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.
* loop.c (decode_loop_info): Use square brackets to print
struct loop_info.lo_init and struct loop_info.reserved arrays.
(decode_loop_info64): Use square brackets to print
struct loop_info64.lo_init array.
* defs.h (evdev_abs): New xlat prototype.
(evdev_ioctl): Remove.
* evdev.c: Do not include "xlat/evdev_abs.h".
Mpersify struct ff_effect and ioctl numbers.
(evdev_ioctl): Mpersify.
* NEWS: Mention this enhancement.
Change the order of ioctl command cheks to match the kernel:
1st, check for fixed-number fixed-length commands,
2nd, check for fixed-number variable-length commands,
3rd, check for multi-number fixed-length commands,
4thm check for multi-number variable-length commands.
* evdev.c (evdev_read_ioctl, evdev_write_ioctl): Reorder
ioctl command checks.
* file_ioctl.c (file_ioctl, print_file_dedupe_range_info): Print first
two elements of info array in abbrev mode.
* tests/btrfs.c (btrfs_test_extent_same_ioctl): Handle newly
printed elements.
* fs_x_ioctl.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* block.c: (block_ioctl): Move parser of FITRIM, FIFREEZE, and FITHAW
to fs_x_ioctl.c.
* defs.h (fs_x_ioctl): New prototype.
* ioctl.c (ioctl_decode): Call fs_x_ioctl for ioctl type 'X'.
* hdio.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* block.c: Do not include <linux/hdreg.h>.
(block_ioctl): Move parser of HDIO_GETGEO to hdio.c.
* defs.h (hdio_ioctl): New prototype.
* ioctl.c (ioctl_decode): Call hdio_ioctl for ioctl type 0x03.
* btrfs.c (btrfs_ioctl): Skip BTRFS_IOC_FILE_EXTENT_SAME case
if this constant is not defined.
* tests/btrfs.c (btrfs_test_extent_same_ioctl): Skip this part
of the test if BTRFS_IOC_FILE_EXTENT_SAME is not defined.
* tests/btrfs.c (btrfs_test_sync_ioctls): Change u64val
to 0xdeadbeefbadc0ded.
(btrfs_test_subvol_ioctls): Likewise. Change bad_pointer
to make their lower 32-bit parts closer to -1U.
* tests/btrfs.c: Include "tests.h" first.
Skip the test if [!HAVE_LINUX_BTRFS_H]. Skip parts of the test
that use struct fiemap if [!HAVE_LINUX_FIEMAP_H].
This is limited to FIEMAP. FICLONE* and FIDEDUPERANGE originated in
btrfs and will be tested there.
* tests/file_ioctl.c: New file.
* tests/file_ioctl.test: New test.
* tests/.gitignore: Add file_ioctl.
* tests/Makefile.am (check_PROGRAMS): Likewise.
(DECODER_TESTS): Add file_ioctl.test.
* btrfs.c (print_btrfs_data_container_logical_ino,
print_btrfs_data_container_ino_path, print_uint64,
print_btrfs_ioctl_search_header, print_objectid_callback,
print_btrfs_ioctl_space_info): New functions.
(btrfs_print_logical_ino_container, btrfs_print_ino_path_container):
Rewrite printers of struct btrfs_data_container.val array
using print_array.
(btrfs_print_qgroup_inherit): Rewrite printer of
struct btrfs_qgroup_inherit.qgroups array using print_array.
(btrfs_print_tree_search): Rewrite printer of
struct btrfs_ioctl_search_header array using print_array.
(btrfs_ioctl): Rewrite printer of
struct btrfs_ioctl_send_args.clone_sources and
struct btrfs_ioctl_space_args.spaces using print_array.
* btrfs.c (btrfs_print_balance_args, btrfs_print_features,
btrfs_print_qgroup_limit, btrfs_print_data_container_header,
btrfs_print_logical_ino_container, btrfs_print_ino_path_container,
btrfs_print_qgroup_inherit): Mark some function arguments as const.
Being a wrapper around <linux/ioctl.h>, <sys/ioctl.h> brings
no benefits for this parser.
* file_ioctl.c: Include <linux/ioctl.h> instead of <sys/ioctl.h>.
* configure.ac (AC_CHECK_HEADERS): Add linux/fiemap.h.
* file_ioctl.c: Do not include <linux/fiemap.h> and xlat/fiemap_*.h
if [!HAVE_LINUX_FIEMAP_H].
(file_ioctl): Do not parse FS_IOC_FIEMAP if [!HAVE_LINUX_FIEMAP_H].
* file_ioctl.c: Include <linux/fiemap.h>, "xlat/fiemap_flags.h",
and "xlat/fiemap_extent_flags.h".
(file_ioctl): Handle FS_IOC_FIEMAP.
* ioctl.c (ioctl_decode): Use file_ioctl for decoding 'f' code ioctls.
On noticeably lagging systems, the parent process may invoke its chdir
call before the tracer gets notified about completion of its last child
process. Add extra sleep in the parent process to win the race.
* tests/attach-f-p.c (main): Change timerid into a static array.
Arm a timer in the parent process and wait for it.
* tests/net-yy-netlink.c (main): Use getsockname instead of recvmsg
as the latter has notible implementation differences that make
the test unreliable.
* tests/net-yy-inet.test: Do not pass arguments to net-yy-inet
as the latter does not accept them.
* tests/net-yy-netlink.test: Do not pass arguments to net-yy-netlink
as the latter does not accept them.
Rather than cast every __u64 or __s64 before printing,
define printing helpers for those types directly.
This complements commit 1f3482beda.
* tests/tests.h (PRI__d64, PRI__u64, PRI__x64): New macros.
We used to declare and define all printers marked with
MPERS_PRINTER_DECL, including ifdef'ed ones. That approach left us
no way to conditionally compile mpersified printers, which was not
a problem until btrfs ioctls appeared on the horizon.
With this change, those mpersified printers that are not going
to be compiled are also won't be declared and won't be added to
struct_printers.
This is implemented by filtering all source files containing
MPERS_PRINTER_DECL markers through CPP. As a nice side effect, this
also lifts an ugly requirement of writing all MPERS_PRINTER_DECL
declarations in a single line.
* README-mpers: Update description of MPERS_PRINTER_DECL syntax.
* defs.h [IN_MPERS_BOOTSTRAP] (MPERS_PRINTER_DECL): Turn into
a recursive variadic macro.
[!IN_MPERS_BOOTSTRAP] (MPERS_PRINTER_DECL): Turn into a variadic macro.
All callers changed.
* Makefile.am (mpers_preproc_files, mpers_printer_decl_pattern):
New variables.
(CLEANFILES): Add $(mpers_preproc_files).
(%.c.mpers.i): New rule.
(printers.h, %_printer_decls.h, %_printer_defs.h): Use
mpers_preproc_files instead of srcdir_mpers_source_files,
use mpers_printer_decl_pattern.
* .gitignore: Add /*.mpers.i.
There is no need to declare mpers printers in defs.h because
they are already declared in printers.h or native_printer_decls.h.
This complements commit 1e56814f77.
* defs.h (v4l2_ioctl): Remove.
* linux/netlink_diag.h: New file.
* Makefile.am (EXTRA_DIST): Add it.
* socketutils.c: Include it and "xlat/netlink_protocols.h".
(netlink_send_query, netlink_parse_response, netlink_print): New
functions.
(print_sockaddr_by_inode): Hook up netlink_print.
Do not assume that pwrite64 syscall fails when only part of the
specified memory buffer is accessible. With some kernels, one
accessible byte at the end of page is enough for pwrite64 syscall
to return 1.
* tests/pread64-pwrite64.c (main): Tweak the test that can lead
to a partial write.
Do not assume that write syscall fails when only part of the specified
memory buffer is accessible. With some kernels, one accessible byte
at the end of page is enough for write syscall to return 1.
* tests/read-write.c (main): Tweak the test that can lead to a partial
write.
* print_statfs.c (print_statfs_flags): Define only
for [HAVE_STRUCT_STATFS_F_FLAGS || HAVE_STRUCT_STATFS64_F_FLAGS].
* tests/inet-cmsg.c (print_origdstaddr): Define only
for [IP_ORIGDSTADDR].
* prctl.c (SYS_FUNC(prctl)): Print 2nd syscall argument
of PR_CAPBSET_DROP, PR_CAPBSET_READ, PR_CAP_AMBIENT, PR_MCE_KILL,
and PR_SET_SECCOMP using printxval_long.
Print 2nd syscall argument of PR_SET_SECUREBITS using printflags_long.
Print 3rd syscall argument of PR_CAP_AMBIENT and PR_MCE_KILL using
printxval_long.
* tests/prctl-seccomp-strict.c (main): Update.
* mem.c (print_mmap): Print 3rd and 4th arguments of syscall
using printflags_long.
(SYS_FUNC(mprotect)): Print 3rd argument of syscall using
printflags_long.
(SYS_FUNC(mremap)): Print 4th argument of syscall using printflags_long.
(SYS_FUNC(remap_file_pages)): Print 3rd and 4th arguments of syscall
using printflags_long.
* tests/remap_file_pages.c (main): Update.
* mtd.c (ubi_ioctl): Explicitly cast struct ubi_mkvol_req.vol_type
to unsigned short to avoid potential sign-extension bug when printing
invalid struct ubi_mkvol_req.vol_type.
Consistently treat operation argument of prctl/arch_prctl syscalls
as int to match the kernel behaviour.
* prctl.c (SYS_FUNC(prctl), SYS_FUNC(arch_prctl)): Assign 1st argument
of syscalls to a variable of type unsigned int and use it in all
subsequent checks and lookups.
* sock.c (print_ifreq): Explicitly cast struct ifreq.ifr_flags
to unsigned short to avoid potential sign-extension bug when printing
invalid struct ifreq.ifr_flags.
* poll.c (print_pollfd): Explicitly cast struct pollfd.events
to unsigned short to avoid potential sign-extension bug when printing
invalid struct pollfd.events.
* ipc_sem.c (print_sembuf): Explicitly cast struct sembuf.sem_flg
to unsigned short to avoid potential sign-extension bug when printing
invalid struct sembuf.sem_flg.
* fcntl.c (print_struct_flock64): Explicitly cast l_type and l_whence
fields to unsigned short to avoid potential sign-extension bug when
printing invalid l_type or l_whence fields.
Consistently treat operation argument of epoll_ctl syscall as int
to match the kernel behaviour.
* epoll.c (SYS_FUNC(epoll_ctl)): Assign 2nd argument of syscall
to a variable of type unsigned int and use it in all subsequent
checks and lookups.
* tests/epoll_ctl.c (invoke_syscall): New function.
(main): Use it.
Consistently treat operation argument of fcntl/fcntl64 syscalls as int
to match the kernel behaviour.
* fcntl.c (print_fcntl, SYS_FUNC(fcntl), SYS_FUNC(fcntl64)): Assign
2nd argument of syscall to a variable of type unsigned int
and use it in all subsequent checks and lookups.
* tests/struct_flock.c (invoke_test_syscall): New function.
(test_flock_einval, test_flock): Use it.
* tests/fcntl.c (test_flock64_einval): Use it.
* tests/fcntl64.c (test_flock64_einval, test_flock64): Use it.
* fcntl.c (print_fcntl, SYS_FUNC(fcntl), SYS_FUNC(fcntl64)):
Cast 2nd argument of xlookup to unsigned long.
* prctl.c (SYS_FUNC(prctl)): Likewise.
* sched.c (SYS_FUNC(sched_getscheduler)): Likewise.
* time.c (do_adjtimex): Likewise.
* ioprio.c (sprint_ioprio): Change type of the argument
and local variables from int to unsigned int.
* keyctl.c (print_keyring_serial_number): Cast 2nd argument
of xlookup to unsigned int.
* net.c (tprint_sock_type): Change type of the argument to unsigned int.
* printmode.c (sprintmode): Likewise.
* printsiginfo.c (printsigval): Change type of si_code argument
to unsigned int.
Prior to v2.6.25-rc3, <linux/videodev2.h> didn't include all
required headers, resulting to omission of V4L2_* constants.
This issue doesn't arise in v4l2.c and tests/ioctl_v4l2.c files
because they already include all necessary headers themselves.
* configure.ac (AC_CHECK_DECLS): Include <sys/time.h>, <linux/ioctl.h>,
and <linux/types.h> before <linux/videodev2.h>.
Since sys/ioctl.h is included unconditionally when needed
and ioctls.h doesn't exist, these checks are redundant.
* configure.ac (AC_CHECK_HEADERS): Remove ioctls.h and sys/ioctl.h.
Include <linux/ioctl.h> for <linux/sockios.h> on alpha, sh, and sh64
architectures because their <asm/sockios.h> use _IOR and _IOW macros
but don't include anything that would define these macros.
Being a wrapper around <linux/ioctl.h>, <sys/ioctl.h> brings
no benefits in this case. <ioctls.h> simply doesn't exist.
* sock.c [ALPHA || SH || SH64]: Include <linux/ioctl.h>
instead of <sys/ioctl.h> or <ioctls.h>.
Being a wrapper around <linux/ioctl.h>, <sys/ioctl.h> brings
no benefits for parser of VIDIOC_* ioctls.
* v4l2.c: Include <linux/ioctl.h> instead of <sys/ioctl.h>.
Remove erroneous use of verbose(tcp) as a printing limit
for already fetched members of siginfo_t structure.
* printsiginfo.h (printsiginfo): Remove boolean argument.
* printsiginfo.c (printsigval): Remove boolean argument, print
si_int and si_ptr members unconditionally. All callers changed.
(print_si_info): Remove boolean argument, print si_utime and si_stime
members unconditionally. All callers changed.
(printsiginfo): Remove boolean argument. All callers changed.
Fix build on systems without HAVE_DECL_V4L2_CTRL_TYPE_STRING or
V4L2_CTRL_CLASS_CAMERA.
* tests/ioctl_v4l2.c (main) [VIDIOC_S_EXT_CTRLS]: Disable part of the
test if [!HAVE_DECL_V4L2_CTRL_TYPE_STRING].
Replace V4L2_CTRL_CLASS_CAMERA with V4L2_CTRL_CLASS_USER.
htole32 function is not portable enough.
* v4l2.c (print_pixelformat): Rewrite initialization of pixel format
union without using of htole32.
* tests/ioctl_v4l2.c (main): Likewise.
The data argument of PTRACE_PEEKSIGINFO request is a pointer
to siginfo_t array.
* printsiginfo.c (print_siginfo_t): New function.
(print_siginfo_array): New mpers printer.
* process.c (SYS_FUNC(ptrace)): Print data argument of
PTRACE_PEEKSIGINFO request using print_siginfo_array.
As flags argument has a long integer type, print it using printflags64
because printflags takes unsigned int argument.
* process.c (SYS_FUNC(ptrace)): Print PTRACE_O_* flags
using printflags64.
The addr argument of PTRACE_PEEKSIGINFO request is a pointer to
a struct ptrace_peeksiginfo_args.
* process.c: Include "xlat/ptrace_peeksiginfo_flags.h".
(SYS_FUNC(ptrace)): Print addr argument of PTRACE_PEEKSIGINFO request
as a struct ptrace_peeksiginfo_args.
* xlat/ptrace_peeksiginfo_flags.in: New file.
For PTRACE_GETSIGMASK and PTRACE_SETSIGMASK requests, addr argument
contains the size of the buffer pointed to by data argument.
For PTRACE_SECCOMP_GET_FILTER request, addr argument contains
an offset counter.
* process.c (SYS_FUNC(ptrace)): Print addr argument of
PTRACE_GETSIGMASK, PTRACE_SETSIGMASK, and PTRACE_SECCOMP_GET_FILTER
requests using %lu format.
On sparc, kernel ignores data argument of PTRACE_GETREGS,
PTRACE_SETREGS, PTRACE_GETFPREGS, and PTRACE_SETFPREGS requests.
* process.c (SYS_FUNC(ptrace)) [SPARC || SPARC64]: Return RVAL_DECODED
right after printing addr argument if request is one of PTRACE_GETREGS,
PTRACE_SETREGS, PTRACE_GETFPREGS, or PTRACE_SETFPREGS.
As kernel ignores addr and data arguments of PTRACE_ATTACH,
PTRACE_INTERRUPT, PTRACE_KILL, and PTRACE_LISTEN request,
we do not print them either.
* process.c (SYS_FUNC(ptrace)): After printing pid argument, return
RVAL_DECODED if request is one of PTRACE_ATTACH, PTRACE_INTERRUPT,
PTRACE_KILL, or PTRACE_LISTEN.
As kernel ignores arguments of PTRACE_TRACEME request,
we do not print them either.
* process.c (SYS_FUNC(ptrace)): If request is PTRACE_TRACEME,
return early with RVAL_DECODED.
As first argument of ptrace syscall has a long integer type, print
it using printxval64 because printxval takes unsigned int argument.
* process.c (SYS_FUNC(ptrace)): Change type of "request" variable
from "long" to "unsigned long", print it using printxval64 instead
of printxval.
* printsiginfo.c (print_si_info): Explicitly cast si_syscall member
of siginfo_t that has type "int" to "unsigned int", to avoid sign
extension when passed to syscall_name function.
* tests/xattr.c: Check decoding of all xattr family syscalls.
* tests/xattr.test: Use run_strace_match_diff.
* tests/xattr.expected: Remove.
* tests/Makefile.am (EXTRA_DIST): Remove it.
* tests/print_quoted_string.c (print_quoted_memory): New function.
(print_quoted_string): Use it.
* tests/tests.h (print_quoted_memory): New prototype.
* xattr.c (print_xattr_val): Do not take insize into account, print it
as unsigned long.
Do not decode xattr values of size larger than XATTR_SIZE_MAX.
Use static buffer for fetching xattr values.
(print_xattr_list): Do not decode string when size is zero.
Decode the ioctls associated with the userfaultfd fd.
Note that they tend to read from and also return result in it's data
structure.
* configure.ac (AC_CHECK_HEADERS): Add linux/userfaultfd.h.
* userfaultfd.c [HAVE_LINUX_USERFAULTFD_H]: Add ioctl decoder.
* defs.h (uffdio_ioctl): New prototype.
* ioctl.c (ioctl_decode) [HAVE_LINUX_USERFAULTFD_H]: Wire in
uffdio_ioctl.
* xlat/uffd_*.in: Create flag xlat for all the IOCTLs.
