d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
a6c54e7eea
freebsd-dev/sys/dev/netmap/if_lem_netmap.h

493 lines
14 KiB
C
Raw Normal View History

Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
/*
It is 2014 and we have a new version of netmap. Most relevant features: - netmap emulation on any NIC, even those without native netmap support. On the ixgbe we have measured about 4Mpps/core/queue in this mode, which is still a lot more than with sockets/bpf. - seamless interconnection of VALE switch, NICs and host stack. If you disable accelerations on your NIC (say em0) ifconfig em0 -txcsum -txcsum you can use the VALE switch to connect the NIC and the host stack: vale-ctl -h valeXX:em0 allowing sharing the NIC with other netmap clients. - THE USER API HAS SLIGHTLY CHANGED (head/cur/tail pointers instead of pointers/count as before). This was unavoidable to support, in the future, multiple threads operating on the same rings. Netmap clients require very small source code changes to compile again. On the plus side, the new API should be easier to understand and the internals are a lot simpler. The manual page has been updated extensively to reflect the current features and give some examples. This is the result of work of several people including Giuseppe Lettieri, Vincenzo Maffione, Michio Honda and myself, and has been financially supported by EU projects CHANGE and OPENLAB, from NetApp University Research Fund, NEC, and of course the Universita` di Pisa.
2014-01-06 12:53:15 +00:00
* Copyright (C) 2011-2014 Matteo Landi, Luigi Rizzo. All rights reserved.
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
A bunch of netmap fixes: USERSPACE: 1. add support for devices with different number of rx and tx queues; 2. add better support for zero-copy operation, adding an extra field to the netmap ring to indicate how many buffers we have already processed but not yet released (with help from Eddie Kohler); 3. The two changes above unfortunately require an API change, so while at it add a version field and some spares to the ioctl() argument to help detect mismatches. 4. update the manual page for the two changes above; 5. update sample applications in tools/tools/netmap KERNEL: 1. simplify the internal structures moving the global wait queues to the 'struct netmap_adapter'; 2. simplify the functions that map kring<->nic ring indexes 3. normalize device-specific code, helps mainteinance; 4. start exploring the impact of micro-optimizations (prefetch etc.) in the ixgbe driver. Use 'legacy' descriptors on the tx ring and prefetch slots gives about 20% speedup at 900 MHz. Another 7-10% would come from removing the explict calls to bus_dmamap* in the core (they are effectively NOPs in this case, but it takes expensive load of the per-buffer dma maps to figure out that they are all NULL. Rx performance not investigated. I am postponing the MFC so i can import a few more improvements before merging.
2012-02-27 19:05:01 +00:00
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
/*
* $FreeBSD$
*
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
* netmap support for: lem
1. Fix the handling of link reset while in netmap more. A link reset now is completely transparent for the netmap client: even if the NIC resets its own ring (e.g. restarting from 0), the client will not see any change in the current rx/tx positions, because the driver will keep track of the offset between the two. 2. make the device-specific code more uniform across different drivers There were some inconsistencies in the implementation of the netmap support routines, now drivers have been aligned to a common code structure. 3. import netmap support for ixgbe . This is implemented as a very small patch for ixgbe.c (233 lines, 11 chunks, mostly comments: in total the patch has only 54 lines of new code) , as most of the code is in an external file sys/dev/netmap/ixgbe_netmap.h , following some initial comments from Jack Vogel about making changes less intrusive. (Note, i have emailed Jack multiple times asking if he had comments on this structure of the code; i got no reply so i assume he is fine with it). Support for other drivers (em, lem, re, igb) will come later. "ixgbe" is now the reference driver for netmap support. Both the external file (sys/dev/netmap/ixgbe_netmap.h) and the device-specific patches (in sys/dev/ixgbe/ixgbe.c) are heavily commented and should serve as a reference for other device drivers. Tested on i386 and amd64 with the pkt-gen program in tools/tools/netmap, the sender does 14.88 Mpps at 1050 Mhz and 14.2 Mpps at 900 MHz on an i7-860 with 4 cores and 82599 card. Haven't tried yet more aggressive optimizations such as adding 'prefetch' instructions in the time-critical parts of the code.
2011-12-05 12:06:53 +00:00
*
A bunch of netmap fixes: USERSPACE: 1. add support for devices with different number of rx and tx queues; 2. add better support for zero-copy operation, adding an extra field to the netmap ring to indicate how many buffers we have already processed but not yet released (with help from Eddie Kohler); 3. The two changes above unfortunately require an API change, so while at it add a version field and some spares to the ioctl() argument to help detect mismatches. 4. update the manual page for the two changes above; 5. update sample applications in tools/tools/netmap KERNEL: 1. simplify the internal structures moving the global wait queues to the 'struct netmap_adapter'; 2. simplify the functions that map kring<->nic ring indexes 3. normalize device-specific code, helps mainteinance; 4. start exploring the impact of micro-optimizations (prefetch etc.) in the ixgbe driver. Use 'legacy' descriptors on the tx ring and prefetch slots gives about 20% speedup at 900 MHz. Another 7-10% would come from removing the explict calls to bus_dmamap* in the core (they are effectively NOPs in this case, but it takes expensive load of the per-buffer dma maps to figure out that they are all NULL. Rx performance not investigated. I am postponing the MFC so i can import a few more improvements before merging.
2012-02-27 19:05:01 +00:00
* For details on netmap support please see ixgbe_netmap.h
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
*/
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
#include <net/netmap.h>
#include <sys/selinfo.h>
#include <vm/vm.h>
#include <vm/pmap.h> /* vtophys ? */
#include <dev/netmap/netmap_kern.h>
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
extern int netmap_adaptive_io;
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
/*
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
* Register/unregister. We are already under netmap lock.
1. Fix the handling of link reset while in netmap more. A link reset now is completely transparent for the netmap client: even if the NIC resets its own ring (e.g. restarting from 0), the client will not see any change in the current rx/tx positions, because the driver will keep track of the offset between the two. 2. make the device-specific code more uniform across different drivers There were some inconsistencies in the implementation of the netmap support routines, now drivers have been aligned to a common code structure. 3. import netmap support for ixgbe . This is implemented as a very small patch for ixgbe.c (233 lines, 11 chunks, mostly comments: in total the patch has only 54 lines of new code) , as most of the code is in an external file sys/dev/netmap/ixgbe_netmap.h , following some initial comments from Jack Vogel about making changes less intrusive. (Note, i have emailed Jack multiple times asking if he had comments on this structure of the code; i got no reply so i assume he is fine with it). Support for other drivers (em, lem, re, igb) will come later. "ixgbe" is now the reference driver for netmap support. Both the external file (sys/dev/netmap/ixgbe_netmap.h) and the device-specific patches (in sys/dev/ixgbe/ixgbe.c) are heavily commented and should serve as a reference for other device drivers. Tested on i386 and amd64 with the pkt-gen program in tools/tools/netmap, the sender does 14.88 Mpps at 1050 Mhz and 14.2 Mpps at 900 MHz on an i7-860 with 4 cores and 82599 card. Haven't tried yet more aggressive optimizations such as adding 'prefetch' instructions in the time-critical parts of the code.
2011-12-05 12:06:53 +00:00
*/
static int
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
lem_netmap_reg(struct netmap_adapter *na, int onoff)
1. Fix the handling of link reset while in netmap more. A link reset now is completely transparent for the netmap client: even if the NIC resets its own ring (e.g. restarting from 0), the client will not see any change in the current rx/tx positions, because the driver will keep track of the offset between the two. 2. make the device-specific code more uniform across different drivers There were some inconsistencies in the implementation of the netmap support routines, now drivers have been aligned to a common code structure. 3. import netmap support for ixgbe . This is implemented as a very small patch for ixgbe.c (233 lines, 11 chunks, mostly comments: in total the patch has only 54 lines of new code) , as most of the code is in an external file sys/dev/netmap/ixgbe_netmap.h , following some initial comments from Jack Vogel about making changes less intrusive. (Note, i have emailed Jack multiple times asking if he had comments on this structure of the code; i got no reply so i assume he is fine with it). Support for other drivers (em, lem, re, igb) will come later. "ixgbe" is now the reference driver for netmap support. Both the external file (sys/dev/netmap/ixgbe_netmap.h) and the device-specific patches (in sys/dev/ixgbe/ixgbe.c) are heavily commented and should serve as a reference for other device drivers. Tested on i386 and amd64 with the pkt-gen program in tools/tools/netmap, the sender does 14.88 Mpps at 1050 Mhz and 14.2 Mpps at 900 MHz on an i7-860 with 4 cores and 82599 card. Haven't tried yet more aggressive optimizations such as adding 'prefetch' instructions in the time-critical parts of the code.
2011-12-05 12:06:53 +00:00
{
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
struct ifnet *ifp = na->ifp;
1. Fix the handling of link reset while in netmap more. A link reset now is completely transparent for the netmap client: even if the NIC resets its own ring (e.g. restarting from 0), the client will not see any change in the current rx/tx positions, because the driver will keep track of the offset between the two. 2. make the device-specific code more uniform across different drivers There were some inconsistencies in the implementation of the netmap support routines, now drivers have been aligned to a common code structure. 3. import netmap support for ixgbe . This is implemented as a very small patch for ixgbe.c (233 lines, 11 chunks, mostly comments: in total the patch has only 54 lines of new code) , as most of the code is in an external file sys/dev/netmap/ixgbe_netmap.h , following some initial comments from Jack Vogel about making changes less intrusive. (Note, i have emailed Jack multiple times asking if he had comments on this structure of the code; i got no reply so i assume he is fine with it). Support for other drivers (em, lem, re, igb) will come later. "ixgbe" is now the reference driver for netmap support. Both the external file (sys/dev/netmap/ixgbe_netmap.h) and the device-specific patches (in sys/dev/ixgbe/ixgbe.c) are heavily commented and should serve as a reference for other device drivers. Tested on i386 and amd64 with the pkt-gen program in tools/tools/netmap, the sender does 14.88 Mpps at 1050 Mhz and 14.2 Mpps at 900 MHz on an i7-860 with 4 cores and 82599 card. Haven't tried yet more aggressive optimizations such as adding 'prefetch' instructions in the time-critical parts of the code.
2011-12-05 12:06:53 +00:00
struct adapter *adapter = ifp->if_softc;
update to the latest netmap snapshot. This includes the following: - use separate memory regions for VALE ports - locking fixes - some simplifications in the NIC-specific routines - performance improvements for the VALE switch - some new features in the pkt-gen test program - documentation updates There are small API changes that require programs to be recompiled (NETMAP_API has been bumped so you will detect old binaries at runtime). In particular: - struct netmap_slot now is 16 bytes to support an extra pointer, which may save one data copy when using VALE ports or VMs; - the struct netmap_if has two extra fields; MFC after: 3 days
2013-11-01 21:21:14 +00:00
EM_CORE_LOCK(adapter);
1. Fix the handling of link reset while in netmap more. A link reset now is completely transparent for the netmap client: even if the NIC resets its own ring (e.g. restarting from 0), the client will not see any change in the current rx/tx positions, because the driver will keep track of the offset between the two. 2. make the device-specific code more uniform across different drivers There were some inconsistencies in the implementation of the netmap support routines, now drivers have been aligned to a common code structure. 3. import netmap support for ixgbe . This is implemented as a very small patch for ixgbe.c (233 lines, 11 chunks, mostly comments: in total the patch has only 54 lines of new code) , as most of the code is in an external file sys/dev/netmap/ixgbe_netmap.h , following some initial comments from Jack Vogel about making changes less intrusive. (Note, i have emailed Jack multiple times asking if he had comments on this structure of the code; i got no reply so i assume he is fine with it). Support for other drivers (em, lem, re, igb) will come later. "ixgbe" is now the reference driver for netmap support. Both the external file (sys/dev/netmap/ixgbe_netmap.h) and the device-specific patches (in sys/dev/ixgbe/ixgbe.c) are heavily commented and should serve as a reference for other device drivers. Tested on i386 and amd64 with the pkt-gen program in tools/tools/netmap, the sender does 14.88 Mpps at 1050 Mhz and 14.2 Mpps at 900 MHz on an i7-860 with 4 cores and 82599 card. Haven't tried yet more aggressive optimizations such as adding 'prefetch' instructions in the time-critical parts of the code.
