2005-01-07 01:45:51 +00:00
|
|
|
/*-
|
2017-11-20 19:43:44 +00:00
|
|
|
* SPDX-License-Identifier: BSD-3-Clause
|
|
|
|
|
*
|
1994-05-24 10:09:53 +00:00
|
|
|
* Copyright (c) 1990, 1991, 1993
|
2011-12-31 07:21:28 +00:00
|
|
|
* The Regents of the University of California. All rights reserved.
|
1994-05-24 10:09:53 +00:00
|
|
|
*
|
|
|
|
|
* This code is derived from the Stanford/CMU enet packet filter,
|
|
|
|
|
* (net/enet.c) distributed as part of 4.3BSD, and code contributed
|
|
|
|
|
* to Berkeley by Steven McCanne and Van Jacobson both of Lawrence
|
|
|
|
|
* Berkeley Laboratory.
|
|
|
|
|
*
|
|
|
|
|
* 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.
|
2017-02-28 23:42:47 +00:00
|
|
|
* 3. Neither the name of the University nor the names of its contributors
|
1994-05-24 10:09:53 +00:00
|
|
|
* may be used to endorse or promote products derived from this software
|
|
|
|
|
* without specific prior written permission.
|
|
|
|
|
*
|
|
|
|
|
* THIS SOFTWARE IS PROVIDED BY THE REGENTS 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 REGENTS 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.
|
|
|
|
|
*
|
|
|
|
|
* @(#)bpf.h 8.1 (Berkeley) 6/10/93
|
1996-08-19 20:28:25 +00:00
|
|
|
* @(#)bpf.h 1.34 (LBL) 6/16/96
|
1994-05-24 10:09:53 +00:00
|
|
|
*
|
1999-08-28 01:08:13 +00:00
|
|
|
* $FreeBSD$
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
|
|
|
|
|
1994-08-21 05:11:48 +00:00
|
|
|
#ifndef _NET_BPF_H_
|
|
|
|
|
#define _NET_BPF_H_
|
|
|
|
|
|
Extract eventfilter declarations to sys/_eventfilter.h
This allows replacing "sys/eventfilter.h" includes with "sys/_eventfilter.h"
in other header files (e.g., sys/{bus,conf,cpu}.h) and reduces header
pollution substantially.
EVENTHANDLER_DECLARE and EVENTHANDLER_LIST_DECLAREs were moved out of .c
files into appropriate headers (e.g., sys/proc.h, powernv/opal.h).
As a side effect of reduced header pollution, many .c files and headers no
longer contain needed definitions. The remainder of the patch addresses
adding appropriate includes to fix those files.
LOCK_DEBUG and LOCK_FILE_LINE_ARG are moved to sys/_lock.h, as required by
sys/mutex.h since r326106 (but silently protected by header pollution prior
to this change).
No functional change (intended). Of course, any out of tree modules that
relied on header pollution for sys/eventhandler.h, sys/lock.h, or
sys/mutex.h inclusion need to be fixed. __FreeBSD_version has been bumped.
2019-05-20 00:38:23 +00:00
|
|
|
#include <sys/_eventhandler.h>
|
2019-05-13 13:45:28 +00:00
|
|
|
#include <sys/ck.h>
|
|
|
|
|
#include <net/dlt.h>
|
|
|
|
|
|
1996-08-19 20:28:25 +00:00
|
|
|
/* BSD style release date */
|
|
|
|
|
#define BPF_RELEASE 199606
|
|
|
|
|
|
|
|
|
|
typedef int32_t bpf_int32;
|
|
|
|
|
typedef u_int32_t bpf_u_int32;
|
Implement flexible BPF timestamping framework.
- Allow setting format, resolution and accuracy of BPF time stamps per
listener. Previously, we were only able to use microtime(9). Now we can
set various resolutions and accuracies with ioctl(2) BIOCSTSTAMP command.
Similarly, we can get the current resolution and accuracy with BIOCGTSTAMP
command. Document all supported options in bpf(4) and their uses.
- Introduce new time stamp 'struct bpf_ts' and header 'struct bpf_xhdr'.
The new time stamp has both 64-bit second and fractional parts. bpf_xhdr
has this time stamp instead of 'struct timeval' for bh_tstamp. The new
structures let us use bh_tstamp of same size on both 32-bit and 64-bit
platforms without adding additional shims for 32-bit binaries. On 64-bit
platforms, size of BPF header does not change compared to bpf_hdr as its
members are already all 64-bit long. On 32-bit platforms, the size may
increase by 8 bytes. For backward compatibility, struct bpf_hdr with
struct timeval is still the default header unless new time stamp format is
explicitly requested. However, the behaviour may change in the future and
all relevant code is wrapped around "#ifdef BURN_BRIDGES" for now.
- Add experimental support for tagging mbufs with time stamps from a lower
layer, e.g., device driver. Currently, mbuf_tags(9) is used to tag mbufs.
The time stamps must be uptime in 'struct bintime' format as binuptime(9)
and getbinuptime(9) do.
Reviewed by: net@
2010-06-15 19:28:44 +00:00
|
|
|
typedef int64_t bpf_int64;
|
|
|
|
|
typedef u_int64_t bpf_u_int64;
|
2023-01-12 20:17:33 +00:00
|
|
|
struct ifnet;
|
1996-08-19 20:28:25 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
2022-06-20 16:02:59 +00:00
|
|
|
* Alignment macros. BPF_WORDALIGN rounds up to the next multiple of
|
|
|
|
|
* BPF_ALIGNMENT.
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
1998-10-04 21:53:59 +00:00
|
|
|
#define BPF_ALIGNMENT sizeof(long)
|
1994-05-24 10:09:53 +00:00
|
|
|
#define BPF_WORDALIGN(x) (((x)+(BPF_ALIGNMENT-1))&~(BPF_ALIGNMENT-1))
|
|
|
|
|
|
|
|
|
|
#define BPF_MAXINSNS 512
|
2000-01-15 19:46:12 +00:00
|
|
|
#define BPF_MAXBUFSIZE 0x80000
|
1994-05-24 10:09:53 +00:00
|
|
|
#define BPF_MINBUFSIZE 32
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Structure for BIOCSETF.
|
|
|
|
|
*/
|
|
|
|
|
struct bpf_program {
|
|
|
|
|
u_int bf_len;
|
|
|
|
|
struct bpf_insn *bf_insns;
|
|
|
|
|
};
|
1995-05-30 08:16:23 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
|
* Struct returned by BIOCGSTATS.
|
|
|
|
|
*/
|
|
|
|
|
struct bpf_stat {
|
|
|
|
|
u_int bs_recv; /* number of packets received */
|
|
|
|
|
u_int bs_drop; /* number of packets dropped */
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
/*
|
1995-05-30 08:16:23 +00:00
|
|
|
* Struct return by BIOCVERSION. This represents the version number of
|
1994-05-24 10:09:53 +00:00
|
|
|
* the filter language described by the instruction encodings below.
