
>    This frame has 1 object(s):
>      [32, 34) 'ext_hdr' (line 487) <== Memory access at offset 34 overflows this variable
>  HINT: this may be a false positive if your program uses some custom stack unwind mechanism, swapcontext or vfork
>        (longjmp and C++ exceptions *are* supported)
>  SUMMARY: AddressSanitizer: stack-buffer-overflow net/eth.c:410:17 in _eth_get_rss_ex_dst_addr
>  Shadow bytes around the buggy address:
>    0x10003df188d0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>    0x10003df188e0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>    0x10003df188f0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>    0x10003df18900: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>    0x10003df18910: 00 00 00 00 00 00 00 00 00 00 00 00 f1 f1 f1 f1
>  =>0x10003df18920:[02]f3 f3 f3 00 00 00 00 00 00 00 00 00 00 00 00
>    0x10003df18930: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>    0x10003df18940: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>    0x10003df18950: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>    0x10003df18960: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>    0x10003df18970: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>  Shadow byte legend (one shadow byte represents 8 application bytes):
>    Addressable:           00
>    Partially addressable: 01 02 03 04 05 06 07
>    Stack left redzone:      f1
>    Stack right redzone:     f3
>  ==2859770==ABORTING
>
>Add the corresponding qtest case with the fuzzer reproducer.
>
>FWIW GCC 11 similarly reported:
>
>  net/eth.c: In function 'eth_parse_ipv6_hdr':
>  net/eth.c:410:15: error: array subscript 'struct ip6_ext_hdr_routing[0]' is partly outside array bounds of 'struct ip6_ext_hdr[1]' [-Werror=array-bounds]
>    410 |     if ((rthdr->rtype == 2) && (rthdr->segleft == 1)) {
>        |          ~~~~~^~~~~~~
>  net/eth.c:485:24: note: while referencing 'ext_hdr'
>    485 |     struct ip6_ext_hdr ext_hdr;
>        |                        ^~~~~~~
>  net/eth.c:410:38: error: array subscript 'struct ip6_ext_hdr_routing[0]' is partly outside array bounds of 'struct ip6_ext_hdr[1]' [-Werror=array-bounds]
>    410 |     if ((rthdr->rtype == 2) && (rthdr->segleft == 1)) {
>        |                                 ~~~~~^~~~~~~~~
>  net/eth.c:485:24: note: while referencing 'ext_hdr'
>    485 |     struct ip6_ext_hdr ext_hdr;
>        |                        ^~~~~~~
>
>Cc: qemu-stable@nongnu.org
>Buglink: https://bugs.launchpad.net/qemu/+bug/1879531
>Reported-by: Alexander Bulekov <alxndr@bu.edu>
>Reported-by: Miroslav Rezanina <mrezanin@redhat.com>
>Fixes: eb700029c78 ("net_pkt: Extend packet abstraction as required by e1000e functionality")
>Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
>---
> net/eth.c                      |  7 ++++-
> tests/qtest/fuzz-e1000e-test.c | 53 ++++++++++++++++++++++++++++++++++
> MAINTAINERS                    |  1 +
> tests/qtest/meson.build        |  1 +
> 4 files changed, 61 insertions(+), 1 deletion(-)
> create mode 100644 tests/qtest/fuzz-e1000e-test.c
>
>diff --git a/net/eth.c b/net/eth.c
>index 77af2b673bb..f0c8dfe8df7 100644
>--- a/net/eth.c
>+++ b/net/eth.c
>@@ -406,7 +406,12 @@ _eth_get_rss_ex_dst_addr(const struct iovec *pkt, int pkt_frags,
>                         struct in6_address *dst_addr)
> {
>     size_t input_size = iov_size(pkt, pkt_frags);
>-    struct ip6_ext_hdr_routing *rthdr = (struct ip6_ext_hdr_routing *) ext_hdr;
>+    struct ip6_ext_hdr_routing *rthdr;
>+
>+    if (input_size < ext_hdr_offset + sizeof(*rthdr)) {
>+        return false;
>+    }
>+    rthdr = (struct ip6_ext_hdr_routing *) ext_hdr;

Also if we check 'input_size', I think this cast keeps having the 2 
bytes buffer overrun issue since 'ext_hdr' contains only the first 2 
bytes.
I think we can remove the 'input_size' check and we should fix in this 
way:

     struct ip6_ext_hdr_routing rthdr;

     bytes_read = iov_to_buf(pkt, pkt_frags, ext_hdr_offset, &rthdr,
                             sizeof(rthdr);
     if (bytes_read < sizeof(rthdr)) {
         return false;
     }

