| Message ID | 1549631126-29067-1-git-send-email-magnus.karlsson@intel.com |
|---|---|
| Headers | show |
| Series | libbpf: adding AF_XDP support | expand |
Hi Magnus, > * In a future release, I am planning on adding a higher level data > plane interface too. This will be based around recvmsg and sendmsg > with the use of struct iovec for batching, without the user having > to know anything about the underlying four rings of an AF_XDP > socket. There will be one semantic difference though from the > standard recvmsg and that is that the kernel will fill in the iovecs > instead of the application. But the rest should be the same as the > libc versions so that application writers feel at home. You might consider recvmmsg() and sendmmsg() (bulk of multi segments packets?) Jean-Mickael
On Mon, Feb 11, 2019 at 7:34 AM Jean-Mickael Guerin <jmg@6wind.com> wrote: > > Hi Magnus, > > > * In a future release, I am planning on adding a higher level data > > plane interface too. This will be based around recvmsg and sendmsg > > with the use of struct iovec for batching, without the user having > > to know anything about the underlying four rings of an AF_XDP > > socket. There will be one semantic difference though from the > > standard recvmsg and that is that the kernel will fill in the iovecs > > instead of the application. But the rest should be the same as the > > libc versions so that application writers feel at home. > > You might consider recvmmsg() and sendmmsg() (bulk of multi segments packets?) Exactly :-). Spot on. /Magnus > Jean-Mickael
On 8 Feb 2019, at 5:05, Magnus Karlsson wrote: > This patch proposes to add AF_XDP support to libbpf. The main reason > for this is to facilitate writing applications that use AF_XDP by > offering higher-level APIs that hide many of the details of the AF_XDP > uapi. This is in the same vein as libbpf facilitates XDP adoption by > offering easy-to-use higher level interfaces of XDP > functionality. Hopefully this will facilitate adoption of AF_XDP, make > applications using it simpler and smaller, and finally also make it > possible for applications to benefit from optimizations in the AF_XDP > user space access code. Previously, people just copied and pasted the > code from the sample application into their application, which is not > desirable. I like the idea of encapsulating the boilerplate logic in a library. I do think there is an important missing piece though - there should be some code which queries the netdev for how many queues are attached, and create the appropriate number of umem/AF_XDP sockets. I ran into this issue when testing the current AF_XDP code - on my test boxes, the mlx5 card has 55 channels (aka queues), so when the test program binds only to channel 0, nothing works as expected, since not all traffic is being intercepted. While obvious in hindsight, this took a while to track down.
On Mon, Feb 11, 2019 at 9:44 PM Jonathan Lemon <jonathan.lemon@gmail.com> wrote: > > On 8 Feb 2019, at 5:05, Magnus Karlsson wrote: > > > This patch proposes to add AF_XDP support to libbpf. The main reason > > for this is to facilitate writing applications that use AF_XDP by > > offering higher-level APIs that hide many of the details of the AF_XDP > > uapi. This is in the same vein as libbpf facilitates XDP adoption by > > offering easy-to-use higher level interfaces of XDP > > functionality. Hopefully this will facilitate adoption of AF_XDP, make > > applications using it simpler and smaller, and finally also make it > > possible for applications to benefit from optimizations in the AF_XDP > > user space access code. Previously, people just copied and pasted the > > code from the sample application into their application, which is not > > desirable. > > I like the idea of encapsulating the boilerplate logic in a library. > > I do think there is an important missing piece though - there should be > some code which queries the netdev for how many queues are attached, and > create the appropriate number of umem/AF_XDP sockets. > > I ran into this issue when testing the current AF_XDP code - on my test > boxes, the mlx5 card has 55 channels (aka queues), so when the test program > binds only to channel 0, nothing works as expected, since not all traffic > is being intercepted. While obvious in hindsight, this took a while to > track down. Yes, agreed. You are not the first one to stumble upon this problem :-). Let me think a little bit on how to solve this in a good way. We need this to be simple and intuitive, as you say. /Magnus > -- > Jonathan
