Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1550434 > unrolled thread

[PATCH] virtio_blk: avoid DMA to stack for the sense buffer

Started byChristoph Hellwig <hch@lst.de>
First post2017-01-04 06:30 +0100
Last post2017-01-09 17:40 +0100
Articles 9 — 5 participants

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  [PATCH] virtio_blk: avoid DMA to stack for the sense buffer Christoph Hellwig <hch@lst.de> - 2017-01-04 06:30 +0100
    Re: [PATCH] virtio_blk: avoid DMA to stack for the sense buffer Jason Wang <jasowang@redhat.com> - 2017-01-04 08:50 +0100
      Re: [PATCH] virtio_blk: avoid DMA to stack for the sense buffer Christoph Hellwig <hch@lst.de> - 2017-01-09 14:40 +0100
        Re: [PATCH] virtio_blk: avoid DMA to stack for the sense buffer Jens Axboe <axboe@kernel.dk> - 2017-01-09 17:00 +0100
    Re: [PATCH] virtio_blk: avoid DMA to stack for the sense buffer 王金浦 <jinpuwang@gmail.com> - 2017-01-04 16:50 +0100
      Re: [PATCH] virtio_blk: avoid DMA to stack for the sense buffer Christoph Hellwig <hch@lst.de> - 2017-01-05 11:00 +0100
        Re: [PATCH] virtio_blk: avoid DMA to stack for the sense buffer 王金浦 <jinpuwang@gmail.com> - 2017-01-05 11:50 +0100
          Re: [PATCH] virtio_blk: avoid DMA to stack for the sense buffer Christoph Hellwig <hch@lst.de> - 2017-01-05 12:20 +0100
    Re: [PATCH] virtio_blk: avoid DMA to stack for the sense buffer "Michael S. Tsirkin" <mst@redhat.com> - 2017-01-09 17:40 +0100

#1550434 — [PATCH] virtio_blk: avoid DMA to stack for the sense buffer

FromChristoph Hellwig <hch@lst.de>
Date2017-01-04 06:30 +0100
Subject[PATCH] virtio_blk: avoid DMA to stack for the sense buffer
Message-ID<sVMj7-di-9@gated-at.bofh.it>
Most users of BLOCK_PC requests allocate the sense buffer on the stack,
so to avoid DMA to the stack copy them to a field in the heap allocated
virtblk_req structure.  Without that any attempt at SCSI passthrough I/O,
including the SG_IO ioctl from userspace will crash the kernel.  Note that
this includes running tools like hdparm even when the host does not have
SCSI passthrough enabled.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/block/virtio_blk.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/block/virtio_blk.c b/drivers/block/virtio_blk.c
index 5545a67..3c3b8f6 100644
--- a/drivers/block/virtio_blk.c
+++ b/drivers/block/virtio_blk.c
@@ -56,6 +56,7 @@ struct virtblk_req {
 	struct virtio_blk_outhdr out_hdr;
 	struct virtio_scsi_inhdr in_hdr;
 	u8 status;
+	u8 sense[SCSI_SENSE_BUFFERSIZE];
 	struct scatterlist sg[];
 };
 
@@ -102,7 +103,8 @@ static int __virtblk_add_req(struct virtqueue *vq,
 	}
 
 	if (type == cpu_to_virtio32(vq->vdev, VIRTIO_BLK_T_SCSI_CMD)) {
-		sg_init_one(&sense, vbr->req->sense, SCSI_SENSE_BUFFERSIZE);
+		memcpy(vbr->sense, vbr->req->sense, SCSI_SENSE_BUFFERSIZE);
+		sg_init_one(&sense, vbr->sense, SCSI_SENSE_BUFFERSIZE);
 		sgs[num_out + num_in++] = &sense;
 		sg_init_one(&inhdr, &vbr->in_hdr, sizeof(vbr->in_hdr));
 		sgs[num_out + num_in++] = &inhdr;
-- 
2.1.4

[toc] | [next] | [standalone]

#1550489

FromJason Wang <jasowang@redhat.com>
Date2017-01-04 08:50 +0100
Message-ID<sVOuB-1CW-3@gated-at.bofh.it>
In reply to#1550434

