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 07:17:39 -05:00
|
|
|
/*
|
|
|
|
|
* Copyright (C) 2011 Universita` di Pisa. All rights reserved.
|
|
|
|
|
*
|
|
|
|
|
* 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.
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* $FreeBSD$
|
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 07:06:53 -05:00
|
|
|
* $Id: if_igb_netmap.h 9802 2011-12-02 18:42:37Z luigi $
|
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 07:17:39 -05:00
|
|
|
*
|
|
|
|
|
* netmap modifications for igb
|
|
|
|
|
* contribured by Ahmed Kooli
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
#include <net/netmap.h>
|
|
|
|
|
#include <sys/selinfo.h>
|
|
|
|
|
#include <vm/vm.h>
|
|
|
|
|
#include <vm/pmap.h> /* vtophys ? */
|
|
|
|
|
#include <dev/netmap/netmap_kern.h>
|
|
|
|
|
|
|
|
|
|
static int igb_netmap_reg(struct ifnet *, int onoff);
|
|
|
|
|
static int igb_netmap_txsync(void *, u_int, int);
|
|
|
|
|
static int igb_netmap_rxsync(void *, u_int, int);
|
|
|
|
|
static void igb_netmap_lock_wrapper(void *, int, u_int);
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
static void
|
|
|
|
|
igb_netmap_attach(struct adapter *adapter)
|
|
|
|
|
{
|
|
|
|
|
struct netmap_adapter na;
|
|
|
|
|
|
|
|
|
|
bzero(&na, sizeof(na));
|
|
|
|
|
|
|
|
|
|
na.ifp = adapter->ifp;
|
|
|
|
|
na.separate_locks = 1;
|
|
|
|
|
na.num_tx_desc = adapter->num_tx_desc;
|
|
|
|
|
na.num_rx_desc = adapter->num_rx_desc;
|
|
|
|
|
na.nm_txsync = igb_netmap_txsync;
|
|
|
|
|
na.nm_rxsync = igb_netmap_rxsync;
|
|
|
|
|
na.nm_lock = igb_netmap_lock_wrapper;
|
|
|
|
|
na.nm_register = igb_netmap_reg;
|
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 07:06:53 -05:00
|
|
|
na.buff_size = NETMAP_BUF_SIZE;
|
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 07:17:39 -05:00
|
|
|
netmap_attach(&na, adapter->num_queues);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* wrapper to export locks to the generic code
|
|
|
|
|
*/
|
|
|
|
|
static void
|
|
|
|
|
igb_netmap_lock_wrapper(void *_a, int what, u_int queueid)
|
|
|
|
|
{
|
|
|
|
|
struct adapter *adapter = _a;
|
|
|
|
|
|
|
|
|
|
ASSERT(queueid < adapter->num_queues);
|
|
|
|
|
switch (what) {
|
|
|
|
|
case NETMAP_CORE_LOCK:
|
|
|
|
|
IGB_CORE_LOCK(adapter);
|
|
|
|
|
break;
|
|
|
|
|
case NETMAP_CORE_UNLOCK:
|
|
|
|
|
IGB_CORE_UNLOCK(adapter);
|
|
|
|
|
break;
|
|
|
|
|
case NETMAP_TX_LOCK:
|
|
|
|
|
IGB_TX_LOCK(&adapter->tx_rings[queueid]);
|
|
|
|
|
break;
|
|
|
|
|
case NETMAP_TX_UNLOCK:
|
|
|
|
|
IGB_TX_UNLOCK(&adapter->tx_rings[queueid]);
|
|
|
|
|
break;
|
|
|
|
|
case NETMAP_RX_LOCK:
|
|
|
|
|
IGB_RX_LOCK(&adapter->rx_rings[queueid]);
|
|
|
|
|
break;
|
|
|
|
|
case NETMAP_RX_UNLOCK:
|
|
|
|
|
IGB_RX_UNLOCK(&adapter->rx_rings[queueid]);
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* support for netmap register/unregisted. We are already under core lock.
|
|
|
|
|
* only called on the first init or the last unregister.