>
>     if ((rthdr->rtype == 2) && (rthdr->segleft == 1)) {
>         size_t bytes_read;
>diff --git a/tests/qtest/fuzz-e1000e-test.c b/tests/qtest/fuzz-e1000e-test.c
>new file mode 100644
>index 00000000000..66229e60964
>--- /dev/null
>+++ b/tests/qtest/fuzz-e1000e-test.c
>@@ -0,0 +1,53 @@
>+/*
>+ * QTest testcase for e1000e device generated by fuzzer
>+ *
>+ * Copyright (c) 2021 Red Hat, Inc.
>+ *
>+ * SPDX-License-Identifier: GPL-2.0-or-later
>+ */
>+
>+#include "qemu/osdep.h"
>+
>+#include "libqos/libqtest.h"
>+
>+/*
>+ * https://bugs.launchpad.net/qemu/+bug/1879531
>+ */
>+static void test_lp1879531_eth_get_rss_ex_dst_addr(void)
>+{
>+    QTestState *s;
>+
>+    s = qtest_init("-nographic -monitor none -serial none -M pc-q35-5.0");
>+
>+    qtest_outl(s, 0xcf8, 0x80001010);
>+    qtest_outl(s, 0xcfc, 0xe1020000);
>+    qtest_outl(s, 0xcf8, 0x80001004);
>+    qtest_outw(s, 0xcfc, 0x7);
>+    qtest_writeb(s, 0x25, 0x86);
>+    qtest_writeb(s, 0x26, 0xdd);
>+    qtest_writeb(s, 0x4f, 0x2b);
>+
>+    qtest_writel(s, 0xe1020030, 0x190002e1);
>+    qtest_writew(s, 0xe102003a, 0x0807);
>+    qtest_writel(s, 0xe1020048, 0x12077cdd);
>+    qtest_writel(s, 0xe1020400, 0xba077cdd);
>+    qtest_writel(s, 0xe1020420, 0x190002e1);
>+    qtest_writel(s, 0xe1020428, 0x3509d807);
>+    qtest_writeb(s, 0xe1020438, 0xe2);
>+    qtest_writeb(s, 0x4f, 0x2b);
>+    qtest_quit(s);
>+}
>+
>+int main(int argc, char **argv)
>+{
>+    const char *arch = qtest_get_arch();
>+
>+    g_test_init(&argc, &argv, NULL);
>+
>+    if (strcmp(arch, "i386") == 0 || strcmp(arch, "x86_64") == 0) {
>+        qtest_add_func("fuzz/test_lp1879531_eth_get_rss_ex_dst_addr",
>+                       test_lp1879531_eth_get_rss_ex_dst_addr);
>+    }
>+
>+    return g_test_run();
>+}
>diff --git a/MAINTAINERS b/MAINTAINERS
>index f22d83c1782..4eb5784ff83 100644
>--- a/MAINTAINERS
>+++ b/MAINTAINERS
>@@ -2003,6 +2003,7 @@ e1000e
> M: Dmitry Fleytman <dmitry.fleytman@gmail.com>
> S: Maintained
> F: hw/net/e1000e*
>+F: tests/qtest/fuzz-e1000e-test.c
>
> eepro100
> M: Stefan Weil <sw@weilnetz.de>
>diff --git a/tests/qtest/meson.build b/tests/qtest/meson.build
>index 58efc46144e..7997d895449 100644
>--- a/tests/qtest/meson.build
>+++ b/tests/qtest/meson.build
>@@ -60,6 +60,7 @@
>   (config_all_devices.has_key('CONFIG_TPM_TIS_ISA') ? ['tpm-tis-test'] : []) +              \
>   (config_all_devices.has_key('CONFIG_TPM_TIS_ISA') ? ['tpm-tis-swtpm-test'] : []) +        \
>   (config_all_devices.has_key('CONFIG_RTL8139_PCI') ? ['rtl8139-test'] : []) +              \
>+  (config_all_devices.has_key('CONFIG_E1000E_PCI_EXPRESS') ? ['fuzz-e1000e-test'] : []) +   \
>   qtests_pci +                                                                              \
>   ['fdc-test',
>    'ide-test',
>-- 
>2.26.2
>

Re: [PATCH v4 6/6] net/eth: Return earlier in _eth_get_rss_ex_dst_addr()

[Stefano Garzarella]

