| Message ID | 20230810081319.65668-4-zhouchuyi@bytedance.com (mailing list archive) |
|---|---|
| State | RFC |
| Headers | show |
| Series | mm: Select victim using bpf_oom_evaluate_task | expand |
| Context | Check | Description |
|---|---|---|
| bpf/vmtest-bpf-next-PR | success | PR summary |
| bpf/vmtest-bpf-next-VM_Test-1 | success | Logs for ${{ matrix.test }} on ${{ matrix.arch }} with ${{ matrix.toolchain_full }} |
| bpf/vmtest-bpf-next-VM_Test-2 | success | Logs for ShellCheck |
| bpf/vmtest-bpf-next-VM_Test-3 | fail | Logs for build for aarch64 with gcc |
| bpf/vmtest-bpf-next-VM_Test-4 | fail | Logs for build for s390x with gcc |
| bpf/vmtest-bpf-next-VM_Test-5 | fail | Logs for build for x86_64 with gcc |
| bpf/vmtest-bpf-next-VM_Test-6 | fail | Logs for build for x86_64 with llvm-16 |
| bpf/vmtest-bpf-next-VM_Test-7 | success | Logs for set-matrix |
| bpf/vmtest-bpf-next-VM_Test-8 | success | Logs for veristat |
| netdev/tree_selection | success | Not a local patch, async |
On 10/08/2023 09:13, Chuyi Zhou wrote: > This patch add a tracepoint to mark the scenario where nothing was > chosen for OOM killer. This would allow BPF programs to catch the fact > that the BPF OOM policy didn't work well. > > Suggested-by: Alan Maguire <alan.maguire@oracle.com> > Signed-off-by: Chuyi Zhou <zhouchuyi@bytedance.com> > --- > include/trace/events/oom.h | 18 ++++++++++++++++++ > mm/oom_kill.c | 1 + > 2 files changed, 19 insertions(+) > > diff --git a/include/trace/events/oom.h b/include/trace/events/oom.h > index 26a11e4a2c36..b6ae1134229c 100644 > --- a/include/trace/events/oom.h > +++ b/include/trace/events/oom.h > @@ -6,6 +6,7 @@ > #define _TRACE_OOM_H > #include <linux/tracepoint.h> > #include <trace/events/mmflags.h> > +#include <linux/oom.h> > > TRACE_EVENT(oom_score_adj_update, > > @@ -151,6 +152,23 @@ TRACE_EVENT(skip_task_reaping, > TP_printk("pid=%d", __entry->pid) > ); > > +TRACE_EVENT(select_bad_process_end, > + would oom_select_bad_process_fail be a better name here? "_end" is kind of neutral, whereas "_fail" indicates something unexpected happened. > + TP_PROTO(struct oom_control *oc), > + > + TP_ARGS(oc), > + > + TP_STRUCT__entry( > + __array(char, policy_name, POLICY_NAME_LEN) > + ), > + > + TP_fast_assign( > + memcpy(__entry->policy_name, oc->policy_name, POLICY_NAME_LEN); > + ), > + > + TP_printk("policy_name=%s", __entry->policy_name) > +); > + > #ifdef CONFIG_COMPACTION > TRACE_EVENT(compact_retry, > > diff --git a/mm/oom_kill.c b/mm/oom_kill.c > index 3239dcdba4d7..af40a1b750fa 100644 > --- a/mm/oom_kill.c > +++ b/mm/oom_kill.c > @@ -1235,6 +1235,7 @@ bool out_of_memory(struct oom_control *oc) > select_bad_process(oc); > /* Found nothing?!?! */ > if (!oc->chosen) { > + trace_select_bad_process_end(oc); > dump_header(oc, NULL); > pr_warn("Out of memory and no killable processes...\n"); > /*
diff --git a/include/trace/events/oom.h b/include/trace/events/oom.h index 26a11e4a2c36..b6ae1134229c 100644 --- a/include/trace/events/oom.h +++ b/include/trace/events/oom.h @@ -6,6 +6,7 @@ #define _TRACE_OOM_H #include <linux/tracepoint.h> #include <trace/events/mmflags.h> +#include <linux/oom.h> TRACE_EVENT(oom_score_adj_update, @@ -151,6 +152,23 @@ TRACE_EVENT(skip_task_reaping, TP_printk("pid=%d", __entry->pid) ); +TRACE_EVENT(select_bad_process_end, + + TP_PROTO(struct oom_control *oc), + + TP_ARGS(oc), + + TP_STRUCT__entry( + __array(char, policy_name, POLICY_NAME_LEN) + ), + + TP_fast_assign( + memcpy(__entry->policy_name, oc->policy_name, POLICY_NAME_LEN); + ), + + TP_printk("policy_name=%s", __entry->policy_name) +); + #ifdef CONFIG_COMPACTION TRACE_EVENT(compact_retry, diff --git a/mm/oom_kill.c b/mm/oom_kill.c index 3239dcdba4d7..af40a1b750fa 100644 --- a/mm/oom_kill.c +++ b/mm/oom_kill.c @@ -1235,6 +1235,7 @@ bool out_of_memory(struct oom_control *oc) select_bad_process(oc); /* Found nothing?!?! */ if (!oc->chosen) { + trace_select_bad_process_end(oc); dump_header(oc, NULL); pr_warn("Out of memory and no killable processes...\n"); /*
This patch add a tracepoint to mark the scenario where nothing was chosen for OOM killer. This would allow BPF programs to catch the fact that the BPF OOM policy didn't work well. Suggested-by: Alan Maguire <alan.maguire@oracle.com> Signed-off-by: Chuyi Zhou <zhouchuyi@bytedance.com> --- include/trace/events/oom.h | 18 ++++++++++++++++++ mm/oom_kill.c | 1 + 2 files changed, 19 insertions(+)