netdev - Re: [Bugme-new] [Bug 38102] New: BUG kmalloc-2048: Poison overwritten

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]

Date:	Wed, 06 Jul 2011 18:56:19 +0200
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	Michael Büsch <m@...s.ch>
Cc:	Neil Horman <nhorman@...driver.com>,
	Alexey Zaytsev <alexey.zaytsev@...il.com>,
	Andrew Morton <akpm@...ux-foundation.org>,
	netdev@...r.kernel.org, Gary Zambrano <zambrano@...adcom.com>,
	bugme-daemon@...zilla.kernel.org,
	"David S. Miller" <davem@...emloft.net>,
	Pekka Pietikainen <pp@...oulu.fi>,
	Florian Schirmer <jolt@...box.org>,
	Felix Fietkau <nbd@...nwrt.org>, Michael Buesch <mb@...sch.de>
Subject: Re: [Bugme-new] [Bug 38102] New: BUG kmalloc-2048: Poison
 overwritten

Le mercredi 06 juillet 2011 à 17:32 +0200, Michael Büsch a écrit :

> You guys are mixing up quite a bit of stuff here...

Well

> 
> The EOT bit has _nothing_ to do with the descriptor pointers.
> It simply marks the last descriptor in the (linear) descriptor
> page, so that it becomes an actual ring:
> 
>    DDDDDDDDDDDDDDDDDDDDDDDDDDDE
>    |                          O
>    |                          T
>    ^--------------------------|
> 
> It doesn't say anything about the read and write pointers
> to the ring.
> 
> The B44_DMARX_PTR is the write-end pointer. It points one entry
> beyond the end of the write area. Then there's the software pointer
> where we keep track of the read position.
> 


Thats not how b44_rx() works :

It writes on DMARX_PTR the last slot that driver _dequeued_ in its NAPI
run. Its not the end of the window that device is allowed to use.

bw32(bp, B44_DMARX_PTR, cons * sizeof(struct dma_desc));

The end of the 'allocated buffers' is in rx_prod. Problem is NIC have no
idea of where is the end of window. We never give rx_prod to NIC.

So NIC actually read old descriptors value. We need to clear them to
avoid memory corruption.


diff --git a/drivers/net/b44.c b/drivers/net/b44.c
index 6c4ef96..ec9773b 100644
--- a/drivers/net/b44.c
+++ b/drivers/net/b44.c
@@ -725,6 +725,7 @@ static void b44_recycle_rx(struct b44 *bp, int src_idx, u32 dest_idx_unmasked)
 			                 DMA_BIDIRECTIONAL);
 
 	ctrl = src_desc->ctrl;
+	src_desc->ctrl = ctrl & cpu_to_le32(DESC_CTRL_EOT);
 	if (dest_idx == (B44_RX_RING_SIZE - 1))
 		ctrl |= cpu_to_le32(DESC_CTRL_EOT);
 	else
@@ -732,6 +733,7 @@ static void b44_recycle_rx(struct b44 *bp, int src_idx, u32 dest_idx_unmasked)
 
 	dest_desc->ctrl = ctrl;
 	dest_desc->addr = src_desc->addr;
+	src_desc->addr = 0;
 
 	src_map->skb = NULL;
 
@@ -1118,6 +1120,7 @@ static void b44_init_rings(struct b44 *bp)
 		if (b44_alloc_rx_skb(bp, -1, i) < 0)
 			break;
 	}
+	bp->rx_prod = i;
 }
 
 /*
@@ -1406,7 +1409,6 @@ static void b44_init_hw(struct b44 *bp, int reset_kind)
 		bw32(bp, B44_DMARX_ADDR, bp->rx_ring_dma + bp->dma_offset);
 
 		bw32(bp, B44_DMARX_PTR, bp->rx_pending);
-		bp->rx_prod = bp->rx_pending;
 
 		bw32(bp, B44_MIB_CTRL, MIB_CTRL_CLR_ON_READ);
 	}



--
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