- 03 Feb, 2018 1 commit
-
-
Georgios Katsikas authored
Signed-off-by:
Georgios Katsikas <katsikas.gp@gmail.com>
-
- 06 Feb, 2017 1 commit
-
-
Tom Barbette authored
This reverts commit 593f7aab. Better solution was found through mainline
-
- 12 Jan, 2017 1 commit
-
-
Tom Barbette authored
-
- 04 Jan, 2017 1 commit
-
-
Tom Barbette authored
-
- 18 Jul, 2016 1 commit
-
-
Eddie Kohler authored
-
- 07 Dec, 2015 1 commit
-
-
Tom Barbette authored
It supports zero-copy through buffer swapping with extra buffer allocated on startup. This more or less double the throughput. The output device also supports a "Full Push" mode, allowing you to build a Click pipeline only with push paths, as the output rings of modern NICs are most of the time enough to absorb transient traffic, allowing to avoid using very slow Queue() elments. Netmap Buffers are now organized in pool much like click pool through a completed implementation of Luigi Rizzo's NetmapBufQ. A shared Ring allows multiple thread-local NetmapBufQ to exchange batches of buffers. The amount of extra buffers allocated on startup is set using NetmapInfo, which is now a real element. NetmapBufQ and other methods are moved to a library in lib/netmapdevice.cc For all those new features, the Netmap integration becomes more different than the other FromDevice/ToDevice methods and this commits moves netmap integration to independent FromNetmapDevice and ToNetmapDevice. This matches the DPDK Integrations. Also, ToNetmapDevice is agnostic, and thereforce could not be kept with ToDevice which is pull-only, or ToDpdkDevice which is push-only. DPDK fake buffer destructor is removed to use the same empty_destructor than Netmap, which won't be called as it is known to be empty.
-
- 22 Oct, 2015 1 commit
-
-
Tom Barbette authored
-
- 25 Aug, 2015 1 commit
-
-
Eddie Kohler authored
-
- 19 May, 2015 1 commit
-
-
Tom Barbette authored
Combining batching and I/O elements. Also took some changes from the original fatclick without wanting to : the support for DPDK pool (using the DPDK descriptor instead of the click Packet descriptor), the Pipeliner which is the QUeue keeping fullpush (see paper) and some minor modifications
-
- 05 May, 2015 2 commits
-
-
Tom Barbette authored
-
Tom Barbette authored
Merge branch numa into queuedevice - Support for numa to automatically assign threads to input device
-
- 15 Mar, 2015 1 commit
-
-
Tom Barbette authored
-
- 23 Dec, 2014 1 commit
-
-
Eddie Kohler authored
Reported by Lars Bro. Yes, we have uintptr_t.
-
- 02 Dec, 2014 1 commit
-
-
Eddie Kohler authored
-
- 28 Jul, 2014 1 commit
-
-
Eddie Kohler authored
Also whitespace
-
- 21 May, 2013 1 commit
-
-
Søren Løvborg authored
Fix configure script, which would incorrectly pass -lpcap to gcc BEFORE the source file, causing problems detecting libpcap features under some circumstances. Signed-off-by:
Søren Løvborg <soeren.j.loevborg@siemens.com>
-
- 02 Mar, 2013 1 commit
-
-
Eddie Kohler authored
With a hand-written "installtree" addition to click-buildtool. Which depends on md5sum existing, oh well. Signed-off-by:
Eddie Kohler <ekohler@gmail.com>
-
- 25 Jan, 2013 1 commit
-
-
Eddie Kohler authored
-
- 20 Jul, 2012 1 commit
-
-
Eddie Kohler authored
Clang takes -W -Wall, and needs -Werror to refuse rvalue references.
-
- 19 Jul, 2012 1 commit
-
-
Eddie Kohler authored
-
- 21 Feb, 2012 1 commit
-
-
Eddie Kohler authored
* Netmap supports zero copy packet transfer (on single threaded machines). * Allow --with-netmap=NETMAPDIRECTORY. * Improve some configury help messages. * Fix ToDevice bug.
-
- 20 Feb, 2012 1 commit
-
-
Eddie Kohler authored
With Luigi Rizzo. We look for the netmap header, and if provided, support "METHOD NETMAP". Netmap is an efficient way to move packets between the kernel and user level using shared memory packet rings. This checkin provides basic netmap support for FromDevice and ToDevice. Signed-off-by:
Eddie Kohler <ekohler@gmail.com>
-
- 16 Feb, 2012 1 commit
-
-
Eddie Kohler authored
-
- 21 Oct, 2011 1 commit
-
-
Eddie Kohler authored
We had turned off all warnings in Linux kernel compilation because otherwise one drowned in errors from the Linux header files, such as "automatic conversion from void * to char *" and such. But there is a way to get user warnings without drowning in kernel warnings: mark the kernel header files as "system" header files, with -isystem. So do that; and update configury to control warning options more carefully. Signed-off-by:
Eddie Kohler <ekohler@gmail.com>
-
- 24 Sep, 2011 2 commits
-
-
Eddie Kohler authored
Changes to CFLAGS in 02bc417e induced the warnings. Signed-off-by:
Eddie Kohler <ekohler@gmail.com>
-
Eddie Kohler authored
-
- 23 Sep, 2011 1 commit
-
-
Eddie Kohler authored
-
- 09 Jul, 2011 1 commit
-
-
Eddie Kohler authored
Using pcap_setdirection. Problem report from Antonie Henning and Robert Sombrutzki.
-
- 06 Jul, 2011 1 commit
-
-
Eddie Kohler authored
Luigi Rizzo reports this can be significantly faster on FreeBSD. Specify the send method via a METHOD keyword. Simultaneously, update FromDevice to support METHOD. Signed-off-by:
Eddie Kohler <ekohler@gmail.com>
-
- 02 Jul, 2011 1 commit
-
-
Eddie Kohler authored
-
- 27 May, 2011 1 commit
-
-
Eddie Kohler authored
-
- 18 May, 2011 1 commit
-
-
Eddie Kohler authored
Inspired by mailing list comments.
-
- 11 May, 2011 1 commit
-
-
Eddie Kohler authored
-
- 27 Apr, 2011 2 commits
-
-
Eddie Kohler authored
Like "g++ -std=gnu++0x" might. Signed-off-by:
Eddie Kohler <ekohler@gmail.com>
-
Eddie Kohler authored
Define it to "const" if not available. Signed-off-by:
Eddie Kohler <ekohler@gmail.com>
-
- 13 Apr, 2011 1 commit
-
-
Eddie Kohler authored
NumericTraits<T> (where T is integral) defines: - const bool is_numeric = true - const bool is_integer = true - const T const_min (minimum value of the type) - const T const_max (maximum value of the type) - const bool is_signed (true if type is signed) - typedef signed_type (signed version of the type) - typedef unsigned_type (unsigned version of the type) NumericTraits<T> (where T is not integral) defines: - const bool is_numeric = false - const bool is_integer = false Signed-off-by:
Eddie Kohler <ekohler@gmail.com>
-
- 02 Mar, 2011 1 commit
-
-
Eddie Kohler authored
-
- 01 Oct, 2010 1 commit
-
-
Eddie Kohler authored
Hopefully this helps Roberto Riggio.
-
- 30 Sep, 2010 1 commit
-
-
Eddie Kohler authored
-
- 21 Sep, 2010 1 commit
-
-
Eddie Kohler authored
is available. Inspired by POSIX 2008 deprecating gettimeofday().
-