IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
* sigreturn.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* defs.h (sprintsigmask_n): New prototype.
(tprintsigmask_addr): New macro.
* signal.c (sprintsigmask_n): Make global.
(tprintsigmask_addr): Remove.
(sys_sigreturn): Move to sigreturn.c.
Implement additional synchronization between parent and child processes
to guarantee that the child starts closing connected socket only after
exiting of the parent's accept() syscall.
This guarantee seems to be necessary to reliably receive UNIX_DIAG_PEER
messages from NETLINK_SOCK_DIAG interface.
* tests/net-accept-connect.c: Implement additional synchronization
between parent and child processes.
* tests/unix-yy-connect.awk: Update.
* configure.ac (ASM_SIGRTMIN): Define to SIGRTMIN from <asm/signal.h>.
* signal.c: Use ASM_SIGRTMIN instead of constants provided by libc.
* tests/sigreturn.c: Use ASM_SIGRTMIN instead of hardcoded value.
Use lower RT_* numbers to support pre-3.18 hppa kernels.
* tests/sigreturn.test: Update regexp.
Fix compilation warning introduced by commit
577be2593d
* signal.c (sys_sigreturn) [POWERPC64]: Fix initialization of signal
mask on 32bit personality.
Fix compilation warnings introduced by commit
5b9b7e1d34:
signal.c:702:45: warning: signed and unsigned type in conditional expression [-Wsign-compare]
syscall.c:746:34: warning: initialization from incompatible pointer type [enabled by default]
* syscall.c [AARCH64] (aarch64_sp_ptr): Add explicit cast.
[AARCH64] (arm_sp_ptr): Change pointer type to unsigned.
* linux/aarch64/arch_regs.h (arm_sp_ptr): Update.
Reported-by: Mike Frysinger <vapier@gentoo.org>
Although sigreturn takes signal mask via frame pointer,
it's more convenient to display signal mask as a syscall argument
rather than an outstanding object of unknown nature:
Before this change:
sigreturn() (mask [USR2 CHLD RT_2 RT_3 RT_4 RT_31 RT_32]) = 0
After this change:
sigreturn({mask=[USR2 CHLD RT_2 RT_3 RT_4 RT_31 RT_32]}) = 0
* signal.c (sys_sigreturn): Display signal mask as a syscall argument.
* linux/64/syscallent.h (139): Use sys_sigreturn for rt_sigreturn
decoding.
* syscall.c [ARM] (arm_regs): Make static.
[ARM] (arm_sp_ptr): New variable.
[AARCH64] (aarch64_sp_ptr, arm_sp_ptr): New variables.
* linux/aarch64/arch_regs.h: New file.
* linux/arm/arch_regs.h (arm_regs): Remove.
(arm_sp_ptr): New declaration.
* signal.c (sys_sigreturn) [ARM]: Use arm_sp_ptr.
[AARCH64]: Print signal mask.
Without this users might get confused why `man strace` describes
an option that doesn't work.
* strace.1: Note that -k is available only if strace is built
with libunwind.
* debian/control (strace, strace-udeb): Add mips64 and mips64el
to architecture list.
Patch by James Cowgill <james410@cowgill.org.uk>
This fixes Debian bug #773375.
In mips n32 abi, like in most of 32-bit architectures, offset
is passed to preadv/pwritev syscalls using two syscall arguments.
* io.c (print_llu_from_low_high_val) [LINUX_MIPSN32]: Remove.
Rather than trigger an ERROR which fails `make check`, go with SKIP
instead. We don't want the testsuite failing due to kernel bugs.
* tests/ipc_sem.c (main): Change 99 to 77.
The s390 hardware can't support an exact si_addr, and the current kernels
don't handle text addresses correctly at all. Until that improves, skip
the test on s390.
* tests/pc.c (main): Return 77 on s390 systems.
As ioctls_inc.h generated natively on x32 differs both from 32/ioctls_inc.h
(by 45 entries) and from linux/64/ioctls_inc.h (by 322 entries), use x32
specific ioctl entries on x32 personality.
* linux/x32/ioctls_inc0.h: Generate from linux v3.19 using ioctls_gen.sh.
* linux/x86_64/ioctlent2.h: Remove.
* linux/x86_64/ioctls_arch2.h: New file.
* linux/x86_64/ioctls_inc2.h: New file.
* Makefile.am (EXTRA_DIST): Remove linux/x86_64/ioctlent2.h, add
linux/x86_64/ioctls_arch2.h and linux/x86_64/ioctls_inc2.h.
* tests/uid.c (main): Ignore uids returned by getresuid to avoid
16bit vs 32bit issues on big endian architectures.
Change real UID instead of effective UID in setresuid test.
* tests/uid16.c (main): Likewise.
* tests/uid32.c (main): Change real UID instead of effective UID
in setresuid test to match uid.c
* tests/uid.awk: Update setresuid regexp.
Reported-by: Mike Frysinger <vapier@gentoo.org>
If the tools we use call older getdents syscalls where d_type isn't
passed back, or the arch is old (like Alpha) and can't pass back in
either version, make sure we don't fail.
* tests/getdents.awk (d_type_dir, d_type_reg): Accept DT_UNKNOWN.
Fix regression introduced by commit v4.9-9-gc3a5c01.
* signal.c (HAVE_SA_RESTORER): New macro.
[HPPA || IA64]: Do not undefine SA_RESTORER.
(old_sigaction, old_sigaction32, new_sigaction): Use HAVE_SA_RESTORER
instead of SA_RESTORER to check whether to define sa_restorer.
(decode_old_sigaction, decode_new_sigaction): Use HAVE_SA_RESTORER in
addition to SA_RESTORER to check whether to use sa_restorer.
* strace.c (startup_child): Check that the length of the command strace
is going to execute does not exceed PATH_MAX limit.
Reported-by: Cheolung Lee <chpie@grayhash.com>
Restore the order of including <sys/reg.h> and <linux/ptrace.h>
headers that was inadvertently changed by commit v4.9-280-g5503dd2.
This should fix build on platforms like m68k where some constants
are defined simultaneously as enums by <sys/reg.h> and as macros
by <asm/ptrace.h>.
* process.c: Include "ptrace.h" after "regs.h".
* signal.c: Likewise.
* syscall.c: Likewise.
* util.c: Likewise.
As time_t is a 64-bit type in x32 personality, it is not correct to
treat it as a long int type.
* time.c (current_time_t_is_compat): New macro.
(sprinttv, sprint_timespec, printitv_bitness, tprint_timex): Use it.
(tprint_timeval, do_sprinttv, sprint_timespec): Use j and uintmax_t
when printing tv_sec/tv_usec.
(tprint_timex): Use j and intmax_t/uintmax_t when printing struct timex
members of type kernel_long_t.
When running 32bit ipc tests on ppc/sparc with a 64bit kernel, the ipc
logic wrongly returns errors. Detect that and throw an error.
* tests/ipc_sem.c: Return 99 when errno is EFAULT.
* tests/ipc_sem.test: Save output to $OUT and show it when failing.
Handle exit status 99 to throw an ERROR.
For a few linux releases (3.1 through 3.3), this define in the exported
headers were broken. Redefine if that's the case.
* ptrace.h [PTRACE_EVENT_STOP]: Redefine if PTRACE_EVENT_STOP is 7.
When files get updated, the bootstrap script should make sure the
parallel dirs are kept in sync.
* bootstrap: Always generate the test$m. Clean up symlinks first.
Don't link in Makefile.in files.
Use the AX_PROG_CC_FOR_BUILD helper to set up build settings when cross
compiling. This way ioctlsort uses the build tools all the time.
* configure.ac: Call AX_PROG_CC_FOR_BUILD.
* Makefile.am (ioctlsort_CC): Set to CC_FOR_BUILD.
(ioctlsort_CPPFLAGS): Change CPPFLAGS to CPPFLAGS_FOR_BUILD.
(ioctlsort_CFLAGS): Change CFLAGS to CFLAGS_FOR_BUILD.
(ioctlsort_LDFLAGS): Change LDFLAGS to LDFLAGS_FOR_BUILD.
(ioctlsort_LD): Delete.
(ioctlsort%): Change to ioctlsort%$(BUILD_EXEEXT). Use ioctlsort_CFLAGS.
* m4/ax_prog_cc_for_build.m4: Import from the autoconf-archive package.
* configure.ac: Check for dladdr in -ldl.
* tests/Makefile.am (pc_LDADD): Adde $(dl_LIBS).
* tests/pc.c: Include "config.h" and <dlfcn.h>.
(main): Use dladdr to find the address to unmap.
As an exception to the general rules that syscall names should match
__NR_* constants defined in kernel headers, show select as "select"
and old select as "oldselect".
* linux/arm/syscallent.h (82): Change name to oldselect.
* linux/hppa/syscallent.h (142): Change name to select.
* linux/mips/syscallent-n32.h (6022): Likewise.
* linux/mips/syscallent-n64.h (5022): Likewise.
* linux/mips/syscallent-o32.h (4142): Likewise.
* linux/sparc/syscallent.h (230): Likewise.
* linux/s390/syscallent.h (1, 180, 181): Change syscall names to match
__NR_* constants defined in arch/s390/include/uapi/asm/unistd.h.
* linux/s390x/syscallent.h: Likewise.
When there were fewer xlat headers, this didn't matter as much.
But now with >200, trying to run them all in parallel can easily
fork bomb smaller systems.
* xlat/gen.sh (main): Do not more jobs than # of cpus.
* linux/hppa/syscallent.h (18, 28, 59, 84, 108, 109, 280):
Change syscall names to match __NR_* constants defined in
arch/parisc/include/uapi/asm/unistd.h.
The parisc kernel has some legacy baggage here and decodes the 64bit
field in the reverse order. Handle it in strace too.
* fanotify.c [HPPA] (sys_fanotify_mark): Reverse the mask bits.
Looks like the similarity of these names caused confusion when generating
the syscall table. getresgid was wired to the wrong output while the other
two were wired up to printargs instead of the real printer.
* linux/ia64/syscallent.h: Connect getresuid to sys_getresuid, getresgid to
sys_getresgid, and setresgid to sys_setresgid.
The ia64 port pulls in the i386 syscall table so it can decode
32bit apps, so we need to enable the 16bit uid parsers for it.
* defs.h (NEED_UID16_PARSERS): Define to 1 for IA64.
* linux/ia64/syscallent.h: Undefine sys_stime to avoid redefine warnings.
There is no guarantee that the backing value for the various time fields
are ints or longs which makes printfs annoying. Lets cast them to the
uintmax_t type so we can use the j flag and not worry about truncation.
* v4l2.c (v4l2_ioctl): Use j and uintmax_t when printing tv_sec/tv_usec.
Assume that the kernel is v2.5.46 or newer, i.e. PTRACE_SETOPTIONS
and PTRACE_O_TRACESYSGOOD|PTRACE_O_TRACEEXEC|PTRACE_O_TRACECLONE
are universally available.
This change removes all code that implemented post-execve SIGTRAP
handling and fork/vfork/clone->CLONE_PTRACE substitution.
* defs.h (TCB_BPTSET, TCB_WAITEXECVE): Remove macros.
(need_fork_exec_workarounds, setbpt, clearbpt): Remove declarations.
* strace.c (need_fork_exec_workarounds,
test_ptrace_setoptions_followfork, test_ptrace_setoptions_for_all):
Remove.
(syscall_trap_sig): Set to (SIGTRAP | 0x80).
(ptrace_setoptions): Set to (PTRACE_O_TRACESYSGOOD | PTRACE_O_TRACEEXEC).
(detach): Do not test for TCB_BPTSET.
(init): Do not call test_ptrace_setoptions_followfork and
test_ptrace_setoptions_for_all. Do not test for TCB_BPTSET.
* syscall.c (syscall_fixup_on_sysenter, internal_fork, internal_exec,
syscall_fixup_for_fork_exec, syscall_fixup_on_sysexit): Remove.
(trace_syscall_entering): Do not test for TCB_WAITEXECVE. Do not call
syscall_fixup_on_sysenter and syscall_fixup_for_fork_exec.
(trace_syscall_exiting): Do not call syscall_fixup_on_sysexit and
syscall_fixup_for_fork_exec.
[IA64] (ia64_ia32mode): Make static.
* linux/ia64/arch_regs.h (ia64_ia32mode): Remove declaration.
* util.c: Do not include "syscall.h".
(arg_setup, get_arg0, get_arg1, set_arg0, set_arg1, restore_arg0,
restore_arg1, arg_finish_change, change_syscall, setbpt, clearbpt):
Remove.
* tests/ptrace_setoptions.test: Remove.
* tests/Makefile.am (TESTS): Remove it.
* tests/pc.c (main): Unmap larger memory region.
Dump process map for debug purposes.
* tests/pc.test: Update regexps to take CLD_DUMPED child process
into account.
* tests/getdents.out: New file.
* tests/Makefile.am (EXTRA_DIST): Add it.
* tests/getdents.test: Check whether awk is capable of matching long
regular expressions.
On some platforms printing of __s64 types with "%lld" format,
or __u64 types with "%llu" or "%llx" formats causes a warning
that __s64/__u64 type is not compatible with long long format.
From my PoV, this is nonsense, but the workaround costs nothing.
* scsi.c (print_sg_io_v4_req, print_sg_io_v4_res): Cast __u64 types
to (unsigned long long).
* v4l2.c (v4l2_ioctl): Likewise. Cast __s64 type to (long long).
On x86_64/x32, sys_sigreturn is defined only for i386 personality.
* signal.c (sys_sigreturn): Define for X32.
(sys_sigreturn) [X86_64 || X32]: Implement for i386 personality only.
* configure.ac: Check whether struct input_absinfo.resolution is defined
in <linux/input.h>.
* evdev.c (SYN_MAX): Add fallback definition.
(abs_ioctl): Wrap use of struct input_absinfo.resolution in #ifdef check.
(keycode_V2_ioctl, mtslots_ioctl, repeat_ioctl): Wrap in #ifdef checks.
(evdev_read_ioctl): Wrap in #ifdef checks the code that uses EV_SW,
EVIOCGREP, EVIOCGKEYCODE_V2, EVIOCGMTSLOTS, EVIOCGPROP, and EVIOCGSW.
(evdev_write_ioctl): Wrap in #ifdef checks the code that uses EVIOCSREP,
EVIOCSKEYCODE_V2, EVIOCSCLOCKID, and EVIOCREVOKE.
This patch introduces decoding for evdev ioctls. For EVIOCSFF the five
first members are decoded, the union is not. The code was compiled and
tested on x86_64 with Linux 3.13.
* evdev.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* defs.h (evdev_ioctl): New prototype.
* ioctl.c (ioctl_decode): Call evdev_ioctl for 'E' ioctl commands.
* xlat/evdev_autorepeat.in: New file.
* xlat/evdev_ff_status.in: New file.
* xlat/evdev_ff_types.in: New file.
* xlat/evdev_keycode.in: New file.
* xlat/evdev_leds.in: New file.
* xlat/evdev_misc.in: New file.
* xlat/evdev_mtslots.in: New file.
* xlat/evdev_prop.in: New file.
* xlat/evdev_relative_axes.in: New file.
* xlat/evdev_snd.in: New file.
* xlat/evdev_switch.in: New file.
* xlat/evdev_sync.in: New file.
Signed-off-by: Etienne Gemsa <etienne.gemsa@lse.epita.fr>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
* strace.c (print_debug_info, maybe_allocate_tcb, print_signalled,
print_exited): Remove "const" qualifier from status variable to
workaround some old implementations of WIFSIGNALED/WIFSTOPPED/WIFEXITED
macros.
* v4l2.c: Include <linux/types.h>.
(v4l2_ioctl): Check that V4L2_CTRL_FLAG_NEXT_CTRL and VIDIOC_S_EXT_CTRLS
macros are defined.
* tests/Makefile.am (AM_CPPFLAGS): Define.
* linux/inotify.h: Remove.
* Makefile.am (EXTRA_DIST): Remove it.
* inotify.c: Do not include <linux/inotify.h>.
* xlat/inotify_flags.in: Add default values.
Linux kernel used to guarantee that the largest errno value will not
exceed 4095, but SECCOMP_RET_ERRNO support introduced by kernel commit
v3.4-rc2-13-gacf3b2c71ed20c53dc69826683417703c2a88059 inadvertently
broke it. The guarantee is back with kernel commit
v3.19-8275-g580c57f1076872ebc2427f898b927944ce170f2d.
* syscall.c (is_negated_errno): Set maximum errno value to 4095.
Considering we're requiring linux-2.6 era kernels, the discussion of
header troubles that plagued 2.2 and 2.4 are no longer relevant.
* README-linux: Delete.
* configure.ac (AC_CHECK_HEADERS): Add linux/securebits.h.
* prctl.c (sys_prctl): Implement full decoding of all PR_* constants
as defined in linux v3.19.
(sys_arch_prctl): Fix decoding of unknown commands.
* xlat/cap.in: New file.
* xlat/pr_mce_kill.in: New file.
* xlat/pr_mce_kill_policy.in: New file.
* xlat/pr_set_mm.in: New file.
* xlat/pr_tsc.in: New file.
* xlat/pr_unalign_flags.in: New file.
* xlat/secbits.in: New file.
* xlat/prctl_options.in: Remove non-linux constants, add fallback
definitions.
* xlat/seccomp_mode.in: Add fallback definitions.
* xlat/gen.sh (gen_header): If the name of constant is specified with a
value, generate a default definition for the constant using this value.
* xlat/timerfdflags.in: Use extended syntax.
* Makefile.am (strace_SOURCES): Add caps0.h and caps1.h.
* caps0.h: New file.
* caps1.h: New file.
* capability.c: Remove CAP_* definitions, include "caps0.h"
and "caps1.h" instead.
Include "xlat/cap_mask0.h" instead of "xlat/capabilities.h".
Include "xlat/cap_mask1.h" instead of "xlat/capabilities1.h".
(print_cap_bits): Update callers.
* xlat/capabilities.in: Rename to xlat/cap_mask0.in.
* xlat/capabilities1.in: Rename to xlat/cap_mask1.in.
* syscall.c (print_pc): Choose instruction pointer format depending
on current_wordsize, not the size of long integer type.
* tests/pc.c: New file.
* tests/pc.test: New test.
* tests/Makefile.am (check_PROGRAMS): Add pc.
(TESTS): Add pc.test.
* tests/.gitignore: Add pc.
* linux/mips/arch_regs.h (struct mips_regs): New structure.
(mips_REG_*): New macros.
* signal.c (sys_sigreturn) [MIPS]: Use mips_REG_SP.
* syscall.c [MIPS] (struct mips_regs): New variable.
[MIPS] (ARCH_REGS_FOR_GETREGS): New macro.
(mips_a3, mips_r2): Remove.
(print_pc) [MIPS]: Use mips_REG_EPC.
(get_scno) [MIPS]: Use mips_REG_V0 and mips_REG_A3.
(get_syscall_args) [MIPS]: Use mips_REG_A[0-5].
(get_syscall_result) [MIPS]: Remove.
(get_error) [MIPS]: Use mips_REG_A3 and mips_REG_V0.
Some gcc versions complain about assigning long* to uint32_t* on i386.
* syscall.c [I386] (i386_esp_ptr): Change type from uint32_t* to long*.
* linux/i386/arch_regs.h (i386_esp_ptr): Likewise.
* linux/x86_64/arch_regs.h: Do not include "i386/arch_regs.h".
(i386_esp_ptr): New prototype.
As get_regs() is now guaranteed to set get_regs_error,
there is no need to reset get_regs_error before get_regs() call.
* strace.c (trace): Call clear_regs() iff !WIFSTOPPED.
As get_regs() is now guaranteed to fetch registers when either
ARCH_REGS_FOR_GETREGSET or ARCH_REGS_FOR_GETREGS is defined, we can
safely replace all checks for architectures where these macros are defined
with a single check for these macros.
* syscall.c (get_syscall_result): Check for [ARCH_REGS_FOR_GETREGSET ||
ARCH_REGS_FOR_GETREGS]. Remove all checks for architectures where
nothing has to be done after get_regs().
* syscall.c [AARCH64, X86_64, X32] (ARCH_REGS_FOR_GETREGSET,
ARCH_IOVEC_FOR_GETREGSET): New macros.
(get_regset): Define iff ARCH_REGS_FOR_GETREGSET is defined; change
return type to long, return ptrace() return code instead of assigning it
to get_regs_error; do not list individual architectures, implement
constant and variable iovec cases depending on ARCH_IOVEC_FOR_GETREGSET.
(get_regs): Assign get_regset() return code to get_regs_error.
PTRACE_GETREGSET and PTRACE_GETREGS methods are equally good on arm and
i386 architectures, but PTRACE_GETREGSET with fallback to PTRACE_GETREGS
is a bit more costly, so choose the method that costs less.
This partially reverts commit v4.7-149-gfaa177e.
* syscall.c [ARM, I386] (ARCH_REGS_FOR_GETREGSET): Remove.
(get_regset) [ARM || I386]: Remove.
(get_regs) [ARM || I386]: Use PTRACE_GETREGS unconditionally.
* regs.h: New file.
* Makefile.am (strace_SOURCES): Add it.
* linux/arch_regs.h: New file.
* linux/alpha/arch_regs.h: Likewise.
* linux/arm/arch_regs.h: Likewise.
* linux/hppa/arch_regs.h: Likewise.
* linux/i386/arch_regs.h: Likewise.
* linux/ia64/arch_regs.h: Likewise.
* linux/mips/arch_regs.h: Likewise.
* linux/powerpc/arch_regs.h: Likewise.
* linux/sh64/arch_regs.h: Likewise.
* linux/sparc/arch_regs.h: Likewise.
* linux/sparc64/arch_regs.h: Likewise.
* linux/tile/arch_regs.h: Likewise.
* linux/x86_64/arch_regs.h: Likewise.
* Makefile.am (EXTRA_DIST): Add them.
* defs.h: Remove definitions of PTRACE_*, REG_*, PT_*, U_REG_*,
and ARM_* macros.
Remove declarations of struct pt_regs variables.
[SPARC]: Do not include <asm/psr.h>.
[SPARC64]: Do not include <asm/psrcompat.h>.
* bjm.c: Do not include <sys/user.h>.
* process.c: Include "regs.h".
Do not include <sys/user.h> and <sys/reg.h>.
[IA64]: Do not include <asm/ptrace_offsets.h> and <asm/rse.h>.
* util.c: Likewise.
* syscall.c: Likewise.
[AARCH64]: Define struct arm_pt_regs and ARM_* macros.
[SPARC]: Include <asm/psr.h>.
[SPARC64]: Include <asm/psrcompat.h>.
* signal.c: Include "regs.h".
Do not include <sys/user.h> and <sys/reg.h>.
[IA64]: Do not include <asm/ptrace_offsets.h>.
* strace.c [IA64]: Do not include <asm/ptrace_offsets.h>.
* xlat/ptrace_events.in: New file.
* wait.c: Include "xlat/ptrace_events.h".
(printstatus): In case of WIFSTOPPED, print 0x80 flag separately from
the stop signal name.
[WIFCONTINUED]: Add WIFCONTINUED support.
Decode PTRACE_EVENT_* events.
The Linux kernel supports two different versions of the SG_IO API,
namely v3 and v4. This patch adds support for version 4 of this API.
