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.
There is a race condition between the moment a netlink socket is created
and the moment it is reported via SOCK_DIAG_BY_FAMILY interface.
Add one more operation on the socket created by create_nl_socket
to increase chances of winning the race.
* tests/create_nl_socket.c (create_nl_socket_ext): Add a getsockopt call
with the created socket.
Starting with linux commit v4.0-rc4~21^2, "SP" register
is available via regs[PTR_SP] member of struct user_pt_regs.
* linux/nios2/arch_regs.h (nios2_sp_ptr): Change type
to "unsigned int *".
* linux/nios2/arch_regs.c (nios2_sp_ptr): Likewise.
Replace .sp with .regs[PTR_SP].
Fixes: http://autobuild.buildroot.net/results/b9f/b9fc25b82f3280872fe1593ac252a8529ba83576
Fixes: v4.16-34-g6117728a ("nios2: export nios2_sp_ptr")
Signed-off-by: Romain Naour <romain.naour@smile.fr>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
Check that syscall counters used for tampering are per-tcb.
* tests/qual_fault.c: Include <fcntl.h> for open(),
<sys/param.h> for PATH_MAX, and <sys/wait.h> for wait().
(expfd): Rename to exp_fd, remove const qualifier and initialization.
(gotfd): Rename to got_fd, remove const qualifier and initialization.
(out_fd): New variable, for the expected strace output.
(open_file): New helper function.
(main): Remove st, add num_procs, proc, exp_prefix, got_prefix,
out_prefix, pid_prefix. Remove stat asserts, update argc assert.
Add num_procs, exp_prefix, got_prefix, out_prefix, pid_prefix
initialization. Add per-process loop, open expfd, gotfd, out_fd, pidfd
in each one, print process's pid to pidfd, print exit message to out_fd.
Add wait() call for each forked process.
* tests/qual_fault.test (N): increase to 100 in order to check
concurrent process execution.
(check_fault_injection): Take additional argument for the process
count, pass it to qual_fault, add filename variables for expected strace
output and pid file, pass them to qual_fault.
Specify -ff parameter to strace invocation.
Compare write and strace output for the each process.
Update all check_fault_injection invocation with process count argument,
add an invocation with 4 concurrent processes.
* tests/init.sh (TIMEOUT_DURATION): Raise from 120 to 300.
Co-authored-by: Dmitry V. Levin <ldv@altlinux.org>
* strace.c (init): Adjust argc along with argv,
check that adjusted argc is sane.
* tests/zeroargc.c: New file.
* tests/.gitignore: Add zeroargc.
* tests/Makefile.am (check_PROGRAMS): Likewise.
* tests/options-syntax.test: Check strace error diagnostics
in case of zero argc.
Emulate program_invocation_name only if it is not provided by libc.
* configure.ac: Check for program_invocation_name variable.
* strace.c (progname): Remove.
[!HAVE_PROGRAM_INVOCATION_NAME] (program_invocation_name): New variable.
(verror_msg, error_msg_and_help): Use it instead of progname.
(init): Initialize program_invocation_name instead of progname.
Introduce a separate header file for generic macros shared between
strace and its tests. Start this unification with ARRAY_SIZE macro.
* defs.h: Include "macros.h".
(ARRAY_SIZE): Move ...
* macros.h: ... to new file.
* Makefile.am (strace_SOURCES): Add it.
* tests/tests.h: Include "macros.h".
(ARRAY_SIZE): Remove. All callers updated.
"sp" register is a member of "scratch" structure (which is nested
in user_regs_struct)
* linux/arc/arch_regs.c: Replace .sp with .scratch.sp .
Fixes: v4.16-31-g134042a5 ("arc: export arc_sp_ptr")
Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
* tests/nlattr_unix_diag_msg.c: Include "test_nlattr.h".
(test_unix_diag_vfs, test_unix_diag_icons, test_unix_diag_rqlen):
Remove.
(init_unix_diag_msg): Add const qualifier to arguments and variables.
(print_unix_diag_msg, print_uint): New functions.
(main): Use them and macros from test_nlattr.h file.
Add const qualifiers to auto variables and function arguments.
This change does not affect the code generated by the compiler,
the purpose of these "const" qualifiers is to highlight the intent.
* netlink_sock_diag.c (decode_unix_diag_vfs, decode_unix_diag_inode,
decode_unix_diag_rqlen): Add const qualifier to tcp, addr, and len
arguments.
* netlink_sock_diag.c (unix_diag_msg_nla_decoders): New array.
(decode_unix_diag_msg): Use it.
* linux/unix_diag.h (unix_diag_vfs, unix_diag_rqlen): New structures.
* tests/nlattr_netlink_diag_msg.c: Include "test_nlattr.h".
(test_netlink_diag_groups, test_netlink_diag_rx_ring,
test_netlink_diag_flags): Remove.
(init_netlink_diag_msg): Add const qualifier to arguments and variables.
(print_netlink_diag_msg, print_xlong): New functions.
(main): Use them and macros from test_nlattr.h file.
Add const qualifiers to auto variables and function arguments.
This change does not affect the code generated by the compiler,
the purpose of these "const" qualifiers is to highlight the intent.
* netlink_sock_diag.c (print_group, decode_netlink_diag_groups,
decode_netlink_diag_ring, decode_netlink_diag_flags): Add const
qualifier to tcp, addr, and len arguments.
* linux/netlink_diag.h (netlink_diag_ring): New structure.
(NDIAG_FLAG_*): New macros.
* netlink_sock_diag.c: Include "xlat/netlink_socket_flags.h".
(print_group, decode_netlink_diag_groups, decode_netlink_diag_ring,
decode_netlink_diag_flags): New functions.
(netlink_diag_msg_nla_decoders): New array.
(decode_netlink_diag_msg): Use it.
* xlat/netlink_socket_flags.in: New file.
* netlink_sock_diag.c (print_inet_diag_sockid): Print idiag_if field
using print_ifindex.
* tests/nlattr_inet_diag_msg.c: Include <net/if.h>.
(IFINDEX_LO): New macro.
(init_inet_diag_msg): Set inet_diag_sockid.idiag_if field.
(print_inet_diag_msg): Update expected output.
* tests/test_nlattr.h: New file.
* tests/Makefile.am (libtests_a_SOURCES): Add it.
* tests/nlattr_inet_diag_msg.c: Include "test_nlattr.h".
(test_inet_diag_meminfo, test_inet_diag_vegasinfo,
test_inet_diag_dctcpinfo, test_inet_diag_bbrinfo): Remove.
(address): New variable.
(init_inet_diag_msg): Remove "address" argument, add const qualifier
to all remaining arguments.
(print_inet_diag_msg, print_uint): New functions.
(main): Use macros from test_nlattr.h file.
* netlink_sock_diag.c (PRINT_FIELD_X): New macro.
(decode_inet_diag_meminfo, decode_tcpvegas_info, decode_tcp_dctcp_info,
decode_tcp_bbr_info): Use it and PRINT_FIELD_U to print structure
fields.
Add const qualifiers to auto variables and function arguments.
This change does not affect the code generated by the compiler,
the purpose of these "const" qualifiers is to highlight the intent.
* netlink_sock_diag.c (print_meminfo, decode_meminfo,
decode_inet_diag_meminfo, decode_tcpvegas_info, decode_tcp_dctcp_info,
decode_tcp_bbr_info): Add const qualifier to tcp, addr, and len
arguments.
* linux/inet_diag.h (inet_diag_meminfo, tcpvegas_info,
tcp_dctcp_info, tcp_bbr_info): New structures.
* linux/sock_diag.h (SK_MEMINFO_VARS): New macro.
* netlink_sock_diag.c: Include <linux/sock_diag.h>.
(print_meminfo, decode_meminfo, decode_inet_diag_meminfo,
decode_tcpvegas_info, decode_tcp_dctcp_info, decode_tcp_bbr_info):
New functions.
(inet_diag_msg_nla_decoders): New array.
(decode_inet_diag_msg): Use it.
This change does not affect the code generated by the compiler,
the purpose of these "const" qualifiers is to highlight the intent.
* nlattr.c (decode_nlattr_with_data, decode_nla_str, decode_nla_strn,
decode_nla_##name): Add const qualifier to tcp, addr, and len arguments.
* linux/packet_diag.h (PACKET_DIAG_*): New enum.
* netlink_sock_diag.c: Include "xlat/packet_diag_attrs.h".
(decode_packet_diag_msg): Use decode_nlattr.
* xlat/packet_diag_attrs.in: New file.
Co-authored-by: Fabien Siron <fabien.siron@epita.fr>
* linux/netlink_diag.h (NETLINK_DIAG_*): New enum.
* netlink_sock_diag.c: Include "xlat/netlink_diag_attrs.h".
(decode_netlink_diag_msg): Use decode_nlattr.
* xlat/netlink_diag_attrs.in: New file.
Co-authored-by: Fabien Siron <fabien.siron@epita.fr>
When the list of headers do not fit into the command line, xargs splits
it into several lists and then passes them to grep. If no headers from
a list match the grep pattern, grep exits with a non-zero status that
causes xargs to exit with a non-zero status, too.
* maint/ioctls_sym.sh: Use "find -exec +" instead of xargs,
ignore its exit status.
While many NLMSG_DONE messages indeed have payload containing
just one integer, there are exceptions. Handle this by passing
payloads of NLMSG_DONE messages to family specific netlink
payload decoders.
* netlink.c (print_nlmsghdr): Do not skip family detection
for nlmsg_type == NLMSG_DONE.
(decode_nlmsg_type): Skip family specific type decoders
for type == NLMSG_DONE.
(decode_nlmsg_flags): Skip family specific decoding of flags
for type == NLMSG_DONE.
(decode_netlink_sock_diag): Skip for nlmsg_type == NLMSG_DONE.
* util.c (printstr_ex): Move the check that -s argument
does not exceed -1U / 4 ...
* strace.c (init): ... here.
* tests/options-syntax.test: Check it.
Add a fallback definition of NETLINK_SOCK_DIAG to netlink.h
instead of adding these fallback definitions to every file.
* netlink.h: Provide a fallback definition of NETLINK_SOCK_DIAG.
* socketutils.c: Include "netlink.h" instead of <linux/netlink.h>,
remove fallback definition of NETLINK_SOCK_DIAG.
* xlat/v4l2_buf_types.in: Add V4L2_BUF_TYPE_META_CAPTURE introduced
by linux kernel commit v4.12-rc1~85^2~71.
* xlat/v4l2_device_capabilities_flags.in: Add V4L2_CAP_META_CAPTURE
introduced by the same commit.
* xlat/v4l2_control_flags.in: Add V4L2_CTRL_FLAG_HAS_PAYLOAD,
V4L2_CTRL_FLAG_EXECUTE_ON_WRITE, and V4L2_CTRL_FLAG_MODIFY_LAYOUT
introduced by linux kernel commits v3.17-rc1~112^2~363,
v4.1-rc1~59^2~1^2~63, and v4.12-rc1~85^2~62, respectively.
* xlat/v4l2_control_types.in: Add V4L2_CTRL_TYPE_U8, V4L2_CTRL_TYPE_U16,
and V4L2_CTRL_TYPE_U32 introduced by linux kernel commits
v3.17-rc1~112^2~343 and v3.17-rc1~112^2~112.
* xlat/sockoptions.in: Add SO_MEMINFO, SO_INCOMING_NAPI_ID,
and SO_COOKIE introduced by linux kernel commits v4.12-rc1~64^3~389,
v4.12-rc1~64^3~341^2, and v4.12-rc1~64^3~221^2~1, respectively.
Despite our efforts made to avoid unaligned access, gcc appears to be
generating instructions that might cause unaligned access when copying
structures to unaligned addresses.
Fix this issue by replacing copying of structures to unaligned addresses
with memcpy calls.
* tests/netlink_sock_diag.c (INIT_STRUCT): New macro.
(test_odd_family_req, test_odd_family_msg, test_unix_diag_req,
test_unix_diag_msg, test_netlink_diag_req, test_netlink_diag_msg,
test_packet_diag_req, test_packet_diag_msg, test_inet_diag_sockid,
test_inet_diag_req, test_inet_diag_req_v2, test_inet_diag_msg,
test_smc_diag_req, test_smc_diag_msg): Use it to initialize potentially
unaligned struct nlmsghdr. Do not access nlh->nlmsg_len.
NETLINK_GENERIC is a protocol stacked on netlink protocol.
When a NETLINK_GENERIC subsystem registers its "family" name in the
kernel at run time, an integer value associated with this name is
assigned by the kernel. This integer value is specified in struct
nlmsghdr.nlmsg_type of subsequent communications with the
NETLINK_GENERIC subsystem.
This change implements decoding of NETLINK_GENERIC message types in
symbolic form, for example:
$ ./strace -yy -e 'sendto' ip tcp_metrics > /dev/null
sendto(4<NETLINK:[GENERIC:12345]>, {{len=20, type=tcp_metrics, ...
+++ exited with 0 +++
Note that type= specifying the NETLINK_GENERIC protocol family
is decoded properly.
* netlink.c (decode_nlmsg_type_generic): New function.
(nlmsg_types): Add NETLINK_GENERIC entry.
Signed-off-by: Masatake YAMATO <yamato@redhat.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
* configure.ac (AC_CHECK_HEADERS): Add linux/genetlink.h.
* defs.h (genl_families_xlat): New prototype.
* socketutils.c [HAVE_LINUX_GENETLINK_H]: Include <linux/genetlink.h>.
[HAVE_LINUX_GENETLINK_H] (genl_send_dump_families,
genl_parse_families_response): New helper functions.
(genl_families_xlat): New function for building the xlat table
for NETLINK_GENERIC decoding.
Signed-off-by: Masatake YAMATO <yamato@redhat.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
xlat tables are usually generated at build time.
printxval is suitable for printing struct nlmsghdr.nlmsg_type for
NETLINK_GENERIC protocol. However, the corresponding xlat table
cannot be defined at build time because the values are registered
by the kernel at run time.
This change introduces dyxlat functions for building xlat tables
dynamically. The primary use case is decoding nlmsg_type but this
interface can be used for other purposes.
* defs.h (struct dyxlat): New opaque data type.
(dyxlat_alloc, dyxlat_free, dyxlat_add_pair): New function declarations.
* dyxlat.c: New file.
* Makefile.am (strace_SOURCES): Add it.
Signed-off-by: Masatake YAMATO <yamato@redhat.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
Enclose macros with multiple statements and macros starting
with "if" statement in a do/while loop.
Do not enclose single statement macros in a do/while loop.
Reported by kernel's checkpatch.pl script.
Change functions declared as taking unspecified number of arguments
of unspecified type to functions that take no arguments.
Reported by kernel's checkpatch.pl script.
Add support for printing control messages containing software and/or
hardware timestamps, which are typically enabled by NTP and PTP
implementations.
* print_timeval.c (print_struct_timeval_data_size): New mpers printer.
* print_timespec.c (print_struct_timespec_data_size,
print_struct_timespec_array_data_size): Likewise.
* msghdr.c (print_scm_timestamp, print_scm_timestampns,
print_scm_timestamping): New functions.
(cmsg_socket_printers): Add them.
* NEWS: Mention this change.
* tests/msg_control.c: Include "xlat.h" and "xlat/scmvals.h".
(test_scm_timestamp, test_scm_timestampns, test_scm_timestamping):
New functions.
(test_sol_socket): Use them.
Signed-off-by: Miroslav Lichvar <mlichvar@redhat.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
* tests/netlink_sock_diag.c: Include <linux/smc_diag.h>.
(SMC_ACTIVE): New macro.
(test_smc_diag_req, test_smc_diag_msg): New functions.
(main): Use them.
* linux/smc_diag.h: New file.
* Makefile.am (EXTRA_DIST): Add it.
* netlink_sock_diag.c: Include <linux/smc_diag.h>,
"xlat/smc_diag_extended_flags.h" and "xlat/smc_states.h".
(decode_smc_diag_req, decode_smc_diag_msg): New functions.
(diag_decoders): Add AF_SMC.
* xlat/smc_diag_extended_flags.in: New file.
* xlat/smc_states.in: Likewise.
* tests/netlink_sock_diag.c: Include <arpa/inet.h>
and <linux/inet_diag.h>.
(test_inet_diag_sockid, test_inet_diag_req, test_inet_diag_req_v2,
test_inet_diag_msg): New functions.
(mian): Use them.
Co-authored-by: Fabien Siron <fabien.siron@epita.fr>
* tests/netlink_sock_diag.c: Include <linux/if_ether.h>
and <linux/packet_diag.h>.
(test_packet_diag_req, test_packet_diag_msg): New functions.
(main): Use them.
Co-authored-by: Fabien Siron <fabien.siron@epita.fr>
* tests/netlink_sock_diag.c: Include <linux/netlink_diag.h>.
(test_netlink_diag_req, test_netlink_diag_msg): New functions.
(main): Use them.
Co-authored-by: Fabien Siron <fabien.siron@epita.fr>
* defs.h (netlink_decoder_t): New typedef.
(DECL_NETLINK): New macro.
(decode_netlink_sock_diag): New prototype.
* netlink.c (netlink_decoders): New array.
(decode_payload): Use it to invoke netlink family specific decoders.
* netlink_sock_diag.c: New file.
* Makefile.am (strace_SOURCES): Add it.
Co-authored-by: Fabien Siron <fabien.siron@epita.fr>
Co-authored-by: JingPiao Chen <chenjingpiao@gmail.com>
* netlink.c (decode_nlmsg_type_default, decode_nlmsg_type_netfilter):
New functions.
(nlmsg_types_decoder_t): New typedef.
(nlmsg_types): Add decoder field.
(decode_nlmsg_type): Use it.
Fix checks for struct v4l2_window.global_alpha and struct
v4l2_sdr_format.buffersize.
* v4l2.c (print_v4l2_format_fmt): Do not assume that
HAVE_STRUCT_V4L2_WINDOW_GLOBAL_ALPHA and
HAVE_STRUCT_V4L2_SDR_FORMAT_BUFFERSIZE are defined.
* tests/ioctl_v4l2.c (init_v4l2_format, dprint_ioctl_v4l2): Likewise.
This is the second patch in series of implementing NETLINK_GENERIC
protocol decoder.
This change allows passing of opaque user data to the callback function.
* socketutils.c (receive_responses): Change the type of proto_name
argument from "const char *" to "void *" to allow passing
of arbitrary data to the callback function.
Rename proto_name to opaque_data, make it the last argument.
Pass it to the callback function as the last argument.
(inet_parse_response, unix_parse_response, netlink_parse_response):
Change the type of proto_name argument from "const char *" to "void *",
rename it to opaque_data, make it the last argument.
Introduce a local variable proto_name to convert the type
of opaque_data argument from "void *" to "const char *".
(unix_get, inet_get, netlink_get): Cast the protocol name from
"const char *" to "void *" for passing to the callback function.
Signed-off-by: Masatake YAMATO <yamato@redhat.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
This is the first patch in series of implementing NETLINK_GENERIC
protocol decoder.
receive_responses was written for decoding kernel responses of type
SOCK_DIAG_BY_FAMILY, other types were ignored.
This change makes the type of netlink response a parameter
so the function can be used for other types of communication.
* socketutils.c (receive_responses): add a new parameter
expected_msg_type to handle types other than SOCK_DIAG_BY_FAMILY.
Signed-off-by: Masatake YAMATO <yamato@redhat.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
Add printflags_ex extension over printflags64 that, similar
to printxvals, takes a NULL-terminated sequence of xlat pointers.
* defs.h (printflags_ex): New prototype.
(printflags64): Turn into a thin wrapper around printflags_ex.
* netlink.c (decode_nlmsg_flags): Use printflags_ex.
* util.c (printflags_ex): New function.
(printflags64): Remove.
* xlat/netlink_get_flags.in: Remove entries listed
in xlat/netlink_flags.in file.
* xlat/netlink_new_flags.in: Likewise.
* tests/netlink_route.c (test_nlmsg_flags): New function, checks
decoding of NETLINK_ROUTE get and new flags.
(main): Use it.
* tests/netlink_sock_diag.c (test_nlmsg_flags): New function, checks
decoding of NETLINK_SOCK_DIAG get flags.
(main): Use it.
* tests/netlink_xfrm.c: (test_nlmsg_flags): New function, checks
decoding of NETLINK_XFRM get and new flags.
(main): Use it.
* netlink.c: Include "xlat/netlink_get_flags.h"
and "xlat/netlink_new_flags.h".
(decode_nlmsg_flags): New function.
(print_nlmsghdr): Use it.
* xlat/netlink_get_flags.in: New file.
* xlat/netlink_new_flags.in: Likewise.
Co-authored-by: Fabien Siron <fabien.siron@epita.fr>
Co-authored-by: Dmitry V. Levin <ldv@altlinux.org>
Some of XFRM_MSG_* constants were introduced later than the header file
where they are defined.
* xlat/nl_xfrm_types.in: Add default values for constants.
This change removes the trace_syscall function. Now, the code that uses
syscall.c trace functions is expected to check whether it is a syscall
entry or exit (with entering(tcp)/exiting(tcp)) itself, and then make
an appropriate sequence of function calls.
* defs.h: Update comment on TCB_INSYSCALL.
(trace_syscall): Remove prototype.
(syscall_entering_decode, syscall_entering_trace,
syscall_entering_finish, syscall_exiting_decode, syscall_exiting_trace,
syscall_exiting_finish): New prototypes.
* strace.c (trace_syscall): New static replacement for old trace_syscall.
* syscall.c (trace_syscall): Remove.
(trace_syscall_entering): Split into ...
(syscall_entering_decode, syscall_entering_trace,
syscall_entering_finish): ... new functions.
(trace_syscall_exiting): Split into ...
(syscall_exiting_decode, syscall_exiting_trace,
syscall_exiting_finish): ... new functions.
nfnetlink was introduced in linux by commit v2.6.14-rc1~1035^2~217,
add a guard for the case when system headers are not fresh enough.
* configure.ac (AC_CHECK_HEADERS): Add linux/netfilter/nfnetlink.h.
* tests/netlink_netfilter.c: Check for HAVE_LINUX_NETFILTER_NFNETLINK_H.
* tests/tests.h (create_nl_socket): Rename to create_nl_socket_ext,
add a string argument.
(create_nl_socket): New macro wrapper around create_nl_socket_ext.
* tests/create_nl_socket.c (create_nl_socket): Rename
to create_nl_socket_ext, add a string argument, include it
in diagnostic messages.
Implement a pluggable mechanism of nlmsg_type decoding.
Add decoding of NETLINK_SOCK_DIAG types.
* linux/sock_diag.h (SOCK_DESTROY): New macro.
* xlat/nl_sock_diag_types.in: New file.
* netlink.c: Include xlat/nl_sock_diag_types.h.
(get_fd_nl_family, decode_nlmsg_type): New functions.
(print_nlmsghdr): Use them for nlmsg_type decoding.
Co-authored-by: Fabien Siron <fabien.siron@epita.fr>
* netlink.c (NL_FAMILY_ERROR, NL_FAMILY_DEFAULT): New constants.
(print_nlmsghdr): Add netlink family argument, change return type
to int, return family.
(decode_nlmsghdr_with_payload, decode_nlmsgerr, decode_payload): Add
netlink family argument. All callers updated.
Add a way to get socket information without printing it.
* defs.h (get_sockaddr_by_inode): New prototype.
* socketutils.c (get_sockaddr_by_inode): New function.
Co-authored-by: Dmitry V. Levin <ldv@altlinux.org>
cache_inode_details gives a way to obtain cache information without
printing it.
* socketutils.c (cache_and_print_inode_details): Rename
to cache_inode_details. Do not print cached information.
(inet_parse_response, unix_parse_response, netlink_parse_response):
Use cache_inode_details instead of cache_and_print_inode_details.
(inet_print, unix_print, netlink_print): Call
print_sockaddr_by_inode_cached.
Co-authored-by: Dmitry V. Levin <ldv@altlinux.org>
Do not assume that an executable is small enough and SO_SNDBUF is large
enough so that the executable could be sendfile'ed into a socket pair.
* tests/sendfile.c (main): Create a regular file of the right size
to avoid hitting SO_SNDBUF limit.
* tests/sendfile64.c: Likewise.
Reported-by: Andreas Schwab <schwab@linux-m68k.org>
The side effect of #include "xlat/netlink_protocols.h" is
NETLINK_* constants properly defined in that header file.
While net.c does not use these constants itself, netlink.c
is going to need them soon.
* net.c: Move inclusion of "xlat/netlink_protocols.h" ...
* netlink.c: ... here.
* netlink.c (decode_payload): Decode NLMSG_DONE messages.
* tests/netlink_protocol.c (test_nlmsg_done): New function
for checking decoding of NLMSG_DONE messages.
(main): Use it.
Change print_sockaddr_by_inode to be the only function exposed,
hide print_sockaddr_by_inode_cached as a part of implementation.
* defs.h (print_sockaddr_by_inode_cached): Remove prototype.
(print_sockaddr_by_inode): Add struct tcb * and descriptor arguments.
* socketutils.c (print_sockaddr_by_inode_cached): Make static.
(print_sockaddr_by_inode): Rename to print_sockaddr_by_inode_uncached,
make static.
(print_sockaddr_by_inode): New function.
* util.c (printfd): Update to use new print_sockaddr_by_inode.
There is nothing unusual about PTRACE_EVENT_EXEC that would require
a special treatment.
* strace.c (next_event): Move PTRACE_EVENT_EXEC handling to the main
switch statement.
As strace does not use WCONTINUED flag, there is no need to handle
WIFCONTINUED status.
* strace.c (print_debug_info) [WIFCONTINUED]: Remove.
(next_event): Remove redundant !WIFSTOPPED check.
This change decouples waiting for next event from reacting to it. This
makes the control flow easier to understand, and serves as a preparation
for implementing a pull-style API for LuaJIT.
* strace.c (enum trace_event): New enum.
(trace): Split into ...
(next_event, dispatch_event): ... new functions.
(main): Use them.
This fixes generation of syscallent.i and subsequent scno.h files
on mips where syscallent.h checks macros defined in config.h.
* scno.am (syscallent.i): Add $(top_builddir)/config.h
to prerequisites, include it.
* NEWS: Mention this fix.
Reported-by: David Robins <dbrobins@i4031.net>
On sparc, unlike all other architectures, old sigaction syscall takes
a negated signal number.
* signal.c (SYS_FUNC(sigaction)) [SPARC || SPARC64]: Enhance printing
of negative signal number argument.
On alpha, struct old_sigaction.sa_flags has a 32-bit type.
On mips, only first word of old_sigaction.sa_mask is read.
On all architectures except alpha and mips that have old sigaction
syscall, struct old_sigaction has sa_restorer field that has to be
decoded.
* signal.c (struct old_sigaction) [ALPHA]: Change sa_flags type
to unsigned int, add ATTRIBUTE_PACKED.
[MIPS]: Dhrink sa_mask array to 1 element.
[!ALPHA && !MIPS]: Define sa_restorer field unconditionally.
(struct old_sigaction32): Likewise.
(decode_old_sigaction) [!current_wordsize]: Initialize sa_restorer field
from old_sigaction32.sa_restorer unconditionally.
[!ALPHA && !MIPS]: Print old_sigaction.sa_restorer if SA_RESTORER flag
is set.
* NEWS: Mention this change.
SA_ONSTACK is the standard name while its alias SA_STACK is not even
defined on many platforms.
* xlat/sigact_flags.in (SA_STACK): Rename to SA_ONSTACK.
Parsers of sigprocmask and osf_sigprocmask syscalls are different
anyway, so rename the function to match the syscall name.
* signal.c [ALPHA] (SYS_FUNC(sigprocmask)): Rename
to SYS_FUNC(osf_sigprocmask).
* linux/alpha/syscallent.h: Update sys_func.
The data collected by -fprofile-arcs during program run is saved
on program exit. When program is terminated by signal, the data
is not saved. However, libgcov provides an interface to dump
the data at runtime.
* m4/ax_code_coverage.m4 (AX_CODE_COVERAGE): Add -DENABLE_COVERAGE_GCOV
to CODE_COVERAGE_CPPFLAGS.
* strace.c [ENABLE_COVERAGE_GCOV] (__gcov_flush): New prototype.
(main) [ENABLE_COVERAGE_GCOV]: Call __gcov_flush() before raise()
and sigprocmask() calls that might cause program termination.
When traced command is terminated by a blocked signal, unblock
that signal to ensure strace termination with the same signal.
