Zack Weinberg [Sun, 10 Mar 2019 02:18:56 +0000 (21:18 -0500)]
Don’t conditionalize declarations of ldiv_t, lldiv_t, __gwchar_t.
The declarations of ldiv_t, lldiv_t, and __gwchar_t are all wrapped in
__foo_t_defined conditionals, despite there being only one header in
glibc that declares them. I checked codesearch.debian.net and only
found one other program that refers to the __foo_t_defined
conditionals: u-boot provides its own definition of ldiv_t if
__ldiv_t_defined is not defined by stdlib.h. I conclude that the
conditionals are not necessary, but the definitions of
__ldiv_t_defined and __lldiv_t_defined should be preserved.
* stdlib/inttypes.h: Unconditionally define __gwchar_t.
Don’t define ____gwchar_t_defined.
* stdlib/stdlib.h: Unconditionally define ldiv_t.
Condition lldiv_t only on __USE_ISOC99.
Zack Weinberg [Sun, 10 Mar 2019 01:54:02 +0000 (20:54 -0500)]
Create bits/types headers for most remaining __T_defined macros.
This doesn't exactly fit the theme but as long as I'm tinkering with
sys/types.h it makes sense to go through and create single-declaration
bits/types/ headers for all of the remaining cases where we have
two or more headers declaring a public type.
The remaining uses of the original __T_defined idiom are:
__error_t_defined in files shared with gnulib, which probably has to
remain as is; ____gwchar_t_defined in inttypes.h, which may not be
necessary anymore and should be addressed separately, and
__ldiv_t_defined and __lldiv_t_defined in stdlib.h, ditto.
Our handling of LFS types is a little inconsistent: some headers
declare both off_t and off64_t (for instance) when
_LARGEFILE64_SOURCE, others will only declare off_t regardless of
_LARGEFILE64_SOURCE. I don't know if this was intentional or not.
I am tempted to centralize responsibility for _LARGEFILE64_SOURCE as
well as _FILE_OFFSET_BITS=64 in bits/types/off_t.h (etc) so that any
header that declares off_t will automatically also declare off64_t
when _LARGEFILE64_SOURCE.
sunrpc/rpc/types.h is special, because it is included in files
compiled by the *build* compiler (cross-rpcgen-objs), and therefore it
cannot unconditionally assume bits/types headers are available. What
I did was have it include the appropriate bits/types headers only if
including sys/types.h did not cause __BIT_TYPES_DEFINED__ to be
defined. This will do the right thing when an installed rpc/types.h
is included by application code compiled without __USE_MISC in effect.
During the build, we rely on the fact that we compile all of our own
code with __USE_MISC in effect. This is fragile, but should be
acceptable for code that's no longer built by default anyway.
scripts/check-obsolete-constructs.py ensures that only sys/types.h and
rpc/types.h include the bits/types/ headers that define obsolete types.
* posix/bits/types/blkcnt64_t.h
* posix/bits/types/blkcnt_t.h
* posix/bits/types/blksize_t.h
* posix/bits/types/dev_t.h
* posix/bits/types/fsblkcnt64_t.h
* posix/bits/types/fsblkcnt_t.h
* posix/bits/types/fsfilcnt64_t.h
* posix/bits/types/fsfilcnt_t.h
* posix/bits/types/fsid_t.h
* posix/bits/types/gid_t.h
* posix/bits/types/id_t.h
* posix/bits/types/ino64_t.h
* posix/bits/types/ino_t.h
* posix/bits/types/intptr_t.h
* posix/bits/types/key_t.h
* posix/bits/types/loff_t.h
* posix/bits/types/mode_t.h
* posix/bits/types/nlink_t.h
* posix/bits/types/off64_t.h
* posix/bits/types/off_t.h
* posix/bits/types/pid_t.h
* posix/bits/types/socklen_t.h
* posix/bits/types/ssize_t.h
* posix/bits/types/suseconds_t.h
* posix/bits/types/uid_t.h
* posix/bits/types/useconds_t.h:
New single-declaration headers for standard types canonically
defined by sys/types.h, sys/socket.h, or inttypes.h but also
exposed by other headers under some circumstances. Code moved
from posix/sys/types.h, socket/sys/socket.h, stdlib/inttypes.h
as appropriate.
* posix/bits/types/uint.h
* posix/bits/types/u_int.h
* posix/bits/types/u_intN_t.h
* posix/bits/types/caddr_t.h
* posix/bits/types/daddr_t.h
* posix/bits/types/loff_t.h
* posix/bits/types/register_t.h:
Similarly, but for obsolete BSD-derived types whose canonical
home is sys/types.h. Some of these headers define more than
one type.
* posix/Makefile (headers): Install the above new headers.
Rewrap the list.
* posix/sys/types.h: All definitions of public types now
accomplished using the above new headers. Consolidate
groups of definitions controlled by the same feature
selection macros.