At least the sg3_utils package supports version 4 of this API. Version
4 of this API is used if /dev/bsg/H:C:I:L is used as device name.
This patch has been tested by inspecting the output of the following
commands:
modprobe scsi_debug dev_size_mb=16 delay=0
dev=$(lsscsi | sed -n '/ scsi_debug /s,^[^/]*,,p')
./strace -eioctl -s256 sg_inq $dev
echo 'Test XDWRITEREAD(10)'
bsg=/dev/bsg/$(lsscsi | sed -n '/ scsi_debug /s,^\[\([^]]\+\)\].*,\1,p')
dd if=/dev/zero bs=512 count=1 |
./strace -eioctl -s256 sg_raw -s 512 -r 1024 $bsg 53 00 00 00 00 01 00 00 01 00
* configure.ac (AC_CHECK_HEADERS): Add linux/bsg.h.
* xlat/bsg_protocol.in: New file.
* xlat/bsg_subprotocol.in: Likewise.
* scsi.c: Include them and <linux/bsg.h>.
(print_sg_io_v3_req, print_sg_io_v3_res, print_sg_io_v4_req,
print_sg_io_v4_res): New functions.
(print_sg_io_req, print_sg_io_res): Use them.
(scsi_ioctl): Update callers.
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
* scsi.c (print_sg_io_req, print_sg_io_res): Add support for decoding
iovecs passed via the SG I/O v3 interface. Decode the "data in" and
"data out" buffers in hex instead of in ASCII.
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
* scsi.c (print_sg_io_res): Only print that part of the data in buffer
that has been initialized by the SCSI LLD driver.
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
Pass pointers to tracee's memory as integers instead of pointers.
* scsi.c (print_sg_io_buffer): Change second argument's type to
unsigned long.
(print_sg_io_req, print_sg_io_res): Update callers.
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
This fixes the bug introduced by commit v4.9-250-ga0a4909.
* net.c (print_setsockopt): Separate SOL_IP case from SOL_PACKET.
Reported-by: Ben Noordhuis <info@bnoordhuis.nl>
If PTRACE_O_TRACEEXEC option is enabled, then post-execve SIGTRAPs are
suppressed and we can safely skip -ENOSYS heuristics. This partially
reverts commits v4.9-258-g8e398b6 and v4.9-261-gd70c41d.
* syscall.c (syscall_fixup_on_sysenter): Skip if PTRACE_O_TRACEEXEC
flag is set in ptrace_setoptions. Remove all SECCOMP_RET_ERRNO
workarounds as they are not needed when PTRACE_O_TRACEEXEC is not
supported by the kernel.
The process stared by strace can exit/terminate only once. When this
happens, forget its pid, so that any other event or action related to a
process with the same pid as the pid of that first process would not be
confused with that first process.
* strace.c (print_signalled, print_exited): Clear strace_child when it
equals to the given pid.
* strace.c (maybe_allocate_tcb): If detach_on_execve is set, do not warn
about !WIFSTOPPED process with pid == strace_child.
* tests/bexecve.test: New test.
* tests/Makefile.am (TESTS): Add it.
This change moves the main loop back to main() and splits trace()
into several functions. There are no algorithmic changes.
* strace.c (print_debug_info, maybe_allocate_tcb, maybe_switch_tcbs,
print_signalled, print_exited, print_stopped, startup_tcb): New
functions.
(trace) Use them. Move the main loop ...
(main): ... here.
* signal.c (printsiginfo): Make the first argument a pointer to a const type.
* defs.h (printsiginfo): Likewise.
(printsiginfo, printsiginfo_at): Always define these prototypes.
* configure.ac (AC_CHECK_TYPES): Remove siginfo_t, as HAVE_SIGINFO_T
is no longer used.
Thanks to SECCOMP_RET_DATA == 0xffff, abnormally large errno
values could be easily seen when a seccomp filter is used, e.g.
BPF_STMT(BPF_RET, SECCOMP_RET_ERRNO | SECCOMP_RET_DATA)
* syscall.c (is_negated_errno): Allow errno values as large as 0xffff.
This is required to process SECCOMP_RET_ERRNO filtered syscalls on
systems where linux kernel does not change the syscall number to -1.
* syscall.c (is_negated_errno): Move before syscall_fixup_on_sysenter.
(syscall_fixup_on_sysenter) [I386 || X32 || X86_64] : Use it to check
AX register.
Use the same function name on all architectures.
* syscall.c [X32] (is_negated_errno_x32): Rename to is_negated_errno.
(get_error) [X32]: Update callers.
Syscall number -1 might be a side effect of SECCOMP_RET_ERRNO filtering.
* syscall.c (syscall_fixup_on_sysenter) [I386 || X32 || X86_64]:
Do not skip syscalls that have number -1.
This allows using both -e read=FD and -e write=FD at the same time
with the same FD.
Without this patch, the -e write=FD is ignored if -e read=FD is used
at the same time.
Signed-off-by: Aurelien Jacobs <aurel@gnuage.org>
The AC_SIN6_SCOPE_ID autoconf macro that was able to define
HAVE_SIN6_SCOPE_ID_LINUX was removed by commit v4.4-98-g64f793a.
* net.c [HAVE_SIN6_SCOPE_ID_LINUX]: Remove.
Ensure that all known SOL_* constants are defined.
Sync the list of known SOL_* constants with linux/socket.h.
* net.c: Provide definitions for all SOL_* constants.
Remove all SOL_* preprocessor conditionals.
(socketlayers): Disconnect from inet_protocols.
* xlat/socketlayers.in: Remove SOL_ICMP, add SOL_ALG and SOL_NFC.
Move the code that is common between getsockopt and setsockopt parsers
to a separate function.
Move printers specific to getsockopt and setsockopt to separate
functions, choose appropriate parsers depending on verbose mode.
Move parsers of specific socket options to separate functions.
* net.c (printicmpfilter): Rename to print_icmp_filter.
(printsockopt): Remove.
(print_sockopt_fd_level_name, print_linger, print_ucred,
print_tpacket_stats, print_getsockopt, print_tpacket_req,
print_packet_mreq, print_setsockopt): New functions.
(sys_getsockopt, sys_setsockopt): Use them.
* block.c (print_blkpg_req): Print blkpg_partition.devname and
blkpg_partition.volname using print_quoted_string.
(block_ioctl): Print blk_user_trace_setup.name using print_quoted_string.
Change printfd() behaviour in show_fd_path mode to quote pathnames
the same way printpathn() does.
* util.c (printfd): Print non-socket paths using print_quoted_string.
* tests/scm_rights-fd.test: Test it.
* tests/scm_rights.c: Update.
string_quote() has proven to be too hard to use, replace it with
print_quoted_string() that does memory allocation and prints the result.
* defs.h (string_quote): Remove.
(QUOTE_0_TERMINATED, QUOTE_OMIT_LEADING_TRAILING_QUOTES): New macros.
(print_quoted_string): New prototype.
* util.c (string_quote): Make static; take "style" flags instead
of "len", treat QUOTE_0_TERMINATED style flag as former (len == -1);
add QUOTE_OMIT_LEADING_TRAILING_QUOTES style flag support.
(ALLOCA_CUTOFF, use_alloca): New macros.
(print_quoted_string): New function.
(printpathn, printstr): Update to new API.
* loop.c (loop_ioctl): Likewise.
* mtd.c (ubi_ioctl): Likewise.
* net.c (print_scm_security): Likewise.
* socketutils.c (unix_parse_response): Likewise.
This fixes an obvious typo which prints "lock" instead of "getpeername"
for the getpeername syscall.
* linux/hppa/syscallent.h (getpeername): Change sys_name to "getpeername".
Print all cmsghdr structures in msg_control array,
not just the first one.
Change output format to be consistent with other parts of strace
where structures and arrays are printed.
Implement decoder for SCM_SECURITY message type.
* net.c (print_scm_rights, print_scm_creds, print_scm_security):
New functions.
(printcmsghdr): Use them. Iterate over all members of the array.
* xlat/scmvals.in: Add SCM_SECURITY.
* tests/scm_rights.c (main): Pass one more descriptor to the receiver.
Set SO_PASSCRED on the receiver part. Reserve enough space to receive
SCM_CREDENTIALS and SCM_RIGHTS.
* tests/scm_rights-fd.test: Update.
When strace supports more than one personality, run test suite
for each personality supported by the host system.
* bootstrap: Create tests-m32 and tests-mx32 subtrees from tests.
* configure.ac: Check for -m32 and -mx32 runtime support.
(AC_CONFIG_FILES): Add tests-m32/Makefile and tests-mx32/Makefile.
* Makefile.am [HAVE_M32_RUNTIME] (TESTS_M32): Define.
[HAVE_MX32_RUNTIME] (TESTS_MX32): Define.
(SUBDIRS): Add $(TESTS_M32) $(TESTS_MX32).
* .gitignore: Add /tests-m32 and /tests-mx32.
* net.c (union_cmsghdr): New union.
[SUPPORTED_PERSONALITIES > 1 && SIZEOF_LONG > 4] (cmsghdr32): New structure.
(printcmsghdr): Properly handle the case of current_wordsize < sizeof(long).
Historically, only 16 bits (8-bit number and 8-bit type) of 32-bit ioctl
commands were used for decoding, which was the source for numerous
annoying collisions like this:
ioctl(0, SNDCTL_TMR_TIMEBASE or SNDRV_TIMER_IOCTL_NEXT_DEVICE or TCGETS, {B38400 opost isig icanon echo ...}) = 0
ioctl(0, MGSL_IOCGPARAMS or MMTIMER_GETRES or MTIOCTOP or SNDCTL_MIDI_MPUMODE, 0x7fffd47f7338) = -1 ENOTTY (Inappropriate ioctl for device)
The solution is to use all 32 bits for decoding, not just "number" and
"type", but also "size" and "direction".
As some architectures override defaults that come from asm-generic/ and
provide alternative definitions for some ioctl commands, we support
per-architecture ioctl definitions and merge them with common
definitions at build time. During the merge, we used to keep both
generic and architecture-specific definitions, now architecture-specific
definitions have precedence over generic ones -- ioctlsort omits
definitions from asm-generic/ for those ioctl names that have different
definitions in asm/.
Additional bits of "direction" are architecture specific -- the number
of bits and their values differ between architectures. To reduce
architecture differences in the source code, we keep "direction" in
symbolic form and compile it in ioctlsort.
Additional bits of "size" are also architecture specific -- not only the
number of bits differ between architectures, but sizes of many types
depend on sizeof(long). To reduce architecture differences in the
source code, we keep 32-bit and 64-bit versions of common ioctl
definitions, and use the appropriate version for each architecture and
personality.
To implement this, the tools for generating ioctl definitions from
kernel headers have been rewritten, and the source format of ioctl
definitions has been extended. The final ioctlent*.h files that are
included by syscall.c are now generated from source ioctls_inc*.h and
ioctls_arch*.h files at build time with ioctlsort.
* ioctl.c (ioctl_lookup): Use all 32 bits of ioctl command code.
* ioctlsort.c: Rewritten.
* linux/32/ioctls_inc.h: New file.
* linux/64/ioctls_inc.h: New file.
* linux/aarch64/ioctls_arch0.h: New file.
* linux/aarch64/ioctls_arch1.h: New file.
* linux/aarch64/ioctls_inc0.h: New file.
* linux/aarch64/ioctls_inc1.h: New file.
* linux/alpha/ioctls_arch0.h: New file.
* linux/alpha/ioctls_inc0.h: New file.
* linux/arc/ioctls_arch0.h: New file.
* linux/arc/ioctls_inc0.h: New file.
* linux/arm/ioctls_arch0.h: New file.
* linux/arm/ioctls_inc0.h: New file.
* linux/avr32/ioctls_arch0.h: New file.
* linux/avr32/ioctls_inc0.h: New file.
* linux/bfin/ioctls_arch0.h: New file.
* linux/bfin/ioctls_inc0.h: New file.
* linux/hppa/ioctls_arch0.h: New file.
* linux/hppa/ioctls_inc0.h: New file.
* linux/i386/ioctls_arch0.h: New file.
* linux/i386/ioctls_inc0.h: New file.
* linux/ia64/ioctls_arch0.h: New file.
* linux/ia64/ioctls_inc0.h: New file.
* linux/m68k/ioctls_arch0.h: New file.
* linux/m68k/ioctls_inc0.h: New file.
* linux/metag/ioctls_arch0.h: New file.
* linux/metag/ioctls_inc0.h: New file.
* linux/microblaze/ioctls_arch0.h: New file.
* linux/microblaze/ioctls_inc0.h: New file.
* linux/mips/ioctls_arch0.h: New file.
* linux/mips/ioctls_inc0.h: New file.
* linux/or1k/ioctls_arch0.h: New file.
* linux/or1k/ioctls_inc0.h: New file.
* linux/powerpc/ioctls_arch0.h: New file.
* linux/powerpc/ioctls_inc0.h: New file.
* linux/powerpc64/ioctls_arch0.h: New file.
* linux/powerpc64/ioctls_arch1.h: New file.
* linux/powerpc64/ioctls_inc0.h: New file.
* linux/powerpc64/ioctls_inc1.h: New file.
* linux/s390/ioctls_arch0.h: New file.
* linux/s390/ioctls_inc0.h: New file.
* linux/s390x/ioctls_arch0.h: New file.
* linux/s390x/ioctls_inc0.h: New file.
* linux/sh/ioctls_arch0.h: New file.
* linux/sh/ioctls_inc0.h: New file.
* linux/sh64/ioctls_arch0.h: New file.
* linux/sh64/ioctls_inc0.h: New file.
* linux/sparc/ioctls_arch0.h: New file.
* linux/sparc/ioctls_inc0.h: New file.
* linux/sparc64/ioctls_arch0.h: New file.
* linux/sparc64/ioctls_arch2.h: New file.
* linux/sparc64/ioctls_inc0.h: New file.
* linux/sparc64/ioctls_inc2.h: New file.
* linux/tile/ioctls_arch0.h: New file.
* linux/tile/ioctls_arch1.h: New file.
* linux/tile/ioctls_inc0.h: New file.
* linux/tile/ioctls_inc1.h: New file.
* linux/x32/ioctls_arch0.h: New file.
* linux/x32/ioctls_arch1.h: New file.
* linux/x32/ioctls_inc0.h: New file.
* linux/x32/ioctls_inc1.h: New file.
* linux/x86_64/ioctls_arch0.h: New file.
* linux/x86_64/ioctls_arch1.h: New file.
* linux/x86_64/ioctls_inc0.h: New file.
* linux/x86_64/ioctls_inc1.h: New file.
* linux/xtensa/ioctls_arch0.h: New file.
* linux/xtensa/ioctls_inc0.h: New file.
* linux/aarch64/ioctlent.h.in: Remove.
* linux/aarch64/ioctlent1.h: Remove.
* linux/alpha/ioctlent.h.in: Remove.
* linux/arc/ioctlent.h.in: Remove.
* linux/arm/ioctlent.h.in: Remove.
* linux/avr32/ioctlent.h.in: Remove.
* linux/bfin/ioctlent.h.in: Remove.
* linux/hppa/ioctlent.h.in: Remove.
* linux/i386/ioctlent.h.in: Remove.
* linux/ia64/ioctlent.h.in: Remove.
* linux/ioctlent.h.in: Remove.
* linux/ioctlent.sh: Remove.
* linux/m68k/ioctlent.h.in: Remove.
* linux/metag/ioctlent.h.in: Remove.
* linux/microblaze/ioctlent.h.in: Remove.
* linux/mips/ioctlent.h.in: Remove.
* linux/mips/ioctlent.sh: Remove.
* linux/or1k/ioctlent.h.in: Remove.
* linux/powerpc/ioctlent.h.in: Remove.
* linux/powerpc64/ioctlent.h: Remove.
* linux/powerpc64/ioctlent1.h: Remove.
* linux/s390/ioctlent.h.in: Remove.
* linux/s390x/ioctlent.h.in: Remove.
* linux/sh/ioctlent.h.in: Remove.
* linux/sh64/ioctlent.h.in: Remove.
* linux/sparc/ioctlent.h.in: Remove.
* linux/sparc64/ioctlent.h.in: Remove.
* linux/sparc64/ioctlent2.h: Remove.
* linux/tile/ioctlent.h.in: Remove.
* linux/tile/ioctlent1.h: Remove.
* linux/x32/ioctlent.h.in: Remove.
* linux/x32/ioctlent1.h: Remove.
* linux/x86_64/ioctlent.h.in: Remove.
* linux/x86_64/ioctlent1.h: Remove.
* linux/xtensa/ioctlent.h.in: Remove.
* linux/x86_64/ioctlent2.h: Include ioctlent0.h instead of ioctlent.h.
* syscall.c (struct_ioctlent ioctlent0): Likewise.
* Makefile.am: Remove all ioctlent-related definitions.
Define the list of ioctlent*.h files that have to be generated by
presence of $(srcdir)/$(OS)/$(ARCH)/ioctls_inc*.h files.
Add rules for ioctlent*.h files generation.
(EXTRA_DIST): Update.
* maint/ioctls_gen.sh: New file.
* maint/ioctls_hex.sh: New file.
* maint/ioctls_sym.sh: New file.
* maint/print_ioctlent.c: New file.
* HACKING-scripts: Update for ioctlent.sh -> ioctls_gen.sh migration.
* .gitignore: Add ioctlent[012].h and ioctls_all[012].h.
* configure.ac (AC_CHECK_HEADERS): Add linux/hiddev.h
and linux/mmtimer.h for tests.
* tests/ioctl.c: New file.
* tests/ioctl.test: New test.
* tests/Makefile.am (check_PROGRAMS): Add ioctl.
(TESTS): Add ioctl.test.
* tests/.gitignore: Add ioctl.
Some ioctl commands from evdev, hiddev, mixer, uinput, spi, and joystick
are parametrized by a size or a number that are variable. This change
implements printing of these parametrized ioctl command names.
* configure.ac (AC_CHECK_HEADERS): Add linux/input.h.
* defs.h (ioctl_decode_number): New prototype.
* io.c (sys_ioctl): Use ioctl_decode_number.
* ioctl.c (ioctl_decode_number): New function.
* xlat/evdev_abs.in: New file.
* xlat/evdev_ev.in: New file.
Signed-off-by: Gabriel Laskar <gabriel@lse.epita.fr>
Most architectures supported by strace have their own subdirectory in
linux directory, but powerpc64 didn't, causing inconveniences.
* configure.ac (arch): Define to "powerpc64" on powerpc64.
* linux/powerpc/errnoent1.h: Move to linux/powerpc64/.
* linux/powerpc/ioctlent1.h: Likewise.
* linux/powerpc/signalent1.h: Likewise.
* linux/powerpc/syscallent1.h: Likewise.
* linux/powerpc64/ioctlent.h: New file.
* linux/powerpc64/syscallent.h: Likewise.
* linux/powerpc64/userent.h: Likewise.
* Makefile.am (EXTRA_DIST): Remove linux/powerpc/ioctlent1.h,
linux/powerpc/syscallent1.h, linux/powerpc/errnoent1.h, and
linux/powerpc/signalent1.h; added linux/powerpc64/errnoent1.h,
linux/powerpc64/ioctlent.h, linux/powerpc64/ioctlent1.h,
linux/powerpc64/signalent1.h, linux/powerpc64/syscallent.h,
linux/powerpc64/syscallent1.h, and linux/powerpc64/userent.h.
* defs.h (ioctl_print_code): New prototype.
* ioctl.c: Include xlat/ioctl_dirs.h.
(ioctl_print_code): New function.
* io.c (sys_ioctl): Use it.
* xlat/ioctl_dirs.in: New file.
ioctlent.sh looks for ioctl command names using grep, and fetches them
with their definitions using sed. Historically, lookup regexps used in
grep and sed are slightly different, which makes the script more
complicated than necessary.
* tests/ipc_msg.c: Include <errno.h>.
(main): Add a test call with IPC_STAT, handle old kernels MSG_STAT behaviour.
* tests/ipc_sem.c: Include <errno.h>.
(main): Add a test call with IPC_STAT, handle old kernels SEM_STAT behaviour.
* tests/ipc_shm.c: Include <errno.h>.
(main): Add a test call with IPC_STAT, handle old kernels SHM_STAT behaviour.
* defs.h (umoven, umovestr): Change type of 3rd argument from "int"
to "unsigned int".
* util.c (umoven, umovestr): Likewise. Adjust the code appropriately.
* tests/inet-accept-connect-send-recv.c (main): Return 77 if inet stream
socket cannot be created or it cannot be bound to INADDR_LOOPBACK.
* tests/net-yy.test: Skip the test if inet-accept-connect-send-recv
returns 77.
Starting with commit v4.9-27-g3e252ca, header file names should not be
part of ioctlent files included by syscall.c
* linux/sparc/ioctlent1.h: Remove ioctl header file names.
* linux/sparc64/ioctlent2.h: Remove all entries, include "ioctlent.h".
According to linux/arch/s390/include/uapi/asm/unistd.h, readdir is
defined both on s390 and s390x, while getrlimit, _llseek, sendfile64,
and fadvise64_64 are defined on s390 only.
* linux/s390x/syscallent.h (readdir): Add.
(getrlimit, _llseek, sendfile64, fadvise64_64): Remove.
Since arc/syscallent.h, metag/syscallent.h, or1k/syscallent.h, and
tile/syscallent1.h share a lot of identical syscall entries described in
<asm-generic/unistd.h>, move these entries to a new file 32/syscallent.h
using the following automated conversion with subsequent correction of
whitespace indentation:
sed -e 's|^[[:space:]]*\({[^}]\+},\)[[:space:]]*/\*[[:space:]]\([[:space:]]*[0-9]\+\)[[:space:]]*\*/|[\2] = \1|' -e '/^[[:space:]]*\[[[:space:]]*2\(4[4-9]\|5[0-9]\)/d' linux/arc/syscallent.h
* linux/32/syscallent.h: New file.
* Makefile.am (EXTRA_DIST): Add it.
* linux/arc/syscallent.h: Include "32/syscallent.h",
remove generic 32-bit syscall entries, use designated initializers.
* linux/metag/syscallent.h: Likewise.
* linux/or1k/syscallent.h: Likewise.
* linux/tile/syscallent1.h: Likewise.
* linux/s390/syscallent.h (fadvise64, fadvise64_64): Change sys_func to
printargs until special parsers for 31-bit 31 s390 emulation wrappers
are written.
Since aarch64/syscallent1.h and tile/syscallent.h share a lot of
identical syscall entries described in <asm-generic/unistd.h>, move
these entries to a new file 64/syscallent.h using the following
automated conversion with subsequent correction of whitespace
indentation:
sed -e 's|^[[:space:]]*\({[^}]\+},\)[[:space:]]*/\*[[:space:]]\([[:space:]]*[0-9]\+\)[[:space:]]*\*/|[\2] = \1|' -e '/^\[[[:space:]]*2\(4[4-9]\|5[0-9]\)\]/d' linux/tile/syscallent.h
* linux/64/syscallent.h: New file.
* Makefile.am (EXTRA_DIST): Add it.
* linux/aarch64/syscallent1.h: Include "64/syscallent.h",
remove generic 64-bit syscall entries, use designated initializers.
* linux/tile/syscallent.h: Likewise.
