[PATCH] ALSA: pcm: Serialize PCM mmap with buffer reallocation to fix page UAF

[PATCH] ALSA: pcm: Serialize PCM mmap with buffer reallocation to fix page UAF

[Yilin Zhang]

snd_pcm_hw_params() and snd_pcm_hw_free() guard buffer reallocation
with an mmap_count check performed under the PCM stream lock, but the
lock is released long before the buffer is actually freed:
snd_pcm_sync_stop(), constraint refinement and do_free_pages() all
happen in between.  snd_pcm_mmap_data(), on the other hand, takes no
lock at all: it validates against the old buffer's state and
dma_bytes, remaps its pages into the VMA, and only then increments
mmap_count.

A concurrent mmap() can therefore slip in between the check and the
free.  remap_pfn_range() installs writable PTEs for the old buffer's
pages without taking page references, and the subsequent
do_free_pages() returns those pages to the page allocator while the
VMA still maps them.  This leaves a stale, writable mapping of freed
pages: a page-level use-after-free that can be leveraged for local
privilege escalation.

Make snd_pcm_mmap_data() participate in the buffer-access scheme
introduced for hw_params/hw_free: acquire runtime->buffer_accessing
before validating and remapping, and release it afterwards.  Buffer
reallocation already fails with -EBUSY while accessors are active,
and the mmap side now fails with -EBUSY while a reallocation is in
progress, so the validate/remap sequence and the check/free sequence
can no longer interleave.

A reproducer that turns this race into a stale writable mapping of
the freed DMA buffer pages is available on request.

Reported-by: Kimi Security Team <bug-report@moonshot.ai>
Fixes: 92ee3c60ec9f ("ALSA: pcm: Fix races among concurrent hw_params and hw_free calls")
Signed-off-by: Yilin Zhang <yilinzhang@moonshot.ai>
---
 sound/core/pcm_native.c | 35 +++++++++++++++++++++++++----------
 1 file changed, 25 insertions(+), 10 deletions(-)

diff --git a/sound/core/pcm_native.c b/sound/core/pcm_native.c
--- a/sound/core/pcm_native.c
+++ b/sound/core/pcm_native.c
@@ -3937,20 +3937,33 @@
 			return -EINVAL;
 	}
 	runtime = substream->runtime;
-	if (runtime->state == SNDRV_PCM_STATE_OPEN)
-		return -EBADFD;
-	if (!(runtime->info & SNDRV_PCM_INFO_MMAP))
-		return -ENXIO;
+	/* don't race with buffer reallocation in hw_params/hw_free */
+	if (!atomic_inc_unless_negative(&runtime->buffer_accessing))
+		return -EBUSY;
+	if (runtime->state == SNDRV_PCM_STATE_OPEN) {
+		err = -EBADFD;
+		goto out;
+	}
+	if (!(runtime->info & SNDRV_PCM_INFO_MMAP)) {
+		err = -ENXIO;
+		goto out;
+	}
 	if (runtime->access == SNDRV_PCM_ACCESS_RW_INTERLEAVED ||
-	    runtime->access == SNDRV_PCM_ACCESS_RW_NONINTERLEAVED)
-		return -EINVAL;
+	    runtime->access == SNDRV_PCM_ACCESS_RW_NONINTERLEAVED) {
+		err = -EINVAL;
+		goto out;
+	}
 	size = area->vm_end - area->vm_start;
 	offset = area->vm_pgoff << PAGE_SHIFT;
 	dma_bytes = PAGE_ALIGN(runtime->dma_bytes);
-	if ((size_t)size > dma_bytes)
-		return -EINVAL;
-	if (offset > dma_bytes - size)
-		return -EINVAL;
+	if ((size_t)size > dma_bytes) {
+		err = -EINVAL;
+		goto out;
+	}
+	if (offset > dma_bytes - size) {
+		err = -EINVAL;
+		goto out;
+	}
 
 	area->vm_ops = &snd_pcm_vm_ops_data;
 	area->vm_private_data = substream;
@@ -3960,6 +3973,8 @@
 		err = snd_pcm_lib_default_mmap(substream, area);
 	if (!err)
 		atomic_inc(&substream->mmap_count);
+out:
+	atomic_dec(&runtime->buffer_accessing);
 	return err;
 }
 EXPORT_SYMBOL(snd_pcm_mmap_data);
-- 
2.43.0

Re: [PATCH] ALSA: pcm: Serialize PCM mmap with buffer reallocation to fix page UAF

[Takashi Iwai]

On Mon, 31 Aug 2026 06:55:06 +0200,
Yilin Zhang wrote:
> 
> snd_pcm_hw_params() and snd_pcm_hw_free() guard buffer reallocation
> with an mmap_count check performed under the PCM stream lock, but the
> lock is released long before the buffer is actually freed:
> snd_pcm_sync_stop(), constraint refinement and do_free_pages() all
> happen in between.  snd_pcm_mmap_data(), on the other hand, takes no
> lock at all: it validates against the old buffer's state and
> dma_bytes, remaps its pages into the VMA, and only then increments
> mmap_count.
> 
> A concurrent mmap() can therefore slip in between the check and the
> free.  remap_pfn_range() installs writable PTEs for the old buffer's
> pages without taking page references, and the subsequent
> do_free_pages() returns those pages to the page allocator while the
> VMA still maps them.  This leaves a stale, writable mapping of freed
> pages: a page-level use-after-free that can be leveraged for local
> privilege escalation.
> 
> Make snd_pcm_mmap_data() participate in the buffer-access scheme
> introduced for hw_params/hw_free: acquire runtime->buffer_accessing
> before validating and remapping, and release it afterwards.  Buffer
> reallocation already fails with -EBUSY while accessors are active,
> and the mmap side now fails with -EBUSY while a reallocation is in
> progress, so the validate/remap sequence and the check/free sequence
> can no longer interleave.
> 
> A reproducer that turns this race into a stale writable mapping of
> the freed DMA buffer pages is available on request.
> 
> Reported-by: Kimi Security Team <bug-report@moonshot.ai>
> Fixes: 92ee3c60ec9f ("ALSA: pcm: Fix races among concurrent hw_params and hw_free calls")
> Signed-off-by: Yilin Zhang <yilinzhang@moonshot.ai>

Applied now.  Thanks.


Takashi