* inet/arpa/inet.h, bits/socket.h
* sysdeps/mach/hurd/bits/socket.h
* sysdeps/unix/sysv/linux/bits/socket.h:
Use bits/types/socklen_t.h.
* dirent/dirent.h: Use bits/types/ino_t.h and bits/types/ino64_t.h.
* grp/grp.h: Use bits/types/gid_t.h.
* io/fcntl.h: Use bits/types/mode_t.h, bits/types/off_t.h,
bits/types/pid_t.h, and bits/types/off64_t.h.
* io/sys/stat.h: Use bits/types/dev_t.h, bits/types/gid_t.h,
bits/types/ino_t.h, bits/types/mode_t.h, bits/types/nlink_t.h,
bits/types/off_t.h, bits/types/uid_t.h, and bits/types/blkcnt_t.h.
* libio/stdio.h: Use bits/types/off_t.h, bits/types/off64_t.h,
and bits/types/ssize_t.h.
* misc/sys/mman.h: Use bits/types/off_t.h and bits/types/mode_t.h.
* misc/sys/select.h: Use bits/types/suseconds_t.h.
* posix/sched.h: Use bits/types/pid_t.h.
* posix/sys/wait.h: Use bits/types/pid_t.h and bits/types/id_t.h.
* posix/unistd.h: Use bits/types/gid_t.h, bits/types/uid_t.h,
bits/types/off_t.h, bits/types/off64_t.h, bits/types/useconds_t.h,
bits/types/intptr_t.h, and bits/types/socklen_t.h.
* pwd/pwd.h: Use bits/types/gid_t.h and bits/types/uid_t.h.
* resource/sys/resource.h: Use bits/types/id_t.h.
* signal/signal.h: Use bits/types/pid_t.h and bits/types/uid_t.h.
* stdlib/monetary.h: Use bits/types/ssize_t.h.
* sysdeps/gnu/utmpx.h: Use bits/types/pid_t.h.
* sysvipc/sys/ipc.h: Use bits/types/uid_t.h, bits/types/gid_t.h,
bits/types/mode_t.h, and bits/types/key_t.h.
* sysvipc/sys/msg.h: Use bits/types/pid_t.h and bits/types/ssize_t.h.
* sysvipc/sys/shm.h: Use bits/types/pid_t.h.
* termios/termios.h: Use bits/types/pid_t.h.
* time/sys/time.h: Use bits/types/suseconds_t.h.
* time/time.h: Use bits/types/pid_t.h.
* sunrpc/rpc/types.h: Consolidate all #includes at the top of
the file. If __BIT_TYPES_DEFINED__ is not defined after
including sys/types.h, also include bits/types/caddr_t.h,
bits/types/daddr_t.h, bits/types/fsid_t.h, and bits/types/u_int.h.
* scripts/check-obsolete-constructs.py (OBSOLETE_TYPE_HDR_RE_): New.
(ObsoleteIndirectDefinitionsAllowed): New; allows inclusion of
bits/types/ headers that define obsolete typedefs, but not
direct definitions of those types.
(ObsoleteNotAllowed, ObsoletePrivateDefinitionsAllowed)
(ObsoletePublicDefinitionsAllowed): Do not allow inclusion of
bits/types/ headers that define obsolete typedefs.
Zack Weinberg [Sat, 9 Mar 2019 23:01:48 +0000 (18:01 -0500)]
Clean up bits/types.h.
This makes four linked changes to bits/types.h. First, we use
__(u)?int(16|32|64)_t to define __[SU](16|32|64)_TYPE. In addition
to reducing the amount of ifdeffage, this means __STD_TYPE is no longer
necessary, since gcc -std=c89 will complain about ‘typedef long long foo_t’
but not ‘typedef __int64_t foo_t’, even if the underlying type
of __int64_t is long long.
Second, we eliminate __UQUAD_TYPE and __SQUAD_TYPE from the set of
macros bits/typesizes.h should use to define __FOO_T_TYPE macros,
since they are always the same as __U64_TYPE and __S64_TYPE
respectively.
Third, we remove __u_char, __u_short, __u_int, __u_long, __u_quad_t,
and __quad_t, we add __uintptr_t, and we define __intmax_t and
__uintmax_t as __int64_t and __uint64_t.
Fourth, we reorganize the list of typedefs into groups by the
standard (if any) that defines them, and sort them alphabetically within
each group.
* posix/bits/types.h: Move #error for __WORDSIZE neither 32 nor 64
to first group of conditionals on __WORDSIZE, and make it more
explicit. Update commentary. Define all __foo_t types with
regular ‘typedef’. Reorganize all __foo_t types into the same
groups that sys/types.h uses.
(__u_char, __u_short, __u_int, __u_long, __quad_t, __u_quad_t)
(__UQUAD_TYPE, __SQUAD_TYPE, __STD_TYPE): Don’t define.
(__S16_TYPE): Define unconditionally as __int16_t.