Reserve the same space for new syscalls on EABI as it is already
reserved on OABI.
* linux/arm/syscallent.h [__ARM_EABI__] (ARM_FIRST_SHUFFLED_SYSCALL):
Raise to 400.
indirect_ipccall used to guess whether currently decoded ipc syscall is
an indirect subcall or not based on generic knowledge, while there is a
robust method to identify indirect ipc subcalls.
* defs.h (TRACE_INDIRECT_SUBCALL): New flag.
* ipc.c (indirect_ipccall): Check it.
* linux/subcall.h: Set it.
Syscall entries for socket and ipc subcalls are the same on all
architectures, or at least they should be the same, so move them
to the single file.
* linux/subcall.h: Add syscall entries for socket and ipc subcalls.
* linux/arm/syscallent.h: Remove entries for socket and ipc subcalls.
* linux/i386/syscallent.h: Likewise.
* linux/m68k/syscallent.h: Likewise.
* linux/powerpc/syscallent.h: Likewise.
* linux/s390/syscallent.h: Likewise.
* linux/s390x/syscallent.h: Likewise.
* linux/sh/syscallent.h: Likewise.
* linux/sh64/syscallent.h: Likewise.
* linux/sparc/syscallent.h: Likewise.
The fanotify_mark syscall takes a 64-bit mask, and on 32-bit
architectures it is split up into two syscall arguments.
* configure.ac (AC_CHECK_FUNCS): Add fanotify_mark.
(AC_CHECK_HEADERS): Add sys/fanotify.h.
* defs.h (getllval): New prototype.
* util.c (getllval): New function based on printllval.
(printllval): Use getllval.
* fanotify.c (sys_fanotify_mark): Use getllval to properly decode
64-bit mask and two syscall arguments followed by it.
* tests/fanotify_mark.c: New file.
* tests/fanotify_mark.test: New test.
* tests/Makefile.am (check_PROGRAMS): Add fanotify_mark.
(TESTS): Add fanotify_mark.test.
* tests/.gitignore: Add fanotify_mark.
* v4l2.c (print_v4l2_format_fmt): Do not use
V4L2_BUF_TYPE_VIDEO_OUTPUT_OVERLAY unless it is defined.
(v4l2_ioctl): Do not use VIDIOC_ENUM_FRAMESIZES and
VIDIOC_ENUM_FRAMEINTERVALS unless they are defined.
The switch-case statement is a gawk-specific feature which is not
necessarily available, let's use traditional if-else statements instead.
* tests/uid.awk: Rewrite without switch-case statements.
NETLINK_INET_DIAG was introduced in linux v2.6.14,
NETLINK_SOCK_DIAG was added much later in linux v3.3.
* net.c [!NETLINK_SOCK_DIAG && NETLINK_INET_DIAG]
(NETLINK_SOCK_DIAG): Define.
* socketutils.c: Likewise.
* tests/netlink_unix_diag.c: Likewise.
Before this change, ioctlent.sh could not recognize unusual constants
defined by linux/aufs_type.h, resulting to ioctls.h with references to
these constants without appropriate definitions in ioctldefs.h.
* linux/ioctlent.sh: Update the regexp that is used to lookup constants
referenced by generated ioctls.h.
* util.c (sprinttime): Increase buffer size.
* tests/stat64-v.test: Try to set modification time of the sample file
to a value set in the distant future.
* file.c (sys_stat, sys_fstat): Collect in one place right before
the definition of printstat64.
(sys_stat64, sys_fstat64): Collect in one place right after the
definition of printstat64.
* file.c (printstat) [X32]: Define.
(printstat64) [X32]: Use "struct stat64" instead of "struct stat".
Use printstat for x32 personality processes.
(printstat64_x32): Remove.
(sys_stat64, sys_fstat64) [X32]: Use printstat64 instead of
printstat64_x32.
* file.c [X32]: Do not define "struct stat" and "struct
__old_kernel_stat" locally, use definitions provided by <asm/stat.h>.
Do not define "struct stat64" exclusively, reuse another definition
added earlier for X86_64.
* printstat.h: New file.
* Makefile.am (strace_SOURCES): Add it.
* file.c: Include "printstat.h".
(do_printstat): Move to printstat.h, parametrize its name
and struct stat.
* file.c (realprintstat): Adjust format strings so that all 64-bit
integer fields are printed with %llu format, and %u format is used for
all the rest. Rename to do_printstat.
(printstat, printoldstat): Update callers.
* tests/stat.c: New file.
* tests/stat32.c: Likewise.
* tests/stat32-v.test: New file.
* tests/stat64-v.test: Likewise.
* tests/Makefile.am (check_PROGRAMS): Add stat and stat32.
(stat_CFLAGS): Define.
(TESTS): Add stat32-v.test and stat64-v.test.
* tests/.gitignore: Add stat and stat32.
There are no symbolic constants defined for struct stat.st_flags,
so decode this rare field as unsigned int.
* xlat/fileflags.in: Remove.
* file.c (realprintstat, printstat64) [HAVE_STRUCT_STAT_ST_FLAGS]:
Decode struct stat.st_flags as unsigned int.
Remove the code that supports struct stat.st_aclcnt and stat.st_level
decoding -- these fields are not defined in system headers.
* configure.ac (AC_CHECK_MEMBERS): Remove struct stat.st_aclcnt and
struct stat.st_level.
* file.c (realprintstat, printstat64)
[HAVE_STRUCT_STAT_ST_ACLCNT || HAVE_STRUCT_STAT_ST_LEVEL]: Remove.
Change all net-accept-connect based tests to use different unix domain
socket addresses, so that these tests could be safely run in parallel.
* tests/net-accept-connect.c: Parametrize unix domain socket address.
* tests/net-fd.test: Pass socket address to net-accept-connect.
* tests/net.test: Likewise.
* tests/unix-yy.test: Likewise.
* tests/unix-yy-accept.awk: Update regexps.
* tests/unix-yy-connect.awk: Likewise.
The current code only decodes TFD_TIMER_ABSTIME when the kernel has
a bunch of newer flags. We have to handle fallbacks for all of them
since the kernel doesn't currently export things to userspace.
* time.c: Include fcntl.h for O_* defines.
(TFD_TIMER_ABSTIME): Move to xlat/timerfdflags.in.
* xlat/timerfdflags.in: Add more definitions.
Prepend -yy output generated for INET/INET6 TCP/UDP sockets with their
protocol name obtained using getxattr.
* socketutils.c (inet_parse_response): Add proto_name argument.
Print proto_name for connected and unconnected sockets.
(receive_responses): Add proto_name argument, pass it to the parser.
(inet_print): Add proto_name argument, pass it to receive_responses.
(unix_parse_response): Add proto_name argument.
(print_sockaddr_by_inode): Pass protocol name to inet_print calls.
* tests/net-yy-accept.awk: Update to match new output format.
* tests/net-yy-connect.awk: Likewise.
Unix domain socket address family name is not universally decoded as
AF_LOCAL, other names like AF_UNIX and AF_FILE are valid synonyms.
Similarly, PF_UNIX and PF_FILE are valid synonyms for PF_LOCAL.
Some test regexps have to be properly initialized before they could
be used, so make sure they are initialized before use.
Since inet_diag and unix_diag interfaces are implemented by different
kernel modules, they have to be tested separately.
* tests/netlink_unix_diag.c: New file.
* tests/unix-yy.test: Use it.
* tests/Makefile.am (check_PROGRAMS): Add it.
* tests/.gitignore: Likewise.
This change extends -yy option to handle unix domain sockets:
their peer addresses will be printed, similar to inet sockets.
For a listening socket, its socket inode and socket path are printed.
For an accepted socket, its socket inode, the peer inode, and the
socket path are printed.
For a client socket, its socket inode and the peer inode are printed.
An example of a server side communication using netcat:
$ ./strace -yy -e network nc -l -U /tmp/example.sock
socket(PF_LOCAL, SOCK_STREAM, 0) = 3
setsockopt(3<UNIX:[14728348]>, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
bind(3<UNIX:[14728348]>, {sa_family=AF_LOCAL, sun_path="/tmp/example.sock"}, 19) = 0
listen(3<UNIX:[14728348,"/tmp/example.sock"]>, 10) = 0
accept(3<UNIX:[14728348,"/tmp/example.sock"]>, {sa_family=AF_LOCAL, NULL}, [2]) = 4<UNIX:[14727246->14727245,"/tmp/example.sock"]>
recvfrom(4<UNIX:[14727246->14727245,"/tmp/example.sock"]>, "INPUT\n", 8192, 0, NULL, NULL) = 6
INPUT
An example of a client side communication using netcat:
$ ./strace -yy -e network nc -U /tmp/example.sock
socket(PF_LOCAL, SOCK_STREAM, 0) = 3
connect(3<UNIX:[14727245]>, {sa_family=AF_LOCAL, sun_path="/tmp/example.sock"}, 19) = 0
getsockopt(3<UNIX:[14727245]>, SOL_SOCKET, SO_ERROR, [0], [4]) = 0
INPUT
...
sendto(3<UNIX:[14727245->14727246]>, "INPUT\n", 6, 0, NULL, 0) = 6
* linux/unix_diag.h: New file.
* socketutils.c (send_query): Rename to inet_send_query.
(parse_response): Rename to inet_parse_response.
(unix_print, unix_send_query, unix_parse_response): New functions.
(receive_responses): Add a new argument named parser: a function for
handling protocol specific data parts of diag messages.
(print_sockaddr_by_inode): Call unix_print.
Replace NETLINK_INET_DIAG with NETLINK_SOCK_DIAG, they are equal
but NETLINK_SOCK_DIAG looks more generic.
Signed-off-by: Masatake YAMATO <yamato@redhat.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
To resolve the peer address of socket, all combinations of families
(AF_INET, AF_INET6) and protocols(IPPROTO_TCP, IPPROTO_UDP) were tried.
This change utilizes the protocol name obtained via getxattr to specify
the right combination.
* socketutils.c (inet_print): New helper function.
(print_sockaddr_by_inode): Use it. Utilize the protocol name
associated with the given inode for resolving the peer socket
address. If the protocol name is NULL, resolve the address
by trying combinations of families and protocols as before.
* defs.h (print_sockaddr_by_inode): Update prototype.
* util.c (printfd): Pass the protocol name associated with
the given path to print_sockaddr_by_inode as the 2nd argument.
Signed-off-by: Masatake YAMATO <yamato@redhat.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
Convert parsers of these syscalls to the same scheme as were applied to
parsers of other uid/gid related syscalls.
That is, define two sets of parsers on architectures that support
(either directly or via multiarch) 16-bit and 32-bit gid getgroups
and setgroups syscalls simultaneously, and reuse essentially the same
code by parametrizing uid_t and names of parser functions.
* groups.c: Remove.
(sys_getgroups, sys_setgroups): Move ...
* uid.c: ... here and parametrize their names.
* Makefile.am (strace_SOURCES): Remove groups.c.
* linux/syscall.h (sys_getgroups32, sys_setgroups32): Remove.
[NEED_UID16_PARSERS] (sys_getgroups16, sys_setgroups16): New prototypes.
* linux/arm/syscallent.h: Rename sys_[gs]etgroups to sys_[gs]etgroups16,
rename sys_[gs]etgroups32 to sys_[gs]etgroups.
* linux/bfin/syscallent.h: Likewise.
* linux/i386/syscallent.h: Likewise.
* linux/m68k/syscallent.h: Likewise.
* linux/microblaze/syscallent.h: Likewise.
* linux/s390/syscallent.h: Likewise.
* linux/sh/syscallent.h: Likewise.
* linux/sh64/syscallent.h: Likewise.
* linux/sparc/syscallent.h: Likewise.
* tests/uid.c: Test for getgroups.
* tests/uid16.c: Likewise.
* tests/uid32.c: Test for getgroups32.
* tests/uid.awk: Test for getgroups/getgroups32 decoding.
* tests/uid.test: Trace getgroups/getgroups32 syscalls.
Define two sets of parsers on architectures that support (either
directly or via multiarch) 16-bit and 32-bit uid/gid syscalls
simultaneously. Since the code in these two sets is essentially
the same and the key difference between them is the size of uid_t,
implement it by parametrizing uid_t and names of parser functions.
* defs.h (NEED_UID16_PARSERS): New macro.
* linux/syscall.h [NEED_UID16_PARSERS] (sys_chown16, sys_fchown16,
sys_getresuid16, sys_getuid16, sys_setfsuid16, sys_setresuid16,
sys_setreuid16, sys_setuid16): New prototypes.
* linux/dummy.h (sys_geteuid16): Alias to sys_getuid16.
(sys_getegid16, sys_getgid16, sys_getresgid16, sys_setfsgid16,
sys_setgid16, sys_setregid16, sys_setresgid16): Alias to corresponding
sys_*uid16 functions.
* uid.c: Stop including <asm/posix_types.h>.
Parametrize uid_t and names of all exported functions.
(get_print_uid): New function.
(sys_getresuid): Use it.
(printuid): Check for (uid_t) -1.
* uid16.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* linux/arm/syscallent.h: Use sys_chown16, sys_fchown16, sys_getegid16,
sys_geteuid16, sys_getgid16, sys_getresgid16, sys_getresuid16,
sys_getuid16, sys_setfsgid16, sys_setfsuid16, sys_setgid16,
sys_setregid16, sys_setresgid16, sys_setresuid16, sys_setreuid16,
and sys_setuid16 parsers for *chown and [gs]et*[gu]id syscall entries.
* linux/bfin/syscallent.h: Likewise.
* linux/i386/syscallent.h: Likewise.
* linux/m68k/syscallent.h: Likewise.
* linux/microblaze/syscallent.h: Likewise.
* linux/s390/syscallent.h: Likewise.
* linux/sparc/syscallent.h: Likewise.
* linux/sh/syscallent.h: Likewise.
* linux/sh64/syscallent.h: Likewise.
* tests/uid16.c: New file.
* tests/uid16.test: New test.
* tests/Makefile.am (CHECK_PROGRAMS): Add uid16.
(TESTS): Add uid16.test.
* tests/.gitignore: Add uid16.
* linux/sh/syscallent.h: Replace printargs with real syscall
parsers in *chown32 and [gs]et*[gu]id32 syscall entries.
* linux/sh64/syscallent.h: Likewise.
* linux/sparc/syscallent.h: Add "32" suffix to names
of [gs]etres[gu]id32 syscalls #108..112.
Fix entry for syscall #112 from setresgid32 to setregid32.
This will make further uid/gid fixes simpler.
* fchownat.c: New file.
* chown.c (sys_fchownat: Move to fchownat.c.
(sys_chown, sys_fchown): Move to uid.c.
* Makefile.am (strace_SOURCES): Remove chown.c, add fchownat.c.
It was not a good idea to treat uid_t as a long int type because
the latter is twice larger than uid_t on 64-bit architectures.
* defs.h (printuid): Change uid argument type from "unsigned long"
to "unsigned int".
* util.c (printuid): Likewise. When uid equals to -1, print "-1".
* tests/uid.awk: New file.
* tests/uid.c: New file.
* tests/uid32.c: Likewise.
* tests/uid.test: New test.
* tests/uid32.test: Likewise.
* tests/Makefile.am (CHECK_PROGRAMS): Add uid and uid32.
(TESTS): Add uid.test and uid32.test.
(EXTRA_DIST): Add uid.awk.
* tests/.gitignore: Add uid and uid32.
* swapon.c: Ensure that SWAP_FLAG_DISCARD_ONCE and
SWAP_FLAG_DISCARD_PAGES are defined.
* xlat/swap_flags.in: Add SWAP_FLAG_DISCARD_ONCE and
SWAP_FLAG_DISCARD_PAGES.
Since sys_prctl is referenced by syscallent files unconditionally,
conditional compilation of sys_prctl depending on prctl availability is
pointless.
* prctl.c (unalignctl_string, sys_prctl): Compile unconditionally.
Introduce XLAT_UOFF macro and use it to automatically transform
struct_user_offsets array into a more readable and compact form.
for n in $(sed -n 's/^[[:space:]]*{[[:space:]]*uoff(\([a-z_0-9]\+\)),.*/\1/p' process.c |sort -u); do
sed -i 's/^\([[:space:]]*\){[[:space:]]*uoff('"$n"'),[[:space:]]*"offsetof(struct user,[[:space:]]*'"$n"')"[[:space:]]*},$/\1XLAT_UOFF('"$n"'),/' process.c
done
* process.c (XLAT_UOFF): New macro.
(struct_user_offsets): Use it.
* clone.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* process.c: Move sys_clone, sys_setns, sys_unshare, sys_fork, and
related code to clone.c.
* groups.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* process.c: Move sys_setgroups, sys_getgroups, sys_setgroups32,
sys_getgroups32, and related code to groups.c.
* wait.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* process.c: Move sys_waitpid, sys_wait4, sys_osf_wait4, sys_waitid and
related code to wait.c.
* sched.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* process.c: Move sys_sched_getscheduler, sys_sched_setscheduler,
sys_sched_getparam, sys_sched_setparam, sys_sched_get_priority_min,
sys_sched_rr_get_interval, and related code to sched.c.
* link.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* file.c: Move sys_link, sys_linkat, sys_unlinkat, sys_symlinkat, and
related code to link.c.
* printmode.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* defs.h (sprintmode): New prototype.
* file.c (sprintmode): Make global and move to printmode.c.
* xattr.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* file.c: Move sys_setxattr, sys_fsetxattr, sys_getxattr, sys_fgetxattr,
sys_listxattr, sys_flistxattr, sys_removexattr, sys_fremovexattr,
and related code to xattr.c.
* sync_file_range.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* file.c: Move sys_sync_file_range, sys_sync_file_range2, and related
code to sync_file_range.c.
Unlike v1 capabilities which are 32-bit, v2 and v3 are 64-bit, but
before this change only lower 32 capability bits were decoded for
v2 and v3.
* xlat/capabilities1.in: New file.
* capability.c: Define v2/v3 CAP_* constants.
Include xlat/capabilities1.h.
(get_cap_header): New function.
(print_cap_header): Update to use get_cap_header result.
(print_cap_data): Decoder higher capability bits for v2 and v3.
(sys_capget, sys_capset): Use get_cap_header, update print_cap_header
and print_cap_data calls.
* tests/caps.c: New file.
* tests/caps.awk: New file.
* tests/caps.test: New test.
* tests/Makefile.am (CHECK_PROGRAMS): Add caps.
(TESTS): Add caps.test.
(EXTRA_DIST): Add caps.awk.
Various versions of <linux/capability.h> used to require different
workarounds to avoid conflicts with types defined by libc headers.
Define all required types and constants locally to fix this issue.
* configure.ac (AC_CHECK_HEADERS): Remove linux/capability.h.
* capability.c: Do not include <linux/capability.h>, remove workarounds
for problematic versions of <linux/capability.h> file.
Define CAP_* and _LINUX_CAPABILITY_VERSION_* constants as enums.
(struct __user_cap_header_struct, struct __user_cap_data_struct): Define.
* xlat/cap_version.in: Add #unconditional.
* xlat/capabilities.in: Likewise.
* cacheflush.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* system.c: Move inclusion of <asm/cachectl.h> to cacheflush.c.
[M68K, BFIN, SH]: Move to cacheflush.c.
* capability.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* system.c: Move inclusion of headers and macro definitions related
to capget and capset decoding to capability.c.
(print_cap_header, print_cap_data, sys_capget, sys_capset): Move
to capability.c.
* sysmips.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* system.c: Fix typo in the check for <linux/utsname.h>.
Move inclusions of <linux/utsname.h> and <asm/sysmips.h> to sysmips.c.
[MIPS]: Likewise.
Since parsers for setdomainname and sethostname syscalls are identical,
replace sys_setdomainname with an alias to sys_sethostname.
* linux/dummy.h (sys_setdomainname): Alias to sys_sethostname.
* linux/syscall.h (sys_setdomainname): Remove.
* process.c (sys_setdomainname): Remove.
Since parsers for getpeername and getsockname syscalls are identical,
replace sys_getpeername with an alias to sys_getsockname.
* linux/dummy.h (sys_getpeername): Alias to sys_getsockname.
* linux/syscall.h (sys_getpeername): Remove.
* net.c (sys_getpeername): Remove.
Since parsers for stime and time syscalls are identical,
replace sys_stime with an alias to sys_time.
* linux/dummy.h (sys_stime): Alias to sys_time.
* linux/syscall.h (sys_stime): Remove.
* time.c (sys_stime): Remove.
Special parsers for mkdir and mkdirat are redundant because
sys_chmod and sys_fchmodat implement the same decoding.
* file.c (decode_mkdir, sys_mkdir, sys_mkdirat): Remove.
* linux/dummy.h (sys_mkdir): Alias to sys_chmod.
(sys_mkdirat): Alias to sys_fchmodat.
* linux/syscall.h (sys_mkdir, sys_mkdirat): Remove.
* pathtrace.c (pathtrace_match): Do not check for sys_mkdirat.
For those socket descriptors that have no associated ip:port pairs
(or when this information is not available), -yy option prints
the same <socket:[INODE]> information as -y option, e.g.
$ strace -e sendto -yy ip l > /dev/null
sendto(3<socket:[23456789]>, ...
This change makes -yy output more informative: instead of just
printing "socket", the name of protocol behind the socket descriptor
will be printed, e.g.
sendto(3<NETLINK:[23456789]>, ...
* configure.ac (AC_CHECK_HEADERS): Add sys/xattr.h.
* tests/net-yy-accept.awk: Update to support protocol names.
* tests/net-yy-connect.awk: Likewise.
* util.c [HAVE_SYS_XATTR_H]: Include <sys/xattr.h>.
(getfdproto): New function.
(printfd): Use it.
Signed-off-by: Masatake YAMATO <yamato@redhat.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
MAXPATHLEN is defined to PATH_MAX, so replace the former with the latter.
* strace.c (startup_child): Replace MAXPATHLEN with PATH_MAX.
* util.c (printpathn, printpath): Likewise.
The freeze/thaw ones are simple, but the trim is an interesting struct.
* block.c (block_ioctl): Handle FIFREEZE/FITHAW/FITRIM.
* ioctl.c (ioctl_decode): Pass 'X' ioctls to block_ioctl.
Since <sys/uio.h> is standardized by POSIX and is present in all
available versions of glibc, it's safe to assume that any usable
libc implementation provides this header file.
* configure.ac (AC_CHECK_HEADERS): Remove sys/uio.h.
* io.c: Include <sys/uio.h> unconditionally.
(tprint_iov_upto, tprint_iov, sys_readv, sys_writev,
print_llu_from_low_high_val, sys_preadv, sys_pwritev): Define
unconditionally.
* net.c: Include <sys/uio.h> unconditionally.
* util.c: Include <sys/uio.h> unconditionally.
(dumpiov): Define unconditionally.
There are two important changes in here:
1. EWOULDBLOCK has been up to kernel 3.14 errorcode #246. Since hppa
folks had problems with EWOULDBLOCK != EAGAIN, this was changed in
kernel 3.14.
2. Starting with kernel 3.18, hppa folks changed some signal numbers in
such a way that we end up with SIGRTMIN == 32, which brings hppa in sync
with other linux ports.
Both were incompatible changes which basically broke hppa ABI, but since
they have been merged into the kernel, we have to follow.
The output format of the stack trace is supposed to be different
depending on whether symbol names are available in the build.