|
|
|
|
|
* bpf understands a program iff kernel_major == filter_major &&
|
|
|
|
|
* kernel_minor >= filter_minor, that is, if the value returned by the
|
|
|
|
|
* running kernel has the same major number and a minor number equal
|
|
|
|
|
* equal to or less than the filter being downloaded. Otherwise, the
|
|
|
|
|
* results are undefined, meaning an error may be returned or packets
|
|
|
|
|
* may be accepted haphazardly.
|
|
|
|
|
* It has nothing to do with the source code version.
|
|
|
|
|
*/
|
|
|
|
|
struct bpf_version {
|
|
|
|
|
u_short bv_major;
|
|
|
|
|
u_short bv_minor;
|
|
|
|
|
};
|
1996-08-19 20:28:25 +00:00
|
|
|
/* Current version number of filter architecture. */
|
1994-05-24 10:09:53 +00:00
|
|
|
#define BPF_MAJOR_VERSION 1
|
|
|
|
|
#define BPF_MINOR_VERSION 1
|
|
|
|
|
|
Introduce support for zero-copy BPF buffering, which reduces the
overhead of packet capture by allowing a user process to directly "loan"
buffer memory to the kernel rather than using read(2) to explicitly copy
data from kernel address space.
The user process will issue new BPF ioctls to set the shared memory
buffer mode and provide pointers to buffers and their size. The kernel
then wires and maps the pages into kernel address space using sf_buf(9),
which on supporting architectures will use the direct map region. The
current "buffered" access mode remains the default, and support for
zero-copy buffers must, for the time being, be explicitly enabled using
a sysctl for the kernel to accept requests to use it.
The kernel and user process synchronize use of the buffers with atomic
operations, avoiding the need for system calls under load; the user
process may use select()/poll()/kqueue() to manage blocking while
waiting for network data if the user process is able to consume data
faster than the kernel generates it. Patchs to libpcap are available
to allow libpcap applications to transparently take advantage of this
support. Detailed information on the new API may be found in bpf(4),
including specific atomic operations and memory barriers required to
synchronize buffer use safely.
These changes modify the base BPF implementation to (roughly) abstrac
the current buffer model, allowing the new shared memory model to be
added, and add new monitoring statistics for netstat to print. The
implementation, with the exception of some monitoring hanges that break
the netstat monitoring ABI for BPF, will be MFC'd.
Zerocopy bpf buffers are still considered experimental are disabled
by default. To experiment with this new facility, adjust the
net.bpf.zerocopy_enable sysctl variable to 1.
Changes to libpcap will be made available as a patch for the time being,
and further refinements to the implementation are expected.
Sponsored by: Seccuris Inc.
In collaboration with: rwatson
Tested by: pwood, gallatin
MFC after: 4 months [1]
[1] Certain portions will probably not be MFCed, specifically things
that can break the monitoring ABI.
2008-03-24 13:49:17 +00:00
|
|
|
/*
|
|
|
|
|
* Historically, BPF has supported a single buffering model, first using mbuf
|
|
|
|
|
* clusters in kernel, and later using malloc(9) buffers in kernel. We now
|
|
|
|
|
* support multiple buffering modes, which may be queried and set using
|
|
|
|
|
* BIOCGETBUFMODE and BIOCSETBUFMODE. So as to avoid handling the complexity
|
|
|
|
|
* of changing modes while sniffing packets, the mode becomes fixed once an
|
|
|
|
|
* interface has been attached to the BPF descriptor.
|
|
|
|
|
*/
|
|
|
|
|
#define BPF_BUFMODE_BUFFER 1 /* Kernel buffers with read(). */
|
|
|
|
|
#define BPF_BUFMODE_ZBUF 2 /* Zero-copy buffers. */
|
|
|
|
|
|
|
|
|
|
/*-
|
|
|
|
|
* Struct used by BIOCSETZBUF, BIOCROTZBUF: describes up to two zero-copy
|
|
|
|
|
* buffer as used by BPF.
|
|
|
|
|
*/
|
|
|
|
|
struct bpf_zbuf {
|
|
|
|
|
void *bz_bufa; /* Location of 'a' zero-copy buffer. */
|
|
|
|
|
void *bz_bufb; /* Location of 'b' zero-copy buffer. */
|
|
|
|
|
size_t bz_buflen; /* Size of zero-copy buffers. */
|
|
|
|
|
};
|
|
|
|
|
|
Implement flexible BPF timestamping framework.
- Allow setting format, resolution and accuracy of BPF time stamps per
listener. Previously, we were only able to use microtime(9). Now we can
set various resolutions and accuracies with ioctl(2) BIOCSTSTAMP command.
Similarly, we can get the current resolution and accuracy with BIOCGTSTAMP
command. Document all supported options in bpf(4) and their uses.
- Introduce new time stamp 'struct bpf_ts' and header 'struct bpf_xhdr'.
The new time stamp has both 64-bit second and fractional parts. bpf_xhdr
has this time stamp instead of 'struct timeval' for bh_tstamp. The new
structures let us use bh_tstamp of same size on both 32-bit and 64-bit
platforms without adding additional shims for 32-bit binaries. On 64-bit
platforms, size of BPF header does not change compared to bpf_hdr as its
members are already all 64-bit long. On 32-bit platforms, the size may
increase by 8 bytes. For backward compatibility, struct bpf_hdr with
struct timeval is still the default header unless new time stamp format is
explicitly requested. However, the behaviour may change in the future and
all relevant code is wrapped around "#ifdef BURN_BRIDGES" for now.
- Add experimental support for tagging mbufs with time stamps from a lower
layer, e.g., device driver. Currently, mbuf_tags(9) is used to tag mbufs.
The time stamps must be uptime in 'struct bintime' format as binuptime(9)
and getbinuptime(9) do.
Reviewed by: net@
2010-06-15 19:28:44 +00:00
|
|
|
#define BIOCGBLEN _IOR('B', 102, u_int)
|
|
|
|
|
#define BIOCSBLEN _IOWR('B', 102, u_int)
|
|
|
|
|
#define BIOCSETF _IOW('B', 103, struct bpf_program)
|
|
|
|
|
#define BIOCFLUSH _IO('B', 104)
|
|
|
|
|
#define BIOCPROMISC _IO('B', 105)
|
|
|
|
|
#define BIOCGDLT _IOR('B', 106, u_int)
|
|
|
|
|
#define BIOCGETIF _IOR('B', 107, struct ifreq)
|
|
|
|
|
#define BIOCSETIF _IOW('B', 108, struct ifreq)
|
|
|
|
|
#define BIOCSRTIMEOUT _IOW('B', 109, struct timeval)
|
|
|
|
|
#define BIOCGRTIMEOUT _IOR('B', 110, struct timeval)
|
|
|
|
|
#define BIOCGSTATS _IOR('B', 111, struct bpf_stat)
|
|
|
|
|
#define BIOCIMMEDIATE _IOW('B', 112, u_int)
|
|
|
|
|
#define BIOCVERSION _IOR('B', 113, struct bpf_version)
|
|
|
|
|
#define BIOCGRSIG _IOR('B', 114, u_int)
|
|
|
|
|
#define BIOCSRSIG _IOW('B', 115, u_int)
|
|
|
|
|
#define BIOCGHDRCMPLT _IOR('B', 116, u_int)
|
|
|
|
|
#define BIOCSHDRCMPLT _IOW('B', 117, u_int)
|
|
|
|
|
#define BIOCGDIRECTION _IOR('B', 118, u_int)
|
|
|
|
|
#define BIOCSDIRECTION _IOW('B', 119, u_int)
|
|
|
|
|
#define BIOCSDLT _IOW('B', 120, u_int)
|
|
|
|
|
#define BIOCGDLTLIST _IOWR('B', 121, struct bpf_dltlist)
|
2005-08-22 19:35:48 +00:00
|
|
|
#define BIOCLOCK _IO('B', 122)
|
Implement flexible BPF timestamping framework.
