diff options
| author | Paul Buetow <paul@buetow.org> | 2026-06-06 09:22:15 +0300 |
|---|---|---|
| committer | Paul Buetow <paul@buetow.org> | 2026-06-06 09:22:15 +0300 |
| commit | 178ca1f256b1e345cad2f506b6b244fb50d8d281 (patch) | |
| tree | b0d44ee4b4052df12177ea69a8f7d8ae1f6d8034 /cmd | |
| parent | 322496c2863d5bc14b0a5e4af16690bf19073cae (diff) | |
test(timerfd): assert timerfd_settime/gettime fd capture end-to-end
The fd-from-air-eventfd-users workload only called timerfd_create then
closed the fd, so timerfd_settime/timerfd_gettime were never exercised
by any integration test. Commit 6ac9fa4 fixed those two syscalls to
KindFd@arg0 (capturing the operating timerfd via fd_event instead of a
null_event), but that fix had no end-to-end coverage.
Extend the workload to arm the still-open timerfd via timerfd_settime
(1s relative expiry, so it never fires) and read it back via
timerfd_gettime before closing. Assert in TestFdFromAirEventfdUsers that
both enter handlers fire (MinCount>=1) and resolve to the "timerfd:"
path prefix, proving arg0 fd is captured rather than null. Locks in the
6ac9fa4 KindFd fix.
splice/tee are NOT touched: retbytes_test.go already asserts
enter_splice/enter_tee plus positive transfer byte counts, which
inherently exercises their arg0 fd capture, so no new coverage is
needed there.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Diffstat (limited to 'cmd')
| -rw-r--r-- | cmd/ioworkload/scenario_ipc.go | 29 |
1 files changed, 28 insertions, 1 deletions
diff --git a/cmd/ioworkload/scenario_ipc.go b/cmd/ioworkload/scenario_ipc.go index af22505..207966e 100644 --- a/cmd/ioworkload/scenario_ipc.go +++ b/cmd/ioworkload/scenario_ipc.go @@ -77,11 +77,38 @@ func fdFromAirEventfdUsers() error { fd, _, _ = syscall.RawSyscall(unix.SYS_SIGNALFD4, ^uintptr(0), uintptr(unsafe.Pointer(&mask)), uintptr(unsafe.Sizeof(mask))) closeIfValid(int(fd)) + // Create a timerfd and, while it is still open, arm it with + // timerfd_settime and read it back with timerfd_gettime. Both of those + // syscalls take the timerfd as arg0 (kind=fd@arg0), so tracing them + // exercises the fd_event capture path fixed in commit 6ac9fa4: the enter + // handlers must resolve arg0 to the registered "timerfd:" descriptor + // rather than emitting a null event. We close the fd only after both + // operations so the descriptor stays registered for the duration. fd, _, _ = syscall.RawSyscall(unix.SYS_TIMERFD_CREATE, uintptr(unix.CLOCK_MONOTONIC), uintptr(unix.TFD_CLOEXEC), 0) - closeIfValid(int(fd)) + if int(fd) >= 0 { + armAndReadTimerfd(int(fd)) + closeIfValid(int(fd)) + } return nil } +// armAndReadTimerfd arms the given timerfd via timerfd_settime and reads its +// current setting back via timerfd_gettime. The timer is set to a one-second +// relative expiration: far enough in the future that it never actually fires +// during the scenario, so its only observable effect is that the two syscalls +// are issued against an already-open timerfd descriptor. +func armAndReadTimerfd(fd int) { + newValue := unix.ItimerSpec{ + Value: unix.Timespec{Sec: 1, Nsec: 0}, + } + syscall.RawSyscall6(unix.SYS_TIMERFD_SETTIME, uintptr(fd), 0, + uintptr(unsafe.Pointer(&newValue)), 0, 0, 0) + + var curValue unix.ItimerSpec + syscall.RawSyscall(unix.SYS_TIMERFD_GETTIME, uintptr(fd), + uintptr(unsafe.Pointer(&curValue)), 0) +} + func closeIfValid(fd int) { if fd >= 0 { _ = syscall.Close(fd) |