On Tue, Mar 09, 2021 at 07:27:09PM +0100, Philippe Mathieu-Daudé wrote:
>Slightly simplify _eth_get_rss_ex_dst_addr() by returning earlier.
>
>Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
>---
> net/eth.c | 16 +++++++---------
> 1 file changed, 7 insertions(+), 9 deletions(-)
>
>diff --git a/net/eth.c b/net/eth.c
>index e984edcfb0b..b44439d31c5 100644
>--- a/net/eth.c
>+++ b/net/eth.c
>@@ -407,23 +407,21 @@ _eth_get_rss_ex_dst_addr(const struct iovec *pkt, int pkt_frags,
> {
>     size_t input_size = iov_size(pkt, pkt_frags);
>     struct ip6_ext_hdr_routing *rthdr;
>+    size_t bytes_read;
>
>     if (input_size < ext_hdr_offset + sizeof(*rthdr)) {
>         return false;
>     }
>     rthdr = (struct ip6_ext_hdr_routing *) ext_hdr;
>
>-    if ((rthdr->rtype == 2) && (rthdr->segleft == 1)) {
>-        size_t bytes_read;
>-
>-        bytes_read = iov_to_buf(pkt, pkt_frags,
>-                                ext_hdr_offset + sizeof(*ext_hdr),
>-                                dst_addr, sizeof(*dst_addr));
>-
>-        return bytes_read == sizeof(*dst_addr);
>+    if ((rthdr->rtype != 2) || (rthdr->segleft != 1)) {
>+        return false;
>     }
>
>-    return false;
>+    bytes_read = iov_to_buf(pkt, pkt_frags, ext_hdr_offset + sizeof(*ext_hdr),
>+                            dst_addr, sizeof(*dst_addr));

Pre-existing issue, but looking at the Routing extension header format 
[1], I think the offset we're using here is wrong.

I had a doubt if the address started at ext_hdr_offset + 4 or 
ext_hdr_offset + 8 but looking in the linux code I think the offset we 
should use is ext_hdr_offset + sizeof(*rthdr).

This is the structure that I found in include/uapi/linux/ipv6.h:

     /*
      *	routing header type 2
      */

     struct rt2_hdr {
     	struct ipv6_rt_hdr	rt_hdr;
     	__u32			reserved;
     	struct in6_addr		addr;

     #define rt2_type		rt_hdr.type
     };

Thanks,
Stefano

[1] https://en.wikipedia.org/wiki/IPv6_packet#Routing

Re: [PATCH v4 4/6] net/eth: Check rt_hdr size before casting to ip6_ext_hdr

[Stefano Garzarella]