- Allow setting format, resolution and accuracy of BPF time stamps per
listener. Previously, we were only able to use microtime(9). Now we can
set various resolutions and accuracies with ioctl(2) BIOCSTSTAMP command.
Similarly, we can get the current resolution and accuracy with BIOCGTSTAMP
command. Document all supported options in bpf(4) and their uses.
- Introduce new time stamp 'struct bpf_ts' and header 'struct bpf_xhdr'.
The new time stamp has both 64-bit second and fractional parts. bpf_xhdr
has this time stamp instead of 'struct timeval' for bh_tstamp. The new
structures let us use bh_tstamp of same size on both 32-bit and 64-bit
platforms without adding additional shims for 32-bit binaries. On 64-bit
platforms, size of BPF header does not change compared to bpf_hdr as its
members are already all 64-bit long. On 32-bit platforms, the size may
increase by 8 bytes. For backward compatibility, struct bpf_hdr with
struct timeval is still the default header unless new time stamp format is
explicitly requested. However, the behaviour may change in the future and
all relevant code is wrapped around "#ifdef BURN_BRIDGES" for now.
- Add experimental support for tagging mbufs with time stamps from a lower
layer, e.g., device driver. Currently, mbuf_tags(9) is used to tag mbufs.
The time stamps must be uptime in 'struct bintime' format as binuptime(9)
and getbinuptime(9) do.
Reviewed by: net@
2010-06-15 19:28:44 +00:00
|
|
|
#define BIOCSETWF _IOW('B', 123, struct bpf_program)
|
|
|
|
|
#define BIOCFEEDBACK _IOW('B', 124, u_int)
|
|
|
|
|
#define BIOCGETBUFMODE _IOR('B', 125, u_int)
|
|
|
|
|
#define BIOCSETBUFMODE _IOW('B', 126, u_int)
|
|
|
|
|
#define BIOCGETZMAX _IOR('B', 127, size_t)
|
|
|
|
|
#define BIOCROTZBUF _IOR('B', 128, struct bpf_zbuf)
|
|
|
|
|
#define BIOCSETZBUF _IOW('B', 129, struct bpf_zbuf)
|
|
|
|
|
#define BIOCSETFNR _IOW('B', 130, struct bpf_program)
|
|
|
|
|
#define BIOCGTSTAMP _IOR('B', 131, u_int)
|
|
|
|
|
#define BIOCSTSTAMP _IOW('B', 132, u_int)
|
2021-07-21 15:41:22 +00:00
|
|
|
#define BIOCSETVLANPCP _IOW('B', 133, u_int)
|
2007-02-26 22:24:14 +00:00
|
|
|
|
|
|
|
|
/* Obsolete */
|
|
|
|
|
#define BIOCGSEESENT BIOCGDIRECTION
|
|
|
|
|
#define BIOCSSEESENT BIOCSDIRECTION
|
|
|
|
|
|
|
|
|
|
/* Packet directions */
|
|
|
|
|
enum bpf_direction {
|
|
|
|
|
BPF_D_IN, /* See incoming packets */
|
|
|
|
|
BPF_D_INOUT, /* See incoming and outgoing packets */
|
|
|
|
|
BPF_D_OUT /* See outgoing packets */
|
|
|
|
|
};
|
1994-05-24 10:09:53 +00:00
|
|
|
|
Implement flexible BPF timestamping framework.
- Allow setting format, resolution and accuracy of BPF time stamps per
listener. Previously, we were only able to use microtime(9). Now we can
set various resolutions and accuracies with ioctl(2) BIOCSTSTAMP command.
Similarly, we can get the current resolution and accuracy with BIOCGTSTAMP
command. Document all supported options in bpf(4) and their uses.
- Introduce new time stamp 'struct bpf_ts' and header 'struct bpf_xhdr'.
The new time stamp has both 64-bit second and fractional parts. bpf_xhdr
has this time stamp instead of 'struct timeval' for bh_tstamp. The new
structures let us use bh_tstamp of same size on both 32-bit and 64-bit
platforms without adding additional shims for 32-bit binaries. On 64-bit
platforms, size of BPF header does not change compared to bpf_hdr as its
members are already all 64-bit long. On 32-bit platforms, the size may
increase by 8 bytes. For backward compatibility, struct bpf_hdr with
struct timeval is still the default header unless new time stamp format is
explicitly requested. However, the behaviour may change in the future and
all relevant code is wrapped around "#ifdef BURN_BRIDGES" for now.
- Add experimental support for tagging mbufs with time stamps from a lower
layer, e.g., device driver. Currently, mbuf_tags(9) is used to tag mbufs.
The time stamps must be uptime in 'struct bintime' format as binuptime(9)
and getbinuptime(9) do.
Reviewed by: net@
2010-06-15 19:28:44 +00:00
|
|
|
/* Time stamping functions */
|
|
|
|
|
#define BPF_T_MICROTIME 0x0000
|
|
|
|
|
#define BPF_T_NANOTIME 0x0001
|
|
|
|
|
#define BPF_T_BINTIME 0x0002
|
|
|
|
|
#define BPF_T_NONE 0x0003
|
|
|
|
|
#define BPF_T_FORMAT_MASK 0x0003
|
|
|
|
|
#define BPF_T_NORMAL 0x0000
|
2011-12-31 07:21:28 +00:00
|
|
|
#define BPF_T_FAST 0x0100
|
|
|
|
|
#define BPF_T_MONOTONIC 0x0200
|
|
|
|
|
#define BPF_T_MONOTONIC_FAST (BPF_T_FAST | BPF_T_MONOTONIC)
|
|
|
|
|
#define BPF_T_FLAG_MASK 0x0300
|
Implement flexible BPF timestamping framework.
- Allow setting format, resolution and accuracy of BPF time stamps per
listener. Previously, we were only able to use microtime(9). Now we can
set various resolutions and accuracies with ioctl(2) BIOCSTSTAMP command.
Similarly, we can get the current resolution and accuracy with BIOCGTSTAMP
command. Document all supported options in bpf(4) and their uses.
- Introduce new time stamp 'struct bpf_ts' and header 'struct bpf_xhdr'.