|
|
|
|
|
*/
|
|
|
|
|
static int
|
|
|
|
|
igb_netmap_reg(struct ifnet *ifp, int onoff)
|
|
|
|
|
{
|
|
|
|
|
struct adapter *adapter = ifp->if_softc;
|
|
|
|
|
struct netmap_adapter *na = NA(ifp);
|
|
|
|
|
int error = 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 07:06:53 -05:00
|
|
|
if (na == NULL)
|
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 07:17:39 -05:00
|
|
|
return EINVAL;
|
|
|
|
|
|
|
|
|
|
igb_disable_intr(adapter);
|
|
|
|
|
|
|
|
|
|
/* Tell the stack that the interface is no longer active */
|
|
|
|
|
ifp->if_drv_flags &= ~(IFF_DRV_RUNNING | IFF_DRV_OACTIVE);
|
|
|
|
|
|
|
|
|
|
if (onoff) {
|
|
|
|
|
ifp->if_capenable |= IFCAP_NETMAP;
|
|
|
|
|
|
|
|
|
|
/* save if_transmit to restore it later */
|
|
|
|
|
na->if_transmit = ifp->if_transmit;
|
|
|
|
|
ifp->if_transmit = netmap_start;
|
|
|
|
|
|
|
|
|
|
igb_init_locked(adapter);
|
|
|
|
|
if ((ifp->if_drv_flags & (IFF_DRV_RUNNING | IFF_DRV_OACTIVE)) == 0) {
|
|
|
|
|
error = ENOMEM;
|
|
|
|
|
goto fail;
|
|
|
|
|
}
|
|
|
|
|
} else {
|
|
|
|
|
fail:
|
|
|
|
|
/* restore if_transmit */
|
|
|
|
|
ifp->if_transmit = na->if_transmit;
|
|
|
|
|
ifp->if_capenable &= ~IFCAP_NETMAP;
|
|
|
|
|
igb_init_locked(adapter); /* also enables intr */
|
|
|
|
|
}
|
|
|
|
|
return (error);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Reconcile kernel and user view of the transmit ring.
|
|
|
|
|
*/
|
|
|
|
|
static int
|
|
|
|
|
igb_netmap_txsync(void *a, u_int ring_nr, int do_lock)
|
|
|
|
|
{
|
|
|
|
|
struct adapter *adapter = a;
|
|
|
|
|
struct tx_ring *txr = &adapter->tx_rings[ring_nr];
|
|
|
|
|
struct netmap_adapter *na = NA(adapter->ifp);
|
|
|
|
|
struct netmap_kring *kring = &na->tx_rings[ring_nr];
|
|
|
|
|
struct netmap_ring *ring = kring->ring;
|
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 07:06:53 -05:00
|
|
|
int j, k, l, n = 0, lim = kring->nkr_num_slots - 1;
|
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 07:17:39 -05:00
|
|
|
|
|
|
|
|
/* generate an interrupt approximately every half ring */
|
|
|
|
|
int report_frequency = kring->nkr_num_slots >> 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 07:06:53 -05:00
|
|
|
k = ring->cur;
|
|
|
|
|
if (k > 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 07:17:39 -05:00
|
|
|
return netmap_ring_reinit(kring);
|
|
|
|
|
|
|
|
|
|
if (do_lock)
|
|
|
|
|
IGB_TX_LOCK(txr);
|
|
|
|
|
bus_dmamap_sync(txr->txdma.dma_tag, txr->txdma.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 07:06:53 -05:00
|
|
|
BUS_DMASYNC_POSTREAD);
|
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 07:17:39 -05:00
|
|
|
|
|
|
|
|
/* update avail to what the hardware knows */
|
|
|
|
|
ring->avail = kring->nr_hwavail;
|
|
|
|
|
|
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 07:06:53 -05:00
|
|
|
j = kring->nr_hwcur; /* netmap ring index */
|
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 07:17:39 -05:00
|
|
|
if (j != k) { /* we have new packets to send */
|
|
|
|
|
u32 olinfo_status = 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 07:06:53 -05:00
|
|
|
int n = 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 07:17:39 -05:00
|
|
|
|
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 07:06:53 -05:00
|
|
|
l = j - kring->nkr_hwofs; /* NIC ring index */
|
|
|
|
|
if (l < 0)
|
|
|
|
|
l += lim + 1;
|
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 07:17:39 -05:00