* m4/ax_code_coverage.m4: Update to serial 16.
* Makefile.am (CODE_COVERAGE_LCOV_OPTIONS): Remove.
(CODE_COVERAGE_GENHTML_OPTIONS): Update, add prefix.
(CODE_COVERAGE_BRANCH_COVERAGE): Set to 1.
* tests/strace-ff.expected: New file.
* tests/Makefile.am (EXTRA_DIST): Add it.
* tests/strace-ff.test: Use it. Check that no other output files
have been created.
This header file was added in glibc-2.3.2,
so it's safe to assume its availability.
* configure.ac (AC_CHECK_HEADERS): Remove sys/epoll.h.
* epoll.c: Include <sys/epoll.h> unconditionally.
Compile [HAVE_SYS_EPOLL_H] code unconditionally.
* tests/epoll_ctl.c: Do not check for HAVE_SYS_EPOLL_H.
* tests/epoll_wait.c: Likewise.
Allow nanosleep(2) to spend 10% more time to make the test suite
more reliable on slow systems.
* tests/count.test: Treat 1.1 seconds as valid output.
* tests/strace-T.expected: Likewise.
* tests/strace-r.expected: Likewise.
* seccomp.c (decode_filter, decode_seccomp_fprog): Remove.
(print_bpf_filter): New function.
(print_seccomp_fprog): Use it via print_array.
* tests/seccomp-filter-v.c (main): Update.
* poll.c (print_pollfd): Change for use as print_array callback.
(decode_poll_entering): Use print_array.
* tests/poll.c: Update.
* tests/poll.test: Update.
* numa.c (print_node): New function.
(print_nodemask): Use it via print_array.
(print_page_array, print_int_array): Remove.
(print_addr): New function.
(print_status, print_int): Change for use as print_array callback.
(SYS_FUNC(move_pages)): Use print_addr, print_int, and print_status
via print_array.
* tests/move_pages.c: Update.
* tests/set_mempolicy.c: Likewise.
* tests/set_mempolicy.test: Likewise.
* epoll.c (print_epoll_event): Change signature for use as print_array
callback, all users updated.
(print_epoll_event_array): Remove.
(epoll_wait_common): Change to use print_array instead
of print_epoll_event_array.
* aio.c (print_iocbp): New function.
(SYS_FUNC(io_submit)): Use it via print_array. Use widen_to_long
to process long int argument.
(print_io_event): Change to print struct io_event fetched elsewhere.
(SYS_FUNC(io_cancel)): Update use of print_io_event.
(SYS_FUNC(io_getevents)): Use print_io_event via print_array.
Use widen_to_long to print long int arguments.
* tests/aio.c (main): Check decoding of io_submit with negative "nr".
* xlat/gen.sh (cond_xlat, gen_header): Do not strip "1<<" prefix
from xlat strings to fix incorrect output.
* tests/caps.awk: Update.
Reported-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
mips o32 differs from all other 32-bit architectures by defining
__kernel_fsid_t structure as an array of long ints.
* tests/xstatfsx.c (print_statfs): Explicitly cast elements
of PRINT_F_FSID array to unsigned int.
As values of some ioctl macros differ between personalities, these
personality specific ioctl macros should be properly defined for
mpersified code. Implement this by redefining personality specific
ioctl macros from ioctlent*.h for each file that includes MPERS_DEFS.
* Makefile.am (ioctl_macros_h): New variable.
(ioctl_macros%.h): New rule.
(BUILT_SOURCES, CLEANFILES): Add $(ioctl_macros_h).
(m%_type_defs.h): #include MPERS_$(mpers_PREFIX)IOCTL_MACROS.
* defs.h [HAVE_M32_MPERS] (MPERS_m32_IOCTL_MACROS): New macro.
[HAVE_MX32_MPERS] (MPERS_mx32_IOCTL_MACROS): Likewise.
Acked-by: Elvira Khabirova <lineprinter0@gmail.com>
There is no need to explicitly initialize big holes after conversion
of syscallent.h files to use designated initializers. For architectures
that have some data at the end of syscallent.h this initialization of
big holes is no-op, for others it just inflates the table unnecessarily.
* linux/arc/syscallent.h: Remove initialization of arch specific block
at the end of table.
* linux/metag/syscallent.h: Likewise.
* linux/nios2/syscallent.h: Likewise.
* linux/or1k/syscallent.h: Likewise.
* linux/tile/syscallent.h: Likewise.
* linux/tile/syscallent1.h: Likewise.
* linux/arm/syscallent.h: Remove explicit initialization of a big hole
before socket subcalls.
* linux/i386/syscallent.h: Likewise.
* linux/m68k/syscallent.h: Likewise.
* linux/mips/syscallent-n32.h: Likewise.
* linux/powerpc/syscallent.h: Likewise.
* linux/powerpc64/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.
* linux/sparc64/syscallent.h: Likewise.
* linux/mips/syscallent-compat.h: Remove explicit initialization
of big holes between different mips ABIs.
* linux/mips/syscallent-n64.h: Remove explicit initialization
of big holes before socket subcalls and between different mips ABIs.
* linux/mips/syscallent-o32.h: Likewise.
Some ioctls have flags fields that are 64-bit. A 32-bit val means
these flags will never be matched or printed.
* xlat.h: Include <stdint.h>.
(struct xlat): Change type of val to uint64_t.
When #val_type directive is encountered, gen.sh starts using
XLAT_TYPE/XLAT_TYPE_PAIR macros instead of XLAT/XLAT_PAIR,
with #val_type's argument as a type.
For example, "#val_type uint64_t" means values of type uint64_t.
* xlat/gen.sh (gen_header, print_xlat, print_xlat_pair):
Add val_type support.
This is essentially a no-op change that makes the following change
easier to read.
* xlat/gen.sh (print_xlat, print_xlat_pair): New functions.
(cond_xlat, gen_header): Use them.
* xlat.h (XLAT_TYPE): New macro, similar to XLAT but casts
to the specified type instead of unsigned int.
(XLAT_TYPE_PAIR): New macro, similar to XLAT_PAIR but casts
to the specified type instead of unsigned int.
Before this change there were two forms of xlat entries: those that use XLAT
or XLAT_END macros, and others verbatim entries. This change converts
the latter to use new XLAT_PAIR macro.
This is necessary for the upcoming change of xlat.val type.
* xlat.h (XLAT_PAIR): New macro.
* xlat/gen.sh (cond_xlat, gen_header): Use it.
As "flags" argument of sprint_open_modes and tprint_open_modes is
involved only in bit operations, change its type to unsigned.
* defs.h.c (sprint_open_modes, tprint_open_modes): Change type
to unsigned int.
* open.c (sprint_open_modes, tprint_open_modes): Likewise.
The "flags" argument of remap_file_page syscall has type "unsigned
long", so it is not correct to load most significant bits with garbage
ans assume they are going to be ignored.
* tests/remap_file_pages.c (main): Remove artificial garbage from flags.
This is necessary for the upcoming change of xlat.val type.
* defs.h (printflags): Rename to printflags64, change type of integer
argument to uint64_t.
(printflags): New static inline function.
* util.c (printflags): Rename to printflags64, change type of integer
argument to uint64_t. Print it using PRIx64 format.
Based on patch by Jeff Mahoney <jeffm@suse.com>.
This is necessary for the upcoming change of xlat.val type.
* defs.h (sprintflags): Rename to sprintflags64, change type of integer
argument to uint64_t.
(sprintflags): New static inline function.
* util.c (sprintflags): Rename to sprintflags64, change type of integer
argument to uint64_t. Print it using PRIx64 format.
Based on patch by Jeff Mahoney <jeffm@suse.com>.
This is necessary for the upcoming change of xlat.val type.
All users of printxvals are already prepared for this change.
* defs.h (printxvals): Change type of integer argument to uint64_t.
* util.c (printxvals): Likewise. Print it using PRIx64 format.
Patch by Jeff Mahoney <jeffm@suse.com>.
This is necessary for the upcoming change of xlat.val type.
All users of addflags are already prepared for this change.
* defs.h (xlat_search): Change type of integer argument to uint64_t.
* util.c (xlat_search): Likewise. Pass it by pointer.
(xlat_bsearch_compare): Treat first argument as a pointer.
Patch by Jeff Mahoney <jeffm@suse.com>.
This is necessary for the upcoming change of xlat.val type.
* defs.h (xlookup): Rename to xlookup64, change type of integer argument
to uint64_t.
(xlookup): New static inline function.
* util.c (xlookup): Rename to xlookup64, change type of integer
argument to uint64_t.
Based on patch by Jeff Mahoney <jeffm@suse.com>.
This is necessary for the upcoming change of xlat.val type.
All users of addflags are already prepared for this change.
* defs.h (addflags): Change type of integer argument to uint64_t.
* util.c (addflags): Likewise. Print it using PRIx64 format.
Based on patch by Jeff Mahoney <jeffm@suse.com>.
* numa.c (get_nodes): Rewrite an rename to print_nodemask.
All callers updated.
(SYS_FUNC(mbind), SYS_FUNC(set_mempolicy), SYS_FUNC(get_mempolicy)):
Print a delimiter before nodemask argument.
Define struct xlat in a separate file so that it could be used later by
tests without inclusion of defs.h header file.
* defs.h (struct xlat, XLAT, XLAT_END): Move ...
* xlat.h: ... here.
* Makefile.am (strace_SOURCES): Add xlat.h.
Do not attempt to print struct osf_statfs as if it was the same as
struct statfs. Since struct osf_statfs has never been decoded properly,
it is probably too late to implement a decoder, so let's just print the
pointer.
* statfs.c [ALPHA] (SYS_FUNC(osf_statfs), SYS_FUNC(osf_fstatfs)): Move ...
* alpha.c: ... here. Replace printstatfs with printaddr.
* statfs.c (print_statfs_flags): New function.
(printstatfs, printstatfs64, printcompat_statfs64): Use it.
* xlat/statfs_flags.in (ST_VALID): Move to the head of the list.
* tests/statfs.expected: Update.
Do not treat failed shmat(SHM_RND) as a test failure.
This change partially reverts commit
bea707377d.
* tests/shmxt.c (main): Use SHM_RND in the second shmat call,
do not treat its potential error as a test failure.
As PRI__*64 macros mirror PRI*64 macros from inttypes.h, follow the
principle of least astonishment and name these macros the same way.
* defs.h (PRI__s64): Rename to PRI__d64.
* mem.c (SYS_FUNC(old_mmap)): Use umove_or_printaddr instead of umoven
to fetch mmap arguments, return RVAL_DECODED when umove_or_printaddr
fails.
* tests/old_mmap.c (main): Check it.
Check decoding of "old mmap" edition of mmap syscall
on those architectures that define it.
* tests/old_mmap.c: New file.
* tests/old_mmap.test: New test.
* tests/.gitignore: Add old_mmap.
* tests/Makefile.am (check_PROGRAMS): Likewise.
(DECODER_TESTS): Add old_mmap.test.
* tests/mincore.c (print_mincore): New function.
(test_mincore): Use it. Check mincore with invalid vec address.
Check mincore with length argument not a multiple of the page size.
(main): Check with DEFAULT_STRLEN pages.
For each given architecture only one of these two syscalls has a
syscallent entry. When each parser is placed into a separate file,
moving to libstrace ensures that only one of them that is need
is linked into strace executable.
* sync_file_range.c (SYS_FUNC(sync_file_range2)): Move ...
* sync_file_range2.c: ... here.
* Makefile.am (strace_SOURCES): Move sync_file_range.c ...
(libstrace_a_SOURCES): ... here. Add sync_file_range2.c.
(strace_SOURCES_c): Add filtered libstrace_a_SOURCES.
Automatically change tests/*.c files using the following script:
for f in tests/*.c; do
grep -q '\<assert(' "$f" ||
sed -i '/# *include *<assert\.h>/d' "$f"
done
Automatically change tests/*.c files using the following script:
for f in tests/*.c; do
grep -Fv errno.h "$f" |
grep -Ewq '(si_)?errno|SOCK_FILTER_DENY_SYSCALL' ||
sed -i '/# *include *<errno\.h>/d' "$f"
done
Automatically change tests/*.c files using the following sed regexp:
sed -i 's/errno == E[[:alnum:]]\+ ? "E[[:alnum:]]\+" : "E[[:alnum:]]\+"/errno2name()/g'
Add a simple translator of errno to its name, so one could
simply use errno2name() instead of complex expressions like
errno == EINVAL ? "EINVAL" : "EFAULT".
* tests.h (errno2name): New prototype.
* errno2name.c: New file.
* Makefile.am (libtests_a_SOURCES): Add it.
hppa seems to be the only architecture that defines ECANCELLED synonym
for ECANCELED constant, other architectures just define ECANCELED.
Change the way how this constant is printed on hppa to be in line
with other architectures.
* linux/hppa/errnoent.h [253]: Rename ECANCELLED to ECANCELED.
timerfd syscall was introduced in v2.6.22-rc1, but in v2.6.25-rc1
it was replaced by timerfd_* syscall family.
* linux/avr32/syscallent.h (timerfd): Fix nargs.
* linux/dummy.h (sys_timerfd): Add stub alias.
* time.c (SYS_FUNC(timerfd)): Remove.
If kernel implementation of strict seccomp mode is buggy, test programs
will be killed by SIGKILL. This is a known problem at least on
x32 and mips. Skip affected tests if this is the case.
This change partially revert commit 9c1a72cd3f.
* tests/seccomp-strict.c (main): Remove workaround for x32.
* tests/prctl-seccomp-strict.c: Likewise.
* tests/seccomp-strict.test: Skip the test if the test program
has been killed by SIGKILL.
* tests/prctl-seccomp-strict.test: Likewise.
With linux kernels older than v3.9-rc1, compat rt_sigpending syscall
could fail with EFAULT in cases where on later kernels it succeeds.
Adjust the test to handle both cases properly.
* tests/rt_sigpending.c (iterate): Stop iterations if rt_sigpending
failed with EFAULT.
With linux kernels older than v3.16-rc1, iovec based compat syscalls may
return EINVAL in some cases where on later kernels they return EFAULT.
Adjust tests to handle both cases properly.
* tests/preadv-pwritev.c: Include <errno.h>.
(main): Print either "EINVAL" or "EFAULT" depending on errno.
* tests/pwritev.c: Likewise.
* tests/readv.c: Likewise.
* uid.c (print_groups): New function.
(SYS_FUNC(setgroups), SYS_FUNC(getgroups)): Use it.
Print first syscall argument using %u format.
* tests/getgroups.c: New file.
* tests/getgroups.test: New test.
* tests/getgroups32.c: New file.
* tests/getgroups32.test: New test.
* tests/setgroups.c: New file.
* tests/setgroups.test: New test.
* tests/setgroups32.c: New file.
* tests/setgroups32.test: New test.
* tests/.gitignore: Add getgroups, getgroups32, setgroups,
and setgroups32.
* tests/Makefile.am (check_PROGRAMS): Likewise.
(DECODER_TESTS): Add getgroups.test, getgroups32.test,
setgroups.test, and setgroups32.test.
* tests/pselect6.c (main): In the last case of printing pselect6
syscall, print the last argument the same way as in other cases.
Reported-by: Steve McIntyre <steve@einval.com>
This code used in two different places, so move it to the library.
* tests/overflowuid.c: New file.
* tests/Makefile.am (libtests_a_SOURCES): Add it.
* tests/tests.h (check_overflowuid, check_overflowgid): New prototypes.
* tests/uid.c (main): Use check_overflowuid.
* tests/uid16.c (main): Likewise.
As x86_64 kernel does not support x32 syscall numbers in
SECCOMP_MODE_STRICT, skip affected tests on x32.
* tests/seccomp-strict.c (main) [__x86_64__ && __ILP32__]: Bail out
using error_msg_and_skip.
* tests/prctl-seccomp-strict.c: Likewise.
AF_* constants appear to be more standardized, and in practice there's
always a one to one relationship between AF_* and PF_*, so let's
use AF_* instead of PF_*.
* xlat/domains.in: Remove.
* net.c: Do not define PF_UNSPEC. Stop including "xlat/domains.h".
Replace PF_* with AF_*. Replace domains with addrfams.
* tests/inet-cmsg.c: Replace PF_INET with AF_INET.
* tests/net-yy-inet.c: Likewise.
* tests/netlink_inet_diag.c: Likewise.
* tests/net-accept-connect.c: Replace PF_LOCAL with AF_LOCAL.
* tests/net-y-unix.c: Likewise.
* tests/net-yy-unix.c: Likewise.
* tests/netlink_unix_diag.c: Likewise.
* tests/net.expected: Replace PF_ with AF_.
* tests/pread64-pwrite64.c (dump_str, print_hex, test_dump):
New functions.
(main): Use them to test dumping of various data.
* tests/pread64-pwrite64.test: Update.
Explicitly declare first 4 arguments of mmap* and all remap_file_pages
arguments as unsigned to avoid potential sign extension issues.
* mem.c (print_mmap, SYS_FUNC(remap_file_pages)): Assign syscall
arguments to local variables of appropriate types.
The second (level) and third (optname) arguments of getsockopt and
setsockopt syscalls are enums, so treat them consistently as enums:
use "unsigned int" type to avoid potential sign extension issues.
* net.c (print_sockopt_fd_level_name, print_getsockopt,
print_setsockopt): Change type of "level" and "name" arguments
from "int" to "unsigned int".
* process.c (print_user_offset_addr): New function for printing offset
address in the tracee's USER area.
(SYS_FUNC(ptrace)): Assign syscall arguments to local variables with
appropriate types. Print pid argument as int. Use switch statement
in the code that handles address argument. Use print_user_offset_addr
to print address in PTRACE_PEEKUSER and PTRACE_POKEUSER requests.
Rather than cast every __u64 or __s64 before printing,
define printing helpers for those types directly.
* defs.h (PRI__s64, PRI__u64, PRI__x64): New macros.
* io.c (SYS_FUNC(preadv)): Call tprint_iov_upto instead
of tprint_iov and specify syscall return value as a data size limit.
* NEWS: Mention it.
* tests/preadv.c (main): Add a test case for preadv short read.
* io.c (print_llu_from_low_high_val): Rename to
print_lld_from_low_high_val, all callers changed.
Print value as a signed integer.
* tests/preadv.c: New file.
* tests/preadv.test: New test.