2011-12-05 12:06:53 +00:00
lem_disable_intr(adapter);
/* Tell the stack that the interface is no longer active */
ifp->if_drv_flags &= ~(IFF_DRV_RUNNING | IFF_DRV_OACTIVE);
#ifndef EM_LEGACY_IRQ // XXX do we need this ?
taskqueue_block(adapter->tq);
taskqueue_drain(adapter->tq, &adapter->rxtx_task);
taskqueue_drain(adapter->tq, &adapter->link_task);
#endif /* !EM_LEGCY_IRQ */
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
/* enable or disable flags and callbacks in na and ifp */
if (onoff) {
nm_set_native_flags(na);
1. Fix the handling of link reset while in netmap more. A link reset now is completely transparent for the netmap client: even if the NIC resets its own ring (e.g. restarting from 0), the client will not see any change in the current rx/tx positions, because the driver will keep track of the offset between the two. 2. make the device-specific code more uniform across different drivers There were some inconsistencies in the implementation of the netmap support routines, now drivers have been aligned to a common code structure. 3. import netmap support for ixgbe . This is implemented as a very small patch for ixgbe.c (233 lines, 11 chunks, mostly comments: in total the patch has only 54 lines of new code) , as most of the code is in an external file sys/dev/netmap/ixgbe_netmap.h , following some initial comments from Jack Vogel about making changes less intrusive. (Note, i have emailed Jack multiple times asking if he had comments on this structure of the code; i got no reply so i assume he is fine with it). Support for other drivers (em, lem, re, igb) will come later. "ixgbe" is now the reference driver for netmap support. Both the external file (sys/dev/netmap/ixgbe_netmap.h) and the device-specific patches (in sys/dev/ixgbe/ixgbe.c) are heavily commented and should serve as a reference for other device drivers. Tested on i386 and amd64 with the pkt-gen program in tools/tools/netmap, the sender does 14.88 Mpps at 1050 Mhz and 14.2 Mpps at 900 MHz on an i7-860 with 4 cores and 82599 card. Haven't tried yet more aggressive optimizations such as adding 'prefetch' instructions in the time-critical parts of the code.
2011-12-05 12:06:53 +00:00
} else {
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
nm_clear_native_flags(na);
1. Fix the handling of link reset while in netmap more. A link reset now is completely transparent for the netmap client: even if the NIC resets its own ring (e.g. restarting from 0), the client will not see any change in the current rx/tx positions, because the driver will keep track of the offset between the two. 2. make the device-specific code more uniform across different drivers There were some inconsistencies in the implementation of the netmap support routines, now drivers have been aligned to a common code structure. 3. import netmap support for ixgbe . This is implemented as a very small patch for ixgbe.c (233 lines, 11 chunks, mostly comments: in total the patch has only 54 lines of new code) , as most of the code is in an external file sys/dev/netmap/ixgbe_netmap.h , following some initial comments from Jack Vogel about making changes less intrusive. (Note, i have emailed Jack multiple times asking if he had comments on this structure of the code; i got no reply so i assume he is fine with it). Support for other drivers (em, lem, re, igb) will come later. "ixgbe" is now the reference driver for netmap support. Both the external file (sys/dev/netmap/ixgbe_netmap.h) and the device-specific patches (in sys/dev/ixgbe/ixgbe.c) are heavily commented and should serve as a reference for other device drivers. Tested on i386 and amd64 with the pkt-gen program in tools/tools/netmap, the sender does 14.88 Mpps at 1050 Mhz and 14.2 Mpps at 900 MHz on an i7-860 with 4 cores and 82599 card. Haven't tried yet more aggressive optimizations such as adding 'prefetch' instructions in the time-critical parts of the code.
2011-12-05 12:06:53 +00:00
}
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
lem_init_locked(adapter); /* also enable intr */
1. Fix the handling of link reset while in netmap more. A link reset now is completely transparent for the netmap client: even if the NIC resets its own ring (e.g. restarting from 0), the client will not see any change in the current rx/tx positions, because the driver will keep track of the offset between the two. 2. make the device-specific code more uniform across different drivers There were some inconsistencies in the implementation of the netmap support routines, now drivers have been aligned to a common code structure. 3. import netmap support for ixgbe . This is implemented as a very small patch for ixgbe.c (233 lines, 11 chunks, mostly comments: in total the patch has only 54 lines of new code) , as most of the code is in an external file sys/dev/netmap/ixgbe_netmap.h , following some initial comments from Jack Vogel about making changes less intrusive. (Note, i have emailed Jack multiple times asking if he had comments on this structure of the code; i got no reply so i assume he is fine with it). Support for other drivers (em, lem, re, igb) will come later. "ixgbe" is now the reference driver for netmap support. Both the external file (sys/dev/netmap/ixgbe_netmap.h) and the device-specific patches (in sys/dev/ixgbe/ixgbe.c) are heavily commented and should serve as a reference for other device drivers. Tested on i386 and amd64 with the pkt-gen program in tools/tools/netmap, the sender does 14.88 Mpps at 1050 Mhz and 14.2 Mpps at 900 MHz on an i7-860 with 4 cores and 82599 card. Haven't tried yet more aggressive optimizations such as adding 'prefetch' instructions in the time-critical parts of the code.
2011-12-05 12:06:53 +00:00
#ifndef EM_LEGACY_IRQ
taskqueue_unblock(adapter->tq); // XXX do we need this ?
#endif /* !EM_LEGCY_IRQ */
update to the latest netmap snapshot. This includes the following: - use separate memory regions for VALE ports - locking fixes - some simplifications in the NIC-specific routines - performance improvements for the VALE switch - some new features in the pkt-gen test program - documentation updates There are small API changes that require programs to be recompiled (NETMAP_API has been bumped so you will detect old binaries at runtime). In particular: - struct netmap_slot now is 16 bytes to support an extra pointer, which may save one data copy when using VALE ports or VMs; - the struct netmap_if has two extra fields; MFC after: 3 days
2013-11-01 21:21:14 +00:00
EM_CORE_UNLOCK(adapter);
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
return (ifp->if_drv_flags & IFF_DRV_RUNNING ? 0 : 1);
1. Fix the handling of link reset while in netmap more. A link reset now is completely transparent for the netmap client: even if the NIC resets its own ring (e.g. restarting from 0), the client will not see any change in the current rx/tx positions, because the driver will keep track of the offset between the two. 2. make the device-specific code more uniform across different drivers There were some inconsistencies in the implementation of the netmap support routines, now drivers have been aligned to a common code structure. 3. import netmap support for ixgbe . This is implemented as a very small patch for ixgbe.c (233 lines, 11 chunks, mostly comments: in total the patch has only 54 lines of new code) , as most of the code is in an external file sys/dev/netmap/ixgbe_netmap.h , following some initial comments from Jack Vogel about making changes less intrusive. (Note, i have emailed Jack multiple times asking if he had comments on this structure of the code; i got no reply so i assume he is fine with it). Support for other drivers (em, lem, re, igb) will come later. "ixgbe" is now the reference driver for netmap support. Both the external file (sys/dev/netmap/ixgbe_netmap.h) and the device-specific patches (in sys/dev/ixgbe/ixgbe.c) are heavily commented and should serve as a reference for other device drivers. Tested on i386 and amd64 with the pkt-gen program in tools/tools/netmap, the sender does 14.88 Mpps at 1050 Mhz and 14.2 Mpps at 900 MHz on an i7-860 with 4 cores and 82599 card. Haven't tried yet more aggressive optimizations such as adding 'prefetch' instructions in the time-critical parts of the code.
2011-12-05 12:06:53 +00:00
}
/*
A bunch of netmap fixes: USERSPACE: 1. add support for devices with different number of rx and tx queues; 2. add better support for zero-copy operation, adding an extra field to the netmap ring to indicate how many buffers we have already processed but not yet released (with help from Eddie Kohler); 3. The two changes above unfortunately require an API change, so while at it add a version field and some spares to the ioctl() argument to help detect mismatches. 4. update the manual page for the two changes above; 5. update sample applications in tools/tools/netmap KERNEL: 1. simplify the internal structures moving the global wait queues to the 'struct netmap_adapter'; 2. simplify the functions that map kring<->nic ring indexes 3. normalize device-specific code, helps mainteinance; 4. start exploring the impact of micro-optimizations (prefetch etc.) in the ixgbe driver. Use 'legacy' descriptors on the tx ring and prefetch slots gives about 20% speedup at 900 MHz. Another 7-10% would come from removing the explict calls to bus_dmamap* in the core (they are effectively NOPs in this case, but it takes expensive load of the per-buffer dma maps to figure out that they are all NULL. Rx performance not investigated. I am postponing the MFC so i can import a few more improvements before merging.
2012-02-27 19:05:01 +00:00
* Reconcile kernel and user view of the transmit ring.
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
*/
static int
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
lem_netmap_txsync(struct netmap_kring *kring, int flags)
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
{
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
struct netmap_adapter *na = kring->na;
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
struct ifnet *ifp = na->ifp;
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
struct netmap_ring *ring = kring->ring;
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
u_int nm_i; /* index into the netmap ring */
u_int nic_i; /* index into the NIC ring */
u_int const lim = kring->nkr_num_slots - 1;
It is 2014 and we have a new version of netmap. Most relevant features: - netmap emulation on any NIC, even those without native netmap support. On the ixgbe we have measured about 4Mpps/core/queue in this mode, which is still a lot more than with sockets/bpf. - seamless interconnection of VALE switch, NICs and host stack. If you disable accelerations on your NIC (say em0) ifconfig em0 -txcsum -txcsum you can use the VALE switch to connect the NIC and the host stack: vale-ctl -h valeXX:em0 allowing sharing the NIC with other netmap clients. - THE USER API HAS SLIGHTLY CHANGED (head/cur/tail pointers instead of pointers/count as before). This was unavoidable to support, in the future, multiple threads operating on the same rings. Netmap clients require very small source code changes to compile again. On the plus side, the new API should be easier to understand and the internals are a lot simpler. The manual page has been updated extensively to reflect the current features and give some examples. This is the result of work of several people including Giuseppe Lettieri, Vincenzo Maffione, Michio Honda and myself, and has been financially supported by EU projects CHANGE and OPENLAB, from NetApp University Research Fund, NEC, and of course the Universita` di Pisa.