On Wed, 13 Feb 2019 12:32:47 +0100 Magnus Karlsson <magnus.karlsson@gmail.com> wrote: > On Mon, Feb 11, 2019 at 9:44 PM Jonathan Lemon <jonathan.lemon@gmail.com> wrote: > > > > On 8 Feb 2019, at 5:05, Magnus Karlsson wrote: > > > > > This patch proposes to add AF_XDP support to libbpf. The main reason > > > for this is to facilitate writing applications that use AF_XDP by > > > offering higher-level APIs that hide many of the details of the AF_XDP > > > uapi. This is in the same vein as libbpf facilitates XDP adoption by > > > offering easy-to-use higher level interfaces of XDP > > > functionality. Hopefully this will facilitate adoption of AF_XDP, make > > > applications using it simpler and smaller, and finally also make it > > > possible for applications to benefit from optimizations in the AF_XDP > > > user space access code. Previously, people just copied and pasted the > > > code from the sample application into their application, which is not > > > desirable. > > > > I like the idea of encapsulating the boilerplate logic in a library. > > > > I do think there is an important missing piece though - there should be > > some code which queries the netdev for how many queues are attached, and > > create the appropriate number of umem/AF_XDP sockets. > > > > I ran into this issue when testing the current AF_XDP code - on my test > > boxes, the mlx5 card has 55 channels (aka queues), so when the test program > > binds only to channel 0, nothing works as expected, since not all traffic > > is being intercepted. While obvious in hindsight, this took a while to > > track down. > > Yes, agreed. You are not the first one to stumble upon this problem > :-). Let me think a little bit on how to solve this in a good way. We > need this to be simple and intuitive, as you say. I see people hitting this with AF_XDP all the time... I had some backup-slides[2] in our FOSDEM presentation[1] that describe the issue, give the performance reason why and propose a workaround. [1] https://github.com/xdp-project/xdp-project/tree/master/conference/FOSDEM2019 [2] https://github.com/xdp-project/xdp-project/blob/master/conference/FOSDEM2019/xdp_building_block.org#backup-slides Alternative work-around * Create as many AF_XDP sockets as RXQs * Have userspace poll()/select on all sockets
On 13 Feb 2019, at 3:32, Magnus Karlsson wrote: > On Mon, Feb 11, 2019 at 9:44 PM Jonathan Lemon > <jonathan.lemon@gmail.com> wrote: >> >> On 8 Feb 2019, at 5:05, Magnus Karlsson wrote: >> >>> This patch proposes to add AF_XDP support to libbpf. The main reason >>> for this is to facilitate writing applications that use AF_XDP by >>> offering higher-level APIs that hide many of the details of the >>> AF_XDP >>> uapi. This is in the same vein as libbpf facilitates XDP adoption by >>> offering easy-to-use higher level interfaces of XDP >>> functionality. Hopefully this will facilitate adoption of AF_XDP, >>> make >>> applications using it simpler and smaller, and finally also make it >>> possible for applications to benefit from optimizations in the >>> AF_XDP >>> user space access code. Previously, people just copied and pasted >>> the >>> code from the sample application into their application, which is >>> not >>> desirable. >> >> I like the idea of encapsulating the boilerplate logic in a library. >> >> I do think there is an important missing piece though - there should >> be >> some code which queries the netdev for how many queues are attached, >> and >> create the appropriate number of umem/AF_XDP sockets. >> >> I ran into this issue when testing the current AF_XDP code - on my >> test >> boxes, the mlx5 card has 55 channels (aka queues), so when the test >> program >> binds only to channel 0, nothing works as expected, since not all >> traffic >> is being intercepted. While obvious in hindsight, this took a while >> to >> track down. > > Yes, agreed. You are not the first one to stumble upon this problem > :-). Let me think a little bit on how to solve this in a good way. We > need this to be simple and intuitive, as you say. Has any investigation been done on using some variant of MPSC implementation as an intermediate form for AF_XDP? E.g.: something like LCRQ or the bulkQ in bpf devmap/cpumap. I'm aware that this would be slightly slower, as it would introduce a lock in the path, but I'd think that having DEVMAP, CPUMAP and XSKMAP all behave the same way would add more flexibility. Ideally, if the configuration matches the underlying hardware, then the implementation would reduce to the current setup (and allow ZC implementations), but a non-matching configuration would still work - as opposed to the current situation.
