* xdpsock poll syscall CPU 100%
@ 2020-02-20 22:49 William Tu
2020-02-21 8:11 ` Magnus Karlsson
2020-02-21 8:28 ` Eelco Chaudron
0 siblings, 2 replies; 8+ messages in thread
From: William Tu @ 2020-02-20 22:49 UTC (permalink / raw)
To: Xdp; +Cc: Karlsson, Magnus, Björn Töpel
Hi,
I'm trying to save some CPU cycles when there is no packet arrives.
I enable the poll syscall option of xdpsock, by doing
$ ./xdpsock -r -p -S -i ens16
sock0@ens160:0 rxdrop xdp-skb poll()
pps pkts 1.00
rx 0 0
tx 0 0
Since there is no packet coming, I though by calling poll()
system call, the xdpsock process will be blocked and CPU utilization
should be way under 100%. However, I'm still seeing 100%
CPU utilization. Am I understanding this correctly?
Thanks,
William
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: xdpsock poll syscall CPU 100%
2020-02-20 22:49 xdpsock poll syscall CPU 100% William Tu
@ 2020-02-21 8:11 ` Magnus Karlsson
2020-02-21 8:28 ` Eelco Chaudron
1 sibling, 0 replies; 8+ messages in thread
From: Magnus Karlsson @ 2020-02-21 8:11 UTC (permalink / raw)
To: William Tu; +Cc: Xdp, Karlsson, Magnus, Björn Töpel
On Thu, Feb 20, 2020 at 11:50 PM William Tu <u9012063@gmail.com> wrote:
>
> Hi,
>
> I'm trying to save some CPU cycles when there is no packet arrives.
> I enable the poll syscall option of xdpsock, by doing
>
> $ ./xdpsock -r -p -S -i ens16
> sock0@ens160:0 rxdrop xdp-skb poll()
> pps pkts 1.00
> rx 0 0
> tx 0 0
>
> Since there is no packet coming, I though by calling poll()
> system call, the xdpsock process will be blocked and CPU utilization
> should be way under 100%. However, I'm still seeing 100%
> CPU utilization. Am I understanding this correctly?
Yes, something seems to be wrong here. Calling poll() should put you
in sleep for the timeout period, here 1000 ms. Let me take a look at
it. What version of the kernel are you running?
/Magnus
> Thanks,
> William
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: xdpsock poll syscall CPU 100%
2020-02-20 22:49 xdpsock poll syscall CPU 100% William Tu
2020-02-21 8:11 ` Magnus Karlsson
@ 2020-02-21 8:28 ` Eelco Chaudron
2020-02-21 8:33 ` Magnus Karlsson
1 sibling, 1 reply; 8+ messages in thread
From: Eelco Chaudron @ 2020-02-21 8:28 UTC (permalink / raw)
To: William Tu; +Cc: Xdp, Karlsson, Magnus, Björn Töpel
On 20 Feb 2020, at 23:49, William Tu wrote:
> Hi,
>
> I'm trying to save some CPU cycles when there is no packet arrives.
> I enable the poll syscall option of xdpsock, by doing
>
> $ ./xdpsock -r -p -S -i ens16
> sock0@ens160:0 rxdrop xdp-skb poll()
> pps pkts 1.00
> rx 0 0
> tx 0 0
>
> Since there is no packet coming, I though by calling poll()
> system call, the xdpsock process will be blocked and CPU utilization
> should be way under 100%. However, I'm still seeing 100%
> CPU utilization. Am I understanding this correctly?
Hi William, I can remember I saw this in the past two with this code. It
had something to do with the way xdpsock waits for the buffers to be
free’ ed by the kernel. What I can remember it had something to do
with the veth interfaces also.
I do remember that I fixed it in the tutorial for AF_XDP:
https://github.com/xdp-project/xdp-tutorial/blob/master/advanced03-AF_XDP/af_xdp_user.c
Maybe you can see if you have the same problem with this example.