2014-01-06 12:53:15 +00:00
u_int const head = kring->rhead;
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
/* generate an interrupt approximately every half ring */
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
u_int report_frequency = kring->nkr_num_slots >> 1;
/* device-specific */
struct adapter *adapter = ifp->if_softc;
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
#ifdef NIC_PARAVIRT
struct paravirt_csb *csb = adapter->csb;
uint64_t *csbd = (uint64_t *)(csb + 1);
#endif /* NIC_PARAVIRT */
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
bus_dmamap_sync(adapter->txdma.dma_tag, adapter->txdma.dma_map,
BUS_DMASYNC_POSTREAD);
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
A bunch of netmap fixes: USERSPACE: 1. add support for devices with different number of rx and tx queues; 2. add better support for zero-copy operation, adding an extra field to the netmap ring to indicate how many buffers we have already processed but not yet released (with help from Eddie Kohler); 3. The two changes above unfortunately require an API change, so while at it add a version field and some spares to the ioctl() argument to help detect mismatches. 4. update the manual page for the two changes above; 5. update sample applications in tools/tools/netmap KERNEL: 1. simplify the internal structures moving the global wait queues to the 'struct netmap_adapter'; 2. simplify the functions that map kring<->nic ring indexes 3. normalize device-specific code, helps mainteinance; 4. start exploring the impact of micro-optimizations (prefetch etc.) in the ixgbe driver. Use 'legacy' descriptors on the tx ring and prefetch slots gives about 20% speedup at 900 MHz. Another 7-10% would come from removing the explict calls to bus_dmamap* in the core (they are effectively NOPs in this case, but it takes expensive load of the per-buffer dma maps to figure out that they are all NULL. Rx performance not investigated. I am postponing the MFC so i can import a few more improvements before merging.
2012-02-27 19:05:01 +00:00
/*
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
* First part: process new packets to send.
reduce the differences between these three files. The three drivers (em, lem and igb) are extremely similar, too bad that the structures use different names and we cannot share the code.
2012-02-15 18:59:26 +00:00
*/
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
nm_i = kring->nr_hwcur;
It is 2014 and we have a new version of netmap. Most relevant features: - netmap emulation on any NIC, even those without native netmap support. On the ixgbe we have measured about 4Mpps/core/queue in this mode, which is still a lot more than with sockets/bpf. - seamless interconnection of VALE switch, NICs and host stack. If you disable accelerations on your NIC (say em0) ifconfig em0 -txcsum -txcsum you can use the VALE switch to connect the NIC and the host stack: vale-ctl -h valeXX:em0 allowing sharing the NIC with other netmap clients. - THE USER API HAS SLIGHTLY CHANGED (head/cur/tail pointers instead of pointers/count as before). This was unavoidable to support, in the future, multiple threads operating on the same rings. Netmap clients require very small source code changes to compile again. On the plus side, the new API should be easier to understand and the internals are a lot simpler. The manual page has been updated extensively to reflect the current features and give some examples. This is the result of work of several people including Giuseppe Lettieri, Vincenzo Maffione, Michio Honda and myself, and has been financially supported by EU projects CHANGE and OPENLAB, from NetApp University Research Fund, NEC, and of course the Universita` di Pisa.
2014-01-06 12:53:15 +00:00
if (nm_i != head) { /* we have new packets to send */
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
#ifdef NIC_PARAVIRT
int do_kick = 0;
uint64_t t = 0; // timestamp
int n = head - nm_i;
if (n < 0)
n += lim + 1;
if (csb) {
t = rdtsc(); /* last timestamp */
csbd[16] += t - csbd[0]; /* total Wg */
csbd[17] += n; /* Wg count */
csbd[0] = t;
}
#endif /* NIC_PARAVIRT */
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
nic_i = netmap_idx_k2n(kring, nm_i);
It is 2014 and we have a new version of netmap. Most relevant features: - netmap emulation on any NIC, even those without native netmap support. On the ixgbe we have measured about 4Mpps/core/queue in this mode, which is still a lot more than with sockets/bpf. - seamless interconnection of VALE switch, NICs and host stack. If you disable accelerations on your NIC (say em0) ifconfig em0 -txcsum -txcsum you can use the VALE switch to connect the NIC and the host stack: vale-ctl -h valeXX:em0 allowing sharing the NIC with other netmap clients. - THE USER API HAS SLIGHTLY CHANGED (head/cur/tail pointers instead of pointers/count as before). This was unavoidable to support, in the future, multiple threads operating on the same rings. Netmap clients require very small source code changes to compile again. On the plus side, the new API should be easier to understand and the internals are a lot simpler. The manual page has been updated extensively to reflect the current features and give some examples. This is the result of work of several people including Giuseppe Lettieri, Vincenzo Maffione, Michio Honda and myself, and has been financially supported by EU projects CHANGE and OPENLAB, from NetApp University Research Fund, NEC, and of course the Universita` di Pisa.
2014-01-06 12:53:15 +00:00
while (nm_i != head) {
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
struct netmap_slot *slot = &ring->slot[nm_i];
u_int len = slot->len;
reduce the differences between these three files. The three drivers (em, lem and igb) are extremely similar, too bad that the structures use different names and we cannot share the code.
2012-02-15 18:59:26 +00:00
uint64_t paddr;
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
void *addr = PNMB(na, slot, &paddr);
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
/* device-specific */
struct e1000_tx_desc *curr = &adapter->tx_desc_base[nic_i];
struct em_buffer *txbuf = &adapter->tx_buffer_area[nic_i];
int flags = (slot->flags & NS_REPORT ||
nic_i == 0 || nic_i == report_frequency) ?
E1000_TXD_CMD_RS : 0;
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
NM_CHECK_ADDR_LEN(na, addr, len);
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
if (slot->flags & NS_BUF_CHANGED) {
1. Fix the handling of link reset while in netmap more. A link reset now is completely transparent for the netmap client: even if the NIC resets its own ring (e.g. restarting from 0), the client will not see any change in the current rx/tx positions, because the driver will keep track of the offset between the two. 2. make the device-specific code more uniform across different drivers There were some inconsistencies in the implementation of the netmap support routines, now drivers have been aligned to a common code structure. 3. import netmap support for ixgbe . This is implemented as a very small patch for ixgbe.c (233 lines, 11 chunks, mostly comments: in total the patch has only 54 lines of new code) , as most of the code is in an external file sys/dev/netmap/ixgbe_netmap.h , following some initial comments from Jack Vogel about making changes less intrusive. (Note, i have emailed Jack multiple times asking if he had comments on this structure of the code; i got no reply so i assume he is fine with it). Support for other drivers (em, lem, re, igb) will come later. "ixgbe" is now the reference driver for netmap support. Both the external file (sys/dev/netmap/ixgbe_netmap.h) and the device-specific patches (in sys/dev/ixgbe/ixgbe.c) are heavily commented and should serve as a reference for other device drivers. Tested on i386 and amd64 with the pkt-gen program in tools/tools/netmap, the sender does 14.88 Mpps at 1050 Mhz and 14.2 Mpps at 900 MHz on an i7-860 with 4 cores and 82599 card. Haven't tried yet more aggressive optimizations such as adding 'prefetch' instructions in the time-critical parts of the code.
2011-12-05 12:06:53 +00:00
/* buffer has changed, reload map */
A bunch of netmap fixes: USERSPACE: 1. add support for devices with different number of rx and tx queues; 2. add better support for zero-copy operation, adding an extra field to the netmap ring to indicate how many buffers we have already processed but not yet released (with help from Eddie Kohler); 3. The two changes above unfortunately require an API change, so while at it add a version field and some spares to the ioctl() argument to help detect mismatches. 4. update the manual page for the two changes above; 5. update sample applications in tools/tools/netmap KERNEL: 1. simplify the internal structures moving the global wait queues to the 'struct netmap_adapter'; 2. simplify the functions that map kring<->nic ring indexes 3. normalize device-specific code, helps mainteinance; 4. start exploring the impact of micro-optimizations (prefetch etc.) in the ixgbe driver. Use 'legacy' descriptors on the tx ring and prefetch slots gives about 20% speedup at 900 MHz. Another 7-10% would come from removing the explict calls to bus_dmamap* in the core (they are effectively NOPs in this case, but it takes expensive load of the per-buffer dma maps to figure out that they are all NULL. Rx performance not investigated. I am postponing the MFC so i can import a few more improvements before merging.
2012-02-27 19:05:01 +00:00
curr->buffer_addr = htole64(paddr);
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
netmap_reload_map(na, adapter->txtag, txbuf->map, addr);
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
}
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
slot->flags &= ~(NS_REPORT | NS_BUF_CHANGED);
/* Fill the slot in the NIC ring. */
A bunch of netmap fixes: USERSPACE: 1. add support for devices with different number of rx and tx queues; 2. add better support for zero-copy operation, adding an extra field to the netmap ring to indicate how many buffers we have already processed but not yet released (with help from Eddie Kohler); 3. The two changes above unfortunately require an API change, so while at it add a version field and some spares to the ioctl() argument to help detect mismatches. 4. update the manual page for the two changes above; 5. update sample applications in tools/tools/netmap KERNEL: 1. simplify the internal structures moving the global wait queues to the 'struct netmap_adapter'; 2. simplify the functions that map kring<->nic ring indexes 3. normalize device-specific code, helps mainteinance; 4. start exploring the impact of micro-optimizations (prefetch etc.) in the ixgbe driver. Use 'legacy' descriptors on the tx ring and prefetch slots gives about 20% speedup at 900 MHz. Another 7-10% would come from removing the explict calls to bus_dmamap* in the core (they are effectively NOPs in this case, but it takes expensive load of the per-buffer dma maps to figure out that they are all NULL. Rx performance not investigated. I am postponing the MFC so i can import a few more improvements before merging.
2012-02-27 19:05:01 +00:00
curr->upper.data = 0;
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
curr->lower.data = htole32(adapter->txd_cmd | len |
A bunch of netmap fixes: USERSPACE: 1. add support for devices with different number of rx and tx queues; 2. add better support for zero-copy operation, adding an extra field to the netmap ring to indicate how many buffers we have already processed but not yet released (with help from Eddie Kohler); 3. The two changes above unfortunately require an API change, so while at it add a version field and some spares to the ioctl() argument to help detect mismatches. 4. update the manual page for the two changes above; 5. update sample applications in tools/tools/netmap KERNEL: 1. simplify the internal structures moving the global wait queues to the 'struct netmap_adapter'; 2. simplify the functions that map kring<->nic ring indexes 3. normalize device-specific code, helps mainteinance; 4. start exploring the impact of micro-optimizations (prefetch etc.) in the ixgbe driver. Use 'legacy' descriptors on the tx ring and prefetch slots gives about 20% speedup at 900 MHz. Another 7-10% would come from removing the explict calls to bus_dmamap* in the core (they are effectively NOPs in this case, but it takes expensive load of the per-buffer dma maps to figure out that they are all NULL. Rx performance not investigated. I am postponing the MFC so i can import a few more improvements before merging.
