Patchwork [net-next] myri10ge: fix truesize underestimation

login
register
mail settings
Submitter Eric Dumazet
Date Oct. 20, 2011, 8:10 p.m.
Message ID <1319141403.2854.17.camel@edumazet-laptop>
Download mbox | patch
Permalink /patch/120870/
State Accepted
Delegated to: David Miller
Headers show

Comments

Eric Dumazet - Oct. 20, 2011, 8:10 p.m.
skb->truesize must account for allocated memory, not the used part of
it. Doing this work is important to avoid unexpected OOM situations.

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
CC: Jon Mason <mason@myri.com>
---
 drivers/net/ethernet/myricom/myri10ge/myri10ge.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)



--
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
Jon Mason - Oct. 20, 2011, 8:33 p.m.
On Thu, Oct 20, 2011 at 3:10 PM, Eric Dumazet <eric.dumazet@gmail.com> wrote:
> skb->truesize must account for allocated memory, not the used part of
> it. Doing this work is important to avoid unexpected OOM situations.
>
> Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>

Acked-by: Jon Mason <mason@myri.com>

> CC: Jon Mason <mason@myri.com>
> ---
>  drivers/net/ethernet/myricom/myri10ge/myri10ge.c |    3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/net/ethernet/myricom/myri10ge/myri10ge.c b/drivers/net/ethernet/myricom/myri10ge/myri10ge.c
> index c970a48..0778edc 100644
> --- a/drivers/net/ethernet/myricom/myri10ge/myri10ge.c
> +++ b/drivers/net/ethernet/myricom/myri10ge/myri10ge.c
> @@ -1210,7 +1210,6 @@ myri10ge_rx_skb_build(struct sk_buff *skb, u8 * va,
>        struct skb_frag_struct *skb_frags;
>
>        skb->len = skb->data_len = len;
> -       skb->truesize = len + sizeof(struct sk_buff);
>        /* attach the page(s) */
>
>        skb_frags = skb_shinfo(skb)->frags;
> @@ -1385,6 +1384,8 @@ myri10ge_rx_done(struct myri10ge_slice_state *ss, int len, __wsum csum,
>        if (skb_frag_size(&skb_shinfo(skb)->frags[0]) <= 0) {
>                skb_frag_unref(skb, 0);
>                skb_shinfo(skb)->nr_frags = 0;
> +       } else {
> +               skb->truesize += bytes * skb_shinfo(skb)->nr_frags;
>        }
>        skb->protocol = eth_type_trans(skb, dev);
>        skb_record_rx_queue(skb, ss - &mgp->ss[0]);
>
>
>
--
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
Eric Dumazet - Oct. 20, 2011, 8:44 p.m.
Le jeudi 20 octobre 2011 à 15:33 -0500, Jon Mason a écrit :
> On Thu, Oct 20, 2011 at 3:10 PM, Eric Dumazet <eric.dumazet@gmail.com> wrote:
> > skb->truesize must account for allocated memory, not the used part of
> > it. Doing this work is important to avoid unexpected OOM situations.
> >
> > Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
> 
> Acked-by: Jon Mason <mason@myri.com>

Thanks for reviewing Jon !


--
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
Andrew Gallatin - Oct. 20, 2011, 8:45 p.m.
On 10/20/11 16:44, Eric Dumazet wrote:
> Le jeudi 20 octobre 2011 à 15:33 -0500, Jon Mason a écrit :
>> On Thu, Oct 20, 2011 at 3:10 PM, Eric Dumazet<eric.dumazet@gmail.com>  wrote:
>>> skb->truesize must account for allocated memory, not the used part of
>>> it. Doing this work is important to avoid unexpected OOM situations.
>>>
>>> Signed-off-by: Eric Dumazet<eric.dumazet@gmail.com>
>>
>> Acked-by: Jon Mason<mason@myri.com>
>
> Thanks for reviewing Jon !
>
>

Please wait a second..  I think the patch is incorrect.

There is already code in myri10ge_rx_skb_build() which
attempts to set the truesize.  However, it sets it to
the used, rather than the allocated size so it is apparently
incorrect.

I'd prefer we fix that code.

Thanks,

Drew
--
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
Eric Dumazet - Oct. 20, 2011, 8:59 p.m.
Le jeudi 20 octobre 2011 à 16:45 -0400, Andrew Gallatin a écrit :
> On 10/20/11 16:44, Eric Dumazet wrote:
> > Le jeudi 20 octobre 2011 à 15:33 -0500, Jon Mason a écrit :
> >> On Thu, Oct 20, 2011 at 3:10 PM, Eric Dumazet<eric.dumazet@gmail.com>  wrote:
> >>> skb->truesize must account for allocated memory, not the used part of
> >>> it. Doing this work is important to avoid unexpected OOM situations.
> >>>
> >>> Signed-off-by: Eric Dumazet<eric.dumazet@gmail.com>
> >>
> >> Acked-by: Jon Mason<mason@myri.com>
> >
> > Thanks for reviewing Jon !
> >
> >
> 
> Please wait a second..  I think the patch is incorrect.
> 
> There is already code in myri10ge_rx_skb_build() which
> attempts to set the truesize.  However, it sets it to
> the used, rather than the allocated size so it is apparently
> incorrect.
> 
> I'd prefer we fix that code.

Well, I believe I did exactly that :)

truesize of initial skb is fine.

Then for everay frag added, you must add to skb-truesize the allocated
memory for this frag.

