This patchset contains the anonymous dup implementation with permissions checking for the ALSA's PCM interface in kernel to enable the restricted DMA sound buffer sharing for the restricted tasks.
The code was tested through qemu and it seems to be pretty stable.
The initial tinyalsa implementation can be found here:
https://github.com/perexg/tinyalsa/commits/anondup
The filtering might be refined. It depends on the real requirements. Perhaps, we may create more ioctl groups. Any comments are more than welcome.
v2 of the patches:
- change clone parameter to subdevice number for the pcm attach - change SNDRV_PCM_PERM_MAX to SNDRV_PCM_PERM_MASK - the tinyalsa implementation was a little updated (restructured)
v3 of the patches:
- group integer declarations in snd_pcm_anonymous_dup() - replaced substream->pcm with pcm in snd_pcm_anonymous_dup() - added SNDRV_PCM_PERM_RW check for read/write/readv/writev syscalls
v4 of the patches:
- more simple restriction control (only two modes - full/buffer) - the tinyalsa implementation follows this change
v5 of the patches:
- merge pcm_..._mmap_allowed fcns to the snd_pcm_mmap_... fcns
Cc: Phil Burk philburk@google.com Cc: Zach Riggle riggle@google.com Cc: Mark Brown broonie@kernel.org Cc: Leo Yan leo.yan@linaro.org Cc: Baolin Wang baolin.wang@linaro.org
Jaroslav Kysela (3): ALSA: pcm: implement the anonymous dup (inode file descriptor) ALSA: pcm: merge pcm_..._mmap_allowed fcns to the snd_pcm_mmap_... fcns ALSA: pcm: implement the mmap buffer mode for the anonymous dup
include/sound/pcm.h | 10 +-- include/uapi/sound/asound.h | 6 +- sound/core/oss/pcm_oss.c | 2 +- sound/core/pcm.c | 13 ++-- sound/core/pcm_compat.c | 1 + sound/core/pcm_native.c | 151 ++++++++++++++++++++++++++++++++------------ 6 files changed, 130 insertions(+), 53 deletions(-)