On Wed, Mar 10, 2021 at 10:05:01AM +0100, Stefano Garzarella wrote:
>On Tue, Mar 09, 2021 at 07:27:07PM +0100, Philippe Mathieu-Daudé wrote:
>>Do not cast our ip6_ext_hdr pointer to ip6_ext_hdr_routing if there
>>isn't enough data in the buffer for a such structure.
>>
>>This fix a 2 bytes buffer overrun in eth_parse_ipv6_hdr() reported
>>by QEMU fuzzer:
>>
>> $ cat << EOF | ./qemu-system-i386 -M pc-q35-5.0 \
>>   -accel qtest -monitor none \
>>   -serial none -nographic -qtest stdio
>> outl 0xcf8 0x80001010
>> outl 0xcfc 0xe1020000
>> outl 0xcf8 0x80001004
>> outw 0xcfc 0x7
>> write 0x25 0x1 0x86
>> write 0x26 0x1 0xdd
>> write 0x4f 0x1 0x2b
>> write 0xe1020030 0x4 0x190002e1
>> write 0xe102003a 0x2 0x0807
>> write 0xe1020048 0x4 0x12077cdd
>> write 0xe1020400 0x4 0xba077cdd
>> write 0xe1020420 0x4 0x190002e1
>> write 0xe1020428 0x4 0x3509d807
>> write 0xe1020438 0x1 0xe2
>> EOF
>> =================================================================
>> ==2859770==ERROR: AddressSanitizer: stack-buffer-overflow on address 0x7ffdef904902 at pc 0x561ceefa78de bp 0x7ffdef904820 sp 0x7ffdef904818
>> READ of size 1 at 0x7ffdef904902 thread T0
>>     #0 0x561ceefa78dd in _eth_get_rss_ex_dst_addr net/eth.c:410:17
>>     #1 0x561ceefa41fb in eth_parse_ipv6_hdr net/eth.c:532:17
>>     #2 0x561cef7de639 in net_tx_pkt_parse_headers hw/net/net_tx_pkt.c:228:14
>>     #3 0x561cef7dbef4 in net_tx_pkt_parse hw/net/net_tx_pkt.c:273:9
>>     #4 0x561ceec29f22 in e1000e_process_tx_desc hw/net/e1000e_core.c:730:29
>>     #5 0x561ceec28eac in e1000e_start_xmit hw/net/e1000e_core.c:927:9
>>     #6 0x561ceec1baab in e1000e_set_tdt hw/net/e1000e_core.c:2444:9
>>     #7 0x561ceebf300e in e1000e_core_write hw/net/e1000e_core.c:3256:9
>>     #8 0x561cef3cd4cd in e1000e_mmio_write hw/net/e1000e.c:110:5
>>
>> Address 0x7ffdef904902 is located in stack of thread T0 at offset 34 in frame
>>     #0 0x561ceefa320f in eth_parse_ipv6_hdr net/eth.c:486
>>
>>   This frame has 1 object(s):
>>     [32, 34) 'ext_hdr' (line 487) <== Memory access at offset 34 overflows this variable
>> HINT: this may be a false positive if your program uses some custom stack unwind mechanism, swapcontext or vfork
>>       (longjmp and C++ exceptions *are* supported)
>> SUMMARY: AddressSanitizer: stack-buffer-overflow net/eth.c:410:17 in _eth_get_rss_ex_dst_addr
>> Shadow bytes around the buggy address:
>>   0x10003df188d0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>>   0x10003df188e0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>>   0x10003df188f0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>>   0x10003df18900: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>>   0x10003df18910: 00 00 00 00 00 00 00 00 00 00 00 00 f1 f1 f1 f1
>> =>0x10003df18920:[02]f3 f3 f3 00 00 00 00 00 00 00 00 00 00 00 00
>>   0x10003df18930: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>>   0x10003df18940: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>>   0x10003df18950: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>>   0x10003df18960: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>>   0x10003df18970: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>> Shadow byte legend (one shadow byte represents 8 application bytes):
>>   Addressable:           00
>>   Partially addressable: 01 02 03 04 05 06 07
>>   Stack left redzone:      f1
>>   Stack right redzone:     f3
>> ==2859770==ABORTING
>>
>>Add the corresponding qtest case with the fuzzer reproducer.
>>
>>FWIW GCC 11 similarly reported:
>>
>> net/eth.c: In function 'eth_parse_ipv6_hdr':
>> net/eth.c:410:15: error: array subscript 'struct ip6_ext_hdr_routing[0]' is partly outside array bounds of 'struct ip6_ext_hdr[1]' [-Werror=array-bounds]
>>   410 |     if ((rthdr->rtype == 2) && (rthdr->segleft == 1)) {
>>       |          ~~~~~^~~~~~~
>> net/eth.c:485:24: note: while referencing 'ext_hdr'
>>   485 |     struct ip6_ext_hdr ext_hdr;
>>       |                        ^~~~~~~
>> net/eth.c:410:38: error: array subscript 'struct ip6_ext_hdr_routing[0]' is partly outside array bounds of 'struct ip6_ext_hdr[1]' [-Werror=array-bounds]
>>   410 |     if ((rthdr->rtype == 2) && (rthdr->segleft == 1)) {
>>       |                                 ~~~~~^~~~~~~~~
>> net/eth.c:485:24: note: while referencing 'ext_hdr'
>>   485 |     struct ip6_ext_hdr ext_hdr;
>>       |                        ^~~~~~~
>>
>>Cc: qemu-stable@nongnu.org
>>Buglink: https://bugs.launchpad.net/qemu/+bug/1879531
>>Reported-by: Alexander Bulekov <alxndr@bu.edu>
>>Reported-by: Miroslav Rezanina <mrezanin@redhat.com>
>>Fixes: eb700029c78 ("net_pkt: Extend packet abstraction as required by e1000e functionality")
>>Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
>>---
>>net/eth.c                      |  7 ++++-
>>tests/qtest/fuzz-e1000e-test.c | 53 ++++++++++++++++++++++++++++++++++
>>MAINTAINERS                    |  1 +
>>tests/qtest/meson.build        |  1 +
>>4 files changed, 61 insertions(+), 1 deletion(-)
>>create mode 100644 tests/qtest/fuzz-e1000e-test.c
>>
>>diff --git a/net/eth.c b/net/eth.c
>>index 77af2b673bb..f0c8dfe8df7 100644
>>--- a/net/eth.c
>>+++ b/net/eth.c
>>@@ -406,7 +406,12 @@ _eth_get_rss_ex_dst_addr(const struct iovec *pkt, int pkt_frags,
>>                        struct in6_address *dst_addr)
>>{
>>    size_t input_size = iov_size(pkt, pkt_frags);
>>-    struct ip6_ext_hdr_routing *rthdr = (struct ip6_ext_hdr_routing *) ext_hdr;
>>+    struct ip6_ext_hdr_routing *rthdr;
>>+
>>+    if (input_size < ext_hdr_offset + sizeof(*rthdr)) {
>>+        return false;
>>+    }
>>+    rthdr = (struct ip6_ext_hdr_routing *) ext_hdr;
>
>Also if we check 'input_size', I think this cast keeps having the 2 
>bytes buffer overrun issue since 'ext_hdr' contains only the first 2 
>bytes.
>I think we can remove the 'input_size' check and we should fix in this 
>way:
>
>    struct ip6_ext_hdr_routing rthdr;
>
>    bytes_read = iov_to_buf(pkt, pkt_frags, ext_hdr_offset, &rthdr,
>                            sizeof(rthdr);
>    if (bytes_read < sizeof(rthdr)) {
>        return false;
>    }
>