2012-02-27 19:05:01 +00:00
(E1000_TXD_CMD_EOP | flags) );
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
bus_dmamap_sync(adapter->txtag, txbuf->map,
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
BUS_DMASYNC_PREWRITE);
nm_i = nm_next(nm_i, lim);
nic_i = nm_next(nic_i, lim);
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
// XXX might try an early kick
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
}
It is 2014 and we have a new version of netmap. Most relevant features: - netmap emulation on any NIC, even those without native netmap support. On the ixgbe we have measured about 4Mpps/core/queue in this mode, which is still a lot more than with sockets/bpf. - seamless interconnection of VALE switch, NICs and host stack. If you disable accelerations on your NIC (say em0) ifconfig em0 -txcsum -txcsum you can use the VALE switch to connect the NIC and the host stack: vale-ctl -h valeXX:em0 allowing sharing the NIC with other netmap clients. - THE USER API HAS SLIGHTLY CHANGED (head/cur/tail pointers instead of pointers/count as before). This was unavoidable to support, in the future, multiple threads operating on the same rings. Netmap clients require very small source code changes to compile again. On the plus side, the new API should be easier to understand and the internals are a lot simpler. The manual page has been updated extensively to reflect the current features and give some examples. This is the result of work of several people including Giuseppe Lettieri, Vincenzo Maffione, Michio Honda and myself, and has been financially supported by EU projects CHANGE and OPENLAB, from NetApp University Research Fund, NEC, and of course the Universita` di Pisa.
2014-01-06 12:53:15 +00:00
kring->nr_hwcur = head;
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
/* synchronize the NIC ring */
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
bus_dmamap_sync(adapter->txdma.dma_tag, adapter->txdma.dma_map,
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE);
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
#ifdef NIC_PARAVIRT
/* set unconditionally, then also kick if needed */
if (csb) {
t = rdtsc();
if (csb->host_need_txkick == 2) {
/* can compute an update of delta */
int64_t delta = t - csbd[3];
if (delta < 0)
delta = -delta;
if (csbd[8] == 0 || delta < csbd[8]) {
csbd[8] = delta;
csbd[9]++;
}
csbd[10]++;
}
csb->guest_tdt = nic_i;
csbd[18] += t - csbd[0]; // total wp
csbd[19] += n;
}
if (!csb || !csb->guest_csb_on || (csb->host_need_txkick & 1))
do_kick = 1;
if (do_kick)
#endif /* NIC_PARAVIRT */
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
/* (re)start the tx unit up to slot nic_i (excluded) */
E1000_WRITE_REG(&adapter->hw, E1000_TDT(0), nic_i);
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
#ifdef NIC_PARAVIRT
if (do_kick) {
uint64_t t1 = rdtsc();
csbd[20] += t1 - t; // total Np
csbd[21]++;
}
#endif /* NIC_PARAVIRT */
1. Fix the handling of link reset while in netmap more. A link reset now is completely transparent for the netmap client: even if the NIC resets its own ring (e.g. restarting from 0), the client will not see any change in the current rx/tx positions, because the driver will keep track of the offset between the two. 2. make the device-specific code more uniform across different drivers There were some inconsistencies in the implementation of the netmap support routines, now drivers have been aligned to a common code structure. 3. import netmap support for ixgbe . This is implemented as a very small patch for ixgbe.c (233 lines, 11 chunks, mostly comments: in total the patch has only 54 lines of new code) , as most of the code is in an external file sys/dev/netmap/ixgbe_netmap.h , following some initial comments from Jack Vogel about making changes less intrusive. (Note, i have emailed Jack multiple times asking if he had comments on this structure of the code; i got no reply so i assume he is fine with it). Support for other drivers (em, lem, re, igb) will come later. "ixgbe" is now the reference driver for netmap support. Both the external file (sys/dev/netmap/ixgbe_netmap.h) and the device-specific patches (in sys/dev/ixgbe/ixgbe.c) are heavily commented and should serve as a reference for other device drivers. Tested on i386 and amd64 with the pkt-gen program in tools/tools/netmap, the sender does 14.88 Mpps at 1050 Mhz and 14.2 Mpps at 900 MHz on an i7-860 with 4 cores and 82599 card. Haven't tried yet more aggressive optimizations such as adding 'prefetch' instructions in the time-critical parts of the code.
2011-12-05 12:06:53 +00:00
}
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
/*
* Second part: reclaim buffers for completed transmissions.
*/
It is 2014 and we have a new version of netmap. Most relevant features: - netmap emulation on any NIC, even those without native netmap support. On the ixgbe we have measured about 4Mpps/core/queue in this mode, which is still a lot more than with sockets/bpf. - seamless interconnection of VALE switch, NICs and host stack. If you disable accelerations on your NIC (say em0) ifconfig em0 -txcsum -txcsum you can use the VALE switch to connect the NIC and the host stack: vale-ctl -h valeXX:em0 allowing sharing the NIC with other netmap clients. - THE USER API HAS SLIGHTLY CHANGED (head/cur/tail pointers instead of pointers/count as before). This was unavoidable to support, in the future, multiple threads operating on the same rings. Netmap clients require very small source code changes to compile again. On the plus side, the new API should be easier to understand and the internals are a lot simpler. The manual page has been updated extensively to reflect the current features and give some examples. This is the result of work of several people including Giuseppe Lettieri, Vincenzo Maffione, Michio Honda and myself, and has been financially supported by EU projects CHANGE and OPENLAB, from NetApp University Research Fund, NEC, and of course the Universita` di Pisa.
2014-01-06 12:53:15 +00:00
if (ticks != kring->last_reclaim || flags & NAF_FORCE_RECLAIM || nm_kr_txempty(kring)) {
kring->last_reclaim = ticks;
1. Fix the handling of link reset while in netmap more. A link reset now is completely transparent for the netmap client: even if the NIC resets its own ring (e.g. restarting from 0), the client will not see any change in the current rx/tx positions, because the driver will keep track of the offset between the two. 2. make the device-specific code more uniform across different drivers There were some inconsistencies in the implementation of the netmap support routines, now drivers have been aligned to a common code structure. 3. import netmap support for ixgbe . This is implemented as a very small patch for ixgbe.c (233 lines, 11 chunks, mostly comments: in total the patch has only 54 lines of new code) , as most of the code is in an external file sys/dev/netmap/ixgbe_netmap.h , following some initial comments from Jack Vogel about making changes less intrusive. (Note, i have emailed Jack multiple times asking if he had comments on this structure of the code; i got no reply so i assume he is fine with it). Support for other drivers (em, lem, re, igb) will come later. "ixgbe" is now the reference driver for netmap support. Both the external file (sys/dev/netmap/ixgbe_netmap.h) and the device-specific patches (in sys/dev/ixgbe/ixgbe.c) are heavily commented and should serve as a reference for other device drivers. Tested on i386 and amd64 with the pkt-gen program in tools/tools/netmap, the sender does 14.88 Mpps at 1050 Mhz and 14.2 Mpps at 900 MHz on an i7-860 with 4 cores and 82599 card. Haven't tried yet more aggressive optimizations such as adding 'prefetch' instructions in the time-critical parts of the code.
2011-12-05 12:06:53 +00:00
/* record completed transmissions using TDH */
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
#ifdef NIC_PARAVIRT
/* host updates tdh unconditionally, and we have
* no side effects on reads, so we can read from there
* instead of exiting.
*/
if (csb) {
static int drain = 0, nodrain=0, good = 0, bad = 0, fail = 0;
u_int x = adapter->next_tx_to_clean;
csbd[19]++; // XXX count reclaims
nic_i = csb->host_tdh;
if (csb->guest_csb_on) {
if (nic_i == x) {
bad++;
csbd[24]++; // failed reclaims
/* no progress, request kick and retry */
csb->guest_need_txkick = 1;
mb(); // XXX barrier
nic_i = csb->host_tdh;
} else {
good++;
}
if (nic_i != x) {
csb->guest_need_txkick = 2;
if (nic_i == csb->guest_tdt)
drain++;
else
nodrain++;
#if 1
if (netmap_adaptive_io) {
/* new mechanism: last half ring (or so)
* released one slot at a time.
* This effectively makes the system spin.
*
* Take next_to_clean + 1 as a reference.
* tdh must be ahead or equal
* On entry, the logical order is
* x < tdh = nic_i
* We first push tdh up to avoid wraps.
* The limit is tdh-ll (half ring).
* if tdh-256 < x we report x;
* else we report tdh-256
*/
u_int tdh = nic_i;
u_int ll = csbd[15];
u_int delta = lim/8;
if (netmap_adaptive_io == 2 || ll > delta)
csbd[15] = ll = delta;
else if (netmap_adaptive_io == 1 && ll > 1) {
csbd[15]--;
}
if (nic_i >= kring->nkr_num_slots) {
RD(5, "bad nic_i %d on input", nic_i);
}
x = nm_next(x, lim);
if (tdh < x)
tdh += lim + 1;
if (tdh <= x + ll) {
nic_i = x;
csbd[25]++; //report n + 1;
} else {
tdh = nic_i;
if (tdh < ll)
tdh += lim + 1;
nic_i = tdh - ll;
csbd[26]++; // report tdh - ll
}
}
#endif
} else {
/* we stop, count whether we are idle or not */
int bh_active = csb->host_need_txkick & 2 ? 4 : 0;
csbd[27+ csb->host_need_txkick]++;
if (netmap_adaptive_io == 1) {
if (bh_active && csbd[15] > 1)
csbd[15]--;
else if (!bh_active && csbd[15] < lim/2)
csbd[15]++;
}
bad--;
fail++;
}
}
RD(1, "drain %d nodrain %d good %d retry %d fail %d",
drain, nodrain, good, bad, fail);
} else
#endif /* !NIC_PARAVIRT */
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
nic_i = E1000_READ_REG(&adapter->hw, E1000_TDH(0));
if (nic_i >= kring->nkr_num_slots) { /* XXX can it happen ? */
D("TDH wrap %d", nic_i);
nic_i -= kring->nkr_num_slots;
1. Fix the handling of link reset while in netmap more. A link reset now is completely transparent for the netmap client: even if the NIC resets its own ring (e.g. restarting from 0), the client will not see any change in the current rx/tx positions, because the driver will keep track of the offset between the two. 2. make the device-specific code more uniform across different drivers There were some inconsistencies in the implementation of the netmap support routines, now drivers have been aligned to a common code structure. 3. import netmap support for ixgbe . This is implemented as a very small patch for ixgbe.c (233 lines, 11 chunks, mostly comments: in total the patch has only 54 lines of new code) , as most of the code is in an external file sys/dev/netmap/ixgbe_netmap.h , following some initial comments from Jack Vogel about making changes less intrusive. (Note, i have emailed Jack multiple times asking if he had comments on this structure of the code; i got no reply so i assume he is fine with it). Support for other drivers (em, lem, re, igb) will come later. "ixgbe" is now the reference driver for netmap support. Both the external file (sys/dev/netmap/ixgbe_netmap.h) and the device-specific patches (in sys/dev/ixgbe/ixgbe.c) are heavily commented and should serve as a reference for other device drivers. Tested on i386 and amd64 with the pkt-gen program in tools/tools/netmap, the sender does 14.88 Mpps at 1050 Mhz and 14.2 Mpps at 900 MHz on an i7-860 with 4 cores and 82599 card. Haven't tried yet more aggressive optimizations such as adding 'prefetch' instructions in the time-critical parts of the code.
