QEMU

[OSS-Fuzz] Issue 29174 sb16: Abrt in audio_bug

Bug #1910603 reported by Alexander Bulekov on 2021-01-07

This bug affects 1 person

Affects		Status	Importance	Assigned to	Milestone
	QEMU	Fix Released	Undecided	Unassigned

Bug Description

=== Reproducer ===
cat << EOF | ../build-system/qemu-system-i386 \
-machine q35 -device sb16,audiodev=snd0 \
-audiodev none,id=snd0 -nographic -nodefaults \
-qtest stdio
outw 0x22c 0x41
outb 0x22c 0x0
outw 0x22c 0x1004
outw 0x22c 0x1c
EOF

=== Stack Trace ===
A bug was just triggered in audio_calloc
Save all your work and restart without audio
I am sorry
Context:
Aborted

#0 raise
#1 abort
#2 audio_bug /src/qemu/audio/audio.c:119:9
#3 audio_calloc /src/qemu/audio/audio.c:154:9
#4 audio_pcm_sw_alloc_resources_out /src/qemu/audio/audio_template.h:116:15
#5 audio_pcm_sw_init_out /src/qemu/audio/audio_template.h:175:11
#6 audio_pcm_create_voice_pair_out /src/qemu/audio/audio_template.h:410:9
#7 AUD_open_out /src/qemu/audio/audio_template.h:503:14
#8 continue_dma8 /src/qemu/hw/audio/sb16.c:216:20
#9 dma_cmd8 /src/qemu/hw/audio/sb16.c:276:5
#10 command /src/qemu/hw/audio/sb16.c:0
#11 dsp_write /src/qemu/hw/audio/sb16.c:949:13
#12 portio_write /src/qemu/softmmu/ioport.c:205:13
#13 memory_region_write_accessor /src/qemu/softmmu/memory.c:491:5
#14 access_with_adjusted_size /src/qemu/softmmu/memory.c:552:18
#15 memory_region_dispatch_write /src/qemu/softmmu/memory.c:0:13
#16 flatview_write_continue /src/qemu/softmmu/physmem.c:2759:23
#17 flatview_write /src/qemu/softmmu/physmem.c:2799:14
#18 address_space_write /src/qemu/softmmu/physmem.c:2891:18
#19 cpu_outw /src/qemu/softmmu/ioport.c:70:5

OSS-Fuzz Report:
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=29174

Tags:

Peter Maydell (pmaydell) on 2021-01-15

tags:

added: fuzzer

Revision history for this message

Thomas Huth (th-huth) wrote on 2021-05-26:

This is still reproducible with the current version of QEMU. Marking this as "Confirmed"

Changed in qemu:
status:	New → Confirmed
tags:	added: audio

Revision history for this message

Philippe Mathieu-Daudé (philmd) wrote on 2021-06-01: [PATCH] hw/audio/sb16: Avoid assertion by restricting I/O sampling rate range

Download full text (5.6 KiB)

While the SB16 seems to work up to 48000 Hz, the "Sound Blaster Series
Hardware Programming Guide" limit the sampling range from 4000 Hz to
44100 Hz (Section 3-9, 3-10: Digitized Sound I/O Programming, tables
3-2 and 3-3).

Later, section 6-15 (DSP Commands) is more specific regarding the 41h /
42h registers (Set digitized sound output sampling rate):

Valid sampling rates range from 5000 to 45000 Hz inclusive.

There is no comment regarding error handling if the register is filled
with an out-of-range value. (See also section 3-28 "8-bit or 16-bit
Auto-initialize Transfer"). Assume limits are enforced in hardware.

This fixes triggering an assertion in audio_calloc():

  #1 abort
  #2 audio_bug audio/audio.c:119:9
  #3 audio_calloc audio/audio.c:154:9
  #4 audio_pcm_sw_alloc_resources_out audio/audio_template.h:116:15
  #5 audio_pcm_sw_init_out audio/audio_template.h:175:11
  #6 audio_pcm_create_voice_pair_out audio/audio_template.h:410:9
  #7 AUD_open_out audio/audio_template.h:503:14
  #8 continue_dma8 hw/audio/sb16.c:216:20
  #9 dma_cmd8 hw/audio/sb16.c:276:5
  #10 command hw/audio/sb16.c:0
  #11 dsp_write hw/audio/sb16.c:949:13
  #12 portio_write softmmu/ioport.c:205:13
  #13 memory_region_write_accessor softmmu/memory.c:491:5
  #14 access_with_adjusted_size softmmu/memory.c:552:18
  #15 memory_region_dispatch_write softmmu/memory.c:0:13
  #16 flatview_write_continue softmmu/physmem.c:2759:23
  #17 flatview_write softmmu/physmem.c:2799:14
  #18 address_space_write softmmu/physmem.c:2891:18
  #19 cpu_outw softmmu/ioport.c:70:5