|
|
|
/* 82575 needs the queue index added */
|
|
|
|
|
if (adapter->hw.mac.type == e1000_82575)
|
|
|
|
|
olinfo_status |= txr->me << 4;
|
|
|
|
|
|
|
|
|
|
while (j != k) {
|
|
|
|
|
struct netmap_slot *slot = &ring->slot[j];
|
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 07:06:53 -05:00
|
|
|
struct igb_tx_buffer *txbuf = &txr->tx_buffers[l];
|
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 07:17:39 -05:00
|
|
|
union e1000_adv_tx_desc *curr =
|
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 07:06:53 -05:00
|
|
|
(union e1000_adv_tx_desc *)&txr->tx_base[l];
|
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 07:17:39 -05:00
|
|
|
void *addr = NMB(slot);
|
|
|
|
|
int flags = ((slot->flags & NS_REPORT) ||
|
|
|
|
|
j == 0 || j == report_frequency) ?
|
|
|
|
|
E1000_ADVTXD_DCMD_RS : 0;
|
|
|
|
|
int len = slot->len;
|
|
|
|
|
|
|
|
|
|
if (addr == netmap_buffer_base || len > NETMAP_BUF_SIZE) {
|
|
|
|
|
if (do_lock)
|
|
|
|
|
IGB_TX_UNLOCK(txr);
|
|
|
|
|
return netmap_ring_reinit(kring);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
slot->flags &= ~NS_REPORT;
|
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 07:06:53 -05:00
|
|
|
// XXX do we need to set the address ?
|
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 07:17:39 -05:00
|
|
|
curr->read.buffer_addr = htole64(vtophys(addr));
|
|
|
|
|
curr->read.olinfo_status =
|
|
|
|
|
htole32(olinfo_status |
|
|
|
|
|
(len<< E1000_ADVTXD_PAYLEN_SHIFT));
|
|
|
|
|
curr->read.cmd_type_len =
|
|
|
|
|
htole32(len | E1000_ADVTXD_DTYP_DATA |
|
|
|
|
|
E1000_ADVTXD_DCMD_IFCS |
|
|
|
|
|
E1000_ADVTXD_DCMD_DEXT |
|
|
|
|
|
E1000_ADVTXD_DCMD_EOP | flags);
|
|
|
|
|
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 07:06:53 -05:00
|
|
|
/* buffer has changed, reload map */
|
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 07:17:39 -05:00
|
|
|
netmap_reload_map(txr->txtag, txbuf->map,
|
|
|
|
|
addr, na->buff_size);
|
|
|
|
|
slot->flags &= ~NS_BUF_CHANGED;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
bus_dmamap_sync(txr->txtag, txbuf->map,
|
|
|
|
|
BUS_DMASYNC_PREWRITE);
|
|
|
|
|
j = (j == lim) ? 0 : j + 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 07:06:53 -05:00
|
|
|
l = (l == lim) ? 0 : l + 1;
|
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 07:17:39 -05:00
|
|
|
n++;
|
|
|
|
|
}
|
|
|
|
|
kring->nr_hwcur = k;
|
|
|
|
|
|
|
|
|
|
/* decrease avail by number of sent packets */
|
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 07:06:53 -05:00
|
|
|
kring->nr_hwavail -= n;
|
|
|
|
|
ring->avail = kring->nr_hwavail;
|
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 07:17:39 -05:00
|
|
|
|
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 07:06:53 -05:00
|
|
|
/* Set the watchdog XXX ? */
|
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 07:17:39 -05:00
|
|
|
txr->queue_status = IGB_QUEUE_WORKING;
|
|
|
|
|
txr->watchdog_time = ticks;
|
|
|
|
|
|
|
|
|
|
bus_dmamap_sync(txr->txdma.dma_tag, txr->txdma.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 07:06:53 -05:00
|
|
|
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 07:17:39 -05:00
|
|
|
|
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 07:06:53 -05:00
|
|
|
E1000_WRITE_REG(&adapter->hw, E1000_TDT(txr->me), l);
|
|
|
|
|
}
|
|
|
|
|
if (n == 0 || kring->nr_hwavail < 1) {