* strace.c (main): Unblock the signal raised for strace termination.
* NEWS: Mention this change.
While strace resets SIGCHLD handler to the default action so that
waitpid definitely works without losing track of children, tracees
should not inherit this change.
* strace.c (struct exec_params): Add child_sa field.
(init): When setting SIGCHLD handler to SIG_DFL, save the old handler.
(exec_or_die): Restore SIGCHLD handler if it was different from SIG_DFL
at startup.
* NEWS: Mention this change.
In interactive mode (-I2), those signals that were blocked at startup
will remain blocked for the whole period of strace execution.
* strace.c (empty_set): Rename to start_set. All users changed.
(init): Initialize start_set and blocked_set to the set of currently
blocked signals.
(startup_attach, trace): Use SIG_SETMASK instead of SIG_BLOCK.
* NEWS: Mention this change.
Old ssetmask syscall takes an argument and returns a value of type int.
Old sgetmask syscall may return an error.
* signal.c (SYS_FUNC(ssetmask)): Explicitly convert the argument
and return value to unsigned int, print it using sprint_old_sigmask_val.
(SYS_FUNC(sgetmask)): Do not print the mask in case of syserror.
* tests/sxetmask.c: New file.
* tests/gen_tests.in (sxetmask): New entry.
* tests/pure_executables.list: Add sxetmask.
* tests/.gitignore: Likewise.
* NEWS: Mention this fix.
Also do not include them into release tarballs. These files are
intended for developers who are expected to use a git repository.
* errnoent.sh: Move to maint subdirectory.
* signalent.sh: Likewise.
* syscallent.sh: Likewise.
* HACKING-scripts: Update.
* Makefile.am (EXTRA_DIST): Remove errnoent.sh, signalent.sh,
and syscallent.sh.
These files are intended for developers who are expected to use
a git repository.
* Makefile.am (EXTRA_DIST): Remove maint/ioctls_gen.sh,
maint/ioctls_hex.sh, maint/ioctls_sym.sh, and maint/print_ioctlent.c.
On some architectures old sigsuspend syscall takes the signal mask from
the 3rd argument, on some from the 1st. And, if it wasn't peculiar
enough, the signal mask is passed by value on all architectures except
mips where it's passed by reference.
* signal.c (SYS_FUNC(sigsuspend)): Take the signal mask from the last
argument.
* linux/alpha/syscallent.h (sigsuspend): Set nargs to 1.
* linux/crisv10/syscallent.h: Likewise.
* linux/mips/syscallent-o32.h: Likewise.
* linux/powerpc/syscallent.h: Likewise.
* linux/powerpc64/syscallent.h: Likewise.
* linux/sh/syscallent.h: Likewise.
* linux/sh64/syscallent.h: Likewise.
* linux/sparc/syscallent.h: Likewise.
* linux/sparc64/syscallent.h: Likewise.
* tests/sigsuspend.c: New file.
* tests/gen_tests.in (sigsuspend): New entry.
* tests/pure_executables.list: Add sigsuspend.
* tests/.gitignore: Likewise.
* NEWS: Mention this fix.
Recently added tests based on pure_executables.list are quite slow
by their nature, the one minute default timeout is no longer enough
for some of not so fast architectures.
* tests/init.sh (TIMEOUT_DURATION): Raise from 60 to 120.
Check for some architecture-specific syscalls which are not present
on most architectures.
* gen_tests.in: Add trace_question test.
* tests/trace_question.in: New file.
* tests/Makefile.am (EXTRA_DIST): Add it.
* v4l2.c: Include "xlat/v4l2_tuner_types.h", "xlat/v4l2_tuner_capabilities.h",
"xlat/v4l2_tuner_rxsubchanses.h", and "xlat/v4l2_tuner_audmodes.h".
(print_v4l2_tuner): New function. Add decoding for VIDIOC_G/S_TUNER's arg.
(v4l2_ioctl): Add new cases for decoding VIDIOC_G/S_TUNER's arg.
* xlat/v4l2_tuner_types.in: Add V4L2_TUNER_RADIO and
V4L2_TUNER_ANALOG_TV introduced by linux kernel commit
2.5.46~39^2~23^2~4.
Add V4L2_TUNER_DIGITAL_TV introduced by linux kernel
commit v2.6.11-rc4~55^2~20.
Add V4L2_TUNER_ADC and V4L2_TUNER_RF introduced by linux kernel commit
v3.15-rc1~85^2~215.
Add V4L2_TUNER_SDR introduced by linux kernel commit v4.4-rc1~118^2~17
(as the replacement of V4L2_TUNER_ADC).
* xlat/v4l2_tuner_capabilities.in: Add V4L2_TUNER_CAP_LOW,
V4L2_TUNER_CAP_NORM, V4L2_TUNER_CAP_STEREO, V4L2_TUNER_CAP_LANG2,
V4L2_TUNER_CAP_SAP, and V4L2_TUNER_CAP_LANG1 introduced by linux kernel
commit 2.5.46~39^2~23^2~4.
Add V4L2_TUNER_CAP_HWSEEK_BOUNDED and V4L2_TUNER_CAP_HWSEEK_WRAP
introduced by linux kernel commit v3.6-rc1~28^2~306.
Add V4L2_TUNER_CAP_RDS introduced by linux commit v2.6.32-rc1~679^2~189.
Add V4L2_TUNER_CAP_RDS_BLOCK_IO and V4L2_TUNER_CAP_RDS_CONTROLS
introduced by linux kernel commit v2.6.37-rc1~64^2~110.
Add V4L2_TUNER_CAP_FREQ_BANDS introduced by linux kernel commit
v3.6-rc1~28^2~46.
Add V4L2_TUNER_CAP_HWSEEK_PROG_LIM introduced by linux kernel commit
v3.6-rc1~28^2~40.
Add V4L2_TUNER_CAP_1HZ introduced by linux kernel commit
v3.15-rc1~85^2~214.
* xlat/v4l2_tuner_audmodes.in: Add V4L2_TUNER_MODE_MONO,
V4L2_TUNER_MODE_STEREO, V4L2_TUNER_MODE_LANG2, V4L2_TUNER_MODE_SAP, and
V4L2_TUNER_MODE_LANG1 introduced by linux kernel commit
2.5.46~39^2~23^2~4.
Add V4L2_TUNER_MODE_LANG1_LANG2 introduced by linux kernel commit
v2.6.17-rc1~739^2~9.
* xlat/v4l2_tuner_rxsubchanses.in: Add V4L2_TUNER_SUB_MONO,
V4L2_TUNER_SUB_STEREO, V4L2_TUNER_SUB_LANG2, V4L2_TUNER_SUB_SAP, and
V4L2_TUNER_SUB_LANG1 introduced by linux kernel commit
2.5.46~39^2~23^2~4.
Add V4L2_TUNER_SUB_RDS introduced by linux kernel commit
v2.6.32-rc1~679^2~189.
Signed-off-by: Edgar Kaziahmedov <edos@linux.com>
* print_time.c (SYS_FUNC(time)): Print timestamp stringification
(via sprinttime) for the value stored in the argument and return value
(return RVAL_STR in the latter case).
* tests/time.c (main): Update expected output.
* print_timespec.c (print_timespec_t_utime): Use sprinttime_nsec
to format a string representation of time, use tprints_comment
to print it as a comment.
* print_timeval.c (print_timeval_t_utime): New function.
(print_timeval_utimes): Use it instead of print_timeval_t.
[ALPHA] (print_timeval32_t_utime): New function.
[ALPHA] (print_timeval32_utimes): Use it instead of print_timeval32_t.
* tests/futimesat.c (print_tv): Update expected output.
* tests/utimensat.c (print_ts): Likewise.
* tests/xutimes.c (print_tv): Likewise.
Co-authored-by: Dmitry V. Levin <ldv@altlinux.org>
It will be needed later.
* tests/tests.h (print_time_t_usec): New prototype.
* tests/print_time.c (print_time_t_usec): New function, a thin wrapper
around print_time_t_ex.
Negative micro/nanoseconds values are treated as invalid by kernel
anyway, and in one case (timespec_valid in include/linux/time.h)
it is even checked by conversion to unsigned long.
* print_timespec.c (timespec_fmt): Change tv_sec format to %lld and
tv_nsec format to %llu.
(print_timespec_t): Cast tv_sec to long long and process tv_nsec with
zero_extend_signed_to_ull.
(sprint_timespec): Likewise.
* print_timeval.c (timeval_fmt): Change tv_sec format to %lld and
tv_usec format to %llu.
(print_timeval_t): Cast tv_sec to long long and process tv_nsec with
zero_extend_signed_to_ull.
(sprint_timeval, print_timeval32_t, sprint_timeval32): Likewise.
* defs.h (TIMESPEC_TEXT_BUFSIZE): Update.
* tests/adjtimex.c (main): Change tv_sec printing format to %lld, cast
it to long long; change tv_usec printing format to %llu, process it with
zero_extend_signed_to_ull.
* tests/clock_nanosleep.c (main): Change tv_sec printing format to %lld,
cast it to long long; change tv_nsec printing format to %llu, process it
with zero_extend_signed_to_ull.
* tests/clock_xettime.c (main): Likewise.
* tests/futex.c (main): Likewise.
* tests/futimesat.c (print_tv): Likewise.
* tests/getrusage.c (invoke_print): Likewise.
* tests/mq_sendrecv.c (do_send, do_recv, main): Likewise.
* tests/nanosleep.c (main): Likewise.
* tests/pselect6.c (main): Likewise.
* tests/restart_syscall.c (main): Likewise.
* tests/rt_sigtimedwait.c (iterate, main): Likewise.
* tests/sched_rr_get_interval.c (main): Likewise.
* tests/semop.c (main): Likewise.
* tests/timer_xettime.c (main): Likewise.
* tests/timerfd_xettime.c (main): Likewise.
* tests/waitid.c (main): Likewise.
* tests/xetitimer.c (main): Likewise.
* tests/xettimeofday.c (main): Likewise.
* tests/xselect.c (main): Likewise.
* tests/xutimes.c (print_tv): Likewise.
* tests/wait4.c (sprint_rusage): Likewise.
* tests/waitid.c (sprint_rusage): Likewise.
* tests/utimensat.c (print_ts): Likewise.
(main): Add check for higher bits of tv_sec/tv_nsec.
Co-authored-by: Dmitry V. Levin <ldv@altlinux.org>
Refactor sprinttime: implement sprinttime_nsec and sprinttime_usec
that handle nanoseconds and microseconds, respectively.
Always print raw values of time data fields, format string
representations of time as comments.
* defs.h (sprinttime): Change argument type from time_t to long long.
(sprinttime_nsec, sprinttime_usec): New prototypes.
* util.c (sprinttime_ex, sprinttime_nsec, sprinttime_usec): New
functions.
(sprinttime): Turn into a thin wrapper around sprinttime_ex.
* stat.h (struct strace_stat): Add has_nsec field.
* fetch_struct_stat.c (HAVE_NSEC): New macro.
(fetch_struct_stat): Initialize has_nsec field with HAVE_NSEC.
* fetch_struct_stat64.c (HAVE_NSEC): New macro.
(fetch_struct_stat64): Initialize has_nsec field with HAVE_NSEC.
* print_struct_stat.c (print_struct_stat) <PRINT_ST_TIME>:
Print raw values of time fields, use sprinttime_nsec to format a string
representation of time, use tprints_comment to print it as a comment.
* statx.c (SYS_FUNC(statx)) <PRINT_FIELD_TIME>: Likewise.
* utime.c (SYS_FUNC(utime)): Print raw values of struct utimbuf.actime
and struct utimbuf.modtime fields, use sprinttime to format a string
representation of time, use tprints_comment to print it as a comment.
* tests/tests.h (print_time_t_nsec): Add int argument.
* tests/print_time.c (print_time_t_ex): New function.
(print_time_t_nsec): Add int argument, turn into a thin wrapper around
print_time_t_ex.
* tests/utime.c (main): Update expected output.
* tests/xstatx.c [!IS_STATX] (HAVE_NSEC): New macro.
[!IS_STATX] (PRINT_ST_TIME), [IS_STATX] (PRINT_FIELD_TIME): Update
expected output.
* NEWS: Mention this timestamps representation improvement.
Co-authored-by: Dmitry V. Levin <ldv@altlinux.org>
stat.sample is created by appropriate executables and never removed.
* tests/gen_tests.in (trace_fstat, trace_lstat, trace_stat,
trace_stat_like): Remove stat.sample creation.
It may be useful in case syscall is not present on the target
architecture.
* qualify.c (qualify_syscall): Add ignore_fail flag, set it to true
in case syscall qualification is led by a question mark(s).
* strace.1 (.SS Filtering): Document ability to use question marks
in syscall qualification.
*xstat syscalls are, according to [1], like their *stat counterparts,
but with additional argument with fixed value.
[1] http://math-atlas.sourceforge.net/devel/assembly/mipsabi32.pdf
* linux/mips/syscallent-compat.h (*_xstat, *_lxstat, bsd43_oldstat):
Add TF flag.
(*_fxstat, bsd43_oldfstat): Add TD flag.
Rewrite sched.test using test_trace_expr().
* tests/gen_tests.in (sched): New entry.
* tests/sched.in: New file, derived from a set of tests in sched.test.
* tests/sched.test: Remove.
* tests/Makefile.am (DECODER_TESTS): Remove sched.test.
(EXTRA_DIST): Add sched.in.
Rewrite regex.test using test_trace_expr().
* tests/gen_tests.in (regex): New entry.
* tests/regex.in: New file, derived from a set of tests in regex.test.
* tests/regex.test: Remove.
* tests/Makefile.am (DECODER_TESTS): Remove regex.test.
(EXTRA_DIST): Add regex.in.
Rewrite clock.test using test_trace_expr().
* tests/gen_tests.in (clock): New entry.
* tests/clock.in: New file, derived from a set of tests in clock.test.
* tests/clock.test: Remove.
* tests/Makefile.am (DECODER_TESTS): Remove clock.test.
(EXTRA_DIST): Add clock.in.
Rewrite trace_statfs.test using test_trace_expr().
* tests/gen_tests.in (trace_statfs): New entry.
* tests/trace_statfs.in: New file, derived from a set of tests
in trace_statfs.test.
* tests/trace_statfs.test: Remove.
* tests/Makefile.am (DECODER_TESTS): Remove trace_statfs.test.
(EXTRA_DIST): Add trace_statfs.in.
This introduces three new helpers:
prog_set_subtract subtracts one program set from another,
test_pure_prog_set runs strace check on a set of programs,
test_trace_expr tests trace expressions.
* tests/init.sh (prog_set_subtract, test_pure_prog_set,
test_trace_expr): New functions.
As tests are now run in separate directories.
* tests/xstatx.c: Use fixed "stat.sample" name for sample file.
* tests/fstat.test: Use it as an argument for -P option.
* tests/gen_tests.in (fstat64, lstat, lstat64, newfstatat, oldlstat,
oldstat, stat, stat64, statx): Likewise.
This solves problems like the need to update test suite on every
New Year's Eve.
* Makefile.am (dist-hook): Generate .year.
* copyright-year-gen: New file.
* configure.ac (copyright_year): New m4 variable, defined as the output
of copyright-year-gen script.
(AC_COPYRIGHT): Use it.
(COPYRIGHT_YEAR): New output variable and preprocessor macro.
* strace.c (print_version): Use COPYRIGHT_YEAR.
* strace.spec.in (%prep): Save the value of COPYRIGHT_YEAR autoconf
variable to .year file.
* tests/strace-V.test (config_year): New variable, derived from config.h.
Add sanity checks for $config_year and use it in expected output.
Reported-by: Andreas Schwab <schwab@suse.de>
* tests/gen_tests.in (pread64-pwrite64): New entry.
* tests/pread64-pwrite64.c: (main): Skip the test if /proc/self/fd/
is not available.
* tests/pread64-pwrite64.test: Remove.
* tests/Makefile.am (DECODER_TESTS): Remove pread64-pwrite64.test.
* tests/gen_tests.in (read-write): New entry.
* tests/read-write.c: (main): Skip the test if /proc/self/fd/
is not available.
* tests/read-write.test: Remove.
* tests/Makefile.am (DECODER_TESTS): Remove read-write.test.
* numa.c (print_status): Do not lose a minus sign when printing
a symbolic error code.
* tests/move_pages.c (print_status_array): Update expected output.
* tests/xutimes.c: New file, based on utimes.c.
* tests/utimes.c [__NR_utimes] (TEST_SYSCALL_NR, TEST_SYSCALL_STR,
TEST_STRUCT): New macros.
[__NR_utimes]: Replace all the rest with include "xutimes.c".
* tests/Makefile.am (EXTRA_DIST): Add xutimes.c.
This follows the rename of print_timeval_item to print_timeval_utimes.
* defs.h [ALPHA] (print_timeval32_pair): Rename
to print_timeval32_utimes.
* print_timeval.c [ALPHA] (print_timeval32_pair): Likewise.
* utimes.c [ALPHA] (SYS_FUNC(osf_utimes)): Replace print_timeval32_pair
with print_timeval32_utimes.
* tests/utimes.c (errstr): New variable.
(print_ts, k_utimes): New functions.
(main): Use them to do more rigorous testing of utimes syscall parser.
* tests/gen_tests.in (utimes): Update -a option.
* tests/futimesat.c (errstr): New variable.
(print_ts, k_futimesat): New functions.
(main): Use them to do more rigorous testing of futimesat syscall parser.
When the whole pair of struct timeval structures cannot be fetched from
tracee's memory, print the address. This behavior follows the kernel's
that uses copy_from_user for the whole pair, too.
* print_timeval.c (print_timeval_item): Remove.
(print_timeval_pair): Rename to print_timeval_utimes, all callers
updated. Fetch the whole timeval_t array with a single
umove_or_printaddr call.
* tests/futimesat.c (main): Update expected output.
* tests/utimes.c (main): Likewise.
The new test also does more rigorous testing.
* tests/utimensat.c: Stop including <assert.h> and <errno.h>, include
<unistd.h> and <asm/unistd.h>, check __NR_utimensat instead
of HAVE_UTIMENSAT.
(errstr): New variable.
(k_utimensat): New function, a thin wrapper around
syscall(__NR_utimensat).
(main): Use them instead of utimensat.
* tests/gen_tests.in (utimensat): New entry.
* tests/utimensat.test: Remove.
* tests/Makefile.am (DECODER_TESTS): Remove it.
* tests/clock.test: Remove futimesat, utime, utimensat, and utimes tests
from the list of exceptions.
* tests/utime.c (main): Do not use time(), use fixed timestamp instead.
* tests/futimesat.c (main): Do not use gettimeofday(), use fixed
timestamps instead.
* tests/utimensat.c (main): Likewise.
* tests/utimes.c (main): Likewise.
* tests/mq_sendrecv.c (MQ_NAME): Remove.
(mq_name): New static variable.
(cleanup): Use it instead of MQ_NAME.
(main): Initialize it dynamically using the pid of process. Use it
instead of MQ_NAME.
* tests/mq.c (NAME): Remove.
(main): Replace the fixed message queue name with a dynamically
made name containing the pid of process.
* tests/gen_tests.in (mq): Update -a option.
* tests/regex.test: Use trace_statfs.test as a template instead
of clock.test and sched.test, not because of more interesting
regular expressions but to fix build on platforms where
clock.test and sched.test are not reenterable.
* qualify.c: Include <regex.h>.
(qualify_syscall_regex): New function.
(qualify_syscall): Use it.
* strace.1: Document -e trace=/regex option.
* NEWS: Mention -e trace=/regex option.
* tests/regex.test: New test.
* tests/Makefile.am (DECODER_TESTS): Add it.
* tests/options-syntax.test: Add checks for invaild regexp
and for regexp that doesn't match a syscall.
* tests/ioctl_nsfs.c (test_user_namespace): Specify an address
in the middle of a mapped page as child_stack address argument
of the clone call.
[IA64] (clone): Update.
glibc on ia64 provides no clone() function, but there is a __clone2
function instead. It is documented but no prototype is provided by
glibc.
* tests/ioctl_nsfs.c [IA64] (__clone2): New prototype.
[IA64] (clone): New macro wrapper around __clone2.
* tests/ioctl_v4l2.c (init_v4l2_format, print_ioctl_v4l2): New functions
to avoid code duplication in VIDIOC_S_FMT and VIDIOC_TRY_FMT tests.
(main) <VIDIOC_G_FMT, VIDIOC_S_FMT, VIDIOC_TRY_FMT>: Use them. Add
checks for remaining V4L2_BUF_TYPE_* types.
Signed-off-by: Edgar Kaziahmedov <edos@linux.com>
* v4l2.c: Include "xlat/v4l2_vbi_flags.h" and "xlat/v4l2_sliced_flags.h".
(struct_v4l2_clip): New typedef. Mpersify it.
(print_v4l2_clip): New function.
(print_v4l2_format_fmt): Use it. Add struct tcb argument. Implement
decoding of the remaining V4L2_BUF_TYPE_* types.
* xlat/v4l2_vbi_flags.in: Add V4L2_VBI_UNSYNC and V4L2_VBI_INTERLACED
introduced by linux kernel commit v2.5.46~39^2~23^2~4.
Add V4L2_VBI_ITU_525_F1_START, V4L2_VBI_ITU_525_F2_START,
V4L2_VBI_ITU_625_F1_START, and V4L2_VBI_ITU_625_F2_START introduced by
linux kernel commit v3.17-rc1~112^2~217.
* xlat/v4l2_sliced_flags.in: Add V4L2_SLICED_TELETEXT_B, V4L2_SLICED_VPS,
V4L2_SLICED_CAPTION_525, V4L2_SLICED_WSS_625, V4L2_SLICED_VBI_525, and
V4L2_SLICED_VBI_625 introduced by linux kernel commit v2.6.14-rc2~64.
* configure.ac (AC_CHECK_DECLS): Add V4L2_BUF_TYPE_SDR_CAPTURE and
V4L2_BUF_TYPE_SDR_OUTPUT.
(AC_CHECK_MEMBERS): Add struct v4l2_window.global_alpha and
struct v4l2_sdr_format.buffersize.
* NEWS: Mention this change.
Signed-off-by: Edgar Kaziahmedov <edos@linux.com>
This saves up to 25% of ptrace syscalls in case of trace filtering.
* syscall.c (trace_syscall_exiting): Do not call gettimeofday
and get_regs for filtered syscalls.
* NEWS: Mention this change.
* tests/get_regs.test: New test.
* tests/Makefile.am (MISC_TESTS): Add it.
strace used to call get_regs on every stop of a tracee, including cases
when the result is going to be discarded. Prepare to change this lavish
practice by delaying get_regs invocation until its result is actually
needed.
* defs.h (get_regs): Remove.
* strace.c (trace): Do not call get_regs.
* syscall.c (get_regs): Add static qualifier.
(trace_syscall_exiting, print_pc, get_scno): Call get_regs.
At this moment there is only one call to get_regs,
but there is going to be more of them, so a guard is needed.
This change also adds a clear_regs call before get_regs,
which essentially reverts commit v4.9-288-ge9bfff6.
* strace.c (trace): Call clear_regs before get_regs.
* syscall.c (get_regs): Skip if get_regs_error != -1.
* syscall.c (trace_syscall_entering, trace_syscall_exiting): Move
definitions of automatic variables closer to their first use. This
change slightly improves readability.
* tests/getpid.c: Skip if __NR_getpid is not defined.
* tests/getppid.c: Skip if __NR_getppid is not defined.
Fixes: fd1041f6 ("tests: add tests for getpid and getppid syscalls")
Make netlink_protocols a global xlat array defined in net.c file.
* defs.h (netlink_protocols): New prototype.
* socketutils.c: Stop including "xlat/netlink_protocols.h".
* xlat/v4l2_buf_flags.in: Add V4L2_BUF_FLAG_PREPARED,
V4L2_BUF_FLAG_NO_CACHE_INVALIDATE, and V4L2_BUF_FLAG_NO_CACHE_CLEAN
introduced by linux kernel commit v3.2-rc1~63^2~55.
Add V4L2_BUF_FLAG_TIMESTAMP_MASK, V4L2_BUF_FLAG_TIMESTAMP_UNKNOWN,
and V4L2_BUF_FLAG_TIMESTAMP_MONOTONIC introduced by linux kernel commit
v3.9-rc1~93^2~536.
Add V4L2_BUF_FLAG_TIMESTAMP_COPY introduced by linux kernel commit
v3.10-rc1~136^2~560.
Add V4L2_BUF_FLAG_TSTAMP_SRC_MASK, V4L2_BUF_FLAG_TSTAMP_SRC_EOF,
and V4L2_BUF_FLAG_TSTAMP_SRC_SOE introduced by linux kernel commit
v3.15-rc1~85^2~186.
Add V4L2_BUF_FLAG_LAST introduced by linux kernel commit
v4.2-rc1~107^2~389.
* xlat/v4l2_buf_types.in: Add V4L2_BUF_TYPE_SDR_CAPTURE
and V4L2_BUF_TYPE_SDR_OUTPUT introduced by linux kernel commits
v3.15-rc1~85^2~213 and v4.4-rc1~118^2~14, respectively.
* xlat/xlat/v4l2_colorspaces.in: Add V4L2_COLORSPACE_ADOBERGB
and V4L2_COLORSPACE_BT2020 introduced by linux kernel commit
v3.19-rc1~29^2~72.
Add V4L2_COLORSPACE_RAW and V4L2_COLORSPACE_DCI_P3 introduced by linux
kernel commits v4.2-rc1~107^2~182 and v4.4-rc1~118^2~120, respectively.
* xlat/v4l2_device_capabilities_flags.in: Add V4L2_CAP_SDR_CAPTURE,
V4L2_CAP_EXT_PIX_FORMAT, and V4L2_CAP_SDR_OUTPUT introduced by linux
kernel commits v3.15-rc1~85^2~210, v3.17-rc1~112^2~327,
and v4.4-rc1~118^2~14, respectively.
* xlat/v4l2_memories.in: Add V4L2_MEMORY_OVERLAY introduced
in linux 2.5.46. Add V4L2_MEMORY_DMABUF introduced by linux kernel
commit v3.8-rc1~115^2^2~52.
Signed-off-by: Edgar Kaziahmedov <edos@linux.com>
* mpers.awk (update_upper_bound): New function, which maintains new "count"
property for storing total element count and updates "upper_bound"
property which now contains string with array dimensions.
(/^DW_AT_upper_bound/, /^DW_AT_count/): Use it.
(what_is) <case "array_type">: Use "count" property in order to calculate
returned_size, do not embody returned string in square brackets.
* mpers_test.sh: Add checks for multidimensional arrays.
* tests/sched.test: Test that trace=%sched doesn't match
anything besides sched_* syscalls by checking it against all test
executables without side effects listed in pure_executables.list
except sched_* and times.
These are executables that could be used by several different tests
simultaneously.
* tests/pure_executables.list: New file.
* tests/Makefile.am (EXTRA_DIST): Add it.
This allows net-yy-unix invocation without arguments.
* tests/net-yy-unix.c (TEST_SOCKET): New macro.
(void): Use it instead of av[1].
* tests/net-yy-unix.test: Do not specify arguments for the test
executable.
This allows net-y-unix invocation without arguments.
* tests/net-y-unix.c (TEST_SOCKET): New macro.
(void): Use it instead of av[1].
* tests/net-y-unix.test: Do not specify arguments for the test
executable.
Most of test executables were ready for the upcoming change,
this change prepares all the rest.
* tests/getdents.c (main): Use fixed name for the sample directory.
* tests/getdents64.c (main): Likewise.
* tests/readdir.c (main): Likewise.
* tests/mknod.c (sample): Change to a pointer.
(main): Use av[0] as a sample.
* tests/symlink.c (main): Use av[0] as a linkpath.
* tests/umode_t.c (sample): Change to a pointer.
(test_syscall): Fix expected output.
(main): Use av[0] as a sample.
Due to a subtle gcc bug that leads to miscompiled aarch64 kernels,
the 3rd argument of sched_getattr syscall is not quite 32-bit
on aarch64 as on other architectures. For more details see
https://sourceforge.net/p/strace/mailman/message/35721703/
* defs.h (print_abnormal_hi): New prototype.
* util.c (print_abnormal_hi): New function.
* sched.c (SYS_FUNC(sched_getattr)) [AARCH64]: Use it.
* tests/sched_xetattr.c (main) [__arm64__ || __aarch64__]: Test it.
