Patchwork [net] net: Fix page seeking for skb_splice_bits().

login
register
mail settings
Submitter Jarek Poplawski
Date Feb. 11, 2009, 9:58 p.m.
Message ID <20090211215806.GA2676@ami.dom.local>
Download mbox | patch
Permalink /patch/22958/
State Accepted
Delegated to: David Miller
Headers show

Comments

Jarek Poplawski - Feb. 11, 2009, 9:58 p.m.
Hi,

So, you both were right there is something wrong with splice for
multipage linear regions, at least theoretically. Since I don't know
if this can really matter in practice (could NUMA slab allocate these
pages from different tables?), I've done it against -net, but maybe
-net-next is enough?

Thanks,
Jarek P.
------------->
net: Fix page seeking for skb_splice_bits().

struct page walking should be done with proper accessor functions, not
directly.

With doubts from David S. Miller and Herbert Xu.

Signed-off-by: Jarek Poplawski <jarkao2@gmail.com>
---

 net/core/skbuff.c |    7 ++++++-
 1 files changed, 6 insertions(+), 1 deletions(-)


--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
David Miller - Feb. 13, 2009, 12:50 a.m.
From: Jarek Poplawski <jarkao2@gmail.com>
Date: Wed, 11 Feb 2009 22:58:06 +0100

> So, you both were right there is something wrong with splice for
> multipage linear regions, at least theoretically. Since I don't know
> if this can really matter in practice (could NUMA slab allocate these
> pages from different tables?), I've done it against -net, but maybe
> -net-next is enough?

I'll add this to net-next-2.6 because I think it's theoretical at
best.  In fact I don't think we would get allocations across
NUMA nodes in this situation.

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

Patch

diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index da74b84..7aacd95 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -1373,8 +1373,13 @@  static inline int spd_fill_page(struct splice_pipe_desc *spd, struct page *page,
 static inline void __segment_seek(struct page **page, unsigned int *poff,
 				  unsigned int *plen, unsigned int off)
 {
+	unsigned long n;
+
 	*poff += off;
-	*page += *poff / PAGE_SIZE;
+	n = *poff / PAGE_SIZE;
+	if (n)
+		*page = nth_page(*page, n);
+
 	*poff = *poff % PAGE_SIZE;
 	*plen -= off;
 }