(__U16_TYPE): Define unconditionally as __uint16_t.
(__S32_TYPE): Define unconditionally as __int32_t.
(__U32_TYPE): Define unconditionally as __uint32_t.
(__S64_TYPE): Define unconditionally as __int64_t.
(__U64_TYPE): Define unconditionally as __uint64_t.
(__intmax_t): Define unconditionally as __int64_t.
(__uintmax_t): Define unconditionally as __uint64_t.
(__uintptr_t): New typedef.
* bits/time64.h
* bits/typesizes.h
* sysdeps/mach/hurd/bits/typesizes.h
* sysdeps/unix/sysv/linux/alpha/bits/typesizes.h
* sysdeps/unix/sysv/linux/generic/bits/typesizes.h
* sysdeps/unix/sysv/linux/mips/bits/typesizes.h
* sysdeps/unix/sysv/linux/s390/bits/typesizes.h
* sysdeps/unix/sysv/linux/sparc/bits/typesizes.h
* sysdeps/unix/sysv/linux/x86/bits/typesizes.h:
Replace all uses of __UQUAD_TYPE with __U64_TYPE, and all
uses of __SQUAD_TYPE with __S64_TYPE.
* posix/sys/types.h, rpc/sys/types.h
(u_char): Define as unsigned char.
(u_short): Define as unsigned short.
(u_int): Define as unsigned int.
(u_long): Define as unsigned long.
(quad_t): Define as __int64_t.
(u_quad_t): Define as __uint64_t.
* stdlib/stdint.h (intptr_t): Define as __intptr_t.
(uintptr_t): Define as __uintptr_t.
* scripts/check-obsolete-constructs.py: Update allowed
definitions for the obsolete types. No longer allow
__STD_TYPE as an alias for typedef.
Zack Weinberg [Sat, 9 Mar 2019 21:01:38 +0000 (16:01 -0500)]
Simplify definition of __time64_t.
bits/time64.h always sets __TIME64_T_TYPE to be __time_t when
__TIMESIZE == 64, so we can unconditionally use __TIME64_T_TYPE to
define __time64_t; we don’t need another conditional in bits/types.h.
Also move the definition of __time64_t next to the definition of
__time_t.
* posix/bits/types.h (__time64_t): Unconditionally define
as __TIME64_T_TYPE. Move definition next to __time_t.
Zack Weinberg [Sat, 9 Mar 2019 16:59:02 +0000 (11:59 -0500)]
Add caddr_t, daddr_t, and loff_t to the set of obsolete typedefs.
caddr_t is a BSD-derived alias for ‘char *’, obsoleted by the
introduction of ‘void *’ in C89 (!) daddr_t is a “disk address,”
but it’s always defined as ‘int’, making it too small for modern
disks and tapes. loff_t is another name for off64_t, from early
drafts of LFS. All three are already only exposed by sys/types.h
under __USE_MISC.
This patch adds them to the set of types that shall not be used in
installed headers (enforced by check-obsolete-constructs.py) and
expunges all remaining uses, internally as well as in installed
headers. Since __DADDR_T_TYPE is always defined as __S32_TYPE, and
daddr_t is obsolete so there’s no need to worry about future
variation, the patch also removes __DADDR_T_TYPE from the set of
macros that bits/typesizes.h is required to define. Instead
bits/types.h always defines __daddr_t as __S32_TYPE, and the
definition is moved to a more logical location within the file, next
to __caddr_t.
It’s always safe to change (__)loff_t to the matching (__)off64_t;
in a few internal files, I removed an unnecessary __ prefix.
daddr_t is only used for struct ustat, which is obsoleted by struct
statvfs and we already don’t declare it in public headers, and for an
ioctl parameter block in sys/mtio.h (which may or may not be obsolete,
I can’t tell). In sys/mtio.h I replaced both uses with ‘int’ to match
the use of bare ‘long int’ for most of the other fields of that
structure. In misc/ustat.c, the definition of struct ustat is not
actually necessary so I removed it entirely. In
sysdeps/unix/sysv/linux/ustat.c a definition is necessary but only
because INLINE_SYSCALL_CALL doesn’t work (on at least x86) when an
argument is a pointer to an incomplete type, so I substituted a dummy
definition.
Most of the internal uses of caddr_t are in the sunrpc and nis
directories, and since most of that code is obsolete, I mechanically
replaced them with char * rather than consider whether void * might
make more sense. Because “const caddr_t foo” is semantically
different from “const char *foo” (in the first case ‘foo’ itself is
const but the memory pointed to isn’t, in the second case the memory
pointed to is const but ‘foo’ isn’t) this change exposed some
const-correctness errors in sunrpc, which I fixed minimally. Outside
of sunrpc and nis, I put a little more thought into whether uses of
caddr_t should be void * instead.
* scripts/check-obsolete-constructs.py: Add caddr_t, daddr_t,
and loff_t to the set of obsolete types forbidden in public
headers.