Just saw that was what you had done in v3 :-)

I think we need to call iov_to_buf 2 times, once to read rthdr and once 
to read dst_addr.

Thanks,
Stefano

Re: [PATCH v4 4/6] net/eth: Check rt_hdr size before casting to ip6_ext_hdr

[Miroslav Rezanina]

----- Original Message -----
> From: "Philippe Mathieu-Daudé" <philmd@redhat.com>
> To: qemu-devel@nongnu.org
> Cc: "Laurent Vivier" <lvivier@redhat.com>, "Dmitry Fleytman" <dmitry.fleytman@gmail.com>, "Miroslav Rezanina"
> <mrezanin@redhat.com>, "Li Qiang" <liq3ea@gmail.com>, "Paolo Bonzini" <pbonzini@redhat.com>, "Jason Wang"
> <jasowang@redhat.com>, "Thomas Huth" <thuth@redhat.com>, "Alexander Bulekov" <alxndr@bu.edu>, "Stefano Garzarella"
> <sgarzare@redhat.com>, "Philippe Mathieu-Daudé" <philmd@redhat.com>, qemu-stable@nongnu.org
> Sent: Tuesday, March 9, 2021 7:27:07 PM
> Subject: [PATCH v4 4/6] net/eth: Check rt_hdr size before casting to ip6_ext_hdr
> 
> Do not cast our ip6_ext_hdr pointer to ip6_ext_hdr_routing if there
> isn't enough data in the buffer for a such structure.
> 
> This fix a 2 bytes buffer overrun in eth_parse_ipv6_hdr() reported
> by QEMU fuzzer:
> 
>   $ cat << EOF | ./qemu-system-i386 -M pc-q35-5.0 \
>     -accel qtest -monitor none \
>     -serial none -nographic -qtest stdio
>   outl 0xcf8 0x80001010
>   outl 0xcfc 0xe1020000
>   outl 0xcf8 0x80001004
>   outw 0xcfc 0x7
>   write 0x25 0x1 0x86
>   write 0x26 0x1 0xdd
>   write 0x4f 0x1 0x2b
>   write 0xe1020030 0x4 0x190002e1
>   write 0xe102003a 0x2 0x0807
>   write 0xe1020048 0x4 0x12077cdd
>   write 0xe1020400 0x4 0xba077cdd
>   write 0xe1020420 0x4 0x190002e1
>   write 0xe1020428 0x4 0x3509d807
>   write 0xe1020438 0x1 0xe2
>   EOF
>   =================================================================
>   ==2859770==ERROR: AddressSanitizer: stack-buffer-overflow on address
>   0x7ffdef904902 at pc 0x561ceefa78de bp 0x7ffdef904820 sp 0x7ffdef904818
>   READ of size 1 at 0x7ffdef904902 thread T0
>       #0 0x561ceefa78dd in _eth_get_rss_ex_dst_addr net/eth.c:410:17
>       #1 0x561ceefa41fb in eth_parse_ipv6_hdr net/eth.c:532:17
>       #2 0x561cef7de639 in net_tx_pkt_parse_headers
>       hw/net/net_tx_pkt.c:228:14
>       #3 0x561cef7dbef4 in net_tx_pkt_parse hw/net/net_tx_pkt.c:273:9
>       #4 0x561ceec29f22 in e1000e_process_tx_desc hw/net/e1000e_core.c:730:29
>       #5 0x561ceec28eac in e1000e_start_xmit hw/net/e1000e_core.c:927:9
>       #6 0x561ceec1baab in e1000e_set_tdt hw/net/e1000e_core.c:2444:9
>       #7 0x561ceebf300e in e1000e_core_write hw/net/e1000e_core.c:3256:9
>       #8 0x561cef3cd4cd in e1000e_mmio_write hw/net/e1000e.c:110:5
> 
>   Address 0x7ffdef904902 is located in stack of thread T0 at offset 34 in
>   frame
>       #0 0x561ceefa320f in eth_parse_ipv6_hdr net/eth.c:486
> 
>     This frame has 1 object(s):
>       [32, 34) 'ext_hdr' (line 487) <== Memory access at offset 34 overflows
>       this variable
>   HINT: this may be a false positive if your program uses some custom stack
>   unwind mechanism, swapcontext or vfork
>         (longjmp and C++ exceptions *are* supported)
>   SUMMARY: AddressSanitizer: stack-buffer-overflow net/eth.c:410:17 in
>   _eth_get_rss_ex_dst_addr
>   Shadow bytes around the buggy address:
