linux-kernel - Re: [PATCH v4 2/5] rpmsg: virtio_rpmsg_bus: fix sg_set_buf() when addr is not a valid kernel address

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <15f1e2b0-877d-9385-96e1-44f2c49aa6bb@ti.com>
Date:   Thu, 24 Aug 2017 17:03:52 -0500
From:   Suman Anna <s-anna@...com>
To:     Loic Pallardy <loic.pallardy@...com>, <bjorn.andersson@...aro.org>,
        <ohad@...ery.com>, <lee.jones@...aro.org>
CC:     <linux-remoteproc@...r.kernel.org>, <linux-kernel@...r.kernel.org>,
        <kernel@...inux.com>, <patrice.chotard@...com>,
        <hugues.fruchet@...com>, Ludovic Barre <ludovic.barre@...com>
Subject: Re: [PATCH v4 2/5] rpmsg: virtio_rpmsg_bus: fix sg_set_buf() when
 addr is not a valid kernel address

Hi Bjorn,

On 03/28/2017 06:49 AM, Loic Pallardy wrote:
> To specify memory for remoteproc, we declare (dma_declare_coherent_memory())
> an area which is ioremap'ed to the vmalloc area.  However, this address is
> not a kernel address so virt_addr_valid(buf) fails.
> 
> Signed-off-by: Ludovic Barre <ludovic.barre@...com>
> Signed-off-by: Loic Pallardy <loic.pallardy@...com>
> Acked-by: Patrice Chotard <patrice.chotard@...com>
> Tested-by: Suman Anna <s-anna@...com>

Can you pick up the first 2 patches from this series for v4.14? This
does enable us to use HighMem CMA pools or carveout pools with
virtio_rpmsg_bus, and have been sitting on the lists for some time in
Acked/Tested status.

regards
Suman

> ---
> Changes since V1:
> - Remove extra line.
> 
> No change since v2.
> 
> ---
>  drivers/rpmsg/virtio_rpmsg_bus.c | 28 +++++++++++++++++++++++++---
>  1 file changed, 25 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/rpmsg/virtio_rpmsg_bus.c b/drivers/rpmsg/virtio_rpmsg_bus.c
> index 3f44a03..1c7cde9 100644
> --- a/drivers/rpmsg/virtio_rpmsg_bus.c
> +++ b/drivers/rpmsg/virtio_rpmsg_bus.c
> @@ -195,6 +195,28 @@ static int virtio_rpmsg_trysend_offchannel(struct rpmsg_endpoint *ept, u32 src,
>  };
>  
>  /**
> + * rpmsg_sg_init - initialize scatterlist according to cpu address location
> + * @sg: scatterlist to fill
> + * @cpu_addr: virtual address of the buffer
> + * @len: buffer length
> + *
> + * An internal function filling scatterlist according to virtual address
> + * location (in vmalloc or in kernel).
> + */
> +static void
> +rpmsg_sg_init(struct scatterlist *sg, void *cpu_addr, unsigned int len)
> +{
> +	if (is_vmalloc_addr(cpu_addr)) {
> +		sg_init_table(sg, 1);
> +		sg_set_page(sg, vmalloc_to_page(cpu_addr), len,
> +			    offset_in_page(cpu_addr));
> +	} else {
> +		WARN_ON(!virt_addr_valid(cpu_addr));
> +		sg_init_one(sg, cpu_addr, len);
> +	}
> +}
> +
> +/**
>   * __ept_release() - deallocate an rpmsg endpoint
>   * @kref: the ept's reference count
>   *
> @@ -606,7 +628,7 @@ static int rpmsg_send_offchannel_raw(struct rpmsg_device *rpdev,
>  			 msg, sizeof(*msg) + msg->len, true);
>  #endif
>  
> -	sg_init_one(&sg, msg, sizeof(*msg) + len);
> +	rpmsg_sg_init(&sg, msg, sizeof(*msg) + len);
>  
>  	mutex_lock(&vrp->tx_lock);
>  
> @@ -731,7 +753,7 @@ static int rpmsg_recv_single(struct virtproc_info *vrp, struct device *dev,
>  		dev_warn(dev, "msg received with no recipient\n");
>  
>  	/* publish the real size of the buffer */
> -	sg_init_one(&sg, msg, vrp->buf_size);
> +	rpmsg_sg_init(&sg, msg, vrp->buf_size);
>  
>  	/* add the buffer back to the remote processor's virtqueue */
>  	err = virtqueue_add_inbuf(vrp->rvq, &sg, 1, msg, GFP_KERNEL);
> @@ -915,7 +937,7 @@ static int rpmsg_probe(struct virtio_device *vdev)
>  		struct scatterlist sg;
>  		void *cpu_addr = vrp->rbufs + i * vrp->buf_size;
>  
> -		sg_init_one(&sg, cpu_addr, vrp->buf_size);
> +		rpmsg_sg_init(&sg, cpu_addr, vrp->buf_size);
>  
>  		err = virtqueue_add_inbuf(vrp->rvq, &sg, 1, cpu_addr,
>  					  GFP_KERNEL);
>