* posix/bits/types.h: Unconditionally define __daddr_t as
__S32_TYPE. Move definition of __daddr_t next to definition
of __caddr_t.
* bits/typesizes.h
* sysdeps/mach/hurd/bits/typesizes.h
* sysdeps/unix/sysv/linux/alpha/bits/typesizes.h
* sysdeps/unix/sysv/linux/generic/bits/typesizes.h
* sysdeps/unix/sysv/linux/mips/bits/typesizes.h
* sysdeps/unix/sysv/linux/s390/bits/typesizes.h
* sysdeps/unix/sysv/linux/sparc/bits/typesizes.h
* sysdeps/unix/sysv/linux/x86/bits/typesizes.h:
Don’t define __DADDR_T_TYPE.
* sysdeps/gnu/sys/mtio.h (struct mtget): Change all uses of
__daddr_t to int.
* misc/ustat.c: Remove definition of struct ustat; only
forward-declare it.
* sysdeps/unix/sysv/linux/ustat.c: Replace fields of
struct ustat with a size-preserving dummy field.
* hurd/Makefile (migheaderpipe): Rewrite loff_t as __off64_t.
* hurd/fd-read.c (_hurd_fd_read): Use off64_t instead of loff_t.
* hurd/fd-write.c (hurd_fd_write): Use off64_t instead of loff_t.
* hurd/hurd/fd.h (_hurd_fd_read, _hurd_fd_write): Declare
using __off64_t instead of __loff_t.
* support/xunistd.h (xcopy_file_range): Declare using off64_t
instead of loff_t.
* sysdeps/unix/sysv/linux/generic/wordsize-32/overflow.h
* sysdeps/unix/sysv/linux/lseek.c
* sysdeps/unix/sysv/linux/lseek64.c
* sysdeps/unix/sysv/linux/splice.c
Throughout, use off64_t instead of loff_t.
* sysdeps/unix/sysv/linux/sys/quota.h
(dqoff): Use __off64_t instead of __loff_t.
(quotactl): Declare using char * instead of caddr_t.
* sysdeps/unix/sysv/linux/test-errno-linux.c
(do_test): Cast to char * instead of caddr_t when calling quotactl.
* elf/dl-map-segments.h (_dl_map_segments): Cast to void *
instead of caddr_t when calling __mprotect and __mmap.
* elf/dl-minimal.c (malloc): Declare page as char *, not caddr_t.
* elf/dl-reloc.c (_dl_relocate_object): Declare textrels.start
as char *, not caddr_t. Cast to char *, not caddr_t, in
pointer arithmetic.
* intl/loadmsgcat.c: Remove two unnecessary casts to caddr_t
when calling munmap. Change a third cast to target void *
instead and add a comment explaining why this one is necessary.
* locale/loadlocale.c (_nl_load_locale): Use NULL instead of
`(caddr_t)0`, and remove an unnecessary cast to caddr_t when
calling munmap.
(_nl_unload_locale): Change casts when calling free and munmap
to target char *, and add a comment explaining why they are
necessary.
* sysdeps/gnu/net/if.h
(struct ifreq): Declare ifru_data as char *, not __caddr_t.
(struct ifconf): Declare ifcu_buf as char *, not __caddr_t.
* sunrpc/xdr_mem.c (xdrmem_create): Cast away const when
setting xdrs->x_private and xdrs->x_base.
* sunrpc/xdr_stdio.c (xdrstdio_getbytes): Correct argument
types in definition to match prototype.
Zack Weinberg [Tue, 19 Feb 2019 02:00:34 +0000 (21:00 -0500)]
sys/types.h: Don’t define u_intN_t or register_t unless __USE_MISC.
sys/types.h unconditionally defines u_int8_t, u_int16_t, u_int32_t,
u_int64_t, and register_t. These are not part of any standard. The
u_intXX_t types are superseded by C99’s uintXX_t types (defined in
stdint.h). I’m not aware of a standardized exact equivalent of
register_t, but also I’ve never seen anyone use it for anything.
I could be persuaded to leave that one alone.
sys/types.h also unconditionally defines int8_t, int16_t, int32_t, and
int64_t, which are the same as the C99 exact-width signed types in
stdint.h. POSIX doesn’t require these to appear in sys/types.h, so in
principle they ought to be brought under __USE_MISC also. But, when I
tried that it broke about two dozen files just in our own source tree,
and POSIX doesn’t *forbid* sys/types.h to define these types, so I
think we should leave them alone.
* posix/sys/types.h (u_int8_t, u_int16_t, u_int32_t, u_int64_t)
(register_t): Move under #ifdef __USE_MISC.
Consolidate adjacent #ifdef __USE_MISC blocks.
* scripts/check_obsolete_constructs.py: Add register_t to the
set of obsolete typedefs that our headers should not use
(but sys/types.h may still define).