>     0x10003df188d0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>     0x10003df188e0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>     0x10003df188f0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>     0x10003df18900: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>     0x10003df18910: 00 00 00 00 00 00 00 00 00 00 00 00 f1 f1 f1 f1
>   =>0x10003df18920:[02]f3 f3 f3 00 00 00 00 00 00 00 00 00 00 00 00
>     0x10003df18930: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>     0x10003df18940: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>     0x10003df18950: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>     0x10003df18960: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>     0x10003df18970: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>   Shadow byte legend (one shadow byte represents 8 application bytes):
>     Addressable:           00
>     Partially addressable: 01 02 03 04 05 06 07
>     Stack left redzone:      f1
>     Stack right redzone:     f3
>   ==2859770==ABORTING
> 
> Add the corresponding qtest case with the fuzzer reproducer.
> 
> FWIW GCC 11 similarly reported:
> 
>   net/eth.c: In function 'eth_parse_ipv6_hdr':
>   net/eth.c:410:15: error: array subscript 'struct ip6_ext_hdr_routing[0]' is
>   partly outside array bounds of 'struct ip6_ext_hdr[1]'
>   [-Werror=array-bounds]
>     410 |     if ((rthdr->rtype == 2) && (rthdr->segleft == 1)) {
>         |          ~~~~~^~~~~~~
>   net/eth.c:485:24: note: while referencing 'ext_hdr'
>     485 |     struct ip6_ext_hdr ext_hdr;
>         |                        ^~~~~~~
>   net/eth.c:410:38: error: array subscript 'struct ip6_ext_hdr_routing[0]' is
>   partly outside array bounds of 'struct ip6_ext_hdr[1]'
>   [-Werror=array-bounds]
>     410 |     if ((rthdr->rtype == 2) && (rthdr->segleft == 1)) {
>         |                                 ~~~~~^~~~~~~~~
>   net/eth.c:485:24: note: while referencing 'ext_hdr'
>     485 |     struct ip6_ext_hdr ext_hdr;
>         |                        ^~~~~~~
> 
> Cc: qemu-stable@nongnu.org
> Buglink: https://bugs.launchpad.net/qemu/+bug/1879531
> Reported-by: Alexander Bulekov <alxndr@bu.edu>
> Reported-by: Miroslav Rezanina <mrezanin@redhat.com>
> Fixes: eb700029c78 ("net_pkt: Extend packet abstraction as required by e1000e
> functionality")
> Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
> ---
>  net/eth.c                      |  7 ++++-
>  tests/qtest/fuzz-e1000e-test.c | 53 ++++++++++++++++++++++++++++++++++
>  MAINTAINERS                    |  1 +
>  tests/qtest/meson.build        |  1 +
>  4 files changed, 61 insertions(+), 1 deletion(-)
>  create mode 100644 tests/qtest/fuzz-e1000e-test.c
> 
> diff --git a/net/eth.c b/net/eth.c
> index 77af2b673bb..f0c8dfe8df7 100644
> --- a/net/eth.c
> +++ b/net/eth.c
> @@ -406,7 +406,12 @@ _eth_get_rss_ex_dst_addr(const struct iovec *pkt, int
> pkt_frags,
>                          struct in6_address *dst_addr)
>  {
>      size_t input_size = iov_size(pkt, pkt_frags);
> -    struct ip6_ext_hdr_routing *rthdr = (struct ip6_ext_hdr_routing *)
> ext_hdr;
> +    struct ip6_ext_hdr_routing *rthdr;
> +
> +    if (input_size < ext_hdr_offset + sizeof(*rthdr)) {
> +        return false;
> +    }
> +    rthdr = (struct ip6_ext_hdr_routing *) ext_hdr;

Hi Philippe,

you're introducing the problem you're trying to fix here. This line
cause warning on GCC 11 and so the build fail with --enable-werror.

return statement has no effect on compiler.

