If some remap_pfn_range() calls succeeded before one failed, we still have
buffer pages mapped into the userspace page tables when we drop the buffer
reference with comedi_buf_map_put(bm). The userspace mappings are only
cleaned up later in the mmap error path.

Fix it by explicitly flushing all mappings in our VMA on the comedi_mmap()
error path.

See commit 79a61cc3fc04 ("mm: avoid leaving partial pfn mappings around in
error case").

Cc: stable@vger.kernel.org
Fixes: ed9eccbe8970 ("Staging: add comedi core")
Signed-off-by: Jann Horn <jannh@google.com>
---
Note: compile-tested only; I don't actually have comedi hardware, and I
don't know anything about comedi.
---
 drivers/comedi/comedi_fops.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/drivers/comedi/comedi_fops.c b/drivers/comedi/comedi_fops.c
index 1b481731df96..0e573df8646f 100644
--- a/drivers/comedi/comedi_fops.c
+++ b/drivers/comedi/comedi_fops.c
@@ -2414,6 +2414,15 @@ static int comedi_mmap(struct file *file, struct vm_area_struct *vma)
 		vma->vm_private_data = bm;
 
 		vma->vm_ops->open(vma);
+	} else {
+		/*
+		 * Leaving behind a partial mapping of a buffer we're about to
+		 * drop is unsafe, see remap_pfn_range_notrack().
+		 * We need to zap the range here ourselves instead of relying
+		 * on the automatic zapping in remap_pfn_range() because we call
+		 * remap_pfn_range() in a loop.
+		 */
+		zap_page_range_single(vma, vma->vm_start, size, NULL);
 	}
 
 done:

---
base-commit: 6485cf5ea253d40d507cd71253c9568c5470cd27
change-id: 20241014-comedi-tlb-400246505961
-- 
Jann Horn <jannh@google.com>

On 14/10/2024 21:50, Jann Horn wrote:
> If some remap_pfn_range() calls succeeded before one failed, we still have
> buffer pages mapped into the userspace page tables when we drop the buffer
> reference with comedi_buf_map_put(bm). The userspace mappings are only
> cleaned up later in the mmap error path.
> 
> Fix it by explicitly flushing all mappings in our VMA on the comedi_mmap()
> error path.
> 
> See commit 79a61cc3fc04 ("mm: avoid leaving partial pfn mappings around in
> error case").
> 
> Cc: stable@vger.kernel.org

Your patched version won't compile before 6.1 so you may want to 
indicate that in the Cc line.

> Fixes: ed9eccbe8970 ("Staging: add comedi core")
> Signed-off-by: Jann Horn <jannh@google.com>
> ---
> Note: compile-tested only; I don't actually have comedi hardware, and I
> don't know anything about comedi.
> ---
>   drivers/comedi/comedi_fops.c | 9 +++++++++
>   1 file changed, 9 insertions(+)
> 
> diff --git a/drivers/comedi/comedi_fops.c b/drivers/comedi/comedi_fops.c
> index 1b481731df96..0e573df8646f 100644
> --- a/drivers/comedi/comedi_fops.c
> +++ b/drivers/comedi/comedi_fops.c
> @@ -2414,6 +2414,15 @@ static int comedi_mmap(struct file *file, struct vm_area_struct *vma)
>   		vma->vm_private_data = bm;
>   
>   		vma->vm_ops->open(vma);
> +	} else {
> +		/*
> +		 * Leaving behind a partial mapping of a buffer we're about to
> +		 * drop is unsafe, see remap_pfn_range_notrack().
> +		 * We need to zap the range here ourselves instead of relying
> +		 * on the automatic zapping in remap_pfn_range() because we call
> +		 * remap_pfn_range() in a loop.
> +		 */
> +		zap_page_range_single(vma, vma->vm_start, size, NULL);
>   	}
>   
>   done:
> 
> ---
> base-commit: 6485cf5ea253d40d507cd71253c9568c5470cd27
> change-id: 20241014-comedi-tlb-400246505961

I guess this doesn't need to be done for the path that calls 
dma_mmap_coherent() since that does not do any range splitting. Would it 
be better to move it into the branch that calls remap_pfn_range() in a loop?

Note that I have no commit access to pulled-from repositories.  Greg-KH 
usually commits them on one of his repos, so could you Cc him too?  Thanks.

-- 
-=( Ian Abbott <abbotti@mev.co.uk> || MEV Ltd. is a company  )=-
-=( registered in England & Wales.  Regd. number: 02862268.  )=-
-=( Regd. addr.: S11 & 12 Building 67, Europa Business Park, )=-
-=( Bird Hall Lane, STOCKPORT, SK3 0XA, UK. || www.mev.co.uk )=-

On Tue, Oct 15, 2024 at 12:19 PM Ian Abbott <abbotti@mev.co.uk> wrote:
> On 14/10/2024 21:50, Jann Horn wrote:
> > If some remap_pfn_range() calls succeeded before one failed, we still have
> > buffer pages mapped into the userspace page tables when we drop the buffer
> > reference with comedi_buf_map_put(bm). The userspace mappings are only
> > cleaned up later in the mmap error path.
> >
> > Fix it by explicitly flushing all mappings in our VMA on the comedi_mmap()
> > error path.
> >
> > See commit 79a61cc3fc04 ("mm: avoid leaving partial pfn mappings around in
> > error case").
> >
> > Cc: stable@vger.kernel.org
>
> Your patched version won't compile before 6.1 so you may want to
> indicate that in the Cc line.

Ah, thanks for pointing that out - I can just use zap_vma_ptes()
instead, which is available in older kernels, that way it will be more
concise and the backport will be easier.

> > Fixes: ed9eccbe8970 ("Staging: add comedi core")
> > Signed-off-by: Jann Horn <jannh@google.com>
> > ---
> > Note: compile-tested only; I don't actually have comedi hardware, and I
> > don't know anything about comedi.
> > ---
> >   drivers/comedi/comedi_fops.c | 9 +++++++++
> >   1 file changed, 9 insertions(+)
> >
> > diff --git a/drivers/comedi/comedi_fops.c b/drivers/comedi/comedi_fops.c
> > index 1b481731df96..0e573df8646f 100644
> > --- a/drivers/comedi/comedi_fops.c
> > +++ b/drivers/comedi/comedi_fops.c
> > @@ -2414,6 +2414,15 @@ static int comedi_mmap(struct file *file, struct vm_area_struct *vma)
> >               vma->vm_private_data = bm;
> >
> >               vma->vm_ops->open(vma);
> > +     } else {
> > +             /*
> > +              * Leaving behind a partial mapping of a buffer we're about to
> > +              * drop is unsafe, see remap_pfn_range_notrack().
> > +              * We need to zap the range here ourselves instead of relying
> > +              * on the automatic zapping in remap_pfn_range() because we call
> > +              * remap_pfn_range() in a loop.
> > +              */
> > +             zap_page_range_single(vma, vma->vm_start, size, NULL);
> >       }
> >
> >   done:
> >
> > ---
> > base-commit: 6485cf5ea253d40d507cd71253c9568c5470cd27
> > change-id: 20241014-comedi-tlb-400246505961
>
> I guess this doesn't need to be done for the path that calls
> dma_mmap_coherent() since that does not do any range splitting. Would it
> be better to move it into the branch that calls remap_pfn_range() in a loop?

Sure, I'll move it up into the branch.

> Note that I have no commit access to pulled-from repositories.  Greg-KH
> usually commits them on one of his repos, so could you Cc him too?  Thanks.

Ack, will do.