The new time stamp has both 64-bit second and fractional parts. bpf_xhdr
has this time stamp instead of 'struct timeval' for bh_tstamp. The new
structures let us use bh_tstamp of same size on both 32-bit and 64-bit
platforms without adding additional shims for 32-bit binaries. On 64-bit
platforms, size of BPF header does not change compared to bpf_hdr as its
members are already all 64-bit long. On 32-bit platforms, the size may
increase by 8 bytes. For backward compatibility, struct bpf_hdr with
struct timeval is still the default header unless new time stamp format is
explicitly requested. However, the behaviour may change in the future and
all relevant code is wrapped around "#ifdef BURN_BRIDGES" for now.
- Add experimental support for tagging mbufs with time stamps from a lower
layer, e.g., device driver. Currently, mbuf_tags(9) is used to tag mbufs.
The time stamps must be uptime in 'struct bintime' format as binuptime(9)
and getbinuptime(9) do.
Reviewed by: net@
2010-06-15 19:28:44 +00:00
|
|
|
#define BPF_T_FORMAT(t) ((t) & BPF_T_FORMAT_MASK)
|
|
|
|
|
#define BPF_T_FLAG(t) ((t) & BPF_T_FLAG_MASK)
|
|
|
|
|
#define BPF_T_VALID(t) \
|
|
|
|
|
((t) == BPF_T_NONE || (BPF_T_FORMAT(t) != BPF_T_NONE && \
|
|
|
|
|
((t) & ~(BPF_T_FORMAT_MASK | BPF_T_FLAG_MASK)) == 0))
|
|
|
|
|
|
2011-12-31 07:21:28 +00:00
|
|
|
#define BPF_T_MICROTIME_FAST (BPF_T_MICROTIME | BPF_T_FAST)
|
|
|
|
|
#define BPF_T_NANOTIME_FAST (BPF_T_NANOTIME | BPF_T_FAST)
|
|
|
|
|
#define BPF_T_BINTIME_FAST (BPF_T_BINTIME | BPF_T_FAST)
|
Implement flexible BPF timestamping framework.
- Allow setting format, resolution and accuracy of BPF time stamps per
listener. Previously, we were only able to use microtime(9). Now we can
set various resolutions and accuracies with ioctl(2) BIOCSTSTAMP command.
Similarly, we can get the current resolution and accuracy with BIOCGTSTAMP
command. Document all supported options in bpf(4) and their uses.
- Introduce new time stamp 'struct bpf_ts' and header 'struct bpf_xhdr'.
The new time stamp has both 64-bit second and fractional parts. bpf_xhdr
has this time stamp instead of 'struct timeval' for bh_tstamp. The new
structures let us use bh_tstamp of same size on both 32-bit and 64-bit
platforms without adding additional shims for 32-bit binaries. On 64-bit
platforms, size of BPF header does not change compared to bpf_hdr as its
members are already all 64-bit long. On 32-bit platforms, the size may
increase by 8 bytes. For backward compatibility, struct bpf_hdr with
struct timeval is still the default header unless new time stamp format is
explicitly requested. However, the behaviour may change in the future and
all relevant code is wrapped around "#ifdef BURN_BRIDGES" for now.
- Add experimental support for tagging mbufs with time stamps from a lower
layer, e.g., device driver. Currently, mbuf_tags(9) is used to tag mbufs.
The time stamps must be uptime in 'struct bintime' format as binuptime(9)
and getbinuptime(9) do.
Reviewed by: net@
2010-06-15 19:28:44 +00:00
|
|
|
#define BPF_T_MICROTIME_MONOTONIC (BPF_T_MICROTIME | BPF_T_MONOTONIC)
|
|
|
|
|
#define BPF_T_NANOTIME_MONOTONIC (BPF_T_NANOTIME | BPF_T_MONOTONIC)
|
|
|
|
|
#define BPF_T_BINTIME_MONOTONIC (BPF_T_BINTIME | BPF_T_MONOTONIC)
|
2011-12-31 07:21:28 +00:00
|
|
|
#define BPF_T_MICROTIME_MONOTONIC_FAST (BPF_T_MICROTIME | BPF_T_MONOTONIC_FAST)
|
|
|
|
|
#define BPF_T_NANOTIME_MONOTONIC_FAST (BPF_T_NANOTIME | BPF_T_MONOTONIC_FAST)
|
|
|
|
|
#define BPF_T_BINTIME_MONOTONIC_FAST (BPF_T_BINTIME | BPF_T_MONOTONIC_FAST)
|
Implement flexible BPF timestamping framework.
- Allow setting format, resolution and accuracy of BPF time stamps per
listener. Previously, we were only able to use microtime(9). Now we can
set various resolutions and accuracies with ioctl(2) BIOCSTSTAMP command.
Similarly, we can get the current resolution and accuracy with BIOCGTSTAMP
command. Document all supported options in bpf(4) and their uses.
- Introduce new time stamp 'struct bpf_ts' and header 'struct bpf_xhdr'.
The new time stamp has both 64-bit second and fractional parts. bpf_xhdr
has this time stamp instead of 'struct timeval' for bh_tstamp. The new
structures let us use bh_tstamp of same size on both 32-bit and 64-bit
platforms without adding additional shims for 32-bit binaries. On 64-bit
platforms, size of BPF header does not change compared to bpf_hdr as its
members are already all 64-bit long. On 32-bit platforms, the size may
increase by 8 bytes. For backward compatibility, struct bpf_hdr with
struct timeval is still the default header unless new time stamp format is
explicitly requested. However, the behaviour may change in the future and
all relevant code is wrapped around "#ifdef BURN_BRIDGES" for now.
- Add experimental support for tagging mbufs with time stamps from a lower
layer, e.g., device driver. Currently, mbuf_tags(9) is used to tag mbufs.
The time stamps must be uptime in 'struct bintime' format as binuptime(9)
and getbinuptime(9) do.
Reviewed by: net@
2010-06-15 19:28:44 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
|
* Structure prepended to each packet.
|
|
|
|
|
*/
|
Implement flexible BPF timestamping framework.
- Allow setting format, resolution and accuracy of BPF time stamps per
listener. Previously, we were only able to use microtime(9). Now we can
set various resolutions and accuracies with ioctl(2) BIOCSTSTAMP command.
Similarly, we can get the current resolution and accuracy with BIOCGTSTAMP
command. Document all supported options in bpf(4) and their uses.
- Introduce new time stamp 'struct bpf_ts' and header 'struct bpf_xhdr'.
The new time stamp has both 64-bit second and fractional parts. bpf_xhdr
has this time stamp instead of 'struct timeval' for bh_tstamp. The new
structures let us use bh_tstamp of same size on both 32-bit and 64-bit
platforms without adding additional shims for 32-bit binaries. On 64-bit
platforms, size of BPF header does not change compared to bpf_hdr as its
members are already all 64-bit long. On 32-bit platforms, the size may
increase by 8 bytes. For backward compatibility, struct bpf_hdr with
struct timeval is still the default header unless new time stamp format is
explicitly requested. However, the behaviour may change in the future and
all relevant code is wrapped around "#ifdef BURN_BRIDGES" for now.