Mirek
>  
>      if ((rthdr->rtype == 2) && (rthdr->segleft == 1)) {
>          size_t bytes_read;
> diff --git a/tests/qtest/fuzz-e1000e-test.c b/tests/qtest/fuzz-e1000e-test.c
> new file mode 100644
> index 00000000000..66229e60964
> --- /dev/null
> +++ b/tests/qtest/fuzz-e1000e-test.c
> @@ -0,0 +1,53 @@
> +/*
> + * QTest testcase for e1000e device generated by fuzzer
> + *
> + * Copyright (c) 2021 Red Hat, Inc.
> + *
> + * SPDX-License-Identifier: GPL-2.0-or-later
> + */
> +
> +#include "qemu/osdep.h"
> +
> +#include "libqos/libqtest.h"
> +
> +/*
> + * https://bugs.launchpad.net/qemu/+bug/1879531
> + */
> +static void test_lp1879531_eth_get_rss_ex_dst_addr(void)
> +{
> +    QTestState *s;
> +
> +    s = qtest_init("-nographic -monitor none -serial none -M pc-q35-5.0");
> +
> +    qtest_outl(s, 0xcf8, 0x80001010);
> +    qtest_outl(s, 0xcfc, 0xe1020000);
> +    qtest_outl(s, 0xcf8, 0x80001004);
> +    qtest_outw(s, 0xcfc, 0x7);
> +    qtest_writeb(s, 0x25, 0x86);
> +    qtest_writeb(s, 0x26, 0xdd);
> +    qtest_writeb(s, 0x4f, 0x2b);
> +
> +    qtest_writel(s, 0xe1020030, 0x190002e1);
> +    qtest_writew(s, 0xe102003a, 0x0807);
> +    qtest_writel(s, 0xe1020048, 0x12077cdd);
> +    qtest_writel(s, 0xe1020400, 0xba077cdd);
> +    qtest_writel(s, 0xe1020420, 0x190002e1);
> +    qtest_writel(s, 0xe1020428, 0x3509d807);
> +    qtest_writeb(s, 0xe1020438, 0xe2);
> +    qtest_writeb(s, 0x4f, 0x2b);
> +    qtest_quit(s);
> +}
> +
> +int main(int argc, char **argv)
> +{
> +    const char *arch = qtest_get_arch();
> +
> +    g_test_init(&argc, &argv, NULL);
> +
> +    if (strcmp(arch, "i386") == 0 || strcmp(arch, "x86_64") == 0) {
> +        qtest_add_func("fuzz/test_lp1879531_eth_get_rss_ex_dst_addr",
> +                       test_lp1879531_eth_get_rss_ex_dst_addr);
> +    }
> +
> +    return g_test_run();
> +}
> diff --git a/MAINTAINERS b/MAINTAINERS
> index f22d83c1782..4eb5784ff83 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -2003,6 +2003,7 @@ e1000e
>  M: Dmitry Fleytman <dmitry.fleytman@gmail.com>
>  S: Maintained
>  F: hw/net/e1000e*
> +F: tests/qtest/fuzz-e1000e-test.c
>  
>  eepro100
>  M: Stefan Weil <sw@weilnetz.de>
> diff --git a/tests/qtest/meson.build b/tests/qtest/meson.build
> index 58efc46144e..7997d895449 100644
> --- a/tests/qtest/meson.build
> +++ b/tests/qtest/meson.build
> @@ -60,6 +60,7 @@
>    (config_all_devices.has_key('CONFIG_TPM_TIS_ISA') ? ['tpm-tis-test'] : [])
>    +              \
>    (config_all_devices.has_key('CONFIG_TPM_TIS_ISA') ? ['tpm-tis-swtpm-test']
>    : []) +        \
>    (config_all_devices.has_key('CONFIG_RTL8139_PCI') ? ['rtl8139-test'] : [])
>    +              \
> +  (config_all_devices.has_key('CONFIG_E1000E_PCI_EXPRESS') ?
> ['fuzz-e1000e-test'] : []) +   \
>    qtests_pci +
>    \
>    ['fdc-test',
>     'ide-test',
> --
> 2.26.2
> 
> 

-- 
Miroslav Rezanina
Software Engineer - Virtualization Team Maintainer

Re: [PATCH v4 4/6] net/eth: Check rt_hdr size before casting to ip6_ext_hdr

[Philippe Mathieu-Daudé]