However, the check only verified the validity of the pointer, not of the
string pointed to (which could be empty).
This commit fixes the check so that the original output:
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x5e000
> /lib/libc-2.10.1.so(_IO_file_doallocate+0x8c) [0x68a38]
> /lib/libc-2.10.1.so(_IO_doallocbuf+0x6c) [0x78574]
> /lib/libc-2.10.1.so(_IO_file_overflow+0x184) [0x7763c]
> /lib/libc-2.10.1.so(_IO_file_xsputn+0x88) [0x76aac]
> /lib/libc-2.10.1.so(_IO_puts+0xc8) [0x6b64c]
> /bin/busybox(+0x0) [0x62c60]
> /bin/busybox(+0x0) [0x4940]
> /bin/busybox(+0x0) [0x499c]
> /bin/busybox(+0x0) [0x4e08]
> /lib/libc-2.10.1.so(__libc_init_first+0x30c) [0x1f84c]
> /lib/libc-2.10.1.so(__libc_start_main+0xd8) [0x1f9f8]
becomes:
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x5e000
> /lib/libc-2.10.1.so(_IO_file_doallocate+0x8c) [0x68a38]
> /lib/libc-2.10.1.so(_IO_doallocbuf+0x6c) [0x78574]
> /lib/libc-2.10.1.so(_IO_file_overflow+0x184) [0x7763c]
> /lib/libc-2.10.1.so(_IO_file_xsputn+0x88) [0x76aac]
> /lib/libc-2.10.1.so(_IO_puts+0xc8) [0x6b64c]
> /bin/busybox() [0x62c60]
> /bin/busybox() [0x4940]
> /bin/busybox() [0x499c]
> /bin/busybox() [0x4e08]
> /lib/libc-2.10.1.so(__libc_init_first+0x30c) [0x1f84c]
> /lib/libc-2.10.1.so(__libc_start_main+0xd8) [0x1f9f8]
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: Masatake YAMATO <yamato@redhat.com>
This patch is similar to what I did in commit
02f9f6b386.
That commit was for sendmsg and recvmsg system calls.
This one is for sendmmsg and recvmmsg system calls.
* defs.h (dumpiov_in_mmsghdr): New declaration.
* net.c (extractmmsghdr): New function derived from printmmsghdr.
(printmmsghdr): Use it.
(dumpiov_in_mmsghdr): New function.
* syscall.c (dumpio) [HAVE_SENDMSG]: Call dumpiov_in_mmsghdr
for recvmmsg and sendmmsg syscalls.
Signed-off-by: Masatake YAMATO <yamato@redhat.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
mmsghrd structure type is defined locally in printmmsghdr function.
However, more functions will refer the definition in modifications for
supporting "-e write=set" and "-e read=set" option for sendmmsg and
recvmmsg system calls.
After this change, the system definition of struct mmsghdr will be used
if configure reports it is available, falling back to the old local
definition.
* configure.ac (AC_CHECK_TYPES): Add struct mmsghdr.
* net.c [!HAVE_STRUCT_MMSGHDR] (struct mmsghdr): Define.
(printmmsghdr): Use previously defined struct mmsghdr.
Signed-off-by: Masatake YAMATO <yamato@redhat.com>
This patch is an initial step for supporting "-e write=set" and
"-e read=set" option for sendmmsg and recvmmsg system calls.
Coverting a data of msghdr32 to msghdr is needed both for
{send,recv}msg and {send,recv}mmsg to decode parameters.
To share the copying code in both decoders, a separate
function named copy_from_msghdr32 is introduced.
* net.c [SUPPORTED_PERSONALITIES > 1 && SIZEOF_LONG > 4]
(copy_from_msghdr32): New function.
(extractmsghdr) [SUPPORTED_PERSONALITIES > 1 && SIZEOF_LONG > 4]: Use it.
Signed-off-by: Masatake YAMATO <yamato@redhat.com>
While most of ioctl-related kernel headers are now exported by kernel's
headers_install, some are still modules_install only. The kernel's
headers installed into /usr/include/ are usually headers_install'ed and
therefore don't contain some internal headers we need. The solution is
to look for modules_install'ed headers for the running kernel, and fall
back to old behavior if they aren't found.
Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
Introduce v4l2.c, a decoder for the arguments of the video-input subset
of the v4l2 ioctl's. This is a combination of
- previous work by Peter Zotov <whitequark@whitequark.org>, found at
https://gist.githubusercontent.com/whitequark/1263207/raw/strace-4.6-v4l2-ioctls.patch
- previous work by William Manley <will@williammanley.net>, found at
http://marc.info/?l=strace&m=139395588520675
- forward port, additions and fixes by Philippe De Muyter <phdm@macqel.be>
As v4l2 is a moving target, I have made v4l2.c compilable with ancient
linux kernels by testing the availability of some macros. It has been
succesfully compiled on linux 3.10, 3.1, 2.6.31 and 2.6.22, and
succesfully used on linux 3.10 with a camera device.
* configure.ac: Check for availabilty of V4L2_* enum constants.
* Makefile.am (strace_SOURCES): Add v4l2.c.
* defs.h (v4l2_ioctl): New prototype.
* ioctl.c (ioctl_decode): Use v4l2_ioctl.
* v4l2.c: New file.
* xlat/v4l2_*.in: New files.
Signed-off-by: Philippe De Muyter <phdm@macqel.be>
Cc: Peter Zotov <whitequark@whitequark.org>
Cc: William Manley <will@williammanley.net>
For two ioctl entries with the same code, if one's name is a prefix
to another's name, keep the entry with a shorter name. Filter out
redundant ioctl entries at ioctlsort stage so that distributed
ioctlent.h.in files will be already filtered.
* linux/ioctlsort.c (is_not_prefix): New function.
(main): Use it.
* linux/ioctlent-filter.awk: Remove.
* Makefile.am (EXTRA_DIST): Remove linux/ioctlent-filter.awk.
($(ioctlent_h)): Don't use linux/ioctlent-filter.awk.
This fixes compilation with musl libc.
This approach was already used in process.c, so I assume it is safe.
* signal.c: Move [HAVE_LINUX_PTRACE_H] code out of [HAVE_SYS_REG_H] check.
* syscall.c: Likewise.
* util.c: Likewise.
Signed-off-by: Maarten ter Huurne <maarten@treewalker.org>
The decoding of SIOCSIFNAME is incorrect. It does not use
the ifr_index field to look things up, but ifr_newname.
* sock.c (sock_ioctl): Split out SIOCSIFNAME from SIOCGIFNAME and
display ifr_newname.
The ifr name fields of the ifreq structure might not be NUL terminated.
If the user makes an ioctl call where they aren't, then strace ends up
reading random content from its own stack. Limit the printf lengths.
* sock.c (sock_ioctl): Add explicit length limits to ifr_name printfs.
On Android, 32-bit arm and x86 use __kernel_mode_t (an unsigned short)
as their mode_t. The open(2) flags are actually an int, so high ones
like O_CLOEXEC get truncated if you coerce them to mode_t.
* defs.h (tprint_open_modes, sprint_open_modes): Change argument type
from mode_t to int.
* file.c (tprint_open_modes, sprint_open_modes): Likewise.
Signed-off-by: Elliott Hughes <enh@google.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
* defs.h (xlat_search): New prototype.
* util.c (xlat_bsearch_compare, xlat_search): New functions.
* file.c (sprintfstype): Use xlat_search for fsmagic lookup.
* xlat/fsmagic.in: Sort by value and mark as not NULL-terminated.
* tests/statfs.c: New file.
* tests/statfs.test: New test.
* tests/Makefile.am (check_PROGRAMS): Add statfs.
(statfs_CFLAGS): Define.
(TESTS): Add statfs.test.
* tests/.gitignore: Add statfs.
When two ore more -y options are given, print local and remote ip:port
pairs associated with socket descriptors. This implementation uses
NETLINK_INET_DIAG for sockaddr lookup; it's based on the patch
prepared by Zubin Mithra as a part of his GSoC 2014 strace project.
* Makefile.am (strace_SOURCES): Add socketutils.c
(EXTRA_DIST): Add linux/inet_diag.h and linux/sock_diag.h.
* defs.h (print_sockaddr_by_inode): New prototype.
* linux/inet_diag.h: New file.
* linux/sock_diag.h: Likewise.
* socketutils.c: Likewise.
* strace.1: Document -yy option.
* strace.c (usage): Likewise.
* util.c (printfd): Use print_sockaddr_by_inode.
Wrap sa_restorer member definitions in #ifdef SA_RESTORER to be
consistent with their use.
If an architecture does not provide sa_restorer members but still
defines SA_RESTORER macro, the latter has to be explicitly undefined.
This change fixes compilation failures like this one:
signal.c: In function 'decode_old_sigaction':
signal.c:631:21: error: 'struct old_sigaction' has no member named 'sa_restorer'
signal.c: In function 'decode_new_sigaction':
signal.c:1224:21: error: 'struct new_sigaction' has no member named 'sa_restorer'
* signal.c (struct old_sigaction, struct old_sigaction32,
struct new_sigaction, struct new_sigaction32):
Wrap sa_restorer member in #ifdef SA_RESTORER.
(decode_old_sigaction, decode_new_sigaction):
Wrap use of sa32.sa_restorer in #ifdef SA_RESTORER.
Signed-off-by: Vicente Olivert Riera <vincent.riera@imgtec.com>
* configure.ac (gl_WARN_ADD): Add -Wsign-compare.
* defs.h (struct tcb): Change 'currpers' type to unsigned.
(struct xlat): Change 'val' type to unsigned
(signame): Add 'const' qualifier to its argument.
(xlookup, printxval): Add 'const' qualifier to the 2nd argument and
change its type to unsigned.
(printpathn): Change the 3rd argument type to unsigned.
(ioctl_lookup): Change 1st argument type to unsigned.
* count.c (call_summary_pers, call_summary): Change 'i' type to unsigned.
* file.c (print_xattr_list): Fix comparisons between signed and unsigned
long values.
* ioctl.c (compare): Fix cast.
(ioctl_lookup): Change 1st argument type to to unsigned.
(ioctl_next_match): Change 'code' type to unsigned.
* mem.c (sys_move_pages): Change 'i' type to unsigned.
* mtd.c (mtd_ioctl): Change 'i' and 'j' types to unsigned.
Print 'i' using %u format string.
* process.c (sys_prctl): Change 'i' type to unsigned.
(printargv): Change 'n' type to unsigned.
(sys_ptrace): Change 'addr' type to unsigned.
* scsi.c (print_sg_io_buffer): Add 'const' qualifier to 'len' argument
and change its type to unsigned. Change 'i' and 'allocated' types
to unsigned.
* signal.c (signame): Add 'const' qualifier to its argument.
Fix comparisons between signed and unsigned values.
(sprintsigmask_n, printsiginfo): Fix comparisons between signed and
unsigned values.
* sock.c (sock_ioctl): Change 'i' and 'nifra' types to unsigned.
* strace.c (expand_tcbtab, alloctcb): Change 'i' type to unsigned.
(detach): Change 'sig' type to unsigned.
(startup_attach): Change 'tcbi' type to unsigned.
(startup_child): Change 'm', 'n', and 'len' types to unsigned.
(init): Use new variable to iterate 'tcbtab'.
(pid2tcb): Change 'i' type to unsigned.
(cleanup): Change 'i' and 'sig' types to unsigned.
* syscall.c (update_personality): Change 'personality' argument type
to unsigned.
(struct qual_options): Change 'bitflag' type to unsigned.
(reallocate_qual): Add 'const' qualifier to its argument and change its
type to unsigned.
(qualify_one): Change 'n' and 'bitflag' arguments types to unsigned.
Add 'const' qualifier to 'n', 'not', and 'pers' arguments.
Change 'p' type to signed int.
(qual_syscall): Change 'bitflag' argument type to unsigned.
Add 'const' qualifier to 'bitflag' and 'not' arguments.
Change 'p' type to signed int.
(qual_signal): Change 'bitflag' argument type to unsigned.
Add 'const' qualifier to 'bitflag' and 'not' arguments.
Change 'i' type to unsigned.
(qual_desc): Change 'bitflag' argument type to unsigned.
Add 'const' qualifier to 'bitflag' and 'not' arguments.
(qualify): Change 'i' type to unsigned.
(get_scno): Change 'currpers' type to unsigned.
Fix a comparison between signed and unsigned values.
* system.c (sys_sysctl): Change 'cnt' and 'max_cnt' types to unsigned.
Fix comparisons between signed and unsigned values.
* util.c (xlookup, printxval): Add 'const' qualifier to 'val' argument
and change its type to unsigned.
(printuid): Fix a comparison between signed and unsigned values.
(printpathn): Change 'n' argument type to unsigned.
(printstr): Change 'size' type to unsigned.
Fix a comparison between signed and unsigned values.
(setbpt): Change 'i' type to unsigned.
* net.c (printsock): Silence a compilation warning.
* reboot.c (sys_reboot): Likewise.
A pointer to fd_set was used uninitialized when nfds == 0.
* desc.c (decode_select): Initialize fds.
Reported-by: Zubin Mithra <zubin.mithra@gmail.com>
Stop using an outdated partial copy of libaio.h, switch back to external
libaio.h from libaio.
This partially reverts commit 2df03c494e.
* aio.c: Drop a partial copy of libaio.h, include <libaio.h> instead.
(print_common_flags): Check for HAVE_STRUCT_IOCB_U_C_FLAGS.
(sys_io_submit): Check for HAVE_DECL_IO_CMD_PWRITE and
HAVE_DECL_IO_CMD_PWRITEV.
* configure.ac: Check for libaio.h and declaration it provides.
The {...} code changes the working dir with `cd`, but the commands outside
of that block expects to be in the original dir. Change to a subshell so
the path outside of this block remains unchanged.
* Makefile.am ($(srcdir)/CREDITS): Change {...} to (...).
When the preadv/pwritev syscalls were added, the ones before it in the
ia64 list were missed, so all the syscalls there and later were not in
the right location (causing things to be decoded incorrectly).
Add the missing syscalls before preadv which also re-aligns all the
syscalls after that point. This fixes the uio.test.
* linux/ia64/syscallent.h: Add syscalls 1310 through 1318.
Looks like ia64 doesn't have sa_restorer either, yet still defines
SA_RESTORER. Deploy the same trick that HPPA is using to make the
test pass.
* signal.c (SA_RESTORER): Undefine when IA64 is defined.
(struct new_sigaction) [IA64]: Disable sa_restorer.
detach-stopped.test is known to fail when PTRACE_SEIZE is not available,
so skip the test in that case.
* tests/detach-stopped.test: Check for "strace -d" output and skip the
test when it says that PTRACE_SEIZE doesn't work.
The preadv/pwritev symbols weren't added to glibc until the 2.10 release,
so trying to build the uio test leads to link failures. Add configure
tests and update uio.test to handle this.
* configure.ac (AC_CHECK_FUNCS): Add preadv/pwritev.
* tests/uio.c: Include config.h.
(main): Check for HAVE_PREADV and HAVE_PWRITEV.
* tests/uio.test: Check exit status of uio helper.
The current detach test code does:
set -e
...
cleanup() {
set +e
kill ...
wait ...
}
...
cleanup
exit 0
The problem is that while `set -e` is disabled for the body of the
cleanup function, it isn't necessarily disabled in the caller scope.
So if the return value of the cleanup function (`wait` in this case)
is non-zero, the script ends up failing overall.
Add an explicit return 0 to the cleanup function so that we don't kill
the overall test pipeline.
* tests/detach-running.test (cleanup): Add return 0.
* tests/detach-sleeping.test (cleanup): Likewise.
* tests/detach-stopped.test (cleanup): Likewise.
This way I don't have to keep reading up on these options and wondering
why the code isn't aborting when the call fails.
* tests/set_ptracer_any.c (main): Note prctl failures are ok.
Previous commit here re-added the bugs trying to be fixed due to a
logic thinko. The patches were tested in isolation and hand merged
later. Oops.
* signal.c (struct new_sigaction): Change || to &&.
Running Linux 3.15 (sparc64) and glibc 2.17 (sparc32) triggers a
rt_sigaction call that does not use SA_RESTORER and has an order
where it inserts a restorer and a size. The current tests don't
support that ordering, so add another regex.
* tests/sigaction.awk: Support no SA_RESTORER and swapped args.
Both these arches have a rt_sigaction syscall that takes 5 args, not 4.
* linux/alpha/syscallent.h (rt_sigaction): Change nargs to 5.
* linux/sparc/syscallent.h (rt_sigaction): Change nargs to 5.
Since the rt_sigaction syscall on hppa doesn't have a sa_restorer,
do not include it in the kernel struct.
We also have to undefine SA_RESTORER so that code doesn't try to
use it. The headers will export this, but the syscall doesn't
actually respect it.
* signal.c (SA_RESTORER): Undefine when HPPA is defined.
(struct new_sigaction): Disable sa_restorer on hppa.
Since the rt_sigaction syscall on alpha doesn't have a sa_restorer,
do not include it in the kernel struct.
* signal.c (struct new_sigaction): Disable sa_restorer on alpha.
This fixes regression introduced by the previous commit.
* io.c (print_llu_from_low_high_val) [SIZEOF_LONG != SIZEOF_LONG_LONG]:
Cast argument to unsigned long before casting it to unsigned long long.
Starting in 3.16, these two syscalls have gotten their own entry
point for x32. See linux 7fd44dacdd803c0bbf38bf478d51d280902bb0f1.
* linux/x32/syscallent.h: Change existing io_{setup,submit} to 64bit,
and add new entry points for x32 specifically.
strace -k does not produce a reliable output on all supported
configurations yet, even basic strace-k.test is known to fail
on some of them.
* strace.c (usage): Document -k option as experimental.
* strace.1: Likewise.
* NEWS: Likewise.
Split stack-fcall.c into several compilation units so that intermediate
function calls would not be optimized out by compiler.
* tests/stack-fcall.c: Move intermediate functions to ...
* tests/stack-fcall-*.c: ... new files.
* tests/Makefile.am (stack_fcall_SOURCES): Add stack-fcall-*.c.
Add two more function calls to the stack. Suggested by Masatake YAMATO.
* tests/stack-fcall.c (f1): Rename to f3.
(f1, f2): New functions.
* tests/strace-k.test: Update.
* unwind.c (build_mmap_cache): Move local variables to the code branch
where they are used. Check return code of sscanf and strdup. Do not
treat unusual memory mappings as fatal errors. Do not skip memory
mappings with path names starting with "[".
* unwind.c (fopen_for_input): Define to fopen64 iff
[_LARGEFILE64_SOURCE && HAVE_FOPEN64], otherwise define it to fopen.
(build_mmap_cache): Use fopen_for_input instead of fopen.
Fix compilation warnings in unwind.c on 32-bit architectures.
On some architectures getuid is actually getuid32, so change the test
to use getpid instead of getuid.
* unwind.c (STACK_ENTRY_SYMBOL_FMT): Explicitly cast function_off_set
to unsigned long.
(queue_put_error): Change the 3rd argument's type to unsigned long.
* tests/stack-fcall.c (f1): Use getpid instead of getuid.
* tests/strace-k.test: Likewise.
* debian/control (strace64): Fix a typo in package description.
Patch by Pascal De Vuyst <pascal.devuyst@gmail.com>.
(strace, strace-udeb): Add x32 to architecture list.
Patch by Guillaume Morin <guillaume@morinfr.org>.
(strace, strace-udeb): Add or1k to architecture list.
Patch by Christian Svensson <debian@cmd.nu>.
(strace, strace-udeb): Add arm64 to architecture list,
and remove defunct arm.
Patch by Wookey <wookey@debian.org>.
This fixes Debian bugs: #697625, #727018, #742235, #749956.
$ groff -ww -mandoc -z strace.1
strace.1:65: warning: macro `IX' not defined
* strace.1: define IX macro as empty for groff.
Change remaining '-' as minus to '\-'.
Have two word spaces after a full stop as an end of sentence.
Use extra space ('\,' or '\/') between roman and italic characters.
Based on patch by Bjarni Ingi Gislason <bjarniig@rhi.hi.is>.
This fixes Debian bug #725987.
Instead of handling stacktrace capturing and mmap cache invalidating in
sys_* functions, handle them uniformly in trace_syscall_entering using
new flags introduced by previous two commits.
The patch is simpler than its older version(v3). The value of
hide_log_until_execve is just ignored. I found the value is nothing
to do with this patch. unwind_cache_invalidate is mentioned only
once in trace_syscall_exiting.
Both are suggested by Dmitry Levin.
Signed-off-by: Masatake YAMATO <yamato@redhat.com>
Add SE flag to execve, exit, and exit_group syscall entries.
Add SI flag to brk, execve, mmap, mprotect, mremap, munmap,
remap_file_pages, shmat, and shmdt syscall entries.
Some system calls require capturing the stack trace before they are
processed in kernel. Typical one is execve. Some system calls require
invalidating mmap cache after they are processed in kernel.
In current implementation these requirements are handled directly by
appropriate syscall handlers. However, it is difficult to keep the
source code maintainable using this approach to cover all system calls
which have such requirements.
A more generic way to implement this is to flag all syscalls that
require special processing, and handle these flags right in
trace_syscall_entering instead of changing syscall handlers.
This patch just defines new flags: STACKTRACE_INVALIDATE_CACHE and
STACKTRACE_CAPTURE_ON_ENTER.
The names of macros are suggested by Dmitry Levin.
Signed-off-by: Masatake YAMATO <yamato@redhat.com>
Here is the benchmark of the dwarf cache.
Target program:
#include <sched.h>
int main(void)
{
unsigned int max = 0x6fff, i;
for (i = 0; i < max; i++)
sched_yield();
return 0;
}
Command line:
./strace -o /dev/null -k a.out
With the dwarf cache:
real 0m12.081s
user 0m3.858s
sys 0m8.194s
Without the dwarf cache:
real 0m22.326s
user 0m5.218s
sys 0m16.952s
Signed-off-by: Masatake YAMATO <yamato@redhat.com>
When a file mmap'ed to the target process is unlink'ed, backtracing the
stack would fail. Current implementation reports it as
"backtracing_error". To avoid confusion, the message is changed to
"expected_backtracing_error".
Here is the reproducer:
$ cat ./p-deleted.c
#include <unistd.h>
int main(int argc, char **argv) {
return unlink(argv[0]) < 0;
}
$ strace -e unlink -k ./p-deleted
unlink("./p-deleted") = 0
> /usr/lib64/libc-2.18.so(unlink+0x7) [0xe7f17]
> /home/yamato/var/strace/t_unwind/p-deleted (deleted)(+0x0) [0x575]
> /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65]
> backtracing_error [0x7ffff1365590]
+++ exited with 0 +++
p-deleted is deleted therefore backtracing_error is reported. This
patch records the deleted marker when making mmap cache and refers the
recorded information in the case "backtracing_error" to switch the
message.
Here is the output of this patch:
$ strace -e unlink -k ./p-deleted
unlink("./p-deleted") = 0
> /usr/lib64/libc-2.18.so(unlink+0x7) [0xe7f17]
> /home/yamato/var/strace/t_unwind/p-deleted (deleted)(+0x0) [0x575]
> /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65]
> expected_backtracing_error [0x7ffff1365590]
+++ exited with 0 +++
This solution is not perfect: if a file is unlink'ed after making the
mmap cache and before unwinding, strace cannot have a chance to record
the deleted marker.