On Wed, Feb 13, 2019 at 9:49 PM Jonathan Lemon <jonathan.lemon@gmail.com> wrote: > > On 13 Feb 2019, at 3:32, Magnus Karlsson wrote: > > > On Mon, Feb 11, 2019 at 9:44 PM Jonathan Lemon > > <jonathan.lemon@gmail.com> wrote: > >> > >> On 8 Feb 2019, at 5:05, Magnus Karlsson wrote: > >> > >>> This patch proposes to add AF_XDP support to libbpf. The main reason > >>> for this is to facilitate writing applications that use AF_XDP by > >>> offering higher-level APIs that hide many of the details of the > >>> AF_XDP > >>> uapi. This is in the same vein as libbpf facilitates XDP adoption by > >>> offering easy-to-use higher level interfaces of XDP > >>> functionality. Hopefully this will facilitate adoption of AF_XDP, > >>> make > >>> applications using it simpler and smaller, and finally also make it > >>> possible for applications to benefit from optimizations in the > >>> AF_XDP > >>> user space access code. Previously, people just copied and pasted > >>> the > >>> code from the sample application into their application, which is > >>> not > >>> desirable. > >> > >> I like the idea of encapsulating the boilerplate logic in a library. > >> > >> I do think there is an important missing piece though - there should > >> be > >> some code which queries the netdev for how many queues are attached, > >> and > >> create the appropriate number of umem/AF_XDP sockets. > >> > >> I ran into this issue when testing the current AF_XDP code - on my > >> test > >> boxes, the mlx5 card has 55 channels (aka queues), so when the test > >> program > >> binds only to channel 0, nothing works as expected, since not all > >> traffic > >> is being intercepted. While obvious in hindsight, this took a while > >> to > >> track down. > > > > Yes, agreed. You are not the first one to stumble upon this problem > > :-). Let me think a little bit on how to solve this in a good way. We > > need this to be simple and intuitive, as you say. > > Has any investigation been done on using some variant of MPSC > implementation > as an intermediate form for AF_XDP? E.g.: something like LCRQ or the > bulkQ > in bpf devmap/cpumap. I'm aware that this would be slightly slower, as > it > would introduce a lock in the path, but I'd think that having DEVMAP, > CPUMAP > and XSKMAP all behave the same way would add more flexibility. Not as far as I know. But adding the option of having a MPSC or even MPMC queues has been on the todo list for a while, however, the current focus of Björn and myself is to upstream the performance improvements from the Linux Plumbers paper, improve ease-of-use, and help Jesper et al. with the per-queue XDP program implementation (which will increase both performance and ease-of-use). If anyone has some spare cycles out there, please go ahead and give MPSC or MPMC queues a try :-). /Magnus > Ideally, if the configuration matches the underlying hardware, then the > implementation would reduce to the current setup (and allow ZC > implementations), > but a non-matching configuration would still work - as opposed to the > current > situation. > -- > Jonathan
On 02/13/2019 12:55 PM, Jesper Dangaard Brouer wrote: > On Wed, 13 Feb 2019 12:32:47 +0100 > Magnus Karlsson <magnus.karlsson@gmail.com> wrote: >> On Mon, Feb 11, 2019 at 9:44 PM Jonathan Lemon <jonathan.lemon@gmail.com> wrote: >>> On 8 Feb 2019, at 5:05, Magnus Karlsson wrote: >>> >>>> This patch proposes to add AF_XDP support to libbpf. The main reason >>>> for this is to facilitate writing applications that use AF_XDP by >>>> offering higher-level APIs that hide many of the details of the AF_XDP >>>> uapi. This is in the same vein as libbpf facilitates XDP adoption by >>>> offering easy-to-use higher level interfaces of XDP >>>> functionality. Hopefully this will facilitate adoption of AF_XDP, make >>>> applications using it simpler and smaller, and finally also make it >>>> possible for applications to benefit from optimizations in the AF_XDP >>>> user space access code. Previously, people just copied and pasted the >>>> code from the sample application into their application, which is not >>>> desirable. >>> >>> I like the idea of encapsulating the boilerplate logic in a library. >>> >>> I do think there is an important missing piece though - there should be >>> some code which queries the netdev for how many queues are attached, and >>> create the appropriate number of umem/AF_XDP sockets. >>> >>> I ran into this issue when testing the current AF_XDP code - on my test >>> boxes, the mlx5 card has 55 channels (aka queues), so when the test program >>> binds only to channel 0, nothing works as expected, since not all traffic >>> is being intercepted. While obvious in hindsight, this took a while to >>> track down. >> >> Yes, agreed. You are not the first one to stumble upon this problem >> :-). Let me think a little bit on how to solve this in a good way. We >> need this to be simple and intuitive, as you say. > > I see people hitting this with AF_XDP all the time... I had some > backup-slides[2] in our FOSDEM presentation[1] that describe the issue, > give the performance reason why and propose a workaround. Magnus, I presume you're going to address this for the initial libbpf merge since the plan is to make it easier to consume for users? Few more minor items in individual patches, will reply there. Thanks, Daniel > [1] https://github.com/xdp-project/xdp-project/tree/master/conference/FOSDEM2019 > [2] https://github.com/xdp-project/xdp-project/blob/master/conference/FOSDEM2019/xdp_building_block.org#backup-slides > > Alternative work-around > * Create as many AF_XDP sockets as RXQs > * Have userspace poll()/select on all sockets >
On Fri, Feb 15, 2019 at 5:48 PM Daniel Borkmann <daniel@iogearbox.net> wrote: > > On 02/13/2019 12:55 PM, Jesper Dangaard Brouer wrote: > > On Wed, 13 Feb 2019 12:32:47 +0100 > > Magnus Karlsson <magnus.karlsson@gmail.com> wrote: > >> On Mon, Feb 11, 2019 at 9:44 PM Jonathan Lemon <jonathan.lemon@gmail.com> wrote: > >>> On 8 Feb 2019, at 5:05, Magnus Karlsson wrote: > >>> > >>>> This patch proposes to add AF_XDP support to libbpf. The main reason > >>>> for this is to facilitate writing applications that use AF_XDP by > >>>> offering higher-level APIs that hide many of the details of the AF_XDP > >>>> uapi. This is in the same vein as libbpf facilitates XDP adoption by > >>>> offering easy-to-use higher level interfaces of XDP > >>>> functionality. Hopefully this will facilitate adoption of AF_XDP, make > >>>> applications using it simpler and smaller, and finally also make it > >>>> possible for applications to benefit from optimizations in the AF_XDP > >>>> user space access code. Previously, people just copied and pasted the > >>>> code from the sample application into their application, which is not > >>>> desirable. > >>> > >>> I like the idea of encapsulating the boilerplate logic in a library. > >>> > >>> I do think there is an important missing piece though - there should be > >>> some code which queries the netdev for how many queues are attached, and > >>> create the appropriate number of umem/AF_XDP sockets. > >>> > >>> I ran into this issue when testing the current AF_XDP code - on my test > >>> boxes, the mlx5 card has 55 channels (aka queues), so when the test program > >>> binds only to channel 0, nothing works as expected, since not all traffic > >>> is being intercepted. While obvious in hindsight, this took a while to > >>> track down. > >> > >> Yes, agreed. You are not the first one to stumble upon this problem > >> :-). Let me think a little bit on how to solve this in a good way. We > >> need this to be simple and intuitive, as you say. > > > > I see people hitting this with AF_XDP all the time... I had some > > backup-slides[2] in our FOSDEM presentation[1] that describe the issue, > > give the performance reason why and propose a workaround. > > Magnus, I presume you're going to address this for the initial libbpf merge > since the plan is to make it easier to consume for users? I think the first thing we need is education and documentation. Have a FAQ or "common mistakes" section in the Documentation. And of course, sending Jesper around the world reminding people about this ;-). To address this on a libbpf interface level, I think the best way is to reprogram the NIC to send all traffic to the queue that you provided in the xsk_socket__create call. This "set up NIC routing" behavior can then be disable with a flag, just as the XDP program loading can be disabled. The standard config of xsk_socket__create will then set up as many things for the user as possible just to get up and running quickly. More advanced users can then disable parts of it to gain more flexibility. Does this sound OK? Do not want to go the route of polling multiple sockets and aggregating the traffic as this will have significant negative performance implications. /Magnus > Few more minor items in individual patches, will reply there. > > Thanks, > Daniel > > > [1] https://github.com/xdp-project/xdp-project/tree/master/conference/FOSDEM2019 > > [2] https://github.com/xdp-project/xdp-project/blob/master/conference/FOSDEM2019/xdp_building_block.org#backup-slides > > > > Alternative work-around > > * Create as many AF_XDP sockets as RXQs > > * Have userspace poll()/select on all sockets > > >