//Eelco
> Thanks,
> William
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: xdpsock poll syscall CPU 100%
2020-02-21 8:28 ` Eelco Chaudron
@ 2020-02-21 8:33 ` Magnus Karlsson
2020-02-21 8:40 ` Eelco Chaudron
0 siblings, 1 reply; 8+ messages in thread
From: Magnus Karlsson @ 2020-02-21 8:33 UTC (permalink / raw)
To: Eelco Chaudron; +Cc: William Tu, Xdp, Karlsson, Magnus, Björn Töpel
On Fri, Feb 21, 2020 at 9:30 AM Eelco Chaudron <echaudro@redhat.com> wrote:
>
>
>
> On 20 Feb 2020, at 23:49, William Tu wrote:
>
> > Hi,
> >
> > I'm trying to save some CPU cycles when there is no packet arrives.
> > I enable the poll syscall option of xdpsock, by doing
> >
> > $ ./xdpsock -r -p -S -i ens16
> > sock0@ens160:0 rxdrop xdp-skb poll()
> > pps pkts 1.00
> > rx 0 0
> > tx 0 0
> >
> > Since there is no packet coming, I though by calling poll()
> > system call, the xdpsock process will be blocked and CPU utilization
> > should be way under 100%. However, I'm still seeing 100%
> > CPU utilization. Am I understanding this correctly?
>
> Hi William, I can remember I saw this in the past two with this code. It
> had something to do with the way xdpsock waits for the buffers to be
> free’ ed by the kernel. What I can remember it had something to do
> with the veth interfaces also.
>
> I do remember that I fixed it in the tutorial for AF_XDP:
> https://github.com/xdp-project/xdp-tutorial/blob/master/advanced03-AF_XDP/af_xdp_user.c
Eelco,
Do you remember exactly what you had to fix in the xdpsock sample?
Your tutorial is quite a rewrite so it is hard for me to tell exactly
which of all the changes that fix this problem. The reason I ask is
that it would be nice to fix this in the sample too.
Thanks: Magnus
> Maybe you can see if you have the same problem with this example.
>
> //Eelco
>
>
> > Thanks,
> > William
>
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: xdpsock poll syscall CPU 100%
2020-02-21 8:33 ` Magnus Karlsson
@ 2020-02-21 8:40 ` Eelco Chaudron
2020-02-21 8:46 ` Magnus Karlsson
0 siblings, 1 reply; 8+ messages in thread
From: Eelco Chaudron @ 2020-02-21 8:40 UTC (permalink / raw)
To: Magnus Karlsson; +Cc: William Tu, Xdp, Karlsson, Magnus, Björn Töpel
On 21 Feb 2020, at 9:33, Magnus Karlsson wrote:
> On Fri, Feb 21, 2020 at 9:30 AM Eelco Chaudron <echaudro@redhat.com>
> wrote:
>>
>>
>>
>> On 20 Feb 2020, at 23:49, William Tu wrote:
>>
>>> Hi,
>>>
>>> I'm trying to save some CPU cycles when there is no packet arrives.
>>> I enable the poll syscall option of xdpsock, by doing
>>>
>>> $ ./xdpsock -r -p -S -i ens16
>>> sock0@ens160:0 rxdrop xdp-skb poll()
>>> pps pkts 1.00
>>> rx 0 0
>>> tx 0 0
>>>
>>> Since there is no packet coming, I though by calling poll()
>>> system call, the xdpsock process will be blocked and CPU utilization
>>> should be way under 100%. However, I'm still seeing 100%
>>> CPU utilization. Am I understanding this correctly?
>>
>> Hi William, I can remember I saw this in the past two with this code.
>> It
>> had something to do with the way xdpsock waits for the buffers to be
>> free’ ed by the kernel. What I can remember it had something to do
>> with the veth interfaces also.
>>
>> I do remember that I fixed it in the tutorial for AF_XDP:
>> https://github.com/xdp-project/xdp-tutorial/blob/master/advanced03-AF_XDP/af_xdp_user.c
>
> Eelco,
>
> Do you remember exactly what you had to fix in the xdpsock sample?
> Your tutorial is quite a rewrite so it is hard for me to tell exactly
> which of all the changes that fix this problem. The reason I ask is
> that it would be nice to fix this in the sample too.