In this version of patch, hardcoded magic number used in comparing "(delete)"
string is replaced with strlen as suggested by Dmitry Levin.
In old version of patch, the deleted entry was thrown away from mmap
cache to avoid to report "backtracing_error". In this patch I keep it,
and just switch the error message.
Inspired by the review comment from Dmitry Levin.
Signed-off-by: Masatake YAMATO <yamato@redhat.com>
captured stacktrace is printed in unwind_tcb_fin if tcp->queue is not
empty. This should happen before printing detached message, so
unwind_tcb_fin is moved to the top of droptcb.
This is implicitly suggested by Dmitry Levin in patch review process.
Signed-off-by: Masatake YAMATO <yamato@redhat.com>
A mmap cache belonging to a tcb was updated when a system call which
changed the memory mapping was called. This implementation was assumed
the mapping was changed only by the tcb. However, this assumption is
incorrect if the target application is multi-threaded; more than two
tcbs can shared the same memory mapping and a tcb can modify it without
being noticed by the others.
This change introduces a global integer variable mmap_cache_generation,
and mmap_cache_generation field to struct tcb. The variable
is incremented each time a process enters a syscall that can modify its
memory mapping. Each tcb records the value of this variable at the
moment if building its mmap cache. Every mmap cache associated with
the given tcb can be validated by comparing its mmap_cache_generation
field with the variable mmap_cache_generation.
This implementation is inefficient. If strace attaches two processes
which don't share the memory mapping, rebuilding mmap cache of a tcb
triggered by another tcb's mmap system call is not necessary.
Signed-off-by: Masatake YAMATO <yamato@redhat.com>
This is the second step for splitting capturing from printing.
New `queue' field is added to tcb. Captured stacktrace is stored here.
The field is initialized/finalized at unwind_tcb_init/unwind_tcb_fin.
New API function unwind_capture_stacktrace is added. This function
captures the currest stack using stracktrace_walker and records it in
tcb. It's printing is delayed to the next call of
unwind_print_stacktrace.
unwind_print_stacktrace is extended. Now it checks queue field of
the given tcb at the start of function. If the function finds a
captured stack trace, the latter is printed using stracktrace_walker.
Currently unwind_capture_stacktrace invocations are added directly to
handlers of mmap, munmap, mprotect, and execve.
Here is the difference of output with/without patch:
(without patch)
execve("./test-fork", ["./test-fork"], [/* 56 vars */]) = 0
> /usr/lib64/ld-2.18.so(check_one_fd.part.0+0x82) [0x11f0]
(with patch)
execve("./test-fork", ["./test-fork"], [/* 54 vars */]) = 0
> /usr/lib64/libc-2.18.so(execve+0x7) [0xbcd27]
> /home/yamato/var/strace/strace(exec_or_die+0x10c) [0x26ac]
> /home/yamato/var/strace/strace(startup_child+0x346) [0x134f6]
> /home/yamato/var/strace/strace(init+0x89f) [0x13dff]
> /home/yamato/var/strace/strace(main+0xa) [0x26ca]
> /usr/lib64/libc-2.18.so(__libc_start_main+0xf5) [0x21d65]
> /home/yamato/var/strace/strace(_start+0x29) [0x2799]
In older version output lines of captured elements were built when
printing. In this version they are built when capturing the stack.
As result, unneeded dynamic memory allocations are avoided.
Suggested by Luca Clementi.
In older version the combination of snprintf and realloc were used.
In this version they are replaced with asprintf.
Suggested by Dmitry Levin.
Signed-off-by: Masatake YAMATO <yamato@redhat.com>
In current implementation, the stack trace is captured and printed at
the same time, in trace_syscall_exiting. This approach cannot
provide user expected information when a system call changes the
memory mapping. In such cases, the stack trace should be captured on
entering syscall and printed on exiting.
As the initial step for splitting capturing from printing, this change
introduces stacktrace_walker utility function. It can be used both for
capturing in trace_syscall_entering and printing in
trace_syscall_exiting.
Signed-off-by: Masatake YAMATO <yamato@redhat.com>
free_libunwind_ui is expected to release all unwind related resources
attached to tcp.
* strace.c (droptcb): Move delete_mmap_cache call ...
* unwind.c (free_libunwind_ui): ... to here.
Signed-off-by: Masatake YAMATO <yamato@redhat.com>
Print the stack trace of the traced process after each system call when
-k option is specified. It is implemented using libunwind to unwind the
stack and to obtain the function name pointed by the IP.
Based on the code that was originally taken from strace-plus
of Philip J. Guo.
* configure.ac: Add --with-libunwind option. Check libunwind support.
* Makefile.am: Add libunwind support.
* defs.h (struct tcb) [USE_LIBUNWIND]: Append libunwind specific fields.
[USE_LIBUNWIND] (stack_trace_enabled, alloc_mmap_cache,
delete_mmap_cache, print_stacktrace): New prototypes.
* mem.c (print_mmap, sys_munmap, sys_mprotect): Add libunwind support.
* process.c (sys_execve): Likewise.
* strace.c (usage, alloctcb, droptcb, init): Likewise.
* syscall.c (trace_syscall_exiting): Likewise.
* unwind.c: New file.
* strace.1: Document -k option.
* xlat/gen.sh: Define all xlat structs not declared in defs.h as static.
Some symbolic constants are not macros, extend #ifdef check to cover
symbolic constants checked by AC_CHECK_DECLS.
Handle complex symbolic constants in SYMBOL|... form.
Handle symbolic constants in 1<<SYMBOL form.
Handle numeric constants.
Implement #unconditional directive that turns off preprocessor checks.
Implement #unterminated directive that turns off adding XLAT_END.
Now that ./xlat/gen.sh has to be run before autoreconf,
replace all autoreconf calls with ./bootstrap call.
* bootstrap: Forward arguments to autoreconf.
* build_static_example.sh: Replace autoreconf call with bootstrap call.
* make-dist: Likewise.
* qemu_multiarch_testing/README: Likewise.
Time spent in system time is not useful where a syscall depends on some
non-CPU resource, eg. typically open() or stat() to a network drive.
This patch adds a new flag (-w) to produce a summary of the time
difference between beginning and end of the system call (ie. latency)
This functionality has been useful to profile slow processes that
are not CPU-bound.
Signed-off-by: Mark Hills <mark.hills@framestore.com>
* count.c (count_syscall): Add const qualifier to timeval argument and
rename it. Store the wall clock time spent while in syscall in separate
timeval variable.
* defs.h (count_syscall): Update prototype.
* syscall.c (trace_syscall_exiting): Update count_syscall invocation.
The argument passed to sethostname and setdomainname syscalls, as well
as the string returned by gethostname syscall, is not a pathname, so
printpathn is not the right method for its decoding.
* process.c (sys_sethostname, sys_setdomainname): Decode 1st argument
using printstr instead of printpathn.
[ALPHA] (sys_gethostname): Likewise.
When strace is built with large file support definitions in CFLAGS (as
may be provided by buildroot) the C library headers may expose a 64-bit
rlim_t even though the struct rlimit fields used by the system call
interface are only 32-bit. The SIZEOF_RLIM_T will then be 8 which
results in bad decoding of the getrlimit and setrlimit syscalls.
This is fixed by replacing unreliable SIZEOF_RLIM_T based checks with
checks for current_wordsize.
Signed-off-by: James Hogan <james.hogan@imgtec.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
I found that I use it quite often. Lets make it so that
after cut-n-pasting it into a file, there is no need
to edit the result (e.g. no need to remove C comment
chars from every line.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Without this patch, strace claims that parent_tidptr == tls, which is
clearly wrong. It is expected that parent_tidptr == child_tidptr.
* process.c [X86_64] (ARG_CTID, ARG_TLS): Take current
personality into account.
Signed-off-by: Elliott Hughes <enh@google.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
We need to handle this situation more like x86-64. 32-bit arm and i386
actually have a common struct stat64, except the arm one must not be
packed. Additionally, on aarch64 the 32-bit personality is personality 0.
Signed-off-by: Elliott Hughes <enh@google.com>
OABI is rarely used in ARM EABI systems nowadays, so disable its support
by default. Add --enable-arm-oabi option to enable ARM OABI support.
* configure.ac: New option --enable-arm-oabi.
* syscall.c (get_scno) [ARM]: Check ENABLE_ARM_OABI macro defined by
configure instead of undocumented STRACE_KNOWS_ONLY_EABI macro.
Patch fb642bb6d6 fixed building with
HAVE_STAT64 for aarch64 with uapi kernel headers but not x86_64.
The workaround needed to be applied to all LP64 architectures, not
just aarch64. This patch fixes that and adds an explanatory comment.
Signed-off-by: Elliott Hughes <enh@google.com>
The only supported domain for socketpair syscall is AF_UNIX, so
no decoding related to other domains is required for socketpair.
* net.c (sys_socketpair): Remove support for PF_INET and PF_IPX domains,
print the protocol argument as is.
Decode siginfo_t more clearly for si_code SI_TIMER.
The 'pid' is actually a POSIX timer id, and the 'uid' is actually the
overrun.
Also factor out the si_value dumping so it's the same for every si_code.
Signed-off-by: Elliott Hughes <enh@google.com>
If an aarch64 strace is tracing a process using the arm personality, it
also needs to call the shuffle_scno function for the ARM-specific
syscalls.
* syscall.c (shuffle_scno): Define on AARCH64.
(get_scno) [AARCH64]: Call shuffle_scno when the tracee is in 32-bit mode.
Signed-off-by: Elliott Hughes <enh@google.com>
Assume that F_SETLK64, F_SETLKW64, and F_GETLK64 are either defined or
not defined altogether.
Do not assume that sizeof(off_t) < sizeof(long long) when F_SETLK64 is
undefined.
This change fixes build with musl libc on x86.
* configure.ac: Define SIZEOF_OFF_T.
* desc.c (USE_PRINTFLOCK64): New macro.
(struct flock64, printflock64): Do not define on X32.
(printflock): Replace X32 specific workaround with SIZEOF_OFF_T check.
Fix printing off_t members of struct flock.
(sys_fcntl): Use USE_PRINTFLOCK64.
F_FREESP and F_FREESP64 fcntl commands are not available in Linux
and therefore the code implementing their decoding is useless.
Besides that, F_FREESP64 decoding is too complicated to support.
* desc.c (fcntlcmds): Remove F_FREESP and F_FREESP64.
Remove F_FREESP64 from the check whether to define struct flock64.
(sys_fcntl): Remove F_FREESP and F_FREESP64 support.
struct sigaction is another structure that contains members
whose size differs between 32-bit and 64-bit personalities.
* signal.c [HAVE_SIGACTION] (old_sigaction32): New structure.
[HAVE_SIGACTION] (decode_old_sigaction): Decode 32-bit struct
old_sigaction on a 64-bit host.
Signed-off-by: Elliott Hughes <enh@google.com>
aarch64's uapi header files have a struct stat but no struct stat64.
To correctly decode a 32-bit process' s struct stat64 we need
HAVE_STAT64, but then the build fails because there is no struct stat64.
Luckily, the aarch64 struct stat is structurally equivalent to the arm
struct stat64, so we can just reuse that.
* file.c [AARCH64] (stat64): Define to stat.
Signed-off-by: Elliott Hughes <enh@google.com>
The generic version of ioctlsort.c became obsolete after commit
v4.6-240-g5afdf12 that removed its last non-Linux users.
* ioctlsort.c: Remove.
* Makefile.am (EXTRA_DIST): Remove ioctlsort.c.
Reported-by: Elliott Hughes <enh@google.com>
stack_t is one of many structures that contain members
whose size differs between 32-bit and 64-bit personalities.
* signal.c (print_stack_t): Decode 32-bit stack_t on a 64-bit host.
Reported-by: Elliott Hughes <enh@google.com>
The sigset_t provided by libc is not quite convenient.
In glibc, sigset_t is an array with space for 1024 bits, which is much
more than required: all architectures supported by Linux have only 64
signals except MIPS, which has 128.
In bionic libc, LP32 sigset_t is only 4 bytes long, which is less than
necessary.
With this change, signal mask is decoded without use of intermediate
sigset_t structure, which saves us some cpu cycles in case of glibc with
its inflated sigset_t, and enables build with libcs where sigset_t is
broken.
Old implementation used to check each signal number in the given signal
mask twice using sigismember().
New implementation is based on popcount and next_set_bit() so it's
noticeably faster.
* configure.ac: Check for __builtin_popcount.
* signal.c: Ensure that NSIG >= 32.
(sprintsigmask, sprintsigmask_long, printsigmask): Remove.
(popcount32, sprintsigmask_n): New functions.
(tprintsigmask_addr, sprintsigmask_val, tprintsigmask_val): New macros.
(print_sigset_addr_len, sys_sigsetmask, sys_sigreturn, sys_siggetmask,
sys_sigsuspend, sys_sigprocmask, decode_new_sigaction): Update to use
new signal mask decoding interface.
* tests/sigaction.c (main): Add a test with almost filled signal mask.
* tests/sigaction.awk: Update.
Add generic tests for fopen64 and fputs_unlocked functions to fix build
with Bionic libc that does not provide them.
* configure.ac (AC_CHECK_FUNCS): Add fopen64 and fputs_unlocked.
* strace.c [_LARGEFILE64_SOURCE]: Use fopen instead of fopen64
if !HAVE_FOPEN64.
Use fputs instead of fputs_unlocked if !HAVE_FPUTS_UNLOCKED.
* vsprintf.c: Use fputs instead of fputs_unlocked
if !HAVE_FPUTS_UNLOCKED.
Reported-by: Elliott Hughes <enh@google.com>
* syscall.c (get_scno) [POWERPC64]: Fix 64-bit process detection
on embedded powerpc.
Signed-off-by: James Yang <james.yang@freescale.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
Add a generic test for <scsi/sg.h> availability to fix build with
Bionic libc that does not provide <scsi/sg.h>.
* configure.ac (AC_CHECK_HEADERS): Add scsi/sg.h.
* ioctl.c (ioctl_decode): Do not call scsi_ioctl if !HAVE_SCSI_SG_H.
* scsi.c: Do not compile scsi ioctl decoding if !HAVE_SCSI_SG_H.
Reported-by: Elliott Hughes <enh@google.com>
The patch originally submitted by Ezequiel García was OK, but I somehow
managed to mangle it so that most of the patch was not applied.
According to Ezequiel García, an architecture port based on the generic
syscall ABI is in progress.
This reverts commit 61e426e87a.
If SA_RESTORER is not defined by libc headers but defined by kernel
headers, use the definition provided by kernel headers for proper
sigaction decoding.
* signal.c [!SA_RESTORER]: Define to ASM_SA_RESTORER if the latter is
defined, regardless of architecure.
Kernel header <asm/signal.h> cannot be included from regular code
because it conflicts with libc headers, but SA_RESTORER is needed in
signal.c, so SA_RESTORER value is forwarded from <asm/signal.h> to
config.h using a configure check.
* configure.ac (ASM_SA_RESTORER): Define if SA_RESTORER is defined
in <asm/signal.h>.
This commit adds strace support for Altera's Nios-II official
kernel port as found in git://git.rocketboards.org/linux-socfpga.git
Notice that this an out-of-tree kernel architectural port, and uses the
legacy (non-generic) system call ABI. In particular, the port doesn't
support PTRACE_GETREGSET, so the implementation is based on PTRACE_GETREGS.
Given it's mandatory for new architectures to support the generic
syscall ABI and PTRACE_GETREGSET, if the nios2 architecure is ever
mainlined, the strace support will have to be re-factored accordingly.
* linux/nios2/ioctlent.h.in: New file.
* linux/nios2/syscallent.h: Likewise.
* Makefile.am (EXTRA_DIST): Add linux/nios2/ioctlent.h.in and
linux/nios2/syscallent.h.
* configure.ac: Add NIOS2 to the list of supported architectures.
* defs.h [NIOS2]: Use register reading system.
* process.c (struct_user_offsets): Add NIOS2 support.
* syscall.c (get_regs, get_scno, get_syscall_args,
get_syscall_result, get_error): Likewise.
* util.c (change_syscall): Likewise.
* mem.c (sys_getpagesize): Define on NIOS2.
* system.c [NIOS2] (sys_cacheflush, sys_nios2cmpxchg): New functions.
Signed-off-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Acked-by: Mike Frysinger <vapier@gentoo.org>
Automatically update all xlat structures using the following sed regexp:
s/^[[:space:]]*{[[:space:]]*0[[:space:]]*,[[:space:]]*NULL[[:space:]]*,\?[[:space:]]*}[[:space:]]*,\?[[:space:]]*/\tXLAT_END/
Automatically convert all xlat structures to XLAT form
using the following sed regexp:
s/^[[:space:]]*{[[:space:]]*\([^",}[:space:]]\+\)[[:space:]]*,[[:space:]]*"\1",\?[[:space:]]*}[[:space:]]*/\tXLAT(\1)/
* time.c (cpuclocknames): New xlat structure.
(printclockname): New function that decodes posix clock names,
including dynamic fd encoded clocks.
(sys_clock_settime, sys_clock_gettime, sys_clock_nanosleep,
sys_clock_adjtime, sys_timer_create, sys_timerfd, sys_timerfd_create):
Use it.
Signed-off-by: Stefan Sørensen <stefan.sorensen@spectralink.com>
Due to automake limitations, some features work only for tests
that end in one of the suffixes listed in TEST_EXTENSIONS.
* tests/detach-running: Rename to detach-running.test.
* tests/detach-sleeping: Rename to detach-sleeping.test.
* tests/detach-stopped: Rename to detach-stopped.test.
* tests/net: Rename to net.test.
* tests/ptrace_setoptions: Rename to ptrace_setoptions.test.
* tests/qual_syscall: Rename to qual_syscall.test.
* tests/sigaction.sh: Rename to sigaction.test.
* tests/stat: Rename to stat.test.
* tests/strace-f: Rename to strace-f.test.
* tests/Makefile.am (TESTS): Update.
(LOG_COMPILER): Rename to TEST_LOG_COMPILER.
MIPS userland uses the same sigaction structure with a full signal
mask for old_sigaction and new_sigaction and does does not have
an sa_restorer field.
These changes have been tested on MIPS O32 big/little endian, MIPS N64
big endian and x86-64.
* signal.c (old_sigaction) [MIPS]: Add definition for MIPS.
(decode_old_sigaction) [MIPS]: Print sa_mask according to its definition.
(new_sigaction) [MIPS]: Add definition for MIPS.
Signed-off-by: Chris Dearman <chris.dearman@imgtec.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
* signal.c [HAVE_SIGACTION] (decode_old_sigaction): New function.
[HAVE_SIGACTION] (sys_sigaction): Use it.
(decode_new_sigaction): New function.
(sys_rt_sigaction): Use it.
Since "struct sigaction" varies between architectures, rt_sigaction
decoding sometimes produces incorrect output. This test is expected
to catch basic rt_sigaction decoding bugs.
Based on a patch proposed by Chris Dearman.
* tests/sigaction.c: New file.
* tests/sigaction.awk: Likewise.
* tests/sigaction.sh: New test.
* tests/Makefile.am (check_PROGRAMS): Add sigaction.
(TESTS): Add sigaction.sh.
(EXTRA_DIST): Add sigaction.awk.
* tests/.gitignore: Add sigaction.
The code base has settled on PTRACE_{PEEK,POKE}USER (with an E) and has
logic in defs.h to make sure it's set sanely. Delete this old logic as
the defs.h takes care of it now.
* process.c: Delete PTRACE_PEEKUSR/PTRACE_POKEUSR defines.
* signal.c: Likewise.
* syscall.c: Delete PTRACE_PEEKUSR define.
* util.c: Likewise.
We used to explicitly list architectures where <sys/ptrace.h> from glibc
is known to provide a valid prototype for ptrace, and use a homegrown
replacement for all the rest. Situation seems to be better nowadays,
glibc is not the only libc available, so let's use ptrace prototype from
<sys/ptrace.h> by default, leaving the replacement for rare broken cases
if any.
* defs.h: Use ptrace prototype workaround iff
NEED_PTRACE_PROTOTYPE_WORKAROUND is defined.
There is only one place left in the code where strace guesses whether
libc provides LFS64 functions and structures. The most natural thing to
do there is to check for _LARGEFILE64_SOURCE - the macro provided by
glibc. Other libc implementations that provide nondegenerate LFS64
interfaces are expected to define this macro as well.
* defs.h (_LFS64_LARGEFILE): Remove.
* strace.c: Use _LARGEFILE64_SOURCE instead of _LFS64_LARGEFILE.
These decoders are referenced by most architectures, there were no
undefined references so far because _LFS64_LARGEFILE appears to be
always defined by glibc when _GNU_SOURCE is defined.
* file.c (sys_truncate64, sys_ftruncate64, sys_getdents64): Define
unconditionally.
* configure.ac (AC_CHECK_TYPES): Add struct flock64.
* desc.c (HAVE_F_SETLK64, HAVE_F_SETLKW64, HAVE_F_GETLK64): New macros.
[!HAVE_STRUCT_FLOCK64] (struct flock64): Define.
(printflock64): Define only if referenced by other code.
(sys_fcntl): Handle F_FREESP64, F_SETLK64, F_SETLKW64, and F_GETLK64 iff
these constants are defined and differ from their non-64bit versions.
* configure.ac (AC_CHECK_DECLS): Add PTRACE_PEEKUSER and
PTRACE_POKEUSER.
* defs.h: Define PTRACE_PEEKUSER and PTRACE_POKEUSER only if
they are not provided by <sys/ptrace.h>.
Reported by John Spencer.
IPPROTO_* macros are defined by <netinet/in.h>, no need to redefine
them.
* net.c (IPPROTO_EGP, IPPROTO_PUP, IPPROTO_IDP, IPPROTO_IGMP,
IPPROTO_RAW, IPPROTO_MAX): Remove.
Reported by John Spencer.
* process.c (sys_sched_setscheduler, sys_sched_getparam,
sys_sched_setparam): Use portable struct sched_param member name
sched_priority instead of glibc specific __sched_priority.
Reported by John Spencer.
SIGEV_THREAD_ID decoding requires access to an internal member of
struct sigevent. There seems to be no portable way to do it besides
adding a configure check.
* configure.ac (AC_CHECK_MEMBERS): Check for
struct sigevent._sigev_un._pad and struct sigevent.__pad.
* time.c (printsigevent): Use an appropriate struct sigevent member
to print thread id.
Reported by John Spencer.
Most of the code already uses long long types unconditionally.
* configure.ac: Remove the check for long long.
* file.c (printstat64): Remove HAVE_LONG_LONG checks, use %llu format
string unconditionally.
This change incorporates a partial copy instead
of using external libaio.h.
Why?
Because we want to properly decode 32-bit aio calls
by 64-bit strace. For that, we need more definitions than
libaio.h provides.
(These defs are not done yet, but will eventually be done).
Keeping our local 32-bit compat defs in sync with libaio.h
_without seeing libaio structs_ is hard/more bug prone.
A smaller benefit is that we don't need libaio installed.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
This helper returns two values (a string and an enum).
The caller prints the string. It's simpler to just print
the string in the caller itself. This eliminates
"return by reference" and more importantly, an intermediate
static string buffer for the string result.
