Re: [PATCH] dmaengine: sun6i: Fix use after free

[Vinod Koul]

On 07-11-19, 19:26, Pan Bian wrote:
> The members in the LLI list is released in an incorrect way. Read and
> store the next member before releasing it to avoid accessing the freed
> memory.
> 
> Fixes: a90e173f3faf ("dmaengine: sun6i: Add cyclic capability")
> 
> Signed-off-by: Pan Bian <bianpan2016@xxxxxxx>
> ---
>  drivers/dma/sun6i-dma.c | 20 ++++++++++++++------
>  1 file changed, 14 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/dma/sun6i-dma.c b/drivers/dma/sun6i-dma.c
> index 06cd7f867f7c..096aad7e75bb 100644
> --- a/drivers/dma/sun6i-dma.c
> +++ b/drivers/dma/sun6i-dma.c
> @@ -687,7 +687,7 @@ static struct dma_async_tx_descriptor *sun6i_dma_prep_slave_sg(
>  	struct sun6i_dma_dev *sdev = to_sun6i_dma_dev(chan->device);
>  	struct sun6i_vchan *vchan = to_sun6i_vchan(chan);
>  	struct dma_slave_config *sconfig = &vchan->cfg;
> -	struct sun6i_dma_lli *v_lli, *prev = NULL;
> +	struct sun6i_dma_lli *v_lli, *next, *prev = NULL;
>  	struct sun6i_desc *txd;
>  	struct scatterlist *sg;
>  	dma_addr_t p_lli;
> @@ -752,8 +752,12 @@ static struct dma_async_tx_descriptor *sun6i_dma_prep_slave_sg(
>  	return vchan_tx_prep(&vchan->vc, &txd->vd, flags);
>  
>  err_lli_free:
> -	for (prev = txd->v_lli; prev; prev = prev->v_lli_next)
> -		dma_pool_free(sdev->pool, prev, virt_to_phys(prev));
> +	v_lli = txd->v_lli;
> +	while (v_lli) {
> +		next = v_lli->v_lli_next;
> +		dma_pool_free(sdev->pool, v_lli, virt_to_phys(v_lli));
> +		v_lli = next;

Have you seen this issue, a panic trace? Has some static checker flagged
this?

I see the code *seems* equivalent. The prev is assigned to txd->v_lli,
checked and then code block will be executed and updated. When do you
see the case when we access freed memory?

-- 
~Vinod