You add frags of a given size (small or big)

In the end, its truesize += bytes * number_of_frags

(bytes being small_size or big_size)


--
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 - Oct. 20, 2011, 9:03 p.m.
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Thu, 20 Oct 2011 22:59:16 +0200

> Le jeudi 20 octobre 2011 à 16:45 -0400, Andrew Gallatin a écrit :
>> On 10/20/11 16:44, Eric Dumazet wrote:
>> > Le jeudi 20 octobre 2011 à 15:33 -0500, Jon Mason a écrit :
>> >> On Thu, Oct 20, 2011 at 3:10 PM, Eric Dumazet<eric.dumazet@gmail.com>  wrote:
>> >>> skb->truesize must account for allocated memory, not the used part of
>> >>> it. Doing this work is important to avoid unexpected OOM situations.
>> >>>
>> >>> Signed-off-by: Eric Dumazet<eric.dumazet@gmail.com>
>> >>
>> >> Acked-by: Jon Mason<mason@myri.com>
>> >
>> > Thanks for reviewing Jon !
>> >
>> >
>> 
>> Please wait a second..  I think the patch is incorrect.
>> 
>> There is already code in myri10ge_rx_skb_build() which
>> attempts to set the truesize.  However, it sets it to
>> the used, rather than the allocated size so it is apparently
>> incorrect.
>> 
>> I'd prefer we fix that code.
> 
> Well, I believe I did exactly that :)
> 
> truesize of initial skb is fine.
> 
> Then for everay frag added, you must add to skb-truesize the allocated
> memory for this frag.
> 
> You add frags of a given size (small or big)
> 
> In the end, its truesize += bytes * number_of_frags
> 
> (bytes being small_size or big_size)

Right, I believe Eric's patch is correct, and I intend to apply it.
--
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
Andrew Gallatin - Oct. 20, 2011, 9:04 p.m.
On 10/20/11 16:59, Eric Dumazet wrote:
> Le jeudi 20 octobre 2011 à 16:45 -0400, Andrew Gallatin a écrit :
>> On 10/20/11 16:44, Eric Dumazet wrote:
>>> Le jeudi 20 octobre 2011 à 15:33 -0500, Jon Mason a écrit :
>>>> On Thu, Oct 20, 2011 at 3:10 PM, Eric Dumazet<eric.dumazet@gmail.com>   wrote:
>>>>> skb->truesize must account for allocated memory, not the used part of
>>>>> it. Doing this work is important to avoid unexpected OOM situations.
>>>>>
>>>>> Signed-off-by: Eric Dumazet<eric.dumazet@gmail.com>
>>>>
>>>> Acked-by: Jon Mason<mason@myri.com>
>>>
>>> Thanks for reviewing Jon !
>>>
>>>
>>
>> Please wait a second..  I think the patch is incorrect.
>>
>> There is already code in myri10ge_rx_skb_build() which
>> attempts to set the truesize.  However, it sets it to
>> the used, rather than the allocated size so it is apparently
>> incorrect.
>>
>> I'd prefer we fix that code.
>
> Well, I believe I did exactly that :)
>
> truesize of initial skb is fine.
>
> Then for everay frag added, you must add to skb-truesize the allocated
> memory for this frag.
>
> You add frags of a given size (small or big)
>
> In the end, its truesize += bytes * number_of_frags
>
> (bytes being small_size or big_size)
>
>

OK, I'm feeling foolish.  I somehow missed the first hunk
of your patch (some local change made it not apply cleanly, and
I did not notice that).

Yes, your patch is correct.  Sorry for the hassle!


Acked-by: Andrew Gallatin <gallatin@myri.com>
--
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
Eric Dumazet - Oct. 20, 2011, 9:06 p.m.
Le jeudi 20 octobre 2011 à 17:04 -0400, Andrew Gallatin a écrit :

> OK, I'm feeling foolish.  I somehow missed the first hunk
> of your patch (some local change made it not apply cleanly, and
> I did not notice that).
> 
> Yes, your patch is correct.  Sorry for the hassle!
> 

No worry, thanks a lot for reviewing !

> 
> Acked-by: Andrew Gallatin <gallatin@myri.com>



--
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 - Oct. 20, 2011, 9:42 p.m.
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Thu, 20 Oct 2011 22:10:03 +0200

> skb->truesize must account for allocated memory, not the used part of
> it. Doing this work is important to avoid unexpected OOM situations.
> 
> Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>

Applied.
--
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/drivers/net/ethernet/myricom/myri10ge/myri10ge.c b/drivers/net/ethernet/myricom/myri10ge/myri10ge.c
index c970a48..0778edc 100644
--- a/drivers/net/ethernet/myricom/myri10ge/myri10ge.c
+++ b/drivers/net/ethernet/myricom/myri10ge/myri10ge.c
@@ -1210,7 +1210,6 @@  myri10ge_rx_skb_build(struct sk_buff *skb, u8 * va,
 	struct skb_frag_struct *skb_frags;
 
 	skb->len = skb->data_len = len;
-	skb->truesize = len + sizeof(struct sk_buff);
 	/* attach the page(s) */
 
 	skb_frags = skb_shinfo(skb)->frags;
@@ -1385,6 +1384,8 @@  myri10ge_rx_done(struct myri10ge_slice_state *ss, int len, __wsum csum,
 	if (skb_frag_size(&skb_shinfo(skb)->frags[0]) <= 0) {
 		skb_frag_unref(skb, 0);
 		skb_shinfo(skb)->nr_frags = 0;
+	} else {
+		skb->truesize += bytes * skb_shinfo(skb)->nr_frags;
 	}
 	skb->protocol = eth_type_trans(skb, dev);
 	skb_record_rx_queue(skb, ss - &mgp->ss[0]);