kernel.git/kernel/bpf/syscall.c, branch linux-5.1.y

bpf: fix unconnected udp hooks

2019-07-03T11:13:43Z

commit 983695fa676568fc0fe5ddd995c7267aabc24632 upstream. Intention of cgroup bind/connect/sendmsg BPF hooks is to act transparently to applications as also stated in original motivation in 7828f20e3779 ("Merge branch 'bpf-cgroup-bind-connect'"). When recently integrating the latter two hooks into Cilium to enable host based load-balancing with Kubernetes, I ran into the issue that pods couldn't start up as DNS got broken. Kubernetes typically sets up DNS as a service and is thus subject to load-balancing. Upon further debugging, it turns out that the cgroupv2 sendmsg BPF hooks API is currently insufficient and thus not usable as-is for standard applications shipped with most distros. To break down the issue we ran into with a simple example: # cat /etc/resolv.conf nameserver 147.75.207.207 nameserver 147.75.207.208 For the purpose of a simple test, we set up above IPs as service IPs and transparently redirect traffic to a different DNS backend server for that node: # cilium service list ID Frontend Backend 1 147.75.207.207:53 1 => 8.8.8.8:53 2 147.75.207.208:53 1 => 8.8.8.8:53 The attached BPF program is basically selecting one of the backends if the service IP/port matches on the cgroup hook. DNS breaks here, because the hooks are not transparent enough to applications which have built-in msg_name address checks: # nslookup 1.1.1.1 ;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.207#53 ;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.208#53 ;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.207#53 [...] ;; connection timed out; no servers could be reached # dig 1.1.1.1 ;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.207#53 ;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.208#53 ;; reply from unexpected source: 8.8.8.8#53, expected 147.75.207.207#53 [...] ; <<>> DiG 9.11.3-1ubuntu1.7-Ubuntu <<>> 1.1.1.1 ;; global options: +cmd ;; connection timed out; no servers could be reached For comparison, if none of the service IPs is used, and we tell nslookup to use 8.8.8.8 directly it works just fine, of course: # nslookup 1.1.1.1 8.8.8.8 1.1.1.1.in-addr.arpa name = one.one.one.one. In order to fix this and thus act more transparent to the application, this needs reverse translation on recvmsg() side. A minimal fix for this API is to add similar recvmsg() hooks behind the BPF cgroups static key such that the program can track state and replace the current sockaddr_in{,6} with the original service IP. From BPF side, this basically tracks the service tuple plus socket cookie in an LRU map where the reverse NAT can then be retrieved via map value as one example. Side-note: the BPF cgroups static key should be converted to a per-hook static key in future. Same example after this fix: # cilium service list ID Frontend Backend 1 147.75.207.207:53 1 => 8.8.8.8:53 2 147.75.207.208:53 1 => 8.8.8.8:53 Lookups work fine now: # nslookup 1.1.1.1 1.1.1.1.in-addr.arpa name = one.one.one.one. Authoritative answers can be found from: # dig 1.1.1.1 ; <<>> DiG 9.11.3-1ubuntu1.7-Ubuntu <<>> 1.1.1.1 ;; global options: +cmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NXDOMAIN, id: 51550 ;; flags: qr rd ra ad; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 1 ;; OPT PSEUDOSECTION: ; EDNS: version: 0, flags:; udp: 512 ;; QUESTION SECTION: ;1.1.1.1. IN A ;; AUTHORITY SECTION: . 23426 IN SOA a.root-servers.net. nstld.verisign-grs.com. 2019052001 1800 900 604800 86400 ;; Query time: 17 msec ;; SERVER: 147.75.207.207#53(147.75.207.207) ;; WHEN: Tue May 21 12:59:38 UTC 2019 ;; MSG SIZE rcvd: 111 And from an actual packet level it shows that we're using the back end server when talking via 147.75.207.20{7,8} front end: # tcpdump -i any udp [...] 12:59:52.698732 IP foo.42011 > google-public-dns-a.google.com.domain: 18803+ PTR? 1.1.1.1.in-addr.arpa. (38) 12:59:52.698735 IP foo.42011 > google-public-dns-a.google.com.domain: 18803+ PTR? 1.1.1.1.in-addr.arpa. (38) 12:59:52.701208 IP google-public-dns-a.google.com.domain > foo.42011: 18803 1/0/0 PTR one.one.one.one. (67) 12:59:52.701208 IP google-public-dns-a.google.com.domain > foo.42011: 18803 1/0/0 PTR one.one.one.one. (67) [...] In order to be flexible and to have same semantics as in sendmsg BPF programs, we only allow return codes in [1,1] range. In the sendmsg case the program is called if msg->msg_name is present which can be the case in both, connected and unconnected UDP. The former only relies on the sockaddr_in{,6} passed via connect(2) if passed msg->msg_name was NULL. Therefore, on recvmsg side, we act in similar way to call into the BPF program whenever a non-NULL msg->msg_name was passed independent of sk->sk_state being TCP_ESTABLISHED or not. Note that for TCP case, the msg->msg_name is ignored in the regular recvmsg path and therefore not relevant. For the case of ip{,v6}_recv_error() paths, picked up via MSG_ERRQUEUE, the hook is not called. This is intentional as it aligns with the same semantics as in case of TCP cgroup BPF hooks right now. This might be better addressed in future through a different bpf_attach_type such that this case can be distinguished from the regular recvmsg paths, for example. Fixes: 1cedee13d25a ("bpf: Hooks for sys_sendmsg") Signed-off-by: Daniel Borkmann Acked-by: Andrey Ignatov Acked-by: Martin KaFai Lau Acked-by: Martynas Pumputis Signed-off-by: Alexei Starovoitov Signed-off-by: Daniel Borkmann Signed-off-by: Greg Kroah-Hartman

