bpf: Add new bpf_task_storage_get proto with no deadlock detection
authorMartin KaFai Lau <martin.lau@kernel.org>
Tue, 25 Oct 2022 18:45:20 +0000 (11:45 -0700)
committerAlexei Starovoitov <ast@kernel.org>
Wed, 26 Oct 2022 06:11:46 +0000 (23:11 -0700)
commit4279adb094a17132423f1271c3d11b593fc2327e
tree0601ed0271392f68e41f4672f63565bf27df3989
parente8b02296a6b8d07de752d6157d863a642117bcd3
bpf: Add new bpf_task_storage_get proto with no deadlock detection

The bpf_lsm and bpf_iter do not recur that will cause a deadlock.
The situation is similar to the bpf_pid_task_storage_lookup_elem()
which is called from the syscall map_lookup_elem.  It does not need
deadlock detection.  Otherwise, it will cause unnecessary failure
when calling the bpf_task_storage_get() helper.

This patch adds bpf_task_storage_get proto that does not do deadlock
detection.  It will be used by bpf_lsm and bpf_iter programs.

Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://lore.kernel.org/r/20221025184524.3526117-6-martin.lau@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
include/linux/bpf.h
kernel/bpf/bpf_task_storage.c
kernel/trace/bpf_trace.c