On 2017年01月04日 13:25, Christoph Hellwig wrote:
> Most users of BLOCK_PC requests allocate the sense buffer on the stack,
> so to avoid DMA to the stack copy them to a field in the heap allocated
> virtblk_req structure.  Without that any attempt at SCSI passthrough I/O,
> including the SG_IO ioctl from userspace will crash the kernel.  Note that
> this includes running tools like hdparm even when the host does not have
> SCSI passthrough enabled.
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>   drivers/block/virtio_blk.c | 4 +++-
>   1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/block/virtio_blk.c b/drivers/block/virtio_blk.c
> index 5545a67..3c3b8f6 100644
> --- a/drivers/block/virtio_blk.c
> +++ b/drivers/block/virtio_blk.c
> @@ -56,6 +56,7 @@ struct virtblk_req {
>   	struct virtio_blk_outhdr out_hdr;
>   	struct virtio_scsi_inhdr in_hdr;
>   	u8 status;
> +	u8 sense[SCSI_SENSE_BUFFERSIZE];
>   	struct scatterlist sg[];
>   };
>   
> @@ -102,7 +103,8 @@ static int __virtblk_add_req(struct virtqueue *vq,
>   	}
>   
>   	if (type == cpu_to_virtio32(vq->vdev, VIRTIO_BLK_T_SCSI_CMD)) {
> -		sg_init_one(&sense, vbr->req->sense, SCSI_SENSE_BUFFERSIZE);
> +		memcpy(vbr->sense, vbr->req->sense, SCSI_SENSE_BUFFERSIZE);
> +		sg_init_one(&sense, vbr->sense, SCSI_SENSE_BUFFERSIZE);
>   		sgs[num_out + num_in++] = &sense;
>   		sg_init_one(&inhdr, &vbr->in_hdr, sizeof(vbr->in_hdr));
>   		sgs[num_out + num_in++] = &inhdr;

Acked-by: Jason Wang <jasowang@redhat.com>

[toc] | [prev] | [next] | [standalone]

#1554312

FromChristoph Hellwig <hch@lst.de>
Date2017-01-09 14:40 +0100
Message-ID<sXIl4-4NO-9@gated-at.bofh.it>
In reply to#1550489
Is someone going to pick the patch up and send it to Linus?  I keep
running into all kinds of boot failures whenever I forget to cherry
pick it into my development trees..

[toc] | [prev] | [next] | [standalone]

#1554435

FromJens Axboe <axboe@kernel.dk>
Date2017-01-09 17:00 +0100
Message-ID<sXKwy-62l-7@gated-at.bofh.it>
In reply to#1554312
On 01/09/2017 06:35 AM, Christoph Hellwig wrote:
> Is someone going to pick the patch up and send it to Linus?  I keep
> running into all kinds of boot failures whenever I forget to cherry
> pick it into my development trees..

I'll add it.

-- 
Jens Axboe

[toc] | [prev] | [next] | [standalone]

#1550941

From王金浦 <jinpuwang@gmail.com>
Date2017-01-04 16:50 +0100
Message-ID<sVVZ7-6Ac-9@gated-at.bofh.it>
In reply to#1550434
Hi Christoph,

2017-01-04 6:25 GMT+01:00 Christoph Hellwig <hch@lst.de>:
> Most users of BLOCK_PC requests allocate the sense buffer on the stack,
> so to avoid DMA to the stack copy them to a field in the heap allocated
> virtblk_req structure.  Without that any attempt at SCSI passthrough I/O,
> including the SG_IO ioctl from userspace will crash the kernel.  Note that
> this includes running tools like hdparm even when the host does not have
> SCSI passthrough enabled.

This sounds scary.
Could you share how to reproduce it, this should go into stable if
it's the case.

Thanks,
Jinpu

>
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  drivers/block/virtio_blk.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/block/virtio_blk.c b/drivers/block/virtio_blk.c
> index 5545a67..3c3b8f6 100644
> --- a/drivers/block/virtio_blk.c
> +++ b/drivers/block/virtio_blk.c
> @@ -56,6 +56,7 @@ struct virtblk_req {
>         struct virtio_blk_outhdr out_hdr;
>         struct virtio_scsi_inhdr in_hdr;
>         u8 status;
> +       u8 sense[SCSI_SENSE_BUFFERSIZE];
>         struct scatterlist sg[];
>  };
>
> @@ -102,7 +103,8 @@ static int __virtblk_add_req(struct virtqueue *vq,
>         }
>
>         if (type == cpu_to_virtio32(vq->vdev, VIRTIO_BLK_T_SCSI_CMD)) {
> -               sg_init_one(&sense, vbr->req->sense, SCSI_SENSE_BUFFERSIZE);
> +               memcpy(vbr->sense, vbr->req->sense, SCSI_SENSE_BUFFERSIZE);
> +               sg_init_one(&sense, vbr->sense, SCSI_SENSE_BUFFERSIZE);
>                 sgs[num_out + num_in++] = &sense;
>                 sg_init_one(&inhdr, &vbr->in_hdr, sizeof(vbr->in_hdr));
>                 sgs[num_out + num_in++] = &inhdr;
> --
> 2.1.4
>

[toc] | [prev] | [next] | [standalone]

#1551836

FromChristoph Hellwig <hch@lst.de>
Date2017-01-05 11:00 +0100
Message-ID<sWd03-1dl-19@gated-at.bofh.it>
In reply to#1550941
On Wed, Jan 04, 2017 at 04:47:03PM +0100, 王金浦 wrote:
> This sounds scary.
> Could you share how to reproduce it, this should go into stable if
> it's the case.