* tests/pwritev.c: New file.
* tests/pwritev.test: New test.
* tests/.gitignore: Add preadv and pwritev.
* tests/Makefile.am (check_PROGRAMS): Likewise.
(preadv_CPPFLAGS, pwritev_CPPFLAGS): New variables.
(DECODER_TESTS): Add preadv.test and pwritev.test.
As virtually every test defines either one or both of OUT and ERR
variables and these definitions are exactly the same, move definitions
of these variables from individual tests to init.sh.
* tests/init.sh (OUT, ERR): New variables.
* tests/ipc.sh: Remove initialization of OUT variable.
* tests/statx.sh: Likewise.
* tests/*.test: Remove initialization of OUT and ERR variables.
Use the executable itself for input by default.
* tests/sendfile.c (main): Use av[0] instead of av[1] for input.
* tests/sendfile64.c (main): Likewise.
* tests/sendfile.test: Invoke run_prog without arguments.
* tests/sendfile64.test: Likewise.
While -P option needs /proc to match descriptors, it does not use
/proc to match pathnames. As these tests do not need descriptor
match, the /proc requirement can be safely lifted.
* tests/open.test: Lift /proc/self/fd/ requirement.
* tests/openat.test: Likewise.
* xlat/clone_flags.in: Rename CLONE_STOPPED to CLONE_NEWCGROUP.
The former was removed by linux kernel commit v2.6.38-rc1~217.
The latter was introduced by linux kernel commit v4.6-rc1~67^2~8.
* xlat/bpf_map_types.in: Add BPF_MAP_TYPE_PERCPU_HASH,
BPF_MAP_TYPE_PERCPU_ARRAY, and BPF_MAP_TYPE_STACK_TRACE introduced by
linux kernel commits v4.6-rc1~91^2~326^2~5, v4.6-rc1~91^2~326^2~4, and
v4.6-rc1~91^2~212^2~1, respectively.
* tests/openat.c (main): Rename fname to sample, change its value
to "openat.sample".
* tests/openat.test: Use -P option to filter out openat syscalls
made by libc and dynamic linker.
* maint/ioctls_sym.sh: Update workarounds for drm/*_drm.h files.
Add workarounds for xen/gntdev.h and net/nfc/nci_core.h files.
Update path of saa6588.h and exynos-fimc.h files.
* mknod.c (decode_mknod): Treat device number argument as unsigned int.
* tests/mknod.c: Include <sys/sysmacros.h> or <sys/mkdev.h>
for definition of makedev macro.
(main): Add a check for a character special file.
Remove remnants of solaris personality support in mknod parser
that resulted to sparc personality being decoded as old solaris
personality. This complements commit v4.10-45-gdf4dd8b and fixes
commit v4.10-46-g588a90f.
* mknod.c (decode_mknod) [SPARC || SPARC64]: Remove.
Pathname containing >, }, or non-printable characters was resulting
in failing of the test.
* tests/scm_rights-fd.test: Fix pathname regex and simplify things.
Split TESTS into three groups: tests of specific decoders, libunwind
tests, and all other tests.
* tests/Makefile.am (TESTS): Move attach-f-p.test, attach-p-cmd.test,
bexecve.test, count-f.test, count.test, detach-running.test,
detach-sleeping.test, detach-stopped.test, dumpio.test,
filter-unavailable.test, fork-f.test, ksysent.test, opipe.test,
pc.test, qual_syscall.test, redirect.test, restart_syscall.test,
strace-f.test, and vfork-f.test to MISC_TESTS.
Move remaining tests except $(LIBUNWIND_TESTS) to DECODER_TESTS.
Add $(DECODER_TESTS) and $(MISC_TESTS).
* signal.c (SYS_FUNC(rt_sigqueueinfo)): Print first syscall argument
using %d format.
(SYS_FUNC(rt_tgsigqueueinfo)): Print first and second syscall arguments
using %d format.
* tests/rt_tgsigqueueinfo.c: New file.
* tests/rt_tgsigqueueinfo.test: New test.
* tests/.gitignore: Add rt_tgsigqueueinfo.
* tests/Makefile.am (check_PROGRAMS): Likewise.
(TESTS): Add rt_tgsigqueueinfo.test.
* signal (print_sigset_addr_len_limit): New function, cloned from
print_sigset_addr_len with added minimal length argument.
Treat length less than minimal length or greater than NSIG/8 as invalid.
Do not align length to 4-byte boundary. Align destination buffer
to 4-byte boundary and initialize it with zeroes so that subsequent
call to sprintsigmask_n will not access uninitialized data.
(print_sigset_addr_len): Turn into a wrapper around
print_sigset_addr_len_limit with current_wordsize
as a minimal length argument.
(SYS_FUNC(rt_sigpending)): Call print_sigset_addr_len_limit
instead of print_sigset_addr_len with 1 as a minimal length argument.
* tests/rt_sigpending.c: New file.
* tests/rt_sigpending.test: New test.
* tests/.gitignore: Add rt_sigpending.
* tests/Makefile.am (check_PROGRAMS): Likewise.
(TESTS): Add rt_sigpending.test.
Unfortunately, 64-bit linux kernel has a bug in matching inode numbers
greater than INT_MAX, so unix_diag exact match is not reliable.
This partially reverts commit 69bfc89770.
* socketutils.c (unix_send_query): Add NLM_F_DUMP to nlmsg_flags,
remove initialization of udiag_cookie.
* tests/netlink_unix_diag.c (send_query): Remove "inode" argument,
add NLM_F_DUMP to nlmsg_flags, remove initialization of udiag_ino
and udiag_cookie.
(check_responses): Remove "inode" argument and its use.
(main): Remove invocation of inode_of_sockfd and passing of listening
socket inode to send_query and check_responses.
* lseek.c (SYS_FUNC(lseek)): Print offset using %lld format.
(SYS_FUNC(llseek)): Likewise. Treat high and low components of offset
as unsigned long integers.
* tests/llseek.c (main): Check that negative offset is printed properly.
* tests/lseek.c (main): Likewise.
* tests/execve.c (main): Use one page gap between two mapped regions
for an address that is more likely to be not yet mapped.
* tests/execveat.c (main): Likewise.
* tests/mmap.c (main): Change output format from regexp to verbatim.
Add tests for madvise, mlockall, mremap, and msync syscalls.
* tests/mmap.test: Convert from match_grep to match_diff.
Add madvise, mlockall, mremap, and msync syscalls to syscall filter.
* socketutils.c (cache_and_print_inode_details): Make "details"
parameter const.
(print_sockaddr_by_inode_cached): Make "e" variable const.
(inet_send_query): Make struct nlmsghdr and struct inet_diag_req_v2
members const. Make "msg" variable const.
(inet_parse_response): Make "data" parameter and diag_msg variable const.
(receive_responses): Make "h" and "rc" variables const.
(unix_send_query): Make struct nlmsghdr and struct unix_diag_req
members const. Make "msg" variable const.
(unix_print): Make all parameters const.
(print_sockaddr_by_inode): Make proto_name parameter const.
* socketutils.c (tcp_v4_print, udp_v4_print, tcp_v6_print,
udp_v6_print): New functions.
(print_sockaddr_by_inode): Make "protocols" array static, replace
"family" and "protocol" members with "print" in its underlying
structure. Iterate through protocols array when the protocol name
is specified.
Check that -f -p attaches and follows all threads of the given pid.
* tests/attach-f-p.c: New file.
* tests/attach-f-p.test: New test.
* tests/.gitignore: Add attach-f-p.
* tests/Makefile.am (check_PROGRAMS): Likewise.
(attach_f_p_LDADD): New variable.
(TESTS): Add attach-f-p.test.
Some files are used conditionally depending on architecture and
configuration. Create a library for the purpose of keeping such
compilation units.
* Makefile.am (strace_LDADD, noinst_LIBRARIES): Add libstrace.a.
(libstrace_a_CPPFLAGS): New variable, set to $(strace_CPPFLAGS).
(libstrace_a_CFLAGS): New variable, set to $(strace_CFLAGS).
(libstrace_a_SOURCES): New variable, set to upeek.c.
* util.c (upeek): Move ...
* upeek.c: ... here.
* tests/utime.c (main): Check decoding of NULL struct utimbuf pointer.
Check decoding of struct utimbuf located at the end of page boundary.
* tests/utime.test: Update.
* tests/sysinfo.c (main): Check decoding of NULL struct sysinfo pointer.
Check decoding of struct sysinfo located at the end of page boundary.
* tests/sysinfo.test: Update.
* tests/adjtimex.c (main): Check decoding of NULL struct timex pointer.
Check decoding of struct timex located at the end of page boundary.
* tests/adjtimex.test: Update.
* strace.c (open_dev_null): Rename to open_dummy_desc. All callers
changed. Create a pipe and return its first descriptor instead of
opening /dev/null.
Otherwise, -oLOGFILE may end up using one of them. This conflicts
with the previous change, which closes stdin/out in strace.
* strace.c (init): If fcntl(F_GETFD) fails on stdin or stdout,
sanitize them: open /dev/null and dup it until fds 0,1,2
are all used up.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Testcase:
{ sleep 1; yes | head -n99999; } | strace -o/dev/null sh -c 'exec <&-; sleep 9'
The writer (head) will be able to perform writes to write end of the pipe.
With strace call removed, head will immediately get SIGPIPE.
This change fixes this: now writer immediately gets SIGPIPE with strace too.
* strace.c (startup_child): Close stdin/out and reopen them to /dev/null.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Check that -o |pipe_command works. In particular, check that
strace waits for completion of piped command.
* tests/opipe.test: New test.
* tests/Makefile.am (TESTS): Add it.
* strace-graph: A clone() followed by exit() with no intervening fork
or execve resulted in the new pid having no seq key. That breaks when
trying to use seq as an array reference. Ensure that clone populates
the new pid with an empty seq entry.
* strace-graph: Set "use strict" and "use warnings" options. Fix errors
detected by strict/warnings. $result is really a string, it can be "-1"
followed by an error text.
* tests/readlink.c (PREFIX, TARGET, LINKPATH): New macros.
(main): Allocate memory for string and buffer passed to the syscall
being tested using tail_alloc mechanism. Create and cleanup test
symlink. Test syscall error path output. Print hexquoted strings
using hexquote_strndup.
* tests/readlinkat.c: Likewise.
* tests/readlink.test: Remove creation and cleanup of test symlinks.
* Makefile.am (strace_CFLAGS): New variable, set to $(AM_CFLAGS).
(libmpers_CFLAGS): New variable, set to $(strace_CFLAGS).
(libmpers_m32_a_CFLAGS, libmpers_mx32_a_CFLAGS): New variables,
set to $(libmpers_CFLAGS).
Unlike the former test that was based on match_grep,
the new one uses match_diff and does more rigorous testing.
* tests/net-y-unix.c: New file.
* tests/net-y-unix.test: New test.
* tests/net-fd.test: Remove.
* tests/net-fd.expected: Remove.
* tests/.gitignore: Add net-y-unix.
* tests/Makefile.am (check_PROGRAMS): Likewise.
(TESTS): Replace net-fd.test with net-y-unix.test.
(net-fd.log): Remove.
(EXTRA_DIST): Remove net-fd.expected.
As -yy parser, compared to -y, needs to do at least 5 extra syscalls
(getxattr, socket, sendmsg, recvmsg, close) to print socket details,
caching results of netlink conversations between strace and kernel
noticeably reduces amount of system time spent by strace.
The caching is safe since sockets do not change their addresses after
successful bind or connect syscall.
* defs.h (string_quote, print_sockaddr_by_inode_cached): New prototypes.
* socketutils.c (cache_entry): New type.
(CACHE_SIZE, CACHE_MASK): New macros.
(cache): New static array.
(cache_and_print_inode_details): New static function.
(print_sockaddr_by_inode_cached): New function.
(inet_parse_response, unix_parse_response): Use
cache_and_print_inode_details.
* util.c (printfd): Use string_quote and print_sockaddr_by_inode_cached.
(string_quote): Remove static qualifier.
* NEWS: Mention this improvement.
* tests/unix-yy.c (main): Update.
* socketutils.c (inet_parse_response, unix_parse_response): Change
return type from bool to int, return -1 on all parse errors except
inode mismatch.
(receive_responses): Stop on the first nlmsg_type that is not
SOCK_DIAG_BY_FAMILY, also stop when the parser returns -1.
* tests/netlink_inet_diag.c (check_responses): Stop on short messages,
on first nlmsg_type that is not SOCK_DIAG_BY_FAMILY, print more verbose
diagnostics for NLMSG_ERROR.
* tests/netlink_unix_diag.c (check_responses): Likewise.
Unlike the former test that was based on match_awk,
the new one uses match_diff and does more rigorous testing.
* tests/net-yy-inet.c: New file.
* tests/net-yy-inet.test: New test.
* tests/net-yy.test: Remove.
* tests/net-yy-accept.awk: Remove.
* tests/net-yy-connect.awk: Remove.
* tests/inet-accept-connect-send-recv.c: Remove.
* tests/.gitignore: Replace inet-accept-connect-send-recv
with net-yy-inet.
* tests/Makefile.am (check_PROGRAMS): Likewise.
(TESTS): Replace net-yy.test with net-yy-inet.test.
(EXTRA_DIST): Remove net-yy-accept.awk and net-yy-connect.awk.
Unlike the former test that was based on match_awk,
the new one uses match_diff and does more rigorous testing.
* tests/net-yy-unix.c: New file.
* tests/net-yy-unix.test: New test.
* tests/unix-yy.test: Remove.
* tests/unix-yy-accept.awk: Remove.
* tests/unix-yy-connect.awk: Remove.
* tests/.gitignore: Add net-yy-unix.
* tests/Makefile.am (check_PROGRAMS): Likewise.
(TESTS): Replace unix-yy.test with net-yy-unix.test.
(EXTRA_DIST): Remove unix-yy-accept.awk and unix-yy-connect.awk.
* mem.c (SYS_FUNC(mincore)): Print only those bytes of the vector
that were written by the kernel.
* NEWS: Mention this fix.
* tests/mincore.c: New file.
* tests/mincore.test: New test.
* tests/Makefile.am (check_PROGRAMS): Add mincore.
(TESTS): Add mincore.test.
* tests/.gitignore: Add mincore.
* strace.c (init): Allow -p option along with a command.
(startup_child): In -D mode, record the parent of the tracer process
as strace_child.
(startup_attach): Save trace_tracer_pid before -D mode fork.
When tracing a command in -f mode, do not check for the command's
threads as it has no threads at this moment.
Never attach to the tracer process.
In -D mode, never attach to the parent of the tracer process,
terminate that process only once at the end of startup_attach,
and reset strace_child.
* strace.1: Document that -p option can be used along with tracing
of a command.
* NEWS: Mention it.
* tests/attach-p-cmd-cmd.c: New file.
* tests/attach-p-cmd-p.c: Likewise.
* tests/attach-p-cmd.test: New test.
* tests/.gitignore: Add attach-p-cmd-cmd and attach-p-cmd-p.
* tests/Makefile.am (check_PROGRAMS): Likewise.
(TESTS): Add attach-p-cmd.test.
This fixes Debian bug #549942.
* tests/sleep.c: New file.
* tests/.gitignore: Add sleep.
* tests/Makefile.am (check_PROGRAMS): Likewise.
* tests/count.test: Use it instead of system sleep utility.
* defs.h (dumpiov_upto): New prototype.
(dumpiov): Change to a wrapper around dumpiov_upto.
* util.c (dumpiov): Rename to dumpiov_upto, add and check data_size
argument.
* io.c (SYS_FUNC(readv)): Call tprint_iov_upto instead
of tprint_iov and specify syscall return value as a data size limit.
* syscall.c (dumpio): In case of SEN_readv, call dumpiov_upto instead
of dumpiov and specify syscall return value as a data size limit.
* NEWS: Mention this fix.
* tests/readv.c: New file.
* tests/readv.test: New test.
* tests/Makefile.am (check_PROGRAMS): Add readv.
(TESTS): Add readv.test.
* tests/.gitignore: Add readv.
* xlat/if_dqinfo_flags.in: New file.
* quota.c: Include "xlat/if_dqinfo_flags.h".
(decode_cmd_data): Print struct if_dqinfo.dqi_flags using printflags.
* quota.c (decode_cmd_data): Change return type to int.
Return 0 on entering Q_GETQUOTA, Q_V1_GETQUOTA, Q_V2_GETQUOTA,
Q_XGETQUOTA, Q_GETFMT, Q_GETINFO, Q_V2_GETINFO, Q_V1_GETSTATS,
Q_V2_GETSTATS, and Q_XGETQSTAT commands, return RVAL_DECODED
for any other command.
(SYS_FUNC(quotactl)): On entering, print third argument of any command.
For any command except Q_QUOTAON and Q_V1_QUOTAON, call decode_cmd_data
and forward its return value.
* quota.c (decode_cmd_data): As quota format value and xfs quota flags
are in-memory constants, print it like an array of one element instead
of a structure.
* quota.c (Q_*): Move to xlat/quotacmds.in.
(USRQUOTA, GRPQUOTA): Move to xlat/quotatypes.in.
(QFMT_VFS_OLD, QFMT_VFS_V0): Move to xlat/quota_formats.in.
(XFS_QUOTA_*): Move to xlat/xfs_quota_flags.in.
(XFS_*_QUOTA): Move to xlat/xfs_dqblk_flags.in.
(QIF_*): Move to xlat/if_dqblk_valid.in.
(IIF_*): Move to xlat/if_dqinfo_valid.in.
This fixes the test on platforms where both __NR_sendmmsg and
HAVE_SENDMMSG are defined but the former is not implemented by the
kernel and the second is implemented as an indirect syscall.
* tests/mmsg.c (LENGTH_OF): New macro.
(send_mmsg, recv_mmsg): Print expected output in case of ENOSYS.
(main): Redirect stdout to a new descriptor. Use LENGTH_OF.
Print expected output.
* tests/mmsg.test: Update.
* tests/mmsg.expected: Remove.
* tests/Makefile.am (EXTRA_DIST): Remove mmsg.expected.
On s390/s390x, syscalls with NR up to 255 can be implemented
directly using "svc NR", for NR >= 256 "svc 0" with %r1=NR is used.
The latter method is allowed for NR < 256, too.
