| Message ID | 20231222122146.65519-1-lulie@linux.alibaba.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | bpf: introduce BPF_MAP_TYPE_RELAY | expand |
On Fri, Dec 22, 2023 at 08:21:43PM +0800, Philo Lu wrote: > The patch set introduce a new type of map, BPF_MAP_TYPE_RELAY, based on > relay interface [0]. It provides a way for persistent and overwritable data > transfer. > > As stated in [0], relay is a efficient method for log and data transfer. > And the interface is simple enough so that we can implement and use this > type of map with current map interfaces. Besides we need a new helper > bpf_relay_output to output data to user, similar with bpf_ringbuf_output. > > We need this map because currently neither ringbuf nor perfbuf satisfies > the requirements of relatively long-term consistent tracing, where the bpf > program keeps writing into the buffer without any bundled reader, and the > buffer supports overwriting. For users, they just run the bpf program to > collect data, and are able to read as need. The detailed discussion can be > found at [1]. > > The buffer is exposed to users as per-cpu files in debugfs, supporting read > and mmap, and it is up to users how to formulate and read it, either > through a program with mmap or just `cat`. Specifically, the files are > created as "/sys/kerenl/debug/<dirname>/<mapname>#cpu", where the <dirname> > is defined with map_update_elem (Note that we do not need to implement > actual elem operators for relay_map). > > If this map is acceptable, other parts including docs, libbpf support, > selftests, and benchmarks (if need) will be added in the following version. looks useful, selftests might be already helpful to see the usage jirka > > [0] > https://github.com/torvalds/linux/blob/master/Documentation/filesystems/relay.rst > [1] > https://lore.kernel.org/bpf/20231219122850.433be151@gandalf.local.home/T/ > > Philo Lu (3): > bpf: implement relay map basis > bpf: implement map_update_elem to init relay file > bpf: introduce bpf_relay_output helper > > include/linux/bpf.h | 1 + > include/linux/bpf_types.h | 3 + > include/uapi/linux/bpf.h | 17 +++ > kernel/bpf/Makefile | 3 + > kernel/bpf/helpers.c | 4 + > kernel/bpf/relaymap.c | 213 ++++++++++++++++++++++++++++++++++++++ > kernel/bpf/syscall.c | 1 + > kernel/bpf/verifier.c | 8 ++ > kernel/trace/bpf_trace.c | 4 + > 9 files changed, 254 insertions(+) > create mode 100644 kernel/bpf/relaymap.c > > -- > 2.32.0.3.g01195cf9f >
On 2023/12/22 22:45, Jiri Olsa wrote: > On Fri, Dec 22, 2023 at 08:21:43PM +0800, Philo Lu wrote: >> The patch set introduce a new type of map, BPF_MAP_TYPE_RELAY, based on >> relay interface [0]. It provides a way for persistent and overwritable data >> transfer. >> >> As stated in [0], relay is a efficient method for log and data transfer. >> And the interface is simple enough so that we can implement and use this >> type of map with current map interfaces. Besides we need a new helper >> bpf_relay_output to output data to user, similar with bpf_ringbuf_output. >> >> We need this map because currently neither ringbuf nor perfbuf satisfies >> the requirements of relatively long-term consistent tracing, where the bpf >> program keeps writing into the buffer without any bundled reader, and the >> buffer supports overwriting. For users, they just run the bpf program to >> collect data, and are able to read as need. The detailed discussion can be >> found at [1]. >> >> The buffer is exposed to users as per-cpu files in debugfs, supporting read >> and mmap, and it is up to users how to formulate and read it, either >> through a program with mmap or just `cat`. Specifically, the files are >> created as "/sys/kerenl/debug/<dirname>/<mapname>#cpu", where the <dirname> >> is defined with map_update_elem (Note that we do not need to implement >> actual elem operators for relay_map). >> >> If this map is acceptable, other parts including docs, libbpf support, >> selftests, and benchmarks (if need) will be added in the following version. > > looks useful, selftests might be already helpful to see the usage > Ok, I will add selftests in the next version. Thanks. > jirka > >> >> [0] >> https://github.com/torvalds/linux/blob/master/Documentation/filesystems/relay.rst >> [1] >> https://lore.kernel.org/bpf/20231219122850.433be151@gandalf.local.home/T/ >> >> Philo Lu (3): >> bpf: implement relay map basis >> bpf: implement map_update_elem to init relay file >> bpf: introduce bpf_relay_output helper >> >> include/linux/bpf.h | 1 + >> include/linux/bpf_types.h | 3 + >> include/uapi/linux/bpf.h | 17 +++ >> kernel/bpf/Makefile | 3 + >> kernel/bpf/helpers.c | 4 + >> kernel/bpf/relaymap.c | 213 ++++++++++++++++++++++++++++++++++++++ >> kernel/bpf/syscall.c | 1 + >> kernel/bpf/verifier.c | 8 ++ >> kernel/trace/bpf_trace.c | 4 + >> 9 files changed, 254 insertions(+) >> create mode 100644 kernel/bpf/relaymap.c >> >> -- >> 2.32.0.3.g01195cf9f >>