提交 e3ac5105 编写于 作者: T Takashi Iwai 提交者: Zheng Zengkai

ALSA: oss: Fix PCM OSS buffer allocation overflow

stable inclusion
from stable-v5.10.109
commit 5ce74ff7059341d8b2f4d01c3383491df63d1898
bugzilla: https://gitee.com/openeuler/kernel/issues/I574AE

Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=5ce74ff7059341d8b2f4d01c3383491df63d1898

--------------------------------

commit efb6402c upstream.

We've got syzbot reports hitting INT_MAX overflow at vmalloc()
allocation that is called from snd_pcm_plug_alloc().  Although we
apply the restrictions to input parameters, it's based only on the
hw_params of the underlying PCM device.  Since the PCM OSS layer
allocates a temporary buffer for the data conversion, the size may
become unexpectedly large when more channels or higher rates is given;
in the reported case, it went over INT_MAX, hence it hits WARN_ON().

This patch is an attempt to avoid such an overflow and an allocation
for too large buffers.  First off, it adds the limit of 1MB as the
upper bound for period bytes.  This must be large enough for all use
cases, and we really don't want to handle a larger temporary buffer
than this size.  The size check is performed at two places, where the
original period bytes is calculated and where the plugin buffer size
is calculated.

In addition, the driver uses array_size() and array3_size() for
multiplications to catch overflows for the converted period size and
buffer bytes.

Reported-by: syzbot+72732c532ac1454eeee9@syzkaller.appspotmail.com
Suggested-by: NLinus Torvalds <torvalds@linux-foundation.org>
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/00000000000085b1b305da5a66f3@google.com
Link: https://lore.kernel.org/r/20220318082036.29699-1-tiwai@suse.deSigned-off-by: NTakashi Iwai <tiwai@suse.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: NYu Liao <liaoyu15@huawei.com>
Reviewed-by: NWei Li <liwei391@huawei.com>
Signed-off-by: NZheng Zengkai <zhengzengkai@huawei.com>
上级 1c4bbee3
......@@ -774,6 +774,11 @@ static int snd_pcm_oss_period_size(struct snd_pcm_substream *substream,
if (oss_period_size < 16)
return -EINVAL;
/* don't allocate too large period; 1MB period must be enough */
if (oss_period_size > 1024 * 1024)
return -ENOMEM;
runtime->oss.period_bytes = oss_period_size;
runtime->oss.period_frames = 1;
runtime->oss.periods = oss_periods;
......@@ -1042,10 +1047,9 @@ static int snd_pcm_oss_change_params_locked(struct snd_pcm_substream *substream)
goto failure;
}
#endif
oss_period_size *= oss_frame_size;
oss_buffer_size = oss_period_size * runtime->oss.periods;
if (oss_buffer_size < 0) {
oss_period_size = array_size(oss_period_size, oss_frame_size);
oss_buffer_size = array_size(oss_period_size, runtime->oss.periods);
if (oss_buffer_size <= 0) {
err = -EINVAL;
goto failure;
}
......
......@@ -61,7 +61,10 @@ static int snd_pcm_plugin_alloc(struct snd_pcm_plugin *plugin, snd_pcm_uframes_t
}
if ((width = snd_pcm_format_physical_width(format->format)) < 0)
return width;
size = frames * format->channels * width;
size = array3_size(frames, format->channels, width);
/* check for too large period size once again */
if (size > 1024 * 1024)
return -ENOMEM;
if (snd_BUG_ON(size % 8))
return -ENXIO;
size /= 8;
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册