summaryrefslogtreecommitdiff
path: root/src/libsystemd-bus
AgeCommit message (Collapse)Author
2013-06-05test-bus-kernel-benchmark: corrected output for memfd bisectHarald Hoyer
2013-06-05bus: benchmark - adjust printf and MAX_SIZEKay Sievers
2013-06-04bus: unmap memfd retrieved from the memfd cacheKay Sievers
2013-06-04bus-benchmark: add performance data output modeKay Sievers
- for now, comment out munmap() check to enable memfd passing - print tab-separated values and header - add memcpy() to fill the memfd, to produce real-world results $ ./test-bus-kernel-benchmark SIZE COPY MEMFD 4194304 370 370 2097152 810 810 1048576 2130 2130 524288 4090 4090 262144 7080 7080 131072 11380 11380 65536 17140 17140 98304 13930 13930 114688 12890 12890 122880 12350 12350 126976 12150 12150 129024 12170 12170 130048 12040 12040 130560 12080 12080 130816 12010 12010 130944 12020 12020 131008 12040 12040 131040 12050 12050 131056 12010 12010 131064 12060 12060 131068 12040 12040 131070 11310 11310 131069 11420 11420 Copying/memfd are equally fast at 131068 bytes $ ./test-bus-kernel-benchmark chart SIZE COPY MEMFD 1 35570 23690 2 36470 23680 4 36160 23520 8 36600 22220 16 33900 20830 32 33990 21360 64 33480 21280 128 34050 20910 256 32950 21750 512 34730 21900 1024 33810 22890 2048 36280 23110 4096 30790 21610 8192 29380 21100 16384 26880 19820 32768 22510 17980 65536 17270 15180 131072 11400 11420 262144 7140 8270 524288 4090 5050 1048576 2110 2780 2097152 800 1140 4194304 350 580
2013-06-04bus-benchmark: lower loop time to 100ms by default, to ensure "make check" ↵Lennart Poettering
finishes more quickly To get useful results you should however specify a much longer time on the command line.
2013-06-04test: determine number of transactions per second rather than time per ↵Lennart Poettering
transaction This way the measurements are not skewed by twoo short total measurement times, and results become stabler.
2013-05-31bus: update for kdbus changesKay Sievers
2013-05-30bus: update for kdbus changesKay Sievers
2013-05-29bus: rework benchmark test to actually yield useful resultsLennart Poettering
2013-05-29bus: when adding memfds to cache and we shorten them, make sure to unmap the ↵Lennart Poettering
remainder
2013-05-29bus: properly unmap mapped areaLennart Poettering
2013-05-24bus: update kdbus.h from upstreamKay Sievers
2013-05-22bus: add benchmark tool to determine the right threshold for copying vs. memfdLennart Poettering
2013-05-20bus: add test for bloom filter prefix matchLennart Poettering
2013-05-20bus: add a more comprehensive test for the bloom filter logicLennart Poettering
2013-05-20bus: calculate bloom filter for matchLennart Poettering
Yay! Filtering using kernel bloom filter matches works now! Yippieh!
2013-05-17bus: do not pass a pointer but a uint64_t address in RECVKay Sievers
2013-05-17bus: add APIs for negotiating what is attached to messagesLennart Poettering
2013-05-17bus: add minimal locking around the memfd cacheLennart Poettering
We want to allow clients to process an sd_bus_message on a different thread than it was received on. Since unreffing a bus message might readd some of its memfds to the memfd cache add some minimal locking around the cache.
2013-05-17bus: return ECHILD as soon as people try to reuse a bus connection across a ↵Lennart Poettering
fork()
2013-05-17bus: actually unmap kdbus pool after useLennart Poettering
2013-05-17bus: keep kernel bus fd around during entire life-time of busLennart Poettering
We need this since we might need to invoke the release ioctl for messages. Since we don't want to add any locking for that we simply keep a reference to the bus and then rely that the fd stays valid all the time.
2013-05-16bus: make bus ref counting atomicLennart Poettering
This is preparation to allow sd_bus_message obejcts to be processed in a different thread from their originating sd_bus object.
2013-05-16bus: synthesize timeout message errors instead of returning error codesLennart Poettering
2013-05-16bus: implement sd_bus_message_append_string_memfd()Lennart Poettering
2013-05-16bus: send memfds as payload only on directed messages and for large partsLennart Poettering
2013-05-16bus: implement receiving side of memfd hookupLennart Poettering
2013-05-16bus: put together messages with memfd payload correctlyLennart Poettering
2013-05-15bus: add support for adding memfds into message payloadLennart Poettering
2013-05-15bus: rework message struct to keep header with fields in same malloc() blockLennart Poettering
This allows us to guarantee that the first payload_vec we pass to the kernel for each message is guaranteed to include the full header and all its field.
2013-05-15bus: add macro for iterating through body parts of a messageLennart Poettering
2013-05-15bus: minor fixesLennart Poettering
2013-05-15bus: seal off memfds when sealing messagesLennart Poettering
2013-05-14bus: fix allocation of body parts from memfdLennart Poettering
2013-05-14bus: properly handle message bodies that are a chain of memory areas rather ↵Lennart Poettering
than a single one
2013-05-14bus: add and use UINT64_TO_PTR()Kay Sievers
2013-05-13bus: test-bus-memfd - use pread()/pwrite()Kay Sievers
2013-05-13bus: update test-bus-memfdKay Sievers
2013-05-12bus: remove KDBUS_MAKE_ACCESS_WORLD, remove (n_payload > 2) checkKay Sievers
KDBUS_MAKE_ACCESS_WORLD is no longer needed, the kernel creates the kdbus device node with the uid/gid of the caller now.
2013-05-11bus: update test-bus-memfdKay Sievers
2013-05-10bus: fix test-bus-memfdKay Sievers
2013-05-10bus: fix error check in memfd codeLennart Poettering
2013-05-10bus: extend memfd testLennart Poettering
2013-05-10bus: add new API for kdbus memfd functionalityLennart Poettering
2013-05-10kdbus: update kdbus.h from upstreamKay Sievers
2013-05-10bus: catch up with latest kdbusLennart Poettering
2013-05-10bus: add sd_bus_message_append_string_space() for zero-copy string appendingLennart Poettering
2013-05-09bus: add API for appending/reading fixed arraysLennart Poettering
2013-05-03Spelling fixesVille Skyttä
2013-05-02kdbus: update kdbus.h from upstreamKay Sievers