Zack Weinberg [Tue, 19 Feb 2019 13:45:22 +0000 (08:45 -0500)]
Define register_t using bits/typesizes.h macros.
Currently register_t is, unlike all other types in sys/types.h,
defined using a GCC extension (__attribute__((mode(word)))), falling
back to ‘int’ if the extension is unavailable. This is a potential
ABI compatibility hazard for people using non-GNU compilers with
glibc. It’s also unnecessary; the bits/typesizes.h mechanism can
handle all of the existing variation in the definition. In most
cases, defining __REGISTER_T_TYPE as __SWORD_TYPE is sufficient.
Special handling is necessary for MIPS n32 and x86-64 x32, where
__SWORD_TYPE is ‘int’ and the appropriate type for register_t is
‘long long’. Unfortunately, this means we need to create a new
bits/typesizes.h variant for linux/mips. This variant is based
on the top-level bits/typesizes.h, not linux/generic/bits/typesizes.h,
to match the existing MIPS ABIs.
Tested using build-many-glibcs. The c++-types test confirms that the
physical type of register_t does not change on any supported platform.
* posix/sys/types.h: Typedef register_t as __register_t.
* posix/bits/types.h: Typedef __register_t using __REGISTER_T_TYPE.
* sysdeps/unix/sysv/linux/mips/bits/typesizes.h:
New file (copied from bits/typesizes.h).
Define __REGISTER_T_TYPE as __SWORD_TYPE for o32 and n64 ABIs.
Define __REGISTER_T_TYPE as __SQUAD_TYPE for n32.
* sysdeps/unix/sysv/linux/x86/bits/typesizes.h:
Define __REGISTER_T_TYPE as __SWORD_TYPE for o32 and 64-bit ABIs.
Define __REGISTER_T_TYPE as __SQUAD_TYPE for x32.
Zack Weinberg [Sun, 24 Feb 2019 00:08:54 +0000 (19:08 -0500)]
Move most headers installed by top-level Makefile to misc/.
The Makefile glue to run tests on installed headers is currently
duplicated between the top-level Makefile and Rules, because the
top-level Makefile doesn't read Rules but does install some headers.
This patch moves most of the headers installed by the top-level Makefile
to misc/ (chosen arbitrarily; I'm open to putting them somewhere else
if reviewers feel it would be better) and removes the duplicated logic
from the top-level Makefile. I believe this also means that none of
the headers in include/ are installed headers anymore.
gnu/lib-names*.h are still generated by code in Makerules and installed
by the top-level Makefile. I tried to move them to misc/ as well, but
that broke the generation process in a manner that didn't make any sense
so I gave up. The tests performed on installed headers are not
especially useful for gnu/lib-names*.h so I think we can live with this
for now.
* include/bits/xopen_lim.h
* include/features.h
* include/gnu-versions.h
* include/gnu/libc-version.h
* include/limits.h
* include/stdc-predef.h
* include/values.h:
Move to misc/ and replace with a trivial wrapper.
* Makefile (headers): Remove all headers moved to misc/.
Don't set a vpath for %.h.
(check-installed-headers-c.out, check-installed-headers-cxx.out)
(check-wrapper-headers.out): Remove rules.
* misc/Makefile (headers): Add all of the above headers and
sort the list.
* sysdeps/mach/hurd/bits/errno.h: Regenerate.
Samuel Thibault [Sat, 4 Jan 2020 18:37:53 +0000 (19:37 +0100)]
htl: Add __errno_location and __h_errno_location
As explained on
https://sourceware.org/ml/libc-alpha/2020-01/msg00049.html
the presence of __errno_location in libpthread.so on GNU/Linux makes
libpthread getting linked in for libstdc++. This aligns on that behavior, to
avoid issues that only GNU/Hurd would get.
Samuel Thibault [Sat, 4 Jan 2020 17:53:23 +0000 (18:53 +0100)]
htl: Move pthread_atfork to libc_nonshared.a
This follows bd60ce86520b ('nptl: Move pthread_atfork to libc_nonshared.a')
with the same rationale: there is no non-libpthread equivalent to be used
for making linking against libpthread optional.
libpthread_nonshared.a is unused after this, so remove it from the
build.
There is no ABI impact because pthread_atfork was implemented using
__register_atfork in libc even before this change.
pthread_atfork has to be a weak alias because pthread_* names are not
reserved in libc.
This patch avoid probing the __NR_clock_getttime64 syscall each time
__clock_gettime64 is issued on a kernel without 64 bit time support.
Once ENOSYS is obtained, only 32-bit clock_gettime are used.
No architecture currently defines the vDSO symbol. On archictures
with 64-bit time_t the HAVE_CLOCK_GETRES_VSYSCALL is renamed to
HAVE_CLOCK_GETRES64_VSYSCALL, it simplifies clock_gettime code.
This patch avoid probing the __NR_clock_getttime64 syscall each time
__clock_gettime64 is issued on a kernel without 64 bit time support.
Once ENOSYS is obtained, only 32-bit clock_gettime are used.