* linux/alpha/syscallent.h (getxgid, osf_stat, osf_lstat, osf_fstat,
osf_sysinfo, sethae): Set nargs according to arch/alpha/kernel/osf_sys.c
from linux v4.10.
* strace-graph: On long running process or heavily forking one (like
compilation), it can happen that a parent get 2 different children with
the same pid. By tracking the currently runnig pid and adding the start
timestamp to the pid, the graph can now handle that case.
Closes: https://github.com/strace/strace/pull/7
Automatically convert tests to use TAIL_ALLOC_OBJECT_CONST_PTR macro
with the following sed expression:
sed -ri \
's/^([[:space:]]*)(([^*]+[[:alnum:]_])( \*)?) *\*( *const )?([^ =]+) = tail_alloc\(sizeof\((\2|\*\6)\)\);/\1TAIL_ALLOC_OBJECT_CONST_PTR(\2, \6);/' \
tests/*.c
Automatically convert tests to use TAIL_ALLOC_OBJECT_VAR_PTR macro
with the following sed expression:
sed -ri \
's/^([[:space:]]*)(([^*]+[[:alnum:]_])( \*)?) *\*([^ =]+) = tail_alloc\(sizeof\((\2|\*\5)\)\);/\1TAIL_ALLOC_OBJECT_VAR_PTR(\2, \5);/' \
tests/{move_pages,net-sockaddr,sendfile}.c
Introduce a new macros for the most widespread use of tail_alloc.
* tests/tests.h (TAIL_ALLOC_OBJECT_CONST_PTR,
TAIL_ALLOC_OBJECT_VAR_PTR): New macros.
Pipelined commands executed in subprocesses (some shells have special
provision for last command in the pipeline, but it is not guaranteed
and not POSIX), so exit codes of these subprocesses can be lost.
In order to avoid creating subprocesses, input for read commands
is better to supply via heredoc and not pipe.
* tests/sched.test: Move test cases to heredoc.
xetitimer.test already implements a comprehensive test of setitimer
syscall parser, no need to duplicate its parts in other tests.
* tests/clock_nanosleep.c (main): Remove setitimer expected output.
* tests/nanosleep.c (main): Likewise.
* tests/clock_nanosleep.test: Remove setitimer from the trace set.
* tests/nanosleep.test: Likewise.
* tests/sched_xetattr.c: Include <unistd.h> and <asm/unistd.h>.
(main): Check that pointer and integer arguments of getitimer
and setitimer syscalls are decoded properly.
* tests/xetitimer.test: Add -a option.
Unlike the former test that was based on match_grep,
the new one uses match_diff and does more rigorous testing.
* tests/sigreturn.c: Rewrite.
* tests/sigreturn.test: Likewise.
As there is no sigreturn syscall on ia64, mips n32, mips n64, x32,
and x86_64, no longer used parsers could be safely removed.
* linux/ia64/arch_sigreturn.c: Remove.
* Makefile.am (EXTRA_DIST): Remove it.
* linux/mips/arch_sigreturn.c (arch_sigreturn): Remove [!LINUX_MIPSO32].
* linux/x86_64/arch_sigreturn.c (arch_sigreturn): Remove.
All architectures have rt_sigreturn, but only old ones also have
sigreturn. Since not all architectures need a parser of sigreturn
syscall, remove the warning and move the parser to libstrace.
* Makefile.am (strace_SOURCES): Move sigreturn.c ...
(libstrace_a_SOURCES): ... here.
* linux/arch_sigreturn.c: Remove warning.
Linux does not prevent a user from creating a lot of nested directories
with length of the absolute path of the deepest one exceeding PATH_MAX,
then chdir'ing into it, creating a file there and mmap'ing it. Since the
length of the prefix preceding the pathname in /proc/[pid]/maps is not
necessary 80 (it's 73 on my machine), an overflow is possible.
* unwind.c (build_mmap_cache): Fix a possible buffer overflow.
Export SP register for later use by get_rt_sigframe_addr.
* linux/riscv/arch_regs.h: New file.
* Makefile.am (EXTRA_DIST): Add it.
* linux/riscv/arch_regs.c (riscv_sp_ptr): New variable.
Export SP register for later use by get_rt_sigframe_addr.
* linux/or1k/arch_regs.h: New file.
* Makefile.am (EXTRA_DIST): Add it.
* linux/or1k/arch_regs.c (or1k_sp_ptr): New variable.
Export SP register for later use by get_rt_sigframe_addr.
* linux/nios2/arch_regs.h: New file.
* Makefile.am (EXTRA_DIST): Add it.
* linux/nios2/arch_regs.c (nios2_sp_ptr): New variable.
Export SP register for later use by get_rt_sigframe_addr.
* linux/metag/arch_regs.h: New file.
* Makefile.am (EXTRA_DIST): Add it.
* linux/metag/arch_regs.c (metag_sp_ptr): New variable.
Export SP register for later use by get_rt_sigframe_addr.
* linux/avr32/arch_regs.h: New file.
* Makefile.am (EXTRA_DIST): Add it.
* linux/avr32/arch_regs.c (avr32_sp_ptr): New variable.
Export SP register for later use by get_rt_sigframe_addr.
* linux/arc/arch_regs.h: New file.
* Makefile.am (EXTRA_DIST): Add it.
* linux/arc/arch_regs.c (arc_sp_ptr): New variable.
As the definition of NSIG_BYTES is based on NSIG defined in <signal.h>,
NSIG_BYTES cannot be used in files that use kernel sigset_t.
Add another wrapper around print_sigset_addr_len_limit function that
takes one argument less than print_sigset_addr_len wrapper. This new
wrapper is going to be used in cases when the length argument is equal
to NSIG_BYTES.
* defs.h (print_sigset_addr): New function prototype.
* signal.c (print_sigset_addr): New function.
* linux/alpha/arch_sigreturn.c (arch_sigreturn): Replace
print_sigset_addr_len with print_sigset_addr.
* linux/arm/arch_sigreturn.c (arch_sigreturn): Likewise.
* linux/crisv10/arch_sigreturn.c (arch_sigreturn): Likewise.
* linux/ia64/arch_sigreturn.c (arch_sigreturn): Likewise.
* linux/microblaze/arch_sigreturn.c (arch_sigreturn): Likewise.
* linux/mips/arch_sigreturn.c (arch_sigreturn): Likewise.
* linux/tile/arch_sigreturn.c (arch_sigreturn): Likewise.
* linux/x86_64/arch_sigreturn.c (arch_sigreturn): Likewise.
(as suggested in
https://www.mail-archive.com/strace-devel@lists.sourceforge.net/msg05147.html )
In order to move them to a different namespace than syscall names. This is
already a problem in case of ipc class which shadows ipc syscall.
Old behaviour is retained in order to preserve backwards compatibility.
* qualify.c (lookup_class) <syscall_class>: Add %-prefixed class definitions.
* strace.1 (.SS Filtering): Add information regarding %-prefixed class
syntax, declare legacy class syntax deprecated.
* tests/net.test: Update to use %-prefixed syscall class name.
* tests/netlink_protocol.test: Likewise.
* tests/qual_fault-exit_group.test: Likewise.
* tests/qual_syscall.test: Likewise.
* tests/scm_rights-fd.test: Likewise.
* tests/sigreturn.test: Likewise.
* tests/uio.test: Likewise.
* NEWS: Mention this change.
* tests/print_time.c (print_time_t): Rename to print_time_t_nsec,
take second argument and print it.
* tests/tests.h (print_time_t): Rename to print_time_t_nsec,
add second argument.
* tests/utime.c (main): Use print_time_t_nsec instead of print_time_t.
* tests/xstatx.c (print_stat): Likewise. Pass nanoseconds
to print_time_t_nsec instead of printing them.
Rename print_time function to print_time_t and move it to libtests.
* tests/print_time.c: New file.
* tests/Makefile.am (libtests_a_SOURCES): Add it.
* tests/tests.h (print_time_t): New prototype.
* tests/print_time.c (print_time): Remove.
(print_stat): Replace print_time with print_time_t.
x32 is the only linux architecture where preadv2 and pwritev2 syscalls
take 5 arguments instead of 6.
* io.c (PREADV2_PWRITEV2_FLAGS_ARG_NO): New macro.
(SYS_FUNC(preadv2), SYS_FUNC(pwritev2)): Use it.
* linux/x32/syscallent.h (preadv2, pwritev2): Change nargs to 5.
* NEWS: Mention this fix.
* tests/preadv2-pwritev2.c (main): Fix invocation of preadv2
and pwritev2 syscalls on x32.
Unlike the former test that was based on match_grep,
the new one uses match_diff and does more rigorous testing.
* tests/ppoll-v.c: New file.
* tests/ppoll-v.test: New test.
* tests/ppoll.c: Rewrite.
* tests/ppoll.test: Likewise.
* tests/ppoll.expected: Remove.
* tests/ppoll-v.expected: Remove.
* tests/.gitignore: Add ppoll-v.
* tests/Makefile.am (check_PROGRAMS): Likewise.
(DECODER_TESTS): Add ppoll-v.test.
(EXTRA_DIST): Remove ppoll.expected and ppoll-v.expected.
* tests/get_sigset_size.c: New file.
* tests/Makefile.am (libtests_a_SOURCES): Add it.
* tests/tests.h (get_sigset_size): New prototype.
* tests/ptrace.c: Do not check for __NR_rt_sigprocmask.
(main): Use get_sigset_size.
* tests/signalfd4.c: Do not check for __NR_rt_sigprocmask.
(get_sigset_size): Remove.
Implement read/write semantics of struct sched_attr.size argument
of sched_setattr syscall. Do not print members of struct sched_attr
besides sched_attr.size when the specified structure size is less than
the minimal size allowed by the kernel.
* sched.c (print_sched_attr): Fetch struct sched_attr.size and use it
as the structure size. Print struct sched_attr.size only when the
structure size is less than SCHED_ATTR_MIN_SIZE.
(SYS_FUNC(sched_setattr)): Call print_sched_attr with zero size
argument. Print struct sched_attr.size returned by the kernel
on exiting syscall in case of E2BIG. Print the last syscall argument
on exiting syscall.
(SYS_FUNC(sched_getattr)): Do not call print_sched_attr with zero size
argument.
* NEWS: Mention it.
* tests/sched_xetattr.c (main): Check it.
* tests/sched_xetattr.c (main): Check that integer arguments
of sched_getattr and sched_setattr syscalls are decoded properly
by adding a few cases with filled higher 32 bits.
Check that pointer argument is decoded properly
on F8ILL_KULONG_SUPPORTED architectures.
Avoid multiple defintions of struct sched_attr by creating a separate
header file with its definition and using it in other places.
* sched_attr.h: New file.
* Makefile.am (strace_SOURCES): Add it.
* sched.c: Include it.
(print_sched_attr): Use it.
* tests/sched_xetattr.c: Include it.
(main): Use it.
Unlike the previous edition of the test that was based on match_grep,
the new one is match_diff based and does more rigorous testing.
* tests/adjtimex.c: Include "xlat.h" and "xlat/schedulers.h",
(sys_sched_getattr, sys_sched_setattr): New functions.
(main): Use them. Update expected output.
* tests/adjtimex.test: Use run_strace_match_diff.
Implement decoding of BPF_F_ALLOW_OVERRIDE flag of BPF_PROG_ATTACH
command introduced by linux kernel commit v4.10-rc7-174-g7f67763.
* configure.ac: Check for union bpf_attr.attach_flags
instead of union bpf_attr.attach_type.
* xlat/bpf_attach_flags.in: New file.
* bpf.c: Include "xlat/bpf_attach_flags.h".
(bpf_prog_attach_detach): Rename print_attach_bpf_fd argument
to print_attach. Add attach_flags field to the structure,
print it in case of BPF_PROG_ATTACH.
* tests/bpf.c: Check for HAVE_UNION_BPF_ATTR_ATTACH_FLAGS
instead of HAVE_UNION_BPF_ATTR_ATTACH_TYPE.
(prog_cgroup): Initialize attach_flags field.
(main): Update expected output.
Unlike the previous edition of the test that was based on match_grep,
the new one is match_diff based and does more rigorous testing.
* tests/adjtimex.c: Include "xlat.h", "xlat/adjtimex_state.h",
and "xlat/adjtimex_status.h".
(main): Update expected output.
* tests/adjtimex.test: Use run_strace_match_diff.
Check that integer arguments of fcntl and fcntl64 syscalls are decoded
properly by casting them to kernel_ulong_t and filling higher bits.
* tests/struct_flock.c (invoke_test_syscall): Cast "fd" and "cmd"
syscall arguments to kernel_ulong_t, fill their higher 32 bits.
capability.c:82:28: error: duplicate "const" declaration specifier
capability.c:110:33: error: duplicate "const" declaration specifier
rt_tgsigqueueinfo.c:42:61: error: duplicate "const" declaration specifier
utime.c:66:23: error: duplicate "const" declaration specifier
waitid.c:147:20: error: duplicate "const" declaration specifier
* capability.c (cap_user_header_t, cap_user_data_t): Remove.
(get_cap_header): Change return type
to "const struct user_cap_header_struct *".
(print_cap_header, print_cap_data): Change the type of last argument
to "const struct user_cap_header_struct * const".
(SYS_FUNC(capget)): Change type of "h" variable
to "const struct user_cap_header_struct *".
(SYS_FUNC(capset)): Change type of "h" variable
to "const struct user_cap_header_struct * const".
* tests/rt_tgsigqueueinfo.c (k_tgsigqueueinfo): Change the type of last
argument to "const void *const".
* tests/utime.c (main): Change the type of "tail_u" variable
to "const struct utimbuf *const".
* tests/waitid.c (do_waitid): Change the type of 3rd argument
to "const siginfo_t *const".
* configure.ac (AC_CHECK_FUNCS): Remove pipe2.
* tests/pipe.c: Include <asm/unistd.h>, skip the test if [!__NR_pipe]
instead of [!HAVE_PIPE2].
(main): Do not call pipe2.
* tests/pipe.test: Skip the test if libc pipe wrapper does not use
pipe syscall, rewrite from match_grep to match_diff.
* tests/pipe.expected: Update expected output.
Unlike pipe.test that is based on match_grep, the new test
is based on match_diff and does more rigorous testing.
* tests/pipe2.c: New file.
* tests/pipe2.test: New test.
* tests/.gitignore: Add pipe2.
* tests/Makefile.am (check_PROGRAMS): Likewise.
(DECODER_TESTS): Add pipe2.test.
Unlike the previous edition of the test that was based on match_grep,
the new one is match_diff based and does more rigorous testing.
* tests/mq.c: Include <stdio.h>.
(NAME): New macro.
(main): Use it. Print expected output.
* tests/mq.expected: Remove.
* tests/Makefile.am (EXTRA_DIST): Remove mq.expected.
* tests/mq.test: Use run_strace_match_diff.
Drop use of INJECT_OPTS_RVAL_DISABLE special inject_opts.rval value,
INJECT_OPTS_RVAL_DEFAULT is enough.
* defs.h (INJECT_OPTS_RVAL_DISABLE): Remove.
* qualify.c (qualify_inject_common): Do not replace
INJECT_OPTS_RVAL_DEFAULT with INJECT_OPTS_RVAL_DISABLE
in inject_opts.rval.
* syscall.c (tamper_with_syscall_entering): Check
for INJECT_OPTS_RVAL_DEFAULT instead of INJECT_OPTS_RVAL_DISABLE.
Follow the example of case-insensitive symbolic signal values in signal
and inject expressions and make symbolic errno values in inject
expressions case-insensitive.
* qualify.c (find_errno_by_name): Use strcasecmp instead of strcmp
to match symbolic errno values.
* tests/qual_fault.c (main): Likewise.
* tests/qual_fault.test: Test case-insensitive symbolic errno match.
Introduce -e inject= as a general injection option,
limit -e fault= option to syscall fault injection.
Change default return code of syscall fault injection to ENOSYS.
* qualify.c (parse_inject_token): Add fault_tokens_only argument,
do not accept retval= and signal= tokens when fault_tokens_only
is set to true.
(parse_inject_expression): Add fault_tokens_only argument,
forward it to parse_inject_token.
(qualify_inject_common): New function.
(qualify_fault): Use it.
(qualify_inject): New function.
(qual_options): New entry.
* strace.1: Describe -e inject= option.
* NEWS: Mention -e inject= option.
* tests/qual_fault-syntax.test: Test that -e fault= option does not
support retval=, signal=, and multiple error= tokens.
* tests/qual_fault.c (DEFAULT_ERRNO): Set to ENOSYS unconditionally.
* tests/qual_inject-retval.test: Replace -e fault= option
with -e inject= option.
* tests/qual_inject-syntax.test: New test.
* tests/Makefile.am (MISC_TESTS): Add it.
Rename files related to fault injection to avoid further confusion.
* tests/fault_injection-exit_group.expected: Rename
to tests/qual_fault-exit_group.expected.
* tests/fault_injection-exit_group.test: Rename
to tests/qual_fault-exit_group.test.
* tests/fault_syntax.test: Rename to tests/qual_fault-syntax.test.
* tests/fault_injection.c: Rename to tests/qual_fault.c.
* tests/fault_injection.test: Rename to tests/qual_fault.test.
* tests/fault_injection-retval.c: Rename to tests/qual_inject-retval.c.
* tests/fault_injection-retval.test: Rename
to tests/qual_inject-retval.test.
* tests/fault_injection.sh: Rename to tests/scno_tampering.sh.
* tests/.gitignore: Update.
* tests/Makefile.am (check_PROGRAMS, DECODER_TESTS, MISC_TESTS,
EXTRA_DIST): Update.
As -e fault= injection syntax has been extended to configure various
kinds of injections besides fault injection, the original option name
bacame too narrow.
This change renames internal constants, types, and functions from
"fault" to more generic inject/tamper ones.
* defs.h (fault_opts): Rename to inject_opts.
(FAULT_OPTS_RVAL_DEFAULT): Rename to INJECT_OPTS_RVAL_DEFAULT.
(FAULT_OPTS_RVAL_DISABLE): Rename to INJECT_OPTS_RVAL_DISABLE.
(struct tcb): Rename fault_vec field to inject_vec.
(TCB_FAULT_INJ): Rename to TCB_TAMPERED.
(QUAL_FAULT): Rename to QUAL_INJECT.
(fault_vec): Rename to inject_vec.
All users changed.
* qualify.c (fault_set): Rename to inject_set.
(parse_fault_token): Rename to parse_inject_token.
(parse_fault_expression): Rename to parse_inject_expression.
All callers changed.
* syscall.c (fault_vec): Rename to inject_vec.
(tcb_fault_opts): Rename to tcb_inject_opts.
(inject_syscall_fault_entering): Rename to tamper_with_syscall_entering.
(update_syscall_fault_exiting): Rename to tamper_with_syscall_exiting.
(syscall_fault_injected): Rename to syscall_tampered.
All callers changed.
Drop local VAL_STR macro in favour of ARG_STR macro from tests/tests.h.
* tests/msg_control.c (VAL_STR): Remove.
(test_sol_socket, test_sol_ip): Replace VAL_STR with ARG_STR.
This fixes a typo in the arch_set_success for the or1k architecture.
* linux/or1k/set_error.c (arch_setsuccess): Rename to arch_set_success.
Fixes: 41d647c ("Implement success injection")
Closes: https://github.com/strace/strace/pull/6
* tests/bpf.c (errstr): New variable.
(sys_bpf): New function.
(map_create, map_any, prog_load, obj_manage, prog_cgroup): Use it.
(main): Update expected output.
* tests/bpf.test: Use run_strace_match_diff.
Implement decoding of BPF_OBJ_PIN, BPF_OBJ_GET, BPF_PROG_ATTACH,
and BPF_PROG_DETACH commands.
* bpf.c: Include "xlat/bpf_attach_type.h".
(bpf_obj_manage, bpf_prog_attach, bpf_prog_detach): New functions.
(SYS_FUNC(bpf)): Use them.
* configure.ac: Check for union bpf_attr.bpf_fd and union
bpf_attr.attach_type.
* xlat/bpf_attach_type.in: New file.
* xlat/bpf_commands.in: Update list of BPF_* command constants.
* xlat/bpf_map_types.in: Update list of BPF_MAP_TYPE_* constants.
* xlat/bpf_prog_types.in: Update list of BPF_PROG_TYPE_* constants.
* tests/bpf.c [HAVE_UNION_BPF_ATTR_BPF_FD] (obj_manage): New function.
[HAVE_UNION_BPF_ATTR_ATTACH_TYPE] (prog_cgroup): Likewise.
(main): Use them.
Use builtin macros provided by gcc >= 3.4 instead of sgidefs.h macros.
* configure.ac <mips>: Do not include <asm/sgidefs.h>, use _ABIO32
instead of _MIPS_SIM_ABI32, _ABIN32 instead of _MIPS_SIM_NABI32,
and _ABI64 instead of _MIPS_SIM_ABI64.
Build fix for MIPS with musl libc.
The MIPS specific header <sgidefs.h> is provided by glibc and uclibc
but not by musl. Regardless of the libc, the kernel headers provide
<asm/sgidefs.h> which provides the same definitions, so use that
instead.
* configure.ac: Replace <sgidefs.h> with <asm/sgidefs.h>.
Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
* Makefile.am (m%_type_defs.h): Change sed regular expression to allow
arbitrary number of whitespace characters between "#" symbol
and "include" directive.
With the switch of fstatat family tests to USE_ASM_STAT variant,
there are no non-USE_ASM_STAT users left. Remove support of unused
non-USE_ASM_STAT case.
* configure.ac (AC_CHECK_MEMBERS): Remove struct stat.st_mtim.tv_nsec.
* tests/fstatat.c (USE_ASM_STAT): Remove.
* tests/fstatx.c: Likewise.
* tests/lstatx.c: Likewise.
* tests/xstatx.c: Assume USE_ASM_STAT.
[!USE_ASM_STAT]: Remove.
Do not use glibc wrappers to call newfstatat and fstatat64.
These wrappers have various problems, e.g. they segfault on sparc64
and mips64 if BOGUS_STRUCT_STAT is not disabled, and they do wrong
conversion of timestamps on mips64.
* tests/fstatat.c (TEST_SYSCALL_INVOKE): Invoke the relevant syscall
directly.
(USE_ASM_STAT): Define.
* tests/fstatat64.c (TEST_BOGUS_STRUCT_STAT): Remove.
(STRUCT_STAT, STRUCT_STAT_STR, STRUCT_STAT_IS_STAT64): Define.
Based on patch by James Cowgill <james410@cowgill.org.uk>.
Remove all definitions of SAMPLE_SIZE that are identical to the fallback
definition in tests/xstatx.c.
* tests/fstat64.c (SAMPLE_SIZE): Remove.
* tests/lstat64.c (SAMPLE_SIZE): Remove.
* tests/stat64.c (SAMPLE_SIZE): Remove.
As ftruncate libc function and our create_sample function that calls
ftruncate both take size argument of type libc_off_t, change the type
of SAMPLE_SIZE constant to libc_off_t.
* tests/fstat.c (SAMPLE_SIZE): Cast to libc_off_t.
* tests/lstat.c (SAMPLE_SIZE): Likewise.
* tests/oldfstat.c (SAMPLE_SIZE): Likewise.
* tests/oldlstat.c (SAMPLE_SIZE): Likewise.
* tests/oldstat.c (SAMPLE_SIZE): Likewise.
* tests/stat.c (SAMPLE_SIZE): Likewise.
* tests/xstatx.c (SAMPLE_SIZE): Likewise.
When cross-compiling, ioctlsort must obtain _IOC_* values from the
host, build's values may be incompatible.
* ioctl_iocdef.c: New file.
* Makefile.am (EXTRA_DIST): Add it.
(ioctl_iocdef.i, ioctl_iocdef.h): New rules.
(ioctlsort%.o): Depend on ioctl_iocdef.h.
(CLEANFILES): Add ioctl_iocdef.h and ioctl_iocdef.i.
* ioctlsort.c: Include "ioctl_iocdef.h" instead of <linux/ioctl.h>.
Signed-off-by: Alexey Neyman <stilor@att.net>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
If the test detects struct stat mismatch, this is likely an error
in definitions of a stat structure that might affect strace as well.
Fail the test instead of skipping it to attract more attention.
* tests/xstatx.c (main): Return 1 instead of 77
in case of struct stat mismatch.
Follow the example of strace.1 and put the description of -k option
into the output format group.
* strace.c (usage): Move description of -k option to the output
format group.
* print_sg_req_info.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* xlat/sg_scsi_reset.in: New file.
* scsi.c: Include "xlat/sg_scsi_reset.h".
(scsi_ioctl): Implement decoding of all SG_* ioctl commands.
Unfortunately, <scsi/sg.h> from libc does not provide all constants
recognized by the kernel.
* xlat/scsi_sg_commands.in: New file.
* scsi.c: Include "xlat/scsi_sg_commands.h" instead of providing
a fallback definition of SG_IO.
Turn print_sg_io_buffer into a universal sg_io buffer printer.
* sg_io_v3.c (print_sg_io_buffer): Add iovec_count argument.
Call tprint_iov_upto when iovec_count is non-zero.
(decode_request, decode_response): Do not call tprint_iov_upto directly,
use print_sg_io_buffer instead.
* sg_io_v4.c: Likewise.
The SCSI driver, starting with kernel commit v2.6.25-rc1~1230^2~78,
translates its "struct request.errors" to ioctl errors after
filling in all the output members of the SG_IO header structure.
As there is no easy way to tell SCSI layer errors from other syscall
errors, decode the structure on exiting syscall in case of syserror,
too.
* scsi.c (scsi_ioctl): Do not call set_tcb_priv_ulong, use
get_tcb_priv_data instead of get_tcb_priv_ulong, call decode_sg_io
unconditionally.
* sg_io_v3.c (decode_request): Save a copy of struct_sg_io_hdr using
set_tcb_priv_data.
(decode_response): Restore it using get_tcb_priv_data. Print its i/o
fields when umove call fails. Pass IOV_DECODE_STR to tprint_iov_upto
unconditionally.
* sg_io_v4.c (decode_request): Save a copy of struct sg_io_v4 using
set_tcb_priv_data.
(decode_response): Restore it using get_tcb_priv_data. Print its i/o
fields when umove call fails. Pass IOV_DECODE_STR to tprint_iov_upto
unconditionally.
The formerly used format string %02x led to misleading output as there
was no clear indication sometimes whether the printed integer was
decimal or hexadecimal.
* sg_io_v3.c (decode_response): Print status and masked_status fields
using %#x format specifier instead of %02x.
The tradition is to print scsi status codes in a hexadecimal form.
* sg_io_v4.c (decode_response): Print driver_status, transport_status,
and device_status fields using %#x format specifier instead of %u.
The tradition is to print scsi tags in a hexadecimal form.
* sg_io_v4.c (decode_request): Print request_tag field
using ("%#" PRI__x64) format string.
(decode_response): Print generated_tag field using ("%#" PRI__x64)
format string.
As the nature of this field is pointer-like, print it
in a hexadecimal form.
* sg_io_v4.c (decode_request): Print usr_ptr field
using ("%#" PRI__x64) format string.
As the kernel does not touch these fields, there is no use
to print them.
* sg_io_v4.c (decode_request): Do not print spare_in field.
(decode_response): Do not print spare_out field.
Print struct sg_io_hdr.info and struct sg_io_v4.info using printflags.
* xlat/sg_io_info.in: New file.
* defs.h (sg_io_info): New xlat prototype.
* sg_io_v3.c: Include "xlat/sg_io_info.h".
(decode_response): Print struct_sg_io_hdr.info using printflags.
* sg_io_v4.c: (decode_response): Print struct sg_io_v4.info
using printflags.
Print struct sg_io_hdr.flags and struct sg_io_v4.flags using printflags.
* xlat/sg_io_flags.in: New file.
* sg_io_v3.c: Include "xlat/sg_io_flags.h".
(decode_request): Print struct_sg_io_hdr.flags using printflags.
* xlat/bsg_flags.in: New file.
* sg_io_v4.c: Include "xlat/bsg_flags.h".
(decode_request): Print struct sg_io_v4.flags using printflags.
struct sg_io_hdr contains pointers and therefore has to be mpersified.
* defs.h (decode_sg_io_v3): Remove prototype.
* sg_io_v3.c (struct_sg_io_hdr): Typedef to struct sg_io_hdr.
Mpersify it.