Since function of the helper is different now,
it is renamed to tprint_lio_opcode().
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
The program may use a very large nr but supply either outright
invalid iocbpp[], or one with far fewer elements than nr.
We used to try reading iocbpp[i] until i == nr.
With this change, we stop on the first failure.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Added next_set_bit() function which finds the next set bit,
properly taking into account word size of the traced process.
Use it in decode_select() instead of fd_isset().
Also, properly round fdsize up to word size of traced process,
not to strace's word size.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
glibc in _FORTIFY_SOURCE mode raises SIGABRT when descriptor greater
or equal to FD_SETSIZE is passed to FD_ISSET. Select family syscalls,
however, can legitimately accept such descriptors. To overcome this
limitation, we have to replace FD_ISSET with an equivalent that imposes
no such restrictions.
* desc.c (fd_isset): New function.
(decode_select): Use it instead of FD_ISSET.
* desc.c (decode_select): Actually print arg[0] as int on entering
syscall. When arg[0] is negative, do not attempt to fetch and decode
descriptor sets on entering syscall, kernel will reject it anyway.
On exiting syscall, stop checking descriptor sets as soon as all
returned descriptors are found.
We used to allocate and fetch bit arrays using a sanitized
length, but then iterate over them with "j < arg[0]" condition,
where arg[0] is not sanitized. This segfaults if arg[0] is huge
or negative. This change fixes this.
Add test/select.c to capture the case.
Signed-off-by: Dr. David Alan Gilbert <dave@treblig.org>
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Since glibc-2.18~39 <sys/ptrace.h> defines ptrace_peeksiginfo_args
which collides with <linux/ptrace.h>.
* configure.ac: Check for `struct ptrace_peeksiginfo_args' in
<sys/ptrace.h>.
* process.c: Work around potential conflict between <sys/ptrace.h>
and <linux/ptrace.h> by redefining ptrace_peeksiginfo_args.
* signal.c: Likewise.
* syscall.c: Likewise.
* util.c: Likewise.
Signed-off-by: Ali Polatel <alip@exherbo.org>
Take #2 on mainlining strace support for ARC (last one was 4.6 based back
in March 2011), see
http://sourceforge.net/p/strace/mailman/message/27210168/
The syscall ABI is asm-generic/unistd.h based (so no legacy syscalls),
hence very similar to metag port.
test/* all seem to work well.
* linux/arc/ioctlent.h.in: New file.
* linux/arc/syscallent.h: Likewise.
* Makefile.am (EXTRA_DIST): Add linux/arc/ioctlent.h.in and
linux/arc/syscallent.h.
* configure.ac: Add ARC to the list of supported architectures.
* defs.h: Add ARC support.
* process.c (struct_user_offsets): Likewise.
* signal.c (sys_sigreturn): Likewise.
* syscall.c (print_pc, get_regset, get_regs, get_scno, get_syscall_args,
get_syscall_result, get_error): Likewise.
* util.c (change_syscall): Likewise.
Signed-off-by: Vineet Gupta <vgupta@synopsys.com>
Acked-by: Mike Frysinger <vapier@gentoo.org>
* file.c (sync_file_range_flags): New xlat structure.
(sys_sync_file_range, sys_sync_file_range2): New functions.
* linux/syscall.h (sys_sync_file_range, sys_sync_file_range2): New
prototypes.
* linux/dummy.h (sys_sync_file_range, sys_sync_file_range2): Remove.
* linux/mips/syscallent-o32.h: Set the number of sync_file_range
arguments to 7.
The fanotify_mark func takes a 64bit mask, so 32bit arches have to split
it up into two fields. When the syscall was added, it was listed as only
having 5 fields total (since that's correct for 64bit systems).
* linux/arm/syscallent.h: Set the number of fanotify_mark arguments to 6.
* linux/bfin/syscallent.h: Likewise.
* linux/hppa/syscallent.h: Likewise.
* linux/i386/syscallent.h: Likewise.
* linux/m68k/syscallent.h: Likewise.
* linux/metag/syscallent.h: Likewise.
* linux/microblaze/syscallent.h: Likewise.
* linux/mips/syscallent-o32.h: Likewise.
* linux/or1k/syscallent.h: Likewise.
* linux/powerpc/syscallent.h: Likewise.
* linux/s390/syscallent.h: Likewise.
* linux/sh/syscallent.h: Likewise.
* linux/sparc/syscallent.h: Likewise.
* linux/tile/syscallent1.h: Likewise.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Acked-by: Chris Metcalf <cmetcalf@tilera.com> [for tile]
In almost all cases, the parameter was 1 ("decode RT sigs too").
This change hardwires it to be 1.
The cases where it was 0 are, and why the change to "decode RT sigs too"
is fine:
IA64: 64-bit arch, all 64 signals fit in one word in any case.
S390[x]: sc.oldmask is a sigset_t on kernel side,
(sigset_t *)&sc.oldmask[0] does contain all 64 signals.
TILE: uc.uc_sigmask is a sigset_t on kernel side,
memcpy(&sigm, &uc.uc_sigmask, NSIG / 8) copies all 64 signals.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Replace print_sigset() with print_sigset_addr_len(),
which takes not only addr, but also len parameter.
This allows us to drop "do we need to print RT signals?" parameter,
and this fixes RT signals printing in many syscalls.
sys_epoll_pwait: print RT signals too, print sigmask size argument.
sys_sigprocmask: print_sigset -> print_sigset_addr_len(current_wordsize),
no change in functionality.
sys_sigpending: use print_sigset_addr_len(current_wordsize)
instead of open-coding it.
sys_rt_sigprocmask: use print_sigset_addr_len instead of open-coding it.
sys_rt_sigpending: ditto.
sys_rt_sigsuspend: ditto.
sys_rt_sigtimedwait: ditto.
do_signalfd: print_sigset -> print_sigset_addr_len. This fixes
RT signals printing (wasn't showing them before).
sys_ppoll: ditto.
copy_sigset_len() is folded into its only user, print_sigset_addr_len(),
and copy_sigset() is gone.
While at it, checked kernel sources and noted where kernel enforces
sigset_size == NSIG / 8 (== sizeof(kernel_sigset_t)),
and where it allows word-sized sigset_size ([rt_]sigpending).
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
In practice, we always copy as many bytes as syscall param says
(8, or 16 on mips). However, malicious program can call sigaction
with wrong sigset size. Such syscall will result in EINVAL,
but we (strace) end up copying 128 bytes (sizeof(sigset_t)),
which copyes some garbage from stack after struct sigaction.
Now we always copy NSIG / 8 bytes (which is 8 bytes, or 16 on mips).
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
ARM in fact has 64 signals (1..64), and NSIG should be 65
(as usual, rememebr that NSIG_libc == NSIG_kernel+1).
I carefully reviewed all usages of NSIG. In syscall.c,
the only usage is:
for (i = 0; i <= NSIG; i++)
if (strcasecmp(s, signame(i) + 3) == 0)...
which is safe even if NSIG is way too big - signame(i)
returns a well-formed string for any i.
In signal.c, memcpy(&sigset, &sc.sc_mask, NSIG / 8) is used by
IA64 and TILE code, so ARM change can't affect it. And final
usage is:
struct new_sigaction::unsigned long sa_mask[NSIG / sizeof(long)];
It will grow on ARM (and become correct in the process).
Its only use is
memcpy(&sigset, &sa.sa_mask, NSIG / 8);
printsigmask(&sigset, 1);
which used to copy garbage in high bits, now it will copy actual data.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
This includes decoding of 32-bit sigreturn by 64-bit strace,
which previously wasn't done.
Added a test for it.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
We were using uname to determine if userspace was 32 or 64bit.
This fails when we have a 64bit kernel and a 32bit userspace.
* configure.ac (powerpc*): Similar to x86, use a runtime test
to determine if we are 32 or 64bit.
Signed-off-by: Anton Blanchard <anton@samba.org>
Fix a number of differing signedness warnings when building on
powerpc.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
This eliminates some rare bugs, such as post-execve SIGTRAP
generation when we attach to a process, and it manages to finish
execve'ing before we set TRACEEXEC option to suppress that.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Before this change, the logic was as follows:
syscall entry:
get_scno:
syscall_mode = GPR2
scno = syscall_mode unless syscall_mode == -ENOSYS
(if -ENOSYS, scn is retrieved by decoding current insn)
fixup:
gpr2 = GPR2
syscall_mode = scno unless syscall_mode == -ENOSYS
if (gpr2 != syscall_mode) stray_entry
syscall exit:
get_res:
gpr2 = GRP2
fixup:
syscall_mode = scno unless syscall_mode == -ENOSYS
if (WAITEXECVE && gpr2 in (-ENOSYS, scno)) gpr2 = 0;
get_error:
gpr2 is retval
Entry fixup's if() can never trigger:
regardless whether GPR2 is -ENOSYS or not, syscall_mode is always
equal to GRP2 value there. So it can be removed.
On sysexit path, syscall mode is never used.
Therefore, syscall_mode variable is deleted. grp2 is read from
GPR2 register in get_scno, redundant read in entry fixup is removed.
As a result, entry fixup's s390 code block vanishes completely.
gpr2 variable is renamed s390_gpr2 to match the convention used
by other arches.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
popen_pid = vfork() does work correctly, but for a subtle reason
that wrong assignment of 0 happens in the child _first_,
and _then_ correct value overwrites it in the parent.
(And in a hyphothetical system where vfork = fork,
popen_pid wouldn't be shared, so it will also be ok.)
However, it's not necessary to be difficult.
This change makes it so that assignment is done only in parent.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
The cleanup sequence in error cases, and on normal code path
was nearly the same, no point in duplicating it.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
PTRACE_GETREGS was added to the ppc kernel in 2.6.23. In order to
provide backward compatibility for very old kernels, add a manual
fallback.
* syscall.c (powerpc_getreg, powerpc_getregs_old): New functions.
(get_regs): Call powerpc_getregs_old if PTRACE_GETREGS is not supported.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
"strace_child = pid" assignment was racing in NOMMU case
because of vfork (no way to know which process would win).
After this change, strace_child is set to nonzero only
in one process after [v]fork.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
We can use tcb::pid == 0 as an indicator of free tcb,
and we already have strace_child variable which holds
pid of our child, if any.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
I stumbeld over this case when I used "exec stace ...".
* strace.c (trace): Do not exit if we see an unknown pid in wait4.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
* strace.c (detach): Use goto's instead of excessive nesting.
Drop sigstop_expected and interrupt_done variables.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
* strace.c (detach): If TCB_IGNORE_ONE_SIGSTOP is set,
di not PTRACE_DETACH - wait for SIGSTOP first.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
* strace.c (detach): Warn if we see ECHILD from waitpid.
Explain in comments that we don't normally expect !WIFSTOPPED
either, and also that PTRACE_CONT failure isn't expected
(the "break" there is a "I'm confused, bailing out" code style)..
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
* strace.c (detach): Change return type from int to void.
Improve error meesages: show PID, tell exactly which operation fails.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
__WALL is available and working at least since 2.4 kernels: 10 years ago.
* strace (detach): Remove workarounds for ancient kernels lacking __WALL.
Add missing EINTR check.
(trace): Remove workarounds for ancient kernels lacking __WALL.
Remove redundant double error printout.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
V3: split SEIZE/!SEIZE code paths to reduce confusion.
Extensively comment every possible case.
Verified that all tests/detach* tests work in both SEIZE and !SEIZE
cases.
* strace.c (detach): If PTRACE_SEIZE API is in use, stop the tracee
using PTRACE_INTERRUPT instead of sending it a SIGSTOP.
In a subsequent waitpid loop, correctly wait and suppress SIGSTOP
on detach if PTRACE_INTERRUPT wasn't used, or wait for any ptrace
stop and detach without suppressing signals.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
* tests/detach-running: Disable /proc based checks when
/proc/self/status is not available.
* tests/detach-sleeping: Likewise.
* tests/detach-stopped: Likewise.
Run each test using a wrapper that does all necessary timeout checks.
* tests/run.sh: New file.
* tests/Makefile.am (EXTRA_DIST, LOG_COMPILER): Add it.
* tests/init.sh (check_strace, check_timeout): Remove.
* tests/detach-stopped: Update callers.
* tests/net: Likewise.
* tests/ptrace_setoptions: Likewise.
* tests/qual_syscall: Likewise.
* tests/stat: Likewise.
* tests/strace-f: Likewise.
Current implementation of -qq does not allow suppressing exit status
messages without suppressing signal delivery messages, which is not
good. There is a traditional "-e signal=none" syntax that can be used
to suppress all signal delivery messages.
This partially reverts commit v4.7-222-g01997cf.
* strace.c (trace): Do not suppress signal delivery messages with -qq.
* strace.1: Update documentation about -qq option.
As of glibc 2.16, "struct siginfo" is no longer supported,
and "siginfo_t" must be used instead.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
These flags support functionality in mmap() that has not been
pushed back to the community, and which may or may not eventually
end up being the final community model. In the interim, having
these flags unconditionally present for "#ifdef TILE" just means
that the TILE build breaks if using the community versions of
the kernel and glibc, so just revert the code until such time
as it may end up in the community.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
The standard name people have adopted is "COPYING" rather than
"COPYRIGHT". Use that as a lot of license scanning tools look for it.
* COPYRIGHT: Rename to ...
* COPYING: ... this.
* Makefile.am (EXTRA_DIST): Rename COPYRIGHT to COPYING.
* README: Likewise.
* strace.spec (%files): Likewise.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
On NOMMU systems in "strace PROG" case, we have no way to attach to
the tracee before it calls execve unless in -D mode. That is, the
first execve call is very likely to be missed, and setting
hide_log_until_execve just results to empty log.
* strace.c (init) [NOMMU_SYSTEM]: Do not set hide_log_until_execve
unless in -D mode.
* strace.c [HAVE_PRCTL]: Include sys/prctl.h.
(startup_child) [HAVE_PRCTL && PR_SET_PTRACER && PR_SET_PTRACER_ANY]:
In -D mode, allow tracing the process that is going to become the
tracee.
Tested with "./strace [-D] [-q] [-bexecve] env true",
all cases seem to work.
* defs.h: Declare new variable: bool hide_log_until_execve.
* strace.c: Define hide_log_until_execve.
Rename skip_startup_execve to skip_one_b_execve.
(startup_child): Do not set skip_one_b_execve = 1 here.
(init): If "strace PROG" case (as opposed to "strace -pPID"),
set skip_one_b_execve and hide_log_until_execve to 1.
(trace): Don't print signal messages if hide_log_until_execve == 1.
* syscall.c (trace_syscall_entering):
Skip syscall printing if hide_log_until_execve == 1.
Reset hide_log_until_execve if we enter execve syscall.
(trace_syscall_exiting): Skip syscall printing if hide_log_until_execve == 1.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Even with the -q flag specified, tracing output is still mixed
with messages about signals and process exit status, which is
often irrelevant. Allow the -q option to be repeated to force
the suppression of signals / exit status info too.
* defs.h: Change 'qflag' from 'bool' to 'unsigned int'.
* strace.1: Document ability to repeat '-q' option.
* strace.c: Allow '-q' to be repeated to quieten process
exit status and signal messages.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
* linux/avr32/syscallent.h: Set the number of sync_file_range arguments
to 6.
* linux/bfin/syscallent.h: Likewise.
* linux/hppa/syscallent.h: Likewise.
* linux/i386/syscallent.h: Likewise.
* linux/m68k/syscallent.h: Likewise.
* linux/microblaze/syscallent.h: Likewise.
* linux/mips/syscallent-o32.h: Likewise.
* linux/or1k/syscallent.h: Likewise.
* linux/s390/syscallent.h: Likewise.
* linux/sh/syscallent.h: Likewise.
* linux/powerpc/syscallent.h: Change sync_file_range to
sync_file_range2, set the number of sync_file_range2 arguments to 6.
* linux/tile/syscallent.h: Change sync_file_range to sync_file_range2,
set the number of sync_file_range2 arguments to 4.
* linux/tile/syscallent1.h: Change sync_file_range to sync_file_range2.
update handler.
* linux/xtensa/syscallent.h: Likewise.
The recent commit (2690fadc8b) shuffled
the mips headers around causing it to check the ABI defines before it
included the header which set those up. Now all mips builds fail with:
In file included from count.c:36:0:
defs.h:48:4: error: #error Unsupported _MIPS_SIM
* defs.h [MIPS]: Move sgidefs.h above _MIPS_SIM check.
The raw llseek syscall is called _llseek from userland, so fix the
naming of it for metag to allow the stat test to pass.
* linux/metag/syscallent.h: Rename llseek syscall to _llseek.
Signed-off-by: James Hogan <james.hogan@imgtec.com>
The metag architecture also uses dedicated syscalls and doesn't need to
multiplex ipc and socket subcalls.
* linux/syscall.h [METAG]: Do not define SYS_socket_subcall and
SYS_ipc_subcall.
Signed-off-by: James Hogan <james.hogan@imgtec.com>
The number of pread and pwrite arguments depends on architecture and
personality. For 64bit and ilp32 it equals to 4, for unaligned
32bit it equals to 5, and for aligned 32bit it equals to 6.
* linux/aarch64/syscallent1.h: Change the number of pread and pwrite
arguments to 4.
* linux/alpha/syscallent.h: Likewise.
* linux/mips/syscallent-n32.h: Likewise.
* linux/mips/syscallent-n64.h: Likewise.
* linux/s390x/syscallent.h: Likewise.
* linux/sh64/syscallent.h: Likewise.
* linux/tile/syscallent.h: Likewise.
* linux/x32/syscallent.h: Likewise.
* linux/x86_64/syscallent.h: Likewise.
* linux/or1k/syscallent.h: Change the number of pread and pwrite
arguments to 5.
The number of preadv and pwritev arguments depends on architecture and
personality. For 64bit and ilp32 it equals to 4, for unaligned 32bit
it equals to 5, and for aligned 32bit it equals to 6.
* linux/aarch64/syscallent1.h: Change the number of preadv and pwritev
arguments to 4.
* linux/alpha/syscallent.h: Likewise.
* linux/ia64/syscallent.h: Likewise.
* linux/mips/syscallent-n32.h: Likewise.
* linux/mips/syscallent-n64.h: Likewise.
* linux/s390x/syscallent.h: Likewise.
* linux/sh64/syscallent.h: Likewise.
* linux/tile/syscallent.h: Likewise.
* linux/x32/syscallent.h: Likewise.
* linux/x86_64/syscallent.h: Likewise.
* linux/powerpc/syscallent.h: Change the number of preadv and pwritev
arguments to 6.
* linux/sh/syscallent.h: Likewise.
* linux/xtensa/syscallent.h: Likewise.
The number of ftruncate64/truncate64 arguments depends on architecture
and personality. For 64bit and ilp32 it equals to 2, for unaligned
32bit it equals to 3, and for aligned 32bit it equals to 4.
* linux/or1k/syscallent.h: Change the number of ftruncate64 and
truncate64 arguments to 3.
* linux/sh/syscallent.h: Likewise.
* linux/sh64/syscallent.h: Change ftruncate64 and truncate64 handlers.
* linux/sparc/syscallent.h: Change ftruncate64 handler, change the
number of its arguments to 3.
* linux/xtensa/syscallent.h: Change the number of ftruncate64 and
truncate64 arguments to 4.
The number of fallocate arguments depends on architecture and
personality. For 64bit and ilp32 it equals to 4,
for 32bit it equals to 6.
* linux/aarch64/syscallent1.h: Change the number of fallocate arguments
to 4.
* linux/alpha/syscallent.h: Likewise.
* linux/mips/syscallent-n32.h: Likewise.
* linux/mips/syscallent-n64.h: Likewise.
* linux/s390x/syscallent.h: Likewise.
* linux/sh64/syscallent.h: Likewise.
* linux/tile/syscallent.h: Likewise.
* linux/x32/syscallent.h: Likewise.
* linux/x86_64/syscallent.h: Likewise.
* linux/microblaze/syscallent.h: Set fallocate handler and flags.
The number of fadvise64 arguments depends on architecture and
personality. For 64bit and ilp32 it equals to 4, for unaligned
32bit it equals to 5, and for aligned 32bit it equals to 6.
The number of fadvise64_64 arguments is 4 for 64bit and ilp32, for
unaligned 32bit and most of aligned 32bit it equals to 6, and for
mips o32 it equals to 7.
* linux/alpha/syscallent.h: Set fadvise64 handler.
* linux/s390/syscallent.h: Likewise.
* linux/hppa/syscallent.h: Set fadvise64_64 handler, change the number
of syscall arguments to 6.
* linux/ia64/syscallent.h: Change the number of fadvise64 arguments
to 4.
* linux/mips/syscallent-n32.h: Likewise.
* linux/tile/syscallent1.h: Change the number of fadvise64 arguments
to 5.
* linux/mips/syscallent-o32.h: Change the number of fadvise64_64
arguments to 7.
* linux/s390x/syscallent.h: Change fadvise64_64 handler. Set fadvise64
handler. Change numbers of fadvise64 and fadvise64_64 arguments to 4.
* linux/sh64/syscallent.h: Change fadvise64_64 handler.
Change numbers of fadvise64 and fadvise64_64 arguments to 4.
* linux/sparc/syscallent.h: Set fadvise64 and fadvise64_64 handlers,
change numbers of their arguments to 5 and 6.
* linux/x86_64/syscallent.h: Change fadvise64 handler.
linux/mips/syscallent.h consisted of several parts that had little in
common. Split them into separate files to ease maintenance.
* linux/mips/syscallent-compat.h: New file.
* linux/mips/syscallent-n32.h: Likewise.
* linux/mips/syscallent-n64.h: Likewise.
* linux/mips/syscallent-o32.h: Likewise.
* linux/mips/syscallent.h: Include them.
* Makefile.am (EXTRA_DIST): Add them.
The number of readahead arguments depends on architecture and
personality. For 64bit and ilp32 it equals to 3, for unaligned 32bit it
equals to 4, and for aligned 32bit it equals to 5.
* linux/aarch64/syscallent1.h: Set number of readahead arguments to 3.
* linux/alpha/syscallent.h: Likewise.
* linux/ia64/syscallent.h: Likewise.
* linux/s390x/syscallent.h: Likewise.
* linux/sh64/syscallent.h: Likewise.
* linux/x32/syscallent.h: Likewise.
* linux/x86_64/syscallent.h: Likewise.
* linux/or1k/syscallent.h: Set number of readahead arguments to 4.
* linux/sh/syscallent.h: Likewise.
* linux/tile/syscallent.h: Likewise.
* linux/xtensa/syscallent.h: Set number of readahead arguments to 5.
The Xtensa architecture also uses dedicated syscalls and doesn't
need to multiplex ipc and socket subcalls.
* linux/syscall.h [XTENSA]: Do not define SYS_socket_subcall and
SYS_ipc_subcall.
Signed-off-by: Chris Zankel <chris@zankel.net>
Newer Linux architectures don't support the stat/stat64 syscalls.
Instead they use fstatat() with AT_FDCWD and an extra flags argument.
Support seeing this output in the 'strace -efile' test.
As part of this change, use "grep -E -x" syntax consistently for
all stat tests, since the number of \(foo\)\? expressions was becoming
pretty unwieldy.