No architecture currently defines the vDSO symbol. On architectures
with 64-bit time_t the HAVE_CLOCK_GETTIME_VSYSCALL is renamed to
HAVE_CLOCK_GETTIME64_VSYSCALL, it simplifies clock_gettime code.
This patch moves the vDSO setup from libc to loader code, just after
the vDSO link_map setup. For static case the initialization
is moved to _dl_non_dynamic_init instead.
Instead of using the mangled pointer, the vDSO data is set as
attribute_relro (on _rtld_global_ro for shared or _dl_vdso_* for
static). It is read-only even with partial relro.
It fixes BZ#24967 now that the vDSO pointer is setup earlier than
malloc interposition is called.
Also, vDSO calls should not be a problem for static dlopen as
indicated by BZ#20802. The vDSO pointer would be zero-initialized
and the syscall will be issued instead.
Checked on x86_64-linux-gnu, i686-linux-gnu, aarch64-linux-gnu,
arm-linux-gnueabihf, powerpc64le-linux-gnu, powerpc64-linux-gnu,
powerpc-linux-gnu, s390x-linux-gnu, sparc64-linux-gnu, and
sparcv9-linux-gnu. I also run some tests on mips.
The code is similar to the one at elf/dl-reloc.c, where it checks for
the l_relro_size from the link_map (obtained from PT_GNU_RELRO header
from program headers) and calls_dl_protected_relro.
For testing I will use the ones proposed by Florian's patch
'elf: Add tests for working RELRO protection' [1].
Checked on x86_64-linux-gnu, i686-linux-gnu, powerpc64le-linux-gnu,
aarch64-linux-gnu, s390x-linux-gnu, and sparc64-linux-gnu. I also
check with --enable-static pie on x86_64-linux-gnu, i686-linux-gnu,
and aarch64-linux-gnu which seems the only architectures where
static PIE is actually working (as per 9d7a3741c9e, on
arm-linux-gnueabihf, powerpc64{le}-linux-gnu, and s390x-linux-gnu
I am seeing runtime issues not related to my patch).
The IFUNC bypass to vDSO is used when USE_IFUNC_TIME is set.
Currently powerpc and x86 defines it. Otherwise the generic
implementation is used, which calls clock_gettime.
Checked on powerpc64le-linux-gnu, powerpc64-linux-gnu,
powerpc-linux-gnu-power4, x86_64-linux-gnu, and i686-linux-gnu.
The IFUNC bypass to vDSO is used when USE_IFUNC_GETTIMEOFDAY is set.
Currently aarch64, powerpc*, and x86 defines it. Otherwise the
generic implementation is used, which calls clock_gettime.
Checked on aarch64-linux-gnu, powerpc64le-linux-gnu,
powerpc64-linux-gnu, powerpc-linux-gnu-power4, x86_64-linux-gnu,
and i686-linux-gnu.
- It requires sync the auto-generated C file with current glibc
implementation;
- It still uses symbol redirections hacks where libc-symbols.h
provide macros that uses compiler builtins
(libc_ifunc_redirected for instance);
- It does not handle all required compiler handling
(inhibit_stack_protector on iFUNC resolver).
- No architecure uses it.
This is the only use of auto-generation syscall which uses a vDSO
plus IFUNC and the current x86 generic implementation already covers
the expected semantic.
linux: Fix vDSO macros build with time64 interfaces
As indicated on libc-help [1] the ec138c67cb commit broke 32-bit
builds when configured with --enable-kernel=5.1 or higher. The
scenario 10 from [2] might also occur in this configuration and
INLINE_VSYSCALL will try to use the vDSO symbol and
HAVE_CLOCK_GETTIME64_VSYSCALL does not set HAVE_VSYSCALL prior its
usage.
Also, there is no easy way to just enable the code to use one
vDSO symbol since the macro INLINE_VSYSCALL is redefined if
HAVE_VSYSCALL is set.
Instead of adding more pre-processor handling and making the code
even more convoluted, this patch removes the requirement of defining
HAVE_VSYSCALL before including sysdep-vdso.h to enable vDSO usage.
The INLINE_VSYSCALL is now expected to be issued inside a
HAVE_*_VSYSCALL check, since it will try to use the internal vDSO
pointers.
Both clock_getres and clock_gettime vDSO code for time64_t were
removed since there is no vDSO setup code for the symbol (an
architecture can not set HAVE_CLOCK_GETTIME64_VSYSCALL).
Checked on i686-linux-gnu (default and with --enable-kernel=5.1),
x86_64-linux-gnu, aarch64-linux-gnu, and powerpc64le-linux-gnu.
I also checked against a build to mips64-linux-gnu and
sparc64-linux-gnu.
The result of INTERNAL_SYSCALL_CANCEL should be checked with
macros INTERNAL_SYSCALL_ERROR_P and INTERNAL_SYSCALL_ERRNO instead
of comparing the result directly.