[*] http://www.baudline.com/solutions/full_duplex/sb16_pci/index.html

Fixes: 85571bc7415 ("audio merge (malc)")
Buglink: https://bugs.launchpad.net/bugs/1910603
OSS-Fuzz Report: https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=29174
Signed-off-by: Philippe Mathieu-Daudé <email address hidden>
---
hw/audio/sb16.c | 14 ++++++++++
tests/qtest/fuzz-sb16-test.c | 52 ++++++++++++++++++++++++++++++++++++
MAINTAINERS | 1 +
tests/qtest/meson.build | 1 +
4 files changed, 68 insertions(+)
create mode 100644 tests/qtest/fuzz-sb16-test.c

diff --git a/hw/audio/sb16.c b/hw/audio/sb16.c
index 8b207004102..5cf121fe363 100644
--- a/hw/audio/sb16.c
+++ b/hw/audio/sb16.c
@@ -115,6 +115,9 @@ struct SB16State {
PortioList portio_list;
};

+#define SAMPLE_RATE_MIN 5000
+#define SAMPLE_RATE_MAX 45000
+
static void SB_audio_callback (void *opaque, int free);

static int magic_of_irq (int irq)
@@ -241,6 +244,17 @@ static void dma_cmd8 (SB16State *s, int mask, int dma_len)
         int tmp = (256 - s->time_const);
         s->freq = (1000000 + (tmp / 2)) / tmp;
     }
+ if (s->freq < SAMPLE_RATE_MIN) {
+ qemu_log_mask(LOG_GUEST_ERROR,
+ "sampling range too low: %d, increasing to %u\n",
+ s->freq, SAMPLE_RATE_MIN);
+ s->freq = SAMPLE_RATE_MIN;
+ } else if (s->freq > SAMPLE_RATE_MAX) {
+ qemu_log_mask(LOG_GUEST_ERROR,
+ "sampling range too high: %d, decreasing to %u\n",
+ s->freq, SAMPLE_RATE_MAX);
+ s->freq = SAMPLE_RATE_MAX;
+ }

if (dma_len != -1) {
...

Later, section 6-15 (DSP Commands) is more specific regarding the 41h /
42h registers (Set digitized sound output sampling rate):

Valid sampling rates range from 5000 to 45000 Hz inclusive.

There is no comment regarding error handling if the register is filled
with an out-of-range value.  (See also section 3-28 "8-bit or 16-bit
Auto-initialize Transfer"). Assume limits are enforced in hardware.

This fixes triggering an assertion in audio_calloc():

[*] http://www.baudline.com/solutions/full_duplex/sb16_pci/index.html

Fixes: 85571bc7415 ("audio merge (malc)")
Buglink: https://bugs.launchpad.net/bugs/1910603
OSS-Fuzz Report: https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=29174
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
---
 hw/audio/sb16.c              | 14 ++++++++++
 tests/qtest/fuzz-sb16-test.c | 52 ++++++++++++++++++++++++++++++++++++
 MAINTAINERS                  |  1 +
 tests/qtest/meson.build      |  1 +
 4 files changed, 68 insertions(+)
 create mode 100644 tests/qtest/fuzz-sb16-test.c

diff --git a/hw/audio/sb16.c b/hw/audio/sb16.c
index 8b207004102..5cf121fe363 100644
--- a/hw/audio/sb16.c
+++ b/hw/audio/sb16.c
@@ -115,6 +115,9 @@ struct SB16State {
     PortioList portio_list;
 };
 