When the syscall number specified directly or indirectly is recognized
by the kernel, i.e. it is less than its NR_syscalls value, it is stored
in %r2 and is available to arch_get_scno via s390_regset.gprs[2].
For syscall numbers >= NR_syscalls this register is set to 0,
but %r1 remains unchanged and could be used by arch_get_scno
via s390_regset.gprs[1] to decide what the syscall number is.
* linux/s390/get_scno.c (arch_get_scno): If s390_regset.gprs[2] is zero,
take syscall number from s390_regset.gprs[1].
* NEWS: Mention this fix.
This fixes Debian bug #485979 and Fedora bug #1298294.
Add clang-3.6/x86_64, clang-3.6/x86, and musl-gcc/x86_64 to the matrix.
* .travis.yml (compiler): Add gcc, clang-3.6, and musl-gcc.
(env): Add matrix, remove global ENABLE_GCC_WERROR.
(matrix): Add exclude, remove include.
* travis-ci.sh: Add support for clang and musl-gcc.
Use --enable-gcc-Werror for gcc build unconditionally.
* mpers.awk: Define default_pointer_size. Initialize byte_size
for DW_TAG_pointer_type to default_pointer_size. Set parent
for DW_TAG_member only. Handle DW_AT_count as an alternative
to DW_AT_upper_bound.
* configure.ac: Sort and group together AC_PROG_*, AC_C_*, AC_TYPE_*,
AC_HEADER_*, AC_CHECK_FUNCS, AC_CHECK_TYPES, AC_CHECK_MEMBERS,
and AC_CHECK_HEADERS.
* tests/scm_rights.c (main): Rewrite without fork.
Place all objects passed to sendmsg and recvmsg at the end
of memory pages followed by inaccessible pages.
* tests/scm_rights-fd.test: Update.
There are weird platforms that define __NR_sendmmsg but at the same time
do not define __NR_recvmmsg. Add a workaround for them.
* configure.ac (AC_CHECK_FUNCS): Add recvmmsg.
* tests/mmsg.c: Check for __NR_recvmmsg || HAVE_RECVMMSG.
Linux kernel commit v4.3-rc1-50-g8d4bd0e corrected uc_sigmask
of the compat signal frame, so remove the old workaround.
* linux/s390/arch_sigreturn.c (arch_sigreturn) [S390]: Remove.
Do not rely on "struct sigevent.__pad" being located at the same address
as "struct sigevent.sigev_notify_thread_id", it's not the case with musl
libc. Do not rely on struct sigevent definition at all to access
sigev_notify_thread_id.
* configure.ac (AC_CHECK_MEMBERS): Remove struct sigevent._sigev_un._pad
and struct sigevent.__pad.
* sigevent.h: New file.
* Makefile.am (strace_SOURCES): Add it.
* print_sigevent.c: Include it.
* print_sigevent.c (struct_sigevent): Remove.
(print_sigevent): Update all struct_sigevent users.
* tests/timer_create.c: Include "sigevent.h".
(main): Use struct_sigevent instead of struct sigevent,
all struct sigevent clients changed.
musl libc defines SIGRTMAX to a function that returns signed int,
which obviously makes gcc -Wsign-compare unhappy.
* signal.c (signame): Cast ASM_SIGRTMAX to unsigned int.
The size of off_t is not something one can rely upon. For example,
musl libc unconditionally defines it as an int64_t type on x86.
A cast to the target type helps to avoid these libc differences.
* configure.ac: Call AC_C_TYPEOF.
* tests/struct_flock.c (TYPEOF_FLOCK_OFF_T): New macro.
(test_flock_einval): Use it instead of off_t.
The types that are used to define msg_namelen member of struct msghdr
and cmsg_len member of struct cmsghdr differ between libcs.
For example, glibc defines them as size_t, while musl libc - as type
socklen_t. Print these fields using %lu format to avoid issues caused
by libc differences.
* tests/inet-cmsg.c (main): Cast mh.msg_controllen and c->cmsg_len
to unsigned long and print them using %lu format.
* tests/tests.h (tail_memdup): New prototype.
* tests/tail_alloc.c (tail_memdup): New function.
* tests/mmsg.c (main): Place all objects passed to sendmmsg and recvmmsg
at the end of memory pages followed by inaccessible pages.
* tests/mmsg.c: Include <sys/syscall.h>.
Check for __NR_sendmmsg as an alternative to HAVE_SENDMMSG.
[!HAVE_STRUCT_MMSGHDR] (struct mmsghdr): Define.
(send_mmsg, recv_mmsg): New functions.
(main): Use them instead of sendmmsg and recvmmsg.
Reported-by: Szabolcs Nagy <nsz@port70.net>
Make the test compatible with musl libc that uses RT_2 internally.
* tests/sigreturn.c (main): Replace RT_2 with RT_5.
* tests/sigreturn.test: Likewise.
Reported-by: Szabolcs Nagy <nsz@port70.net>
Workaround limitations of settimeofday wrapper provided by musl libc.
* tests/xettimeofday.c (main): Call settimeofday using syscall().
Reported-by: Szabolcs Nagy <nsz@port70.net>
Print the second return value of getxpid, getxuid, and getxgid syscalls
that return a pair of values using the same mechanism as pipe syscall.
* alpha.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* linux/alpha/syscallent.h [20]: Change SEN(getpid) to SEN(getxpid).
[24]: Change SEN(getuid) to SEN(getxuid).
[47]: Change SEN(getgid) to SEN(getxgid).
* NEWS: Mention this enhancement.
* tests/uid.awk: Update for getxgid output change.
* tests/uid.test: Cleanup.
* tests/getxxid.c: New file.
* tests/getxxid.test: New test.
* tests/Makefile.am (check_PROGRAMS): Add getxxid.
(TESTS): Add getxxid.test.
* tests/.gitignore: Add getxxid.
Use chdir syscalls instead of pwrite64 syscalls as a method of inserting
text markers into strace output because the former are more portable.
* tests/fork-f.c (logit): Rename to logit_. Use chdir instead of pwrite.
(logit): New macro wrapper around logit_.
(main): Update expected output.
* tests/vfork-f.c: Likewise.
* tests/fork-f.test: Update.
This change complements commit v4.9-359-gd93d9f8 by fixing
RVAL_UDECIMAL case.
The only syscall that appears to be affected is the times syscall.
* syscall.c (trace_syscall_exiting): In case of RVAL_UDECIMAL,
when current personality is 32-bit, print 32-bit return code.
* NEWS: Mention this fix.
Reported-by: Steve McIntyre <steve@einval.com>
On some platforms, e.g. hppa glibc, vfork() is implemented using fork
syscall, so the test cannot rely on the parent process remaining blocked
until the child process either terminates or calls execve.
* tests/vfork-f.c (main): Explicitly block the parent until the child
either terminates or calls execve.
Print cpu_set_t as a set of integers, similar to the way
fd_set is printed as a set of descriptors.
* affinity.c: Include <sched.h>.
(get_cpuset_size): New function.
(print_affinitylist): Rewrite using get_cpuset_size and next_set_bit.
(sched_getaffinity, sched_setaffinity): Print first two args as ints.
* NEWS: Mention this enhancement.
* tests/sched_xetaffinity.c: New file.
* tests/sched_xetaffinity.test: New test.
* tests/Makefile.am (check_PROGRAMS): Add sched_xetaffinity.
(TESTS): Add sched_xetaffinity.test.
* tests/.gitignore: Add sched_xetaffinity.
Reported-by: Joe Korty <joe.korty@ccur.com>
Also, since -k option does not support multiple personalities,
add strace-k.test to XFAIL_TESTS when running mpers tests.
* tests/Makefile.am (MPERS_NAME, LIBUNWIND_TESTS, XFAIL_TESTS_,
XFAIL_TESTS_m32, XFAIL_TESTS_mx32, XFAIL_TESTS): New variables.
[USE_LIBUNWIND] (LIBUNWIND_TESTS): Add strace-k.test.
(TESTS): Replace strace-k.test with $(LIBUNWIND_TESTS).
* bootstrap: Substitute MPERS_NAME variable in tests-*/Makefile.am.
* tests/strace-k.test: Do not check for -k option availability.
These manual tests are obsoleted by fork-f.test and vfork-f.test.
* test/fork.c: Remove.
* test/vfork.c: Remove.
* test/Makefile (PROGS): Remove fork and vfork.
* test/.gitignore: Remove fork and vfork.
Check how strace -f follows fork and vfork syscalls.
* tests/fork-f.c: New file.
* tests/fork-f.test: New test.
* tests/vfork-f.c: New file.
* tests/vfork-f.test: New test.
* tests/Makefile.am (check_PROGRAMS): Add fork-f and vfork-f.
(TESTS): Add fork-f.test and vfork-f.test.
* tests/.gitignore: Add fork-f and vfork-f.
On architectures that use dedicated registers to report syscall errors,
check for syscall error condition regardless of SYSCALL_NEVER_FAILS flag
in the syscall entry. On architectures that use negated errno
semantics, there is no way to distinguish a syscall error from a valid
return code that looks like a syscall error, and the check for
SYSCALL_NEVER_FAILS flag remains the only way to tell one case
from another.
* linux/alpha/get_error.c (get_error): Ignore check_errno.
* linux/ia64/get_error.c: Likewise.
* linux/mips/get_error.c: Likewise.
* linux/nios2/get_error.c: Likewise.
* linux/powerpc/get_error.c: Likewise.
* linux/sparc/get_error.c: Likewise.
* linux/sparc64/get_error.c: Likewise.
Despite being marked as SYSCALL_NEVER_FAILS, personality syscall still
might report a failure on some architectures due to obscure kernel
bugs. Check for this unlikely case and let users know that their
kernel has gone bananas.
* personality.c (SYS_FUNC(personality)): Do not parse syscall return
value in case of syserror.
This should not be a user visible change unless the kernel does
something completely unexpected. In the latter case, let user see
what is going on.
* uid.c (SYS_FUNC(getuid), SYS_FUNC(setfsuid)): Do not truncate
tcp->u_rval to uid_t on exiting syscall, return RVAL_DECODED.
* xlat/personality_options.in: Split into ...
* xlat/personality_types.in: ... personality types and ...
* xlat/personality_flags.in: ... personality flags.
* personality.c: Include "xlat/personality_types.h"
and "xlat/personality_flags.h" instead of "xlat/personality_options.h".
(SYS_FUNC(personality)): Print PER_MASK part of personality as
a symbolic value, and the rest of personality as a set of flags.
* tests/personality.c (main): Add more test cases.
* tests/personality.test: Update.
Value 0xffffffff of personality syscall first argument has a special
meaning, it is not a set of personality flags.
* personality.c (SYS_FUNC(personality)): If personality equals
to 0xffffffff, print it verbatim.
* tests/personality.c (main): Test it.
* personality.c (SYS_FUNC(personality)): On entering syscall,
print first argument not as a hex value but as a set of flags.
On exiting syscall, print return code as a set of flags.
* NEWS: Mention this enhancement.
Starting with commit
v2.6.29-6609-g11d06b2a1e5658f448a308aa3beb97bacd64a940, personality
syscall never fails to set the personality, but before commit
v2.6.35-rc1-372-g485d527 it still could return an error.
Starting with that commit, personality syscall never returns an error.
* linux/*/syscallent.h: Set SYSCALL_NEVER_FAILS flag in personality
syscall entries.
Starting with commit v4.9-46-g681452b that removed all references
to acl.h, we no longer use acl development library.
* .travis.yml (addons.apt.packages): Remove libacl1-dev.
* strace.spec (BuildRequires): Remove libacl-devel.
Before this change, all files that exist since 20th century had
copyright headers, while most files that appeared later didn't. This
change fixes the inconsistency by adding missing copyright headers.
It doesn't mean that copyright headers became maintained. In my view,
git history provides much better information on this subject and is much
more accurate than copyright headers.
The syscall_name argument was subject to macro expansion because
it was passed down to other macros before it was prefixed.
musl libc defines lfs64 names as macros (e.g. fstat64 as fstat)
so SYS_FUNC(fstat64) was expanded to sys_fstat.
This change adds the prefix before the name is passed to other macros,
i.e. the argument of SYS_FUNC_NAME is already prefixed with sys_.
* defs.h (SYS_FUNC): Add sys_ prefix to SYS_FUNC_NAME's argument.
(SYS_FUNC_NAME): Do not add sys_ prefix to MPERS_FUNC_NAME's argument.
* linux/ia64/syscallent.h (SYS_FUNC_NAME): Do not add sys_ prefix
to MPERS_FUNC_NAME's argument.
* syscall.c (SEN_NAME): Remove.
(SEN): Replace SEN_NAME call with its definition. Add sys_ prefix
to SYS_FUNC_NAME's argument.
As glibc times wrapper on x32 is known to wrongly truncate and then sign
extend the syscall return value, invoke the syscall directly on x32.
* tests/times.c (main) [__NR_times && __x86_64__ && __ILP32__]:
Invoke times syscall using inline asm.
* mpers.awk (array_seq): New function.
(what_is): Use it for printing filler names. Tweak return types.
* mpers_test.sh: Use a more complicated sample type to test mpers
machinery.
Make mpers.awk check for potential index loops. Such loops should not
normally happen, but mpers.awk will not go into infinite recursion if
they do.
* mpers.awk (enter, leave): New functions.
(what_is): Use them.
Make mpers.awk check that accessed elements of the main array that
are expected to have non-empty values really meet this requirement.
* mpers.awk (array_get): New function.
(what_is, END): Use it.
* mpers_test.sh: Use a more complicated type to test mpers machinery.
Reported-by: Mike Frysinger <vapier@gentoo.org>
As arm kernels used to overwrite ARM_r0 with -EINTR after
ERESTART_RESTARTBLOCK, update regexp to workaround this.
* tests/restart_syscall.c (main) [__arm__]: Add 0xfffffffc as a valid
alternative value for the first argument of restarted nanosleep syscall.
Follow the x86_64 example and define registers in arm_pt_regs
as unsigned, to avoid potential sign extension bugs.
This also fixes --enable-gcc-Werror build.
* linux/aarch64/arch_regs.c (arm_pt_regs): Change uregs's type
from int to uint32_t.
Reported-by: Sergey Bolshakov <sbolshakov@altlinux.org>
Check that syscall names and numbers defined in syscallent files
match kernel __NR_* constants defined by <asm/unistd.h>.
Tested on various platforms, including the following combinations
of architectures and kernel headers:
x86_64: 4.4-rc, 4.3, 4.2, 4.1, 3.19, 3.17, 3.16, 3.12, 3.11,
3.10-rhel, 2.6.32-rhel, 2.6.27-sle, 2.6.18-rhel
i386: 4.4-rc, 4.3, 4.2, 4.1, 3.19, 3.17, 3.16, 3.11,
3.10-rhel, 2.6.32-rhel, 2.6.27-sle, 2.6.18-rhel
aarch64: 4.4-rc
alpha: 4.3
arm eabi: 4.4-rc, 4.2, 4.1
hppa: 3.18
ia64: 3.18
mips o32: 4.1
ppc: 3.18
ppc64: 3.10-rhel, 2.6.32-rhel
s390: 3.18
s390x: 3.18
sparc: 4.1
sparc: 3.18
x32: 3.19
The only platform which is known at this moment to fail the test
is CentOS-5 provided by OBS. On x86_64 instance it fails with
error: "prlimit64" syscall #300 is "fanotify_init" in syscallent.h
and on i586 it similarly fails with
error: "prlimit64" syscall #338 is "fanotify_init" in syscallent.h
So this is a real platform bug that is not likely to be worked around
on the strace side.
* tests/ksysent.c: New file.
* tests/ksysent.sed: Likewise.
* tests/ksysent.test: New test.
* tests/Makefile.am (AM_CPPFLAGS): Add -I$(builddir).
(check_PROGRAMS): Add ksysent.
(TESTS): Add ksysent.test.
(EXTRA_DIST): Add ksysent.sed.
(ksysent.h): New rule.
(BUILT_SOURCES, CLEANFILES): Add ksysent.h.
* tests/.gitignore: Add ksysent.
Printing of ia32 syscalls on ia64 must have been broken for a long time.
Do not pretend that it works, print ia32 syscalls using printargs
instead.
* linux/ia64/syscallent.h: Remove all redefinitions of sys_* macros.
Stop including"../dummy.h". Include "../i386/syscallent.h" with
SYS_FUNC_NAME macro temporarily redirected to printargs.
* tests/Makefile.am (fstat64_CFLAGS): Rename to fstat64_CPPFLAGS,
replace $(AM_CFLAGS) with $(AM_CPPFLAGS).
(fstatat64_CFLAGS): Rename to fstatat64_CPPFLAGS,
replace $(AM_CFLAGS) with $(AM_CPPFLAGS).
(ftruncate64_CFLAGS): Rename to ftruncate64_CPPFLAGS,
replace $(AM_CFLAGS) with $(AM_CPPFLAGS).
(lstat64_CFLAGS): Rename to lstat64_CPPFLAGS,
replace $(AM_CFLAGS) with $(AM_CPPFLAGS).
(mmap64_CFLAGS): Rename to mmap64_CPPFLAGS,
replace $(AM_CFLAGS) with $(AM_CPPFLAGS).
(newfstatat_CFLAGS): Rename to newfstatat_CPPFLAGS,
replace $(AM_CFLAGS) with $(AM_CPPFLAGS).
(stat64_CFLAGS): Rename to stat64_CPPFLAGS,
replace $(AM_CFLAGS) with $(AM_CPPFLAGS).
(statfs_CFLAGS): Rename to statfs_CPPFLAGS,
replace $(AM_CFLAGS) with $(AM_CPPFLAGS).
(truncate64_CFLAGS): Rename to truncate64_CPPFLAGS,
replace $(AM_CFLAGS) with $(AM_CPPFLAGS).
(uio_CFLAGS): Rename to uio_CPPFLAGS,
replace $(AM_CFLAGS) with $(AM_CPPFLAGS).
As tests-m*32/Makefile.am files now refer to proper arch specific
directories, it's possible to relocate x32 <asm/stat.h> replacement
to its arch specific directory.
* linux/x86_64/asm_stat.h: Move ...
* linux/x32/asm_stat.h: ... here, remove x32 guard.
* Makefile.am (strace_SOURCES): Remove linux/x86_64/asm_stat.h.
* tests/Makefile.am (ARCH_MFLAGS, AM_LDFLAGS): New variables.