This patch adds a new macro, libm_alias_finite, to define all _finite
symbol. It sets all _finite symbol as compat symbol based on its first
version (obtained from the definition at built generated first-versions.h).
The <fn>f128_finite symbols were introduced in GLIBC 2.26 and so need
special treatment in code that is shared between long double and float128.
It is done by adding a list, similar to internal symbol redifinition,
on sysdeps/ieee754/float128/float128_private.h.
Alpha also needs some tricky changes to ensure we still emit 2 compat
symbols for sqrt(f).
Rafał Lużyński [Mon, 30 Dec 2019 10:58:18 +0000 (11:58 +0100)]
Multiple locales: Add date_fmt (bug 24054)
It is not specified what should be the content of d_t_fmt and date_fmt
but in the built-in C locale those fields have only one difference:
date_fmt contains "%Z" (the current time zone) while d_t_fmt does not.
For most of the locales this commit does the following operation:
copy d_t_fmt to date_fmt, and then remove "%Z" from d_t_fmt.
If "%Z" was originally missing from d_t_fmt add it to date_fmt.
It also corrects comments where necessary.
Exceptions:
* In bo_CN, dz_BT, and km_KH "%Z" has not been added to date_fmt because
it was too difficult. In these locales date_fmt has been set to the
copy of d_t_fmt.
* In en_DK "%Z" has not been removed from d_t_fmt in order to preserve
the conformance with the standard mentioned in the comment.
The command to identify and initially edit the locales that need the
update was:
for i in `grep -lw d_t_fmt *`
do
if ! grep -qw date_fmt $i ; then
awk '/d_t_fmt/ { print $0; gsub("d_t_fmt", "date_fmt"); } //{ print $0 }' < $i > $i.next
mv $i.next $i
fi
done
Hurd uses an empty prefix, so the linker scripts end up in /lib, the
find command picked them up, and stripping them failed because they
are not ELF files.
Florian Weimer [Thu, 2 Jan 2020 09:18:37 +0000 (10:18 +0100)]
Linux: Remove pread/pread64, pwrite/pwrite64 kludges from <sysdep.h>
Since the switch away from auto-generated wrappers for these system
calls, the kludge is already included in the C source file of the
system call wrapper.
This command uses pre-built compilers to re-install the Linux headers
from the current sources into a temporary location and runs glibc's
“make update-syscalls-lists” against that. This updates the glibc
source tree with the current system call numbers.
The new classes GlibcPolicyForCompiler and GlibcPolicyForBuild allow
customization of the Glibc.build_glibc method, replacing the existing
for_compiler flag.
Florian Weimer [Thu, 2 Jan 2020 09:18:22 +0000 (10:18 +0100)]
Linux: Use system call tables during build
Use <arch-syscall.h> instead of <asm/unistd.h> to obtain the system
call numbers. A few direct includes of <asm/unistd.h> need to be
removed (if the system call numbers are already provided indirectly
by <sysdep.h>) or replaced with <sys/syscall.h>.
Current Linux headers for alpha define the required system call names,
so most of the _NR_* hacks are no longer needed. For the 32-bit arm
architecture, eliminate the INTERNAL_SYSCALL_ARM macro, now that we
have regular system call names for cacheflush and set_tls. There are
more such cleanup opportunities for other architectures, but these
cleanups are required to avoid macro redefinition errors during the
build.
For ia64, it is desirable to use <asm/break.h> directly to obtain
the break number for system calls (which is not a system call number
itself). This requires replacing __BREAK_SYSCALL with
__IA64_BREAK_SYSCALL because the former is defined as an alias in
<asm/unistd.h>, but not in <asm/break.h>.
Florian Weimer [Thu, 2 Jan 2020 09:18:10 +0000 (10:18 +0100)]
Linux: Add tables with system call numbers
The new tables are currently only used for consistency checks
with the installed kernel headers and the architecture-independent
system call names table. They are based on Linux 5.4.
The goal is to use these architecture-specific tables to ensure
that system call wrappers are available irrespective of the version
of the installed kernel headers.
The tables are formatted in the form of C header files so that they
can be used directly in an #include directive, without external
preprocessing. (External preprocessing of a plain table file
would introduce cross-subdirectory dependency issues.) However,
the intent is that they can still be treated as tables and can be
processed by simple tools.
The irregular system call names on 32-bit arm add a complication.
The <fixup-asm-unistd.h> header is introduced to work around that,
and the system calls are listed under regular names in the
<arch-syscall.h> file.
A make target, update-syscalls-list, is added to patch the glibc
sources with data from the current kernel headers.
Joseph Myers [Wed, 1 Jan 2020 00:21:22 +0000 (00:21 +0000)]
Update copyright dates not handled by scripts/update-copyrights.