2011-12-05 12:06:53 +00:00
}
It is 2014 and we have a new version of netmap. Most relevant features: - netmap emulation on any NIC, even those without native netmap support. On the ixgbe we have measured about 4Mpps/core/queue in this mode, which is still a lot more than with sockets/bpf. - seamless interconnection of VALE switch, NICs and host stack. If you disable accelerations on your NIC (say em0) ifconfig em0 -txcsum -txcsum you can use the VALE switch to connect the NIC and the host stack: vale-ctl -h valeXX:em0 allowing sharing the NIC with other netmap clients. - THE USER API HAS SLIGHTLY CHANGED (head/cur/tail pointers instead of pointers/count as before). This was unavoidable to support, in the future, multiple threads operating on the same rings. Netmap clients require very small source code changes to compile again. On the plus side, the new API should be easier to understand and the internals are a lot simpler. The manual page has been updated extensively to reflect the current features and give some examples. This is the result of work of several people including Giuseppe Lettieri, Vincenzo Maffione, Michio Honda and myself, and has been financially supported by EU projects CHANGE and OPENLAB, from NetApp University Research Fund, NEC, and of course the Universita` di Pisa.
2014-01-06 12:53:15 +00:00
adapter->next_tx_to_clean = nic_i;
kring->nr_hwtail = nm_prev(netmap_idx_n2k(kring, nic_i), lim);
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
}
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
return 0;
}
/*
1. Fix the handling of link reset while in netmap more. A link reset now is completely transparent for the netmap client: even if the NIC resets its own ring (e.g. restarting from 0), the client will not see any change in the current rx/tx positions, because the driver will keep track of the offset between the two. 2. make the device-specific code more uniform across different drivers There were some inconsistencies in the implementation of the netmap support routines, now drivers have been aligned to a common code structure. 3. import netmap support for ixgbe . This is implemented as a very small patch for ixgbe.c (233 lines, 11 chunks, mostly comments: in total the patch has only 54 lines of new code) , as most of the code is in an external file sys/dev/netmap/ixgbe_netmap.h , following some initial comments from Jack Vogel about making changes less intrusive. (Note, i have emailed Jack multiple times asking if he had comments on this structure of the code; i got no reply so i assume he is fine with it). Support for other drivers (em, lem, re, igb) will come later. "ixgbe" is now the reference driver for netmap support. Both the external file (sys/dev/netmap/ixgbe_netmap.h) and the device-specific patches (in sys/dev/ixgbe/ixgbe.c) are heavily commented and should serve as a reference for other device drivers. Tested on i386 and amd64 with the pkt-gen program in tools/tools/netmap, the sender does 14.88 Mpps at 1050 Mhz and 14.2 Mpps at 900 MHz on an i7-860 with 4 cores and 82599 card. Haven't tried yet more aggressive optimizations such as adding 'prefetch' instructions in the time-critical parts of the code.
2011-12-05 12:06:53 +00:00
* Reconcile kernel and user view of the receive ring.
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
*/
static int
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
lem_netmap_rxsync(struct netmap_kring *kring, int flags)
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
{
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
struct netmap_adapter *na = kring->na;
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
struct ifnet *ifp = na->ifp;
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
struct netmap_ring *ring = kring->ring;
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
u_int nm_i; /* index into the netmap ring */
u_int nic_i; /* index into the NIC ring */
It is 2014 and we have a new version of netmap. Most relevant features: - netmap emulation on any NIC, even those without native netmap support. On the ixgbe we have measured about 4Mpps/core/queue in this mode, which is still a lot more than with sockets/bpf. - seamless interconnection of VALE switch, NICs and host stack. If you disable accelerations on your NIC (say em0) ifconfig em0 -txcsum -txcsum you can use the VALE switch to connect the NIC and the host stack: vale-ctl -h valeXX:em0 allowing sharing the NIC with other netmap clients. - THE USER API HAS SLIGHTLY CHANGED (head/cur/tail pointers instead of pointers/count as before). This was unavoidable to support, in the future, multiple threads operating on the same rings. Netmap clients require very small source code changes to compile again. On the plus side, the new API should be easier to understand and the internals are a lot simpler. The manual page has been updated extensively to reflect the current features and give some examples. This is the result of work of several people including Giuseppe Lettieri, Vincenzo Maffione, Michio Honda and myself, and has been financially supported by EU projects CHANGE and OPENLAB, from NetApp University Research Fund, NEC, and of course the Universita` di Pisa.
2014-01-06 12:53:15 +00:00
u_int n;
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
u_int const lim = kring->nkr_num_slots - 1;
Sync netmap sources with the version in our private tree. This commit contains large contributions from Giuseppe Lettieri and Stefano Garzarella, is partly supported by grants from Verisign and Cisco, and brings in the following: - fix zerocopy monitor ports and introduce copying monitor ports (the latter are lower performance but give access to all traffic in parallel with the application) - exclusive open mode, useful to implement solutions that recover from crashes of the main netmap client (suggested by Patrick Kelsey) - revised memory allocator in preparation for the 'passthrough mode' (ptnetmap) recently presented at bsdcan. ptnetmap is described in S. Garzarella, G. Lettieri, L. Rizzo; Virtual device passthrough for high speed VM networking, ACM/IEEE ANCS 2015, Oakland (CA) May 2015 http://info.iet.unipi.it/~luigi/research.html - fix rx CRC handing on ixl - add module dependencies for netmap when building drivers as modules - minor simplifications to device-specific routines (*txsync, *rxsync) - general code cleanup (remove unused variables, introduce macros to access rings and remove duplicate code, Applications do not need to be recompiled, unless of course they want to use the new features (monitors and exclusive open). Those willing to try this code on stable/10 can just update the sys/dev/netmap/*, sys/net/netmap* with the version in HEAD and apply the small patches to individual device drivers. MFC after: 1 month Sponsored by: (partly) Verisign, Cisco
2015-07-10 05:51:36 +00:00
u_int const head = kring->rhead;
update to the latest netmap snapshot. This includes the following: - use separate memory regions for VALE ports - locking fixes - some simplifications in the NIC-specific routines - performance improvements for the VALE switch - some new features in the pkt-gen test program - documentation updates There are small API changes that require programs to be recompiled (NETMAP_API has been bumped so you will detect old binaries at runtime). In particular: - struct netmap_slot now is 16 bytes to support an extra pointer, which may save one data copy when using VALE ports or VMs; - the struct netmap_if has two extra fields; MFC after: 3 days
2013-11-01 21:21:14 +00:00
int force_update = (flags & NAF_FORCE_READ) || kring->nr_kflags & NKR_PENDINTR;
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
/* device-specific */
struct adapter *adapter = ifp->if_softc;
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
#ifdef NIC_PARAVIRT
struct paravirt_csb *csb = adapter->csb;
uint32_t csb_mode = csb && csb->guest_csb_on;
uint32_t do_host_rxkick = 0;
#endif /* NIC_PARAVIRT */
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
It is 2014 and we have a new version of netmap. Most relevant features: - netmap emulation on any NIC, even those without native netmap support. On the ixgbe we have measured about 4Mpps/core/queue in this mode, which is still a lot more than with sockets/bpf. - seamless interconnection of VALE switch, NICs and host stack. If you disable accelerations on your NIC (say em0) ifconfig em0 -txcsum -txcsum you can use the VALE switch to connect the NIC and the host stack: vale-ctl -h valeXX:em0 allowing sharing the NIC with other netmap clients. - THE USER API HAS SLIGHTLY CHANGED (head/cur/tail pointers instead of pointers/count as before). This was unavoidable to support, in the future, multiple threads operating on the same rings. Netmap clients require very small source code changes to compile again. On the plus side, the new API should be easier to understand and the internals are a lot simpler. The manual page has been updated extensively to reflect the current features and give some examples. This is the result of work of several people including Giuseppe Lettieri, Vincenzo Maffione, Michio Honda and myself, and has been financially supported by EU projects CHANGE and OPENLAB, from NetApp University Research Fund, NEC, and of course the Universita` di Pisa.
2014-01-06 12:53:15 +00:00
if (head > lim)
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
return netmap_ring_reinit(kring);
reduce the differences between these three files. The three drivers (em, lem and igb) are extremely similar, too bad that the structures use different names and we cannot share the code.
2012-02-15 18:59:26 +00:00
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
#ifdef NIC_PARAVIRT
if (csb_mode) {
force_update = 1;
csb->guest_need_rxkick = 0;
}
#endif /* NIC_PARAVIRT */
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
/* XXX check sync modes */
bus_dmamap_sync(adapter->rxdma.dma_tag, adapter->rxdma.dma_map,
BUS_DMASYNC_POSTREAD | BUS_DMASYNC_POSTWRITE);
A bunch of netmap fixes: USERSPACE: 1. add support for devices with different number of rx and tx queues; 2. add better support for zero-copy operation, adding an extra field to the netmap ring to indicate how many buffers we have already processed but not yet released (with help from Eddie Kohler); 3. The two changes above unfortunately require an API change, so while at it add a version field and some spares to the ioctl() argument to help detect mismatches. 4. update the manual page for the two changes above; 5. update sample applications in tools/tools/netmap KERNEL: 1. simplify the internal structures moving the global wait queues to the 'struct netmap_adapter'; 2. simplify the functions that map kring<->nic ring indexes 3. normalize device-specific code, helps mainteinance; 4. start exploring the impact of micro-optimizations (prefetch etc.) in the ixgbe driver. Use 'legacy' descriptors on the tx ring and prefetch slots gives about 20% speedup at 900 MHz. Another 7-10% would come from removing the explict calls to bus_dmamap* in the core (they are effectively NOPs in this case, but it takes expensive load of the per-buffer dma maps to figure out that they are all NULL. Rx performance not investigated. I am postponing the MFC so i can import a few more improvements before merging.
2012-02-27 19:05:01 +00:00
/*
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
* First part: import newly received packets.
reduce the differences between these three files. The three drivers (em, lem and igb) are extremely similar, too bad that the structures use different names and we cannot share the code.
2012-02-15 18:59:26 +00:00
*/
A bunch of netmap fixes: USERSPACE: 1. add support for devices with different number of rx and tx queues; 2. add better support for zero-copy operation, adding an extra field to the netmap ring to indicate how many buffers we have already processed but not yet released (with help from Eddie Kohler); 3. The two changes above unfortunately require an API change, so while at it add a version field and some spares to the ioctl() argument to help detect mismatches. 4. update the manual page for the two changes above; 5. update sample applications in tools/tools/netmap KERNEL: 1. simplify the internal structures moving the global wait queues to the 'struct netmap_adapter'; 2. simplify the functions that map kring<->nic ring indexes 3. normalize device-specific code, helps mainteinance; 4. start exploring the impact of micro-optimizations (prefetch etc.) in the ixgbe driver. Use 'legacy' descriptors on the tx ring and prefetch slots gives about 20% speedup at 900 MHz. Another 7-10% would come from removing the explict calls to bus_dmamap* in the core (they are effectively NOPs in this case, but it takes expensive load of the per-buffer dma maps to figure out that they are all NULL. Rx performance not investigated. I am postponing the MFC so i can import a few more improvements before merging.
2012-02-27 19:05:01 +00:00
if (netmap_no_pendintr || force_update) {
add some definition and driver changes in preparation for two upcoming features: semi-transparent mode: when a device is opened in this mode, the user program will be able to mark slots that must be forwarded to the "other" side (i.e. from NIC to host stack, or viceversa), and the forwarding will occur automatically at the next netmap syscall. This saves the need to open another file descriptor and do the forwarding manually. direct-forwarding mode: when operating with a VALE port, the user can specify in the slot the actual destination port, overriding the forwarding decision made by a lookup of the destination MAC. This can be useful to implement packet dispatchers. No API changes will be introduced. No new functionality in this patch yet.