(decode_sg_io_v3): Wrap into MPERS_PRINTER_DECL.
* scsi.c (print_sg_io_v3_res): Detect and print changes
of struct sg_io_hdr.interface_id field between entering and exiting
syscall.
(print_sg_io_v4_res): Detect and print changes of struct sg_io_v4.guard
field between entering and exiting syscall.
* scsi.c (print_sg_io_v3_req): Always print the name
of struct sg_io_hdr.interface_id field.
(print_sg_io_v4_req): Always print the name of struct sg_io_v4.guard
field.
(print_sg_io_req): Print unknown interface id as an integer.
(scsi_ioctl): Save interface id on entering syscall and use it
on exiting syscall.
The formerly used format string %x led to misleading output as there was
no clear indication sometimes whether the printed integer was decimal or
hexadecimal.
* aio.c (startup_child, test_ptrace_seize): Print status
using %#x format specifier instead of %x.
The formerly used format string %x led to misleading output as there was
no clear indication sometimes whether the printed integer was decimal or
hexadecimal.
* aio.c (print_common_flags): Print struct iocb.aio_flags using %#x
format specifier instead of %x.
* tests/aio.c (main): Likewise.
As printstr_ex has got QUOTE_FORCE_HEX flag support, there is no need
to implement it manually any longer.
* scsi.c (print_uchar): Remove.
(print_sg_io_buffer): Rewrite using printstr_ex(QUOTE_FORCE_HEX).
Since there is no reason to interpret the value returned by getrandom as
ASCII string, it makes sense to always print it as a hex-escaped string.
* getrandom.c (SYS_FUNC(getrandom)): Use printstr_ex instead
of printstrn, set QUOTE_FORCE_HEX in user_style parameter in order
to force hex-escaped string formatting.
* tests/getrandom.test: Remove no longer needed -xx flag as the string
is always printed in hexadecimal format now.
Suggested-by: JingPiao Chen <chenjingpiao@gmail.com>
This could be useful in cases when some binary data should not be
interpreted as an ASCII string, but rather as an array of byte values.
* defs.h (QUOTE_FORCE_HEX): New macro constant.
* util.c (quote_string): Enable use_hex when QUOTE_FORCE_HEX is set
in user_style parameter.
The case of unset verbose is handled by umoven_or_printaddr, moreover,
exiting at this point is plain wrong and leads to last argument not
being printed while it should be.
* term.c (decode_termios, decode_termio, decode_winsize, decode_ttysize,
decode_modem_flags): Do not check verbose flags explicitly and do not
exit early if it is not set.
Due to a bug in glibc readahead syscall wrapper on MIPS n64,
this wrapper cannot be used in readahead.test.
* tests/readahead.c [HAVE_READAHEAD && LINUX_MIPSN64 && glibc < 2.25]:
Undefine HAVE_READAHEAD.
Reported-by: James Cowgill <james410@cowgill.org.uk>
* tests/ioctl_v4l2.c (cc0, cc1, cc2, cc3, fourcc): New macros.
(main): Use them. Rewrite expected output without assumptions
on any particular magic data.
As these functions behave like memset, it's more convenient to have
the first argument of type void * like memset.
* tests/fill_memory.c (fill_memory, fill_memory_ex): Change the type
of first argument from "char *" to "void *".
* tests/tests.h (fill_memory, fill_memory_ex): Likewise.
Drop custom vfprintf implementation that was disabled by default.
Apparently nobody tests strace with this vfprintf enabled, otherwise
multiple uses of format specifiers not supported by this custom vfprintf
would not left unnoticed.
The GNU C library is not the only libc available, so those who want
faster implementations of libc functions are encouraged to try building
strace with other libc implementations.
* vsprintf.c: Remove.
* Makefile.am (strace_SOURCES): Remove vsprintf.c.
* defs.h (USE_CUSTOM_PRINTF, strace_vfprintf): Remove.
* strace.c (tprintf): Replace strace_vfprintf with vfprintf.
The verbose flag is responsible for dereferencing of addresses, and it
is perfectly handled by umove_or_printaddr; moreover, this early exit
leads to incorrect formatting of last argument.
* loop.c (loop_ioctl): Do not check verbose flag explicitly and do not
exit early if it is not set.
lo_encrypt_key is a sized string with its size specified by
lo_encrypt_key_size field, so take lo_encrypt_key_size into account.
* loop.c (decode_loop_info, decode_loop_info64): Use minimum of
LO_KEY_SIZE and lo_encrypt_key_size field value as lo_encrypt_key size.
It is converted to an unsigned value in kernel (see loop_info64_from_old
in drivers/block/loop.c), so let's print it that way despite its type.
* loop.c (decode_loop_info): Print lo_encrypt_key_size as an uint32_t
value.
* tests/ioctl.c (main): Additional check for unknown ioctl command
decoding.
* tests/ioctl_evdev.c (main): Likewise.
* tests/ioctl_dm.c: Include <linux/ioctl.h> for _IOC_SIZE and _IOC_NR.
(main) <dummy_dm_ioctl1, dummy_dm_ioctl2, dummy_dm_arg>: New constants.
Add more checks for unknown command formatting.
Specifying output width for size argument is rather misleading - it can
be up to 14 bits in size. The use of explicit "0x" prefix is
discouraged because 0x0 looks ugly. Usage of width specification along
with alternate form flag leads to inconsistent output for 0.
Let's use just plain %#x qualifiers instead.
* ioctl.c (ioctl_print_code): Change printf qualifiers for ioctl command
type, number and size to "%#x".
* tests/ioctl.c (main): Update expected test output.
* tests/ioctl_dm.c (main): Likewise.
* tests/ioctl_evdev.c (main): Likewise.
They are of char type, so "hh" length modifier should be used (otherwise
they are printed incorrectly as they are sign extended implicitly when
passed as printf arguments in case char is a signed type which is true
for most compilers).
* loop.c (decode_loop_info): Change "%#x" printf qualifier to "%#hhx"
in printing routine for array elements of the "reserved" field of
loop_info structure.
* strace.c: Remove __NR_tkill checks, assume it is always defined.
* util.c: Remove __NR_process_vm_readv checks and fallback definitions,
assume it is always defined.
Generate scno.h using only those tools that are available for cross
build. This allows to use scno.h during the build of strace itself.
* scno.am: New file.
* scno.head: New file.
* Makefile.am (EXTRA_DIST): Add it.
* tests/generate_scno.c: Remove.
* tests/.gitignore: Remove generate_scno.
* tests/Makefile.am (check_PROGRAMS): Likewise.
(CLEANFILES): Remove scno.h.
(scno.h, $(scno_dependants)): Remove rules.
(scno_dependants): Remove.
($(objects)): New rule.
Include ../scno.am.
* COPYING: Update copyright year number range.
* configure.ac: Likewise.
* strace.c: Likewise.
* tests/strace-V.test: Use the current year for the check.
Limit the number of files where <asm/unistd.h> is included
to only those few that need it.
* defs.h: Do not include <asm/unistd.h>.
* clone.c: Include <asm/unistd.h>.
* strace.c: Likewise.
* syscall.c: Likewise.
* util.c: Likewise.
* tests/generate_scno.c: Include "tests.h", "sysent.h", <stdio.h>,
<string.h>, and <asm/unistd.h> instead of "defs.h" and "syscall.h".
* tests/ksysent.c: Likewise.
* tests/nsyscalls.c: Include "tests.h", "sysent.h", <errno.h>,
<stdio.h>, and <unistd.h> instead of "defs.h" and "syscall.h".
struct ifreq and struct ifconf are ABI-dependent and should have been
mpersified.
* defs.h (sock_ioctl): Remove prototype.
* sock.c (struct_ifconf, struct_ifreq): New typedefs.
(DEF_MPERS_TYPE(struct_ifconf), DEF_MPERS_TYPE(struct_ifreq),
MPERS_DEFS): New includes.
(print_ifreq, print_ifc_len, print_ifconf_ifreq): Replace struct ifreq
with struct_ifreq.
(decode_ifconf): Replace struct ifconf with struct_ifconf, struct ifreq
with struct_fireq.
(sock_ioctl): Wrap into MPERS_PRINTER_DECL.
Since these are architecture-independent and have been added over the
kernel's lifetime, it makes sense to have these definitions in strace
itself.
* xlat/loop_cmds.in: New file.
* loop.c: Include "xlat/loop_cmds.h".
(loop_ioctl): Eliminate conditional compilation based on availability
of LOOP_* constant definitions.
* xlat/gen.sh: Implement #stop directive that stops preprocessor
at the line it has been encountered during the first pass. This could
be used, for example, to provide fallback definitions for constants
without defining xlat tables.
struct loop_info contains fields of type long, so it definitely should
have been mpersified.
* defs.h (loop_ioctl): Remove prototype.
* loop.c (struct_loop_info): New typedef.
Include DEF_MPERS_TYPE(struct_loop_info) and MPERS_DEFS.
(decode_loop_info): Replace struct loop_info with struct_loop_info.
Cast lo_inode, lo_init[0], and lo_init[1] to kernel_ulong_t and print
them using appropriate format specifiers.
(loop_ioctl): Wrap into MPERS_PRINTER_DECL.
Co-authored-by: Dmitry V. Levin <ldv@altlinux.org>
* print_dev_t.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* defs.h (print_dev_t): New prototype.
* dm.c: Do not include <sys/sysmacros.h>.
(dm_decode_device, dm_print_dev, dm_decode_dm_name_list): Use
print_dev_t function for printing device numbers.
* mknod.c: Do not include <sys/sysmacros.h>.
(decode_mknod): Use print_dev_t function for printing device number.
* print_struct_stat.c: Do not include <sys/sysmacros.h>.
(print_struct_stat): Use print_dev_t function for printing device
numbers.
Co-authored-by: Dmitry V. Levin <ldv@altlinux.org>
* tests/seccomp-strict.c (main): Change the type of addr variable
containg the invalid address from unsigned long to kernel_ulong_t,
print it using %#llx format.
Set the publicly shown copyright holder to "The strace developers.".
* COPYING: Update copyright information.
* configure.ac: Add AC_COPYRIGHT.
* strace.c (print_version): New function.
(init): Use it.
* tests/strace-V.test: Update expected output.
This extends the fault injection capability with :signal=SIG option
which injects a signal on entering each syscall from the specified set.
:signal and :error options are complementary, if they are both specified
the syscall will be fault injected as usual and the specified signal
will be delivered to the tracee.
* defs.h (struct fault_opts): Change the type of err field to int16_t,
add signo field.
(trace_syscall): Add a pointer argument.
* qualify.c: Include "nsig.h".
(parse_fault_token): Handle signal= option.
(qualify_fault): Update default fault_opts.
* strace.c (trace): Forward signal number from trace_syscall
to ptrace_restart(PTRACE_SYSCALL).
* syscall.c (inject_syscall_fault_entering): Add pointer argument,
save there the signal number specified by fault options. Do not inject
a syscall fault unless instructed by fault options.
(update_syscall_fault_exiting): Update the error code injection check.
(trace_syscall_entering): Add pointer argument, forward it to
inject_syscall_fault_entering.
(trace_syscall): Add pointer argument, forward it to
trace_syscall_entering.
Signed-off-by: Seraphime Kirkovski <kirkseraph@gmail.com>
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
* defs.h (widen_to_ulong): Remove.
(truncate_kulong_to_current_wordsize): New static inline function.
* io.c (do_preadv, do_pwritev): Use it instead of widen_to_ulong.
The demise of HAVE_STRUCT_TCB_EXT_ARG opens the way for a simpler
implementation.
* io.c (print_lld_from_low_high_val): Merge [SIZEOF_LONG > 4
&& SIZEOF_LONG == SIZEOF_LONG_LONG]
and [SIZEOF_KERNEL_LONG_T > SIZEOF_LONG] cases
into a single [SIZEOF_KERNEL_LONG_T > 4] case.
[SIZEOF_KERNEL_LONG_T == 4]: Use direct casts to long long
instead of zero_extend_signed_to_ull.
The demise of HAVE_STRUCT_TCB_EXT_ARG opens the way for a simpler
implementation.
* util.c (getllval): Merge [SIZEOF_LONG > 4
&& SIZEOF_LONG == SIZEOF_LONG_LONG]
and [SIZEOF_KERNEL_LONG_T > SIZEOF_LONG] cases
into a single [SIZEOF_KERNEL_LONG_T > 4] case.
Since there is no compat handler for x32/n32, kernel returns kernel's
long as a result of the call.
Interesting, does any callee out there actually care to have enough room
allocated for kernel long?
* prctl.c (SYS_FUNC(prctl)): Use printnum_kptr instead of printnum_ptr
to print the value stored at address pointed by addr2 argument
of PR_GET_TID_ADDRESS operation.
In cases where no compat for x32/n32 is implemented, pointer is of
kernel's long size, so it should be printed appropriately.
* defs.h [!current_klongsize] (printnum_addr_klong_int): New prototype.
(printnum_kptr): New macro.
* util.c [!current_klongsize] (printnum_addr_klong_int): New function.
Originally, printnum_long_int was used, but it prints NULL incorrectly.
* defs.h (DECL_PRINTNUM_ADDR): New macro.
(DECL_PRINTNUM_ADDR(int), DECL_PRINTNUM_ADDR(int64)): New prototypes.
[!current_wordsize] (printnum_addr_long_int): New prototype.
[!current_wordsize] (printnum_ptr): Use it.
[current_wordsize > 4] (printnum_ptr): Use printnum_addr_int64.
[current_wordsize == 4] (printnum_ptr) Use printnum_addr_int.
* util.c (DEF_PRINTNUM_ADDR): New macro.
(DEF_PRINTNUM_ADDR(int, unsigned int),
DEF_PRINTNUM_ADDR(int64, uint64_t)): New macro instantiations that
provide printnum_addr_int and printnum_addr_int64, accordingly.
[!current_wordsize] (printnum_addr_long_int): New function.
* xet_robust_list.c (sprintaddr): New function.
(main): Use it, update expected output.
* tests/ipc_sem.c (main): Update expected output.
Co-authored-by: Dmitry V. Levin <ldv@altlinux.org>
Do not silently truncate tracee addresses to current_wordsize.
After transition from long to kernel_ulong_t there should be no sign
extension issues with tracee addresses, and invalid addresses should
be printed properly.
* util.c (umoven): Check [SIZEOF_KERNEL_LONG_T > 4] instead
of [SIZEOF_LONG > 4], reject addresses that do not fit into
current_wordsize.
* tests/umoven-illptr.c: New file.
* tests/umoven-illptr.test: New test.
* tests/.gitignore: Add umoven-illptr.
* tests/Makefile.am (check_PROGRAMS): Likewise.
(DECODER_TESTS): Add umoven-illptr.test.
Do not silently truncate tracee addresses to current_wordsize.
After transition from long to kernel_ulong_t there should be no sign
extension issues with tracee addresses, and invalid addresses should
be printed properly.
* util.c (umovestr): Check [SIZEOF_KERNEL_LONG_T > 4] instead
of [SIZEOF_LONG > 4], reject addresses that do not fit into
current_wordsize.
* tests/umovestr-illptr.c: New file.
* tests/umovestr-illptr.test: New test.
* tests/.gitignore: Add umovestr-illptr.
* tests/Makefile.am (check_PROGRAMS): Likewise.
(DECODER_TESTS): Add umovestr-illptr.test.
After transition from long to kernel_ulong_t the implementation
could be made a bit simpler.
* lseek.c (SYS_FUNC(llseek)): Use direct cast to long long instead
of zero_extend_signed_to_ull.
After transition from long to kernel_ulong_t there are no mo
complications that required to keep a separate implementation
for x32 and mips n32.
* lseek.c (SYS_FUNC(lseek)): Merge two different implementations
into a single one.
After transition from long to kernel_ulong_t we no longer need
two versions of printaddr function.
* defs.h (printaddr): Remove.
(printaddr_klu): Rename to printaddr.
* util.c (printaddr_klu): Likewise.
* bjm.c (SYS_FUNC(init_module)): Replace printaddr_klu with printaddr.
* mem.c (do_mprotect): Likewise.
Check decoding of prctl operations that take the second syscall argument
as a pointer to an integer.
* tests/prctl-arg2-intptr.c: New file.
* tests/prctl-arg2-intptr.test: New test.
* tests/.gitignore: Add prctl-arg2-intptr.
* tests/Makefile.am (check_PROGRAMS): Likewise.
(DECODER_TESTS): Add prctl-arg2-intptr.test.
Remove temporary macros created for transition from long
to kernel_ulong_t.
Automatically replace PRI_kr[dux] with PRI_kl[dux] using
$ git grep -l 'PRI_kr[dux]' | xargs sed -ri 's/PRI_kr([dux])/PRI_kl\1/g'
* defs.h (PRI_krd, PRI_kru, PRI_krx): Remove. All users updated.
Remove temporary types created for transition from long
to kernel_ulong_t.
Automatically replace kernel_scno_t and kernel_ureg_t with
kernel_ulong_t using
$ git grep -El 'kernel_(scno|ureg)_t' |
xargs sed -ri 's/kernel_(scno|ureg)_t/kernel_ulong_t/g'
* kernel_types.h (kernel_scno_t, kernel_ureg_t): Remove.
All users updated.
* defs.h (struct tcb): Change u_rval type from long to kernel_long_t.
(PRI_krd, PRI_kru, PRI_krx): Redirect to PRI_kld, PRI_klu, and PRI_klx.
* kernel_types.h (kernel_scno_t, kernel_ureg_t): Redirect
to kernel_ulong_t.
Prepare for the change of kernel_ureg_t and kernel_scno_t from
unsigned long to kernel_ulong_t.
* defs.h (PRI_krd, PRI_kru, PRI_krx): New temporary macros.
All users updated.
When sizeof(kernel_ureg_t) > sizeof(long), the tracee address passed
to vm_read_mem could be silently truncated. Detect this situation
and return EIO when the tracee address does not fit into unsigned long.
* util.c (vm_read_mem): Save raddr argument into a temporary variable
truncated_raddr of type unsigned long. Set errno to EIO and return -1
when truncated_raddr does not equal to raddr.
Change type of sa_handler and sa_restorer fields from function pointers
to unsigned long. This saves the code from a lot of casts between
pointers and integers.
Rename __sa_handler to sa_handler__ as names starting with underscore
symbol are reserved and should not be used.
* signal.c (old_sigaction): Rename __sa_handler to sa_handler__,
all users updated. Change the type of sa_handler__ and sa_restorer
from pointer to unsigned long.
(old_sigaction32): Rename __sa_handler to sa_handler__, all users
updated.
(decode_old_sigaction): Remove now redundant casts from
old_sigaction32.sa_handler__ to old_sigaction.sa_handler__,
from old_sigaction32.sa_restorer to old_sigaction.sa_restorer,
and from old_sigaction.sa_handler__ to unsigned long. Print
old_sigaction.sa_restorer using printaddr.
* signal.c (new_sigaction): Rename __sa_handler to sa_handler__,
all users updated. Change the type of sa_handler__ and sa_restorer
from pointer to unsigned long.
(new_sigaction32): Rename __sa_handler to sa_handler__, all users
updated.
(decode_new_sigaction): Remove now redundant casts from
new_sigaction32.sa_handler__ to new_sigaction.sa_handler__,
from new_sigaction32.sa_restorer to new_sigaction.sa_restorer,
and from new_sigaction.sa_handler__ to unsigned long. Print
new_sigaction.sa_restorer using printaddr.
The flags argument is defined in kernel as unsigned long,
so change its decoding to avoid truncation.
* clone.c (SYS_FUNC(clone)): Change flags type from unsigned long
to kernel_ureg_t, print it using printflags64 instead of printflags.
After the change of tcb.u_arg type from long to kernel_ureg_t it is safe
to stop using printxval_long proxy and pass syscall arguments directly
to printxval64.
* defs.h (printxval_long): Remove.
* fcntl.c (print_fcntl): Replace printxval_long with printxval64.
* mtd.c (mtd_ioctl): Likewise.
* numa.c (SYS_FUNC(mbind)): Likewise.
* or1k_atomic.c [OR1K] (SYS_FUNC(or1k_atomic)): Likewise.
* sysmips.c [MIPS] (SYS_FUNC(sysmips)): Likewise.
* term.c (term_ioctl): Likewise.
After the change of tcb.u_arg type from long to kernel_ureg_t it is safe
to stop using printflags_long proxy and pass syscall arguments directly
to printflags64.
* defs.h (printflags_long): Remove.
* fcntl.c (print_fcntl): Replace printflags_long with printflags64.
* numa.c (SYS_FUNC(get_mempolicy)): Likewise.
* perf.c (SYS_FUNC(perf_event_open)): Likewise.
* sram_alloc.c [BFIN] (SYS_FUNC(sram_alloc)): Likewise.
* defs.h (upeek, upoke): Change offset argument type from long
to unsigned long. Change res argument type from long to kernel_ureg_t.
* upeek.c (upeek): Likewise.
* upoke.c (upoke.c): Likewise.
* syscall.c (print_pc) [ARCH_PC_PEEK_ADDR]: Change pc type from long
to kernel_ureg_t.
* linux/aarch64/arch_regs.c (arm_sp_ptr): Remove redundant cast.
* linux/metag/get_syscall_args.c (get_syscall_args): Likewise.
* linux/sh/get_syscall_result.c (get_syscall_result_regs): Likewise.
* linux/sh64/get_syscall_result.c (get_syscall_result_regs): Likewise.
* linux/powerpc/getregs_old.c (getregs_old): Remove redundant casts.
* linux/alpha/arch_getrval2.c (getrval2): Change r20 type from long
to unsigned long.
* linux/alpha/arch_regs.c (alpha_r0, alpha_a3): Change type from long
to unsigned long.
* linux/bfin/arch_regs.c (bfin_r0): Likewise.
* linux/crisv10/arch_regs.c (cris_r10): Likewise.
* linux/hppa/arch_regs.c (hppa_r28): Likewise.
* linux/ia64/arch_regs.c (IA64_PSR_IS): Likewise.
* linux/microblaze/arch_regs.c (microblaze_r3): Likewise.
* linux/sh/arch_regs.c (sh_r0): Likewise.
* linux/sh64/arch_regs.c (sh64_r9): Likewise.
* linux/xtensa/arch_regs.c (xtensa_a2): Likewise.
* linux/alpha/arch_sigreturn.c (arch_sigreturn): Change addr type
from long to unsigned long.
* linux/microblaze/arch_sigreturn.c (arch_sigreturn): Likewise.
* linux/alpha/get_scno.c (arch_get_scno): Update for the change
of signedness.
* linux/arc/get_syscall_args.c (get_syscall_args): Change arc_args type
from pointer to long to pointer to unsigned long.
* linux/arm/arch_regs.c (arm_sp_ptr): Change type from pointer to long
to pointer to unsigned long.
* linux/arm/arch_regs.h (arm_sp_ptr): Likewise.
* linux/i386/arch_regs.c (i386_esp_ptr): Likewise.
* linux/i386/arch_regs.h (i386_esp_ptr): Likewise.
* linux/m68k/arch_regs.c (m68k_usp_ptr): Likewise.
* linux/m68k/arch_regs.h (m68k_usp_ptr): Likewise.
* linux/ia64/get_syscall_args.c (get_syscall_args): Use umove
instead of umoven.
* linux/sh/arch_getrval2.c (getrval2): Change val type from long
to unsigned long.
* linux/arm/arch_sigreturn.c (arch_sigreturn): Change addr variable
type from long to unsigned long.
* linux/ia64/arch_sigreturn.c (arch_sigreturn): Likewise.
* linux/m68k/arch_sigreturn.c (arch_sigreturn): Likewise.
* linux/powerpc/arch_sigreturn.c (arch_sigreturn): Likewise.
* linux/s390/arch_sigreturn.c (arch_sigreturn): Likewise.
* linux/sparc/arch_sigreturn.c (arch_sigreturn): Likewise.
* linux/tile/arch_sigreturn.c (arch_sigreturn): Likewise.
* linux/crisv10/arch_sigreturn.c (arch_sigreturn): Change the type
of addr and regs variables from long to unsigned long.
* linux/mips/arch_sigreturn.c (arch_sigreturn): Change addr variable
type from long to kernel_ureg_t.
* linux/x86_64/arch_sigreturn.c (arch_sigreturn): Change the type
of offset and addr variables from long to kernel_ureg_t.
* kexec.c (SYS_FUNC(kexec_load)): Change n variable type
from unsigned long to kernel_ureg_t. Use printxval64 instead
of printxval_long. Use printflags64 instead of printflags_long.
* kexec.c (print_seg): Change the type of *seg and seg_buf variables
from unsigned long to kernel_ureg_t.
(print_kexec_segments): Change len argument type from unsigned long
to kernel_ureg_t. Change seg variable type from unsigned long
to kernel_ureg_t.
* defs.h (decode_netlink): Change len argument type from unsigned long
to kernel_ureg_t.
* netlink.c (fetch_nlmsghdr): Likewise.
(decode_nlmsghdr_with_payload): Likewise. Change nlmsg_len variable
type from unsigned long to unsigned int.
(decode_netlink): Change the type of len argument and next_len variable
from unsigned long to kernel_ureg_t. Change nlmsg_len variable type
from unsigned long to unsigned int.
* ipc_msg.c (tprint_msgsnd): Change count argument type
from unsigned long to kernel_ureg_t. Change flags argument type
from unsigned long to unsigned int.
(tprint_msgrcv): Change the type of count and msgtyp arguments
from unsigned long to kernel_ureg_t.
(fetch_msgrcv_args): Change *pair type from unsigned long
to kernel_ureg_t.
(SYS_FUNC(msgrcv)): Change pair type from unsigned long
to kernel_ureg_t.
* print_msgbuf.c (tprint_msgbuf): Change count argument type
from unsigned long to kernel_ureg_t.
* defs.h (dumpiov_upto): Change data_size argument type
from unsigned long to kernel_ureg_t.
* util.c (dumpiov_upto): Change the type of data_size argument
and iov_len variable from unsigned long to kernel_ureg_t.
As the kernel does not accept control messages longer than the value
of int sysctl_optmem_max kernel variable exported as
/proc/sys/net/core/optmem_max, we do not print them either.
Since unsigned int has enough room to contain any valid control message
size, change all internal cmsghdr size types from size_t
to unsigned int.
* msghdr.c (print_scm_rights, print_scm_creds, print_scm_security,
print_cmsg_ip_pktinfo, print_cmsg_uint, print_cmsg_uint8_t,
print_cmsg_ip_opts, print_cmsg_ip_recverr, print_cmsg_ip_origdstaddr,
cmsg_printer, print_cmsg_type_data): Change data_len argument type
from size_t to unsigned int.
(cmsg_socket_printers): Change min_len type from size_t to unsigned int.
(decode_msg_control): Change the type of in_control_len argument,
cmsg_len, and len variables from size_t to kernel_ureg_t.
Change the type of cmsg_size, control_len, and buf_len variables
from size_t to unsigned int.
* defs.h (dumpiov_in_msghdr): Change data_size argument type
from unsigned long to kernel_ureg_t.
* msghdr.h (print_struct_msghdr): Likewise.
* msghdr.c (print_struct_msghdr): Likewise. Cast msghdr.msg_iovlen
and msghdr.msg_controllen to kernel_ureg_t instead of unsigned long.
(decode_msghdr, dumpiov_in_msghdr): Change data_size argument type
from unsigned long to kernel_ureg_t.
(SYS_FUNC(sendmsg)): Pass -1 instead of -1UL as data_size argument
to decode_msghdr.
* mmsghdr.c (print_struct_mmsghdr): Pass (kernel_ureg_t) -1 instead
of -1UL as data_size argument to print_struct_msghdr.
(dumpiov_in_mmsghdr): Cast msghdr.msg_iovlen to kernel_ureg_t instead
of unsigned long.
* syscall.c (dumpio): Pass -1 instead of -1UL as data_size argument
to dumpiov_in_msghdr
* defs.h (tprint_iov_upto): Change the type of len and data_size
arguments from unsigned long to kernel_ureg_t.
(tprint_iov): Change the type of len argument from unsigned long
to kernel_ureg_t.
* io.c (print_iovec_config): Change data_size type from unsigned long
to kernel_ureg_t.
(print_iovec): Change the type of *iov, iov_buf, and len variables
from unsigned long to kernel_ureg_t.
(tprint_iov_upto): Change the type of len and data_size arguments, and
the type of iov variable from unsigned long to kernel_ureg_t.