- Add experimental support for tagging mbufs with time stamps from a lower
layer, e.g., device driver. Currently, mbuf_tags(9) is used to tag mbufs.
The time stamps must be uptime in 'struct bintime' format as binuptime(9)
and getbinuptime(9) do.
Reviewed by: net@
2010-06-15 19:28:44 +00:00
|
|
|
struct bpf_ts {
|
|
|
|
|
bpf_int64 bt_sec; /* seconds */
|
|
|
|
|
bpf_u_int64 bt_frac; /* fraction */
|
|
|
|
|
};
|
|
|
|
|
struct bpf_xhdr {
|
|
|
|
|
struct bpf_ts bh_tstamp; /* time stamp */
|
|
|
|
|
bpf_u_int32 bh_caplen; /* length of captured portion */
|
|
|
|
|
bpf_u_int32 bh_datalen; /* original length of packet */
|
|
|
|
|
u_short bh_hdrlen; /* length of bpf header (this struct
|
|
|
|
|
plus alignment padding) */
|
|
|
|
|
};
|
|
|
|
|
/* Obsolete */
|
1994-05-24 10:09:53 +00:00
|
|
|
struct bpf_hdr {
|
|
|
|
|
struct timeval bh_tstamp; /* time stamp */
|
1996-08-19 20:28:25 +00:00
|
|
|
bpf_u_int32 bh_caplen; /* length of captured portion */
|
|
|
|
|
bpf_u_int32 bh_datalen; /* original length of packet */
|
1994-05-24 10:09:53 +00:00
|
|
|
u_short bh_hdrlen; /* length of bpf header (this struct
|
|
|
|
|
plus alignment padding) */
|
|
|
|
|
};
|
1999-12-29 04:46:21 +00:00
|
|
|
#ifdef _KERNEL
|
Implement flexible BPF timestamping framework.
- Allow setting format, resolution and accuracy of BPF time stamps per
listener. Previously, we were only able to use microtime(9). Now we can
set various resolutions and accuracies with ioctl(2) BIOCSTSTAMP command.
Similarly, we can get the current resolution and accuracy with BIOCGTSTAMP
command. Document all supported options in bpf(4) and their uses.
- Introduce new time stamp 'struct bpf_ts' and header 'struct bpf_xhdr'.
The new time stamp has both 64-bit second and fractional parts. bpf_xhdr
has this time stamp instead of 'struct timeval' for bh_tstamp. The new
structures let us use bh_tstamp of same size on both 32-bit and 64-bit
platforms without adding additional shims for 32-bit binaries. On 64-bit
platforms, size of BPF header does not change compared to bpf_hdr as its
members are already all 64-bit long. On 32-bit platforms, the size may
increase by 8 bytes. For backward compatibility, struct bpf_hdr with
struct timeval is still the default header unless new time stamp format is
explicitly requested. However, the behaviour may change in the future and
all relevant code is wrapped around "#ifdef BURN_BRIDGES" for now.
- Add experimental support for tagging mbufs with time stamps from a lower
layer, e.g., device driver. Currently, mbuf_tags(9) is used to tag mbufs.
The time stamps must be uptime in 'struct bintime' format as binuptime(9)
and getbinuptime(9) do.
Reviewed by: net@
2010-06-15 19:28:44 +00:00
|
|
|
#define MTAG_BPF 0x627066
|
|
|
|
|
#define MTAG_BPF_TIMESTAMP 0
|
1994-05-24 10:09:53 +00:00
|
|
|
#endif
|
|
|
|
|
|
Introduce support for zero-copy BPF buffering, which reduces the
overhead of packet capture by allowing a user process to directly "loan"
buffer memory to the kernel rather than using read(2) to explicitly copy
data from kernel address space.
The user process will issue new BPF ioctls to set the shared memory
buffer mode and provide pointers to buffers and their size. The kernel
then wires and maps the pages into kernel address space using sf_buf(9),
which on supporting architectures will use the direct map region. The
current "buffered" access mode remains the default, and support for
zero-copy buffers must, for the time being, be explicitly enabled using
a sysctl for the kernel to accept requests to use it.
The kernel and user process synchronize use of the buffers with atomic
operations, avoiding the need for system calls under load; the user
process may use select()/poll()/kqueue() to manage blocking while
waiting for network data if the user process is able to consume data
faster than the kernel generates it. Patchs to libpcap are available
to allow libpcap applications to transparently take advantage of this
support. Detailed information on the new API may be found in bpf(4),
including specific atomic operations and memory barriers required to
synchronize buffer use safely.
These changes modify the base BPF implementation to (roughly) abstrac
the current buffer model, allowing the new shared memory model to be
added, and add new monitoring statistics for netstat to print. The
implementation, with the exception of some monitoring hanges that break
the netstat monitoring ABI for BPF, will be MFC'd.
Zerocopy bpf buffers are still considered experimental are disabled
by default. To experiment with this new facility, adjust the
net.bpf.zerocopy_enable sysctl variable to 1.
Changes to libpcap will be made available as a patch for the time being,
and further refinements to the implementation are expected.
Sponsored by: Seccuris Inc.
In collaboration with: rwatson
Tested by: pwood, gallatin
MFC after: 4 months [1]
[1] Certain portions will probably not be MFCed, specifically things
that can break the monitoring ABI.
2008-03-24 13:49:17 +00:00
|
|
|
/*
|
|
|
|
|
* When using zero-copy BPF buffers, a shared memory header is present
|
|
|
|
|
* allowing the kernel BPF implementation and user process to synchronize
|
|
|
|
|
* without using system calls. This structure defines that header. When
|
|
|
|
|
* accessing these fields, appropriate atomic operation and memory barriers
|
|
|
|
|
* are required in order not to see stale or out-of-order data; see bpf(4)
|
|
|
|
|
* for reference code to access these fields from userspace.
|
|
|
|
|
*
|
|
|
|
|
* The layout of this structure is critical, and must not be changed; if must
|
|
|
|
|
* fit in a single page on all architectures.
|
|
|
|
|
*/
|
|
|
|
|
struct bpf_zbuf_header {
|
|
|
|
|
volatile u_int bzh_kernel_gen; /* Kernel generation number. */
|
|
|
|
|
volatile u_int bzh_kernel_len; /* Length of data in the buffer. */
|
|
|
|
|
volatile u_int bzh_user_gen; /* User generation number. */
|
|
|
|
|
u_int _bzh_pad[5];
|
|
|
|
|
};
|
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
1996-01-30 23:02:38 +00:00
|
|
|
* The instruction encodings.
|
2015-01-06 18:58:31 +00:00
|
|
|
*
|
|
|
|
|
* Please inform tcpdump-workers@lists.tcpdump.org if you use any
|
|
|
|
|
* of the reserved values, so that we can note that they're used
|
|
|
|
|
* (and perhaps implement it in the reference BPF implementation
|
|
|
|
|
* and encourage its implementation elsewhere).