(AM_CPPFLAGS): Use ARCH_MFLAGS.
* bootstrap: In tests-m32/Makefile.am and tests-mx32/Makefile.am,
add -m32 and -mx32, respectively, to ARCH_MFLAGS instead of AM_CFLAGS.
* configure.ac (arch_m32): Set to sparc on sparc64, powerpc on
powerpc64, arm on aarch64, i386 on x86_64 and x32, $arch in other cases.
(arch_mx32): Set to x32 on x86_64, $arch in other cases.
(AC_SUBST): Add arch_m32 and arch_mx32.
* bootstrap: Substitute @arch@ with @arch_m32@ in tests-m32/Makefile.am.
Substitute @arch@ with @arch_mx32@ in tests-mx32/Makefile.am.
* tests/times.c: Include <sys/syscall.h>.
(main): On systems where user's and kernel's long types are the same,
prefer direct times syscall over libc's times function because
the latter is more prone to return value truncation.
* tests/times.c (main): Reduce cpu time consumption fourfold,
make the parent process consume less cpu time than the child process
so that members of the structure returned by times syscall would be
more likely to contain different values.
As times syscall returns kernel's long value, it has to be printed as
RVAL_LUDECIMAL on systems where long type is less than kernel's long.
* times.c (SYS_FUNC(times)) [RVAL_LUDECIMAL && !IN_MPERS]:
Return RVAL_LUDECIMAL instead of RVAL_UDECIMAL.
As the kernel recognizes F_GETLK64, F_SETLK64, and F_SETLKW64 commands
in fcntl64 syscall only, do not parse their structures in fcntl parser.
* xlat/fcntlcmds.in: Move F_GETLK64, F_SETLK64, and F_SETLKW64 ...
* xlat/fcntl64cmds.in: ... here.
* fcntl.c: Include "xlat/fcntl64cmds.h".
(print_fcntl): Move printing of first two syscall arguments
and handling of F_GETLK64, F_SETLK64, and F_SETLKW64 commands ...
(SYS_FUNC(fcntl), SYS_FUNC(fcntl64)): ... here.
* tests/fcntl.c: New file, based on struct_flock.c.
* tests/fcntl64.c: Likewise.
* tests/struct_flock.c (test_flock_einval, create_sample): New functions.
(test_flock): Use test_flock_einval.
(test_flock64, main): Remove.
* tests/fcntl.test: New test.
* tests/fcntl64.test: Likewise.
* tests/struct_flock.test: Remove.
* tests/Makefile.am (check_PROGRAMS): Add fcntl and fcntl64,
remove struct_flock.
(TESTS): Add fcntl.test and fcntl64.test, remove struct_flock.test.
(EXTRA_DIST) Add struct_flock.c.
* tests/.gitignore: Add fcntl and fcntl64, remove struct_flock.
* tests/stat.c: Replace with a wrapper around lstatx.c
* tests/stat.test: Replace with a wrapper around statx.sh.
* tests/stat32-v.test: Remove.
* tests/stat32.c: Remove.
* tests/stat64-v.test: Remove.
* tests/stat64.c: New file.
* tests/stat64.test: New test.
* tests/Makefile.am (check_PROGRAMS): Replace stat32 with stat64.
(stat_CFLAGS): Rename to stat64_CFLAGS.
(TESTS): Replace stat32-v.test and stat64-v.test with stat64.test.
* tests/.gitignore: Replace stat32 with stat64.
Parametrize code from fstatat.c to support <asm/stat.h>
as an alternative to <sys/stat.h> and move it to a separate file.
* tests/statx.sh: New file, based on fstatat64.test.
* tests/fstatat64.test: Use it.
* tests/newfstatat.test: Likewise.
* tests/xstatx.c: New file, based on fstatat.c.
* tests/fstatat.c: Use it.
* tests/Makefile.am (EXTRA_DIST): Add statx.sh and xstatx.c.
* tests/fstatat64.c (FSTATAT_NAME): Change to TEST_SYSCALL_NAME.
* tests/newfstatat.c: Likewise.
For some ioctls, like from drm, the identification of the correct ioctl
is done by custom code. Specifying IOCTL_NUMBER_STOP_LOOKUP on
return of ioctl_decode_command_number() disables further calls to
ioctl_lookup().
* defs.h (IOCTL_NUMBER_UNKNOWN, IOCTL_NUMBER_HANDLED,
IOCTL_NUMBER_STOP_LOOKUP): Add macros representing ioctl number state
decoding.
* ioctl.c (SYS_FUNC(ioctl)): Skip ioctl_lookup() when
ioctl_decode_command_number() returns a value with
IOCTL_NUMBER_STOP_LOOKUP flag is set.
Suggested-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Signed-off-by: Gabriel Laskar <gabriel@lse.epita.fr>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
The syscall number of restart_syscall returned by the kernel in case
of ERESTART_RESTARTBLOCK is broken on x32 from the beginning, see
https://lkml.org/lkml/2015/11/30/790
Until that improves, skip the test on x32.
* tests/restart_syscall.c (main) [__x86_64__ && __ILP32__]: Return 77.
On x32, tcp->ext_arg is not initialized for i386 personality,
so tcp->u_arg has to be used instead.
* lseek.c (SYS_FUNC(lseek)) [X32]: Handle "current_personality == 1" case.
Some old kernel headers, Ubuntu 14.04 in particular, provide
<asm/stat.h> editions that are completely wrong for x32.
Workaround this issue by providing a replacement.
* linux/x86_64/asm_stat.h: New file.
* linux/x32/asm_stat.h: Likewise.
* Makefile.am (strace_SOURCES): Add them.
While -P option needs /proc to match descriptors, it does not use
/proc to match pathnames. As this test does not need descriptor
match, the /proc requirement can be safely lifted.
* tests/fstatat64.test: Lift /proc/self/fd/ requirement.
Move inclusion of arch specific files that define static functions to
the end of syscall.c.
* syscall.c (get_syscall_result_regs, get_error, getregs_old):
New forward declarations.
(arch_get_scno): Move forward.
Move inclusion of "get_scno.c", "get_syscall_args.c",
"get_syscall_result.c", "get_error.c", and "getregs_old.c"
to the end of file.
Apparently, there are only two types of instruction pointer printers
depending on the architecture: those that print a register that was
fetched earlier, and those that fetch a register themselves using upeek.
With this change, architectures of the first type have ARCH_PC_REG
defined in their arch_regs.c file, architectures of the first type
have ARCH_PC_PEEK_ADDR defined there, and the common code in syscall.c
uses these macros to print the instruction pointer.
* Makefile.am (EXTRA_DIST): Remove linux/*/print_pc.c.
* linux/*/print_pc.c: Remove.
* linux/aarch64/arch_regs.c(ARCH_PC_REG): Define macro.
* linux/arc/arch_regs.c(ARCH_PC_REG): Likewise.
* linux/arm/arch_regs.c(ARCH_PC_REG): Likewise.
* linux/avr32/arch_regs.c(ARCH_PC_REG): Likewise.
* linux/i386/arch_regs.c(ARCH_PC_REG): Likewise.
* linux/ia64/arch_regs.c(ARCH_PC_REG): Likewise.
* linux/metag/arch_regs.c(ARCH_PC_REG): Likewise.
* linux/mips/arch_regs.c(ARCH_PC_REG): Likewise.
* linux/nios2/arch_regs.c(ARCH_PC_REG): Likewise.
* linux/or1k/arch_regs.c(ARCH_PC_REG): Likewise.
* linux/powerpc64/arch_regs.c(ARCH_PC_REG): Likewise.
* linux/powerpc/arch_regs.c(ARCH_PC_REG): Likewise.
* linux/s390/arch_regs.c(ARCH_PC_REG): Likewise.
* linux/s390x/arch_regs.c(ARCH_PC_REG): Likewise.
* linux/sparc64/arch_regs.c(ARCH_PC_REG): Likewise.
* linux/sparc/arch_regs.c(ARCH_PC_REG): Likewise.
* linux/tile/arch_regs.c(ARCH_PC_REG): Likewise.
* linux/x32/arch_regs.c(ARCH_PC_REG): Likewise.
* linux/x86_64/arch_regs.c(ARCH_PC_REG): Likewise.
* linux/alpha/arch_regs.c(ARCH_PC_PEEK_ADDR): Define macro.
* linux/bfin/arch_regs.c(ARCH_PC_PEEK_ADDR): Likewise.
* linux/crisv10/arch_regs.c(ARCH_PC_PEEK_ADDR): Likewise.
* linux/crisv32/arch_regs.c(ARCH_PC_PEEK_ADDR): Likewise.
* linux/hppa/arch_regs.c(ARCH_PC_PEEK_ADDR): Likewise.
* linux/m68k/arch_regs.c(ARCH_PC_PEEK_ADDR): Likewise.
* linux/microblaze/arch_regs.c(ARCH_PC_PEEK_ADDR): Likewise.
* linux/sh64/arch_regs.c(ARCH_PC_PEEK_ADDR): Likewise.
* linux/sh/arch_regs.c(ARCH_PC_PEEK_ADDR): Likewise.
* linux/xtensa/arch_regs.c(ARCH_PC_PEEK_ADDR): Likewise.
* syscall.c (print_pc): Stop including "print_pc.c".
Use ARCH_PC_REG or ARCH_PC_PEEK_ADDR.
Make get_syscall_result.c files more self-contained. While they are
still being included by syscall.c, the latter no longer provides
a function framework for them.
* linux/alpha/get_syscall_result.c (get_syscall_result_regs): Define.
* linux/bfin/get_syscall_result.c: Likewise.
* linux/crisv10/get_syscall_result.c: Likewise.
* linux/hppa/get_syscall_result.c: Likewise.
* linux/m68k/get_syscall_result.c: Likewise.
* linux/microblaze/get_syscall_result.c: Likewise.
* linux/sh/get_syscall_result.c: Likewise.
* linux/sh64/get_syscall_result.c: Likewise.
* linux/xtensa/get_syscall_result.c: Likewise.
* syscall.c [!ARCH_REGS_FOR_GETREGSET && !ARCH_REGS_FOR_GETREGS]:
Define USE_GET_SYSCALL_RESULT_REGS macro.
[USE_GET_SYSCALL_RESULT_REGS]: Include "get_syscall_result.c".
(get_syscall_result) [USE_GET_SYSCALL_RESULT_REGS]: Stop including
"get_syscall_result.c". Call get_syscall_result_regs.
Requested-by: Denys Vlasenko <dvlasenk@redhat.com>
Make arch_sigreturn.c files more self-sustained. While they are
still being included by sigreturn.c, the latter no longer defines
arch_sigreturn function.
* linux/alpha/arch_sigreturn.c (arch_sigreturn): Define.
* linux/arm/arch_sigreturn.c: Likewise.
* linux/crisv10/arch_sigreturn.c: Likewise.
* linux/i386/arch_sigreturn.c: Likewise.
* linux/ia64/arch_sigreturn.c: Likewise.
* linux/m68k/arch_sigreturn.c: Likewise.
* linux/microblaze/arch_sigreturn.c: Likewise.
* linux/mips/arch_sigreturn.c: Likewise.
* linux/powerpc/arch_sigreturn.c: Likewise.
* linux/s390/arch_sigreturn.c: Likewise.
* linux/sparc/arch_sigreturn.c: Likewise.
* linux/tile/arch_sigreturn.c: Likewise.
* linux/x32/arch_sigreturn.c: Remove code.
Include "x86_64/arch_sigreturn.c".
* linux/x86_64/arch_sigreturn.c: Stop including "x32/arch_sigreturn.c".
Include "i386/arch_sigreturn.c" with arch_sigreturn temporarily defined
to i386_arch_sigreturn.
(arch_sigreturn): Define. Add x32 personality support there.
* sigreturn.c: Remove arch_sigreturn header and footer.
Requested-by: Denys Vlasenko <dvlasenk@redhat.com>
* kernel_types.h: New file.
* Makefile.am (strace_SOURCES): Add it.
* configure.ac (AC_CHECK_TYPES): Check for __kernel_long_t
and __kernel_ulong_t in <asm/posix_types.h>.
* dirent.c: Stop including <dirent.h>.
(kernel_dirent): Remove.
Include "kernel_types.h".
(print_old_dirent, SYS_FUNC(getdents)): Print variables
of kernel_ulong_t type using %Lu format.
* syscall.c (kernel_long_t, kernel_ulong_t): Remove.
Include "kernel_types.h".
* tests/getdents.c (kernel_dirent): Remove.
Include "kernel_types.h".
(print_dirent): Print variables of kernel_ulong_t type using %Lu format.
* defs.h (error_msg_and_help): New prototype.
* strace.c (error_msg_and_help): New function.
(error_opt_arg, init): Use it.
(usage): Remove unnecessary arguments.
* count.c: Use the new function.
x86 is the only personality supported on x86_64
that has 32-bit syscall arguments.
* io.c (print_llu_from_low_high_val) [X86_64]: Fix
current_personality == 2 case.
* fetch_struct_flock.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* fcntl.c (print_struct_flock64): New function.
(printflock64): Use fetch_struct_flock64 and print_struct_flock64.
(printflock): Use fetch_struct_flock and print_struct_flock64.
* fcntl.c (USE_PRINTFLOCK64): Remove macro.
(printflock64): Define unconditionally.
(SYS_FUNC(fcntl)): Do not check for USE_PRINTFLOCK64,
F_DUPFD_CLOEXEC, F_NOTIFY, F_SETLEASE, and F_GETLEASE macros.
Include <linux/fcntl.h> for proper definitions of F_* constants
and flock structures.
* configure.ac (AC_CHECK_TYPES): Check for struct flock, struct flock64,
struct __kernel_flock, and struct __kernel_flock64 in <linux/fcntl.h>.
* flock.h: New file.
* Makefile.am (strace_SOURCES): Add it.
* fcntl.c: Include it instead of <fcntl.h>.
(struct flock64): Remove.
(printflock): Use struct_kernel_flock instead of struct flock.
(printflock64): Use struct_kernel_flock64 instead of struct flock64.
Some systems (like Bionic) omit support for SysV related code. That
means no C library headers for strace to include. Add configure tests
to probe the headers from the kernel and use them when they are
available.
It might make more sense to never rely on the C library's headers as
there is no guarantee or requirement that the structure layout between
apps and the C library match that what is passed to the kernel.
* configure.ac (AC_CHECK_HEADERS): Check for linux/ipc.h,
linux/mqueue.h, linux/msg.h, linux/sem.h, linux/shm.h,
sys/ipc.h, sys/msg.h, sys/sem.h, and sys/shm.h.
* ipc_defs.h: Include <sys/ipc.h> or <linux/ipc.h> depending
on what is available.
* ipc_msg.c: Replace <sys/ipc.h> with "ipc_defs.h".
Fallback to <linux/msg.h> when available.
* ipc_msgctl.c: Include <sys/msg.h>, <asm/msgbuf.h>, or <linux/msg.h>
based on what is available. Note missing support for old ipc structs.
* ipc_sem.c: Include <sys/sem.h> or <linux/sem.h> depending
on what is available. Only decode sembuf when available.
* ipc_shm.c: Fallback to <linux/shm.h> when available.
* ipc_shmctl.c: Include <sys/shm.h>, <asm/shmbuf.h>, or <linux/shm.h>
based on what is available. Note missing support for old ipc structs.
* print_mq_attr.c: Fallback to <linux/mqueue.h> when available.
These system calls are not implemented in the Linux kernel.
* linux/dummy.h(getpmsg, putpmsg): Move to the list of deprecated
syscalls.
* Makefile.am (strace_SOURCES): Remove stream.c.
* stream.c: Remove.
* xlat/pmsgflags.in: Remove.
Since query_module syscall is present only in kernels before Linux 2.6
and strace does not support those kernels, there is no use to keep this
parser any longer.
* bjm.c: Do not include <sys/*>, "xlat/qm_which.h",
and "xlat/modflags.h".
(MOD_*, QM_*, module_symbol, module_info, SYS_FUNC(query_module)):
Remove.
* linux/dummy.h (sys_query_module): Add stub alias.
* xlat/modflags.in: Remove.
* xlat/qm_which.in: Remove.
Do not pass sizeof(struct cmsghdr) to every cmsg printer.
* net.c (printcmsghdr): Pass pointer to cmsg_data and length
of cmsg_data to print_cmsg_type_data.
(print_cmsg_type_data): Update prototype. Pass pointer to cmsg_data
and length of cmsg_data to cmsg printers.
(print_scm_rights, print_scm_creds, print_scm_security): Update.
Avoid duplicate definitions of xlat structures in files
compiled in mpers mode.
Each xlat file defines the corresponding xlat struct with
either global or local visibility using the following rules:
- if xlat struct declaration is available in defs.h,
a global definition is provided in regular mode,
and nothing is provided in mpers mode;
- otherwise, if xlat file is included by a mpers source file,
a global definition is provided in regular mode
(unless no mpers mode is supported on this architecture,
in that case, a local definition is provided instead),
and a declaration is provided in mpers mode;
- otherwise, a local definition is provided in regular mode,
and an error message is printed in mpers mode.
Fallback definitions of constants provided by xlat files
remain available in all modes.
* bootstrap: Invoke generate_mpers_am.sh before xlat/gen.sh.
* defs.h (adjtimex_modes, adjtimex_status, sigev_value): Remove.
* generate_mpers_am.sh: Generate mpers_xlat.h.
* Makefile.am (EXTRA_DIST): Add mpers_xlat.h.
* print_sigevent.c: Include "xlat/sigev_value.h" unconditionally.
* print_timex.c: Include "xlat/adjtimex_modes.h" and
"xlat/adjtimex_status.h" unconditionally.
* xlat/gen.sh (cond_xlat): Move printing of fallback definitions ...
(cond_def): ... here.
(gen_header): Check also mpers_xlat.h for global declarations
of xlat structures.
Process input file twice, first time print directives
only, second time print everything.
Wrap struct xlat into #ifdef IN_MPERS ... #else ... #endif.
* xlat/getrandom_flags.in: Cleanup.
* defs.h (dirent_types): New xlat prototype.
* dirent.c: Stop including "xlat/direnttypes.h".
(kernel_dirent): New typedef. Mpersify it.
(print_old_dirent): Use it instead of old_dirent_t.
(SYS_FUNC(getdents)): Use it instead of struct kernel_dirent.
