Operating System: Ubuntu 16.04.4 Kernel: 4.16.0-041600-generic Running ureadahead from terminal as root with # strace ureadahead --verbose --force-trace / produces the output: [..] openat(3, "events/fs/do_sys_open/enable", O_RDWR) = -1 ENOENT (No such file or directory) close(3) = 0 write(2, "ureadahead: Error while tracing:"..., 59ureadahead: Error while tracing: No such file or directory [..] Looking into this the directory/file in fact does not exist: "/sys/kernel/debug/tracing/events/fs/do_sys_open/enable" The reason I became aware of this is because systemd is reporting that ureadahead.service is failing: * ureadahead.service - Read required files in advance Loaded: loaded (/lib/systemd/system/ureadahead.service; enabled; vendor preset: enabled) Active: failed (Result: exit-code) since Sun 2018-05-27 23:41:51 EDT; 3h 38min ago Process: 282 ExecStart=/sbin/ureadahead (code=exited, status=5) Main PID: 282 (code=exited, status=5) May 27 23:41:51 UbuntuHP ureadahead[282]: ureadahead: Error while tracing: No such file or directory May 27 23:41:51 UbuntuHP ureadahead[282]: Counted 4 CPUs May 27 23:41:51 UbuntuHP systemd[1]: ureadahead.service: Main process exited, code=exited, status=5/NOTINSTALLED May 27 23:41:51 UbuntuHP systemd[1]: ureadahead.service: Unit entered failed state. May 27 23:41:51 UbuntuHP systemd[1]: ureadahead.service: Failed with result 'exit-code'. May 27 23:42:59 UbuntuHP systemd[1]: Stopped Read required files in advance.
One reported solution on stackexchange was to apply a patch to the kernel : https://unix.stackexchange.com/questions/308801/ureadahead-error-while-tracing-possible-kernel-patch-problem Others are reporting this bug as well on launchpad with different kernel.. https://bugs.launchpad.net/ubuntu/+source/ureadahead/+bug/1763584
I don't know the Ubuntu kernel, but does your kernel have this commit? 1c758a2202a6 ("tracing/x86: Update syscall trace events to handle new prefixed syscall func names") ?? -- Steve
(In reply to Steven Rostedt from comment #2) > I don't know the Ubuntu kernel, but does your kernel have this commit? > > 1c758a2202a6 ("tracing/x86: Update syscall trace events to handle new > prefixed syscall func names") > > ?? > > -- Steve I completely forgot about this issue, but also my email address changed since I commented. However, I just recently did a custom compiled PREEMPT-RT kernel for version 5.6.19-rt. Using this patch: http://update.rootmaster.info/linux-kernel/projects/rt/5.6/ ) This is the content of ~/dev/kernel/linux5.6.19/arch/x86/include/asm/ftrace.h in regards to this bug report and your question about the commit, which seems its not implemented here: ========================================================== /* SPDX-License-Identifier: GPL-2.0 */ #ifndef _ASM_X86_FTRACE_H #define _ASM_X86_FTRACE_H #ifdef CONFIG_FUNCTION_TRACER #ifndef CC_USING_FENTRY # error Compiler does not support fentry? #endif # define MCOUNT_ADDR ((unsigned long)(__fentry__)) #define MCOUNT_INSN_SIZE 5 /* sizeof mcount call */ #ifdef CONFIG_DYNAMIC_FTRACE #define ARCH_SUPPORTS_FTRACE_OPS 1 #endif #define HAVE_FUNCTION_GRAPH_RET_ADDR_PTR #ifndef __ASSEMBLY__ extern atomic_t modifying_ftrace_code; extern void __fentry__(void); static inline unsigned long ftrace_call_adjust(unsigned long addr) { /* * addr is the address of the mcount call instruction. * recordmcount does the necessary offset calculation. */ return addr; } /* * When a ftrace registered caller is tracing a function that is * also set by a register_ftrace_direct() call, it needs to be * differentiated in the ftrace_caller trampoline. To do this, we * place the direct caller in the ORIG_AX part of pt_regs. This * tells the ftrace_caller that there's a direct caller. */ static inline void arch_ftrace_set_direct_caller(struct pt_regs *regs, unsigned long addr) { /* Emulate a call */ regs->orig_ax = addr; } #ifdef CONFIG_DYNAMIC_FTRACE struct dyn_arch_ftrace { /* No extra data needed for x86 */ }; #define FTRACE_GRAPH_TRAMP_ADDR FTRACE_GRAPH_ADDR #endif /* CONFIG_DYNAMIC_FTRACE */ #endif /* __ASSEMBLY__ */ #endif /* CONFIG_FUNCTION_TRACER */ #ifndef __ASSEMBLY__ #if defined(CONFIG_FUNCTION_TRACER) && defined(CONFIG_DYNAMIC_FTRACE) extern void set_ftrace_ops_ro(void); #else static inline void set_ftrace_ops_ro(void) { } #endif #define ARCH_HAS_SYSCALL_MATCH_SYM_NAME static inline bool arch_syscall_match_sym_name(const char *sym, const char *name) { /* * Compare the symbol name with the system call name. Skip the * "__x64_sys", "__ia32_sys" or simple "sys" prefix. */ return !strcmp(sym + 3, name + 3) || (!strncmp(sym, "__x64_", 6) && !strcmp(sym + 9, name + 3)) || (!strncmp(sym, "__ia32_", 7) && !strcmp(sym + 10, name + 3)); } #ifndef COMPILE_OFFSETS #if defined(CONFIG_FTRACE_SYSCALLS) && defined(CONFIG_IA32_EMULATION) #include <linux/compat.h> /* * Because ia32 syscalls do not map to x86_64 syscall numbers * this screws up the trace output when tracing a ia32 task. * Instead of reporting bogus syscalls, just do not trace them. * * If the user really wants these, then they should use the * raw syscall tracepoints with filtering. */ #define ARCH_TRACE_IGNORE_COMPAT_SYSCALLS 1 static inline bool arch_trace_is_compat_syscall(struct pt_regs *regs) { return in_32bit_syscall(); } #endif /* CONFIG_FTRACE_SYSCALLS && CONFIG_IA32_EMULATION */ #endif /* !COMPILE_OFFSETS */ #endif /* !__ASSEMBLY__ */ #endif /* _ASM_X86_FTRACE_H */ ========================================================== Please note I'm not working with the kernel from the linux repository (git), it's the zipped source code found here: https://cdn.kernel.org/pub/linux/kernel/v5.x/ *** However from the kernel repository which I checked I did see: demon@Bionic18:~/dev/kernel/linux$ git log 1c758a2202a6 commit 1c758a2202a6b4624d0703013a2c6cfa6e7455aa Author: Steven Rostedt (VMware) <rostedt@goodmis.org> Date: Tue Apr 17 17:41:28 2018 -0400 *** Thank you for your assistance. Please advice on what should be done to properly get ureadahead to function.