[PATCH 2.6-BK] Fix stray pointer in e100

[PATCH 2.6-BK] Fix stray pointer in e100

[Deepak Saxena]

[-- Attachment #1: Type: text/plain, Size: 401 bytes --]


e100_alloc_cbs() allocates the cb's but does not set cb->skb = NULL
which means that the following check in e100_tx_clean() will execute
even though cb->skb is not really a valid pointer an we OOPs:

	if(likely(cb->skb != NULL)) {
		...
		nic->net_stats.tx_bytes += cb->skb->len;

	}

Attached patch fixes the issue.

~Deepak

-- 
Deepak Saxena - dsaxena at plexity dot net - http://www.plexity.net/

[-- Attachment #2: patch-e100-cb-alloc --]
[-- Type: text/plain, Size: 414 bytes --]

diff -Nru a/drivers/net/e100.c b/drivers/net/e100.c
--- a/drivers/net/e100.c	Fri Mar 12 14:26:45 2004
+++ b/drivers/net/e100.c	Fri Mar 12 14:26:45 2004
@@ -1346,6 +1346,7 @@
 		cb->dma_addr = nic->cbs_dma_addr + i * sizeof(struct cb);
 		cb->link = cpu_to_le32(nic->cbs_dma_addr +
 			((i+1) % count) * sizeof(struct cb));
+		cb->skb = NULL;
 	}
 
 	nic->cb_to_use = nic->cb_to_send = nic->cb_to_clean = nic->cbs;

RE: [PATCH 2.6-BK] Fix stray pointer in e100

[Feldman, Scott]

> e100_alloc_cbs() allocates the cb's but does not set cb->skb 
> = NULL which means that the following check in 
> e100_tx_clean() will execute even though cb->skb is not 
> really a valid pointer an we OOPs:

Thanks Deepak.  I found one other place where we need to set
cb->skb=NULL.  Patch sent to Jeff upstream.

-scott