* mem.c (print_mmap): Change the type of len, prot, and flags
from unsigned long to kernel_ureg_t. Use printflags64 instead
of printflags_long. Use printxval64 instead of printxval_long.
(SYS_FUNC(old_mmap)): Check for current_klongsize instead
of [AARCH64 || X86_64] to take X32 into account.
(SYS_FUNC(mremap)): Use printflags64 instead of printflags_long.
(SYS_FUNC(remap_file_pages)): Change the type of size, prot, pgoff, and
flags from unsigned long to kernel_ureg_t. Use printflags64 instead
of printflags_long. Use printxval64 instead of printxval_long.
[POWERPC] (SYS_FUNC(subpage_prot)): Change the type of len and nmemb
from unsigned long to kernel_ureg_t.
* numa.c (print_node, print_addr): Add kernel_ureg_t support.
(print_nodemask): Change type of maxnodes argument from unsigned long
to kernel_ureg_t.
(SYS_FUNC(move_pages)): Change npages type from unsigned long
to kernel_ureg_t. Change buf type from long to kernel_ureg_t.
* defs.h (print_sigset_addr_len): Change len argument type from long
to kernel_ureg_t.
* signal.c (print_sigset_addr_len_limit, print_sigset_addr_len): Likewise.
* process_vm.c (SYS_FUNC(process_vm_readv),
SYS_FUNC(process_vm_writev)): Change the type of local_iovcnt,
remote_iovcnt, and flags from unsigned long to kernel_ureg_t.
* xattr.c (print_xattr_val): Change type of insize and size arguments
from unsigned long to kernel_ureg_t.
(print_xattr_list): Change type of size argument from unsigned long
to kernel_ureg_t.
Do not treat len == -1 in any special way. All users who need the
string to be handled as a NUL-terminated string set QUOTE_0_TERMINATED
bit in user_style flags already.
* util.c (printstr_ex): Remove (len == -1) check.
* tests/printstr.c: New file.
* tests/printstr.test: New test.
* tests/.gitignore: Add printstr.
* tests/Makefile.am (check_PROGRAMS): Likewise.
(DECODER_TESTS): Add printstr.test.
The second argument of sethostname syscall is not an unsigned long but
unsigned int. The kernel does not look at the string argument when
the length argument is too long.
* hostname.c [HAVE_LINUX_UTSNAME_H]: Include <linux/utsname.h>.
[!__NEW_UTS_LEN] (__NEW_UTS_LEN): Define.
(SYS_FUNC(sethostname)): Treat the second argument as unsigned int.
Print the first argument as a pointer when the second argument exceeds
__NEW_UTS_LEN.
* tests/sethostname.c [HAVE_LINUX_UTSNAME_H]: Include <linux/utsname.h>.
[!__NEW_UTS_LEN] (__NEW_UTS_LEN): Define.
(main): Use it. Check that the second argument of sethostname
is handled as unsigned int. Check that the first argument is printed
as a pointer when the second argument exceeds __NEW_UTS_LEN.
As about half of all printstr uses cases are invocations with the last
argument set to -1, create a dedicated function for this use case.
* defs.h (printstr): Rename to printstrn. All callers updated.
(printstr): New static inline function. All callers updated.
* process.c (print_user_offset_addr): Change address argument type
from unsigned long to kernel_ureg_t.
(SYS_FUNC(ptrace)): Change addr variable type from unsigned long
to kernel_ureg_t.
* defs.h [ALPHA] (printrusage32): Change address argument type
from long to kernel_ureg_t.
* printrusage.c (printrusage): Likewise.
[ALPHA] (printrusage32): Likewise.
* wait.c (printwaitn): Change the type of tracee address argument
of print_rusage from long to kernel_ureg_t.
* defs.h [ALPHA] (sprint_timeval32, print_timeval32,
print_timeval32_pair, print_itimerval32): Change address argument type
from long to kernel_ureg_t.
* desc.c (decode_select): Change the type of tracee address argument
of print_tv_ts and sprint_tv_ts from long to kernel_ureg_t.
* print_timespec.c (print_timespec, sprint_timespec,
print_timespec_utime_pair, print_itimerspec): Change address argument
type from long to kernel_ureg_t.
* print_timeval.c (print_timeval, print_timeval_pair, sprint_timeval,
print_itimerval, print_timeval32, print_timeval32_pair,
print_itimerval32, sprint_timeval32): Likewise.
* rtc.c (decode_rtc_time, decode_rtc_wkalrm, decode_rtc_pll_info):
Change addr type from long to kernel_ureg_t.
(rtc_ioctl): Change arg type from long to kernel_ureg_t.
* mtd.c (decode_erase_info_user, decode_erase_info_user64,
decode_mtd_oob_buf, decode_mtd_oob_buf64, decode_otp_info,
decode_otp_select, decode_mtd_write_req, decode_mtd_info_user,
decode_nand_oobinfo, decode_nand_ecclayout_user, decode_mtd_ecc_stats):
Change addr type from long to kernel_ureg_t.
(mtd_ioctl): Change arg type from long to kernel_ureg_t.
* defs.h (DECL_IOCTL): Change arg type from long to kernel_ureg_t.
* dm.c (dm_known_ioctl, dm_ioctl): Likewise.
* file_ioctl.c (file_ioctl): Likewise.
* fs_x_ioctl.c (fs_x_ioctl): Likewise.
* ioctl.c (ioctl_decode): Likewise.
* loop.c (decode_loop_info, decode_loop_info64): Change addr type
from long to kernel_ureg_t.
(loop_ioctl): Change arg type from long to kernel_ureg_t.
* ptp.c (ptp_ioctl): Likewise.
* scsi.c (print_sg_io_v3_req, print_sg_io_v3_res, print_sg_io_v4_req,
print_sg_io_v4_res, scsi_ioctl): Likewise.
* sock.c (print_ifreq, sock_ioctl): Likewise.
(decode_ifconf): Change addr type from long to kernel_ureg_t.
* term.c (decode_termios, decode_termio, decode_winsize, decode_ttysize,
decode_modem_flags): Likewise.
(term_ioctl): Change arg type from long to kernel_ureg_t.
* ubi.c (ubi_ioctl): Likewise.
* userfaultfd.c (uffdio_ioctl): Likewise.
* defs.h [HAVE_STRUCT_USER_DESC] (print_user_desc): Change address
argument type from long to kernel_ureg_t.
* ldt.c [HAVE_STRUCT_USER_DESC] (print_user_desc): Likewise.
* bpf.c (bpf_map_create, bpf_map_update_elem, bpf_map_delete_elem,
bpf_map_io, bpf_prog_load, SYS_FUNC(bpf)): Change address argument
type from long to kernel_ureg_t.
* defs.h (print_sigset_addr_len) Change address argument type from long
to kernel_ureg_t.
* signal.c (print_sigset_addr_len_limit, print_sigset_addr_len,
decode_old_sigaction, decode_new_sigaction): Likewise.
* defs.h (DECL_PRINTNUM, DECL_PRINTPAIR, printnum_long_int): Change
address argument type from long to kernel_ureg_t.
* util.c (DEF_PRINTNUM, DEF_PRINTPAIR, printnum_long_int): Likewise.
* defs.h (umovestr, umoven, umoven_or_printaddr,
umoven_or_printaddr_ignore_syserror): Change tracee address argument
type from long to kernel_ureg_t.
(print_array): Change the type of tracee address argument of umoven_func
from long to kernel_ureg_t.
* util.c (vm_read_mem, umovestr, umoven, umoven_or_printaddr,
umoven_or_printaddr_ignore_syserror): Change tracee address argument
type from long to kernel_ureg_t.
(print_array): Change the type of tracee address argument of umoven_func
from long to kernel_ureg_t.
* desc.c (umove_kulong_array_or_printaddr): Change tracee address
argument type from long to kernel_ureg_t.
* mmsghdr.c (fetch_struct_mmsghdr_or_printaddr): Likewise.
* printsiginfo.c (printsigval): Cast siginfo.si_ptr pointer
to kernel_ureg_t instead of unsigned long.
(print_si_info): Cast siginfo.si_addr and siginfo.si_call_addr pointers
to kernel_ureg_t instead of unsigned long.
* msghdr.c (print_struct_msghdr): Cast msghdr.msg_name pointer
to kernel_ureg_t instead of long.
(dumpiov_in_msghdr): Cast msghdr.msg_iov pointer to kernel_ureg_t
instead of long.
* process.c (SYS_FUNC(ptrace)): Remove syserror checks as both
print_siginfo_array and print_seccomp_fprog do the right thing
by using umoven_or_printaddr.
* mem.c (print_mmap, SYS_FUNC(remap_file_pages),
SYS_FUNC(subpage_prot)): Change the type of variables containing tracee
addresses from unsigned long to kernel_ureg_t.
* pathtrace.c (upathmatch): Change address argument type
from unsigned long to kernel_ureg_t.
(pathtrace_match): Change the type of all tracee address variables
from unsigned long to kernel_ureg_t.
* dm.c (dm_decode_dm_target_spec, dm_decode_dm_target_deps,
dm_decode_dm_name_list, dm_decode_dm_target_versions,
dm_decode_dm_target_msg, dm_decode_string): Change address argument type
from unsigned long to kernel_ureg_t.
* defs.h (decode_netlink) Change address argument type
from unsigned long to kernel_ureg_t.
* netlink.c (fetch_nlmsghdr, decode_nlmsghdr_with_payload,
decode_netlink): Likewise.
* defs.h (print_seccomp_filter, print_seccomp_fprog): Change
address argument type from unsigned long to kernel_ureg_t.
* seccomp.c (print_seccomp_filter, print_seccomp_fprog,
decode_seccomp_set_mode_strict): Likewise.
* defs.h (tprint_iov, tprint_iov_upto): Change address argument type
from unsigned long to kernel_ureg_t.
* io.c (tprint_iov, tprint_iov_upto): Likewise.
* defs.h (print_array): Change start_addr type from unsigned long
to kernel_ureg_t.
* util.c (print_array): Likewise. Change the type of all tracee
address variables from unsigned long to kernel_ureg_t.
In all places where LONG_LONG is used the destination already has type
unsigned long long, so it is quite natural to change LONG_LONG as well.
* defs.h (LONG_LONG): Rename to ULONG_LONG, change return type
from long long to unsigned long long. All callers updated.
* mem.c (SYS_FUNC(old_mmap)): Remove redundant cast
from kernel_ureg_t to unsigned long.
[S390] (SYS_FUNC(old_mmap_pgoff)): Remove redundant cast
from unsigned int to unsigned long.
* desc.c (umove_kulong_array_or_printaddr): Remove redundant cast
from uint32_t to kernel_ulong_t.
(SYS_FUNC(pselect6)): Remove redundant casts from kernel_ulong_t
to unsigned long.
* print_sigevent.c (print_sigevent): Remove redundant cast from
struct_sigevent.sigev_value.sival_ptr to unsigned long as the former
is not a pointer but an unsigned integer.
* tests/xattr.c (DEFAULT_STRLEN): New macro.
(main): Use it to limit the length passed to print_quoted_memory.
Print ellipsis when the list returned by listxattr is too long.
This fixes the display when using `-s` to limit the string size
of the values displayed by {get,set}xattr.
* xattr.c (print_xattr_val): Remove static buffer. Use printstr_ex
instead of print_quoted_string.
Reported-by: Марк Коренберг <socketpair@gmail.com>
Signed-off-by: Gabriel Laskar <gabriel@lse.epita.fr>
* btrfs.c (btrfs_ioctl): Cast btrfs_ioctl_send_args.clone_sources
and btrfs_ioctl_vol_args_v2.qgroup_inherit pointers to kernel_ureg_t
instead of unsigned long.
* ioctl.c (compare): Cast pointer to uintptr_t before assigning
it to an integer.
(ioctl_lookup): Cast integer to uintptr_t before passing it
as a pointer.
* travis-build.sh: Use DISTCHECK_CONFIGURE_FLAGS instead
of CHECK_CONFIGURE_FLAGS and ENABLE_GCC_WERROR.
[$CC == gcc]: Add --enable-gcc-Werror to DISTCHECK_CONFIGURE_FLAGS
instead of ENABLE_GCC_WERROR.
[$CHECK == coverage]: Add --enable-code-coverage
to DISTCHECK_CONFIGURE_FLAGS instead of CHECK_CONFIGURE_FLAGS.
Remove the workaround for broken ptrace prototypes
that seems to be unneeded for quite a long time.
* defs.h [NEED_PTRACE_PROTOTYPE_WORKAROUND]: Remove.
Provide scno.h with fallback definitions of all __NR_* constants known
to strace for the personbality being tested.
Some tests need just __NR_* constants without any kernel support.
Such tests can include "scno.h" after <asm/unistd.h> to get definitions
of these constants when system headers are too old.
* tests/generate_scno.c: New file.
* tests/.gitignore: Add generate_scno.
* tests/Makefile.am (check_PROGRAMS): Likewise.
(scno.h): New rule.
(BUILT_SOURCES, CLEANFILES): Add scno.h.
Automatic argument truncation makes use of this function redundant.
* kexec.c (SYS_FUNC(kexec_load)): Do not invoke widen_to_ulong.
* process_vm.c (SYS_FUNC(process_vm_readv),
SYS_FUNC(process_vm_writev)): Likewise.
x32 syscalls starting with number 512 are compat syscalls that operate
on compat types. Set COMPAT_SYSCALL_TYPES flag to these syscall entries
and make get_syscall_args truncate arguments of syscalls that have this
flag set, to avoid unpleasant alternative of adding multiple
widen_to_ulong invocations to individual syscall parsers.
preadv, pwritev, preadv2, and pwritev2 are exceptions: while these
syscalls operate on compat types, the offset argument has a regular
64-bit type, so the automatic argument truncation cannot be applied.
* defs.h (COMPAT_SYSCALL_TYPES): New macro.
* syscall.c (CST): New macro, defined to COMPAT_SYSCALL_TYPES.
* linux/ia64/syscallent.h: Do not include "../i386/syscallent.h"
unless [CST > 0].
* linux/x32/syscallent.h [512..533, 536..545]: Add CST flag.
* linux/x86_64/get_syscall_args.c (get_syscall_args): Truncate arguments
of compat syscalls to 32-bit values.
* tests/ksysent.c (CST): New macro, defined to 0.
* tests/nsyscalls.c (CST): Likewise.
As there are no compat versions of mprotect or pkey_mprotect syscalls
in x32 personality, "addr", "len", and "prot" syscall arguments have
kernel_ulong_t type and should be printed accordingly.
* mem.c (do_mprotect): Retrieve 1st, 2nd and 3rd arguments using
getarg_klu, print them using printaddr_klu, PRI_klu format,
and printflags64, correspondingly.
In x32 personality these syscalls use compat type for the 3rd argument,
therefore it has to be truncated there.
* io.c (do_preadv, do_pwritev): Truncate tcp->u_arg[2] using
widen_to_ulong.
On x32 p(read|write)v(|2) syscalls are implemented via compat, which
truncates len argument to unsigned int.
* tests/preadv2-pwritev2.c (main): Print vlen argument as unsigned long.
There is no compat (and no 32-bit version, ever) for x32 ABI, so
kernel's long type should be used for arguments' values.
The only remaining thing is a pointer to cmdline string, but it is
currently broken on x32 anyway.
* kexec.c (SYS_FUNC(kexec_file_load)): Print the cmdline_len argument
as kernel long, retrieve it via getarg_klu; use printflags64 for
printing the flags argument, retrieve it via getarg_klu.
current_klongsize checks are more universal, therefore they are usually
shorter and easier to comprehend.
* desc.c (umove_kulong_array_or_printaddr): Check for current_klongsize
instead of current_personality and current_wordsize.
* io.c (print_lld_from_low_high_val): Likewise.
* lseek.c (SYS_FUNC(lseek)): Likewise.
* resource.c (decode_rlimit): Likewise.
* syscall.c (is_negated_errno): Likewise.
* util.c (getllval, getarg_klu): Likewise.
* defs.h (printaddr_ull): Rename to printaddr_klu, change argument
type from unsigned long long to kernel_ulong_t. All callers updated.
(getarg_ull): Rename to getarg_klu, change return value type
from unsigned long long to kernel_ulong_t. All callers updated.
(PRI_kl, PRI_kld, PRI_klu, PRI_klx): New macros.
* bjm.c (SYS_FUNC(init_module)): Print kernel_ulong_t type using
PRI_klu format.
* desc.c (SYS_FUNC(pselect6)): Likewise.
* fadvise.c (SYS_FUNC(fadvise64)): Likewise.
* lookup_dcookie.c (SYS_FUNC(lookup_dcookie)): Likewise.
* mq.c (SYS_FUNC(mq_timedsend), SYS_FUNC(mq_timedreceive)): Likewise.
* kcmp.c (SYS_FUNC(kcmp)): Print kernel_ulong_t type using
PRI_klx format.
* keyctl.c (SYS_FUNC(keyctl)): Likewise.
* pkeys.c (SYS_FUNC(pkey_alloc)): Likewise.
* prctl.c (print_prctl_args, SYS_FUNC(prctl), SYS_FUNC(arch_prctl)):
Print kernel_ulong_t type using PRI_kld, PRI_klu, or PRI_klx format.
* util.c (printaddr_ull): Rename to printaddr_klu, change argument
type from unsigned long long to kernel_ulong_t, print it using
PRI_klx format.
(getarg_ull): Rename to getarg_klu, change return value type
from unsigned long long to kernel_ulong_t, print it using
PRI_klx format.
* block.c (struct_blkpg_partition): Change type of start and length
fields from long long to int64_t.
(print_blkpg_req): Print them using PRId64 format.
Provide SIZEOF_KERNEL_LONG_T macro in addition to SIZEOF_LONG and
SIZEOF_LONG_LONG for checking of kernel_long_t and kernel_ulong_t
types at the time of preprocessing.
* configure.ac (AC_CHECK_SIZEOF): Add kernel_long_t.
This function has never been used in the code, and in perspective
the whole ext_arg thing will go away.
* defs.h (getarg_ll): Remove prototype.
* util.c (getarg_ll): Remove.
This is the first step in the direction of revisiting current practice
of indiscriminate use of signed types for syscall arguments and memory
addresses.
* kernel_types.h (kernel_ureg_t): New type, typedef to unsigned long.
* defs.h (struct tcb): Change type of u_arg from long to kernel_ureg_t.
[HAVE_STRUCT_TCB_EXT_ARG]: Change type of ext_arg from long long
to unsigned long long.
* desc.c (decode_select): Change type of syscall arguments from long
to kernel_ureg_t.
(SYS_FUNC(oldselect)): Change type of select_args from long
to kernel_ureg_t.
* io.c (print_lld_from_low_high_val): Remove no longer needed cast
of syscall arguments to unsigned long.
* lseek.c (SYS_FUNC(lseek)): Cast syscall argument from unsigned long
to long.
* mem.c (print_mmap): Change type of syscall arguments from long
to kernel_ureg_t.
(SYS_FUNC(old_mmap), SYS_FUNC(old_mmap_pgoff)): Change type of u_arg
from long to kernel_ureg_t.
(SYS_FUNC(mmap), SYS_FUNC(mmap_pgoff), SYS_FUNC(mmap_pgoff)): Remove
no longer needed cast of syscall arguments to unsigned long.
* pathtrace.c (pathtrace_match): Change type of args and select_args
from long to kernel_ureg_t.
* util.c (getarg_ull): Remove no longer needed casts of syscall
arguments to unsigned types.
Use an unsigned type for syscall numbers as they are not intended
for signed arithmetics. Introduce kernel_scno_t as a typedef
to unsigned long, that could be changed later to kernel_ulong_t.
* kernel_types.h (kernel_scno_t): New type, typedef to unsigned long.
* defs.h (struct tcb): Change type of scno field from long
to kernel_scno_t.
(syscall_name): Change argument type from long to kernel_scno_t.
(scno_in_range, scno_is_valid): Change argument type from unsigned long
to kernel_scno_t.
* linux/aarch64/set_scno.c (arch_set_scno): Change scno argument type
from long to kernel_scno_t.
* linux/alpha/set_scno.c (arch_set_scno): Likewise.
* linux/arc/set_scno.c (arch_set_scno): Likewise.
* linux/arm/set_scno.c (arch_set_scno): Likewise.
* linux/avr32/set_scno.c (arch_set_scno): Likewise.
* linux/bfin/set_scno.c (arch_set_scno): Likewise.
* linux/crisv10/set_scno.c (arch_set_scno): Likewise.
* linux/hppa/set_scno.c (arch_set_scno): Likewise.
* linux/i386/set_scno.c (arch_set_scno): Likewise.
* linux/ia64/set_scno.c (arch_set_scno): Likewise.
* linux/m68k/set_scno.c (arch_set_scno): Likewise.
* linux/metag/set_scno.c (arch_set_scno): Likewise.
* linux/microblaze/set_scno.c (arch_set_scno): Likewise.
* linux/mips/set_scno.c (arch_set_scno): Likewise.
* linux/nios2/set_scno.c (arch_set_scno): Likewise.
* linux/or1k/set_scno.c (arch_set_scno): Likewise.
* linux/powerpc/set_scno.c (arch_set_scno): Likewise.
* linux/riscv/set_scno.c (arch_set_scno): Likewise.
* linux/s390/set_scno.c (arch_set_scno): Likewise.
* linux/sh/set_scno.c (arch_set_scno): Likewise.
* linux/sh64/set_scno.c (arch_set_scno): Likewise.
* linux/sparc/set_scno.c (arch_set_scno): Likewise.
* linux/tile/set_scno.c (arch_set_scno): Likewise.
* linux/x86_64/set_scno.c (arch_set_scno): Likewise.
* linux/xtensa/set_scno.c (arch_set_scno): Likewise.
* linux/aarch64/get_scno.c (arch_get_scno): Change scno variable type
from long to kernel_scno_t.
* linux/alpha/get_scno.c (arch_get_scno): Likewise.
* linux/arm/get_scno.c (arch_get_scno): Likewise.
* linux/sh/get_scno.c (arch_get_scno): Likewise.
* linux/x86_64/get_scno.c (arch_get_scno): Likewise.
* syscall.c (arch_set_scno): Likewise.
(shuffle_scno): Change return type from long to kernel_scno_t.
(syscall_name): Change argument type from long to kernel_scno_t.
* .travis.yml (after_success): Remove.
(matrix): Create an entry for the regular x86_64 target.
* travis-success.sh: Merge into ...
* travis-build.sh: ... this file. Disable optimization
for coverage builds.
* travis-install.sh: Rename COVERAGE=true to CHECK=coverage.
* xlat/pr_dumpable.in: New file.
* prctl.c: Include "xlat/pr_dumpable.h".
(SYS_FUNC(prctl)): Decode returned value of PR_GET_DUMPABLE operation
and the second syscall argument of PR_SET_DUMPABLE operation
as SUID_DUMP_* constants.
As new constants have been added gradually over kernel releases, define
all constant values to make newer constants decoded properly when strace
is built with older kernel headers.
* xlat/netlink_protocols.in: Add default values for constants.
* net.c [!NETLINK_SOCK_DIAG && NETLINK_INET_DIAG]: Remove.
* socketutils.c: Likewise.
* tests/sched_xetaffinity.c (errstr): New variable.
(getaffinity, setaffinity): Initialize it.
(main): Use it. Allow getaffinity return value to be less than
its cpusetsize argument.
Workaround a bug introduced in GNU grep 2.27, for details see
https://lists.gnu.org/archive/html/bug-grep/2016-12/msg00011.html
* tests/init.sh (grep_pid_status): New function.
* tests/detach-running.test: Use it instead of grep to search
in /proc/$pid/status.
* tests/detach-sleeping.test: Likewise.
* tests/detach-stopped.test: Likewise.
* strace.c (init) [USE_LIBUNWIND]: Call unwind_tcb_init after
unwind_init for all allocated tcb structures if stack trace is enabled.
* unwind.c (unwind_tcb_init): Skip if tcb is already initialized.
Co-authored-by: Dmitry V. Levin <ldv@altlinux.org>
* tests/fault_injection.sh: New file.
* tests/Makefile.am (EXTRA_DIST): Add it.
* tests/fault_injection-exit_group.test: Use it.
* tests/fault_injection.test: Likewise.
* tests/init.sh (kernel_version_code,
require_min_kernel_version_or_skip): New functions.
* tests/threads-execve.test: Use require_min_kernel_version_or_skip.
Commit dc427d50d9 enabled a number of extra
warning flags through configure.ac. The configure script will determine
dynamically if CC supports these flags before adding them to WARN_CFLAGS.
ioctlsort is compiled with CC_FOR_BUILD, rather than CC. Nevertheless,
the flags passed to this compilation also include WARN_CFLAGS (through
AM_CFLAGS). This is incorrect: in a cross-compilation use case, CC
and CC_FOR_BUILD are not the same. The former is the cross-compiler,
the latter is the host compiler. Both may be of different versions
and support different warning flags.
In particular, this posed a problem when cross-compiling with a host
compiler gcc 4.1, which does not support all the new flags:
/usr/bin/gcc -DHAVE_CONFIG_H -I./linux/arm -I./linux/arm -I./linux
-I./linux -I. -I. -I/host-sysroot/usr/include -Wall -Wempty-body
-Wformat-security -Wignored-qualifiers -Winit-self -Wlogical-op
-Wmissing-parameter-type -Wnested-externs -Wold-style-declaration
-Wold-style-definition -Wsign-compare -Wtype-limits -Wwrite-strings -O2
-I/host-sysroot/usr/include -DIOCTLSORT_INC=\"ioctls_all0.h\" -c -o
ioctlsort0.o ./ioctlsort.c
cc1: error: unrecognized command line option "-Wempty-body"
cc1: error: unrecognized command line option "-Wignored-qualifiers"
cc1: error: unrecognized command line option "-Wlogical-op"
cc1: error: unrecognized command line option "-Wmissing-parameter-type"
cc1: error: unrecognized command line option "-Wold-style-declaration"
cc1: error: unrecognized command line option "-Wtype-limits"
make[2]: *** [ioctlsort0.o] Error 1
* Makefile.am (AM_CFLAGS_FOR_BUILD, AM_CPPFLAGS_FOR_BUILD): New
variables.
(ioctlsort_CPPFLAGS): Use AM_CPPFLAGS_FOR_BUILD instead of AM_CPPFLAGS.
(ioctlsort_CFLAGS): Use AM_CFLAGS_FOR_BUILD instead of AM_CFLAGS.
(ioctlsort_LDFLAGS): Use AM_LDFLAGS_FOR_BUILD instead of AM_LDFLAGS.
* m4/ax_prog_cc_for_build.m4 (AX_PROG_CC_FOR_BUILD): Redirect
WARN_CFLAGS and gl_unknown_warnings_are_errors, call st_WARN_CFLAGS,
substitute WARN_CFLAGS_FOR_BUILD.
Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
* m4/warnings.m4 (gl_COMPILER_OPTION_IF): Change gl_Flags to use
_AC_LANG_PREFIX[]FLAGS in unquoted form. Change gl_Warn to use gl_Flags
instead of _AC_LANG_ABBREV. Change notification message to mention
the variable name of the compiler.
(gl_UNKNOWN_WARNINGS_ARE_ERRORS): Do nothing if
gl_unknown_warnings_are_errors is set.
(gl_WARN_ADD): Change the use of gl_UNKNOWN_WARNINGS_ARE_ERRORS from
AC_REQUIRE to a regular call.
Some of these variables cannot be redirected using traditional
pushdef/popdef mechanism because of complex constructions like
[ac_cv_[]_AC_LANG_ABBREV[]_compiler_gnu] employed by autoconf macros.
Use st_SAVE_VAR/st_RESTORE_VAR to workaround this issue.
* m4/ax_prog_cc_for_build.m4: Redirect ac_cv_prog_cc_stdc,
ac_cv_prog_cc_c11, ac_cv_prog_cc_c99, ac_cv_prog_cc_c89,
am_cv_prog_cc_c_o, and am_cv_CC_dependencies_compiler_type using
pushdef/popdef.
Redirect ac_c_decl_warn_flag, ac_c_preproc_warn_flag, ac_c_werror_flag,
ac_compile, ac_compiler_gnu, ac_cpp, ac_cv_c_compiler_gnu,
ac_cv_c_decl_report, and ac_link using st_SAVE_VAR/st_RESTORE_VAR.