Rename direnttypes to dirent_types.
(SYS_FUNC(getdents64)): Move ...
* dirent64.c: ... here. Rename direnttypes to dirent_types.
Include "xlat/dirent_types.h".
* Makefile.am (strace_SOURCES): Add dirent64.c.
* xlat/direnttypes.in: Rename to xlat/dirent_types.in.
Array should be enclosed by square brakets, and elements should be
separated by commas.
* dirent.c (SYS_FUNC(getdents), SYS_FUNC(getdents64)): Fix typos
in array output.
* tests/getdents.awk: Update regexps to match fixed output.
* tests/getdents.out: Update output.
Signed-off-by: Gabriel Laskar <gabriel@lse.epita.fr>
* net.c (print_scm_rights, print_scm_creds, print_scm_security): Change
to return void. Move printing of struct cmsghdr closing '}' ...
(printcmsghdr): ... here. Move type and data decoders to ...
(print_cmsg_type_data): ... new function.
Starting with commit v4.6-281-g7af9f35, we implicitly assume that
HAVE_SENDMSG is always defined. Therefore, the check for sendmsg
availability is redundant and could be safely removed.
* configure.ac (AC_CHECK_FUNCS): Remove sendmsg.
* linux/dummy.h [!HAVE_SENDMSG] (sys_recvmsg, sys_sendmsg): Remove
stub aliases.
* net.c [HAVE_SENDMSG]: Define unconditionally.
* syscall.c (dumpio) [HAVE_SENDMSG]: Likewise.
Starting with linux commit v4.3-rc3-33-ga342361, powerpc has
direct sysv ipc syscalls in addition to traditional ipc syscall.
* linux/powerpc/syscallent.h (semop, semget, semctl, semtimedop, msgsnd,
msgrcv, msgget, msgctl, shmat, shmdt, shmget, shmctl): New entries.
If the set of headers are unable to produce a valid list, printflags
will try to pass NULL to tprints which crashes. Add a sanity check
for this edge case.
* util.c (printflags): Check xlat->str is not NULL.
mpers_DEFS has to be forwarded to mpers.sh so that config.h would be
properly included by defs.h at preprocessor stage.
* Makefile.am (mpers-m%.stamp): Add the whole $(mpers_sh_opts) set of
options to mpers.sh's CPPFLAGS.
Starting with linux commit v4.3-rc3-1-g5b3f33e, m68k has direct
socket syscalls in addition to traditional socketcall syscall.
* linux/m68k/syscallent.h (socket, socketpair, bind, connect, listen,
accept4, getsockopt, setsockopt, getsockname, getpeername, sendto,
sendmsg, recvfrom, recvmsg, shutdown, recvmmsg, sendmmsg): New entries.
Starting with linux commit v4.2-rc1-64-g9dea5dc, x86 has direct
socket syscalls in addition to traditional socketcall syscall.
* linux/i386/syscallent.h (socket, socketpair, bind, connect, listen,
accept4, getsockopt, setsockopt, getsockname, getpeername, sendto,
sendmsg, recvfrom, recvmsg, shutdown): New entries.
There is an ambiguousness in *_printer_defs.h rules definition:
$ grep '^[^ ]*_defs\.h: ' Makefile.am
m%_defs.h: $(srcdir_mpers_source_files)
%_printer_defs.h: $(srcdir_mpers_source_files)
While it appears to work with GNU make, it has to be fixed nevertheless.
* Makefile.am (m%_defs.h): Rename to m%_type_defs.h.
(mpers_m32_targets): Rename m32_defs.h to m32_type_defs.h.
(mpers_mx32_targets): Rename mx32_defs.h to mx32_type_defs.h.
* mpers_type.h: Rename m32_defs.h to m32_type_defs.h,
mx32_defs.h to mx32_type_defs.h.
* .gitignore: Likewise.
Reported-by: Elliott Hughes <enh@google.com>
* tests/uio.c (main): Use descriptor number 0 in pread/pwrite
and preadv/pwritev syscalls.
* tests/uio.expected: Update regexps.
Reported-by: Lennart Sorensen <lsorense@csclub.uwaterloo.ca>
In files generated by mpers scripts, includes directives are taken from
original files where the type definition was done. This causes to
include defs.h in multiple files. defs.h includes printers.h header,
which is a generated header.
This patch add an explicit dependency to printers.h for the mpers
scripts.
* Makefile.am (mpers-m%.stamp): Add printers.h to order-only
prerequisites.
Signed-off-by: Gabriel Laskar <gabriel@lse.epita.fr>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
When a command number was decoded through ioctl_decode_command_number(),
there was no check for conflicts with other potential ioctls numbers.
For example:
ioctl(fd, MCE_GET_RECORD_LEN, &i);
output:
ioctl(3, MIXER_READ(1), 0x7ffddce74a58) = 0
instead of:
ioctl(3, MIXER_READ(1) or MCE_GET_RECORD_LEN, 0x7ffee435ce08) = 0
* ioctl.c (SYS_FUNC(ioctl)): Fix ioctl command number decoding
in case of conflicts.
* tests/ioctl.c (main): Add a case for command number conflicts.
Signed-off-by: Gabriel Laskar <gabriel@lse.epita.fr>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
* seccomp_fprog.h: New file.
* fetch_seccomp_fprog.c: New file.
* Makefile.am (strace_SOURCES): Add them.
* seccomp.c: Include "seccomp_fprog.h".
(print_seccomp_filter): Use fetch_seccomp_fprog.
* tests/select.c Do not include <string.h>.
Include <sys/syscall.h>.
[!__NR_select]: Skip the test.
(main): Remove pselect6 support, test just select syscall.
* tests/select.test: Remove all pselect6 checks, test just
select syscall.
* tests/pselect6.c: Do not include <assert.h> and <string.h>.
Include <sys/syscall.h>.
[!__NR_pselect6]: Skip the test.
(main): Skip the test instead of assertions.
Add some padding to the timespec structure.
* print_time.c (sprint_timeval): New mpers printer.
[ALPHA] (sprint_timeval32): New function.
* defs.h [ALPHA] (sprint_timeval32): New prototype.
(bitness_t, TIMEVAL_TEXT_BUFSIZE, printtv_bitness, sprinttv): Remove.
* desc.c (decode_select): Replace bitness parameter with two printers.
Use them instead of printtv_bitness and sprinttv.
(sys_oldselect, sys_select): Pass print_timeval and sprint_timeval
to decode_select.
[ALPHA] (sys_osf_select): Pass print_timeval32 and sprint_timeval32
to decode_select.
(pselect6): Pass print_timespec and sprint_timespec to decode_select.
* time.c (UTIME_NOW, UTIME_OMIT, current_time_t_is_compat,
struct timeval32, printtv_bitness, do_sprinttv, sprinttv): Remove.
* time.c (print_timezone): new function.
(sys_gettimeofday, sys_settimeofday): Use it instead of print_timeval
to print struct timezone.
[ALPHA] (sys_osf_gettimeofday, sys_osf_settimeofday): Use it
instead of print_timeval32 to print struct timezone.
* tests/xettimeofday.c: New file.
* tests/xettimeofday.test: New test.
* tests/Makefile.am (check_PROGRAMS): Add xettimeofday.
(TESTS): Add xettimeofday.
* tests/.gitignore: Add xettimeofday.
* defs.h (is_erestart): New prototype.
* syscall.c (is_erestart): New function.
* time.c (sys_nanosleep): Use is_erestart,
temporarily_clear_syserror, and restore_cleared_syserror.
* strace.c (trace): Clear the whole siginfo_t variable before passing it
to ptrace(PTRACE_GETSIGINFO) because the latter may initialize only a
part of the structure.
* print_time.c (print_itimerspec): New mpers printer.
* time.c (sys_timer_settime, sys_timer_gettime, sys_timerfd,
sys_timerfd_settime, sys_timerfd_gettime): Use it instead of printitv.
* defs.h (adjtimex_modes, adjtimex_status): New xlat prototypes.
* print_timex.c (struct_timex): New typedef. Mpersify it.
[IN_MPERS]: Do not include "xlat/adjtimex_modes.h"
and "xlat/adjtimex_status.h".
[SUPPORTED_PERSONALITIES > 1]: Remove.
(tprint_timeval): Remove.
(tprint_timex): Rename to print_timex. Mpersify it.
* time.c (do_adjtimex): Rename tprint_timex to print_timex.
* tests/adjtimex.c: New file.
* tests/adjtimex.test: New test.
* tests/Makefile.am (check_PROGRAMS): Add adjtimex.
(TESTS): Add adjtimex.test.
* tests/.gitignore: Add adjtimex.
* print_sigevent.c (print_sigevent): Always print names
of struct_sigevent fields along with their values.
Print signal name of sigev_signo for those sigev_notify values
that require valid signal numbers.
* tests/timer_create.c: New file.
* tests/timer_create.test: New test.
* tests/Makefile.am (check_PROGRAMS): Add timer_create.
(TESTS): Add timer_create.test.
* tests/.gitignore: Add timer_create.
* defs.h (sigev_value): New prototype.
(printsigevent): Remove.
* print_sigevent.c (struct_sigevent): New typedef. Mpersify it.
[IN_MPERS]: Do not include "xlat/sigev_value.h".
(printsigevent32): Remove.
(printsigevent): Rename to print_sigevent. Mpersify it.
* mq.c (sys_mq_notify): Rename printsigevent to print_sigevent.
* time.c (sys_timer_create): Likewise.
* defs.h (umove_ulong_array_or_printaddr): New prototype.
* util.c (umove_ulong_array_or_printaddr): New function.
* desc.c (sys_pselect6): Use it instead of open-coding
a proxy struct parser for different personalities.
* io.c (tprint_iov_upto): Use it instead of open-coding
a struct iovec parser for different personalities.
* kexec.c (print_kexec_segments): Use it instead of open-coding
a struct kexec_segment parser for different personalities.
(print_seg): Pass arguments using an array.
Rename the recently added function to reflect what kind of data type
it actually deals with: unsigned long int, without sign extension.
* defs.h (umove_long_or_printaddr): Rename to umove_ulong_or_printaddr,
chage 3rd argument to unsigned long *.
* util.c (umove_long_or_printaddr): Likewise.
* aio.c (sys_io_submit): Update callers.
On some systems missing gnu/stubs-*.h files make -m32/-mx32 compilation
checks fail. As we want to support multiple personalities despite of
this limitation, workaround the check by creating empty gnu stub files
if necessary.
* m4/mpers.m4 (st_MPERS): Check for the gnu stub file. Create an empty
gnu stub file if the system one is missing.
* Makefile.am (DISTCLEANFILES): Add gnu/stubs-32.h and gnu/stubs-x32.h.
Fix multiple personalities support in parser of times syscall by
mpersifying struct tms.
* times.c (tms_t): New typedef. Mpersify it.
(sys_times): Use it instead of struct tms.
Fix multiple personalities support in parser of sysinfo syscall by
mpersifying struct sysinfo.
* sysinfo.c (sysinfo_t): New typedef. Mpersify it.
(sys_sysinfo): Use it instead of struct sysinfo.
Make parser of msgctl syscall print struct msqid_ds.
* ipc_msgctl.c (msqid_ds_t): New typedef. Mpersify it.
(print_msqid_ds): New function. Use msqid_ds_t.
(sys_msgctl): Use print_msqid_ds.
* tests/ipc_msg.c: Update for struct msqid_ds support.
Make parser of shmctl syscall print struct shmid_ds.
* ipc_shmctl.c (shmid_ds_t): New typedef. Mpersify it.
(print_shmid_ds): New function. Use shmid_ds_t.
(sys_shmctl): Use print_shmid_ds.
* tests/ipc_shm.c: Update for struct shmid_ds support.
Fix multiple personalities support in tprint_msgsnd and tprint_msgrcv
by converting tprint_msgbuf into a mpersified printer.
* print_msgbuf.c (msgbuf_t): New typedef. Mpersify it.
(tprint_msgbuf): Use it instead of long.
Wrap function name into MPERS_PRINTER_DECL.
* ipc_msg.c (tprint_msgsnd, tprint_msgrcv): Use mpersified
tprint_msgbuf.
Fix multiple personalities support in parsers of mq_open and
mq_getsetattr by converting printmqattr into a mpersified printer.
* print_mq_attr.c (mq_attr_t): New typedef. Mpersify it.
(printmqattr): Use it instead of struct mq_attr. Wrap function name
into MPERS_PRINTER_DECL.
* mq.c (mq_open, mq_getsetattr): Use mpersified printmqattr.
Fix multiple personalities support in parsers of getrusage, waitpid,
wait4, and waitid syscalls by converting printrusage into a mpersified
printer.
* defs.h (printrusage): Remove.
* printrusage.c (rusage_t): New typedef. Mpersify it.
(printrusage): Use it instead of struct rusage.
Wrap function name into MPERS_PRINTER_DECL.
* resource.c (sys_getrusage): Use mpersified printrusage.
* wait.c (printwaitn, sys_waitid): Likewise.
Fix multiple personalities support in parsers of ptrace,
rt_sigqueueinfo, rt_tgsigqueueinfo, rt_sigtimedwait and waitid syscalls
by converting printsiginfo_at into a mpersified printer.
* printsiginfo.c (siginfo_t): Mpersify.
(printsiginfo_at): Wrap function name into MPERS_PRINTER_DECL.
* defs.h (printsiginfo_at): Remove.
* signal.c (print_sigqueueinfo, sys_rt_sigtimedwait): Use mpersified
printsiginfo_at.
* process.c (sys_ptrace): Likewise.
* wait.c (sys_waitid): Likewise.
Fix multiple personalities support in parser of sigaltstack syscall by
mpersifying stack_t typedef.
* sigaltstack.c (stack_t): Mpersify.
(print_stack_t): Use it.
Fix multiple personalities support in parser of utime syscall by using
mpersified utimbuf_t typedef.
* utime.c (utimbuf_t): New typedef. Mpersify it.
(sys_utime): Use it instead of a locally defined union.
Fix multiple personalities support in parsers of utimes, futimesat,
and utimensat syscalls by introducing two mpersified printers:
print_timeval_pair and print_timespec_utime_pair.
* print_time.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* utimes.c (decode_utimes): Remove.
(sys_utimes, sys_futimesat): Use print_timeval_pair instead
of decode_utimes.
(sys_utimensat): Use print_timespec_utime_pair instead of decode_utimes.
Add a subsystem for semi-automatical definition of how parsers should
work with personality-dependent (mpers) types of tracee's data. Create
auxiliary libraries containing mpers syscall parsers and printer
functions, one library for each possible nonnative target personality.
Currently some parsers do not handle differences in definitions of data
types between personalities, namely LP64 and ILP32. When
this is the case, long integers, pointers, and all compound
types containing long and pointer members may be printed incorrectly,
because of differences in sizes, offsets and alignments.
Since in most cases these are the only differences in desired behaviour
of parsers and printers for different personalities, a correct way
would be to compile one source code into multiple parsers, differing
only in definitions of mpers types.
To get a definition of a given type for nonnative personality
a very basic .c file containing a declaration of a variable of this type
is being compiled for this personality (using -m32 or -mx32 compiler
flag). Information about the type is then being extracted from
this binary's DWARF debug info with an awk script and put
into a corresponding header file. Resulting headers are being used to
compile mpers variations of syscall parsers and printer functions.
In addition to syscall parsers, there can occur a need to create mpers
printing functions, which then can be called from many places
in the code (for example, printsiginfo_at). Such functions (printers)
are marked in a special manner.
For each possible nonnative target personality a library is being
created, containing mpers variations of syscall parsers and printers.
Only syscall parsers from files marked in a special manner and specially
marked functions from such files are being recompiled and included
in these libraries.
generate_mpers_am.sh is called by bootstrap to find the files
from strace_SOURCES which include MPERS_DEFS. During compilation,
these files are being inspected for inclusions of DEF_MPERS_TYPE,
and nonnative variations of each included type are being generated
by an awk script.
Mpers parser names are being modified during inclusions of syscallent
headers for nonnative personalities. Pointers to printers are
being stored in structs struct_printers, and a master
pointer printers is being updated on every set_personality.
* README-mpers: New README explaining how to use mpers support.
* empty.h: New empty file.
* generate_mpers_am.sh: New file.
* mpers.awk: Likewise.
* mpers.sh: Likewise.
* mpers_test.sh: Likewise.
* mpers_type.h: Likewise.
* Makefile.am (strace_SOURCES): Add empty.h and mpers_type.h.
(strace_CPPFLAGS, strace_LDFLAGS, strace_LDADD): Move to the beginning
of the file.
(strace_LDADD, noinst_LIBRARIES): Add libmpers-%.a.
(EXTRA_DIST): Add mpers.awk, mpers.sh, mpers_test.sh.
(BUILT_SOURCES, CLEANFILES): Add new generated files:
native_printer_decls.h, native_printer_defs.h, printers.h,
[HAVE_M32_MPERS] $(mpers_m32_targets), and [HAVE_MX32_MPERS]
$(mpers_mx32_targets).
(mpers_NAME, mpers_PREFIX, mpers_DEFS, mpers_INCLUDES, mpers_CPPFLAGS,
mpers_sh_opts, libmpers_CPPFLAGS, libmpers_m[x]32_a_SOURCES,
libmpers_m[x]32_a_CPPFLAGS, mpers_m[x]32_targets): New variables.
(mpers-m[x]32.stamp, m[x]32_defs.h, m[x]32_funcs.h, printers.h,
%_printer_decls.h, %_printer_defs.h, clean-local,
native_printer_decls.h, native_printer_defs.h, $mpers_m[x]32_targets):
New targets.
* bootstrap: Add generate_mpers_am.sh.
* configure.ac: Add AC_PROG_RANLIB.
* m4/mpers.m4: Add HAVE_MPERS variable. Add $st_cv_mpers checks.
* defs.h: Include mpers_type.h.
Include printers.h, native_printer_decls.h, define MPERS_PRINTER_NAME.
Redefine SYS_FUNC_NAME. Define MPERS_PRINTER_DECL.
[HAVE_M32_MPERS]: define PERSONALITY1_INCLUDE_FUNCS,
PERSONALITY1_INCLUDE_PRINTERS_DECLS, PERSONALITY1_INCLUDE_PRINTERS_DEFS
for X86_64, X32.