>
> Thanks: Magnus
From an earlier email conversation we had this is where it looped in my
case:
>>>>> One other thing I noticed, which I need to research is that if I
>>>>> use
>>>>> rx_drop() function from /xdpsock_user.c it loops a lot in:
>>>>>
>>>>> while (ret != rcvd) {
>>>>> if (ret < 0) {
>>>>> exit(-1);
>>>>> }
>>>>> ret = xsk_ring_prod__reserve(&xsk->umem-
>>> fq, rcvd, &idx_fq);
>>>>>
>>>>> }
>>>>>
>>>>> As ret return 0, until (it looks like) I send more packets. So
>>>>> even
>>>>> in the poll() mode, it uses 100% cpu after sending a single
>>>>> packet.
>>>>> Note this is with the default Fedora Kernel, as I’m working on
>>>>> this
>>>>> from my laptop. Does this sound familiar? If not I’ll dig into
>>>>> it
>>>>> once I’m back.
>>>>
>>>> The xdpsock test is a busypolling test, to compare against DPDK
>>>> speeds. For real use-cases, I think people will want to trade-off
>>>> latency vs. burning CPU.
>>>
>>> I understand the use case, but even with the xdpsock test program,
>>> if
>>> I send a single packet it’s not received, or at least not when
>>> it's
>>> sent. It takes 16 (or a multiple of it) before the get
>>> detected/processed. I think it’s because of the
>>> xsk_ring_prod__reserve(), but I’ll try to debug it more today and
>>> to
>>> understand the APIs better.
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: xdpsock poll syscall CPU 100%
2020-02-21 8:40 ` Eelco Chaudron
@ 2020-02-21 8:46 ` Magnus Karlsson
2020-02-21 9:23 ` Eelco Chaudron
2020-02-21 18:40 ` William Tu
0 siblings, 2 replies; 8+ messages in thread
From: Magnus Karlsson @ 2020-02-21 8:46 UTC (permalink / raw)
To: Eelco Chaudron; +Cc: William Tu, Xdp, Karlsson, Magnus, Björn Töpel
On Fri, Feb 21, 2020 at 9:40 AM Eelco Chaudron <echaudro@redhat.com> wrote:
>
>
>
> On 21 Feb 2020, at 9:33, Magnus Karlsson wrote:
>
> > On Fri, Feb 21, 2020 at 9:30 AM Eelco Chaudron <echaudro@redhat.com>
> > wrote:
> >>
> >>
> >>
> >> On 20 Feb 2020, at 23:49, William Tu wrote:
> >>
> >>> Hi,
> >>>
> >>> I'm trying to save some CPU cycles when there is no packet arrives.
> >>> I enable the poll syscall option of xdpsock, by doing
> >>>
> >>> $ ./xdpsock -r -p -S -i ens16
> >>> sock0@ens160:0 rxdrop xdp-skb poll()
> >>> pps pkts 1.00
> >>> rx 0 0
> >>> tx 0 0
> >>>
> >>> Since there is no packet coming, I though by calling poll()
> >>> system call, the xdpsock process will be blocked and CPU utilization
> >>> should be way under 100%. However, I'm still seeing 100%
> >>> CPU utilization. Am I understanding this correctly?
> >>
> >> Hi William, I can remember I saw this in the past two with this code.
> >> It
> >> had something to do with the way xdpsock waits for the buffers to be
> >> free’ ed by the kernel. What I can remember it had something to do
> >> with the veth interfaces also.
> >>
> >> I do remember that I fixed it in the tutorial for AF_XDP:
> >> https://github.com/xdp-project/xdp-tutorial/blob/master/advanced03-AF_XDP/af_xdp_user.c
> >
> > Eelco,
> >
> > Do you remember exactly what you had to fix in the xdpsock sample?
> > Your tutorial is quite a rewrite so it is hard for me to tell exactly
> > which of all the changes that fix this problem. The reason I ask is
> > that it would be nice to fix this in the sample too.
> >
> > Thanks: Magnus
>
> From an earlier email conversation we had this is where it looped in my
> case:
Thanks Eelco. Yes, the xdpsock sample is too simplistic in this case.
I will put this on my backlog to fix so that we do not have this
problem in the future. I might take some inspiration from your code
:-). Hope you do not mind.