* tests/stat: Update stat/stat64 check regexp to handle architectures
that use fstatat instead. Use "grep -E -x" syntax consistently.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
ARM OABI and ARM EABI have different function parameters passing rules,
see commit v4.6-11-g7a5b08f for details.
* util.c (printllval): Do not align argument number in case of ARM OABI.
* file.c (sys_fadvise64_64): Likewise.
The mips/o32 ABI and xtensa arch also do 64bit aligning, so add it to the
printllval list for this.
Normally for sys_fadvise64_64 we'd handle the same list of arches, but
mips/o32 ABI is funky and doesn't shift -- it has 7 args. So just add
xtensa to it.
* file.c (sys_fadvise64_64): Add XTENSA to the shifted list.
* util.c (printllval): Add LINUX_MIPSO32 and XTENSA to the shifted list.
The tile arch just has the dedicated syscalls, so disable
the logic for parsing subcalls.
* linux/syscall.h [TILE]: Do not define SYS_socket_subcall and
SYS_ipc_subcall.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
The Blackfin arch does not have a socketcall or ipc subcall
(it has dedicated syscalls broken out), so disable the logic for it.
* linux/syscall.h [BFIN]: Do not define SYS_socket_subcall and
SYS_ipc_subcall.
Since the length fields with the ubi volnames are signed 16bit values,
make sure we clamp that number to the size of the buffer we've allocated
on the stack to prevent buffer overflows.
* mtd.c (ubi_ioctl): Clamp length to string_quote to 0/UBI_MAX_VOLUME_NAME.
Check the return of string_quote and tweak the output accordingly.
Been playing with UBI of late and it'd help if I could see what it was
doing. Not entirely sure about the decoding of UBI_IOCVOLUP -- it takes
a pointer to a 64bit value, not a strict.
* util.c (MAX, MIN): Move to ...
* defs.h (MAX, MIN): ... here.
(CLAMP): Define.
(ubi_ioctl): New prototype.
* ioctl.c (ioctl_decode): Call ubi_ioctl when code is 'o' or 'O'.
* mtd.c: Include ubi user API header.
(ubi_volume_types): New enum decode.
(ubi_volume_props): Likewise.
(ubi_ioctl): Define.
* linux/ubi-user.h: Import from linux 3.8.
I used these to develop the mtd/ubi ioctl decoders.
* test/.gitignore: Add mtd/ubi.
* test/Makefile: Declare all tests in a PROGS var.
Declare build targets .PHONY.
(all): Depend on $(PROGS) instead of hardcoded list.
(clean): Remove $(PROGS) instead of hardcoded list.
* test/mtd.c: New file.
* test/ubi.c: Likewise.
These funcs use things like wait/write/read/strcmp but sometimes
don't include the right header for them.
* test/Makefile: Add -Wall to CFLAGS.
* test/clone.c: Include unistd.h.
* test/fork.c: Include sys/wait.h.
* test/sig.c: Include unistd.h.
* test/sigkill_rain.c: Include sys/wait.h.
* test/vfork.c: Include sys/wait.h.
* test/wait_must_be_interruptible.c: Include string.h
* linux/ia64/syscallent.h: Undefine sys_oldlstat, sys_lstat and
sys_lstat64 before redefining them to printargs.
Reported-by: Mike Frysinger <vapier@gentoo.org>
* desc.c (printflock) [X32]: Add special handling required for
this architecture with sizeof(long) < sizeof(off_t).
* file.c [X32] (struct stat64): Add __attribute__((packed)).
[X32] (HAVE_STAT64): Define.
(printstat) [X32]: Redirect to printstat64.
(printstat64) [X32]: Use "struct stat" instead of "struct stat64".
[X32] (realprintstat64): Rename to printstat64_x32.
(sys_stat64, sys_fstat64) [X32]: Remove second definitions of these
functions. Call printstat64_x32 instead of printstat64
* linux/x32/syscallent.h: Fix handlers for truncate and ftruncate.
The metag ABI has stacks growing upwards so clone expects the stack
argument to point at the bottom of the stack instead of the top.
* test/threaded_execve.c [__metag__]: Define clone2 differently
to avoid a segfault at runtime.
Signed-off-by: James Hogan <james.hogan@imgtec.com>
* test/skodic.c (main): Don't use MAP_FIXED since valid virtual addresses
vary between architectures (as far as I can tell the use of MAP_FIXED is
not relevant to the test). Also don't assume the file desriptor returned
by open call is 3 when passing it to mmap.
Signed-off-by: James Hogan <james.hogan@imgtec.com>
This is a 64bit arch w/no personalities, so fix printing.
It can now trace a simple call like readahead:
$ cat test.c
main(){readahead(-1, (unsigned long long)1 << 50, 0);}
$ gcc test.c
$ ./strace -ereadahead ./a.out
readahead(-1, 1125899906842624, 0) = -1 EBADF (Bad file descriptor)
* util.c (printllval): Handle S390X define.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Building with automake-1.13 throws an error:
configure.ac:6: error: 'AM_CONFIG_HEADER': this macro is obsolete.
You should use the 'AC_CONFIG_HEADERS' macro instead.
* configure.ac (AM_CONFIG_HEADER): Rename to AC_CONFIG_HEADERS.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
* system.c (sys_sram_alloc): The 2nd argument of sram_alloc syscall
is a bit field, not a specific value, so decode it as such.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
This makes output formatting more consistent with pread()/pwrite(),
which print their offset parameters in decimal.
* desc.c (sys_io_submit): Change offset output format from %llx to %lld.
Signed-off-by: Zev Weiss <zev@bewilderbeest.net>
The file linux/ioctlent.h is generated in the build directory, so
"../ioctlent.h" won't find it.
* linux/aarch64/ioctlent1.h: Remove "../" from include file name.
* linux/powerpc/ioctlent1.h: Likewise.
* linux/tile/ioctlent1.h: Likewise.
* linux/x32/ioctlent1.h: Likewise.
* linux/x86_64/ioctlent1.h: Likewise.
* linux/x86_64/ioctlent2.h: Likewise.
Signed-off-by: Andreas Schwab <schwab@suse.de>
strace already has a mechanism to use fopen64 for output when the 64-bit
version of fopen is available on 32-bit architectures. Apply this
mechanism for other three functions to make strace fully adopted for
64-bit types.
* strace.c (struct_stat, stat_file, struct_dirent, read_dir,
struct_rlimit, set_rlimit): New macros.
(startup_attach): Use read_dir.
(startup_child): Use struct_stat and stat_file.
(main): Use struct_rlimit and set_rlimit.
struct dirent from libc should not be used for umove'ing into because it
contains fixed size d_name.
* file.c (printdir): Rename to print_old_dirent.
[SH64]: Decode using struct kernel_dirent.
[!SH64]: Decode using an open-coded struct with 32-bit d_ino and d_off.
(sys_readdir): Update.
sendfile parser used to decode off_t* parameter as a pointer to host
off_t type. With this change, it is decoded as a pointer to target long
type.
* io.c (print_off_t): New function.
(sys_sendfile): Use it.
* linux/aarch64/syscallent1.h: Use sys_sendfile64 for sendfile decoding.
* linux/tile/syscallent.h: Likewise.
* linux/x32/syscallent.h: Likewise.
* linux/x86_64/syscallent.h: Likewise.
* linux/hppa/syscallent.h: Use sys_sendfile64 for sendfile64 decoding.
* linux/metag/syscallent.h: Correct sendfile syscall name.
* linux/or1k/syscallent.h: Likewise.
* configure.ac: When libaio.h is available, check for
struct iocb.u.c.flags, IO_CMD_PWRITE and IO_CMD_PWRITEV.
* desc.c (print_common_flags): Check for HAVE_STRUCT_IOCB_U_C_FLAGS.
(sys_io_submit): Check for HAVE_DECL_IO_CMD_PWRITE and
HAVE_DECL_IO_CMD_PWRITEV.
Busybox version of timeout doesn't have --version option.
Relax check_prog to handle those cases.
* tests/init.sh (check_prog): Use type builtin to check program
availability.
Signed-off-by: Maxin B. John <maxin.john@enea.com>
* syscall.c [I386 || ARM || OR1K || METAG] (ARCH_REGS_FOR_GETREGSET):
New macro.
(get_regset): Implement for AARCH64, METAG, OR1K and X32.
(get_regs) [AARCH64 || METAG || OR1K || X32]: Use it.
count_syscall() was calling setitimer/getitimer once in order to find
smallest "tick" OS uses in time accounting, in order to use it
for syscalls which apparently spent less than that time in syscall.
The code assumed that this "tick" is not zero... but it is zero
on linux-3.6.11. Which means that this hack doesn't work...
At least this change prevents this measurement from being done
_repeatedly_, by initializing one_tick to -1, not 0.
While at it, added comments in count_syscall() explaining what we are doing.
Optimized call_summary_pers() a bit, by eliminating redundant tv -> float
conversions, and prevented 0.0/0.0 which was resulting in "% time"
being shown as "-nan" if total CPU time spent was 0.000000
(try "strace -c /bin/true").
The code seems to seriously underestimate CPU usage:
"strace -c ls -lR /usr/share >/dev/null" shows total time spent
in syscalls to be only ~10..20% of what "time ls -lR /usr/share >/dev/null"
shows.
It might be useful to have a mode where we show wall clock time
spent in syscalls, not CPU time. It might also be more accurate.
text data bss dec hex filename
245019 676 5708 251403 3d60b strace_old
244923 684 5676 251283 3d593 strace
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
We don't call setlocale, thus we always use C locale.
But libc supports various other locales, and therefore
its ctype interface is general and at times inefficient.
For example, in glibc these macros result in function call,
whereas for e.g. isprint(c) just c >= ' ' && c <= 0x7e
suffices.
By open-coding ctype checks (we have only 4 of them)
we avoid function calls, we get smaller code:
text data bss dec hex filename
245127 680 5708 251515 3d67b strace_old
245019 676 5708 251403 3d60b strace
and we don't link in ctype tables (beneficial for static builds).
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
PAGE_SHIFT couldn't be reliably obtained at compile time,
thanks to Chris Metcalf for the hint.
* mem.c: Do not include <sys/user.h>.
[SH64] Do not include <asm/page.h>.
(get_pagesize): New function.
(sys_mmap_pgoff, sys_old_mmap_pgoff): Use it.
While at it, added a small optimization of not remembering
the path twice if it happens to be the same.
text data bss dec hex filename
245111 680 10860 256651 3ea8b strace_old
245075 680 9804 255559 3e647 strace
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
or1k was missing TM on many memory-related syscalls
sys_lookup_dcookie is 3-arg on 64-bit arches, and isn't TF
sys_recvmsg is 3-arg on all arches
sys_nfsservctl is 3-arg on all arches
sys_timerfd_create is 2-arg on all arches
sys_[f]truncate64 is 4-arg or 3-arg, never 5-arg
truncate64 is TF
sys_[l]lseek is TD
fstat[64] is TD
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Add support for Imagination Technologies Meta architecture (the
architecture/ABI is usually referred to as metag in code). The Meta
Linux kernel port is in the process of being upstreamed for v3.9 so it
uses generic system call numbers.
sys_lookup_dcookie writes a filename to buffer argument, so I've set
TF flag.
nfsservctl appears to be set to sys_ni_syscall in asm-generic/unistd.h
so I've left it blank.
truncate64/ftruncate64/pread64/pwrite64/readahead have unaligned 64bit
args which are packed tightly on metag, so less arguments on metag.
fchdir/llseek takes a file descriptor so s/TF/TD/
sync_file_range has 2 64bit args so uses 6 args, so s/4/6/
timerfd_create/msgget/msgctl/msgrcv/semget/segtimedop/semop/shmget/
shmctl/shmat/shmdt/recvmsg/migrate_pages have different number of args.
oldgetrlimit is just getrlimit for metag.
add TM flag to various memory syscalls.
metag doesn't directly use sys_mmap_pgoff for mmap2.
prlimit64/process_vm_readv/process_vm_writev take a pid so add TP flag.
fanotify_init doesn't appear to take a file descriptor so remove TD.
Add kcmp syscall.
Signed-off-by: James Hogan <james.hogan@imgtec.com>
Cc: Christian Svensson <blue@cmd.nu>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Without this fix the tilegx build fails when it hits the new #error
about SIZEOF_LONG > 4.
* util.c (printllval): Fix printing of long long values on TILE.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
To not waste an option letter for just one trick,
extend -b to take a parameter:
"on which syscalls do you want to detach?".
Currently supports only execve.
While at it, fixed (by removing non-Linux and stale info)
and extended manpage text about -f.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Cleanup sloppy error handling.
First, EFAULT kind of errors from process_vm_readv by itself is not
something unusual, so a warning message will not be issued unless a
short read is detected.
Second, clients of umoven and umovestr are not prepared to detect and
handle short reads that can happen in these functions. The most safe
way to handle them is to return an error code.
* util.c (umoven, umovestr): Cleanup handling of errors coming from
process_vm_readv and PTRACE_PEEKDATA.
This fixes regression introduced by commit v4.7-111-g9cbc15b.
* syscall.c: Merge all nsyscalls* enums into single enum.
Likewise merge nerrnos*, nsignals*, and nioctlents* into single enums.
There is no need to print literal strings through "%s".
Only untrusted strings such as filenames need that.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
umoven() uses process_vm_readv() when available but it returns the
return value of that syscall, which is the number of bytes copied,
while its callers expect it to simply return zero on success.
It was causing syscalls that take a user-space argument to print
the abbreviated version, e.g.:
epoll_ctl(5, EPOLL_CTL_ADD, 10, {...})
Instead of:
epoll_ctl(5, EPOLL_CTL_ADD, 10, {EPOLLIN, {u32=10, u64=10}})
* util.c (umoven): Make umove[n] report success as 0, not >=0,
stop returning "success" on partial reads.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
With -D, strdup'ing of pathname is necessary only on NOMMU.
Don't set skip_startup_execve to 1 if NOMMU and not in daemonized mode
(try "strace [-D] -b env echo HI" to see whether we detach on correct execve).
Fix test_ptrace_FOO shortcuts on NOMMU to always assume success
and _properly_ set all variables.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
This is a 14 year old bug (!).
It wasn't biting us merely because outstr[80] was static, thus ended up
in bss and whatever was after it "accidentally" provided the NUL byte.
When dumpstr was changed to use on-stack buffer, the bug reared its ugly head.
This is a rewrite which is smaller and should be significantly faster
for _long_ strings.
text data bss dec hex filename
244627 680 10860 256167 3e8a7 strace.t9/strace
244563 680 10860 256103 3e867 strace.ta/strace
* util.c (dumpstr): Rewrite to be faster and smaller.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
MAX_QUALS was 2048, even though most arches used less than 500 entries
in it. MAX_QUALS had to be maintained by hand to be higher than syscall
count. It also limited the highest possible fd to track.
This change makes qual_flagsN[] arrays start sized to the required minimum
(number of syscalls) and grow dynamically if user requested
-e read=BIGNUM. As a precaution, BIGNUM should be < 2^15, but this limit
can be raised with no cost for normal strace invocations.
qual_flags is now a define to qual_vec[current_personality].
As a bonus, this patch aliases sysent, errnoent, signalent, ioctlent
names in one-personality arches to their corresponding <foo>0 arrays,
removing one indirection level.
text data bss dec hex filename
244471 700 12928 258099 3f033 strace.t7/strace
244627 680 10860 256167 3e8a7 strace.t8/strace
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
This is a preparatory mass replace patch with no code changes.
The future change will need to typedef sysent to sysent0,
which results in compile failures when "struct sysent" string
gets mangled into "struct sysent0".
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
We have hundreds of uses of these macros.
Result is more efficient and 1.1 kbyte shorter code:
text data bss dec hex filename
245579 700 12928 259207 3f487 strace.t5/strace
244471 700 12928 258099 3f033 strace.t6/strace
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
By adding tcp->s_ent pointer tot syscall table entry,
we can replace sysent[tcp->scno] references by tcp->s_ent.
More importantly, we may ensure that tcp->s_ent is always valid,
regardless of tcp->scno value. This allows us to drop
SCNO_IS_VALID(tcp->scno) checks before we access syscall
table entry.
We can optimize (qual_flags[tcp->scno] & QUAL_foo) checks
with a similar technique.
Resulting code shrink:
text data bss dec hex filename
245975 700 19072 265747 40e13 strace.t3/strace
245703 700 19072 265475 40d03 strace.t4/strace
* count.c (count_syscall): Use cheaper SCNO_IN_RANGE() check.
* defs.h: Add "int qual_flg" and "const struct sysent *s_ent"
to struct tcb. Remove "int u_nargs" from it.
Add UNDEFINED_SCNO constant which will mark undefined scnos
in tcp->qual_flg.
* pathtrace.c (pathtrace_match): Drop SCNO_IS_VALID check.
Use tcp->s_ent instead of sysent[tcp->scno].
* process.c (sys_prctl): Use tcp->s_ent->nargs instead of tcp->u_nargs.
(sys_waitid): Likewise.
* strace.c (init): Add compile-time check that DEFAULT_QUAL_FLAGS
constant is consistent with init code.
* syscall.c (decode_socket_subcall): Use tcp->s_ent->nargs
instead of tcp->u_nargs. Set tcp->qual_flg and tcp->s_ent.
(decode_ipc_subcall): Likewise.
(printargs): Use tcp->s_ent->nargs instead of tcp->u_nargs.
(printargs_lu): Likewise.
(printargs_ld): Likewise.
(get_scno): [MIPS,ALPHA] Use cheaper SCNO_IN_RANGE() check.
If !SCNO_IS_VALID, set tcp->s_ent and tcp->qual_flg to default values.
(internal_fork): Use tcp->s_ent instead of sysent[tcp->scno].
(syscall_fixup_for_fork_exec): Remove SCNO_IS_VALID check.
Use tcp->s_ent instead of sysent[tcp->scno].
(get_syscall_args): Likewise.
(get_error): Drop SCNO_IS_VALID check where it is redundant.
(dumpio): Drop SCNO_IS_VALID check where it is redundant.
Use tcp->s_ent instead of sysent[tcp->scno].
(trace_syscall_entering): Use (tcp->qual_flg & UNDEFINED_SCNO) instead
of SCNO_IS_VALID check. Use tcp->s_ent instead of sysent[tcp->scno].
Drop SCNO_IS_VALID check where it is redundant.
Print undefined syscall name with undefined_scno_name(tcp).
(trace_syscall_exiting): Likewise.
* util.c (setbpt): Use tcp->s_ent instead of sysent[tcp->scno].
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
ARM in fact _is_ one personality.
We had two personalities for it because it has a handful of
syscalls with huge scnos (0x000f00xx).
Extending syscall table to have [0x000f0005] index is of course
not a good idea.
Someone decided to handle that by having a separate personality
just for these syscalls.
But multi-personality arch does a bit more work in other parts.
This patch is another alternative: "move" 0x000f00nn syscalls
down to the entries just above last ordinary syscall,
by manipulating scno if it falls into the 0x000f00xx range.
In order to not worsen genuine undefined scnos' printing,
the code remaps scno back to actual value before printing
"syscall_NNN" string.
* defs.h: Remove multi-reprsonality defines from ARM.
* syscall.c (shuffle_scno): New function.
(undefined_scno_name): New function.
(get_scno): [ARM] Replace personality setting with scno shuffling.
(trace_syscall_entering): Print unknown syscall name using
undefined_scno_name().
(trace_syscall_exiting): Likewise.
* linux/arm/syscallent.h: Add ARM specific syscalls at the end.
* linux/arm/errnoent1.h: Deleted.
* linux/arm/ioctlent1.h: Deleted.
* linux/arm/signalent1.h: Deleted.
* linux/arm/syscallent1.h: Deleted.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
While at it, fix some comments which say
"we use i386 syscalls/ioctls/errnos" but in reality
common ones are used.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
With constant current_wordsize == 4 and 32-bit longs,
gcc was spooked by "1ul << (8 * current_wordsize)" = "1ul << 32".
Make such places conditional on SIZEOF_LONG > 4.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
pathname[] was getting destroyed, execve of garbage pathname
failing, and to top it off, the tracer's stack was also
smashed and trecer segfaulted.
* strace.c (exec_or_die): New function.
(startup_child): Don't use pathname[] contents after vfork,
make a malloced copy instead. Explain "NOMMU + -D bug"
and how we work around it.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
And while using it, I discovered that -D doesn't work too well
on NOMMU. Added a comment about it.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Previous code merges too many similar, but different ways
of decoding mmap. For example, sys_old_mmap is "params in memory"
API... except SH[64], where it is "params in regs",
i.e. what sys_mmap ("new mmap") function does on other arches!
It's much simpler when every mmap handler has same API regardless
of arch. Where API means whether params are in regs or in memory,
and whether offset is in bytes, pages, or 4k blocks.
Then we just insert correct function pointers into
arch syscall tables.
It turns out there are four common mmap APIs over
all architectures which exist in Linux kernel,
and one outlier for S390.
A number of mmap decoders were plain wrong in arch tables.
For example, BFIN has no old_mmap. It returns ENOSYS.
I checked kernel sources for all arches nad fixed the tables.
There was dead code for x86_64 for old_mmap:
x86_64 has no old_mmap.
* mem.c: Refactor mmap functions so that we have five mmap syscall
handlers, each with the fixed API (not varying by arch).
* pathtrace.c (pathtrace_match): Adjust sys_func == mmap_func checks.
* linux/syscall.h: Declare new mmap syscall handler functions.
* linux/arm/syscallent.h: mmap2 is sys_mmap_pgoff.
* linux/avr32/syscallent.h: mmap is sys_mmap_pgoff.
* linux/bfin/syscallent.h: old_mmap is ENOSYS, mmap2 is sys_mmap_pgoff.
* linux/hppa/syscallent.h: mmap2 is sys_mmap_4koff.
* linux/i386/syscallent.h: mmap2 is sys_mmap_pgoff.
* linux/ia64/syscallent.h: mmap2 is sys_mmap_pgoff.
* linux/m68k/syscallent.h: mmap2 is sys_mmap_pgoff.
* linux/microblaze/syscallent.h: old_mmap is sys_mmap, mmap2 is sys_mmap_pgoff.
* linux/mips/syscallent.h: mmap is sys_mmap_4kgoff.
* linux/or1k/syscallent.h: mmap2 is sys_mmap_pgoff.
* linux/powerpc/syscallent.h: mmap2 is sys_mmap_4kgoff.
* linux/s390/syscallent.h: mmap2 is sys_old_mmap_pgoff.
* linux/s390x/syscallent.h: mmap is sys_old_mmap and thus has 1 arg.
* linux/sh/syscallent.h: old_mmap2 is sys_mmap, mmap2 is sys_mmap_4koff.
* linux/sh64/syscallent.h: Likewise.
* linux/sparc/syscallent1.h: mmap is TD|TM.
* linux/tile/syscallent1.h: mmap2 is sys_mmap_4koff.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
No wonder that it is unused.
It's code looked quite questionable.
* mem.c (sys_mmap64): Remove this function.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
While looking at mmap mess, did experimenting in order
to figure out what gets used when.
Tried building armv4tl, armv5l, armv6l, mips, mipsel, i686,
x86_64 and none of they have long long off_t,
which isn't suprprising: we aren't using glibc defines
which enable that.
Moreover, we SHOULD NOT use off_t in syscall decode!