bpf: add map_lookup_elem_sys_only for lookups from syscall side

2019-05-25T16:16:49Z

commit c6110222c6f49ea68169f353565eb865488a8619 upstream. Add a callback map_lookup_elem_sys_only() that map implementations could use over map_lookup_elem() from system call side in case the map implementation needs to handle the latter differently than from the BPF data path. If map_lookup_elem_sys_only() is set, this will be preferred pick for map lookups out of user space. This hook is used in a follow-up fix for LRU map, but once development window opens, we can convert other map types from map_lookup_elem() (here, the one called upon BPF_MAP_LOOKUP_ELEM cmd is meant) over to use the callback to simplify and clean up the latter. Signed-off-by: Daniel Borkmann Acked-by: Martin KaFai Lau Signed-off-by: Alexei Starovoitov Signed-off-by: Greg Kroah-Hartman

bpf: Try harder when allocating memory for large maps

2019-03-18T15:48:25Z

It has been observed that sometimes a higher order memory allocation for BPF maps fails when there is no obvious memory pressure in a system. E.g. the map (BPF_MAP_TYPE_LRU_HASH, key=38, value=56, max_elems=524288) could not be created due to vmalloc unable to allocate 75497472B, when the system's memory consumption (in MB) was the following: Total: 3942 Used: 837 (21.24%) Free: 138 Buffers: 239 Cached: 2727 Later analysis [1] by Michal Hocko showed that the vmalloc was not trying to reclaim memory from the page cache and was failing prematurely due to __GFP_NORETRY. Considering dcda9b0471 ("mm, tree wide: replace __GFP_REPEAT by __GFP_RETRY_MAYFAIL with more useful semantic") and [1], we can replace __GFP_NORETRY with __GFP_RETRY_MAYFAIL, as it won't invoke OOM killer and will try harder to fulfil allocation requests. Unfortunately, replacing the body of the BPF map memory allocation function with the kvmalloc_node helper function is not an option at this point in time, given 1) kmalloc is non-optional for higher order allocations, and 2) passing __GFP_RETRY_MAYFAIL to the kmalloc would stress the slab allocator too much for large requests. The change has been tested with the workloads mentioned above and by observing oom_kill value from /proc/vmstat. [1]: https://lore.kernel.org/bpf/20190310071318.GW5232@dhcp22.suse.cz/ Signed-off-by: Martynas Pumputis Acked-by: Yonghong Song Cc: Michal Hocko Signed-off-by: Daniel Borkmann Link: https://lore.kernel.org/bpf/20190318153940.GL8924@dhcp22.suse.cz/

Merge branch 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2019-03-06T15:59:36Z

Pull perf updates from Ingo Molnar: "Lots of tooling updates - too many to list, here's a few highlights: - Various subcommand updates to 'perf trace', 'perf report', 'perf record', 'perf annotate', 'perf script', 'perf test', etc. - CPU and NUMA topology and affinity handling improvements, - HW tracing and HW support updates: - Intel PT updates - ARM CoreSight updates - vendor HW event updates - BPF updates - Tons of infrastructure updates, both on the build system and the library support side - Documentation updates. - ... and lots of other changes, see the changelog for details. Kernel side updates: - Tighten up kprobes blacklist handling, reduce the number of places where developers can install a kprobe and hang/crash the system. - Fix/enhance vma address filter handling. - Various PMU driver updates, small fixes and additions. - refcount_t conversions - BPF updates - error code propagation enhancements - misc other changes" * 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (238 commits) perf script python: Add Python3 support to syscall-counts-by-pid.py perf script python: Add Python3 support to syscall-counts.py perf script python: Add Python3 support to stat-cpi.py perf script python: Add Python3 support to stackcollapse.py perf script python: Add Python3 support to sctop.py perf script python: Add Python3 support to powerpc-hcalls.py perf script python: Add Python3 support to net_dropmonitor.py perf script python: Add Python3 support to mem-phys-addr.py perf script python: Add Python3 support to failed-syscalls-by-pid.py perf script python: Add Python3 support to netdev-times.py perf tools: Add perf_exe() helper to find perf binary perf script: Handle missing fields with -F +.. perf data: Add perf_data__open_dir_data function perf data: Add perf_data__(create_dir|close_dir) functions perf data: Fail check_backup in case of error perf data: Make check_backup work over directories perf tools: Add rm_rf_perf_data function perf tools: Add pattern name checking to rm_rf perf tools: Add depth checking to rm_rf perf data: Add global path holder ...