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
2015-01-06 18:58:31 +00:00
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* The upper 8 bits of the opcode aren't used. BSD/OS used 0x8000.
|
2009-03-21 20:43:56 +00:00
|
|
|
*/
|
2015-01-06 18:58:31 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/* instruction classes */
|
|
|
|
|
#define BPF_CLASS(code) ((code) & 0x07)
|
|
|
|
|
#define BPF_LD 0x00
|
|
|
|
|
#define BPF_LDX 0x01
|
|
|
|
|
#define BPF_ST 0x02
|
|
|
|
|
#define BPF_STX 0x03
|
|
|
|
|
#define BPF_ALU 0x04
|
|
|
|
|
#define BPF_JMP 0x05
|
|
|
|
|
#define BPF_RET 0x06
|
|
|
|
|
#define BPF_MISC 0x07
|
|
|
|
|
|
|
|
|
|
/* ld/ldx fields */
|
|
|
|
|
#define BPF_SIZE(code) ((code) & 0x18)
|
|
|
|
|
#define BPF_W 0x00
|
|
|
|
|
#define BPF_H 0x08
|
|
|
|
|
#define BPF_B 0x10
|
2015-01-06 18:58:31 +00:00
|
|
|
/* 0x18 reserved; used by BSD/OS */
|
1994-05-24 10:09:53 +00:00
|
|
|
#define BPF_MODE(code) ((code) & 0xe0)
|
|
|
|
|
#define BPF_IMM 0x00
|
|
|
|
|
#define BPF_ABS 0x20
|
|
|
|
|
#define BPF_IND 0x40
|
|
|
|
|
#define BPF_MEM 0x60
|
|
|
|
|
#define BPF_LEN 0x80
|
|
|
|
|
#define BPF_MSH 0xa0
|
2015-01-06 18:58:31 +00:00
|
|
|
/* 0xc0 reserved; used by BSD/OS */
|
|
|
|
|
/* 0xe0 reserved; used by BSD/OS */
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
|
/* alu/jmp fields */
|
|
|
|
|
#define BPF_OP(code) ((code) & 0xf0)
|
|
|
|
|
#define BPF_ADD 0x00
|
|
|
|
|
#define BPF_SUB 0x10
|
|
|
|
|
#define BPF_MUL 0x20
|
|
|
|
|
#define BPF_DIV 0x30
|
|
|
|
|
#define BPF_OR 0x40
|
|
|
|
|
#define BPF_AND 0x50
|
|
|
|
|
#define BPF_LSH 0x60
|
|
|
|
|
#define BPF_RSH 0x70
|
|
|
|
|
#define BPF_NEG 0x80
|
2015-01-06 18:58:31 +00:00
|
|
|
#define BPF_MOD 0x90
|
|
|
|
|
#define BPF_XOR 0xa0
|
|
|
|
|
/* 0xb0 reserved */
|
|
|
|
|
/* 0xc0 reserved */
|
|
|
|
|
/* 0xd0 reserved */
|
|
|
|
|
/* 0xe0 reserved */
|
|
|
|
|
/* 0xf0 reserved */
|
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
#define BPF_JA 0x00
|
|
|
|
|
#define BPF_JEQ 0x10
|
|
|
|
|
#define BPF_JGT 0x20
|
|
|
|
|
#define BPF_JGE 0x30
|
|
|
|
|
#define BPF_JSET 0x40
|
2015-01-06 18:58:31 +00:00
|
|
|
/* 0x50 reserved; used on BSD/OS */
|
|
|
|
|
/* 0x60 reserved */
|
|
|
|
|
/* 0x70 reserved */
|
|
|
|
|
/* 0x80 reserved */
|
|
|
|
|
/* 0x90 reserved */
|
|
|
|
|
/* 0xa0 reserved */
|
|
|
|
|
/* 0xb0 reserved */
|
|
|
|
|
/* 0xc0 reserved */
|
|
|
|
|
/* 0xd0 reserved */
|
|
|
|
|
/* 0xe0 reserved */
|
|
|
|
|
/* 0xf0 reserved */
|
1994-05-24 10:09:53 +00:00
|
|
|
#define BPF_SRC(code) ((code) & 0x08)
|
|
|
|
|
#define BPF_K 0x00
|
|
|
|
|
#define BPF_X 0x08
|
|
|
|
|
|
|
|
|
|
/* ret - BPF_K and BPF_X also apply */
|
|
|
|
|
#define BPF_RVAL(code) ((code) & 0x18)
|
|
|
|
|
#define BPF_A 0x10
|
2015-01-06 18:58:31 +00:00
|
|
|
/* 0x18 reserved */
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
|
/* misc */
|
|
|
|
|
#define BPF_MISCOP(code) ((code) & 0xf8)
|
|
|
|
|
#define BPF_TAX 0x00
|
2015-01-06 18:58:31 +00:00
|
|
|
/* 0x08 reserved */
|
|
|
|
|
/* 0x10 reserved */
|
|
|
|
|
/* 0x18 reserved */
|
|
|
|
|
/* #define BPF_COP 0x20 NetBSD "coprocessor" extensions */
|
|
|
|
|
/* 0x28 reserved */
|
|
|
|
|
/* 0x30 reserved */
|
|
|
|
|
/* 0x38 reserved */
|
|
|
|
|
/* #define BPF_COPX 0x40 NetBSD "coprocessor" extensions */
|
|
|
|
|
/* also used on BSD/OS */
|
|
|
|
|
/* 0x48 reserved */
|
|
|
|
|
/* 0x50 reserved */
|
|
|
|
|
/* 0x58 reserved */
|
|
|
|
|
/* 0x60 reserved */
|
|
|
|
|
/* 0x68 reserved */
|
|
|
|
|
/* 0x70 reserved */
|
|
|
|
|
/* 0x78 reserved */
|
1994-05-24 10:09:53 +00:00
|
|
|
#define BPF_TXA 0x80
|
2015-01-06 18:58:31 +00:00
|
|
|
/* 0x88 reserved */
|
|
|
|
|
/* 0x90 reserved */
|
|
|
|
|
/* 0x98 reserved */
|
|
|
|
|
/* 0xa0 reserved */
|
|
|
|
|
/* 0xa8 reserved */
|
|
|
|
|
/* 0xb0 reserved */
|
|
|
|
|
/* 0xb8 reserved */
|
|
|
|
|
/* 0xc0 reserved; used on BSD/OS */
|
|
|
|
|
/* 0xc8 reserved */
|
|
|
|
|
/* 0xd0 reserved */
|
|
|
|
|
/* 0xd8 reserved */
|
|
|
|
|
/* 0xe0 reserved */
|
|
|
|
|
/* 0xe8 reserved */
|
|
|
|
|
/* 0xf0 reserved */
|
|
|
|
|
/* 0xf8 reserved */
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* The instruction data structure.
|
|
|
|
|
*/
|
|
|
|
|
struct bpf_insn {
|
1996-08-19 20:28:25 +00:00
|
|
|
u_short code;
|
|
|
|
|
u_char jt;
|
|
|
|
|
u_char jf;
|
|
|
|
|
bpf_u_int32 k;
|
1994-05-24 10:09:53 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Macros for insn array initializers.