This is going to be needed for the upcoming AX_PROG_CC_FOR_BUILD
change.
* configure.ac: Call st_WARN_CFLAGS before AX_PROG_CC_FOR_BUILD.
Move all gl_WARN_ADD calls ...
* m4/st_warn_cflags.m4: ... here.
Due to probabilistic nature of the test, try it several times.
* tests/threads-execve.c (NUMBER_OF_ITERATIONS): Change to 1.
* tests/threads-execve.test: Iterate up to 10 times.
If strace -ttt is running too long, it might happen that time stamps
before and after its invocation differ for more than a second.
Adjust expected output to handle this rare but possible case.
* tests/strace-ttt.test: Allow any time stamp between start and finish
of strace invocation.
If strace -tt is running too long, it might happen that time stamps
before and after its invocation differ for more than a second.
Adjust expected output to handle this rare but possible case.
* tests/strace-tt.test: Allow any time stamp between start and finish
of strace invocation.
Introduce a synchronization mechanism between the tracee and its peer.
* tests/redirect.test: Change timeout file to $OUT. Let the tracee
remove $LOG, wait for $LOG removal in its peer.
Introduce a synchronization mechanism between attach-f-p and strace.
* tests/attach-f-p-cmd.c: New file.
* tests/Makefile.am (check_PROGRAMS): Add attach-f-p-cmd.
* tests/attach-f-p.c: Include <sys/stat.h>.
(main): Write to stdout instead of descriptor 3. Wait for the peer
writing to stdout.
* tests/attach-f-p.test: Assume that ./attach-f-p-cmd works.
Use $OUT for expected output. Use attach-f-p-cmd for unlocking $OUT.
Introduce a synchronization mechanism between attach-p-cmd-p
and attach-p-cmd-cmd processes.
* tests/attach-p-cmd-cmd.c: Include <errno.h> and <sys/stat.h>.
(main): Wait for the lock directory creation by attach-p-cmd-p.
* tests/attach-p-cmd-p.c: Include <errno.h> and <sys/stat.h>.
(main): Create a lock directory and wait for its removal
by attach-p-cmd-cmd.
* tests/attach-p-cmd.test: Assume that test programs work.
* defs.h (signal_set): New variable prototypes.
(qualify_signals): New function prototypes.
(QUAL_SIGNAL): Change to a value greater than 0xff.
(QUAL_FAULT): Change to a lower value.
* qualify.c (signal_set): New variable.
(sigstr_to_uint, qualify_signals): New functions.
* syscall.c (qual_signal): Remove.
(qual_options): Replace qual_signal with NULL.
(qualify): Use qualify_signals.
* strace.c (print_signalled, print_stopped): Use is_number_in_set
with signal_set argument.
As a side effect, this also fixes support of negated sets
of descriptors.
* defs.h (struct number_set): New forward declaration.
(read_set, write_set): New variable prototypes.
(is_number_in_set, qualify_read, qualify_write): New function
prototypes.
(QUAL_READ, QUAL_WRITE): Change to values greater than 0xff.
* qualify.c: New file.
* Makefile.am (strace_SOURCES): Add it.
* syscall.c (qual_desc): Remove.
(qual_options): Replace qual_desc with NULL.
(qualify): Use qualify_read and qualify_write.
(dumpio): Use is_number_in_set.
* tests/options-syntax.test: Check invalid sets of descriptors.
* tests/readv.test: Check dumping of negated sets of descriptors.
valgrind is not as transparent as strace wrt standard descriptors,
so if strace is wrapped with valgrind, the test would check valgrind
transparency properties and fail imminently.
* tests/redirect-fds.test: Skip the test when $STRACE shows a presence
of valgrind.
As valgrind is not quite happy about the way how get_cpuset_size uses
sched_getaffinity, add a suppression file for this case.
* valgrind.supp: New file.
* Makefile.am (EXTRA_DIST): Add it.
Change the test to allow wrappers like STRACE="valgrind ./strace".
* tests/options-syntax.test: Strip all words but last from $STRACE
variable in all contexts where the name of strace executable is expected
in its output.
libc is perfectly capable of allocating memory for its buffers,
so let it do its job.
* strace.c (init): Do not allocate memory for setvbuf invocation.
As this is the type used in kernel.
* uid.c (SYS_FUNC(setgroups), SYS_FUNC(getgroups)): Print size parameter
as "%d".
* tests/getgroups.c: Update expected output.
* tests/setgroups.c: Likewise.
Co-authored-by: Elvira Khabirova <lineprinter0@gmail.com>
* uid.c (setfsuid, get_print_uid, print_gid): Use printuid.
* tests/setfsugid.c (printuid): New function.
(main): Use it to print UID/GID values.
* tests/setgroups.c: Likewise.
* tests/setugid.c (ugid2int): Remove.
(printuid): New function.
(main): Use it to print UID/GID values.
Co-authored-by: Elvira Khabirova <lineprinter0@gmail.com>
The string literal "__NR_syscall_4294967295" is semantically incorrect
as there is no such constant defined.
* syscall.c (syscall_name): Return NULL if there is no syscall
corresponding to the given number.
* defs.h (syscall_name): Document this behaviour.
* printsiginfo.c (print_si_info): Print syscall name with "__NR_" prefix
only if there is a syscall corresponding to si_syscall number; print
a plain syscall number otherwise.
* tests/ptrace.c (main): Update expected output.
* print_statfs.c: Print f_fsid as a structure containing field "val"
which, in turn, is an array of two elements that used to be printed
as members of f_fsid structure.
* tests/statfs.expected: Update expected output.
* tests/xstatfsx.c (print_statfs): Likewise.
* swapon.c (SYS_FUNC(swapon)): Do not check for non-zero the value
of the prio variable in order to print it.
* tests/swap.c (main): Update expected output.
This makes debugging of failing test a bit easier.
* tests/init.sh (match_grep): Add the cnt variable, increment it on every
pattern line read, print it as a prefix for non-matched pattern.
It is a number which consists of several xlat values, not a structure.
* futex.c (SYS_FUNC(futex)): Modify output of the val3 argument
of the FUTEX_WAKE_OP futex syscall command.
* tests/futex.c (main): Update expected output.
Co-authored-by: Elvira Khabirova <lineprinter0@gmail.com>
* syscall.c (dumpio): Explicitly cast the first argument of syscall
to "int", the same way as the kernel does.
* tests/read-write.c: Include <asm/unistd.h> and "kernel_types.h".
(k_read, k_write): New functions.
(test_dump, main): Use them.
* tests/preadv2-pwritev2.c: Include "kernel_types.h".
(main): Use kernel_ulong_t instead of unsigned long as types
of arguments of preadv2 and pwritev2 syscalls.
* util.c (sprinttime): Make date output conform to ISO 8601.
* tests/utime.c (print_tm): Update expected output.
* tests/xstatx.c (print_time): Likewise.
Change the parser of fault expression to follow the POLA:
-e fault=!SET
means that all syscalls except those from SET are subject to fault
injection;
-e fault=!SET1:error=ERRNO1, -e fault=!SET2:error=ERRNO2
means that all syscalls except those from SET2 are subject to fault
injection with error code ERRNO2, and all syscalls from SET2 that are
not in SET1 are subject to fault injection with error code ERRNO1.
* syscall.c (qualify_scno, qualify_syscall_class, qualify_syscall_name):
Handle negated QUAL_FAULT case differently.
* tests/fault_syntax.test: Add checks of negated sets.
* utime.c (SYS_FUNC(utime)): Print struct utimbuf as a structure,
not as an array (use curly brackets instead of square ones).
Add field names to the output.
* tests/utime.c (main): Update expected output.
This is necessary for the upcoming change in the output format
of the val3 argument of the FUTEX_WAKE_OP futex command.
* defs.h (printxvals, printxval_searchn): Change return type to int.
(printxval64, printxval, printxval_long): Likewise. Forward the value
returned by printxvals call.
* util.c (printxvals, printxval_searchn): Change return type to int,
return 1 if xlat value has been found, 0 otherwise.
This is necessary for the upcoming change in the output format
of the val3 argument of the FUTEX_WAKE_OP futex command.
* util.c (printxvals): Do not print dflt if it is NULL.
* linux/aarch64/set_scno.c: Note the kernel commit that introduced
NT_ARM_SYSTEM_CALL support.
* linux/arm/set_scno.c: Note the kernel commit that introduced
PTRACE_SET_SYSCALL support.
* linux/sparc/set_scno.c: Note the kernel commit that introduced
reloading from the syscall number register.
While -r option implies -t, it should not literally have the effect
of -t option, e.g. -rr should not increment tflag twice.
* strace.c (init): In the getopt loop, do not increment tflag for each
occurrence of -r option. After the getopt loop, if rflag is set, set
tflag to 1. Issue a warning that -tt has no effect with -r.
* print_timeval.c (print_timeval_item): New function.
(print_timeval_pair): Use it as a print_function in print_array call.
* tests/futimesat.c (main): Update to test new behaviour.
* tests/utimes.c (main): Likewise.
* defs.h [ALPHA] (timeval32_t): New type.
[ALPHA] (print_timeval32_t): New prototype.
* print_timeval.c [ALPHA] (timeval32_t): Remove.
[ALPHA] (print_timeval32_t): Remove static keyword.
* printrusage.c [ALPHA] (printrusage32): Use timeval32_t instead
of local definition of struct timeval32. Use print_timeval32_t.
After commit v4.14-136-g151d1d1 that removed the last object common
for timespec and timeval parsers there is no need to keep them all
in a single file.
* print_timespec.c: New file.
* print_timeval.c: Likewise.
* Makefile.am (strace_SOURCES): Add them.
* print_time.c: Move all timespec related code to print_timespec.c.
Move all timeval related code to print_timeval.c.
Do not assume that some syscalls do not generate syscall-exit-stops.
When syscalls fail for any reason they may generate syscall-exit-stops.
The solution is to wait for an actual exit reported by PTRACE_EVENT_EXIT
and print the end of unfinished exiting syscall properly.
* exit.c: Remove.
* Makefile.am (strace_SOURCES): Remove exit.c.
* linux/dummy.h (sys_exit): Alias to printargs_d.
* strace.c (ptrace_setoptions): Add PTRACE_O_TRACEEXIT bit.
(print_event_exit): New function.
(trace): Use it in case of PTRACE_EVENT_EXIT.
* syscall.c (trace_syscall_entering): Remove special handling
of SEN_exit.
As printargs is invoked as a generic syscall decoder only and
it is not supposed to print anything on exiting, change printargs
to return RVAL_DECODED so it would not be called on exiting at all.
* util.c (printargs): Print args unconditionally, return RVAL_DECODED.
There is little use in injections of faults into syscalls made by strace.
* syscall.c (trace_syscall_entering): Clear QUAL_FAULT bit from
tcp->qual_flg when tcp->flags has TCB_HIDE_LOG bit set.
A simultaneous use of -p option and tracing of a command available
since commit v4.11-183-gfa8c286 introduces a race condition because
the flags whether the first exec has happened are global.
Fix the race by moving hide_log_until_execve and hide_log_until_execve
global variables to TCB_HIDE_LOG and TCB_SKIP_DETACH_ON_FIRST_EXEC bits
in struct tcb.flags, correspondingly.
* defs.h (TCB_HIDE_LOG, TCB_SKIP_DETACH_ON_FIRST_EXEC, hide_log):
New macros.
(hide_log_until_execve): Remove prototype.
* strace.c (skip_one_b_execve, hide_log_until_execve): Remove.
(startup_child): Set TCB_HIDE_LOG and TCB_SKIP_DETACH_ON_FIRST_EXEC bits
in the allocated tcb structure.
(init): Remove initialization of hide_log_until_execve and
skip_one_b_execve.
(print_stopped): Use hide_log() instead of hide_log_until_execve.
(trace): Check and clear TCB_SKIP_DETACH_ON_FIRST_EXEC flag instead
of skip_one_b_execve.
* syscall.c (trace_syscall_entering): Clear TCB_HIDE_LOG flag instead of
hide_log_until_execve.
(trace_syscall_entering, trace_syscall_exiting): Check hide_log()
instead of hide_log_until_execve.
Make expressions like fault=SYSCALL1,SYSCALL2:error=EPERM work
as documented, i.e. fail both SYSCALL1 and SYSCALL2 with EPERM.
* syscall.c (parse_fault_expression): Remove const qualifier from
"name" and "token: variables, as well as from the return value.
(qual_fault): Remove const qualifier from "name" variables.
Split "name" into comma delimited tokens and pass each token
to individual qual_syscall_ex call.
(qualify): For QUAL_FAULT options, pass the whole option value
to their qualify methods without prior splitting into comma
delimited tokens.
* tests/fault_injection.test: Check it.
* tests/fault_syntax.test: Check empty syscall sets.
* maint/ioctls_sym.sh: Skip asm/core_*.h. Filter out from
asm-generic/ioctls.h those macros that are defined using unavailable
struct termios2 on alpha and powerpc.
* maint/ioctls_sym.sh: Remove pre-include linux/kvm.h workarounds.
Skip linux/kvm.h on those architectures that have no asm/kvm.h file.
Introduce post-process workarounds, add post-process workarounds
for linux/kvm.h.
Kernel code does this when compares it with -1 (apart from comparing it with
PR_SET_PTRACER_ANY, which is also -1) in security/yama/yama_lsm.c.
* prctl.c (SYS_FUNC(prctl)) <PR_SET_PTRACER>: Cast arg2 to int before
comparing it with -1.
The prctl syscall does not use compat on x32/n32, so argument types
are in fact kernel_ulong_t.
* prctl.c (print_prctl_args): Use getarg_ull, print as "%#llx" instead
of "%#lx".
(SYS_FUNC(prctl)): Use getarg_ull to store call arguments in arg2, arg3,
arg4, arg5 variables of type unsigned long long. Use them in printing
routines.
(SYS_FUNC(arch_prctl)): Use getarg_ull to store call argument in addr
variable of type unsigned long long. Print it as "%#llx" instead of
"%#lx".
This is done in order to fix build error on old distributions where
struct blkpg_ioctl_arg declaration is broken:
In file included from block.c:35:
/usr/include/linux/blkpg.h:36: error: expected ':', ',', ';', '}' or '__attribute__' before '*' token
block.c: In function 'print_blkpg_req':
block.c:113: error: 'struct_blkpg_ioctl_arg' has no member named 'data'
(from SLE10)
* block.c: Include <linux/ioctl.h> instead of <linux/blkpg.h>.
(BLKPG, BLKPG_DEVNAMELTH, BLKPG_VOLNAMELTH): New macros.
(struct_blkpg_ioctl_arg, struct_blkpg_partition): New definitions,
copied from blkpg_ioctl_arg and blkpg_partition structures defined
in <linux/blkpg.h>.
Since UAPI header declares them under CONFIG_CHECKPOINT_RESTORE
and they would never be defined in user space.
* xlat/archvals.in: Add numeric values to ARCH_MAP_VDSO_* constants.
There are no users of prctl function in prctl.c, and all the rest
is provided by <linux/prctl.h> anyway.
* prctl.c: Include <linux/prctl.h> instead of <sys/prctl.h>.
* prctl.c (SYS_FUNC(prctl)): Use printstr_ex with QUOTE_0_TERMINATED
instead of printstr to print the argument of PR_SET_NAME and PR_GET_NAME
operations.
* util.c (printstr_ex): Simplify handling of size == 0 case.
Do not artificially decrement size when QUOTE_0_TERMINATED bit is set.
Ensure that str[size] byte is non-zero if it hasn't been fetched.
When user_style has QUOTE_0_TERMINATED bit set, printstr_ex prints
the fetched string as a NUL-terminated.
After this change, the string is being fetched as a NUL-terminated, too.
* util.c (printstr_ex): Use umovestr instead of umoven
if QUOTE_0_TERMINATED bit is set.
* xlat/v4l2_device_capabilities_flags.in: Add V4L2_CAP_TOUCH introduced
by linux kernel commit v4.9-rc4~7^2~422.
* xlat/v4l2_input_types.in: Add V4L2_CAP_TOUCH introduced
by linux kernel commit v4.9-rc4~7^2~422.
Fix the following build error on SLE 11 SP4:
btrfs.c: In function 'btrfs_test_dev_replace_ioctl':
btrfs.c:1570: error: unknown field 'start' specified in initializer
* tests/btrfs.c (btrfs_test_dev_replace_ioctl): Move initialization
of struct btrfs_ioctl_dev_replace_args.start.srcdevid out of the
designated initializer.
Do not assume that the string returned by signame starts with "SIG"
prefix, this is not always the case.
* syscall.c (qual_signal): Skip signame return value
when it does not have "SIG" prefix.
Change the way how subsequent -e fault= expressions are interpreted
to implement a cumulative behavior. For example,
-e fault=file:when=3+ -e fault=chdir
now specifies that all chdir syscalls and 3+ file related syscalls
except chdir are subject for fault injection.
* syscall.c (qualify): Do not reset qual_vec for QUAL_FAULT.
* tests/fault_injection.test: Check it.
Introduce new -e fault=EXPR syntax that can be used to specify a subset
of syscalls that are subject of syscall fault injection, an error code
that has to be injected, and a frequency of injection.
The expression specifying syscall fault injection has the following
format: SET[:error=ERRNO][:when=FIRST[+[STEP]]]
where only SET is a required part and all the rest is optional.
The method used to implement syscall fault injection is the following:
on entering syscall the syscall number is substituted by an invalid
syscall number -1, and on exiting syscall the error code returned by
the kernel is substituted with the error code specified in the fault
expression.
This implementaion is based on the prototype developed
by Nahim El Atmani as a part of his GSoC 2016 strace project.
* defs.h (struct fault_opts): New forward declaration.
(struct tcb): Add fault_vec field.
(TCB_FAULT_INJ, QUAL_FAULT): New macros.
* strace.1: Document -e fault expression syntax.
* strace.c (usage): Mention -e fault expression.
(droptcb): Deallocate fault_vec member.
* syscall.c (qual_fault, arch_set_scno, arch_set_error): New prototypes.
(qual_options): Add "fault" option.
(struct fault_opts): New structure.
(num_faults): New variable.
(fault_vec): New array.
(syscall_fault_injected, tcb_fault_opts, reallocate_fault,
find_errno_by_name, qual_syscall_ex, strip_prefix, parse_fault_token,
parse_fault_expression, qual_fault, inject_syscall_fault_entering,
update_syscall_fault_exiting): New functions.
(qual_syscall): Use qual_syscall_ex.
(qualify_one): Add argument: a pointer to struct fault_opts, all callers
changed. Copy struct fault_opts from the pointer to fault_vec.
Use reallocate_fault.
(qualify_scno, qualify_syscall_class, qualify_syscall_name): Add
argument: a pointer to struct fault_opts.
(qualify): Use reallocate_fault. Do not check "all" class for
QUAL_FAULT qualifier.
(lookup_class): Check for "all" class.
(trace_syscall_entering): Use inject_syscall_fault_entering.
(trace_syscall_exiting): Use update_syscall_fault_exiting. Clear
TCB_FAULT_INJ flag along with TCB_INSYSCALL. Print " (INJECTED)" suffix
when the syscall has been injected successfully.
[ARCH_REGS_FOR_GETREGSET && !HAVE_GETREGS_OLD]
(ptrace_setregset): New function.
(ptrace_setregset_or_setregs): Define to ptrace_setregset.
[ARCH_REGS_FOR_GETREGS && !HAVE_GETREGS_OLD]
(ptrace_setregs): New function.
(ptrace_setregset_or_setregs): Define to ptrace_setregs.
[ptrace_setregset_or_setregs] (set_regs): New function.
Include "set_scno.c" and "set_error.c"
* NEWS: Mention this enhancement.
The incomplete CRIS support introduced by commit v4.5.18-77-gea0e6e8
should not have been merged because it lacks essential parts, e.g.
syscall tables.
This change adds missing syscall tables for crisv10 and crisv32.
* linux/crisv10/syscallent.h: New file.
* linux/crisv32/syscallent.h: Likewise.
* Makefile.am (EXTRA_DIST): Add them.
* defs.h (string_to_uint_ex): New prototype.
(string_to_uint_upto): New function, a thin wrapper around
string_to_uint_ex.
* util.c (string_to_uint_ex): New function.
(string_to_uint): Change into a thin wrapper around string_to_uint_upto.
* strace.c (init): Use string_to_uint_upto.
* syscall.c (qualify_scno, qual_signal, qual_desc): Use
string_to_uint_upto instead of string_to_uint.
Split qual_syscall into qualify_scno, qualify_syscall_class,
and qualify_syscall_name.
This might be needed later to implement syscall fault injection.
* syscall.c (qualify_scno, qualify_syscall_class, qualify_syscall_name):
New functions.
(qual_syscall): Use them.
ioctl_block.c:48: warning: 'init_magic' defined but not used
* tests/ioctl_block.c (init_magic): Define only for
[BLKTRACESETUP && HAVE_STRUCT_BLK_USER_TRACE_SETUP].
alarm.c: In function 'main':
alarm.c:41: warning: integer constant is too large for 'long' type
aio.c: In function 'main':
aio.c:162: warning: integer constant is too large for 'long' type
btrfs.c: In function 'btrfs_test_sync_ioctls':
btrfs.c:202: warning: integer constant is too large for 'long' type
btrfs.c: In function 'btrfs_test_subvol_ioctls':
btrfs.c:289: warning: integer constant is too large for 'long' type
btrfs.c:290: warning: integer constant is too large for 'long' type
dup2.c: In function 'main':
dup2.c:12: warning: integer constant is too large for 'long' type
dup2.c:13: warning: integer constant is too large for 'long' type
dup3.c: In function 'main':
dup3.c:13: warning: integer constant is too large for 'long' type
dup3.c:14: warning: integer constant is too large for 'long' type
epoll_create.c: In function 'main':
epoll_create.c:12: warning: integer constant is too large for 'long' type
epoll_ctl.c: In function 'invoke_syscall':
epoll_ctl.c:14: warning: integer constant is too large for 'long' type
faccessat.c: In function 'main':
faccessat.c:13: warning: integer constant is too large for 'long' type
fchdir.c: In function 'main':
fchdir.c:12: warning: integer constant is too large for 'long' type
struct_flock.c: In function 'invoke_test_syscall':
struct_flock.c:48: warning: integer constant is too large for 'long' type
struct_flock.c: In function 'test_flock_einval':
struct_flock.c:58: warning: integer constant is too large for 'long' type
struct_flock.c:59: warning: integer constant is too large for 'long' type
fcntl64.c: In function 'test_flock64_einval':
fcntl64.c:44: warning: integer constant is too large for 'long' type
fcntl64.c:45: warning: integer constant is too large for 'long' type
fcntl.c: In function 'test_flock64_einval':
fcntl.c:44: warning: integer constant is too large for 'long' type
fcntl.c:45: warning: integer constant is too large for 'long' type
fdatasync.c: In function 'main':
fdatasync.c:12: warning: integer constant is too large for 'long' type
flock.c: In function 'main':
flock.c:13: warning: integer constant is too large for 'long' type
xstatx.c: In function 'main':
xstatx.c:255: warning: integer constant is too large for 'long' type
xstatx.c:290: warning: integer constant is too large for 'long' type
xstatx.c:290: warning: integer constant is too large for 'long' type
xstatx.c:290: warning: integer constant is too large for 'long' type
xstatx.c:290: warning: integer constant is too large for 'long' type
xstatx.c:290: warning: integer constant is too large for 'long' type
xstatx.c:290: warning: integer constant is too large for 'long' type
xstatx.c:290: warning: integer constant is too large for 'long' type
xstatx.c:290: warning: integer constant is too large for 'long' type
xstatx.c:290: warning: integer constant is too large for 'long' type
xstatx.c:294: warning: integer constant is too large for 'long' type
xstatx.c:294: warning: integer constant is too large for 'long' type
xstatx.c:294: warning: integer constant is too large for 'long' type
xstatx.c:294: warning: integer constant is too large for 'long' type
xstatx.c:294: warning: integer constant is too large for 'long' type
xstatx.c:294: warning: integer constant is too large for 'long' type
xstatx.c:294: warning: integer constant is too large for 'long' type
xstatx.c:294: warning: integer constant is too large for 'long' type
xstatx.c:294: warning: integer constant is too large for 'long' type
xstatfsx.c: In function 'main':
xstatfsx.c:113: warning: integer constant is too large for 'long' type
fsync.c: In function 'main':
fsync.c:12: warning: integer constant is too large for 'long' type
ftruncate64.c: In function 'main':
ftruncate64.c:39: warning: integer constant is too large for 'long' type
futimesat.c: In function 'main':
futimesat.c:52: warning: integer constant is too large for 'long' type
futimesat.c:59: warning: integer constant is too large for 'long' type
get_mempolicy.c: In function 'main':
get_mempolicy.c:83: warning: integer constant is too large for 'long' type
get_mempolicy.c:84: warning: integer constant is too large for 'long' type
get_mempolicy.c:85: warning: integer constant is too large for 'long' type
get_mempolicy.c:86: warning: integer constant is too large for 'long' type
getdents.c: In function 'main':
getdents.c:109: warning: integer constant is too large for 'long' type
getdents.c:110: warning: integer constant is too large for 'long' type
getdents.c:114: warning: integer constant is too large for 'long' type
getdents64.c: In function 'main':
getdents64.c:114: warning: integer constant is too large for 'long' type
getdents64.c:115: warning: integer constant is too large for 'long' type
getdents64.c:119: warning: integer constant is too large for 'long' type
xgetrlimit.c: In function 'main':
xgetrlimit.c:71: warning: integer constant is too large for 'long' type
ioctl_block.c:44: warning: integer constant is too large for 'long' type
ioctl_block.c: In function 'main':
ioctl_block.c:136: warning: integer constant is too large for 'long' type
ioctl_block.c:137: warning: integer constant is too large for 'long' type
ioctl_block.c:161: warning: integer constant is too large for 'long' type
ioctl_block.c:170: warning: integer constant is too large for 'long' type
ioctl_block.c:171: warning: integer constant is too large for 'long' type
ioctl_evdev.c:42: warning: integer constant is too large for 'long' type
ioctl_mtd.c:46: warning: integer constant is too large for 'long' type
ioctl_rtc.c:40: warning: integer constant is too large for 'long' type
ioperm.c: In function 'main':
ioperm.c:12: warning: integer constant is too large for 'long' type
ioctl_v4l2.c: In function 'main':
ioctl_v4l2.c:409: warning: integer constant is too large for 'long' type
ipc.c: In function 'ipc_call':
ipc.c:52: warning: integer constant is too large for 'long' type
kill.c: In function 'main':
kill.c:59: warning: integer constant is too large for 'long' type
kill.c:62: warning: integer constant is too large for 'long' type
kill.c:63: warning: integer constant is too large for 'long' type
kill.c:68: warning: integer constant is too large for 'long' type
linkat.c: In function 'main':
linkat.c:14: warning: integer constant is too large for 'long' type
linkat.c:15: warning: integer constant is too large for 'long' type
mbind.c: In function 'main':
mbind.c:41: warning: integer constant is too large for 'long' type
mbind.c:43: warning: integer constant is too large for 'long' type
mbind.c:44: warning: integer constant is too large for 'long' type
migrate_pages.c: In function 'main':
migrate_pages.c:41: warning: integer constant is too large for 'long' type
In file included from mkdirat.c:10:
umode_t.c: In function 'test_syscall':
umode_t.c:47: warning: integer constant is too large for 'long' type
mknod.c: In function 'call_mknod':
mknod.c:16: warning: integer constant is too large for 'long' type
mknod.c: In function 'main':
mknod.c:23: warning: integer constant is too large for 'long' type
mknod.c:49: warning: integer constant is too large for 'long' type
mknod.c:56: warning: integer constant is too large for 'long' type
mknodat.c:12: warning: integer constant is too large for 'long' type
mknodat.c: In function 'call_mknodat':
mknodat.c:17: warning: integer constant is too large for 'long' type
mknodat.c: In function 'main':
mknodat.c:24: warning: integer constant is too large for 'long' type
mknodat.c:50: warning: integer constant is too large for 'long' type
mknodat.c:57: warning: integer constant is too large for 'long' type
mmap.c: In function 'main':
mmap.c:51: warning: integer constant is too large for 'long' type
move_pages.c: In function 'print_stat_pages':
move_pages.c:139: warning: integer constant is too large for 'long' type
move_pages.c: In function 'print_move_pages':
move_pages.c:166: warning: integer constant is too large for 'long' type
move_pages.c: In function 'main':
move_pages.c:188: warning: integer constant is too large for 'long' type
mq_sendrecv.c: In function 'main':
mq_sendrecv.c:238: warning: integer constant is too large for 'long' type
mq_sendrecv.c:401: warning: integer constant is too large for 'long' type
mq_sendrecv.c:403: warning: integer constant is too large for 'long' type
nsyscalls.c: In function 'test_syscall':
nsyscalls.c:64: warning: integer constant is too large for 'long' type
nsyscalls.c:65: warning: integer constant is too large for 'long' type
nsyscalls.c:66: warning: integer constant is too large for 'long' type
nsyscalls.c:67: warning: integer constant is too large for 'long' type
nsyscalls.c:68: warning: integer constant is too large for 'long' type
nsyscalls.c:69: warning: integer constant is too large for 'long' type
preadv-pwritev.c: In function 'main':
preadv-pwritev.c:162: warning: comparison between signed and unsigned
prlimit64.c: In function 'main':
prlimit64.c:65: warning: integer constant is too large for 'long' type
prlimit64.c:70: warning: integer constant is too large for 'long' type
remap_file_pages.c: In function 'main':
remap_file_pages.c:42: warning: integer constant is too large for 'long' type
remap_file_pages.c:43: warning: integer constant is too large for 'long' type
remap_file_pages.c:45: warning: integer constant is too large for 'long' type
renameat.c: In function 'main':
renameat.c:15: warning: integer constant is too large for 'long' type
renameat.c:16: warning: integer constant is too large for 'long' type
rt_sigqueueinfo.c: In function 'main':
rt_sigqueueinfo.c:41: warning: integer constant is too large for 'long' type
rt_tgsigqueueinfo.c: In function 'k_tgsigqueueinfo':
rt_tgsigqueueinfo.c:45: warning: integer constant is too large for 'long' type
rt_tgsigqueueinfo.c:46: warning: integer constant is too large for 'long' type
rt_tgsigqueueinfo.c:47: warning: integer constant is too large for 'long' type
rt_tgsigqueueinfo.c: In function 'main':
rt_tgsigqueueinfo.c:67: warning: integer constant is too large for 'long' type
sendfile.c: In function 'main':
sendfile.c:110: warning: integer constant is too large for 'long' type
sendfile64.c: In function 'main':
sendfile64.c💯 warning: integer constant is too large for 'long' type
set_mempolicy.c: In function 'main':
set_mempolicy.c:129: warning: integer constant is too large for 'long' type
set_mempolicy.c:130: warning: integer constant is too large for 'long' type
setrlimit.c: In function 'main':
setrlimit.c:44: warning: integer constant is too large for 'long' type
socketcall.c: In function 'test_socketcall':
socketcall.c:58: warning: integer constant is too large for 'long' type
splice.c: In function 'main':
splice.c:41: warning: integer constant is too large for 'long' type
splice.c:42: warning: integer constant is too large for 'long' type
splice.c:45: warning: integer constant is too large for 'long' type
splice.c:46: warning: integer constant is too large for 'long' type
symlinkat.c: In function 'main':
symlinkat.c:12: warning: integer constant is too large for 'long' type
sync_file_range.c: In function 'main':
sync_file_range.c:42: warning: integer constant is too large for 'long' type
sync_file_range.c:43: warning: integer constant is too large for 'long' type
syslog.c: In function 'main':
syslog.c:14: warning: integer constant is too large for 'long' type
tee.c: In function 'main':
tee.c:41: warning: integer constant is too large for 'long' type
tee.c:42: warning: integer constant is too large for 'long' type
timer_create.c: In function 'main':
timer_create.c:52: warning: integer constant is too large for 'long' type
timer_create.c:84: warning: integer constant is too large for 'long' type
timer_create.c:85: warning: integer constant is too large for 'long' type
truncate64.c: In function 'main':
truncate64.c:41: warning: integer constant is too large for 'long' type
xgetrlimit.c: In function 'main':
xgetrlimit.c:71: warning: integer constant is too large for 'long' type
umode_t.c: In function 'test_syscall':
umode_t.c:46: warning: integer constant is too large for 'long' type
unlinkat.c: In function 'main':
unlinkat.c:13: warning: integer constant is too large for 'long' type
waitpid.c: In function 'main':
waitpid.c:43: warning: integer constant is too large for 'long' type
waitid.c: In function 'poison':
waitid.c:141: warning: integer constant is too large for 'long' type
xetpriority.c: In function 'main':
xetpriority.c:15: warning: integer constant is too large for 'long' type
xetpriority.c:19: warning: integer constant is too large for 'long' type
xetpriority.c:20: warning: integer constant is too large for 'long' type
xetpgid.c: In function 'main':
xetpgid.c:43: warning: integer constant is too large for 'long' type
xetpgid.c:47: warning: integer constant is too large for 'long' type
xetpgid.c:48: warning: integer constant is too large for 'long' type
(SLES 11SP4, i386)
* tests/alarm.c (main): Add proper suffix to integer constant.