Merge git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next

2019-03-04T18:14:31Z

Daniel Borkmann says: ==================== pull-request: bpf-next 2019-03-04 The following pull-request contains BPF updates for your *net-next* tree. The main changes are: 1) Add AF_XDP support to libbpf. Rationale is to facilitate writing AF_XDP applications by offering higher-level APIs that hide many of the details of the AF_XDP uapi. Sample programs are converted over to this new interface as well, from Magnus. 2) Introduce a new cant_sleep() macro for annotation of functions that cannot sleep and use it in BPF_PROG_RUN() to assert that BPF programs run under preemption disabled context, from Peter. 3) Introduce per BPF prog stats in order to monitor the usage of BPF; this is controlled by kernel.bpf_stats_enabled sysctl knob where monitoring tools can make use of this to efficiently determine the average cost of programs, from Alexei. 4) Split up BPF selftest's test_progs similarly as we already did with test_verifier. This allows to further reduce merge conflicts in future and to get more structure into our quickly growing BPF selftest suite, from Stanislav. 5) Fix a bug in BTF's dedup algorithm which can cause an infinite loop in some circumstances; also various BPF doc fixes and improvements, from Andrii. 6) Various BPF sample cleanups and migration to libbpf in order to further isolate the old sample loader code (so we can get rid of it at some point), from Jakub. 7) Add a new BPF helper for BPF cgroup skb progs that allows to set ECN CE code point and a Host Bandwidth Manager (HBM) sample program for limiting the bandwidth used by v2 cgroups, from Lawrence. 8) Enable write access to skb->queue_mapping from tc BPF egress programs in order to let BPF pick TX queue, from Jesper. 9) Fix a bug in BPF spinlock handling for map-in-map which did not propagate spin_lock_off to the meta map, from Yonghong. 10) Fix a bug in the new per-CPU BPF prog counters to properly initialize stats for each CPU, from Eric. 11) Add various BPF helper prototypes to selftest's bpf_helpers.h, from Willem. 12) Fix various BPF samples bugs in XDP and tracing progs, from Toke, Daniel and Yonghong. 13) Silence preemption splat in test_bpf after BPF_PROG_RUN() enforces it now everywhere, from Anders. 14) Fix a signedness bug in libbpf's btf_dedup_ref_type() to get error handling working, from Dan. 15) Fix bpftool documentation and auto-completion with regards to stream_{verdict,parser} attach types, from Alban. ==================== Signed-off-by: David S. Miller

Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net

2019-03-02T20:54:35Z

bpf: drop refcount if bpf_map_new_fd() fails in map_create()

2019-03-01T15:04:29Z

In bpf/syscall.c, map_create() first set map->usercnt to 1, a file descriptor is supposed to return to userspace. When bpf_map_new_fd() fails, drop the refcount. Fixes: bd5f5f4ecb78 ("bpf: Add BPF_MAP_GET_FD_BY_ID") Signed-off-by: Peng Sun Acked-by: Martin KaFai Lau Signed-off-by: Alexei Starovoitov Signed-off-by: Daniel Borkmann

Merge branch 'linus' into perf/core, to pick up fixes

2019-02-28T07:27:17Z

Signed-off-by: Ingo Molnar

bpf: expose program stats via bpf_prog_info

2019-02-27T16:22:50Z

Return bpf program run_time_ns and run_cnt via bpf_prog_info Signed-off-by: Alexei Starovoitov Acked-by: Andrii Nakryiko Signed-off-by: Daniel Borkmann

bpf: enable program stats

2019-02-27T16:22:50Z

JITed BPF programs are indistinguishable from kernel functions, but unlike kernel code BPF code can be changed often. Typical approach of "perf record" + "perf report" profiling and tuning of kernel code works just as well for BPF programs, but kernel code doesn't need to be monitored whereas BPF programs do. Users load and run large amount of BPF programs. These BPF stats allow tools monitor the usage of BPF on the server. The monitoring tools will turn sysctl kernel.bpf_stats_enabled on and off for few seconds to sample average cost of the programs. Aggregated data over hours and days will provide an insight into cost of BPF and alarms can trigger in case given program suddenly gets more expensive. The cost of two sched_clock() per program invocation adds ~20 nsec. Fast BPF progs (like selftests/bpf/progs/test_pkt_access.c) will slow down from ~10 nsec to ~30 nsec. static_key minimizes the cost of the stats collection. There is no measurable difference before/after this patch with kernel.bpf_stats_enabled=0 Signed-off-by: Alexei Starovoitov Signed-off-by: Daniel Borkmann