2013-01-17 22:14:58 +00:00
uint16_t slot_flags = kring->nkr_slot_flags;
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
nic_i = adapter->next_rx_desc_to_check;
nm_i = netmap_idx_n2k(kring, nic_i);
A bunch of netmap fixes: USERSPACE: 1. add support for devices with different number of rx and tx queues; 2. add better support for zero-copy operation, adding an extra field to the netmap ring to indicate how many buffers we have already processed but not yet released (with help from Eddie Kohler); 3. The two changes above unfortunately require an API change, so while at it add a version field and some spares to the ioctl() argument to help detect mismatches. 4. update the manual page for the two changes above; 5. update sample applications in tools/tools/netmap KERNEL: 1. simplify the internal structures moving the global wait queues to the 'struct netmap_adapter'; 2. simplify the functions that map kring<->nic ring indexes 3. normalize device-specific code, helps mainteinance; 4. start exploring the impact of micro-optimizations (prefetch etc.) in the ixgbe driver. Use 'legacy' descriptors on the tx ring and prefetch slots gives about 20% speedup at 900 MHz. Another 7-10% would come from removing the explict calls to bus_dmamap* in the core (they are effectively NOPs in this case, but it takes expensive load of the per-buffer dma maps to figure out that they are all NULL. Rx performance not investigated. I am postponing the MFC so i can import a few more improvements before merging.
2012-02-27 19:05:01 +00:00
for (n = 0; ; n++) {
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
struct e1000_rx_desc *curr = &adapter->rx_desc_base[nic_i];
A bunch of netmap fixes: USERSPACE: 1. add support for devices with different number of rx and tx queues; 2. add better support for zero-copy operation, adding an extra field to the netmap ring to indicate how many buffers we have already processed but not yet released (with help from Eddie Kohler); 3. The two changes above unfortunately require an API change, so while at it add a version field and some spares to the ioctl() argument to help detect mismatches. 4. update the manual page for the two changes above; 5. update sample applications in tools/tools/netmap KERNEL: 1. simplify the internal structures moving the global wait queues to the 'struct netmap_adapter'; 2. simplify the functions that map kring<->nic ring indexes 3. normalize device-specific code, helps mainteinance; 4. start exploring the impact of micro-optimizations (prefetch etc.) in the ixgbe driver. Use 'legacy' descriptors on the tx ring and prefetch slots gives about 20% speedup at 900 MHz. Another 7-10% would come from removing the explict calls to bus_dmamap* in the core (they are effectively NOPs in this case, but it takes expensive load of the per-buffer dma maps to figure out that they are all NULL. Rx performance not investigated. I am postponing the MFC so i can import a few more improvements before merging.
2012-02-27 19:05:01 +00:00
uint32_t staterr = le32toh(curr->status);
int len;
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
#ifdef NIC_PARAVIRT
if (csb_mode) {
if ((staterr & E1000_RXD_STAT_DD) == 0) {
/* don't bother to retry if more than 1 pkt */
if (n > 1)
break;
csb->guest_need_rxkick = 1;
wmb();
staterr = le32toh(curr->status);
if ((staterr & E1000_RXD_STAT_DD) == 0) {
break;
} else { /* we are good */
csb->guest_need_rxkick = 0;
}
}
} else
#endif /* NIC_PARAVIRT */
A bunch of netmap fixes: USERSPACE: 1. add support for devices with different number of rx and tx queues; 2. add better support for zero-copy operation, adding an extra field to the netmap ring to indicate how many buffers we have already processed but not yet released (with help from Eddie Kohler); 3. The two changes above unfortunately require an API change, so while at it add a version field and some spares to the ioctl() argument to help detect mismatches. 4. update the manual page for the two changes above; 5. update sample applications in tools/tools/netmap KERNEL: 1. simplify the internal structures moving the global wait queues to the 'struct netmap_adapter'; 2. simplify the functions that map kring<->nic ring indexes 3. normalize device-specific code, helps mainteinance; 4. start exploring the impact of micro-optimizations (prefetch etc.) in the ixgbe driver. Use 'legacy' descriptors on the tx ring and prefetch slots gives about 20% speedup at 900 MHz. Another 7-10% would come from removing the explict calls to bus_dmamap* in the core (they are effectively NOPs in this case, but it takes expensive load of the per-buffer dma maps to figure out that they are all NULL. Rx performance not investigated. I am postponing the MFC so i can import a few more improvements before merging.
2012-02-27 19:05:01 +00:00
if ((staterr & E1000_RXD_STAT_DD) == 0)
break;
len = le16toh(curr->length) - 4; // CRC
if (len < 0) {
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
RD(5, "bogus pkt (%d) size %d nic idx %d", n, len, nic_i);
A bunch of netmap fixes: USERSPACE: 1. add support for devices with different number of rx and tx queues; 2. add better support for zero-copy operation, adding an extra field to the netmap ring to indicate how many buffers we have already processed but not yet released (with help from Eddie Kohler); 3. The two changes above unfortunately require an API change, so while at it add a version field and some spares to the ioctl() argument to help detect mismatches. 4. update the manual page for the two changes above; 5. update sample applications in tools/tools/netmap KERNEL: 1. simplify the internal structures moving the global wait queues to the 'struct netmap_adapter'; 2. simplify the functions that map kring<->nic ring indexes 3. normalize device-specific code, helps mainteinance; 4. start exploring the impact of micro-optimizations (prefetch etc.) in the ixgbe driver. Use 'legacy' descriptors on the tx ring and prefetch slots gives about 20% speedup at 900 MHz. Another 7-10% would come from removing the explict calls to bus_dmamap* in the core (they are effectively NOPs in this case, but it takes expensive load of the per-buffer dma maps to figure out that they are all NULL. Rx performance not investigated. I am postponing the MFC so i can import a few more improvements before merging.
2012-02-27 19:05:01 +00:00
len = 0;
}
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
ring->slot[nm_i].len = len;
ring->slot[nm_i].flags = slot_flags;
A bunch of netmap fixes: USERSPACE: 1. add support for devices with different number of rx and tx queues; 2. add better support for zero-copy operation, adding an extra field to the netmap ring to indicate how many buffers we have already processed but not yet released (with help from Eddie Kohler); 3. The two changes above unfortunately require an API change, so while at it add a version field and some spares to the ioctl() argument to help detect mismatches. 4. update the manual page for the two changes above; 5. update sample applications in tools/tools/netmap KERNEL: 1. simplify the internal structures moving the global wait queues to the 'struct netmap_adapter'; 2. simplify the functions that map kring<->nic ring indexes 3. normalize device-specific code, helps mainteinance; 4. start exploring the impact of micro-optimizations (prefetch etc.) in the ixgbe driver. Use 'legacy' descriptors on the tx ring and prefetch slots gives about 20% speedup at 900 MHz. Another 7-10% would come from removing the explict calls to bus_dmamap* in the core (they are effectively NOPs in this case, but it takes expensive load of the per-buffer dma maps to figure out that they are all NULL. Rx performance not investigated. I am postponing the MFC so i can import a few more improvements before merging.
2012-02-27 19:05:01 +00:00
bus_dmamap_sync(adapter->rxtag,
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
adapter->rx_buffer_area[nic_i].map,
BUS_DMASYNC_POSTREAD);
nm_i = nm_next(nm_i, lim);
nic_i = nm_next(nic_i, lim);
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
}
A bunch of netmap fixes: USERSPACE: 1. add support for devices with different number of rx and tx queues; 2. add better support for zero-copy operation, adding an extra field to the netmap ring to indicate how many buffers we have already processed but not yet released (with help from Eddie Kohler); 3. The two changes above unfortunately require an API change, so while at it add a version field and some spares to the ioctl() argument to help detect mismatches. 4. update the manual page for the two changes above; 5. update sample applications in tools/tools/netmap KERNEL: 1. simplify the internal structures moving the global wait queues to the 'struct netmap_adapter'; 2. simplify the functions that map kring<->nic ring indexes 3. normalize device-specific code, helps mainteinance; 4. start exploring the impact of micro-optimizations (prefetch etc.) in the ixgbe driver. Use 'legacy' descriptors on the tx ring and prefetch slots gives about 20% speedup at 900 MHz. Another 7-10% would come from removing the explict calls to bus_dmamap* in the core (they are effectively NOPs in this case, but it takes expensive load of the per-buffer dma maps to figure out that they are all NULL. Rx performance not investigated. I am postponing the MFC so i can import a few more improvements before merging.
2012-02-27 19:05:01 +00:00
if (n) { /* update the state variables */
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
#ifdef NIC_PARAVIRT
if (csb_mode) {
if (n > 1) {
/* leave one spare buffer so we avoid rxkicks */
nm_i = nm_prev(nm_i, lim);
nic_i = nm_prev(nic_i, lim);
n--;
} else {
csb->guest_need_rxkick = 1;
}
}
#endif /* NIC_PARAVIRT */
It is 2014 and we have a new version of netmap. Most relevant features: - netmap emulation on any NIC, even those without native netmap support. On the ixgbe we have measured about 4Mpps/core/queue in this mode, which is still a lot more than with sockets/bpf. - seamless interconnection of VALE switch, NICs and host stack. If you disable accelerations on your NIC (say em0) ifconfig em0 -txcsum -txcsum you can use the VALE switch to connect the NIC and the host stack: vale-ctl -h valeXX:em0 allowing sharing the NIC with other netmap clients. - THE USER API HAS SLIGHTLY CHANGED (head/cur/tail pointers instead of pointers/count as before). This was unavoidable to support, in the future, multiple threads operating on the same rings. Netmap clients require very small source code changes to compile again. On the plus side, the new API should be easier to understand and the internals are a lot simpler. The manual page has been updated extensively to reflect the current features and give some examples. This is the result of work of several people including Giuseppe Lettieri, Vincenzo Maffione, Michio Honda and myself, and has been financially supported by EU projects CHANGE and OPENLAB, from NetApp University Research Fund, NEC, and of course the Universita` di Pisa.
2014-01-06 12:53:15 +00:00
ND("%d new packets at nic %d nm %d tail %d",
n,
adapter->next_rx_desc_to_check,
netmap_idx_n2k(kring, adapter->next_rx_desc_to_check),
kring->nr_hwtail);
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
adapter->next_rx_desc_to_check = nic_i;
Mechanically convert to if_inc_counter().
2014-09-19 03:51:26 +00:00
// if_inc_counter(ifp, IFCOUNTER_IPACKETS, n);
It is 2014 and we have a new version of netmap. Most relevant features: - netmap emulation on any NIC, even those without native netmap support. On the ixgbe we have measured about 4Mpps/core/queue in this mode, which is still a lot more than with sockets/bpf. - seamless interconnection of VALE switch, NICs and host stack. If you disable accelerations on your NIC (say em0) ifconfig em0 -txcsum -txcsum you can use the VALE switch to connect the NIC and the host stack: vale-ctl -h valeXX:em0 allowing sharing the NIC with other netmap clients. - THE USER API HAS SLIGHTLY CHANGED (head/cur/tail pointers instead of pointers/count as before). This was unavoidable to support, in the future, multiple threads operating on the same rings. Netmap clients require very small source code changes to compile again. On the plus side, the new API should be easier to understand and the internals are a lot simpler. The manual page has been updated extensively to reflect the current features and give some examples. This is the result of work of several people including Giuseppe Lettieri, Vincenzo Maffione, Michio Honda and myself, and has been financially supported by EU projects CHANGE and OPENLAB, from NetApp University Research Fund, NEC, and of course the Universita` di Pisa.