On 3/10/21 10:36 AM, Miroslav Rezanina wrote:
> ----- Original Message -----
>> From: "Philippe Mathieu-Daudé" <philmd@redhat.com>
>> To: qemu-devel@nongnu.org
>> Cc: "Laurent Vivier" <lvivier@redhat.com>, "Dmitry Fleytman" <dmitry.fleytman@gmail.com>, "Miroslav Rezanina"
>> <mrezanin@redhat.com>, "Li Qiang" <liq3ea@gmail.com>, "Paolo Bonzini" <pbonzini@redhat.com>, "Jason Wang"
>> <jasowang@redhat.com>, "Thomas Huth" <thuth@redhat.com>, "Alexander Bulekov" <alxndr@bu.edu>, "Stefano Garzarella"
>> <sgarzare@redhat.com>, "Philippe Mathieu-Daudé" <philmd@redhat.com>, qemu-stable@nongnu.org
>> Sent: Tuesday, March 9, 2021 7:27:07 PM
>> Subject: [PATCH v4 4/6] net/eth: Check rt_hdr size before casting to ip6_ext_hdr
>>
>> Do not cast our ip6_ext_hdr pointer to ip6_ext_hdr_routing if there
>> isn't enough data in the buffer for a such structure.
>>

>> FWIW GCC 11 similarly reported:
>>
>>   net/eth.c: In function 'eth_parse_ipv6_hdr':
>>   net/eth.c:410:15: error: array subscript 'struct ip6_ext_hdr_routing[0]' is
>>   partly outside array bounds of 'struct ip6_ext_hdr[1]'
>>   [-Werror=array-bounds]
>>     410 |     if ((rthdr->rtype == 2) && (rthdr->segleft == 1)) {
>>         |          ~~~~~^~~~~~~
>>   net/eth.c:485:24: note: while referencing 'ext_hdr'
>>     485 |     struct ip6_ext_hdr ext_hdr;
>>         |                        ^~~~~~~
>>   net/eth.c:410:38: error: array subscript 'struct ip6_ext_hdr_routing[0]' is
>>   partly outside array bounds of 'struct ip6_ext_hdr[1]'
>>   [-Werror=array-bounds]
>>     410 |     if ((rthdr->rtype == 2) && (rthdr->segleft == 1)) {
>>         |                                 ~~~~~^~~~~~~~~
>>   net/eth.c:485:24: note: while referencing 'ext_hdr'
>>     485 |     struct ip6_ext_hdr ext_hdr;
>>         |                        ^~~~~~~
>>
>> Cc: qemu-stable@nongnu.org
>> Buglink: https://bugs.launchpad.net/qemu/+bug/1879531
>> Reported-by: Alexander Bulekov <alxndr@bu.edu>
>> Reported-by: Miroslav Rezanina <mrezanin@redhat.com>
>> Fixes: eb700029c78 ("net_pkt: Extend packet abstraction as required by e1000e
>> functionality")
>> Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
>> ---
>>  net/eth.c                      |  7 ++++-
>>  tests/qtest/fuzz-e1000e-test.c | 53 ++++++++++++++++++++++++++++++++++
>>  MAINTAINERS                    |  1 +
>>  tests/qtest/meson.build        |  1 +
>>  4 files changed, 61 insertions(+), 1 deletion(-)
>>  create mode 100644 tests/qtest/fuzz-e1000e-test.c
>>
>> diff --git a/net/eth.c b/net/eth.c
>> index 77af2b673bb..f0c8dfe8df7 100644
>> --- a/net/eth.c
>> +++ b/net/eth.c
>> @@ -406,7 +406,12 @@ _eth_get_rss_ex_dst_addr(const struct iovec *pkt, int
>> pkt_frags,
>>                          struct in6_address *dst_addr)
>>  {
>>      size_t input_size = iov_size(pkt, pkt_frags);
>> -    struct ip6_ext_hdr_routing *rthdr = (struct ip6_ext_hdr_routing *)
>> ext_hdr;
>> +    struct ip6_ext_hdr_routing *rthdr;
>> +
>> +    if (input_size < ext_hdr_offset + sizeof(*rthdr)) {
>> +        return false;
>> +    }
>> +    rthdr = (struct ip6_ext_hdr_routing *) ext_hdr;
> 
> Hi Philippe,
> 
> you're introducing the problem you're trying to fix here. This line
> cause warning on GCC 11 and so the build fail with --enable-werror.

Oh, I thought we had switched to testing on Fedora rawhide, but this
never made it to mainstream [*], which is why my testing didn't notice
this was not fixing.

I'll keep working on this, thanks.

[*] https://www.mail-archive.com/qemu-devel@nongnu.org/msg774720.html