* tests/aio.c (main): Likewise.
* tests/btrfs.c (btrfs_test_sync_ioctls, btrfs_test_subvol_ioctls): Likewise.
* tests/dup2.c (main): Likewise.
* tests/dup3.c (main): Likewise.
* tests/epoll_create.c (main): Likewise.
* tests/epoll_ctl.c (invoke_syscall): Likewise.
* tests/faccessat.c (main): Likewise.
* tests/fchdir.c (main): Likewise.
* tests/fcntl.c (test_flock64_einval): Likewise.
* tests/fcntl64.c (test_flock64_einval): Likewise.
* tests/fdatasync.c (main): Likewise.
* tests/flock.c (main): Likewise.
* tests/fstat64.c (SAMPLE_SIZE): Likewise.
* tests/fstat.c (SAMPLE_SIZE): Likewise.
* tests/fsync.c (main): Likewise.
* tests/ftruncate64.c (main): Likewise.
* tests/futimesat.c (main): Likewise.
* tests/get_mempolicy.c (main): Likewise.
* tests/getdents.c (main): Likewise.
* tests/getdents64.c (main): Likewise.
* tests/ioctl_block.c (lmagic, main): Likewise.
* tests/ioctl_evdev.c (lmagic): Likewise.
* tests/ioctl_mtd.c (lmagic): Likewise.
* tests/ioctl_rtc.c (lmagic): Likewise.
* tests/ioctl_v4l2.c (main): Likewise.
* tests/ioperm.c (main): Likewise.
* tests/ipc.c (ipc_call): Likewise.
* tests/kill.c (main): Likewise.
* tests/linkat.c (main): Likewise.
* tests/lstat64.c (SAMPLE_SIZE): Likewise.
* tests/lstat.c (SAMPLE_SIZE): Likewise.
* tests/mbind.c (main): Likewise.
* tests/migrate_pages.c (main): Likewise.
* tests/mkdirat.c (TEST_SYSCALL_PREFIX_ARGS): Likewise.
* tests/mknod.c (call_mknod, main): Likewise.
* tests/mknodat.c (fd, call_mknodat, main): Likewise.
* tests/mmap.c (main): Likewise.
* tests/move_pages.c (print_stat_pages, print_move_pages): Likewise.
* tests/mq_sendrecv.c (main): Likewise.
* tests/nsyscalls.c (test_syscall): Likewise.
* tests/prlimit64.c (main): Likewise.
* tests/remap_file_pages.c (main): Likewise.
* tests/renameat.c (main): Likewise.
* tests/rt_sigqueueinfo.c (main): Likewise.
* tests/rt_tgsigqueueinfo.c (k_tgsigqueueinfo, main): Likewise.
* tests/sendfile.c (main): Likewise.
* tests/sendfile64.c (main): Likewise.
* tests/set_mempolicy.c (main): Likewise.
* tests/setrlimit.c (main): Likewise.
* tests/socketcall.c (test_socketcall): Likewise.
* tests/splice.c (main): Likewise.
* tests/stat64.c (SAMPLE_SIZE): Likewise.
* tests/stat.c (SAMPLE_SIZE): Likewise.
* tests/struct_flock.c (invoke_test_syscall, test_flock_einval): Likewise.
* tests/symlinkat.c (main): Likewise.
* tests/sync_file_range.c (main): Likewise.
* tests/syslog.c (main): Likewise.
* tests/tee.c (main): Likewise.
* tests/timer_create.c (main): Likewise.
* tests/truncate64.c (main): Likewise.
* tests/umode_t.c (test_syscall): Likewise.
* tests/unlinkat.c (main): Likewise.
* tests/waitid.c (main): Likewise.
* tests/waitpid.c (main): Likewise.
* tests/xetpgid.c (main): Likewise.
* tests/xetpriority.c (main): Likewise.
* tests/xgetrlimit.c (main): Likewise.
* tests/xstatfsx.c (main) [CHECK_ODD_SIZE]: Likewise.
* tests/xstatx.c (main): Likewise.
preadv-pwritev.c: In function 'main':
preadv-pwritev.c:162: warning: comparison between signed and unsigned
readv.c: In function 'main':
readv.c:148: warning: comparison between signed and unsigned
recvmsg.c: In function 'main':
recvmsg.c:148: warning: comparison between signed and unsigned
* tests/preadv-pwritev.c (main): Cast r_len to int.
* tests/readv.c (main): Likewise.
* tests/recvmsg.c (main): Likewise.
This change moves remaining arch specific getregs_old code into
appropriate arch subdirectories and removes unnecessary code
duplication.
* linux/getregs_old.h: New file.
* linux/powerpc/getregs_old.h: Likewise.
* linux/powerpc64/getregs_old.h: Likewise.
* linux/x86_64/getregs_old.h: Likewise.
* Makefile.am (EXTRA_DIST): Add them.
* syscall.c: Include "getregs_old.h".
[X86_64 || POWERPC]: Remove.
[ARCH_REGS_FOR_GETREGSET] (ptrace_getregset_or_getregs): Define
to ptrace_getregset.
[ARCH_REGS_FOR_GETREGS] (ptrace_getregset_or_getregs): Define
to ptrace_getregs.
(get_regs): Check for ptrace_getregset_or_getregs instead
of ARCH_REGS_FOR_GETREGSET and ARCH_REGS_FOR_GETREGS. Use
ptrace_getregset_or_getregs instead of ptrace_getregset and
ptrace_getregs. Check for HAVE_GETREGS_OLD instead of X86_64
and POWERPC. Use use_getregs_old instead of getregset_support
and old_kernel.
In order to avoid name clash on distributions which decide to include
kernel headers instead of UAPI ones. For example, on SLES 11SP4:
gcc -DHAVE_CONFIG_H -I./linux/x86_64 -I./linux -I. -Wall -O2 -c net.c
In file included from /usr/include/linux/sysctl.h:25:0,
from /usr/include/linux/netfilter.h:6,
from /usr/include/linux/netfilter_arp.h:8,
from /usr/include/linux/netfilter_arp/arp_tables.h:14,
from net.c:60:
net.c: In function 'print_packet_mreq':
gcc_compat.h:59:27: error: negative width in bit-field '<anonymous>'
# define MUST_BE_ARRAY(a) BUILD_BUG_ON_ZERO(!SAME_TYPE((a), &(a)[0]))
^
defs.h:76:53: note: in expansion of macro 'MUST_BE_ARRAY'
#define ARRAY_SIZE(a) (sizeof(a) / sizeof((a)[0]) + MUST_BE_ARRAY(a))
^
net.c:747:22: note: in expansion of macro 'ARRAY_SIZE'
if (mreq.mr_alen > ARRAY_SIZE(mreq.mr_address))
^
due to the fact BUILD_BUG_ON_ZERO is already defined in <linux/kernel.h>.
* gcc_compat.h (BUILD_BUG_ON_ZERO): Rename to FAIL_BUILD_ON_ZERO.
(MUST_BE_ARRAY): Update usage.
Why gcc didn't complain about it? It is definitely an index clash.
* linux/x32/syscallent.h: Change initializer indices for empty records from
[327 ... 511] to [329 ... 511].
This will be needed to implement fault injection on those architectures
that lack PTRACE_SETREGSET/PTRACE_SETREGS support.
* defs.h (upoke): New prototype.
* upoke.c: New file.
* Makefile.am (libstrace_a_SOURCES): Add it.
This enables printing size-limited (expectedly) ASCIZ strings.
This is done by increasing umoven size limit for sized strings by one
byte above max_strlen (enabling copying possible NUL byte in case len is
greater than max_strlen) and decreasing size after copying by one byte
in case QUOTE_0_TERMINATED is set (due to user_style or usage of len of
-1). As a result, there is one excess byte for string_quote in case
QUOTE_0_TERMINATED is set so string_quote can check for NUL termination
of strings up to size bytes in size (which is len or max_strlen, whatever
is greater).
The catch here is that when string is not properly NUL-terminated and
QUOTE_0_TERMINATED is provided in user_style and len is less than
max_strlen then last non-NUL byte is not printed. But ellipsis is
printed instead, being indication that string is not terminated
properly. QUOTE_OMIT_TRAILING_0 should be used instead in case this
behaviour is not intended.
* util.c (printstr_ex): Copy one excess byte in case of non-negative len
provided and it is more than max_strlen; handle case of max_strlen of 0
in case QUOTE_0_TERMINATED is set separately; check for need of printing
ellipsis by checking resulting style against QUOTE_0_TERMINATED.
* linux/perf_event_struct.h: New file, definition of struct perf_event_attr
from Linux 4.8
* Makefile.am (strace_SOURCES): Add it.
* xlat/hw_breakpoint_len.in: New file.
* xlat/hw_breakpoint_type.in: Likewise.
* xlat/perf_attr_size.in: Likewise.
* xlat/perf_branch_sample_type.in: Likewise.
* xlat/perf_event_read_format.in: Likewise.
* xlat/perf_event_sample_format.in: Likewise.
* xlat/perf_hw_cache_id.in: Likewise.
* xlat/perf_hw_cache_op_id.in: Likewise.
* xlat/perf_hw_cache_op_result_id.in: Likewise.
* xlat/perf_hw_id.in: Likewise.
* xlat/perf_sw_ids.in: Likewise.
* xlat/perf_type_id.in: Likewise.
* perf.c [HAVE_LINUX_PERF_EVENT_H]: Remove inclusion of
<linux/perf_event.h>.
(PRINT_XLAT): New macro for simplifying printing values from
sorted xlat.
(printxlat_search): New function, wrapper for xlat_search with behaviour
similar to printval.
(print_perf_event_attr): New print_event_attr structure fetching and
printing function.
(SYS_FUNC(perf_event_open)): Use print_perf_event_attr for displaying
attr argument contents.
* tests/perf_event_open.c: Renamed to
tests/perf_event_open_nonverbose.c.
* tests/perf_event_open.test: add -e verbose=none in order to preserve
output format being checked (in case verbose output is enabled,
contents of attr arguments are shown now); rename to
tests/perf_event_open_nonverbose.test; add -e trace=perf_event_open.
* util.c (printxval_searchn): New function.
* defs.h (printxval_searchn): New prototype.
(printxval_search): New helper macro useful in conjunction with static
xlat arrays.
Also, use u_rval for determining message size.
* mq.c (SYS_FUNC(mq_timedreceive)): Check for syserror and non-negative
u_rval before calling printstr on msg_ptr argument, or use printaddr otherwise;
provide u_rval as string size.
mq_flags field of struct mq_attr can contain only O_NONBLOCK flag (other
flags are treated as invalid by mq_getsetattr). Moreover, this field is
ignored by mq_open at all (O_NONBLOCK is set via oflag, not via attr
attribute).
* xlat/mq_attr_flags.in: New file.
* print_mq_attr.c: Include xlat/mq_attr_flags.h.
(printmqattr): New parameter, bool decode_flags; cast members of struct
mq_attr to long long type; use mq_attr_flags xlat for printing mq_flags;
print mq_flags as flags only in case decode_flags parameter is set
to true.
* mq.c (SYS_FUNC(mq_open)): Specify value of false for decode_flags
parameter of printmqattr call.
(SYS_FUNC(mq_getsetattr)): Specify value of true for decode_flags
parameter of printmqattr call.
* tests/mq.expected: Update expected output.
* mq.c (SYS_FUNC(mq_timedsend)): Change conversion specifier from "%lu"
to "%llu", use getarg_ull for obtaining msg_len parameter.
(SYS_FUNC(mq_timedreceive)): Likewise.
It is defined as int in <linux/posix_types.h>.
* mq.c (SYS_FUNC(mq_timedsend)): Use "%d" conversion specifier instead
of "%ld" for printing mqdes (0th parameter), cast parameter value to
int.
(SYS_FUNC(mq_timedreceive)): Likewise.
(SYS_FUNC(mq_notify)): Likewise.
(SYS_FUNC(mq_getsetattr)): Likewise.
* tests/file_handle.c: replace fcntl.h include with asm/unistd.h.
[MAX_HANDLE_SZ]: change to defined __NR_name_to_handle_at
&& defined __NR_open_by_handle_at, add fcntl.h include
[!MAX_HANDLE_SZ]: Add definition of MAX_HANDLE_SZ and struct
file_handle.
(main): Change name_to_handle_at and open_by_handle_at calls to syscall.
Since the only two flags supported are AT_SYMLINK_FOLLOW and
AT_EMPTY_PATH.
* xlat/name_to_handle_at_flags.in: New file.
* file_handle.c (SYS_FUNC(name_to_handle_at)): Use name_to_handle_at_flags
for printing flags parameter.
Curiously, kexec_load uses compat on x32/n32, so its parameters should
be 4 bytes in size on these ABIs.
* kexec.c (SYS_FUNC(kexec_load)): Use widen_to_ulong for casting
parameters to proper size on x32/n32.
Kernel declares flags parameter as long and looks like x32 and n32
implement no compat for this call.
* clone.c (SYS_FUNC(unshare)): Use getarg_ull and printflags64 for
obtaining and parsing flags parameter.
Not all clone flags could be passed to unshare (see check_unshare_flags
function in kernel/fork.c).
* xlat/unshare_flags.in: New file.
* clone.c (SYS_FUNC(unshare)): Use unshare_flags for printing flags
parameter.
nstype is not flag set but specific type value against which it is
compared in kernel (see kernel/nsproxy.c, SYSCALL_DEFINE2(setns, ...)).
* xlat/setns_types.in: New file.
* clone.c: Use printxval and setns_types for nstype parameter.
As init_module has no compat wrapper, its first parameter is a pointer
and the second is length, they both have the same size as kernel_ulong_t
type.
* bjm.c (SYS_FUNC(init_module)): Obtain first two parameters using
getarg_ull, print address argument using printaddr_ull, print length
argument using "%llu" conversion specifier.
* defs.h (printaddr_ull): New prototype.
(printaddr): Change to a static inline wrapper around printaddr_ull.
* util.c (printaddr): Rename to printaddr_ull, change argument type
to unsigned long long, change print format to %#llx.
* bjm.c (MODULE_INIT_IGNORE_MODVERSIONS, MODULE_INIT_IGNORE_VERMAGIC):
Remove.
* xlat/module_init_flags.in: Add values for
MODULE_INIT_IGNORE_MODVERSIONS and MODULE_INIT_IGNORE_VERMAGIC records.
Since create_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 (SYS_FUNC(create_module)): Remove.
* linux/dummy.h (sys_create_module): Add stub alias.
This check had been done by print_array inside tprint_iov anyway.
* process_vm.c (SYS_FUNC(process_vm_readv)): Use tprint_iov_upto for
printing local_iov; do not check for syserror, provide decode_iov
parameter to tprint_iov{,_upto} based on its value instead.
This enables printing of iovec arrays even in case of failed syscall
(failed syscall doesn't mean that iovec itself is inaccessible and
useless). One caveat here is that we should explicitly provide proper
IOV_DECODE_* value based on syserror value in case printing is performed
on exiting; we can't simply override it to IOV_DECODE_ADDR on exiting
when syserror is up, since this may be called by code which tries to
print iovec containing local data, which should be perfectly accessible
(on the other hand, there are no cases of such behaviour at the moment).
Since iovecs themselves are printed even if syscall has failed now,
preadv test is updated to reflect this. It is notable, though, that this
is the only place where this case is checked.
* io.c (tprint_iov_upto): Specify umoven_or_printaddr_ignore_syserror
instead of umoven_or_printaddr as umoven_func parameter.
(SYS_FUNC(readv), do_preadv): Specify decode_iov parameter value
based on syserror(tcp) value.
* scsi.c: (print_sg_io_v3_res, print_sg_io_v4_res): Likewise.
* tests/preadv.c: Update expected output for the case when preadv
with singe-item iovec failed.
This change makes behaviour of umoven_or_printaddr_ignore_syserror in
line with umoven_or_printaddr when verbose flag is disabled.
* v4l2.c (umoven_or_printaddr_ignore_syserror): Simply call printaddr
when verbose flag is unset.
It is size_t and without compat on x32/n32.
* lookup_dcookie.c (SYS_FUNC(lookup_dcookie)): Change conversion
specifier from "%lu" to "%llu", use getarg_ull for len argument
retrieval.
Analogous to offsetof but returns structure offset after the specified
field. Useful for checking whether specific field is present in obtained
data or specifying amount of data to copy based on the (last) field needed.
* defs.h (offsetofend): New macro.
Since new values have been added gradually over various kernel versions,
it's better to define them explicitly in order to avoid situations when
strace built with older kernel headers cannot decode some recently
defined values.
* xlat/clocknames.in: Add values for constants.
* tests/quotactl.h (fill_memory_ex): Add period parameter,
use it as a divisor in non-constant part of value.
(fill_memory): Specify period of 0x80 to fill_memory_ex call.
* tests/quotactl-xfs.c (main): Likewise.
This potentially fixes keyctl decoder for x32 personality.
* keyctl.c (keyctl_update_key, keyctl_read_key, keyctl_instantiate_key,
keyctl_instantiate_key_iov, keyctl_dh_compute): Change addr and len
arguments to kernel_ulong_t. Print len using %llu format.
(keyctl_keyring_search): Change addr1 and addr2 arguments
to kernel_ulong_t.
(print_dh_params): Change addr argument to kernel_ulong_t.
(SYS_FUNC(keyctl)): Retrieve arguments via getarg_ull, pass them
to the appropriate handlers.
* keyctl.c (keyctl_update_key): Change addr and len arguments
to unsigned.
(keyctl_read_key): Likewise.
(keyctl_instantiate_key): Likewise.
(keyctl_instantiate_key_iov): Likewise.
(keyctl_dh_compute): Likewise.
(keyctl_keyring_search): Change addr1 and addr2 arguments to unsigned.
(keyctl_chown_key): Change user and group arguments to unsigned.
(keyctl_get_persistent): Change uid argument to unsigned.
keyctl_read_key had subtle bug by treating out buffer in KEYCTL_READ
as NUL-terminated, which is not true. We fix it by adding parameter to
keyctl_read_key signalising whether buffer is NUL-terminated and using
printstr_ex for printing (expectedly) NUL-terminated strings.
* keyctl.c (keyctl_read_key): Add has_nul parameter. Do not use -1 as
string len. Use printstr_ex for buffer output with user style depending
on has_nul value.
(SYS_FUNC(keyctl)): Specify has_nul parameter to keyctl_read_key
by comparing cmd value with KEYCTL_READ.
This is useful for providing QUOTE_OMIT_TRAILING_0 and maybe other
flags.
* defs.h (printstr_ex): New prototype.
(printstr): Change to a wrapper around printstr_ex with zero user style.
* util.c (printstr): Rename to ...
(printstr_ex) ... new function, add user_style argument which is or'ed
with computed style.
It is useful in cases strings with size provided are expected to be
NUL-terminated but are not trustworthy enough to call just plain
printstr(str, -1).
* defs.h (QUOTE_OMIT_TRAILING_0): New constant definition.
* util.c (string_quote): Swallow terminating NUL if
QUOTE_OMIT_TRAILING_0 is set.
* keyctl.c (struct keyctl_dh_params): New structure.
(print_dh_params, keyctl_dh_compute): New functions.
(SYS_FUNC(keyctl)): Add support for KEYCTL_DH_COMPUTE cmd value.
* syscall.c (trace_syscall_exiting): Use err_name() instead
of open-coding it. Print unrecognized errno values using %lu format
instead of ERRNO_%lu as the latter prodices an invalid constant.
This is the type actually used for the error code on architectures
that use a dedicated register.
* defs.h (struct tcb): Change the type of u_error to unsigned long.
* syscall.c (trace_syscall_exiting): Change the type of u_error variable
to unsigned long, print it using %lu format, drop no longer needed
explicit cast to unsigned long.
(saved_u_error): Change type to unsigned long.
Since this command doesn't have any additional arguments, the comma does
not needed. Since this is the only command which lacks additional
arguments, it's better to add special case for it rather than add
printing of comma to all other commands.
* keyctl.c (SYS_FUNC(keyctl)): Add check for command not being
KEYCTL_SESSION_TO_PARENT when printing comma dividing cmd argument
from the rest.
Otherwise it is erroneous on x32, for example.
* keyctl.c (SYS_FUNC(keyctl)): Use "%#llx" conversion specifier
and getarg_ull for fallback argument printing.
UID/GID are unsigned except special -1 value (which is also special in
context of specific keyctl commands), so special printing function
should be used.
* keyctl.c (keyctl_chown_key, keyctl_get_persistent): Use printuid
instead of printf with "%d" conversion for printing UID/GID.
Otherwise it can be depleted and print_iovec starts printing empty
strings.
* io.c (print_iovec): Interpret c->data_size of -1 as unlimited data
and do not decrease it in this case.
Enhance abbrev=, raw=, and verbose= to accept the same syntax as trace=.
For example, this allows such syntax as -e verbose=file.
* syscall.c (lookup_class): Define before qual_syscall.
(qualify): Move the loop based on lookup_class ...
(qual_syscall): ... here.
* tests/qual_syscall.test: Check it.
fadvise_64_64 decoder used before this change prints different sign
of the "len" argument (loff_t vs size_t).
* linux/x32/syscallent.h (fadvise64): Replace SEN(fadvise64_64)
with SEN(fadvise64).
Since its type is size_t, it is 64-bit wide on x32 and special care
should be taken in order to obtain it.
* fadvise.c (SYS_FUNC(fadvise64)): Use getarg_ull for obtaining value
of "len" syscall argument.
These allow retrieving specific argument in full taking into account
peculiarities of runtimes which employ tcp->ext_arg (e.g. x32).
* defs.h (getarg_ll, getarg_ull): New prototypes.
* util.c (getarg_ll, getarg_ull): New functions.
(printargs): Use getarg_ull.
The kernel declares fadvise64 as
long sys_fadvise64(int fd, loff_t offset, size_t len, int advice);
* fadvise.c (SYS_FUNC(fadvise64)): Change conversion specifier from
"%ld" to "%lu" for printing len argument since kernel expects argument
of type size_t.
This change is no-op yet, but things will change when decoder
of fadvise64 syscall get fixed.
* linux/sh64/syscallent.h (fadvise64_64): Replace SEN(fadvise64)
with SEN(fadvise64_64).
* linux/sparc64/syscallent.h (fadvise64_64): Likewise.
Since types of these arguments are off_t and kernel actually expects
signed values in order to fail when negative values are provided,
lets display these values as signed.
* fallocate.c (SYS_FUNC(fallocate)): Change conversion specifier for
printing "offset" and "len" syscall arguments from %llu to %lld.
Support for these quota subcommands by linux kernels has been dropped
long time ago (between 2.5.16 and 2.5.17), so lets drop it in order
to minimize amount of code which should be subjected to testing and
structured conversion.
* quota.c (struct v1_dqblk, struct v2_dqblk, struct v2_dqinfo,
struct v1_dqstats, struct v2_dqstats): Remove.
(decode_cmd_data): Remove handling of Q_V1_GETQUOTA, Q_V1_SETQUOTA,
Q_V2_GETQUOTA, Q_V2_SETQUOTA, Q_V2_GETINFO, Q_V2_SETINFO, Q_V1_GETSTATS,
and Q_V2_GETSTATS subcommands.
(SYS_FUNC(quotactl)): Remove handling of Q_V1_QUOTAON subcommand.
Since the only difference between 32-bit and 64-bit environments
regarding this structure is its padding, lets just add packed attribute
to its definition instead of going full mpers.
* quota.c (struct if_dqblk): Add ATTRIBUTE_PACKED.