+#define SAMPLE_RATE_MIN 5000
+#define SAMPLE_RATE_MAX 45000
+
 static void SB_audio_callback (void *opaque, int free);
 
 static int magic_of_irq (int irq)
@@ -241,6 +244,17 @@ static void dma_cmd8 (SB16State *s, int mask, int dma_len)
         int tmp = (256 - s->time_const);
         s->freq = (1000000 + (tmp / 2)) / tmp;
     }
+    if (s->freq < SAMPLE_RATE_MIN) {
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "sampling range too low: %d, increasing to %u\n",
+                      s->freq, SAMPLE_RATE_MIN);
+        s->freq = SAMPLE_RATE_MIN;
+    } else if (s->freq > SAMPLE_RATE_MAX) {
+        qemu_log_mask(LOG_GUEST_ERROR,
+                      "sampling range too high: %d, decreasing to %u\n",
+                      s->freq, SAMPLE_RATE_MAX);
+        s->freq = SAMPLE_RATE_MAX;
+    }
 
     if (dma_len != -1) {
         s->block_size = dma_len << s->fmt_stereo;
diff --git a/tests/qtest/fuzz-sb16-test.c b/tests/qtest/fuzz-sb16-test.c
new file mode 100644
index 00000000000..51030cd7dc4
--- /dev/null
+++ b/tests/qtest/fuzz-sb16-test.c
@@ -0,0 +1,52 @@
+/*
+ * QTest fuzzer-generated testcase for sb16 audio device
+ *
+ * Copyright (c) 2021 Philippe Mathieu-Daudé <f4bug@amsat.org>
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#include "qemu/osdep.h"
+#include "libqos/libqtest.h"
+
+/*
+ * This used to trigger the assert in audio_calloc
+ * https://bugs.launchpad.net/qemu/+bug/1910603
+ */
+static void test_fuzz_sb16_0x1c(void)
+{
+    QTestState *s = qtest_init("-M q35 -display none "
+                               "-device sb16,audiodev=snd0 "
+                               "-audiodev none,id=snd0");
+    qtest_outw(s, 0x22c, 0x41);
+    qtest_outb(s, 0x22c, 0x00);
+    qtest_outw(s, 0x22c, 0x1004);
+    qtest_outw(s, 0x22c, 0x001c);
+    qtest_quit(s);
+}
+
+static void test_fuzz_sb16_0x91(void)
+{
+    QTestState *s = qtest_init("-M pc -display none "
+                               "-device sb16,audiodev=none "
+                               "-audiodev id=none,driver=none");
+    qtest_outw(s, 0x22c, 0xf141);
+    qtest_outb(s, 0x22c, 0x00);
+    qtest_outb(s, 0x22c, 0x24);
+    qtest_outb(s, 0x22c, 0x91);
+    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) {
+        qtest_add_func("fuzz/test_fuzz_sb16/1c", test_fuzz_sb16_0x1c);
+        qtest_add_func("fuzz/test_fuzz_sb16/91", test_fuzz_sb16_0x91);
+   }
+
+   return g_test_run();
+}
diff --git a/MAINTAINERS b/MAINTAINERS
index 5f55404f2fa..7edb26d2293 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -2213,6 +2213,7 @@ F: qapi/audio.json
 F: tests/qtest/ac97-test.c
 F: tests/qtest/es1370-test.c
 F: tests/qtest/intel-hda-test.c
+F: tests/qtest/fuzz-sb16-test.c
 
 Block layer core
 M: Kevin Wolf <kwolf@redhat.com>
diff --git a/tests/qtest/meson.build b/tests/qtest/meson.build
index c3a223a83d6..b03e8541700 100644
--- a/tests/qtest/meson.build
+++ b/tests/qtest/meson.build
@@ -20,6 +20,7 @@
 qtests_generic = \
   (config_all_devices.has_key('CONFIG_MEGASAS_SCSI_PCI') ? ['fuzz-megasas-test'] : []) + \
   (config_all_devices.has_key('CONFIG_VIRTIO_SCSI') ? ['fuzz-virtio-scsi-test'] : []) + \
+  (config_all_devices.has_key('CONFIG_SB16') ? ['fuzz-sb16-test'] : []) + \
   [
   'cdrom-test',
   'device-introspect-test',
-- 
2.26.3

Revision history for this message

Alexander Bulekov (a1xndr) wrote on 2021-06-19:

OSS-Fuzz confirms this is fixed: https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=30574#c4