From: David Laight
Date: Tue, 27 May 2014 15:23:14 +
> True, you still need to sort out how to handle multiple ethernet
> frames in the same 4k page, and arbitrary page boundaries within a
> frame.
Multiple ethernet frames in the same 4K page is used by several ethernet
drivers already. NIU
From: David Miller
> From: David Laight
> Date: Fri, 23 May 2014 08:52:13 +
>
> > The hardware will put multiple ethernet frames into a single USB bulk data
> > message. To handle this the driver generates a URB that is (hopefully) long
> > enough for the longest USB message (typically 32k is
From: David Laight
Date: Fri, 23 May 2014 08:52:13 +
> The hardware will put multiple ethernet frames into a single USB bulk data
> message. To handle this the driver generates a URB that is (hopefully) long
> enough for the longest USB message (typically 32k is assumed to be enough).
> The U
From: Jim Baxter
Date: Fri, 23 May 2014 10:21:19 +0100
> On 22/05/14 21:58, David Miller wrote:
>> From: Jim Baxter
>> Date: Thu, 22 May 2014 21:21:55 +0100
>>
>>> If there are multiple clones for an skb all pointing at the same data,
>>> will that distort the memory used when they all have the
On Fri, 2014-05-23 at 08:44 -0700, Rick Jones wrote:
> If you are measuring performance with the likes of netperf, you should
> be able to get an idea of the performance effect from the change in
> service demand (CPU consumed per unit of work) even if the maximum
> throughput remains capped.
On 05/23/2014 02:33 AM, Bjørn Mork wrote:
Jim Baxter writes:
I'll create and test a patch for the cdc_ncm host driver unless someone
else wants to do that. I haven't really played with the gadget driver
before, so I'd prefer if someone knowing it (Jim maybe?) could take care
of it. If not, th
On Fri, 2014-05-23 at 15:30 +, David Laight wrote:
> From: Eric Dumazet
> ...
> > TCP stack uses fast clones, and current stack gives them a truesize of
> > 2048 + sizeof(sk_buff), while it really should be 2048 +
> > 2*sizeof(sk_buff)
> >
> > Luckily, GSO/TSO tends to reduce the error, as skb
From: Eric Dumazet
...
> TCP stack uses fast clones, and current stack gives them a truesize of
> 2048 + sizeof(sk_buff), while it really should be 2048 +
> 2*sizeof(sk_buff)
>
> Luckily, GSO/TSO tends to reduce the error, as skbs overhead is lower.
Doesn't that affect the tx side - where the tru
On 23/05/14 14:47, Eric Dumazet wrote:
> On Fri, 2014-05-23 at 12:13 +0100, Jim Baxter wrote:
>
>> What are the side effects of changing the truesize, if the original
>> uncloned skb has the full truesize then isn't the potential memory usage
>> still counted for the avoidance of OOM?
>
> Nope. T
On Fri, 2014-05-23 at 11:33 +0200, Bjørn Mork wrote:
> Jim Baxter writes:
>
> >> I'll create and test a patch for the cdc_ncm host driver unless someone
> >> else wants to do that. I haven't really played with the gadget driver
> >> before, so I'd prefer if someone knowing it (Jim maybe?) could t
On Fri, 2014-05-23 at 12:13 +0100, Jim Baxter wrote:
> What are the side effects of changing the truesize, if the original
> uncloned skb has the full truesize then isn't the potential memory usage
> still counted for the avoidance of OOM?
Nope. This can be disastrous.
A malicious remote peer ca
On 23/05/14 11:45, David Laight wrote:
> From: Bjørn Mork [mailto:bj...@mork.no]
>> David Laight writes:
> ...
>>> Note that some of the usb ethernet drivers allocate large skb then
>>> lie about the truesize.
>>
>> Hmm, which drivers are these?
>
> $ grep truesize linux/drivers/net/usb/*
>
> as
From: Bjørn Mork [mailto:bj...@mork.no]
> David Laight writes:
...
> > Note that some of the usb ethernet drivers allocate large skb then
> > lie about the truesize.
>
> Hmm, which drivers are these?
$ grep truesize linux/drivers/net/usb/*
asix_88179_178a, smsc95xx, sr9700.
David
David Laight writes:
> Since USB bulk data are terminated by a short fragment there is actually
> no need for the URB be long enough for the full message. Provided the
> URB are multiples of the USB message size (1k for USB 3) the message
> can be received into multiple URB - the driver just has
Jim Baxter writes:
>> I'll create and test a patch for the cdc_ncm host driver unless someone
>> else wants to do that. I haven't really played with the gadget driver
>> before, so I'd prefer if someone knowing it (Jim maybe?) could take care
>> of it. If not, then I can always make an attempt u
From: Jim Baxter
> On 22/05/14 21:58, David Miller wrote:
> > From: Jim Baxter
> > Date: Thu, 22 May 2014 21:21:55 +0100
> >
> >> If there are multiple clones for an skb all pointing at the same data,
> >> will that distort the memory used when they all have the same truesize?