On 02/18/2019 09:20 AM, Magnus Karlsson wrote: > On Fri, Feb 15, 2019 at 5:48 PM Daniel Borkmann <daniel@iogearbox.net> wrote: >> >> On 02/13/2019 12:55 PM, Jesper Dangaard Brouer wrote: >>> On Wed, 13 Feb 2019 12:32:47 +0100 >>> Magnus Karlsson <magnus.karlsson@gmail.com> wrote: >>>> On Mon, Feb 11, 2019 at 9:44 PM Jonathan Lemon <jonathan.lemon@gmail.com> wrote: >>>>> On 8 Feb 2019, at 5:05, Magnus Karlsson wrote: >>>>> >>>>>> This patch proposes to add AF_XDP support to libbpf. The main reason >>>>>> for this is to facilitate writing applications that use AF_XDP by >>>>>> offering higher-level APIs that hide many of the details of the AF_XDP >>>>>> uapi. This is in the same vein as libbpf facilitates XDP adoption by >>>>>> offering easy-to-use higher level interfaces of XDP >>>>>> functionality. Hopefully this will facilitate adoption of AF_XDP, make >>>>>> applications using it simpler and smaller, and finally also make it >>>>>> possible for applications to benefit from optimizations in the AF_XDP >>>>>> user space access code. Previously, people just copied and pasted the >>>>>> code from the sample application into their application, which is not >>>>>> desirable. >>>>> >>>>> I like the idea of encapsulating the boilerplate logic in a library. >>>>> >>>>> I do think there is an important missing piece though - there should be >>>>> some code which queries the netdev for how many queues are attached, and >>>>> create the appropriate number of umem/AF_XDP sockets. >>>>> >>>>> I ran into this issue when testing the current AF_XDP code - on my test >>>>> boxes, the mlx5 card has 55 channels (aka queues), so when the test program >>>>> binds only to channel 0, nothing works as expected, since not all traffic >>>>> is being intercepted. While obvious in hindsight, this took a while to >>>>> track down. >>>> >>>> Yes, agreed. You are not the first one to stumble upon this problem >>>> :-). Let me think a little bit on how to solve this in a good way. We >>>> need this to be simple and intuitive, as you say. >>> >>> I see people hitting this with AF_XDP all the time... I had some >>> backup-slides[2] in our FOSDEM presentation[1] that describe the issue, >>> give the performance reason why and propose a workaround. >> >> Magnus, I presume you're going to address this for the initial libbpf merge >> since the plan is to make it easier to consume for users? > > I think the first thing we need is education and documentation. Have a > FAQ or "common mistakes" section in the Documentation. And of course, > sending Jesper around the world reminding people about this ;-). > > To address this on a libbpf interface level, I think the best way is > to reprogram the NIC to send all traffic to the queue that you > provided in the xsk_socket__create call. This "set up NIC routing" > behavior can then be disable with a flag, just as the XDP program > loading can be disabled. The standard config of xsk_socket__create > will then set up as many things for the user as possible just to get > up and running quickly. More advanced users can then disable parts of > it to gain more flexibility. Does this sound OK? Do not want to go the > route of polling multiple sockets and aggregating the traffic as this > will have significant negative performance implications. I think that is fine, I would probably make this one a dedicated API call in order to have some more flexibility than just simple flag. E.g. once nfp AF_XDP support lands at some point, I could imagine that this call resp. a drop-in replacement API call for more advanced steering could also take an offloaded BPF prog fd, for example, which then would program the steering on the NIC [0]. Seems at least there's enough complexity on its own to have a dedicated API for it. Thoughts? Thanks, Daniel [0] https://patchwork.ozlabs.org/cover/910614/
On Mon, Feb 18, 2019 at 10:38 AM Daniel Borkmann <daniel@iogearbox.net> wrote: > > On 02/18/2019 09:20 AM, Magnus Karlsson wrote: > > On Fri, Feb 15, 2019 at 5:48 PM Daniel Borkmann <daniel@iogearbox.net> wrote: > >> > >> On 02/13/2019 12:55 PM, Jesper Dangaard Brouer wrote: > >>> On Wed, 13 Feb 2019 12:32:47 +0100 > >>> Magnus Karlsson <magnus.karlsson@gmail.com> wrote: > >>>> On Mon, Feb 11, 2019 at 9:44 PM Jonathan Lemon <jonathan.lemon@gmail.com> wrote: > >>>>> On 8 Feb 2019, at 5:05, Magnus Karlsson wrote: > >>>>> > >>>>>> This patch proposes to add AF_XDP support to libbpf. The main reason > >>>>>> for this is to facilitate writing applications that use AF_XDP by > >>>>>> offering higher-level APIs that hide many of the details of the AF_XDP > >>>>>> uapi. This is in the same vein as libbpf facilitates XDP adoption by > >>>>>> offering easy-to-use higher level interfaces of XDP > >>>>>> functionality. Hopefully this will facilitate adoption of AF_XDP, make > >>>>>> applications using it simpler and smaller, and finally also make it > >>>>>> possible for applications to benefit from optimizations in the AF_XDP > >>>>>> user space access code. Previously, people just copied and pasted the > >>>>>> code from the sample application into their application, which is not > >>>>>> desirable. > >>>>> > >>>>> I like the idea of encapsulating the boilerplate logic in a library. > >>>>> > >>>>> I do think there is an important missing piece though - there should be > >>>>> some code which queries the netdev for how many queues are attached, and > >>>>> create the appropriate number of umem/AF_XDP sockets. > >>>>> > >>>>> I ran into this issue when testing the current AF_XDP code - on my test > >>>>> boxes, the mlx5 card has 55 channels (aka queues), so when the test program > >>>>> binds only to channel 0, nothing works as expected, since not all traffic > >>>>> is being intercepted. While obvious in hindsight, this took a while to > >>>>> track down. > >>>> > >>>> Yes, agreed. You are not the first one to stumble upon this problem > >>>> :-). Let me think a little bit on how to solve this in a good way. We > >>>> need this to be simple and intuitive, as you say. > >>> > >>> I see people hitting this with AF_XDP all the time... I had some > >>> backup-slides[2] in our FOSDEM presentation[1] that describe the issue, > >>> give the performance reason why and propose a workaround. > >> > >> Magnus, I presume you're going to address this for the initial libbpf merge > >> since the plan is to make it easier to consume for users? > > > > I think the first thing we need is education and documentation. Have a > > FAQ or "common mistakes" section in the Documentation. And of course, > > sending Jesper around the world reminding people about this ;-). > > > > To address this on a libbpf interface level, I think the best way is > > to reprogram the NIC to send all traffic to the queue that you > > provided in the xsk_socket__create call. This "set up NIC routing" > > behavior can then be disable with a flag, just as the XDP program > > loading can be disabled. The standard config of xsk_socket__create > > will then set up as many things for the user as possible just to get > > up and running quickly. More advanced users can then disable parts of > > it to gain more flexibility. Does this sound OK? Do not want to go the > > route of polling multiple sockets and aggregating the traffic as this > > will have significant negative performance implications. > > I think that is fine, I would probably make this one a dedicated API call > in order to have some more flexibility than just simple flag. E.g. once > nfp AF_XDP support lands at some point, I could imagine that this call > resp. a drop-in replacement API call for more advanced steering could > also take an offloaded BPF prog fd, for example, which then would program > the steering on the NIC [0]. Seems at least there's enough complexity on > its own to have a dedicated API for it. Thoughts? I agree that there is probably enough complexity to warrant adding a higher level API to deal with this problem (flow steering). But there are likely a number of cases we have not thought that would complicate it even further. This is why I suggest that this functionality should be in its own patch set that I can devote some time and thought to. IMO, the current patch set and functionality does already lower the bar of entry significantly and has a value even without hiding or controlling the steering of traffic. What I would like to do in this patch set is to add a FAQ section in Documentation/networking/af_xdp.rst explaining this problem. Something like: "Q: Why am I not seeing any traffic? A: Check these four things.....". Could add some text in the libbpf README referring to this document also. Opinions? Thanks: Magnus > Thanks, > Daniel > > [0] https://patchwork.ozlabs.org/cover/910614/
This patch proposes to add AF_XDP support to libbpf. The main reason for this is to facilitate writing applications that use AF_XDP by offering higher-level APIs that hide many of the details of the AF_XDP uapi. This is in the same vein as libbpf facilitates XDP adoption by offering easy-to-use higher level interfaces of XDP functionality. Hopefully this will facilitate adoption of AF_XDP, make applications using it simpler and smaller, and finally also make it possible for applications to benefit from optimizations in the AF_XDP user space access code. Previously, people just copied and pasted the code from the sample application into their application, which is not desirable. The proposed interface is composed of two parts: * Low-level access interface to the four rings and the packet * High-level control plane interface for creating and setting up umems and AF_XDP sockets. This interface also loads a simple XDP program that routes all traffic on a queue up to the AF_XDP socket. The sample program has been updated to use this new interface and in that process it lost roughly 300 lines of code. I cannot detect any performance degradations due to the use of this library instead of the previous functions that were inlined in the sample application. But I did measure this on a slower machine and not the Broadwell that we normally use. The rings are now called xsk_ring and when a producer operates on it. It is xsk_ring_prod and for a consumer it is xsk_ring_cons. This way we can get some compile time error checking that the rings are used correctly. Comments and contenplations: * The current behaviour is that the library loads an XDP program (if requested to do so) but the clean up of this program is left to the application. It would be possible to implement this cleanup in the library, but it would require state to be kept on netdev level, which there is none at the moment, and the synchronization of this between processes. All this adding complexity. But when we get an XDP program per queue id, then it becomes trivial to also remove the XDP program when the application exits. This proposal from Jesper, Björn and others will also improve the performance of libbpf, since most of the XDP program code can be removed when that feature is supported. * In a future release, I am planning on adding a higher level data plane interface too. This will be based around recvmsg and sendmsg with the use of struct iovec for batching, without the user having to know anything about the underlying four rings of an AF_XDP socket. There will be one semantic difference though from the standard recvmsg and that is that the kernel will fill in the iovecs instead of the application. But the rest should be the same as the libc versions so that application writers feel at home. Patch 1: adds AF_XDP support in libbpf Patch 2: updates the xdpsock sample application to use the libbpf functions. Changes v3 to v4: * Dropped the pr_*() patch in favor of Yonghong Song's patch set * Addressed the review comments of Daniel Borkmann, mainly leaking of file descriptors at clean up and making the data plane APIs all static inline (with the exception of xsk_umem__get_data that uses an internal structure I do not want to expose). * Fixed the netlink callback as suggested by Maciej Fijalkowski. * Removed an unecessary include in the sample program as spotted by Ilia Fillipov. Changes v2 to v3: * Added automatic loading of a simple XDP program that routes all traffic on a queue up to the AF_XDP socket. This program loading can be disabled. * Updated function names to be consistent with the libbpf naming convention * Moved all code to xsk.[ch] * Removed all the XDP program loading code from the sample since this is now done by libbpf * The initialization functions now return a handle as suggested by Alexei * const statements added in the API where applicable. Changes v1 to v2: * Fixed cleanup of library state on error. * Moved API to initial version * Prefixed all public functions by xsk__ instead of xsk_ * Added comment about changed default ring sizes, batch size and umem size in the sample application commit message * The library now only creates an Rx or Tx ring if the respective parameter is != NULL Note that for zero-copy to work on FVL you need the following patch: https://lore.kernel.org/netdev/1548770597-16141-1-git-send-email-magnus.karlsson@intel.com/. For ixgbe, you need a similar patch called found here: https://lore.kernel.org/netdev/CAJ8uoz1GJBmC0GFbURvEzY4kDZZ6C7O9+1F+gV0y=GOMGLobUQ@mail.gmail.com/. I based this patch set on bpf-next commit a4021a3579c5 ("tools/bpf: add log_level to bpf_load_program_attr") Thanks: Magnus Magnus Karlsson (2): libbpf: add support for using AF_XDP sockets samples/bpf: convert xdpsock to use libbpf for AF_XDP access samples/bpf/Makefile | 1 - samples/bpf/xdpsock.h | 11 - samples/bpf/xdpsock_kern.c | 56 --- samples/bpf/xdpsock_user.c | 839 ++++++++++++-------------------------- tools/include/uapi/linux/if_xdp.h | 78 ++++ tools/lib/bpf/Build | 2 +- tools/lib/bpf/Makefile | 5 +- tools/lib/bpf/README.rst | 11 +- tools/lib/bpf/libbpf.map | 7 + tools/lib/bpf/xsk.c | 742 +++++++++++++++++++++++++++++++++ tools/lib/bpf/xsk.h | 205 ++++++++++ 11 files changed, 1306 insertions(+), 651 deletions(-) delete mode 100644 samples/bpf/xdpsock.h delete mode 100644 samples/bpf/xdpsock_kern.c create mode 100644 tools/include/uapi/linux/if_xdp.h create mode 100644 tools/lib/bpf/xsk.c create mode 100644 tools/lib/bpf/xsk.h -- 2.7.4