Step 1: Build your kernel with CONFIG_VMAP_STACK=y
Step 2: issue a SG_IO ioctl, e.g. sg_inq /dev/vda

[toc] | [prev] | [next] | [standalone]

#1551875

From王金浦 <jinpuwang@gmail.com>
Date2017-01-05 11:50 +0100
Message-ID<sWdMm-1OF-3@gated-at.bofh.it>
In reply to#1551836
2017-01-05 10:57 GMT+01:00 Christoph Hellwig <hch@lst.de>:
> On Wed, Jan 04, 2017 at 04:47:03PM +0100, 王金浦 wrote:
>> This sounds scary.
>> Could you share how to reproduce it, this should go into stable if
>> it's the case.
>
> Step 1: Build your kernel with CONFIG_VMAP_STACK=y
> Step 2: issue a SG_IO ioctl, e.g. sg_inq /dev/vda
>
Thanks, so it's only relevant to kernel > 4.9, as  CONFIG_VMAP_STACK
only introduced in 4.9 kernel.

Regards,
Jinpu

[toc] | [prev] | [next] | [standalone]

#1551898

FromChristoph Hellwig <hch@lst.de>
Date2017-01-05 12:20 +0100
Message-ID<sWefo-2gj-7@gated-at.bofh.it>
In reply to#1551875
On Thu, Jan 05, 2017 at 11:37:46AM +0100, 王金浦 wrote:
> Thanks, so it's only relevant to kernel > 4.9, as  CONFIG_VMAP_STACK
> only introduced in 4.9 kernel.

kernel >= 4.9, but otherwise, yes.

[toc] | [prev] | [next] | [standalone]

#1554470

From"Michael S. Tsirkin" <mst@redhat.com>
Date2017-01-09 17:40 +0100
Message-ID<sXL9h-6vS-39@gated-at.bofh.it>
In reply to#1550434
On Wed, Jan 04, 2017 at 08:25:05AM +0300, Christoph Hellwig wrote:
> Most users of BLOCK_PC requests allocate the sense buffer on the stack,
> so to avoid DMA to the stack copy them to a field in the heap allocated
> virtblk_req structure.  Without that any attempt at SCSI passthrough I/O,
> including the SG_IO ioctl from userspace will crash the kernel.  Note that
> this includes running tools like hdparm even when the host does not have
> SCSI passthrough enabled.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>  drivers/block/virtio_blk.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/block/virtio_blk.c b/drivers/block/virtio_blk.c
> index 5545a67..3c3b8f6 100644
> --- a/drivers/block/virtio_blk.c
> +++ b/drivers/block/virtio_blk.c
> @@ -56,6 +56,7 @@ struct virtblk_req {
>  	struct virtio_blk_outhdr out_hdr;
>  	struct virtio_scsi_inhdr in_hdr;
>  	u8 status;
> +	u8 sense[SCSI_SENSE_BUFFERSIZE];
>  	struct scatterlist sg[];
>  };
>  
> @@ -102,7 +103,8 @@ static int __virtblk_add_req(struct virtqueue *vq,
>  	}
>  
>  	if (type == cpu_to_virtio32(vq->vdev, VIRTIO_BLK_T_SCSI_CMD)) {
> -		sg_init_one(&sense, vbr->req->sense, SCSI_SENSE_BUFFERSIZE);
> +		memcpy(vbr->sense, vbr->req->sense, SCSI_SENSE_BUFFERSIZE);
> +		sg_init_one(&sense, vbr->sense, SCSI_SENSE_BUFFERSIZE);

I would prefer sizeof(vbr->sense) here to avoid duplication.
Otherwise:

Acked-by: Michael S. Tsirkin <mst@redhat.com>


>  		sgs[num_out + num_in++] = &sense;
>  		sg_init_one(&inhdr, &vbr->in_hdr, sizeof(vbr->in_hdr));
>  		sgs[num_out + num_in++] = &inhdr;
> -- 
> 2.1.4

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web