2014-01-06 12:53:15 +00:00
kring->nr_hwtail = nm_i;
A bunch of netmap fixes: USERSPACE: 1. add support for devices with different number of rx and tx queues; 2. add better support for zero-copy operation, adding an extra field to the netmap ring to indicate how many buffers we have already processed but not yet released (with help from Eddie Kohler); 3. The two changes above unfortunately require an API change, so while at it add a version field and some spares to the ioctl() argument to help detect mismatches. 4. update the manual page for the two changes above; 5. update sample applications in tools/tools/netmap KERNEL: 1. simplify the internal structures moving the global wait queues to the 'struct netmap_adapter'; 2. simplify the functions that map kring<->nic ring indexes 3. normalize device-specific code, helps mainteinance; 4. start exploring the impact of micro-optimizations (prefetch etc.) in the ixgbe driver. Use 'legacy' descriptors on the tx ring and prefetch slots gives about 20% speedup at 900 MHz. Another 7-10% would come from removing the explict calls to bus_dmamap* in the core (they are effectively NOPs in this case, but it takes expensive load of the per-buffer dma maps to figure out that they are all NULL. Rx performance not investigated. I am postponing the MFC so i can import a few more improvements before merging.
2012-02-27 19:05:01 +00:00
}
kring->nr_kflags &= ~NKR_PENDINTR;
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
}
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
/*
* Second part: skip past packets that userspace has released.
*/
nm_i = kring->nr_hwcur;
It is 2014 and we have a new version of netmap. Most relevant features: - netmap emulation on any NIC, even those without native netmap support. On the ixgbe we have measured about 4Mpps/core/queue in this mode, which is still a lot more than with sockets/bpf. - seamless interconnection of VALE switch, NICs and host stack. If you disable accelerations on your NIC (say em0) ifconfig em0 -txcsum -txcsum you can use the VALE switch to connect the NIC and the host stack: vale-ctl -h valeXX:em0 allowing sharing the NIC with other netmap clients. - THE USER API HAS SLIGHTLY CHANGED (head/cur/tail pointers instead of pointers/count as before). This was unavoidable to support, in the future, multiple threads operating on the same rings. Netmap clients require very small source code changes to compile again. On the plus side, the new API should be easier to understand and the internals are a lot simpler. The manual page has been updated extensively to reflect the current features and give some examples. This is the result of work of several people including Giuseppe Lettieri, Vincenzo Maffione, Michio Honda and myself, and has been financially supported by EU projects CHANGE and OPENLAB, from NetApp University Research Fund, NEC, and of course the Universita` di Pisa.
2014-01-06 12:53:15 +00:00
if (nm_i != head) {
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
nic_i = netmap_idx_k2n(kring, nm_i);
It is 2014 and we have a new version of netmap. Most relevant features: - netmap emulation on any NIC, even those without native netmap support. On the ixgbe we have measured about 4Mpps/core/queue in this mode, which is still a lot more than with sockets/bpf. - seamless interconnection of VALE switch, NICs and host stack. If you disable accelerations on your NIC (say em0) ifconfig em0 -txcsum -txcsum you can use the VALE switch to connect the NIC and the host stack: vale-ctl -h valeXX:em0 allowing sharing the NIC with other netmap clients. - THE USER API HAS SLIGHTLY CHANGED (head/cur/tail pointers instead of pointers/count as before). This was unavoidable to support, in the future, multiple threads operating on the same rings. Netmap clients require very small source code changes to compile again. On the plus side, the new API should be easier to understand and the internals are a lot simpler. The manual page has been updated extensively to reflect the current features and give some examples. This is the result of work of several people including Giuseppe Lettieri, Vincenzo Maffione, Michio Honda and myself, and has been financially supported by EU projects CHANGE and OPENLAB, from NetApp University Research Fund, NEC, and of course the Universita` di Pisa.
2014-01-06 12:53:15 +00:00
for (n = 0; nm_i != head; n++) {
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
struct netmap_slot *slot = &ring->slot[nm_i];
small code cleanup in preparation for future modifications in the memory allocator used by netmap. No functional change, two small bug fixes: - in if_re.c add a missing bus_dmamap_sync() - in netmap.c comment out a spurious free() in an error handling block
2012-01-10 19:57:23 +00:00
uint64_t paddr;
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
void *addr = PNMB(na, slot, &paddr);
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
struct e1000_rx_desc *curr = &adapter->rx_desc_base[nic_i];
struct em_buffer *rxbuf = &adapter->rx_buffer_area[nic_i];
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
if (addr == NETMAP_BUF_BASE(na)) /* bad buf */
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
goto ring_reset;
reduce the differences between these three files. The three drivers (em, lem and igb) are extremely similar, too bad that the structures use different names and we cannot share the code.
2012-02-15 18:59:26 +00:00
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
if (slot->flags & NS_BUF_CHANGED) {
(This commit only touches code within the DEV_NETMAP blocks) Introduce some functions to map NIC ring indexes into netmap ring indexes and vice versa. This way we can implement the bound checks only in one place (and hopefully in a correct way). On passing, make the code and comments more uniform across the various drivers.
2012-02-15 23:13:29 +00:00
/* buffer has changed, reload map */
A bunch of netmap fixes: USERSPACE: 1. add support for devices with different number of rx and tx queues; 2. add better support for zero-copy operation, adding an extra field to the netmap ring to indicate how many buffers we have already processed but not yet released (with help from Eddie Kohler); 3. The two changes above unfortunately require an API change, so while at it add a version field and some spares to the ioctl() argument to help detect mismatches. 4. update the manual page for the two changes above; 5. update sample applications in tools/tools/netmap KERNEL: 1. simplify the internal structures moving the global wait queues to the 'struct netmap_adapter'; 2. simplify the functions that map kring<->nic ring indexes 3. normalize device-specific code, helps mainteinance; 4. start exploring the impact of micro-optimizations (prefetch etc.) in the ixgbe driver. Use 'legacy' descriptors on the tx ring and prefetch slots gives about 20% speedup at 900 MHz. Another 7-10% would come from removing the explict calls to bus_dmamap* in the core (they are effectively NOPs in this case, but it takes expensive load of the per-buffer dma maps to figure out that they are all NULL. Rx performance not investigated. I am postponing the MFC so i can import a few more improvements before merging.
2012-02-27 19:05:01 +00:00
curr->buffer_addr = htole64(paddr);
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
netmap_reload_map(na, adapter->rxtag, rxbuf->map, addr);
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
slot->flags &= ~NS_BUF_CHANGED;
}
A bunch of netmap fixes: USERSPACE: 1. add support for devices with different number of rx and tx queues; 2. add better support for zero-copy operation, adding an extra field to the netmap ring to indicate how many buffers we have already processed but not yet released (with help from Eddie Kohler); 3. The two changes above unfortunately require an API change, so while at it add a version field and some spares to the ioctl() argument to help detect mismatches. 4. update the manual page for the two changes above; 5. update sample applications in tools/tools/netmap KERNEL: 1. simplify the internal structures moving the global wait queues to the 'struct netmap_adapter'; 2. simplify the functions that map kring<->nic ring indexes 3. normalize device-specific code, helps mainteinance; 4. start exploring the impact of micro-optimizations (prefetch etc.) in the ixgbe driver. Use 'legacy' descriptors on the tx ring and prefetch slots gives about 20% speedup at 900 MHz. Another 7-10% would come from removing the explict calls to bus_dmamap* in the core (they are effectively NOPs in this case, but it takes expensive load of the per-buffer dma maps to figure out that they are all NULL. Rx performance not investigated. I am postponing the MFC so i can import a few more improvements before merging.
2012-02-27 19:05:01 +00:00
curr->status = 0;
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
bus_dmamap_sync(adapter->rxtag, rxbuf->map,
1. Fix the handling of link reset while in netmap more. A link reset now is completely transparent for the netmap client: even if the NIC resets its own ring (e.g. restarting from 0), the client will not see any change in the current rx/tx positions, because the driver will keep track of the offset between the two. 2. make the device-specific code more uniform across different drivers There were some inconsistencies in the implementation of the netmap support routines, now drivers have been aligned to a common code structure. 3. import netmap support for ixgbe . This is implemented as a very small patch for ixgbe.c (233 lines, 11 chunks, mostly comments: in total the patch has only 54 lines of new code) , as most of the code is in an external file sys/dev/netmap/ixgbe_netmap.h , following some initial comments from Jack Vogel about making changes less intrusive. (Note, i have emailed Jack multiple times asking if he had comments on this structure of the code; i got no reply so i assume he is fine with it). Support for other drivers (em, lem, re, igb) will come later. "ixgbe" is now the reference driver for netmap support. Both the external file (sys/dev/netmap/ixgbe_netmap.h) and the device-specific patches (in sys/dev/ixgbe/ixgbe.c) are heavily commented and should serve as a reference for other device drivers. Tested on i386 and amd64 with the pkt-gen program in tools/tools/netmap, the sender does 14.88 Mpps at 1050 Mhz and 14.2 Mpps at 900 MHz on an i7-860 with 4 cores and 82599 card. Haven't tried yet more aggressive optimizations such as adding 'prefetch' instructions in the time-critical parts of the code.
2011-12-05 12:06:53 +00:00
BUS_DMASYNC_PREREAD);
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
#ifdef NIC_PARAVIRT
if (csb_mode && csb->host_rxkick_at == nic_i)
do_host_rxkick = 1;
#endif /* NIC_PARAVIRT */
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
nm_i = nm_next(nm_i, lim);
nic_i = nm_next(nic_i, lim);
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
}
It is 2014 and we have a new version of netmap. Most relevant features: - netmap emulation on any NIC, even those without native netmap support. On the ixgbe we have measured about 4Mpps/core/queue in this mode, which is still a lot more than with sockets/bpf. - seamless interconnection of VALE switch, NICs and host stack. If you disable accelerations on your NIC (say em0) ifconfig em0 -txcsum -txcsum you can use the VALE switch to connect the NIC and the host stack: vale-ctl -h valeXX:em0 allowing sharing the NIC with other netmap clients. - THE USER API HAS SLIGHTLY CHANGED (head/cur/tail pointers instead of pointers/count as before). This was unavoidable to support, in the future, multiple threads operating on the same rings. Netmap clients require very small source code changes to compile again. On the plus side, the new API should be easier to understand and the internals are a lot simpler. The manual page has been updated extensively to reflect the current features and give some examples. This is the result of work of several people including Giuseppe Lettieri, Vincenzo Maffione, Michio Honda and myself, and has been financially supported by EU projects CHANGE and OPENLAB, from NetApp University Research Fund, NEC, and of course the Universita` di Pisa.
