netdev - Re: [PATCH 4/4] bgmac: fix DMA rx corruption

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

Message-ID: <1428834675.25985.344.camel@edumazet-glaptop2.roam.corp.google.com>
Date:	Sun, 12 Apr 2015 03:31:15 -0700
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	Felix Fietkau <nbd@...nwrt.org>
Cc:	netdev@...r.kernel.org, zajec5@...il.com, hauke@...ke-m.de
Subject: Re: [PATCH 4/4] bgmac: fix DMA rx corruption

On Sun, 2015-04-12 at 12:08 +0200, Felix Fietkau wrote:
> The driver needs to inform the hardware about the first invalid (not yet
> filled) rx slot, by writing its DMA descriptor pointer offset to the
> BGMAC_DMA_RX_INDEX register.
> 
> This register was set to a value exceeding the rx ring size, effectively
> allowing the hardware constant access to the full ring, regardless of
> which slots are initialized.
> 
> Fix this by updating the register in bgmac_dma_rx_setup_desc.
> 
> Signed-off-by: Felix Fietkau <nbd@...nwrt.org>
> ---
>  drivers/net/ethernet/broadcom/bgmac.c | 14 +++++++-------
>  1 file changed, 7 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/net/ethernet/broadcom/bgmac.c b/drivers/net/ethernet/broadcom/bgmac.c
> index e332de8..856ceee 100644
> --- a/drivers/net/ethernet/broadcom/bgmac.c
> +++ b/drivers/net/ethernet/broadcom/bgmac.c
> @@ -380,6 +380,12 @@ static void bgmac_dma_rx_setup_desc(struct bgmac *bgmac,
>  	dma_desc->addr_high = cpu_to_le32(upper_32_bits(ring->slots[desc_idx].dma_addr));
>  	dma_desc->ctl0 = cpu_to_le32(ctl0);
>  	dma_desc->ctl1 = cpu_to_le32(ctl1);
> +
> +	desc_idx = (desc_idx + 1) % BGMAC_RX_RING_SLOTS;
> +
> +	bgmac_write(bgmac, ring->mmio_base + BGMAC_DMA_RX_INDEX,
> +			ring->index_base +
> +			desc_idx * sizeof(struct bgmac_dma_desc));
>  }
>  
>  static int bgmac_dma_rx_read(struct bgmac *bgmac, struct bgmac_dma_ring *ring,
> @@ -394,9 +400,7 @@ static int bgmac_dma_rx_read(struct bgmac *bgmac, struct bgmac_dma_ring *ring,
>  	end_slot &= BGMAC_DMA_RX_STATDPTR;
>  	end_slot /= sizeof(struct bgmac_dma_desc);
>  
> -	ring->end = end_slot;
> -
> -	while (ring->start != ring->end) {
> +	while (ring->start != end_slot) {
>  		struct device *dma_dev = bgmac->core->dma_dev;
>  		struct bgmac_slot_info *slot = &ring->slots[ring->start];
>  		struct bgmac_rx_header *rx = slot->buf + BGMAC_RX_BUF_OFFSET;
> @@ -693,10 +697,6 @@ static void bgmac_dma_init(struct bgmac *bgmac)
>  		for (j = 0; j < ring->num_slots; j++)
>  			bgmac_dma_rx_setup_desc(bgmac, ring, j);
>  

Missing dma_wmb() here  (or legacy wmb() for stable kernels)

> -		bgmac_write(bgmac, ring->mmio_base + BGMAC_DMA_RX_INDEX,
> -			    ring->index_base +
> -			    ring->num_slots * sizeof(struct bgmac_dma_desc));
> -


This might be better for performance to perform one single bgmac_write()
at the end of bgmac_dma_rx_read(), and leave this one in place as well,
not for performance since this is slow path, but correctness.

Also I am surprised there is no memory barrier to make sure changes to
dma_desc are committed to memory ?

I would use dma_wmb() before bgmac_write(), like the wmb() done in
bgmac_dma_tx_add()

Thanks



--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html