> >
> > What is dist
On 22/05/14 21:58, David Miller wrote:
> From: Jim Baxter
> Date: Thu, 22 May 2014 21:21:55 +0100
>
>> If there are multiple clones for an skb all pointing at the same data,
>> will that distort the memory used when they all have the same truesize?
>
> What is distorted about it?
>
> Each clone
> But although the problem is the same, I believe the driver in question
> isn't the one I have been looking at recently. The posted code snippet
> was from the NCM gadget driver (drivers/usb/gadget/f_ncm.c), isn't that
> right Jim?
Yes this is the NCM Gadget driver.
>
> Yes, judging by this dis
From: Eric Dumazet
> On Thu, 2014-05-22 at 20:07 +0100, Jim Baxter wrote:
>
> >
> > I have been investigating a network issue with bursts of network traffic
> > over USB CDC-NCM, the issue is that the kernel is dropping packets
> > because sk_rcvqueues_full() returns true due to skb2->truesize is
David Miller writes:
> From: Eric Dumazet
> Date: Thu, 22 May 2014 14:03:21 -0700
>
>> On Thu, 2014-05-22 at 13:58 -0700, Eric Dumazet wrote:
>>
>>> I would set rx_max (rx_urb_size) to SKB_MAX_HEAD(0) so that you do not
>>> use high order allocations.
>>
>> Correction, that would need SKB_MAX_H
From: Eric Dumazet
Date: Thu, 22 May 2014 14:03:21 -0700
> On Thu, 2014-05-22 at 13:58 -0700, Eric Dumazet wrote:
>
>> I would set rx_max (rx_urb_size) to SKB_MAX_HEAD(0) so that you do not
>> use high order allocations.
>
> Correction, that would need SKB_MAX_HEAD(NET_SKB_PAD + NET_IP_ALIGN),
On Thu, 2014-05-22 at 13:58 -0700, Eric Dumazet wrote:
> I would set rx_max (rx_urb_size) to SKB_MAX_HEAD(0) so that you do not
> use high order allocations.
Correction, that would need SKB_MAX_HEAD(NET_SKB_PAD + NET_IP_ALIGN),
because drivers/net/usb/usbnet.c calls __netdev_alloc_skb_ip_align().
From: Jim Baxter
Date: Thu, 22 May 2014 21:21:55 +0100
> If there are multiple clones for an skb all pointing at the same data,
> will that distort the memory used when they all have the same truesize?
What is distorted about it?
Each clone references exactly that much backing memory.
--
To uns
On Thu, 2014-05-22 at 20:07 +0100, Jim Baxter wrote:
>
> I have been investigating a network issue with bursts of network traffic
> over USB CDC-NCM, the issue is that the kernel is dropping packets
> because sk_rcvqueues_full() returns true due to skb2->truesize is always
> 32960 instead of SKB_
On 22/05/14 20:21, David Miller wrote:
> From: Jim Baxter
> Date: Thu, 22 May 2014 20:07:41 +0100
>
>> My questions are:
>>
>> 1) Which buffer size does truesize represent, is it the total buffer or
>> just the data related to the relevant skb?
>>
>> 2) If truesize is for the skb it is contained
On Thu, 2014-05-22 at 21:21 +0100, Jim Baxter wrote:
> OK, so it is the value of the memory that has been allocated for the SKB.
> If there are multiple clones for an skb all pointing at the same data,
> will that distort the memory used when they all have the same truesize?
Its always better to
On 22/05/14 20:59, Eric Dumazet wrote:
> On Thu, 2014-05-22 at 20:39 +0100, Jim Baxter wrote:
>
>> I now think that the correct solution here is to create a new smaller
>> skb and copy the data from the sub packets into it.
>
> For low speed devices, this is indeed the best way.
>
> (this is cal
On Thu, 2014-05-22 at 20:39 +0100, Jim Baxter wrote:
> I now think that the correct solution here is to create a new smaller
> skb and copy the data from the sub packets into it.
For low speed devices, this is indeed the best way.
(this is called copybreak in some nic drivers)
--
To unsubscri
On 22/05/14 20:25, Vlad Yasevich wrote:
> On 05/22/2014 03:07 PM, Jim Baxter wrote:
>>
>> skb2 = skb_clone(skb, GFP_ATOMIC);
>> if (skb2 == NULL)
>> goto err;
>>
>> if (!skb_pull(skb2, index)) {
>> ret = -EOVERFLOW;
>> goto err;
>> }
>
> This assumes that yo
On 05/22/2014 03:07 PM, Jim Baxter wrote:
> Hi, I was hoping you can help me with some questions.
>
> I have been investigating a network issue with bursts of network traffic
> over USB CDC-NCM, the issue is that the kernel is dropping packets
> because sk_rcvqueues_full() returns true due to skb2
From: Jim Baxter
Date: Thu, 22 May 2014 20:07:41 +0100
> My questions are:
>
> 1) Which buffer size does truesize represent, is it the total buffer or
> just the data related to the relevant skb?
>
> 2) If truesize is for the skb it is contained within should it be
> updated during the call to
31 matches
Mail list logo