2014-01-06 12:53:15 +00:00
kring->nr_hwcur = head;
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
bus_dmamap_sync(adapter->rxdma.dma_tag, adapter->rxdma.dma_map,
1. Fix the handling of link reset while in netmap more. A link reset now is completely transparent for the netmap client: even if the NIC resets its own ring (e.g. restarting from 0), the client will not see any change in the current rx/tx positions, because the driver will keep track of the offset between the two. 2. make the device-specific code more uniform across different drivers There were some inconsistencies in the implementation of the netmap support routines, now drivers have been aligned to a common code structure. 3. import netmap support for ixgbe . This is implemented as a very small patch for ixgbe.c (233 lines, 11 chunks, mostly comments: in total the patch has only 54 lines of new code) , as most of the code is in an external file sys/dev/netmap/ixgbe_netmap.h , following some initial comments from Jack Vogel about making changes less intrusive. (Note, i have emailed Jack multiple times asking if he had comments on this structure of the code; i got no reply so i assume he is fine with it). Support for other drivers (em, lem, re, igb) will come later. "ixgbe" is now the reference driver for netmap support. Both the external file (sys/dev/netmap/ixgbe_netmap.h) and the device-specific patches (in sys/dev/ixgbe/ixgbe.c) are heavily commented and should serve as a reference for other device drivers. Tested on i386 and amd64 with the pkt-gen program in tools/tools/netmap, the sender does 14.88 Mpps at 1050 Mhz and 14.2 Mpps at 900 MHz on an i7-860 with 4 cores and 82599 card. Haven't tried yet more aggressive optimizations such as adding 'prefetch' instructions in the time-critical parts of the code.
2011-12-05 12:06:53 +00:00
BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE);
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
/*
* IMPORTANT: we must leave one free slot in the ring,
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
* so move nic_i back by one unit
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
*/
It is 2014 and we have a new version of netmap. Most relevant features: - netmap emulation on any NIC, even those without native netmap support. On the ixgbe we have measured about 4Mpps/core/queue in this mode, which is still a lot more than with sockets/bpf. - seamless interconnection of VALE switch, NICs and host stack. If you disable accelerations on your NIC (say em0) ifconfig em0 -txcsum -txcsum you can use the VALE switch to connect the NIC and the host stack: vale-ctl -h valeXX:em0 allowing sharing the NIC with other netmap clients. - THE USER API HAS SLIGHTLY CHANGED (head/cur/tail pointers instead of pointers/count as before). This was unavoidable to support, in the future, multiple threads operating on the same rings. Netmap clients require very small source code changes to compile again. On the plus side, the new API should be easier to understand and the internals are a lot simpler. The manual page has been updated extensively to reflect the current features and give some examples. This is the result of work of several people including Giuseppe Lettieri, Vincenzo Maffione, Michio Honda and myself, and has been financially supported by EU projects CHANGE and OPENLAB, from NetApp University Research Fund, NEC, and of course the Universita` di Pisa.
2014-01-06 12:53:15 +00:00
nic_i = nm_prev(nic_i, lim);
Update to the current version of netmap. Mostly bugfixes or features developed in the past 6 months, so this is a 10.1 candidate. Basically no user API changes (some bugfixes in sys/net/netmap_user.h). In detail: 1. netmap support for virtio-net, including in netmap mode. Under bhyve and with a netmap backend [2] we reach over 1Mpps with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode. 2. (kernel) add support for multiple memory allocators, so we can better partition physical and virtual interfaces giving access to separate users. The most visible effect is one additional argument to the various kernel functions to compute buffer addresses. All netmap-supported drivers are affected, but changes are mechanical and trivial 3. (kernel) simplify the prototype for *txsync() and *rxsync() driver methods. All netmap drivers affected, changes mostly mechanical. 4. add support for netmap-monitor ports. Think of it as a mirroring port on a physical switch: a netmap monitor port replicates traffic present on the main port. Restrictions apply. Drive carefully. 5. if_lem.c: support for various paravirtualization features, experimental and disabled by default. Most of these are described in our ANCS'13 paper [1]. Paravirtualized support in netmap mode is new, and beats the numbers in the paper by a large factor (under qemu-kvm, we measured gues-host throughput up to 10-12 Mpps). A lot of refactoring and additional documentation in the files in sys/dev/netmap, but apart from #2 and #3 above, almost nothing of this stuff is visible to other kernel parts. Example programs in tools/tools/netmap have been updated with bugfixes and to support more of the existing features. This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline. A lot of this code has been contributed by my colleagues at UNIPI, including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella. MFC after: 3 days.
2014-08-16 15:00:01 +00:00
#ifdef NIC_PARAVIRT
/* set unconditionally, then also kick if needed */
if (csb)
csb->guest_rdt = nic_i;
if (!csb_mode || do_host_rxkick)
#endif /* NIC_PARAVIRT */
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
E1000_WRITE_REG(&adapter->hw, E1000_RDT(0), nic_i);
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
}
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
return 0;
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
ring_reset:
return netmap_ring_reinit(kring);
Bring in support for netmap, a framework for very efficient packet I/O from userspace, capable of line rate at 10G, see http://info.iet.unipi.it/~luigi/netmap/ At this time I am bringing in only the generic code (sys/dev/netmap/ plus two headers under sys/net/), and some sample applications in tools/tools/netmap. There is also a manpage in share/man/man4 [1] In order to make use of the framework you need to build a kernel with "device netmap", and patch individual drivers with the code that you can find in sys/dev/netmap/head.diff The file will go away as the relevant pieces are committed to the various device drivers, which should happen in a few days after talking to the driver maintainers. Netmap support is available at the moment for Intel 10G and 1G cards (ixgbe, em/lem/igb), and for the Realtek 1G card ("re"). I have partial patches for "bge" and am starting to work on "cxgbe". Hopefully changes are trivial enough so interested third parties can submit their patches. Interested people can contact me for advice on how to add netmap support to specific devices. CREDITS: Netmap has been developed by Luigi Rizzo and other collaborators at the Universita` di Pisa, and supported by EU project CHANGE (http://www.change-project.eu/) The code is distributed under a BSD Copyright. [1] In my opinion is a bad idea to have all manpage in one directory. We should place kernel documentation in the same dir that contains the code, which would make it much simpler to keep doc and code in sync, reduce the clutter in share/man/ and incidentally is the policy used for all of userspace code. Makefiles and doc tools can be trivially adjusted to find the manpages in the relevant subdirs.
2011-11-17 12:17:39 +00:00
}
A bunch of netmap fixes: USERSPACE: 1. add support for devices with different number of rx and tx queues; 2. add better support for zero-copy operation, adding an extra field to the netmap ring to indicate how many buffers we have already processed but not yet released (with help from Eddie Kohler); 3. The two changes above unfortunately require an API change, so while at it add a version field and some spares to the ioctl() argument to help detect mismatches. 4. update the manual page for the two changes above; 5. update sample applications in tools/tools/netmap KERNEL: 1. simplify the internal structures moving the global wait queues to the 'struct netmap_adapter'; 2. simplify the functions that map kring<->nic ring indexes 3. normalize device-specific code, helps mainteinance; 4. start exploring the impact of micro-optimizations (prefetch etc.) in the ixgbe driver. Use 'legacy' descriptors on the tx ring and prefetch slots gives about 20% speedup at 900 MHz. Another 7-10% would come from removing the explict calls to bus_dmamap* in the core (they are effectively NOPs in this case, but it takes expensive load of the per-buffer dma maps to figure out that they are all NULL. Rx performance not investigated. I am postponing the MFC so i can import a few more improvements before merging.
2012-02-27 19:05:01 +00:00
static void
lem_netmap_attach(struct adapter *adapter)
{
struct netmap_adapter na;
bzero(&na, sizeof(na));
na.ifp = adapter->ifp;
update to the latest netmap snapshot. This includes the following: - use separate memory regions for VALE ports - locking fixes - some simplifications in the NIC-specific routines - performance improvements for the VALE switch - some new features in the pkt-gen test program - documentation updates There are small API changes that require programs to be recompiled (NETMAP_API has been bumped so you will detect old binaries at runtime). In particular: - struct netmap_slot now is 16 bytes to support an extra pointer, which may save one data copy when using VALE ports or VMs; - the struct netmap_if has two extra fields; MFC after: 3 days
2013-11-01 21:21:14 +00:00
na.na_flags = NAF_BDG_MAYSLEEP;
A bunch of netmap fixes: USERSPACE: 1. add support for devices with different number of rx and tx queues; 2. add better support for zero-copy operation, adding an extra field to the netmap ring to indicate how many buffers we have already processed but not yet released (with help from Eddie Kohler); 3. The two changes above unfortunately require an API change, so while at it add a version field and some spares to the ioctl() argument to help detect mismatches. 4. update the manual page for the two changes above; 5. update sample applications in tools/tools/netmap KERNEL: 1. simplify the internal structures moving the global wait queues to the 'struct netmap_adapter'; 2. simplify the functions that map kring<->nic ring indexes 3. normalize device-specific code, helps mainteinance; 4. start exploring the impact of micro-optimizations (prefetch etc.) in the ixgbe driver. Use 'legacy' descriptors on the tx ring and prefetch slots gives about 20% speedup at 900 MHz. Another 7-10% would come from removing the explict calls to bus_dmamap* in the core (they are effectively NOPs in this case, but it takes expensive load of the per-buffer dma maps to figure out that they are all NULL. Rx performance not investigated. I am postponing the MFC so i can import a few more improvements before merging.
2012-02-27 19:05:01 +00:00
na.num_tx_desc = adapter->num_tx_desc;
na.num_rx_desc = adapter->num_rx_desc;
na.nm_txsync = lem_netmap_txsync;
na.nm_rxsync = lem_netmap_rxsync;
na.nm_register = lem_netmap_reg;
split netmap code according to functions: - netmap.c base code - netmap_freebsd.c FreeBSD-specific code - netmap_generic.c emulate netmap over standard drivers - netmap_mbq.c simple mbuf tailq - netmap_mem2.c memory management - netmap_vale.c VALE switch simplify devce-specific code
2013-12-15 08:37:24 +00:00
na.num_tx_rings = na.num_rx_rings = 1;
netmap_attach(&na);
A bunch of netmap fixes: USERSPACE: 1. add support for devices with different number of rx and tx queues; 2. add better support for zero-copy operation, adding an extra field to the netmap ring to indicate how many buffers we have already processed but not yet released (with help from Eddie Kohler); 3. The two changes above unfortunately require an API change, so while at it add a version field and some spares to the ioctl() argument to help detect mismatches. 4. update the manual page for the two changes above; 5. update sample applications in tools/tools/netmap KERNEL: 1. simplify the internal structures moving the global wait queues to the 'struct netmap_adapter'; 2. simplify the functions that map kring<->nic ring indexes 3. normalize device-specific code, helps mainteinance; 4. start exploring the impact of micro-optimizations (prefetch etc.) in the ixgbe driver. Use 'legacy' descriptors on the tx ring and prefetch slots gives about 20% speedup at 900 MHz. Another 7-10% would come from removing the explict calls to bus_dmamap* in the core (they are effectively NOPs in this case, but it takes expensive load of the per-buffer dma maps to figure out that they are all NULL. Rx performance not investigated. I am postponing the MFC so i can import a few more improvements before merging.
2012-02-27 19:05:01 +00:00
}
- change the buffer size from a constant to a TUNABLE variable (hw.netmap.buf_size) so we can experiment with values different from 2048 which may give better cache performance. - rearrange the memory allocation code so it will be easier to replace it with a different implementation. The current code relies on a single large contiguous chunk of memory obtained through contigmalloc. The new implementation (not committed yet) uses multiple smaller chunks which are easier to fit in a fragmented address space.
2012-02-08 11:43:29 +00:00
/* end of file */
Reference in New Issue Copy Permalink