| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
| |
For clarity, as before, the last modified date stands for changes more
substantial than trivial editorial and formatting fixups. That, for
example, would be at least a change of the described behaviour, return
value, semantics or backward compatibility.
[skip ci]
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is a pain to detect, because the PF_PACKET socket code appears to
get separate "interface went down" and "interface went away"
notifications in my "unplug a USB Wi-Fi adapter" tests on my VMware
Fusion Ubuntu 18.04 virtual machine (5.3.0 kernel), and the first
notification delivers a wakeup and returns ENETDOWN while the second
notificaiton delivers *no* wakeup and sets the ifindex member of the
struct packet_sock for the socket, so there's nothing we can test after
the wakeup that's guaranteed to indicate that the interface has
disappeared.
So what we have to do is remember the ENETDOWN but not return it as an
error, and then arrange to periodically check whether the interface is
still there; if it isn't, we *then* return the "interface went away"
error, and, if we see traffic or see that the interface is up, we clear
the remembered ENETDOWN and stop doing the periodic checks.
This is tricky, because it needs to work not only for blocking pcap_t's,
where we're in a loop doing poll() calls, so we can keep checking within
the loop, but also for non-blocking pcap_t's on which the caller is
doing select()/poll()/epoll_wait().
In order to make *that* work, we need to tweak the semantics of
pcap_get_required_select_timeout() so that it's not guaranteed that it
will always return the same value, so that it should be called within
event loops rather than called once outside the event loop. Normally,
there is no timeout required for Linux PF_PACKET sockets, but when we're
doing the periodic tests, the timeout is required.
While we're doing that, we make the return value of
pcap_get_required_select_timeout() a const pointer - there was no good
reason for the caller to modify it (it doesn't belong to the caller).
If poll() returns POLLERR, use getsockopt(SO_ERROR) to get the socket
error, rather than a read().
Update the documentation to reflect this, and make various other
cleanups (including documenting the error return value for
pcap_get_selectable_fd() to -1 rather than PCAP_ERROR - it's not an
error code, it's just a specific error value). Also note that, for
kqueues on *BSD/macOS and for select/poll on Linux, the timeout needn't
be used as a timeout for the call - you can have a timer, so that when
that *particular* timer fires, you try calling pcap_dispatch() on the
pcap_t to which it corresponds.
Update selpolltest to add more capabilities needed when testing this on
Linux.
This should address GitHub issue #859 and pull request #858.
|
| |
|
|
|
|
|
|
| |
Use the BSD house style, in which, in
foobar() returns 17 on success and 137 on failure.
"foobar" is boldfaced but "()" isn't.
|
| | |
|
| |
|
|
|
|
|
| |
All manpage references such as pcap_create(3PCAP) will now be formatted
with the identifier (e.g. "pcap_create") in **bold** and the section
name (e.g. "(3PCAP)") in roman (default) face. This is how most manpages
seem to be formatted and makes things more consistent.
|
| |
|
|
| |
[skip ci]
|
| |
|
|
| |
[skip ci]
|
| |
|
|
|
|
|
|
| |
In the man pages that since the previous timestamp had meaningful (i.e.
not typos or whitespace fixups) changes set the timestamp to the date
of such last meaningful change.
[skip ci]
|
| |
|
|
|
|
| |
For the epoll mechanism, the call to wait for events is epoll_wait()
(and epoll_pwait()); for the kqueue mechanism, the call to wait for
events is kevent() (and additional calls in Darwin-based systems).
|
| |
|
|
|
|
|
|
|
|
|
| |
DAG adapters don't support blocking until packets arrive; a program
using select()/poll()/epoll()/kqueue-based event loop must periodically
poll DAG pcap_t's to see if they have packets available. Add
pcap_get_required_select_timeout(), which returns a pointer to a struct
timeval containing the appropriate timeout interval if polling is
required and NULL if polling isn't required or isn't supported.
Clean up the pcap_get_selectable_fd() documentation while we're at it.
|
| |
|
|
|
|
|
|
|
| |
That should make it clearer that the timeout does *not* guarantee that a
call that reads packets will return within N milliseconds even if no
packets arrive.
In the pcap_open_live() and pcap_set_timeout() calls, point to the
detailed description of the packet buffer timeout in pcap(3PCAP).
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Note, in more places, that if you call pcap_dispatch() on a pcap_t for
which there's a read timeout, it might return 0 if the read timeout
expires and there are no packets to be read - but that this behavior is
not guaranteed (so write your code to be able to handle it if it does
happen but not to depend on it happening).
Note also that a select()/poll()/etc. on the selectable descriptor for
the pcap_t might report the descriptor as readable if the read timeout
expires, even if there are no packets available to read - but that it
might not (so write your code to be able to handle it if it does happen
but not to depend on it happening).
Also, note that pcap_t's start out blocking, so they don't think that a
0 return from pcap_dispatch() means it's non-blocking and that they need
to call pcap_setnonblock() to put it in blocking mode.
|
| |
|
|
|
| |
This change reflects only meaningful (i.e. not purely editorial) changes
in the text.
|
| |
|
|
|
|
| |
This change removes CVS keywords that express that the file belongs to
libpcap repository. All such keywords represented the revision and
timestamp by the end of 2008 or even older.
|
| |
|
|
|
|
|
|
| |
Indicate what releases of FreeBSD, NetBSD, OpenBSD, and Mac OS X first
fixed select() to handle the timeout on BPF devices. (DragonFly BSD was
forked from FreeBSD after the fix, so it always had the fix.)
More precisely describe what the problem is.
|
| |
|
|
|
|
|
|
| |
pcap-linux.c memory-mapped mode.
Always doing a poll() if there are no packets available allows us to
catch disappearing interfaces in memory-mapped mode, so remove the
caveats about Linux.
|
| |
|
|
|
| |
Describe what happens with select() and poll() on Linux if the device on
which you're capturing goes away.
|
| |
|
|
|
|
| |
concepts to the pcap(3PCAP) man page, refer people to the pcap(3PCAP)
man page from the man pages for libpcap functions, and clean up some
errors.
|
| |
|
|
|
|
|
| |
to 1.0, might as well go with the place where Red Hat stuck the header
at one point and where the header "officially" resides.
(We should put a "backwards compatibility" note into pcap.3pcap.)
|
|
|
functions plus an overall man page for libpcap, and put them all into
section 3PCAP. That means you can actually do "man pcap_open_live" and
get something meaningful, rather than having to do "man pcap" and then
scroll through all the other stuff in the man page.
|