[HAVE_MX32_MPERS]: define PERSONALITY2_INCLUDE_FUNCS,
PERSONALITY2_INCLUDE_PRINTERS_DECLS, PERSONALITY2_INCLUDE_PRINTERS_DEFS
for X86_64.
Add fallback definitions of
PERSONALITY1_INCLUDE_FUNCS, PERSONALITY2_INCLUDE_FUNCS,
PERSONALITY0_INCLUDE_PRINTERS_DECLS, PERSONALITY0_INCLUDE_PRINTERS_DEFS,
PERSONALITY1_INCLUDE_PRINTERS_DECLS, PERSONALITY1_INCLUDE_PRINTERS_DEFS,
PERSONALITY2_INCLUDE_PRINTERS_DECLS, PERSONALITY2_INCLUDE_PRINTERS_DEFS.
* syscall.c: Include PERSONALITY1_INCLUDE_FUNCS,
PERSONALITY2_INCLUDE_FUNCS,
PERSONALITY0_INCLUDE_PRINTERS_DECLS, PERSONALITY0_INCLUDE_PRINTERS_DEFS,
PERSONALITY1_INCLUDE_PRINTERS_DECLS, PERSONALITY1_INCLUDE_PRINTERS_DEFS,
PERSONALITY2_INCLUDE_PRINTERS_DECLS, PERSONALITY2_INCLUDE_PRINTERS_DEFS.
(printers): New struct. Update it when needed.
* .gitignore: Add libmpers-m32.a, libmpers-mx32.a, m32_defs.h,
m32_funcs.h, m32_printer_decls.h, m32_printer_defs.h, mpers-m32,
mpers-m32.stamp, mpers-mx32, mpers-mx32.stamp, mpers.am, mx32_defs.h,
mx32_funcs.h, mx32_printer_decls.h, mx32_printer_defs.h,
native_printer_decls.h, native_printer_defs.h, and printers.h.
The first argument of match_diff should be the file with expected
output, the second argument - the file with actual output.
* tests/aio.test: Swap match_diff arguments.
* tests/restart_syscall.test: Likewise.
* tests/seccomp.test: Likewise.
* tests/umovestr2.test: Likewise.
Introduce a new helper function to fetch tracee's long integers
and use it to fix multiple personalities support in io_submit parser.
* defs.h (umove_long_or_printaddr): New prototype.
* util.c (umove_long_or_printaddr): New function.
* aio.c (sys_io_submit): Use it to fetch tracee's pointers.
* tests/aio.c: New file.
* tests/aio.test: New test.
* tests/Makefile.am (check_PROGRAMS): Add aio.
(TESTS): Add aio.test.
* tests/.gitignore: Add aio.
Currently the code assumes the set of valid options between getsockopt
and setsockopt are exactly the same and thus maintains one list. The
kernel unfortunately does not do this -- it allows for different opts
between the get and set functions. See the {g,s}et_opt{min,max} fields
in the various netfilter subcores.
To support this, extend the printxval function to take multiple sets of
xlats as varargs. Then we add the new get/set lists, and pass them down
in the net code when decoding things.
A simple example is iptables; before:
getsockopt(4, SOL_IP, 0x40 /* IP_??? */, ...) = 0
getsockopt(4, SOL_IP, 0x41 /* IP_??? */, ...) = 0
after:
getsockopt(4, SOL_IP, IPT_SO_GET_INFO, ...) = 0
getsockopt(4, SOL_IP, IPT_SO_GET_ENTRIES, ...) = 0
If these were setsockopt calls, then 0x40 & 0x41 would be
IPT_SO_SET_REPLACE & IPT_SO_SET_ADD_COUNTERS.
* configure.ac: Check for netfilter headers.
* defs.h (printxvals): New prototype.
(printxval): Change to a define.
* net.c: Include netfilter headers and new sockopts headers.
(print_sockopt_fd_level_name): Add a is_getsockopt argument. Change SOL_IP
and SOL_IPV6 decoding to use printxvals, and use is_getsockopt to pass more
xlats down.
(getsockopt): Call print_sockopt_fd_level_name with is_getsockopt as true.
(setsockopt): Call print_sockopt_fd_level_name with is_getsockopt as false.
* util.c (printxval): Rename to ...
(printxvals): ... this. Rewrite to be varargs based.
* xlat/getsockipoptions.in: New xlat list.
* xlat/getsockipv6options.in, xlat/setsockipoptions.in,
xlat/setsockipv6options.in: Likewise.
When msgrcv syscall is an ipc subcall, msgp (pointer to struct msgbuf)
and msgtyp (message type) syscall arguments are passed via proxy
structure which definition significantly depends on tracee's
wordsize.
* ipc_msg.c (fetch_msgrcv_args): New function.
(sys_msgrcv): Use it.
* tests/sendfile.c: New file.
* tests/sendfile64.c: New file.
* tests/sendfile.test: New test.
* tests/sendfile64.test: New test.
* tests/Makefile.am (check_PROGRAMS): Add sendfile and sendfile64.
(TESTS): Add sendfile.test and sendfile64.test.
* tests/.gitignore: Add sendfile and sendfile64.
When sendfile is called with a valid pointer to a file offset variable,
kernel updates this variable on successfull exit from syscall.
* sendfile.c (sys_sendfile, sys_sendfile64): Print tcp->u_arg[2]
on exiting syscall as well as on entering.
When receiving SIGSYS, the si_syscall & si_arch fields are set to known
values, so make sure we decode their values into the symbol settings.
This makes stracing seccomp failures much easier.
* defs.h (syscall_name): New prototype.
* printsiginfo.c: Include linux/audit.h and xlat/audit_arch.h.
(print_si_info): Decode si_syscall & si_arch for SIGSYS.
* syscall.c (undefined_scno_name): Delete.
(syscall_name): New function.
(trace_syscall_entering): Change undefined_scno_name to syscall_name.
(trace_syscall_exiting): Likewise.
* xlat/audit_arch.in: New file.
struct sysinfo members totalhigh, freehigh and mem_unit are present
since Linux 2.3.48.
* configure.ac (AC_CHECK_MEMBERS): Remove struct sysinfo checks.
* sysinfo.c (sys_sysinfo): Remove HAVE_STRUCT_SYSINFO_* checks.
* net.c (sys_socket): Set RVAL_FD flag in the return code.
* tests/net-yy-accept.awk: Update socket and bind regexps.
* tests/unix-yy-accept.awk: Likewise.
* tests/net-yy-connect.awk: Update socket and connect regexps.
* tests/unix-yy-connect.awk: Likewise.
* ipc.c: Move MSG_STAT and MSG_INFO definitions to xlat/msgctl_flags.in.
Move SHM_STAT and SHM_INFO definitions to xlat/shmctl_flags.in.
Move SEM_STAT and SEM_INFO definitions to xlat/semctl_flags.in.
Use SIGHUP instead of SIGABRT, the number of the former
is more portable across architectures.
* tests/ppoll.c (test2): Replace SIGABRT with SIGHUP.
* tests/ppoll.expected: Update regexp.
* tests/ppoll-v.expected: Likewise.
As struct sel_arg_struct is an array of 32-bit values,
fetch it using an intermediate array on 64-bit architectures.
* desc.c (sys_oldselect): Rename args to long_args.
[SIZEOF_LONG == 4] Alias oldselect_args to long_args.
[SIZEOF_LONG != 4] Introduce oldselect_args to fetch oldselect args.
* linux/sh64/syscallent.h (oldselect): Remove.
* execve.c (printargc): Do not assume that host and target pointers have
the same size.
* tests/execve.c: New file.
* tests/execve.expected: Likewise.
* tests/execve-v.expected: Likewise.
* tests/execve.test: New test.
* tests/Makefile.am (check_PROGRAMS): Add execve.
(TESTS): Add execve.test.
(EXTRA_DIST): Add execve.expected and execve-v.expected.
* tests/.gitignore: Add execve.
The use of printargs introduced by commit v4.10-179-g210a6b6 is wrong:
when the first argument is already printed, printargs cannot be called.
* prctl.c (print_prctl_args): New function.
(sys_prctl): Use it.
On some obscure systems, -lpthread is more likely to work than -pthread.
* tests/Makefile.am (filter_unavailable_LDFLAGS): Rename to
filter_unavailable_LDADD and change to -lpthread.
Change linux/mips/syscallent-compat.h to use designated initializers
like normal syscallent.h files.
* linux/mips/syscallent-compat.h: Use designated initializers.
* aio.c (sys_io_setup): Use printnum_int64 and printnum_long.
(sys_io_destroy, sys_io_submit): Update for RVAL_DECODED.
(sys_io_submit): Use umove_or_printaddr and printaddr.
Enclose iocb array in square brackets.
Change to traditional structure output format.
(print_io_event): New function.
(sys_io_cancel, sys_io_getevents): Use it.
(sys_io_cancel): Use umove_or_printaddr and printaddr.
(sys_io_getevents): Use printaddr.
* desc.c (printflock64, printflock): Use umove_or_printaddr.
(decode_select): Do not use xmalloc. Use umoven_or_printaddr
and printaddr.
(sys_oldselect): Use printaddr.
(sys_epoll_ctl): Use umoven_or_printaddr and printaddr.
(print_epoll_event_array): New function.
(sys_epoll_wait_common): Use it.
(sys_pselect6): Use umove_or_printaddr.
(sys_eventfd2): Use printaddr.
* dirent.c (print_old_dirent): Use umove_or_printaddr.
(sys_readdir): Use printaddr.
(sys_getdents, sys_getdents64): Do not use xmalloc for dirents,
treat malloc failure the same way as umoven failure. Use printaddr.
* io.c (sys_read, sys_pread): Use printaddr.
(tprint_iov_upto): Do not fetch data in case of syserror.
Use printaddr.
(sys_readv, sys_preadv): Remove redundant check for syserror.
(print_off_t): Use printnum_int and printnum_long.
(print_loff_t): Use umove_or_printaddr.
* ipc.c (sys_msgctl): Use printaddr.
(tprint_msgbuf): New function.
(tprint_msgsnd, tprint_msgrcv): Use it.
(sys_msgrcv): Use umove_or_printaddr.
(tprint_sembuf): Rename to tprint_sembuf_array.
(tprint_sembuf): New function.
(tprint_sembuf_array): Use it.
(sys_semop, sys_semtimedop): Update callers.
(sys_shmctl, sys_shmat, sys_shmdt): Use printaddr.
(sys_mq_open, printmqattr): Use printaddr and umove_or_printaddr.
* tests/ipc_msg.c (main): Update msgctl IPC_RMID regexp.
* tests/ipc_shm.c (main): Update shmctl IPC_RMID regexp.
* kexec.c (print_seg): New function.
(print_kexec_segments): Use it. Use printaddr and umove_or_printaddr.
(kexec_load): Use printaddr. Update for RVAL_DECODED.
* printsiginfo.h: New file.
* Makefile.am (strace_SOURCES): Add it.
* defs.h: Do not include <signal.h>.
(printsiginfo): Remove.
* printsiginfo.c: Include <signal.h> and "printsiginfo.h".
* strace.c: Likewise.
* sigaltstack.c: Include <signal.h>.
* signal.c: Likewise.
* sigreturn.c: Likewise.
* time.c: Likewise.
* quota.c (decode_cmd_data): Use umove_or_printaddr.
(sys_quotactl): Do not disable the parser in verbose mode.
Decode all "set" commands on entering syscall.
* sched.c (sys_sched_setscheduler, sys_sched_getparam,
sys_sched_setparam): Use printnum_int for sched_param decoding.
(sched_rr_get_interval): Use printaddr.
* signal.c (print_sigset_addr_len, decode_old_sigaction):
Use printaddr and umove_or_printaddr.
(sys_signal): Use printaddr.
(sys_sigprocmask, sys_sigpending, sys_rt_sigprocmask,
sys_rt_sigpending): Remove redundant checks for syserror.
(decode_new_sigaction): Use umove_or_printaddr.
* time.c (sprinttv): Do not fetch data in case syserror.
Print address instead of {...} in case of umove failure.
(sys_gettimeofday, sys_osf_gettimeofday, sys_adjtime, sys_clock_gettime,
clock_nanosleep): Remove now redundant checks for syserror.
Make the test work on alpha that has getxuid syscall instead of getuid.
* tests/uid.awk (BEGIN): Update getuid regexp to match both getuid
and getxuid syscalls.
* tests/uid.c (main): Allow __NR_getxuid as an alternative
to __NR_getuid.
* tests/uid.test: If getuid syscall is not available, probe for
getxuid syscall.
* uid.c (get_print_uid): Use umoven_or_printaddr.
(sys_getresuid): Do not check for syserror, get_print_uid now does
the right thing.
(sys_setgroups, sys_getgroups): Add const qualifier to local variables
that are assigned once. Use printaddr.
* wait.c (waitid): Do not check for NULL the address passed
to printrusage.
(printwaitn): Likewise.
Use printaddr and umove_or_printaddr for fetching wait status.
After this change, umoven_or_printaddr and its callers, including
printnum_* and printpair_*, will not fetch data in !verbose mode.
* util.c (umoven_or_printaddr): Do not call umoven if !verbose.
With this change, printnum_* printers will not attempt to fetch data
in case of exiting(tcp) && syserror(tcp).
All "at exiting" decoders that call these printers have been made
ready for this change in previous commits.
* util.c (DEF_PRINTNUM, DEF_PRINTPAIR): Use umove_or_printaddr.
* prctl.c (prctl_exit): Use printnum_int and printnum_long.
(sys_arch_prctl): Do not check for syserror.
Decode as much data on entering syscall as possible.
* term.c (decode_termios, decode_termio, decode_winsize, decode_ttysize,
decode_modem_flags): New functions.
(term_ioctl): Use them. Update for RVAL_DECODED.
* pathtrace.c (pathtrace_match): Change all comparisons with function
pointers to a single SEN based switch.
* syscall.c (dumpio, trace_syscall_entering): Change all comparisons
with function pointers to comparisons with SENs.
* ipc.c (sys_ipc): Remove.
* sock.c (sys_socketcall): Likewise.
* linux/dummy.h (sys_ipc, sys_socketcall): New aliases.
* Makefile.am (BUILT_SOURCES, CLEANFILES): Add sen.h.
(syscallent_names, syscallent_patterns, syscallent_files): New variables.
(sen.h): New rule.
* defs.h (struct sysent): Add sen field.
* generate_sen.sh: New file.
* linux/syscall.h: Include "sen.h".
* syscall.c (SEN_NAME): New temporary macro.
(SEN): Use it to prepend syscall entry number.
* syscall.c (SEN): New temporary macro.
* linux/dummy.h (sys_printargs): New macro.
* linux/subcall.h: Wrap all sys_func entries using SEN.
* linux/*/syscallent*.h: Likewise.
Fix SG_IO ioctl output in case of failed umove
* scsi.c: Include <linux/ioctl.h> instead of <sys/ioctl.h>.
Update for RVAL_DECODED.
(print_sg_io_v3_req, print_sg_io_v3_res, print_sg_io_v4_req,
print_sg_io_v4_res): Do not print address in case of failed umove.
(scsi_ioctl): Use printaddr.
Decode as much data on entering syscall as possible.
* xlat/mtd_file_mode_options.in: New file.
* mtd.c: Include <linux/ioctl.h> instead of <sys/ioctl.h>.
Include "xlat/mtd_file_mode_options.h".
Update for RVAL_DECODED.
(mtd_ioctl, ubi_ioctl): Use printnum_int and umove_or_printaddr.
Decode as much data on entering syscall as possible.
* sock.c: Update for RVAL_DECODED.
(print_addr): Remove.
(print_ifreq_addr, print_ifreq, print_ifc_len, decode_ifconf): New
functions.
(sock_ioctl) Use them. Use umove_or_printaddr.
* block.c: Update for RVAL_DECODED.
Define BLKROTATIONAL and BLKZEROOUT.
(block_ioctl): Decode them. Use printnum_short, printnum_int,
printnum_long, printpair_int64, and umove_or_printaddr.
(print_blkpg_req, block_ioctl): Use umove_or_printaddr.
Decode as much data on entering syscall as possible.
* loop.c: Include <linux/ioctl.h> instead of <sys/ioctl.h>.
Update for RVAL_DECODED.
(decode_loop_info, decode_loop_info64): New functions.
(loop_ioctl): Use them. Decode LOOP_SET_STATUS, LOOP_SET_STATUS64,
LOOP_SET_FD, LOOP_CHANGE_FD, LOOP_CTL_ADD and LOOP_CTL_REMOVE on
entering syscall.
Print LOOP_SET_FD and LOOP_CHANGE_FD arguments using printfd.
* rtc.c: Update for RVAL_DECODED.
(print_rtc): Rename to print_rtc_time.
(decode_rtc_time, decode_rtc_wkalrm): New functions.
(rtc_ioctl): Use them.
[RTC_VL_READ]: Decode RTC_VL_READ.
Use printnum_long for RTC_IRQP_READ and RTC_EPOCH_READ decoding.
Add functions to fetch and print pairs of integer types.
Note that these printers do not attempt to fetch data
in case of exiting(tcp) && syserror(tcp).
printnum_* printers will follow as soon as all callers are made ready
for this change.
* defs.h (printpair_int, printpair_long, printpair_int64): New
prototypes.
* util.c (DEF_PRINTPAIR): New macro.
(printpair_int, printpair_long, printpair_int64): New functions.
Add printnum_short and printnum_int64 in addition to already existing
printnum_int and printnum_long.
* defs.h (printnum_short, printnum_int64): New prototypes.
* util.c (DEF_PRINTNUM): New macro.
(printnum_int, printnum_long): Use DEF_PRINTNUM.
(printnum_short, printnum_int64): New functions.
printaddr is a simple function implementing "print NULL or address"
idiom.
umoven_or_printaddr is a wrapper around umoven that has the same return
value semantics as umoven but also prints the address when the data
is not going to be fetched (in case of exiting(tcp) && syserror(tcp))
or cannot be fetched (umoven fails).
umove_or_printaddr is a macro wrapper around umoven_or_printaddr that
mirrors umove wrapper around umoven.
* defs.h (printaddr, umoven_or_printaddr): New prototypes.
(umove_or_printaddr): New macro.
* util.c (printaddr, umoven_or_printaddr): New functions.
Implement a method for "on entering" parsers to return
"done with decoding" information to their callers.