Q_XQUOTASYNC is no-op, but it does not require id/addr arguments anyway.
* quota.c (decode_cmd_data): Add Q_SYNC and Q_XQUOTASYNC to subcommand
dispatch switch since id/addr arguments for these subcommands are known
to be ignored.
As id argument is not used for Q_QUOTAOFF, Q_GETFMT, Q_V2_GETINFO,
Q_GETINFO, Q_V2_SEETINFO, Q_SETINFO, Q_SYNC, Q_XQUOTAON, Q_XQUOTAOFF,
Q_XQUOTARM, Q_XGETQSTAT, Q_XGETQSTATV, Q_XQUOTASYNC, Q_V1_GETSTATS, and
Q_V2_GETSTATS subcommands, do not print it for these subcommands.
* quota.c (decode_cmd_data): Add id argument parameter, add printing
of id argument for Q_GETQUOTA, Q_SETQUOTA, Q_GETNEXTQUOTA,
Q_V1_GETQUOTA, Q_V1_SETQUOTA, Q_V2_GETQUOTA, Q_V2_SETQUOTA,
Q_XGETQUOTA, Q_XGETNEXTQUOTA, Q_XSETQLIM, and unknown subcommands.
(SYS_FUNC(quotactl)): Remove printing of id argument, pass it
to decode_cmd_data.
* quota.c (SYS_FUNC(quotactl)): Since id call argument is used for
user/group/project ID which all have special semantics for -1,
print it the same way as UIDs are printed.
Introduce VERBOSE macro (defaults to 0) that is expected to be defined
to 1 by code testing "strace -v" output.
* tests/tests.h [!VERBOSE] (VERBOSE): New macro.
O_TMPFILE reqires the mode argument (just like O_CREAT), so print it.
* open.c (STRACE_O_TMPFILE): New macro.
(decode_open): Print the mode argument when O_TMPFILE flag is set.
* tests/open.c (main): Check it.
Fixes RH#1377846.
Add *_MAGIC constants defined for some relatively widespread
non-mainline filesystems.
* xlat/fsmagic.in: Add AUFS_SUPER_MAGIC, GPFS_SUPER_MAGIC,
VZFS_SUPER_MAGIC, and ZFS_SUPER_MAGIC constants.
Add fs *_MAGIC constants exported by linux uapi.
* xlat/fsmagic.in: Add BFS_MAGIC, GFS2_MAGIC, and ROMFS_MAGIC constants
defined in linux/bfs_fs.h, linux/gfs2_ondisk.h, and linux/romfs_fs.h,
respectively.
musl libc forwards semctl command argument for 8 known commands only,
for all the rest it passes 0 instead.
* tests/ipc_sem.c (main): Update semctl expected output.
* tests/ipc_msg.c: Additional msgget (parameter format) and msgctl
(parameter format, decoding of struct msqid_ds in IPC_SET/IPC_STAT
commands) checks.
* tests/ipc_sem.c: Additional semget and semctl checks.
* tests/ipc_shm.c: Additional shmget and shmctl checks.
* tests/semop.c: Additional semop checks. Add checks for semtimedop.
* tests/semop.test: Add explicit -e parameter in order to trace both
semop and semtimedop.
* tests/shmxt.c: Additional shmat and shmdt checks.
This is needed due to apparent deviation in spelling of commit author
name in several commits.
* .mailmap: Add canonical name for Eugene Syromyatnikov.
* linux/64/ioctls_inc.h: Update from linux v4.8 using ioctls_gen.sh.
* linux/32/ioctls_inc_align32.h: Likewise.
* linux/32/ioctls_inc_align64.h: Likewise.
According to POSIX, nsops argument of semop syscall has type size_t,
but the kernel treats nsops argument of semop and semtimedop syscalls
as unsigned int.
* ipc_sem.c (tprint_sembuf_array): Change type of "count" argument
from unsigned long to unsigned int, print it using %u format.
Also fix io_cancel and io_destroy checks which did not check correct
printing of context argument properly.
* tests/aio.c (main): Update syscall checks in order to use newly
defined bogus_ctx constant.
As io_setup syscall allocates some memory using do_mmap_pgoff, and
io_destroy deallocates this memory using vm_munmap, set TRACE_MEMORY
flag for all sysentries of io_setup and io_destroy using the following
oneliner:
sed -ri '/io_setup|io_destroy/ s/0,/TM,/' linux/*/syscallent*.h
* linux/*/syscallent*.h (io_setup, io_destroy): Change sys_flags to TM.
It was incorrectly assumed that __*64 types are long long on all
platforms, despite strace having specially crafted macros in order
to handle precisely this architecture discrepancy.
The commit fixes this oversight.
* tests/aio.c (main): Use PRI__*64 macros for correct format conversion
specifiers for __*64-typed values.
* ipc_msgctl.c (SYS_FUNC(msgctl)): As msqid argument is treated as int
by the kernel, cast it to int and print it using %d format.
* ipc_sem.c (SYS_FUNC(semop), SYS_FUNC(semtimedop)): Likewise,
for semid argument.
(SYS_FUNC(semget)): Likewise, for nsems argument.
(SYS_FUNC(semctl)): Likewise, for semid and semnum arguments.
* ipc_shm.c (SYS_FUNC(shmat)): Likewise, for shmid argument.
* ipc_shmctl.c (SYS_FUNC(shmctl)): Likewise.
* ipc_msg.c (SYS_FUNC(msgget)): As key_t type in the kernel
is __kernel_key_t (i.e. int), cast key_t argument to int
and print it using %#x format.
* ipc_sem.c (SYS_FUNC(semget)): Likewise.
* ipc_shm.c (SYS_FUNC(shmget)): Likewise.
* tests/ipc_msg.c (main): Test it.
* tests/ipc_sem.c (main): Likewise.
* tests/ipc_shm.c (main): Likewise.
Kernel headers declare this field as s32, and strace prints it with %d
specifier.
* tests/sched_xetattr.c (main): Change type of sched_nice field of struct
sched_attr to int32_t, update format specifiers accordingly.
New sprintrc_grep function is sprintrc function equivalent suitable for
tests where grep-base pattern matching is employed.
* tests/tests.h (sprintrc_grep): New prototype.
* tests/sprintrc.c (enum sprintrc_fmt): New sprintrc format enumeration.
(sprintrc_ex): New function, renamed from sprintrc and updated to
support different formats.
(sprintrc): Change to use sprintrc_ex with SPRINTRC_FMT_RAW.
(sprintrc_grep): New function, calls sprintrc_ex with SPRINTRC_FMT_GREP.
Use tail_alloc with precise size of the structure.
* tests/sched_xetattr.c (main): Eliminate usage of anonymous union type.
Rename sched to sched_attr. Change type of sched_attr to struct
pointer. Use tail_alloc for sched_attr allocation, update printf
statements accrodingly.
* tests/aio.c (main): Do not pass 64-bit aio_context_t to io_submit
and io_getevents until strace learns how to print 64-bit pointers on x32
and on x86_64 for x32 personality.
As aio_context_t is treated by the kernel as a pointer,
print it using printaddr.
* aio.c (SYS_FUNC(io_setup)): Print the pointer to aio_context_t
argument using printnum_ptr.
(SYS_FUNC(io_destroy), SYS_FUNC(io_submit), SYS_FUNC(io_cancel),
SYS_FUNC(io_getevents)): Print aio_context_t argument using printaddr.
* tests/aio.c (sprint_aio_context_t): Remove.
(main): Update expected output.
This helps to avoid warnings like
"integer constant is too large for ‘long’ type"
reported by some versions of gcc on 32-bit platforms.
* tests/aio.c (main): Add ULL suffix to 64-bit constants
and cast them to unsigned long.
* tests/sched_xetscheduler.c (main): Check for decoding of invalid PID
in sched_getscheduler and sched_setscheduler, invalid address
of sched_param structure, and invalid policy value.
struct iocb contains two fields with fd semantics: aio_fildes and
aio_resfd. It is quite reasonable to use the appropriate function for
printing them (apart from just "%d").
* aio.c (print_common_flags): Add struct tcb pointer to parameter list;
use printfd for printing aio_resfd field.
(print_iocb_header): Add struct tcb pointer to parameter list;
use printfd for printing aio_fildes field.
(print_iocb): Provide tcp argument to print_iocb_header
and print_common_flags.
(SYS_FUNC(io_cancel)): Likewise.
According to documentation and kernel's syscalls.h, its type is size_t,
so "%lu" format should be used instead of "%ld".
* readahead.c (SYS_FUNC(readahead)): Fix conversion specifier
for the "count" argument.
Use zero_extend_signed_to_ull and sign_extend_unsigned_to_ll macros
instead of explicit casts with unpredictable sign extension semantics.
* tests/xstatx.c (print_time, main): Use zero_extend_signed_to_ull
instead of explicit cast.
(print_stat): Use zero_extend_signed_to_ull and
sign_extend_unsigned_to_ll instead of explicit casts.
* xlat/sa_handler_values.in: New file.
* signal.c: Include "xlat/sa_handler_values.h".
(get_sa_handler_str, print_sa_handler): New functions.
(SYS_FUNC(sigsetmask), SYS_FUNC(signal), decode_new_sigaction): Use them.
* xlat/socksctpoptions.in: Add SCTP_PR_SUPPORTED, SCTP_DEFAULT_PRINFO,
and SCTP_PR_ASSOC_STATUS introduced by linux kernel commits
v4.8-rc1~140^2~148^2~5, v4.8-rc1~140^2~148^2~4, and
v4.8-rc1~140^2~148^2~3, respectively.
* xlat/bpf_map_types.in: Add BPF_MAP_TYPE_CGROUP_ARRAY introduced
by linux kernel commit v4.8-rc1~140^2~212^2~2.
* xlat/bpf_prog_types.in: Add BPF_PROG_TYPE_XDP introduced
by linux kernel commit v4.8-rc1~140^2~64^2~10.
The syscall parsers guarded by HAVE_SIGACTION check have to be compiled
regardless of libc sigaction function availability.
* configure.ac (AC_CHECK_FUNCS): Remove sigaction.
* signal.c: Remove HAVE_SIGACTION checks.
As configure.ac does not check for linux/futex.h, HAVE_LINUX_FUTEX_H is
never defined and therefore the inclusion of <linux/futex.h> guarded by
HAVE_LINUX_FUTEX_H makes no sense.
Moreover, <linux/futex.h> used to have an incorrect definition
of FUTEX_WAIT_BITSET_PRIVATE and FUTEX_WAKE_BITSET_PRIVATE:
since kernel commit v2.6.24-6320-gcd68998 where these definitions
were initially introduced and up to v2.6.31-7082-gf8d1e54 where they
were finally fixed these macros had been incorrectly defined via
FUTEX_WAIT_BITS and FUTEX_WAKE_BITS instead of FUTEX_WAIT_BITSET
and FUTEX_WAKE_BITSET, and these incorrect definitions made their way
into some distributions still in use.
* futex.c [HAVE_LINUX_FUTEX_H]: Remove.
This is analogous to timeout argument omitting in FUTEX_WAKE_BITSET
command.
* futex.c (SYS_FUNC(futex)): Remove common printing of val argument.
Add printing of val argument for all futex commands except
FUTEX_LOCK_PI, FUTEX_UNLOCK_PI, and FUTEX_TRYLOCK_PI.
* xlat/futexops.in: Add FUTEX_WAIT|FUTEX_CLOCK_REALTIME and
FUTEX_WAIT_PRIVATE|FUTEX_CLOCK_REALTIME values supported by linux kernel
since commit v4.5-rc1~172^2.
Use alternate form for printing hexadecimal numbers to avoid confusion.
Use printaddr to print uaddr as the latter is interpreted as a pointer
in all currently supported futex operations.
* futex.c (SYS_FUNC(futex)): Fix formatting of unknown command
argument.
Since obsolete FUTEX_FD command is known and used to have some expected
argument format, print FUTEX_FD using that format.
* futex.c (SYS_FUNC(futex)): Handle FUTEX_FD command.
According to the initial and current (v4.7) kernel implementations,
in FUTEX_WAKE_OP case the compare function does not have
FUTEX_OP_OPARG_SHIFT flag and occupies 4 bits starting with bit 24.
* futex.c (SYS_FUNC(futex)): Do not print FUTEX_OP_OPARG_SHIFT
for 27th bit of val3 in FUTEX_WAKE_OP case.
As ptrace.h already ensures that PTRACE_LISTEN is defined,
there is no need to check this fact in other places.
* strace.c (ptrace_restart): Do not check that PTRACE_LISTEN is defined.
SPARC has a different PTRACE_DETACH value correctly defined in
sys/ptrace.h, but linux/ptrace.h clobbers it with the standard one.
PTRACE_SUNDETACH is also defined to the correct value by sys/ptrace.h,
so use that instead.
* strace.c (detach) [SPARC]: Move redefinition of PTRACE_DETACH
to PTRACE_SUNDETACH ...
* ptrace.h [SPARC || SPARC64]: ... here.
glibc, starting with commit glibc-2.24-28-gdbab657, has deprecated
inclusion of <sys/sysmacros.h> by <sys/types.h>. The method used
for deprecation breaks AC_HEADER_MAJOR: this autoconf macro no longer
defines MAJOR_IN_SYSMACROS, which consequently breaks build.
Let's assume that all systems supported by strace provide major, minor,
and makedev macros via <sys/sysmacros.h>.
* configure.ac (AC_HEADER_MAJOR): Remove.
* mknod.c [MAJOR_IN_SYSMACROS, MAJOR_IN_MKDEV]: Remove.
Include <sys/sysmacros.h> unconditionally.
* print_struct_stat.c: Likewise.
* tests/mknod.c: Likewise.
* tests/mknodat.c: Likewise.
* tests/xstatx.c: Likewise.
This complements commit v4.11-719-gfb0c609.
* btrfs.c (struct_btrfs_ioctl_vol_args_v2): New type. Mpersify it.
(btrfs_print_qgroup_inherit): Change qgi_addr type to unsigned long.
(btrfs_ioctl): Replace struct btrfs_ioctl_vol_args_v2 with
struct_btrfs_ioctl_vol_args_v2.
This fixes mpersing of unions containing nameless members,
e.g. struct btrfs_ioctl_vol_args_v2.
* mpers.awk (what_is): Print names of union_type members as is.
Previous commit effectively changed types of st_atime, st_ctime, and
st_mtime members of struct stat and struct stat64 to signed integers,
making the mips64 workaround obsolete.
This reverts commit 3fb84bfc79.
* tests/xstatx.c (create_sample) [__mips64]: Remove.
This allows testing of MPERS_IS_* macros in pre-MPERS_DEFS parts
of source code.
* Makefile.am (mpers-m%.stamp): Add -DMPERS_IS_$(mpers_NAME) to CFLAGS
passed to mpers.sh.
* mpers_test.sh: Likewise.
* file.c: Move redefinition of types that might be used
to define struct stat ...
* linux/asm_stat.h: ... here.
* tests/xstatx.c: Remove redefinition of stat types.
For the reason described in commit v4.10-517-gcfde1e3, a correct
definition of struct stat for x32 personality is necessary to enable
"asm_stat.h" with -mx32 on x86_64.
* linux/x32/asm_stat.h: Rename to ...
* linux/x86_64/asm_stat.h: ... new file.
* Makefile.am (strace_SOURCES): Add it.
* linux/x32/asm_stat.h: New file, include "x86_64/asm_stat.h".
Assume that the check for struct stat.st_mtime_nsec is enough.
Likewise, do not check for struct stat.st_{a,c}tim.tv_nsec.
* configure.ac (AC_CHECK_MEMBERS): Remove struct stat.st_atime_nsec,
struct stat.st_ctime_nsec, struct stat.st_atim.tv_nsec,
and struct stat.st_ctim.tv_nsec.
* file.c (HAVE_STRUCT_STAT_ST_ATIME_NSEC,
HAVE_STRUCT_STAT_ST_CTIME_NSEC): Remove.
* printstat.h (HAVE_STRUCT_STAT_ST_ATIME_NSEC,
HAVE_STRUCT_STAT_ST_CTIME_NSEC): Replace with
HAVE_STRUCT_STAT_ST_MTIME_NSEC.
* tests/xstatx.c: Likewise.
For some reason, struct stat and struct stat64 are different on sparc64.
This change fixes decoding of struct stat64 related syscalls for sparc64
personality, sparc32 personality on sparc64 needs more work.
* file.c (printstat64) [SPARC64]: Do not use printstat.
(SYS_FUNC(newfstatat)): Likewise.
This is needed in order to differentiate it from the value returned
by the call. As a consequence, it enables revealing possible bugs
in syscall parsers, for example, when the value read on syscall
entering and not on syscall exiting, as it was the case with
timerfd_settime parser.
* tests/timer_xettime.c (main): Fill old.its field with value different
from the expected one upon call return.
* tests/timerfd_xettime.c: Likewise.
This is done similar to timer_settime syscall parser.
* time.c (SYS_FUNC(timerfd_settime)): Retrieve old_value argument
on exiting and not on entering. Return 0 instead of RVAL_DECODED
since the call hasn't been decoded in full on entering.
This change prevents scattering of ll-related hacks and simplifies
pread/pwrite syscalls parsers' logic a bit.
* util.c (getllval): Add fixup for arg_no for SuperH when argument
number is equal to 3.
* io.c (PREAD_OFFSET_ARG): Remove.
(SYS_FUNC(pread)): Always use argument number 3 for "count" argument
printing.
(SYS_FUNC(pwrite)): Likewise.
* linux/sparc/arch_sigreturn.c (arch_sigreturn): Parametrize
member types of struct signal_frame.
* linux/sparc64/arch_sigreturn.c (sparc64_arch_sigreturn,
sparc32_arch_sigreturn): New functions.
(arch_sigreturn): Use them.
On sparc64, unlike all other architectures where semctl is an indirect
ipc subcall, the forth argument is passed directly.
* ipc_sem.c (SYS_FUNC(semctl)) [SPARC64]: Print 4th argument without
indirection in case of native personality.
The structures defined in asm/termbits.h have the same size
on sparc and sparc64.
* linux/sparc64/ioctls_arch0.h (TCGETS, TCGETS2, TCSETS, TCSETS2,
TCSETSF, TCSETSF2, TCSETSW, TCSETSW2): Sync with
linux/sparc/ioctls_arch0.h
As st_flags, st_fstype, and st_gen members of struct stat are not filled
by the kernel, there is no use supporting them.
* configure.ac (AC_CHECK_MEMBERS): Remove struct stat.st_flags,
struct stat.st_fstype, and struct stat.st_gen.
* file.c [STAT32_PERSONALITY, HAVE_STRUCT_STAT64]: Do not undefine
HAVE_STRUCT_STAT_ST_FLAGS, HAVE_STRUCT_STAT_ST_FSTYPE,
and HAVE_STRUCT_STAT_ST_GEN.
* printstat.h (DO_PRINTSTAT): Do not check for
HAVE_STRUCT_STAT_ST_FLAGS, HAVE_STRUCT_STAT_ST_FSTYPE,
and HAVE_STRUCT_STAT_ST_GEN.
Our test suite already assumes that struct stat contains st_blksize,
st_blocks, and st_rdev members, and there haven't been any complaints.
* configure.ac (AC_CHECK_MEMBERS): Remove struct stat.st_blksize,
struct stat.st_blocks, and struct stat.st_rdev.
* printstat.h (DO_PRINTSTAT): Do not check for
HAVE_STRUCT_STAT_ST_BLKSIZE, HAVE_STRUCT_STAT_ST_BLOCKS,
and HAVE_STRUCT_STAT_ST_RDEV.
This fixes compilation warning that ARCH_PC_REG is redefined.
* linux/powerpc64/arch_regs.c (ARCH_PC_REG): Undefine before
the new definition.
* linux/sparc64/arch_regs.c (ARCH_PC_REG): Likewise.
* linux/s390x/arch_regs.c (ARCH_PC_REG): Remove, it is already defined
in just included linux/s390/arch_regs.c.
* linux/x32/arch_regs.c (ARCH_PC_REG): Remove, it is already defined
in just included linux/x86_64/arch_regs.c.
* init.sh (run_prog, run_prog_skip_if_failed): When the program fails,
add its exit code to the diagnostic message.
(run_strace): When strace fails, add its exit code to the diagnostic
message.
(run_strace_merge): When strace-log-merge fails, add its exit code
to the diagnostic message.
As a side effect of commit v4.11-211-g0736d4e, strace used to leak
placeholders for standard descriptors to tracees thus affecting their
behaviour. Fix this by setting close-on-exec flag on placeholder
descriptors.
* strace.c (open_dummy_desc): Set close-on-exec flag on the descriptor
that is going to be returned to the caller.
(fd_is_placeholder): New array.
(ensure_standard_fds_opened, redirect_standard_fds): New functions.
(startup_child): Use redirect_standard_fds.
(init): Use ensure_standard_fds_opened.
On some architectures including alpha, <asm/stat.h> provides a
definition of struct stat that has no st_atime_nsec, st_mtime_nsec, and
st_ctime_nsec fields. At the same time, struct stat64 always has these
fields. Fix tests to take this difference into account.
* tests/fstat64.c (STRUCT_STAT_IS_STAT64): New macro, defined to 1.
* tests/lstat64.c (STRUCT_STAT_IS_STAT64): Likewise.
* tests/stat64.c (STRUCT_STAT_IS_STAT64): Likewise.
* tests/xstatx.c [!STRUCT_STAT] (STRUCT_STAT_IS_STAT64): New macro,
defined to 0.
[USE_ASM_STAT && STRUCT_STAT_IS_STAT64]:
(HAVE_STRUCT_STAT_ST_ATIME_NSEC, HAVE_STRUCT_STAT_ST_CTIME_NSEC,
HAVE_STRUCT_STAT_ST_MTIME_NSEC): Redefine to 1.
On mips64 the F_GETLK and F_SETLKW64 constants have identical values which
causes the "wrong" constant to be printed by strace.
tests/fcntl.c (test_flock64): Do not test F_SETLKW64 on mips64.
The "syscall" syscall only exists on o32 and causes strace to error out on
64-bit mips ABIs. Pass MIPS_ABI from the configure script through to
nsyscalls.test so the MIPS ABI can be checked.
* configure.ac (MIPS_ABI): Substitute into output files.
* tests/Makefile.am (MIPS_ABI): Export via AM_TEST_LOG_FLAGS.
* tests/nsyscalls.test: Restrict special mips handling to mips o32.
For historical reasons the kernel struct stat represents times as unsigned
32-bit integers on mips64. Therefore, while it's possible to give a file a
timestamp before 1970 with futimens, reading the same timestamp through
struct stat will give a positive time (around 2106).
Workaround by using positive timestamps for testing on mips64.
* tests/xstatx.c (create_sample): Use positive timestamps on mips64.
* tests/mlock2.c (main): Pass unsigned long arguments to mlock2 syscall
explicitly, to avoid unwanted sign extension issues.
Based on patch by James Cowgill <james410@cowgill.org.uk>
Migrate to __NR_* the last user of SYS_* macros provided
by <sys/syscall.h>.
* clone.c [IA64] (ARG_STACKSIZE, ARG_PTID, ARG_CTID, ARG_TLS): Replace
SYS_clone2 with __NR_clone2.
Remove remains of solaris personality support.
This complements commit v4.10-45-gdf4dd8b.
* file.c [SPARC || SPARC64]: Remove the code related to struct solstat.
(printstat, printoldstat) [SPARC || SPARC64]: Remove.
When the syscall number returned by arch_get_scno is a mapped indirect
subcall (i.e. mapped subcall of socketcall or ipc syscall), do not
mistakenly treat it as a valid indirect subcall.
* defs.h (SCNO_IS_VALID): Treat scno with TRACE_INDIRECT_SUBCALL flag
as invalid.
* syscall.c (syscall_name): Do no shuffle scno.
(trace_syscall_entering, trace_syscall_exiting): Use
tcp->s_ent->sys_name instead of syscall_name.
(get_scno): In case of invalid syscall, allocate a dynamic struct sysent
containing an appropriate .sys_name.
* tests/nsyscalls.c (main) [SYS_socket_subcall]: Check decoding
of direct syscall number SYS_socket_subcall+1.
(main) [SYS_ipc_subcall]: Check decoding of direct syscall number
SYS_ipc_subcall+1.
Fix inconsistency between syscall and ioctl entries on sparc64.
Make layout of personalities on sparc64 the same as on other
architectures that support two personalities.
* defs.h [SPARC64] (PERSONALITY0_WORDSIZE): Change to 8.
[SPARC64] (PERSONALITY1_WORDSIZE): Change to 4.
* linux/sparc64/get_scno.c (arch_get_scno): Swap personality numbers.
* linux/sparc64/ioctls_arch1.h: Rename to ioctls_arch0.h.
* linux/sparc64/ioctls_arch0.h: Rename to ioctls_arch1.h.
* linux/sparc64/ioctls_inc0.h: Rename to ioctls_inc1.h.
* linux/sparc64/ioctls_inc1.h: Rename to ioctls_inc0.h.
This commit is an attempt to unify usage of include guards (in top-level
headers, at least). As a side note, different files with *.h extension
have different semantics: for example, printargs.h is included multiple
times in order to generate slightly varying code depending on values of
macro definitions - maybe it's better to change extension of such files
to something like *.inc.
* defs.h: Add #include guard.
* flock.h: Likewise.
* ipc_defs.h: Likewise.
* mpers_type.h: Likewise.
* printsiginfo.h: Likewise.
* ptrace.h: Likewise.
* regs.h: Likewise.
* seccomp_fprog.h: Likewise.
* gcc_compat.h: Rename the macro used for #include guard.
* msghdr.h: Likewise.
* sigevent.h: Likewise.
* kernel_types.h: Comment the trailing part of #include guard.
* xlat.h: Add missing macro definition for #include guard.
There seems to be no readily available binary packages of musl that are
sufficiently up to date to be used to build and test strace, so prepare
a suitable musl from source.
* travis-install.sh (musl-gcc): Do not add an extra repository,
do not install musl-tools and linux-musl-dev packages.
Download, build, and install a suitable musl revision instead.
* tests/attach-f-p.c Simplify communications between threads
by replacing per-thread timers with pipes.
* tests/attach-f-p.test: Do not ignore signals.
* tests/signalfd4.c (get_sigset_size): New function.
(main): Use it. Print expected output.
* tests/signalfd4.test: Use run_strace_match_diff.
* tests/signalfd4.expected: Remove.
* tests/Makefile.am (EXTRA_DIST): Remove it.
* strace.c (ptrace_attach_cmd): New variable.
(ptrace_attach_or_seize): Save last ptrace command's name
to ptrace_attach_cmd.
(attach_tcb, startup_child): Use it in error diagnostics.
This script has been created by commit v4.8-180-g0ed617b
for the one-shot job of automatic convertion of xlat structures
from *.c files to xlat/*.in files.
* generate_xlat_in.sh: Remove.
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 - the linux syscall tracer
=================================
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.
This is [strace](https://strace.io) -- a diagnostic, debugging and instructional userspace utility with a traditional command-line interface for Linux. It is used to monitor and tamper with 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).
See the file [NEWS](https://raw.githubusercontent.com/strace/strace/master/NEWS) for information on what has changed in recent versions.
strace is released under a Berkeley-style license at the request of Paul Kranenburg; see the file [COPYING](COPYING) for details.
Please send bug reports and enhancements to [the strace mailing list](https://lists.sourceforge.net/lists/listinfo/strace-devel).
See the file [NEWS](NEWS) for information on what has changed in recent versions.
Please read the file [INSTALL-git](INSTALL-git.md) for installation instructions.
The user discussion and development of strace take place on [the strace mailing list](https://lists.sourceforge.net/lists/listinfo/strace-devel) -- everyone is welcome to post bug reports, feature requests, comments and patches to strace-devel@lists.sourceforge.net. The mailing list archives are available at https://sourceforge.net/p/strace/mailman/strace-devel/ and other archival sites.
list="$(sed -n '/^strace_SOURCES[[:space:]]*=/,/^[[:space:]]*# end of strace_SOURCES/ s/^[[:space:]]*\([[:alnum:]][^.]*\.c\)[[:space:]]*\\$/\1/p' Makefile.am |
list="$(sed -r -n '/^strace_SOURCES[[:space:]]*=/,/^[[:space:]]*# end of strace_SOURCES/ s/^[[:space:]]*([[:alnum:]][^.]*\.c)[[:space:]]*\\$/\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.