/Magnus
> >>>>> One other thing I noticed, which I need to research is that if I
> >>>>> use
> >>>>> rx_drop() function from /xdpsock_user.c it loops a lot in:
> >>>>>
> >>>>> while (ret != rcvd) {
> >>>>> if (ret < 0) {
> >>>>> exit(-1);
> >>>>> }
> >>>>> ret = xsk_ring_prod__reserve(&xsk->umem-
> >>> fq, rcvd, &idx_fq);
> >>>>>
> >>>>> }
> >>>>>
> >>>>> As ret return 0, until (it looks like) I send more packets. So
> >>>>> even
> >>>>> in the poll() mode, it uses 100% cpu after sending a single
> >>>>> packet.
> >>>>> Note this is with the default Fedora Kernel, as I’m working on
> >>>>> this
> >>>>> from my laptop. Does this sound familiar? If not I’ll dig into
> >>>>> it
> >>>>> once I’m back.
> >>>>
> >>>> The xdpsock test is a busypolling test, to compare against DPDK
> >>>> speeds. For real use-cases, I think people will want to trade-off
> >>>> latency vs. burning CPU.
> >>>
> >>> I understand the use case, but even with the xdpsock test program,
> >>> if
> >>> I send a single packet it’s not received, or at least not when
> >>> it's
> >>> sent. It takes 16 (or a multiple of it) before the get
> >>> detected/processed. I think it’s because of the
> >>> xsk_ring_prod__reserve(), but I’ll try to debug it more today and
> >>> to
> >>> understand the APIs better.
>
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: xdpsock poll syscall CPU 100%
2020-02-21 8:46 ` Magnus Karlsson
@ 2020-02-21 9:23 ` Eelco Chaudron
2020-02-21 18:40 ` William Tu
1 sibling, 0 replies; 8+ messages in thread
From: Eelco Chaudron @ 2020-02-21 9:23 UTC (permalink / raw)
To: Magnus Karlsson; +Cc: William Tu, Xdp, Karlsson, Magnus, Björn Töpel
On 21 Feb 2020, at 9:46, Magnus Karlsson wrote:
> On Fri, Feb 21, 2020 at 9:40 AM Eelco Chaudron <echaudro@redhat.com>
> wrote:
>>
>>
>>
>> On 21 Feb 2020, at 9:33, Magnus Karlsson wrote:
>>
>>> On Fri, Feb 21, 2020 at 9:30 AM Eelco Chaudron <echaudro@redhat.com>
>>> wrote:
>>>>
>>>>
>>>>
>>>> On 20 Feb 2020, at 23:49, William Tu wrote:
>>>>
>>>>> Hi,
>>>>>
>>>>> I'm trying to save some CPU cycles when there is no packet
>>>>> arrives.
>>>>> I enable the poll syscall option of xdpsock, by doing
>>>>>
>>>>> $ ./xdpsock -r -p -S -i ens16
>>>>> sock0@ens160:0 rxdrop xdp-skb poll()
>>>>> pps pkts 1.00
>>>>> rx 0 0
>>>>> tx 0 0
>>>>>
>>>>> Since there is no packet coming, I though by calling poll()
>>>>> system call, the xdpsock process will be blocked and CPU
>>>>> utilization
>>>>> should be way under 100%. However, I'm still seeing 100%
>>>>> CPU utilization. Am I understanding this correctly?
>>>>
>>>> Hi William, I can remember I saw this in the past two with this
>>>> code.
>>>> It
>>>> had something to do with the way xdpsock waits for the buffers to
>>>> be
>>>> free’ ed by the kernel. What I can remember it had something to
>>>> do
>>>> with the veth interfaces also.
>>>>
>>>> I do remember that I fixed it in the tutorial for AF_XDP:
>>>> https://github.com/xdp-project/xdp-tutorial/blob/master/advanced03-AF_XDP/af_xdp_user.c
>>>
>>> Eelco,
>>>
>>> Do you remember exactly what you had to fix in the xdpsock sample?
>>> Your tutorial is quite a rewrite so it is hard for me to tell
>>> exactly
>>> which of all the changes that fix this problem. The reason I ask is
>>> that it would be nice to fix this in the sample too.