If a syscall parser called from trace_syscall_entering returns a value
with RVAL_DECODED bit set, record this value in tcp->sys_func_rval
and use it in trace_syscall_exiting instead of calling the parser on
exiting syscall.
If an ioctl parser returned RVAL_DECODED, this would tell
SYS_FUNC(ioctl) that the decoding is finished but fallback printing of
the ioctl command argument is needed, while e.g. RVAL_DONE+1 would mean
that the decoding is finished and no fallback decoding is needed.
* defs.h (struct tcb): Add sys_func_rval.
(RVAL_DECODED): New macro.
* syscall.c (trace_syscall_entering): Initialize tcp->sys_func_rval
along with setting TCB_INSYSCALL flag.
(trace_syscall_exiting): Use tcp->sys_func_rval instead of calling
tcp->s_ent->sys_func when the former has RVAL_DECODED flag set.
Reset tcp->sys_func_rval along with clearing TCB_INSYSCALL flag.
* ioctl.c (sys_ioctl): Set RVAL_DECODED flag on exiting.
Print ioctl argument iff RVAL_DECODED is set and fallback decoding
is requested.
If stat.st_mode returned by syscall is 0, it means that the definition
of struct stat in <asm/stat.h> is incorrect. This is the case e.g. on
mips n32 where the only available stat syscall is 64-bit but at the same
time struct stat defined by <asm/stat.h> is 32-bit.
* tests/stat.c (main) [NR_stat]: Return 77 if st_mode is 0.
* aio.c (print_common_flags): Use #ifdef in the check for
HAVE_STRUCT_IOCB_U_C_FLAGS.
* defs.h: Use #ifdef in the check for HAVE_LITTLE_ENDIAN_LONG_LONG.
* net.c: Use #ifdef in the check for HAVE_SENDMSG.
* syscall.c (dumpio): Likewise.
* printstat.h (DO_PRINTSTAT): Use #ifdef in checks for
HAVE_STRUCT_STAT_ST_* macros.
* util.c: Use #ifdef in checks for HAVE_SYS_XATTR_H.
* mount.c (sys_mount): Do not decode type and data strings for
MS_SHARED, MS_PRIVATE, MS_SLAVE, and MS_UNBINDABLE mount flags
that do not imply valid strings.
* configure.ac (AC_CHECK_FUNCS): Add inet_pton.
* net.c (print_mreq, print_mreq6): New functions.
(print_setsockopt): Use them to decode IP_ADD_MEMBERSHIP,
IP_DROP_MEMBERSHIP, IPV6_ADD_MEMBERSHIP, IPV6_DROP_MEMBERSHIP,
IPV6_JOIN_ANYCAST, and IPV6_LEAVE_ANYCAST.
* tests/ip_mreq.c: New file.
* tests/ip_mreq.expected: Likewise.
* tests/ip_mreq.test: New test.
* tests/Makefile.am (check_PROGRAMS): Add ip_mreq.
(TESTS): Add ip_mreq.test.
(EXTRA_DIST): ip_mreq.expected.
* tests/.gitignore: Add ip_mreq.
Based on patch by Ben Noordhuis <info@bnoordhuis.nl>.
Introduce wrappers to the following functions that do memory allocation:
malloc, calloc, realloc, strdup.
This commit is a follow-up to the related discussions in strace-devel ML:
http://sourceforge.net/p/strace/mailman/message/33618180/http://sourceforge.net/p/strace/mailman/message/33733470/
* defs.h (xmalloc, xcalloc, xreallocarray, xstrdup): New prototypes.
* xmalloc.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* count.c (count_syscall, call_summary_pers): Use xcalloc.
* desc.c (decode_select): Use xmalloc.
* dirent.c (sys_getdents, sys_getdents64): Likewise.
* net.c (sys_recvmmsg): Use xstrdup.
* pathtrace.c (storepath): Use xreallocarray.
(pathtrace_match): Use xmalloc.
* strace.c (die_out_of_memory): Move to xmalloc.c.
(expand_tcbtab): Use xcalloc and xreallocarray.
(startup_child): Use xstrdup.
(init): Use xmalloc, xcalloc, and xstrdup.
* syscall.c (reallocate_qual): Use xreallocarray.
(qualify): Use xstrdup.
* unwind.c (unwind_tcb_init): Use xmalloc.
(build_mmap_cache): Use xcalloc, xreallocarray, and xstrdup.
(get_symbol_name): Use xreallocarray.
(stacktrace_walk, queue_put): Use xmalloc.
* util.c (printstr): Use xmalloc.
* vsprintf.c (strace_vfprintf): Likewise.
Some ioctl constants are defined to different values on those 32-bit
architectures where structures containing a 64-bit field are aligned on
a 8-byte boundary and have a size multiple of 8 bytes, and on other
32-bit architectures.
* configure.ac: Check the size of a structure containing a 64-bit field.
* linux/32/ioctls_inc.h: Rename to ...
* linux/32/ioctls_inc_align32.h: ... this.
* linux/32/ioctls_inc_align64.h: New file.
* linux/32/ioctls_inc.h: Include either 32/ioctls_inc_align32.h
or ioctls_inc_align64.h depending on SIZEOF_STRUCT_I64_I32.
* Makefile.am (EXTRA_DIST): Add linux/32/ioctls_inc_align32.h and
* linux/32/ioctls_inc_align64.h.
* tests/ioctl.c (main): Test for VIDIOC_ENUMINPUT.
* tests/ioctl.expected: Check VIDIOC_ENUMINPUT decoding.
Reported-by: Philippe De Muyter <phdm@macqel.be>
This commit adds strace support for Altera's Nios-II. The architecture
is supported by Linux since v3.19, and it implements the generic syscall
ABI.
* configure.ac: Add nios2 support.
* cacheflush.c [NIOS2] (sys_cacheflush): New function.
* linux/nios2/*: New Nios-II specific files.
* Makefile.am (EXTRA_DIST): Add them.
Signed-off-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Fix bexecve.test when the executable being run does not match the
native architecture.
* tests/bexecve.test: Use "set_ptracer_any" instead of "sh".
Allow for a personality switch notification in the strace output.
Introduce SYS_FUNC macro to declare and define all syscall parsers.
* Makefile.am (BUILT_SOURCES, CLEANFILES): Add sys_func.h.
(sys_func.h): New rule.
* defs.h (SYS_FUNC_NAME, SYS_FUNC): New macros.
* linux/syscall.h: Include "sys_func.h".
[NEED_UID16_PARSERS]: Use SYS_FUNC to declare uid16 syscall parsers.
Remove other declarations.
* linux/alpha/syscallent.h (160, 161): Add sys_ prefix to osf_statfs
and osf_fstatfs syscall parsers.
* *.c: Use SYS_FUNC to define syscall parsers.
Some of our tests use GNU awk @include extension, but not all currently
used versions of GNU awk support it. Skip these tests if gawk does not
provide @include support.
* tests/init.sh (check_gawk): New function.
(match_awk): Use it.
The regular expression generated by umovestr2.c on systems with large
page size is too big for GNU grep. Rewrite the test to use diff
instead of grep.
* tests/umovestr2.test: Use match_diff instead of match_grep.
* tests/umovestr2.c (main): Convert output from regexp to plain text.
* defs.h (get_pagesize): New prototype.
* mem.c (get_pagesize) Make global.
* util.c (PAGMASK): Remove.
(vm_read_mem): New process_vm_readv proxy function.
(umoven, umovestr): Use it.
(umovestr): Read chunks up to pagesize at a time.
process_vm_readv() and ptrace(PTRACE_PEEKDATA) have inconsistent access
control rules wrt traced processes: process_vm_readv() is more likely to
fail with EPERM than ptrace(PTRACE_PEEKDATA) when tracing a process that
has execve'd a privileged executable.
* util.c (umoven, umovestr): If process_vm_readv returned EPERM,
fall back to ptrace(PTRACE_PEEKDATA).
Reported-by: Andrew Guertin <lists@dolphinling.net>
Define macros for gcc attributes that are already in use
or going to be used soon.
* defs.h (GNUC_PREREQ, ATTRIBUTE_NORETURN, ATTRIBUTE_FORMAT,
ATTRIBUTE_ALIGNED, ATTRIBUTE_PACKED, ATTRIBUTE_MALLOC,
ATTRIBUTE_NOINLINE, ATTRIBUTE_ALLOC_SIZE): New macros.
Fix a compilation failure with musl libc.
* evdev.c: Include <linux/ioctl.h>.
* ioctl.c: Include <linux/ioctl.h> instead of <asm/ioctl.h>.
* ioctlsort.c: Likewise.
Reported-by: Dima Krasner <dima@dimakrasner.com>
Acked-by: Mike Frysinger <vapier@gentoo.org>
* printsiginfo.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* defs.h (printsiginfo): Change second argument's type from int to bool.
* signal.c: Stop defining siginfo_t related constants.
Move inclusion of xlat/sig*_codes.h files to printsiginfo.c
(printsigsource, printsigval, printsiginfo_at): Move to printsiginfo.c.
(printsiginfo): Change second argument's type from int to bool,
split, and move to printsiginfo.c.
Personality 1 for sparc and sparc64 appears to be an old attempt
to support stracing Solaris binaries.
It stalled after the only syscall, solaris_open, was covered:
all other solaris_foo's are printargs.
This change deletes personality 1. Now sparc is an one-personality
arch, and sparc64 is a two-personality one.
For clarity, pure renaming of personality 2 files to personality 1 for sparc64
is performed in the next commit - this prevents a state where syscallent1.h
seems to indergo a lot of changes, where in fact in is deleted, and then
replaced by syscallent2.h
Therefore, tree at this commit will not build.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
Starting with commit v4.10-25-g8497b62,
arm$ ./strace true
pid 1234 stray syscall exit
Segmentation fault
Fix this by updating ARM syscall sanity check.
In particular, get_scno() should not set TCB_INSYSCALL flag because
other code assumes that s_ent is properly initialized when this flag
is set.
* linux/arm/get_scno.c: Check syscall number after fetching.
Do not apply the check to SCNO_IN_RANGE syscalls.
Do not set TCB_INSYSCALL flag.
Extend diagnostics and move it to debug level.
Split code that use arch-specific registers to separate arch files.
* sigreturn.c (sys_sigreturn): Move arch-specific code
to linux/*/arch_sigreturn.c, include "arch_sigreturn.c".
* linux/arch_sigreturn.c: New file.
* Makefile.am (EXTRA_DIST): Add linux/arch_sigreturn.c
and linux/*/arch_sigreturn.c files.
Split code that use arch-specific registers to separate arch files.
* syscall.c: Move definitions of variables containing fetched registers
to linux/*/arch_regs.c files.
[HAVE_GETRVAL2] (getrval2): Move arch-specific code
to linux/*/arch_getrval2.c, include "arch_getrval2.c".
(print_pc): Move arch-specific code to linux/*/print_pc.c files,
include "print_pc.c".
[X86_64] (x86_64_getregs_old): Rename to getregs_old, move to
linux/x86_64/getregs_old.c, include "getregs_old.c".
[POWERPC] (powerpc_getregs_old): Rename to getregs_old, move to
linux/powerpc/getregs_old.c, include "getregs_old.c".
(get_regs) [X86_64, POWERPC]: Update callers.
(get_scno): Move arch-specific code to linux/*/get_scno.c,
include "get_scno.c".
(get_syscall_args): Move arch-specific code
to linux/*/get_syscall_args.c, include "get_syscall_args.c".
(get_error): Move arch-specific code to linux/*/get_error.c,
include "get_error.c".
(get_syscall_result): Move arch-specific code
to linux/*/get_syscall_result.c, include "get_syscall_result.c".
* Makefile.am (EXTRA_DIST): Add new linux/*/*.c files.
Starting with commit v4.10-25-g8497b62, get_scno() can be called outside
trace_syscall_entering(), so move the get_regs_error check from
trace_syscall_entering() to get_scno().
* syscall.c (trace_syscall_entering): Move get_regs_error check ...
(get_scno): ... here.
When signal is received, or if we have attached to a process,
current syscall (if process is in one) gets restarted.
Some syscalls are restarted via "restart_syscall()" mechanism.
On such sycalls, we don't show _which_ syscall gets restarted.
IOW: users want to see "resuming interrupted nanosleep"
instead of "resuming interrupted call" when they attach to "sleep 999".
Kernel does expose this information. The only thing we need is
to fetch syscall# on attach, and save it.
This patch does this. It adds tcp->s_prev_ent, which is
a pointer to struct_sysent of the previous syscall of this tracee.
It can be NULL.
sys_restart_syscall() is made to use it when the message is generated.
To similarly handle restart_syscall() *after signals*, not just
on attach, on each syscall exit patch saves exited syscall's data
in the same member (tcp->s_prev_ent).
Example:
$ sleep 3 & strace -p $!
Process 8728 attached
restart_syscall(<... resuming interrupted nanosleep ...>) = 0
_exit(0) = ?
+++ exited with 0 +++
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
* linux/sparc/syscallent.h (mmap2): Decode with sys_mmap_4koff,
not sys_mmap.
* linux/sparc64/syscallent2.h (mmap2): Decode with sys_mmap_pgoff,
not sys_mmap.
Factor out awk code used in several tests to match.awk.
* tests/match.awk: New file.
* tests/Makefile.am (EXTRA_DIST): Add it.
* tests/caps.awk: Use it.
* tests/getdents.awk: Likewise.
* tests/getrandom.awk: Likewise.
* tests/select.awk: Likewise.
* tests/sigaction.awk: Likewise.
* tests/init.sh (match_awk): Use gawk not awk. Define AWKPATH.
* tests/getdents.test: Likewise.
According to arch/arm/kernel/sys_oabi-compat.c,
struct statfs64 has extra padding with EABI.
* statfs.c [AARCH64 || defined ARM] (COMPAT_STATFS64_PADDED_SIZE):
Define.
(do_statfs64_fstatfs64): New function, factored out from sys_statfs64.
[COMPAT_STATFS64_PADDED_SIZE]: Check it in addition to
sizeof(struct compat_statfs64).
(sys_statfs64, sys_fstatfs64): Use do_statfs64_fstatfs64.
Reported-and-tested-by: Elliott Hughes <enh@google.com>
Factor out shell code used in several tests to common functions.
* tests/fanotify_mark.expected: New file.
* tests/ioctl.expected: New file.
* tests/net-fd.expected: New file.
* tests/net.expected: New file.
* tests/statfs.expected: New file.
* tests/sun_path.expected: New file.
* tests/uio.expected: New file.
* tests/ipc.sh: New file.
* tests/Makefile.am (EXTRA_DIST): Add them.
* tests/init.sh (dump_log_and_fail_with, run_prog,
run_prog_skip_if_failed, run_strace, run_strace_merge,
match_awk, match_diff, match_grep): New functions.
* tests/*.test: Use them.
Newer architectures have no chown syscall, so use fchown* syscalls
for testing printuid.
* tests/uid.test: Use fchown instead of chown.
* tests/uid.c: Test __NR_fchown instead of __NR_chown.
(main): Use __NR_fchown instead of __NR_chown.
* tests/uid32.c: Test __NR_fchown32 instead of __NR_chown32.
(main): Use __NR_fchown32 instead of __NR_chown32.
* tests/uid16.c: Test __NR_fchown and __NR_fchown32 instead
of __NR_chown and __NR_chown32.
(main): Use __NR_fchown instead of __NR_chown.
* tests/uid.awk: Update regexp.
Newer architectures have no stat syscall, so stat() is implemented there
using newfstatat syscall.
* tests/stat.c (STAT_FNAME): Rename to STAT_PREFIX. Update callers.
[_FILE_OFFSET_BITS == 64] (STAT_PREFIX): Add newfstatat support.
(main) [!NR_stat]: Add newfstatat support.
Reported-by: Andreas Schwab <schwab@suse.de>
STRUCT_STAT.st_[acm]time are declared as unsigned int for some
personalities, while time_t is signed.
* printstat.h (DO_PRINTSTAT): If st_[acm]time have the same size as int,
explicitly cast them to int.
* tests/stat64-v.test: Test that negative time_t is decoded properly.
Reported-by: Andreas Schwab <schwab@suse.de>
On architectures where the semctl call is implemented by the ipc syscall
the 4th argument is passed by reference.
* ipc.c (sys_semctl): Handle the indirect ipc subcall case.
* tests/ipc_sem.c (main): Optionally match indirection
in the 4th argument of semctl calls.
Reported-by: Andreas Schwab <schwab@suse.de>
* tests/ipc_msg.c (main): Optionally match "IPC_64|" in the third
argument of the ipc call.
* tests/ipc_sem.c (main): Likewise.
* tests/ipc_shm.c (main): Likewise.
Properly use union semun as argument of semctl.
* tests/ipc_sem.c (main): Properly use union semun as argument of
semctl. Don't handle EFAULT specially.
* tests/ipc_sem.test: Revert last change.
* linux/aarch64/ioctls_inc0.h: Rename from ioctls_inc1.h.
* linux/aarch64/ioctls_inc1.h: Rename from ioctls_inc0.h.
* linux/aarch64/ioctls_arch0.h: Rename from ioctls_arch1.h.
* linux/aarch64/ioctls_arch1.h: Rename from ioctls_arch0.h.
* tests/select.awk (BEGIN): Update regexps to match both select
and pselect6 syscalls.
* tests/select.test: Probe for both select and pselect6 syscall.
This is [strace](http://strace.io) -- a diagnostic, debugging and instructional userspace utility for Linux. It is used to monitor interactions between processes and the Linux kernel, which include system calls, signal deliveries, and changes of process state. The operation of strace is made possible by the kernel feature known as [ptrace](http://man7.org/linux/man-pages/man2/ptrace.2.html).
strace is released under a Berkeley-style license at the request of Paul Kranenburg; see the file [COPYING](https://raw.githubusercontent.com/strace/strace/master/COPYING) for details.
See the file [NEWS](https://raw.githubusercontent.com/strace/strace/master/NEWS) for information on what has changed in recent versions.
Please send bug reports and enhancements to [the strace mailing list](https://lists.sourceforge.net/lists/listinfo/strace-devel).
list="$(sed -n '/^strace_SOURCES[[:space:]]*=/,/^[[:space:]]*# end of strace_SOURCES/ s/^[[:space:]]*\([[:alnum:]][^.]*\.c\)[[:space:]]*\\$/\1/p' Makefile.am |
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.
