On 12/08/2018 12:14 PM, Ilias Apalodimas wrote:
> On Sat, Dec 08, 2018 at 09:11:53PM +0100, Jesper Dangaard Brouer wrote:
>>>
>>> I want to make sure you guys thought about splice() stuff, and
>>> skb_try_coalesce(), and GRO, and skb cloning, and ...
>>
>> Thanks for the pointers. To Ilias,
On Sat, Dec 08, 2018 at 10:14:47PM +0200, Ilias Apalodimas wrote:
> On Sat, Dec 08, 2018 at 09:11:53PM +0100, Jesper Dangaard Brouer wrote:
> > >
> > > I want to make sure you guys thought about splice() stuff, and
> > > skb_try_coalesce(), and GRO, and skb cloning, and ...
> >
> > Thanks for
On Sat, Dec 08, 2018 at 12:21:10PM -0800, David Miller wrote:
> From: Ilias Apalodimas
> Date: Sat, 8 Dec 2018 16:57:28 +0200
>
> > The patchset speeds up the mvneta driver on the default network
> > stack. The only change that was needed was to adapt the driver to
> > using the page_pool API.
From: Ilias Apalodimas
Date: Sat, 8 Dec 2018 16:57:28 +0200
> The patchset speeds up the mvneta driver on the default network
> stack. The only change that was needed was to adapt the driver to
> using the page_pool API. The speed improvements we are seeing on
> specific workloads (i.e 256b <
On Sat, Dec 08, 2018 at 09:11:53PM +0100, Jesper Dangaard Brouer wrote:
> >
> > I want to make sure you guys thought about splice() stuff, and
> > skb_try_coalesce(), and GRO, and skb cloning, and ...
>
> Thanks for the pointers. To Ilias, we need to double check skb_try_coalesce()
> code path,
On Sat, 8 Dec 2018 11:26:56 -0800
Eric Dumazet wrote:
> On 12/08/2018 06:57 AM, Ilias Apalodimas wrote:
> > Hi Eric,
> This patch is changing struct sk_buff, and is thus per-definition
> controversial.
>
> Place a new member 'mem_info' of type struct xdp_mem_info, just
From: Jesper Dangaard Brouer
Date: Sat, 8 Dec 2018 12:36:10 +0100
> The annoying part is actually that depending on the kernel config
> options CONFIG_XFRM, CONFIG_NF_CONNTRACK and CONFIG_BRIDGE_NETFILTER,
> whether there is a cache-line split, where mem_info gets moved into the
> next
On 12/08/2018 06:57 AM, Ilias Apalodimas wrote:
> Hi Eric,
This patch is changing struct sk_buff, and is thus per-definition
controversial.
Place a new member 'mem_info' of type struct xdp_mem_info, just after
members (flags) head_frag and pfmemalloc, And not in
> I got other concerns on the patchset though. Like how much memory is
> it 'ok' to keep mapped keeping in mind we are using the streaming
> DMA API. Are we going to affect anyone else negatively by doing so ?
For mvneta, you can expect the target to have between 512Mbyte to
3G. You can take a
Hi Eric,
> >> This patch is changing struct sk_buff, and is thus per-definition
> >> controversial.
> >>
> >> Place a new member 'mem_info' of type struct xdp_mem_info, just after
> >> members (flags) head_frag and pfmemalloc, And not in between
> >> headers_start/end to ensure skb_copy() and
On Sat, 8 Dec 2018 04:29:17 -0800
Eric Dumazet wrote:
> On 12/08/2018 01:57 AM, Florian Westphal wrote:
> > Jesper Dangaard Brouer wrote:
> >> From: Ilias Apalodimas
> >>
> >> This patch is changing struct sk_buff, and is thus per-definition
> >> controversial.
> >>
> >> Place a new member
On 12/08/2018 04:29 AM, Eric Dumazet wrote:
>
> But I do not get why the patch is needed.
>
> Adding extra cost for each skb destruction is costly.
>
> I though XDP was all about _not_ having skbs.
>
> Please let's do not slow down the non XDP stack only to make XDP more
> appealing.
>
On 12/08/2018 01:57 AM, Florian Westphal wrote:
> Jesper Dangaard Brouer wrote:
>> From: Ilias Apalodimas
>>
>> This patch is changing struct sk_buff, and is thus per-definition
>> controversial.
>>
>> Place a new member 'mem_info' of type struct xdp_mem_info, just after
>> members (flags)
On Sat, 8 Dec 2018 10:57:58 +0100
Florian Westphal wrote:
> Jesper Dangaard Brouer wrote:
> > From: Ilias Apalodimas
> >
> > This patch is changing struct sk_buff, and is thus per-definition
> > controversial.
> >
> > Place a new member 'mem_info' of type struct xdp_mem_info, just after
> >
Jesper Dangaard Brouer wrote:
> From: Ilias Apalodimas
>
> This patch is changing struct sk_buff, and is thus per-definition
> controversial.
>
> Place a new member 'mem_info' of type struct xdp_mem_info, just after
> members (flags) head_frag and pfmemalloc, And not in between
>
15 matches
Mail list logo