|
|
|
|
|
int delta;
|
|
|
|
|
|
|
|
|
|
/* record completed transmission using TDH */
|
|
|
|
|
l = E1000_READ_REG(&adapter->hw, E1000_TDH(ring_nr));
|
|
|
|
|
if (l >= kring->nkr_num_slots) /* XXX can it happen ? */
|
|
|
|
|
l -= kring->nkr_num_slots;
|
|
|
|
|
delta = l - txr->next_to_clean;
|
|
|
|
|
if (delta) {
|
|
|
|
|
/* new tx were completed */
|
|
|
|
|
if (delta < 0)
|
|
|
|
|
delta += kring->nkr_num_slots;
|
|
|
|
|
txr->next_to_clean = l;
|
|
|
|
|
kring->nr_hwavail += delta;
|
|
|
|
|
ring->avail = kring->nr_hwavail;
|
|
|
|
|
}
|
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 07:17:39 -05:00
|
|
|
}
|
|
|
|
|
if (do_lock)
|
|
|
|
|
IGB_TX_UNLOCK(txr);
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Reconcile kernel and user view of the receive ring.
|
|
|
|
|
*/
|
|
|
|
|
static int
|
|
|
|
|
igb_netmap_rxsync(void *a, u_int ring_nr, int do_lock)
|
|
|
|
|
{
|
|
|
|
|
struct adapter *adapter = a;
|
|
|
|
|
struct rx_ring *rxr = &adapter->rx_rings[ring_nr];
|
|
|
|
|
struct netmap_adapter *na = NA(adapter->ifp);
|
|
|
|
|
struct netmap_kring *kring = &na->rx_rings[ring_nr];
|
|
|
|
|
struct netmap_ring *ring = kring->ring;
|
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 07:06:53 -05:00
|
|
|
int j, k, l, n, lim = kring->nkr_num_slots - 1;
|
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 07:17:39 -05:00
|
|
|
|
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 07:06:53 -05:00
|
|
|
k = ring->cur;
|
|
|
|
|
if (k > 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 07:17:39 -05:00
|
|
|
return netmap_ring_reinit(kring);
|
|
|
|
|
|
|
|
|
|
if (do_lock)
|
|
|
|
|
IGB_RX_LOCK(rxr);
|
|
|
|
|
|
|
|
|
|
/* Sync the ring. */
|
|
|
|
|
bus_dmamap_sync(rxr->rxdma.dma_tag, rxr->rxdma.dma_map,
|
|
|
|
|
BUS_DMASYNC_POSTREAD | BUS_DMASYNC_POSTWRITE);
|
|
|
|
|
|
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 07:06:53 -05:00
|
|
|
l = rxr->next_to_check;
|
|
|
|
|
j = l + kring->nkr_hwofs;
|
|
|
|
|
if (j > lim)
|
|
|
|
|
j -= lim + 1;
|
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 07:17:39 -05:00
|
|
|
for (n = 0; ; n++) {
|
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 07:06:53 -05:00
|
|
|
union e1000_adv_rx_desc *curr = &rxr->rx_base[l];
|
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 07:17:39 -05:00
|
|
|
uint32_t staterr = le32toh(curr->wb.upper.status_error);
|
|
|
|
|
|
|
|
|
|
if ((staterr & E1000_RXD_STAT_DD) == 0)
|
|
|
|
|
break;
|
|
|
|
|
ring->slot[j].len = le16toh(curr->wb.upper.length);
|
|
|
|
|
|
|
|
|
|
bus_dmamap_sync(rxr->ptag,
|
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 07:06:53 -05:00
|
|
|
rxr->rx_buffers[l].pmap, BUS_DMASYNC_POSTREAD);
|
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 07:17:39 -05:00
|
|
|
j = (j == lim) ? 0 : j + 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 07:06:53 -05:00
|
|
|
l = (l == lim) ? 0 : l + 1;
|
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 07:17:39 -05:00
|
|
|
}
|
|
|
|
|
if (n) {
|
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 07:06:53 -05:00
|
|
|
rxr->next_to_check = l;
|
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 07:17:39 -05:00
|
|
|
kring->nr_hwavail += n;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* skip past packets that userspace has already processed,
|
|
|
|
|
* making them available for reception.