>>>
>>> Thanks: Magnus
>>
>> From an earlier email conversation we had this is where it looped in
>> my
>> case:
>
> Thanks Eelco. Yes, the xdpsock sample is too simplistic in this case.
> I will put this on my backlog to fix so that we do not have this
> problem in the future. I might take some inspiration from your code
> :-). Hope you do not mind.
Its GPL so take whatever you need ;)
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: xdpsock poll syscall CPU 100%
2020-02-21 8:46 ` Magnus Karlsson
2020-02-21 9:23 ` Eelco Chaudron
@ 2020-02-21 18:40 ` William Tu
1 sibling, 0 replies; 8+ messages in thread
From: William Tu @ 2020-02-21 18:40 UTC (permalink / raw)
To: Magnus Karlsson
Cc: Eelco Chaudron, Xdp, Karlsson, Magnus, Björn Töpel
On Fri, Feb 21, 2020 at 12:46 AM Magnus Karlsson
<magnus.karlsson@gmail.com> wrote:
>
> On Fri, Feb 21, 2020 at 9:40 AM Eelco Chaudron <echaudro@redhat.com> wrote:
> >
> >
> >
> > On 21 Feb 2020, at 9:33, Magnus Karlsson wrote:
> >
> > > On Fri, Feb 21, 2020 at 9:30 AM Eelco Chaudron <echaudro@redhat.com>
> > > wrote:
> > >>
> > >>
> > >>
> > >> On 20 Feb 2020, at 23:49, William Tu wrote:
> > >>
> > >>> Hi,
> > >>>
> > >>> I'm trying to save some CPU cycles when there is no packet arrives.
> > >>> I enable the poll syscall option of xdpsock, by doing
> > >>>
> > >>> $ ./xdpsock -r -p -S -i ens16
> > >>> sock0@ens160:0 rxdrop xdp-skb poll()
> > >>> pps pkts 1.00
> > >>> rx 0 0
> > >>> tx 0 0
> > >>>
> > >>> Since there is no packet coming, I though by calling poll()
> > >>> system call, the xdpsock process will be blocked and CPU utilization
> > >>> should be way under 100%. However, I'm still seeing 100%
> > >>> CPU utilization. Am I understanding this correctly?
> > >>
> > >> Hi William, I can remember I saw this in the past two with this code.
> > >> It
> > >> had something to do with the way xdpsock waits for the buffers to be
> > >> free’ ed by the kernel. What I can remember it had something to do
> > >> with the veth interfaces also.
> > >>
> > >> I do remember that I fixed it in the tutorial for AF_XDP:
> > >> https://github.com/xdp-project/xdp-tutorial/blob/master/advanced03-AF_XDP/af_xdp_user.c
> > >
> > > Eelco,
> > >
> > > Do you remember exactly what you had to fix in the xdpsock sample?
> > > Your tutorial is quite a rewrite so it is hard for me to tell exactly
> > > which of all the changes that fix this problem. The reason I ask is
> > > that it would be nice to fix this in the sample too.
> > >
> > > Thanks: Magnus
> >
> > From an earlier email conversation we had this is where it looped in my
> > case:
>
> Thanks Eelco. Yes, the xdpsock sample is too simplistic in this case.
> I will put this on my backlog to fix so that we do not have this
> problem in the future. I might take some inspiration from your code
> :-). Hope you do not mind.
>
> /Magnus
>
Hi Magnus and Eelco,
Thanks for your reply.
I run Eelco's af_xdp_user.c and indeed, with poll syscall, the CPU is
way under 100%. But I'm still figuring out the difference and where to
fix in xdpsock_user.c.
Regards,
William
^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2020-02-21 18:41 UTC | newest]
Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-02-20 22:49 xdpsock poll syscall CPU 100% William Tu
2020-02-21 8:11 ` Magnus Karlsson
2020-02-21 8:28 ` Eelco Chaudron
2020-02-21 8:33 ` Magnus Karlsson
2020-02-21 8:40 ` Eelco Chaudron
2020-02-21 8:46 ` Magnus Karlsson
2020-02-21 9:23 ` Eelco Chaudron
2020-02-21 18:40 ` William Tu
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.