|
|
|
|
|
*/
|
|
|
|
|
#define BPF_STMT(code, k) { (u_short)(code), 0, 0, k }
|
|
|
|
|
#define BPF_JUMP(code, k, jt, jf) { (u_short)(code), jt, jf, k }
|
|
|
|
|
|
2003-01-20 19:08:46 +00:00
|
|
|
/*
|
|
|
|
|
* Structure to retrieve available DLTs for the interface.
|
|
|
|
|
*/
|
|
|
|
|
struct bpf_dltlist {
|
|
|
|
|
u_int bfl_len; /* number of bfd_list array */
|
|
|
|
|
u_int *bfl_list; /* array of DLTs */
|
|
|
|
|
};
|
|
|
|
|
|
1999-12-29 04:46:21 +00:00
|
|
|
#ifdef _KERNEL
|
Introduce support for zero-copy BPF buffering, which reduces the
overhead of packet capture by allowing a user process to directly "loan"
buffer memory to the kernel rather than using read(2) to explicitly copy
data from kernel address space.
The user process will issue new BPF ioctls to set the shared memory
buffer mode and provide pointers to buffers and their size. The kernel
then wires and maps the pages into kernel address space using sf_buf(9),
which on supporting architectures will use the direct map region. The
current "buffered" access mode remains the default, and support for
zero-copy buffers must, for the time being, be explicitly enabled using
a sysctl for the kernel to accept requests to use it.
The kernel and user process synchronize use of the buffers with atomic
operations, avoiding the need for system calls under load; the user
process may use select()/poll()/kqueue() to manage blocking while
waiting for network data if the user process is able to consume data
faster than the kernel generates it. Patchs to libpcap are available
to allow libpcap applications to transparently take advantage of this
support. Detailed information on the new API may be found in bpf(4),
including specific atomic operations and memory barriers required to
synchronize buffer use safely.
These changes modify the base BPF implementation to (roughly) abstrac
the current buffer model, allowing the new shared memory model to be
added, and add new monitoring statistics for netstat to print. The
implementation, with the exception of some monitoring hanges that break
the netstat monitoring ABI for BPF, will be MFC'd.
Zerocopy bpf buffers are still considered experimental are disabled
by default. To experiment with this new facility, adjust the
net.bpf.zerocopy_enable sysctl variable to 1.
Changes to libpcap will be made available as a patch for the time being,
and further refinements to the implementation are expected.
Sponsored by: Seccuris Inc.
In collaboration with: rwatson
Tested by: pwood, gallatin
MFC after: 4 months [1]
[1] Certain portions will probably not be MFCed, specifically things
that can break the monitoring ABI.
2008-03-24 13:49:17 +00:00
|
|
|
#ifdef MALLOC_DECLARE
|
|
|
|
|
MALLOC_DECLARE(M_BPF);
|
|
|
|
|
#endif
|
|
|
|
|
#ifdef SYSCTL_DECL
|
|
|
|
|
SYSCTL_DECL(_net_bpf);
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Rotate the packet buffers in descriptor d. Move the store buffer into the
|
2015-07-31 21:43:27 +00:00
|
|
|
* hold slot, and the free buffer into the store slot. Zero the length of the
|
|
|
|
|
* new store buffer. Descriptor lock should be held. One must be careful to
|
|
|
|
|
* not rotate the buffers twice, i.e. if fbuf != NULL.
|
Introduce support for zero-copy BPF buffering, which reduces the
overhead of packet capture by allowing a user process to directly "loan"
buffer memory to the kernel rather than using read(2) to explicitly copy
data from kernel address space.
The user process will issue new BPF ioctls to set the shared memory
buffer mode and provide pointers to buffers and their size. The kernel
then wires and maps the pages into kernel address space using sf_buf(9),
which on supporting architectures will use the direct map region. The
current "buffered" access mode remains the default, and support for
zero-copy buffers must, for the time being, be explicitly enabled using
a sysctl for the kernel to accept requests to use it.
The kernel and user process synchronize use of the buffers with atomic
operations, avoiding the need for system calls under load; the user
process may use select()/poll()/kqueue() to manage blocking while
waiting for network data if the user process is able to consume data
faster than the kernel generates it. Patchs to libpcap are available
to allow libpcap applications to transparently take advantage of this
support. Detailed information on the new API may be found in bpf(4),
including specific atomic operations and memory barriers required to
synchronize buffer use safely.
These changes modify the base BPF implementation to (roughly) abstrac
the current buffer model, allowing the new shared memory model to be
added, and add new monitoring statistics for netstat to print. The
implementation, with the exception of some monitoring hanges that break
the netstat monitoring ABI for BPF, will be MFC'd.
Zerocopy bpf buffers are still considered experimental are disabled
by default. To experiment with this new facility, adjust the
net.bpf.zerocopy_enable sysctl variable to 1.
Changes to libpcap will be made available as a patch for the time being,
and further refinements to the implementation are expected.
Sponsored by: Seccuris Inc.
In collaboration with: rwatson
Tested by: pwood, gallatin
MFC after: 4 months [1]
[1] Certain portions will probably not be MFCed, specifically things
that can break the monitoring ABI.
2008-03-24 13:49:17 +00:00
|
|
|
*/
|
|
|
|
|
#define ROTATE_BUFFERS(d) do { \
|
|
|
|
|
(d)->bd_hbuf = (d)->bd_sbuf; \
|
|
|
|
|
(d)->bd_hlen = (d)->bd_slen; \
|
|
|
|
|
(d)->bd_sbuf = (d)->bd_fbuf; \
|
|
|
|
|
(d)->bd_slen = 0; \
|
|
|
|
|
(d)->bd_fbuf = NULL; \
|
|
|
|
|
bpf_bufheld(d); \
|
|
|
|
|
} while (0)
|
|
|
|
|
|
2006-06-02 19:59:33 +00:00
|
|
|
/*
|
|
|
|
|
* Descriptor associated with each attached hardware interface.
|
2015-04-20 22:08:11 +00:00
|
|
|
* Part of this structure is exposed to external callers to speed up
|
|
|
|
|
* bpf_peers_present() calls.
|
2006-06-02 19:59:33 +00:00
|
|
|
*/
|
2015-04-20 22:08:11 +00:00
|
|
|
struct bpf_if;
|
2019-05-13 13:45:28 +00:00
|
|
|
CK_LIST_HEAD(bpfd_list, bpf_d);
|
2015-04-20 22:08:11 +00:00
|
|
|
|
|
|
|
|
struct bpf_if_ext {
|
2019-05-13 13:45:28 +00:00
|
|
|
CK_LIST_ENTRY(bpf_if) bif_next; /* list of all interfaces */
|
|
|
|
|
struct bpfd_list bif_dlist; /* descriptor list */
|
2006-06-02 19:59:33 +00:00
|
|
|
};
|
|
|
|
|
|
Introduce support for zero-copy BPF buffering, which reduces the
overhead of packet capture by allowing a user process to directly "loan"
buffer memory to the kernel rather than using read(2) to explicitly copy
data from kernel address space.
