Debian Codegraph @

upstream/debian-unstable/man2/FUTEX_TRYLOCK_PI.2const

raw
1 .\" Copyright, the authors of the Linux man-pages project
2 .\"
3 .\" %%%LICENSE_START(FREELY_REDISTRIBUTABLE)
4 .\" may be freely modified and distributed
5 .\" %%%LICENSE_END
6 .\"
7 .TH FUTEX_TRYLOCK_PI 2const 2025-09-21 "Linux man-pages 6.18"
8 .SH NAME
9 FUTEX_TRYLOCK_PI \- try to lock a priority-inheritance futex
10 .SH LIBRARY
11 Standard C library
12 .RI ( libc ,\~ \-lc )
13 .SH SYNOPSIS
14 .nf
15 .BR "#include <linux/futex.h>" " /* Definition of " FUTEX_* " constants */"
16 .BR "#include <sys/syscall.h>" " /* Definition of " SYS_* " constants */"
17 .B #include <unistd.h>
18 .P
19 .BI "long syscall(SYS_futex, uint32_t *" uaddr ", FUTEX_TRYLOCK_PI);"
20 .fi
21 .SH DESCRIPTION
22 This operation tries to acquire the lock at
23 .IR uaddr .
24 It is invoked when a user-space atomic acquire did not
25 succeed because the futex word was not 0.
26 .P
27 Because the kernel has access to more state information than user space,
28 acquisition of the lock might succeed if performed by the
29 kernel in cases where the futex word
30 (i.e., the state information accessible to use-space) contains stale state
31 .RB ( FUTEX_WAITERS
32 and/or
33 .BR FUTEX_OWNER_DIED ).
34 This can happen when the owner of the futex died.
35 User space cannot handle this condition in a race-free manner,
36 but the kernel can fix this up and acquire the futex.
37 .\" Paraphrasing a f2f conversation with Thomas Gleixner about the
38 .\" above point (Aug 2015): ###
39 .\" There is a rare possibility of a race condition involving an
40 .\" uncontended futex with no owner, but with waiters. The
41 .\" kernel-user-space contract is that if a futex is nonzero, you must
42 .\" go into kernel. The futex was owned by a task, and that task dies
43 .\" but there are no waiters, so the futex value is non zero.
44 .\" Therefore, the next locker has to go into the kernel,
45 .\" so that the kernel has a chance to clean up. (CMXCH on zero
46 .\" in user space would fail, so kernel has to clean up.)
47 .\" Darren Hart (Oct 2015):
48 .\" The trylock in the kernel has more state, so it can independently
49 .\" verify the flags that user space must trust implicitly.
50 .\"
51 .SH RETURN VALUE
52 On error,
53 \-1 is returned,
54 and
55 .I errno
56 is set to indicate the error.
57 .P
58 On success,
59 .B FUTEX_TRYLOCK_PI
60 Returns 0 if the futex was successfully locked.
61 .SH ERRORS
62 See
63 .BR futex (2).
64 .TP
65 .B EAGAIN
66 The futex owner thread ID of
67 .I uaddr
68 is about to exit,
69 but has not yet handled the internal state cleanup.
70 Try again.
71 .TP
72 .B EDEADLK
73 The futex word at
74 .I uaddr
75 is already locked by the caller.
76 .TP
77 .B EINVAL
78 The kernel detected an inconsistency between the user-space state at
79 .I uaddr
80 and the kernel state.
81 This indicates either state corruption
82 or that the kernel found a waiter on
83 .I uaddr
84 which is waiting via
85 .BR FUTEX_WAIT (2const)
86 or
87 .BR FUTEX_WAIT_BITSET (2const).
88 .TP
89 .B ENOMEM
90 The kernel could not allocate memory to hold state information.
91 .TP
92 .B ENOSYS
93 A run-time check determined that the operation is not available.
94 The PI-futex operations are not implemented on all architectures and
95 are not supported on some CPU variants.
96 .TP
97 .B EPERM
98 The caller is not allowed to attach itself to the futex at
99 .IR uaddr .
100 (This may be caused by a state corruption in user space.)
101 .TP
102 .B ESRCH
103 The thread ID in the futex word at
104 .I uaddr
105 does not exist.
106 .SH STANDARDS
107 Linux.
108 .SH HISTORY
109 Linux 2.6.18.
110 .\" commit c87e2837be82df479a6bae9f155c43516d2feebc
111 .SH SEE ALSO
112 .BR futex (2)