|
|
|
|
|
* advance nr_hwcur and issue a bus_dmamap_sync on the
|
|
|
|
|
* buffers so it is safe to write to them.
|
|
|
|
|
* Also increase nr_hwavail
|
|
|
|
|
*/
|
|
|
|
|
j = kring->nr_hwcur;
|
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 07:06:53 -05:00
|
|
|
l = kring->nr_hwcur - kring->nkr_hwofs;
|
|
|
|
|
if (l < 0)
|
|
|
|
|
l += lim + 1;
|
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 07:17:39 -05:00
|
|
|
if (j != k) { /* userspace has read some packets. */
|
|
|
|
|
n = 0;
|
|
|
|
|
while (j != k) {
|
|
|
|
|
struct netmap_slot *slot = ring->slot + j;
|
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 07:06:53 -05:00
|
|
|
union e1000_adv_rx_desc *curr = &rxr->rx_base[l];
|
|
|
|
|
struct igb_rx_buf *rxbuf = rxr->rx_buffers + l;
|
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 07:17:39 -05:00
|
|
|
void *addr = NMB(slot);
|
|
|
|
|
|
|
|
|
|
if (addr == netmap_buffer_base) { /* bad buf */
|
|
|
|
|
if (do_lock)
|
|
|
|
|
IGB_RX_UNLOCK(rxr);
|
|
|
|
|
return netmap_ring_reinit(kring);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
curr->wb.upper.status_error = 0;
|
|
|
|
|
curr->read.pkt_addr = htole64(vtophys(addr));
|
|
|
|
|
if (slot->flags & NS_BUF_CHANGED) {
|
|
|
|
|
netmap_reload_map(rxr->ptag, rxbuf->pmap,
|
|
|
|
|
addr, na->buff_size);
|
|
|
|
|
slot->flags &= ~NS_BUF_CHANGED;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
bus_dmamap_sync(rxr->ptag, rxbuf->pmap,
|
|
|
|
|
BUS_DMASYNC_PREREAD);
|
|
|
|
|
|
|
|
|
|
j = (j == lim) ? 0 : j + 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 07:06:53 -05:00
|
|
|
l = (l == lim) ? 0 : l + 1;
|
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 07:17:39 -05:00
|
|
|
n++;
|
|
|
|
|
}
|
|
|
|
|
kring->nr_hwavail -= n;
|
|
|
|
|
kring->nr_hwcur = ring->cur;
|
|
|
|
|
bus_dmamap_sync(rxr->rxdma.dma_tag, rxr->rxdma.dma_map,
|
|
|
|
|
BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE);
|
|
|
|
|
/* IMPORTANT: we must leave one free slot in the ring,
|
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 07:06:53 -05:00
|
|
|
* so move l 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 07:17:39 -05:00
|
|
|
*/
|
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 07:06:53 -05:00
|
|
|
l = (l == 0) ? lim : l - 1;
|
|
|
|
|
E1000_WRITE_REG(&adapter->hw, E1000_RDT(rxr->me), l);
|
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 07:17:39 -05:00
|
|
|
}
|
|
|
|
|
/* tell userspace that there are new packets */
|
|
|
|
|
ring->avail = kring->nr_hwavail ;
|
|
|
|
|
if (do_lock)
|
|
|
|
|
IGB_RX_UNLOCK(rxr);
|
|
|
|
|
return 0;
|
|
|
|
|
}
|