The user process will issue new BPF ioctls to set the shared memory
buffer mode and provide pointers to buffers and their size. The kernel
then wires and maps the pages into kernel address space using sf_buf(9),
which on supporting architectures will use the direct map region. The
current "buffered" access mode remains the default, and support for
zero-copy buffers must, for the time being, be explicitly enabled using
a sysctl for the kernel to accept requests to use it.
The kernel and user process synchronize use of the buffers with atomic
operations, avoiding the need for system calls under load; the user
process may use select()/poll()/kqueue() to manage blocking while
waiting for network data if the user process is able to consume data
faster than the kernel generates it. Patchs to libpcap are available
to allow libpcap applications to transparently take advantage of this
support. Detailed information on the new API may be found in bpf(4),
including specific atomic operations and memory barriers required to
synchronize buffer use safely.
These changes modify the base BPF implementation to (roughly) abstrac
the current buffer model, allowing the new shared memory model to be
added, and add new monitoring statistics for netstat to print. The
implementation, with the exception of some monitoring hanges that break
the netstat monitoring ABI for BPF, will be MFC'd.
Zerocopy bpf buffers are still considered experimental are disabled
by default. To experiment with this new facility, adjust the
net.bpf.zerocopy_enable sysctl variable to 1.
Changes to libpcap will be made available as a patch for the time being,
and further refinements to the implementation are expected.
Sponsored by: Seccuris Inc.
In collaboration with: rwatson
Tested by: pwood, gallatin
MFC after: 4 months [1]
[1] Certain portions will probably not be MFCed, specifically things
that can break the monitoring ABI.
2008-03-24 13:49:17 +00:00
|
|
|
void bpf_bufheld(struct bpf_d *d);
|
2002-03-19 21:54:18 +00:00
|
|
|
int bpf_validate(const struct bpf_insn *, int);
|
2002-11-14 23:24:13 +00:00
|
|
|
void bpf_tap(struct bpf_if *, u_char *, u_int);
|
2023-01-12 20:17:33 +00:00
|
|
|
void bpf_tap_if(struct ifnet *, u_char *, u_int);
|
2002-11-14 23:24:13 +00:00
|
|
|
void bpf_mtap(struct bpf_if *, struct mbuf *);
|
2023-01-12 20:17:33 +00:00
|
|
|
void bpf_mtap_if(struct ifnet *, struct mbuf *);
|
2003-12-28 03:56:00 +00:00
|
|
|
void bpf_mtap2(struct bpf_if *, void *, u_int, struct mbuf *);
|
2023-01-12 20:17:33 +00:00
|
|
|
void bpf_mtap2_if(struct ifnet *, void *, u_int, struct mbuf *);
|
2002-03-19 21:54:18 +00:00
|
|
|
void bpfattach(struct ifnet *, u_int, u_int);
|
2002-11-14 23:24:13 +00:00
|
|
|
void bpfattach2(struct ifnet *, u_int, u_int, struct bpf_if **);
|
2002-03-19 21:54:18 +00:00
|
|
|
void bpfdetach(struct ifnet *);
|
2016-04-11 10:00:38 +00:00
|
|
|
#ifdef VIMAGE
|
|
|
|
|
int bpf_get_bp_params(struct bpf_if *, u_int *, u_int *);
|
|
|
|
|
#endif
|
The advent of if_detach, allowing interface removal at runtime, makes it
possible for a panic to occur if BPF is in use on the interface at the
time of the call to if_detach. This happens because BPF maintains pointers
to the struct ifnet describing the interface, which is freed by if_detach.
To correct this problem, a new call, bpfdetach, is introduced. bpfdetach
locates BPF descriptor references to the interface, and NULLs them. Other
BPF code is modified so that discovery of a NULL interface results in
ENXIO (already implemented for some calls). Processes blocked on a BPF
call will also be woken up so that they can receive ENXIO.
Interface drivers that invoke bpfattach and if_detach must be modified to
also call bpfattach(ifp) before calling if_detach(ifp). This is relevant
for buses that support hot removal, such as pccard and usb. Patches to
all effected devices will not be committed, only to if_wi.c, due to
testing limitations. To reproduce the crash, load up tcpdump on you
favorite pccard ethernet card, and then eject the card. As some pccard
drivers do not invoke if_detach(ifp), this bug will not manifest itself
for those drivers.
Reviewed by: wes
2000-03-19 05:42:34 +00:00
|
|
|
|
2002-03-19 21:54:18 +00:00
|
|
|
void bpfilterattach(int);
|
2008-08-26 16:12:49 +00:00
|
|
|
u_int bpf_filter(const struct bpf_insn *, u_char *, u_int, u_int);
|
2002-11-14 23:24:13 +00:00
|
|
|
|
2006-06-02 19:59:33 +00:00
|
|
|
static __inline int
|
|
|
|
|
bpf_peers_present(struct bpf_if *bpf)
|
|
|
|
|
{
|
2015-04-20 22:08:11 +00:00
|
|
|
struct bpf_if_ext *ext;
|
2006-06-02 19:59:33 +00:00
|
|
|
|
2015-04-20 22:08:11 +00:00
|
|
|
ext = (struct bpf_if_ext *)bpf;
|
2019-05-13 13:45:28 +00:00
|
|
|
if (!CK_LIST_EMPTY(&ext->bif_dlist))
|
2006-06-03 18:48:14 +00:00
|
|
|
return (1);
|
2006-06-03 06:37:00 +00:00
|
|
|
return (0);
|
2006-06-02 19:59:33 +00:00
|
|
|
}
|
|
|
|
|
|
2023-01-12 20:17:33 +00:00
|
|
|
#define BPF_TAP(_ifp,_pkt,_pktlen) \
|
|
|
|
|
bpf_tap_if((_ifp), (_pkt), (_pktlen))
|
|
|
|
|
#define BPF_MTAP(_ifp,_m) \
|
|
|
|
|
bpf_mtap_if((_ifp), (_m))
|
|
|
|
|
#define BPF_MTAP2(_ifp,_data,_dlen,_m) \
|
|
|
|
|
bpf_mtap2_if((_ifp), (_data), (_dlen), (_m))
|
1994-05-24 10:09:53 +00:00
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Number of scratch memory words (for BPF_LD|BPF_MEM and BPF_ST).
|
|
|
|
|
*/
|
|
|
|
|
#define BPF_MEMWORDS 16
|
|
|
|
|
|
2013-10-28 07:45:03 +00:00
|
|
|
/* BPF attach/detach events */
|
|
|
|
|
typedef void (*bpf_track_fn)(void *, struct ifnet *, int /* dlt */,
|
|
|
|
|
int /* 1 =>'s attach */);
|
|
|
|
|
EVENTHANDLER_DECLARE(bpf_track, bpf_track_fn);
|
|
|
|
|
|
2002-11-14 23:24:13 +00:00
|
|
|
#endif /* _NET_BPF_H_ */
|
