[net-next PATCH] net: Reserve skb headroom and set skb->dev even if using __alloc_skb

[net-next PATCH] net: Reserve skb headroom and set skb->dev even if using __alloc_skb

[Alexander Duyck]

When I had inlined __alloc_rx_skb into __netdev_alloc_skb and
__napi_alloc_skb I had overlooked the fact that there was a return in the
__alloc_rx_skb.  As a result we weren't reserving headroom or setting the
skb->dev in certain cases.  This change corrects that by adding a couple of
jump labels to jump to depending on __alloc_skb either succeeding or failing.

Fixes: 9451980a6646 ("net: Use cached copy of pfmemalloc to avoid accessing page")
Reported-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
---
 net/core/skbuff.c |   20 ++++++++++++++++----
 1 file changed, 16 insertions(+), 4 deletions(-)

diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index d67e612bf0ef..f3fe9bd9e672 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -414,8 +414,12 @@ struct sk_buff *__netdev_alloc_skb(struct net_device *dev, unsigned int len,
 	len += NET_SKB_PAD;
 
 	if ((len > SKB_WITH_OVERHEAD(PAGE_SIZE)) ||
-	    (gfp_mask & (__GFP_WAIT | GFP_DMA)))
-		return __alloc_skb(len, gfp_mask, SKB_ALLOC_RX, NUMA_NO_NODE);
+	    (gfp_mask & (__GFP_WAIT | GFP_DMA))) {
+		skb = __alloc_skb(len, gfp_mask, SKB_ALLOC_RX, NUMA_NO_NODE);
+		if (!skb)
+			goto skb_fail;
+		goto skb_success;
+	}
 
 	len += SKB_DATA_ALIGN(sizeof(struct skb_shared_info));
 	len = SKB_DATA_ALIGN(len);
@@ -445,9 +449,11 @@ struct sk_buff *__netdev_alloc_skb(struct net_device *dev, unsigned int len,
 		skb->pfmemalloc = 1;
 	skb->head_frag = 1;
 
+skb_success:
 	skb_reserve(skb, NET_SKB_PAD);
 	skb->dev = dev;
 
+skb_fail:
 	return skb;
 }
 EXPORT_SYMBOL(__netdev_alloc_skb);
@@ -475,8 +481,12 @@ struct sk_buff *__napi_alloc_skb(struct napi_struct *napi, unsigned int len,
 	len += NET_SKB_PAD + NET_IP_ALIGN;
 
 	if ((len > SKB_WITH_OVERHEAD(PAGE_SIZE)) ||
-	    (gfp_mask & (__GFP_WAIT | GFP_DMA)))
-		return __alloc_skb(len, gfp_mask, SKB_ALLOC_RX, NUMA_NO_NODE);
+	    (gfp_mask & (__GFP_WAIT | GFP_DMA))) {
+		skb = __alloc_skb(len, gfp_mask, SKB_ALLOC_RX, NUMA_NO_NODE);
+		if (!skb)
+			goto skb_fail;
+		goto skb_success;
+	}
 
 	len += SKB_DATA_ALIGN(sizeof(struct skb_shared_info));
 	len = SKB_DATA_ALIGN(len);
@@ -499,9 +509,11 @@ struct sk_buff *__napi_alloc_skb(struct napi_struct *napi, unsigned int len,
 		skb->pfmemalloc = 1;
 	skb->head_frag = 1;
 
+skb_success:
 	skb_reserve(skb, NET_SKB_PAD + NET_IP_ALIGN);
 	skb->dev = napi->dev;
 
+skb_fail:
 	return skb;
 }
 EXPORT_SYMBOL(__napi_alloc_skb);

Re: [net-next PATCH] net: Reserve skb headroom and set skb->dev even if using __alloc_skb

[Kevin Hilman]

Alexander Duyck <alexander.h.duyck@redhat.com> writes:

> When I had inlined __alloc_rx_skb into __netdev_alloc_skb and
> __napi_alloc_skb I had overlooked the fact that there was a return in the
> __alloc_rx_skb.  As a result we weren't reserving headroom or setting the
> skb->dev in certain cases.  This change corrects that by adding a couple of
> jump labels to jump to depending on __alloc_skb either succeeding or failing.
>
> Fixes: 9451980a6646 ("net: Use cached copy of pfmemalloc to avoid accessing page")
> Reported-by: Felipe Balbi <balbi@ti.com>
> Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>

Tested this on top of next-20150513 on an ARM/OMAP
(am335x-boneblack.dts) an it fixes the boot problem for me.

Tested-by: Kevin Hilman <khilman@linaro.org>

Kevin

Re: [net-next PATCH] net: Reserve skb headroom and set skb->dev even if using __alloc_skb

[Felipe Balbi]

[-- Attachment #1: Type: text/plain, Size: 1015 bytes --]

On Wed, May 13, 2015 at 02:35:51PM -0700, Kevin Hilman wrote:
> Alexander Duyck <alexander.h.duyck@redhat.com> writes:
> 
> > When I had inlined __alloc_rx_skb into __netdev_alloc_skb and
> > __napi_alloc_skb I had overlooked the fact that there was a return in the
> > __alloc_rx_skb.  As a result we weren't reserving headroom or setting the
> > skb->dev in certain cases.  This change corrects that by adding a couple of
> > jump labels to jump to depending on __alloc_skb either succeeding or failing.
> >
> > Fixes: 9451980a6646 ("net: Use cached copy of pfmemalloc to avoid accessing page")
> > Reported-by: Felipe Balbi <balbi@ti.com>
> > Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
> 
> Tested this on top of next-20150513 on an ARM/OMAP
> (am335x-boneblack.dts) an it fixes the boot problem for me.
> 
> Tested-by: Kevin Hilman <khilman@linaro.org>

Yeah, I know it's too late, but I also tested on my AM437x SK.

Tested-by: Felipe Balbi <balbi@ti.com>

-- 
balbi

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

Re: [net-next PATCH] net: Reserve skb headroom and set skb->dev even if using __alloc_skb

[David Miller]

From: Alexander Duyck <alexander.h.duyck@redhat.com>
Date: Wed, 13 May 2015 13:34:13 -0700

> When I had inlined __alloc_rx_skb into __netdev_alloc_skb and
> __napi_alloc_skb I had overlooked the fact that there was a return in the
> __alloc_rx_skb.  As a result we weren't reserving headroom or setting the
> skb->dev in certain cases.  This change corrects that by adding a couple of
> jump labels to jump to depending on __alloc_skb either succeeding or failing.
> 
> Fixes: 9451980a6646 ("net: Use cached copy of pfmemalloc to avoid accessing page")
> Reported-by: Felipe Balbi <balbi@ti.com>
> Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>

Applied, thanks.