Its size depends on libc, not on arch! I.e. it is essentially
unpredictable and can even in theory vary on the same arch
with different libc.
We should use longs or long longs, in a way which matches
architectural ABI for the given syscall. There are usually
*at most* two permutations, no need to add yet another variable
(sizeof(off_t)) to the mix.
This change removes almost all HAVE_LONG_LONG_OFF_T conditionals,
which will reveal further possible simplifications.
* mem.c: Remove code conditional on HAVE_LONG_LONG_OFF_T.
As a result, never remap sys_mmap64 to sys_mmap.
(print_mmap): Compile unconditionally.
(sys_old_mmap): Compile unconditionally.
(sys_mmap): Compile unconditionally.
* io.c (sys_sendfile): Add a FIXME comment.
* file.c: Remove code conditional on HAVE_LONG_LONG_OFF_T.
As a result, never remap sys_*stat64 to sys_*stat etc.
(sys_truncate): Compile unconditionally.
(realprintstat): Likewise.
(sys_stat): Likewise.
(sys_fstat): Likewise.
(sys_lstat): Likewise.
* desc.c (printflock): Likewise.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
* mem.c (sys_mmap): Ensure unsigned expansion of tcp->u_arg[5].
Add page shift of offset for I386.
Use tcp->ext_arg[5] as offset for X32.
(sys_old_mmap): [X32] Remove this function, X32 doesn't use is.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
* mem.c: Move "define sys_mmap64 sys_mmap" from the top
to the only place it affects.
(print_mmap): Make offset argument unsigned, for safer implicit conversions.
(sys_old_mmap): [IA64] use unsigned narrow_arg[].
Cast u_arg[5] (offset param) to unsigned long, to prevent erroneous signed
expansion.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
We use printllval without HAVE_LONG_LONG guards in many places,
but define it only if HAVE_LONG_LONG. This means that
on !HAVE_LONG_LONG systems we won't build for some time now.
* defs.h: Remove HAVE_LONG_LONG guard around LONG_LONG() macro
and printllval() function declaration.
* util.c: Remove HAVE_LONG_LONG guard around printllval()
function definition.
(printllval): Add compile-time error check for using wrong
if branch. Explain places where we deliberately use mismatched
types for printf formats.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
I found hard way why the code was using "_whence" name.
* file.c: Rename whence[] to whence_codes[].
(sys_lseek): Fix printxval() to use whence_codes[].
(sys_lseek32): Likewise.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
The code doesn't look fully correct to me, but I need to experiment
on actual x32 machine before I start "fixing" things.
For now, add comments, and optimize out one tprints() call...
* file.c (sys_lseek): Rename '_whence' as 'whence'.
Merge printing of ", " into subsequent tprintf.
(sys_lseek32): Likewise.
(sys_llseek): Likewise.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
* linux/mips/syscallent.h: Remove trailing empty elements.
Compactify huge stretches of NULL elements in the middle.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
* syscall.c (get_scno): [TILE] Remove TCB_WAITEXECVE check,
it is never true on Tile, and stopped compiling when
TCB_WAITEXECVE define was removed for Tile.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Zero-extension is slightly more common that sign-extension:
all pointers are zero-extended, and some other params are unsigned.
Whereas signed ones (fds, pids, etc) are often treated as
_32-bit ints_ even by kernel, so just unconditionally casting
such tcp->u_arg[N] to int works.
* syscall.c (get_syscall_args): [X86] Zero-extend 32-bit args
instead of sign-extension.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
* defs.h: Define widen_to_long() macro.
* signal.c (sys_kill): Use it instead of open-coding it.
(sys_tgkill): Use widen_to_long() on pids.
* resource.c (decode_rlimit): Formatting fix.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
On x86_64:
text data bss dec hex filename
435661 26628 47424 509713 7c711 strace_old
435501 26612 47440 509553 7c671 strace_new_clever_wordsize
On x32 and arm it should be even better, current_wordsize becomes
a constant there.
* defs.h: Declare current_wordsize as a variable if needed,
else declare as a constant define.
Remove declatation of personality_wordsize[].
* syscall.c: Make personality_wordsize[] static.
Declare current_wordsize as a variable if needed.
(set_personality): Set current_wordsize only if non-constant.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Unlike PTRACE_GETREGS, this new method detects 32-bit processes
reliably, without checking segment register values which
are undocumented and aren't part of any sort of API.
While at it, also fixed x32 detection to use __X32_SYSCALL_BIT,
as it should have been from the beginning.
* defs.h: Declare os_release and KERNEL_VERSION.
* strace.c: Make os_release non-static, remove KERNEL_VERSION define.
* syscall.c: New struct i386_user_regs_struct,
static union x86_regs_union and struct iovec x86_io.
(printcall): Use i386_regs or x86_64_regs depending on x86_io.iov_len.
(get_regs): On x86 and kernels 2.6.30+, use PTRACE_GETREGSET,
on earlier kernels fall back to old method.
(get_scno): [X86] Determine personality based on regset size
on scno & __X32_SYSCALL_BIT.
(syscall_fixup_on_sysenter): Use i386_regs or x86_64_regs depending
on x86_io.iov_len.
(get_syscall_args): Likewise.
(get_error): Likewise.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
This makes it easier to grep for them.
* syscall.c: Rename variables:
r0 -> bfin_r0,alpha_r0,sh_r0.
a3 -> mips_a3.
r2 -> mips_r2.
(get_scno): Use new variable names.
(get_syscall_result): Likewise.
(get_error): Likewise.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
* defs.h: Declare new function printsiginfo_at(tcp, addr).
* process.c (sys_waitid): Use printsiginfo_at().
(sys_ptrace): Likewise.
* signal.c: (printsiginfo_at): Implement this new function.
(sys_rt_sigsuspend): Use printsiginfo_at().
(sys_rt_sigtimedwait): Likewise.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
* process.c: Add start_code and start_data members of struct user
in struct_user_offsets[], where appropriate.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
* util.c (change_syscall): For MICROBLAZE, replace code
with dummy "return 0" and a comment explaining why that is ok
for this architecture.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
This is a leftover from sysenter/sysexit split.
I can't run-test it, but from code inspection it seems to be correct.
* syscall.c (get_scno): Remove stray syscall result reading for s390[x].
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
The maze of ifdefs/ifndefs was scaring new contributors.
Format it so that every arch has its own ifdef block.
* process.c: Deobfuscate definitions of struct user offsets.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
* defs.h: Do not define TCB_WAITEXECVE for AARCH64.
* util.c (change_syscall): For AARCH64 and X32, replace code
with dummy "return 0" and a comment explaining why that is ok
for these architectures.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
* defs.h: Do not define TCB_WAITEXECVE for AVR32, BFIN and TILE.
* util.c (change_syscall): For AVR32, BFIN and TILE, replace code
with dummy "return 0" and a comment explaining why that is ok
for these architectures.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
* net.c (printmsghdr): If current_wordsize is 4 and long is wider than it,
read 32-bit struct msghdr and expand it into a native one before using it.
(printmmsghdr): Likewise for struct mmsghdr.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
X32's return value is 64-bit. We were truncating it to 32-bit long
before checking for -errno.
* syscall.c (is_negated_errno_x32): New function.
(get_error): Use is_negated_errno_x32 for X32 architecture.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
We sometimes use LINUXSPARC and sometimes (more often)
use "defined(SPARC) || defined(SPARC64)". Better to use
one construct consistently.
LINUX_MIPS64 is altogether unused.
* defs.h: Remove LINUXSPARC and LINUX_MIPS64 defines.
Move PTRACE_xxx compat defines up, before arch-specific
machinery. Use defined(SPARC) || defined(SPARC64)
instead of LINUXSPARC.
* file.c: Use defined(SPARC) || defined(SPARC64) instead of LINUXSPARC.
* signal.c: Likewise.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
* syscall.c: Move PT_FLAGS_COMPAT define to its only user, get_scno().
Rename arm_regs_union's fields to names less likely to collide with
system defines.
(get_regs): Use sizeof(arm_regs_union) instead of sizeof(aarch64_regs).
This should be the same, but htis way it's cleaner.
Remove __X32_SYSCALL_MASK and use __X32_SYSCALL_BIT instead.
Explain 64-bit check in X32 build better.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
strace used to hang when decoding sigaltstack called with invalid
stack_t pointers because of mishandling umove() return code.
* signal.c (print_stack_t): Handle unfetchable stack_t properly.
Change return type to void.
(sys_sigaltstack): Update print_stack_t() usage.
Reported-by: kawillia@ucalgary.ca
If you call glibc's syscall wrapper like so:
static int name[] = { CTL_NET, NET_IPV4, NET_IPV4_LOCAL_PORT_RANGE };
int buffer[2] = { 32768, 61000 };
size_t size = sizeof(buffer);
sysctl(name, 3, 0, 0, buffer, size);
(note that oldval/oldlenp are NULL).
The current strace code complains like so:
_sysctl({{CTL_NET, NET_IPV4, NET_IPV4_LOCAL_PORT_RANGE, 38}, 3, process_vm_readv: Bad address
(nil), 0, 0x7fffe23c3960, 8}) = -1 EACCES (Permission denied)
Since passing NULL for the old values is valid, handle that explicitly.
This also simplifies the code a bit by splitting up the handling of the
new and old args so that we only handle the new args once.
Now the output looks like:
_sysctl({{CTL_NET, NET_IPV4, NET_IPV4_LOCAL_PORT_RANGE, 38}, 3, NULL, 0, 0x7fff8c0c91b0, 8) = -1 EACCES (Permission denied)
* system.c (sys_sysctl): Check if info.oldval is NULL first. Move the
processing of oldlen/info.newval/info.newlen out so they always get
executed. Fix the format strings so we use %lu for unsigned long rather
than a mix of %ld and %lu.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
It is not likely anyone uses glibc older that that:
glibc 2.1.1 was released in 1999
* net.c: Remove test for glibc >= 2.1.
* signal.c: Remove code which is compiled only for glibc < 2.1.
* util.c: Likewise.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
* syscall.c: Remove "static long pc" variable.
(get_scno): Use an automatic long variable instead of a static.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
* file.c (sys_utime): Don't call abort() if wordsize is strange.
Instead, warn user about it.
* desc.c (printflock): Use the same message string as in sys_utime.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
* signal.c: Stop using __sighandler_t glibc'ism. Remove SA_HANDLER macro.
Explain why we can't use "sa_handler" as a field name.
(sys_sigaction): Use __sa_handler instead of SA_HANDLER macro.
(sys_rt_sigaction): Likewise.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
By putting aarch64_regs and arm_regs into a union,
register copying is eliminated.
No need to check and change personality on syscall exit.
* defs.h: Remove unused NUM_ARM_REGS define. Fix indentation.
* syscall.c: Put aarch64_regs and arm_regs into a union.
(update_personality): Shorten bitness message.
(printcall): Add commented-out PC printing.
(get_regs): Remove now-unnecessary 64-to-32 bits register copying.
(get_syscall_result): Drop personality changing code.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
* syscall.c: Remove static long psr.
(get_scno): Use local psr variable.
(get_syscall_result): Likewise.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
This change should speed up strace by a tiny bit.
More importantly, it makes it much more clear that
fork and exec fixups are not necessary for any reasonably
recent kernels. IOW: syscall_fixup_for_fork_exec() and its callees
are all dead code.
* defs.h: Declare new need_fork_exec_workarounds flag variable.
* strace.c: Define need_fork_exec_workarounds flag variable.
(test_ptrace_setoptions_followfork): Return 0/1 as success/fail indicator.
(test_ptrace_setoptions_for_all): Likewise.
(init): Set need_fork_exec_workarounds to TRUE if needed.
* syscall.c: Rename internal_syscall() to syscall_fixup_for_fork_exec().
(trace_syscall_entering): Call syscall_fixup_for_fork_exec() only if
need_fork_exec_workarounds == TRUE.
(trace_syscall_exiting): Likewise.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
tilegx support has been in the kernel since 3.0.
In addition, fix some issues with the tilepro support already
present in strace, primarily the decision to use the
<asm/unistd.h> numbering space for system calls.
* defs.h [TILE]: Include <asm/ptrace.h> and provide an extern
struct pt_regs tile_regs for efficiency. Provide compat 32-bit
personality via SUPPORTED_PERSONALITIES, PERSONALITY0_WORDSIZE,
PERSONALITY1_WORDSIZE, and DEFAULT_PERSONALITY.
* linux/tile/errnoent1.h: New file, includes linux/errnoent.h.
* linux/tile/ioctlent1.h: New file, includes linux/ioctlent.h.
* linux/tile/signalent1.h: New file, includes linux/signalent.h.
* linux/tile/syscallent.h: Update with new asm-generic syscalls.
The version previously committed was the from the first tile patch
to LKML, which subsequently was changed to use <asm-generic/unistd.h>.
* linux/tile/syscallent1.h: Copy from linux/tile/syscallent.h.
* mem.c (addtileflags) [TILE]: use %ld properly for a "long" variable.
* process.c [TILE]: Choose clone arguments correctly and properly
suppress all "struct user" related offsets in user_struct_offsets.
* signal.c [TILE]: Use tile_regs not upeek.
* syscall.c (update_personality) [TILE]: Print mode.
(PT_FLAGS_COMPAT) [TILE]: Provide if not in system headers.
(tile_regs) [TILE]: Define 'struct pt_regs' variable to hold state.
(get_regs) [TILE]: use PTRACE_GETREGS to set tile_regs rather than using upeek.
(get_scno) [TILE]: Set personality.
(get_syscall_args) [TILE]: Use tile_regs.
(get_syscall_result) [TILE]: Update tile_regs.
(get_error) [TILE]: Use tile_regs.
(printcall) [TILE]: Print pc.
(arg0_offset, arg1_offset, restore_arg0, restore_arg1) [TILE]:
Properly handle tile call semantics and support tilegx.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
strace -i was fetching PC with a separate PEEKUSER
despite having GETREGS data:
ptrace(PTRACE_GETREGS, 22331, 0, 0x8087f00) = 0
ptrace(PTRACE_PEEKUSER, 22331, 4*EIP, [0x80dd7b7]) = 0
write(3, "[080dd7b7] ioctl(0, SNDCTL_TMR_T"..., 82) = 82
ptrace(PTRACE_SYSCALL, 22331, 0, SIG_0) = 0
Now it does this:
ptrace(PTRACE_GETREGS, 22549, 0, 0x8087ea0) = 0
write(3, "[080dd7b7] ioctl(0, SNDCTL_TMR_T"..., 82) = 82
ptrace(PTRACE_SYSCALL, 22549, 0, SIG_0) = 0
Analogous improvement in sys_sigreturn() is also implemented.
* defs.h: Declare extern struct pt_regs regs for SPARC[64] and ARM.
Declare clear_regs(), get_regs() and get_regs_error flag variable.
* strace.c (trace): Call get_regs(pid) as soon as we know the tcb
and that it is stopped.
* syscall.c (get_regs): New function. Used to fetch registers early,
just after tracee has stopped.
(printcall): Move it here from util.c. Use global regs.REG data,
if available on the arch, instead of re-fetching it.
(get_scno): Use global regs.REG data.
(get_syscall_result): Likewise.
* signal.c (sys_sigreturn): Likewise.
* util.c (printcall): Moved to syscall.c.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
The s390 and s390x pass semtimedop arguments differently from other
architectures. sys_semtimedop parser was fixed for s390 by commit
v4.6-177-ge0f5fd8, and s390x requires the same fix.
* linux/ipc.c (sys_semtimedop): Fix timespec decoding on s390x.
Signed-off-by: Stanislav Brabec <sbrabec@suse.cz>
Looking at the implementation of wrapped semtimedop() call inside glibc
and kernel, I started to believe that timeval should be located in
tcp->u_arg[4] and not tcp->u_arg[5]. Fortunately, tcp->u_arg[5] now
works correctly as well, due to side effects of decode_ipc_subcall().
declaration in header:
int semtimedop(semid, *sops, nsops, *timeout);
0 1 2 3
sys_ipc arguments in glibc on all patforms except s390*:
semid, (int) nsops, 0, CHECK_N (sops, nsops), timeout
0 1 2 3 4
We have to use indexes: 0 3 1 4
sys_ipc arguments on s390*:
semid, (int) nsops, timeout, sops
0 1 2 3
We have to use indexes: 0 3 1 2
* ipc.c (sys_semtimedop) [!S390]: Fix timeval argument index in
indirect_ipccall case.
Add a state argument to change_syscall() so that SPARC can modify that
instead of read-modify-writing the whole register set. This function is
always called within an arg_setup/arg_finish_change sequence which on
certain architectures like SPARC will also be doing a read-modify-write.
This prevents the second write (from arg_finish_change) from undoing the
effects of the change_syscall call.
* util.c (change_syscall): Move below definition of arg_setup_state.
Add state argument.
[SPARC || SPARC64] Change to set state->u_regs[U_REG_G1] rather than
read-modify-writing it with PTRACE_GETREGS and PTRACE_SETREGS.
(setbpt, clearbpt): Pass state argument to change_syscall.
Signed-off-by: James Hogan <james.hogan@imgtec.com>
* defs.h [AARCH64]: Copy in the definition of arm_pt_regs and the
accessor macros, so it's possible to build on AArch64 without
ARM system headers. Set SUPPORTED_PERSONALITIES to 2.
Define PERSONALITY0_WORDSIZE and PERSONALITY1_WORDSIZE.
Set DEFAULT_PERSONALITY to 1.
* linux/aarch64/errnoent1.h: New file, includes generic errnoent.h.
* linux/aarch64/ioctlent1.h: New file, includes generic ioctlent.h.
* linux/aarch64/signalent1.h: New file, includes generic signalent.h.
* linux/aarch64/syscallent1.h: Rename from linux/aarch64/syscallent.h.
* linux/aarch64/syscallent.h: New file, includes arm/syscallent.h.
* syscall.c [AARCH64]: Define aarch64_regs.
(update_personality) [AARCH64]: Add debug output.
(get_scno) [AARCH64]: Determine if we're in ARM or AArch64 mode by
checking the size of the returned uio structure from PTRACE_GETREGSET
and interpret the structure accordingly.
(get_syscall_result): Likewise.
(get_syscall_args): Merge the AArch64 and ARM sections so that on
AArch64 we can fall back to supporting the ARM personality.
(get_error): Likewise.
Signed-off-by: Steve McIntyre <steve.mcintyre@linaro.org>
AArch64 has been included in linux from 3.7 onwards.
Add support for AArch64 in strace, tested on linux in a simulator.
* configure.ac: Support AArch64.
* defs.h [AARCH64]: Include <sys/ptrace.h>, define TCB_WAITEXECVE.
* ipc.c (indirect_ipccall): Support AArch64.
* process.c (struct_user_offsets): Likewise.
* syscall.c [AARCH64]: Include <asm/ptrace.h>, <sys/uio.h>, and
<elf.h>. Define struct user_pt_regs regs.
(get_scno, get_syscall_result): Support AArch64 using PTRACE_GETREGSET.
(get_syscall_args, get_error): Support AArch64.
* linux/aarch64/ioctlent.h.in: New file.
* linux/aarch64/syscallent.h: New file, based on linux 3.7 version of
asm-generic/unistd.h.
Signed-off-by: Steve McIntyre <steve.mcintyre@linaro.org>
New definitions match updates in Linux 3.4 and Linux 3.5 respectively.
* linux/errnoent.h (ERRNO_517): Change to EPROBE_DEFER.
(ERRNO_518): Change to EOPENSTALE.
Signed-off-by: Steve McIntyre <steve.mcintyre@linaro.org>
Since someone can invoke these entry points directly with syscall(),
at least decode their name and show that they're 64bit versions rather
than just showing syscall_###.
* linux/x32/syscallent.h: Sync all missing entries below 312 with x86_64.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
strace used to honor fflush(3) return value in trace_syscall_entering
which resulted to tracees not being PTRACE_SYSCALL'ed which in turn
caused nasty hangups like this one:
$ strace -o'|:' pwd
|:: Broken pipe
There is little strace can do in case of fflush(3) returning EOF, and
hangup is certainly not the best solution for the issue.
* syscall.c (trace_syscall_entering): Ignore fflush(3) return value.
Discovered by running test/sigkill_rain under strace.
* util.c (umoven): Do not emit error message if process_vm_readv
fails with ESRCH.
(umovestr): LikeWise.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Starting with linux 3.6 (and backported to earlier kernels), these two
syscalls have changed numbers (moving from native to compat entry points).
Update the strace syscall list accordingly.
* linux/x32/syscallent.h: Move setsockopt from 54 to 541, and move
getsockopt from 55 to 542.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
In case of normal strace termination, when the trace output is
redirected to a file or a pipe, close it and wait for the pipe
process termination.
* strace.c (main): Before normal exit, close shared_log when it
differs from stderr, and wait for popen_pid termination.
* defs.h: Define USE_SEIZE to 1. Remove PTRACE_SEIZE_DEVEL
and PTRACE_EVENT_STOP1.
* strace.c (ptrace_attach_or_seize): Replace PTRACE_SEIZE_DEVEL
with 0.
(trace): Do not check for PTRACE_EVENT_STOP1.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
This syncs with the syscall table as it is in linux 3.4.
* linux/x32/syscallent.h (59): Fix comment typo.
(78): Add missing getdents entry.
(174): Delete create_module entry (not in the kernel).
(181, 182, 183, 184, 185): Add missing entries.
(524, 536, 539, 540): Fix spacing.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
* system.c (sys_sysctl): Cast pointer to long, not size_t,
when we intend to use it as an address. Set oldlen to 0 prior
to reading into it - we want to have deterministic result
if read fails.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
The documented formats are %zu and %zd, but since our (normally disabled)
"fast" printf code doesn't support those too, I convert them to %lu and %ld.
* bjm.c (sys_query_module): Convert %Zd usages to %lu.
* system.c (sys_sysctl): Likewise.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
We no longer track parent/child relationship between tracees.
Therefore, we no longer suspend new tracee until parent is seen
exiting form [v]fork/clone. The comment is obsolete.
* strace.c (trace): Remove outdated comment.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
On 64bit systems with a single personality, they used to be sizeof(long),
which has type "long", not "int", which complicates printf formats.
* defs.h: Ensure that PERSONALITY0_WORDSIZE;s tyoe is int.
This in turn makes sure current_wordsize is also an int.
* count.c (call_summary): Revert the change which added cast to int.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
This fixes Fedora bug 659382.
Low risk: this code is not supposed to be used on any non-acient kernel.
* util.c (clearbpt): Restore syscall number too.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
If you have a newer glibc which provides process_vm_readv, but it is built
against older kernel headers which lack __NR_process_vm_readv, the library
will contain a stub implementation that just returns ENOSYS. Autoconf
checks for this case explicitly and will declare it as unavailable. So we
end up in a case where the headers provide the prototype, but autoconf has
not defined HAVE_PROCESS_VM_READV, so we hit the same build failure again:
util.c:738:16: error: static declaration of 'process_vm_readv' follows non-static declaration
/usr/include/bits/uio.h:58:16: note: previous declaration of 'process_vm_readv' was here
So rename our local function to something unique, and add a define so the
callers all hit the right place.
* util.c (strace_process_vm_readv): Rename from process_vm_readv.
(process_vm_readv): Define to strace_process_vm_readv.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