I've updated copyright dates in glibc for 2020. This is the patch for
the changes not generated by scripts/update-copyrights and subsequent
build / regeneration of generated files. As well as the usual annual
updates, mainly dates in --version output (minus libc.texinfo which
previously had to be handled manually but is now successfully updated
by update-copyrights), there is a fix to
sysdeps/unix/sysv/linux/powerpc/bits/termios-c_lflag.h where a typo in
the copyright notice meant it failed to be updated automatically.
Please remember to include 2020 in the dates for any new files added
in future (which means updating any existing uncommitted patches you
have that add new files to use the new copyright dates in them).
Rafał Lużyński [Mon, 30 Dec 2019 10:42:46 +0000 (11:42 +0100)]
lv_LV locale: Correct the time part of d_t_fmt (bug 25324)
Currently d_t_fmt formats time as "plkst. %H un %M". A quick Google
search says that "plkst." means "o’clock" and "un" means "and".
Also this format does not display seconds.
CLDR does not mention anything like that. We have no reason to use
anything different than "%H:%M:%S".
Jeremie Koenig [Sun, 29 Dec 2019 16:59:55 +0000 (17:59 +0100)]
hurd: Global signal disposition
This adds _hurd_sigstate_set_global_rcv used by libpthread to enable
POSIX-confirming behavior of signals on a per-thread basis.
This also provides a sigstate destructor _hurd_sigstate_delete, and a
global process signal state, which needs to be locked and check when
global disposition is enabled, thus the addition of _hurd_sigstate_lock
_hurd_sigstate_actions _hurd_sigstate_pending _hurd_sigstate_unlock helpers.
This also updates all the glibc code accordingly.
This also drops support for get_int(INIT_SIGMASK), which did not make sense
any more since we do not have a single signal thread any more.
During fork/spawn, this also reinitializes the child global sigstate's
lock. That cures an issue that would very rarely cause a deadlock in the
child in fork, tries to unlock ss' critical section lock at the end of
fork. This will typically (always?) be observed in /bin/sh, which is not
surprising as that is the foremost caller of fork.
To reproduce an intermediate state, add an endless loop if
_hurd_global_sigstate is locked after __proc_dostop (cast through
volatile); that is, while still being in the fork's parent process.
When that triggers (use the libtool testsuite), the signal thread has
already locked ss (which is _hurd_global_sigstate), and is stuck at
hurdsig.c:685 in post_signal, trying to lock _hurd_siglock (which the
main thread already has locked and keeps locked until after
__task_create). This is the case that ss->thread == MACH_PORT_NULL, that
is, a global signal. In the main thread, between __proc_dostop and
__task_create is the __thread_abort call on the signal thread which would
abort any current kernel operation (but leave ss locked). Later in fork,
in the parent, when _hurd_siglock is unlocked in fork, the parent's
signal thread can proceed and will unlock eventually the global sigstate.
In the client, _hurd_siglock will likewise be unlocked, but the global
sigstate never will be, as the client's signal thread has been configured
to restart execution from _hurd_msgport_receive. Thus, when the child
tries to unlock ss' critical section lock at the end of fork, it will
first lock the global sigstate, will spin trying to lock it, which can
never be successful, and we get our deadlock.
Options seem to be:
* Move the locking of _hurd_siglock earlier in post_signal -- but that
may generally impact performance, if this locking isn't generally
needed anyway?
On the other hand, would it actually make sense to wait here until we
are not any longer in a critical section (which is meant to disable
signal delivery anyway (but not for preempted signals?))?
* Clear the global sigstate in the fork's child with the rationale that
we're anyway restarting the signal thread from a clean state. This
has now been implemented.
Why has this problem not been observed before Jérémie's patches? (Or has
it? Perhaps even more rarely?) In _S_msg_sig_post, the signal is now
posted to a *global receiver thread*, whereas previously it was posted to
the *designated signal-receiving thread*. The latter one was in a
critical section in fork, so didn't try to handle the signal until after
leaving the critical section? (Not completely analyzed and verified.)
Another question is what the signal is that is being received
during/around the time __proc_dostop executes.
Jeremie Koenig [Sun, 29 Dec 2019 16:18:04 +0000 (17:18 +0100)]
hurd: Signal code refactoring
This should not change the current behavior, although this fixes a few
minor bugs which were made apparent in the process of global signal
disposition work:
- Split into more functions
- Scope variables more restrictively
- Split out inner functions
- refactor check_pending_signals
- make sigsuspend POSIX-conformant.
- fix uninitialized act value.
Not thoroughly tested, but manual testing as well as glibc tests look fine, and
manual -lpthread testing also looks fine (within the given bounds for a new
stack to be used with makecontext).
ldbl-128ibm-compat: Do not mix -mabi=*longdouble and -mlong-double-128
Some compiler versions, e.g. GCC 7, complain when -mlong-double-128 is
used together with -mabi=ibmlongdouble or -mabi=ieeelongdouble,
producing the following error message:
ldbl-128ibm-compat: Compiler flags for stdio functions
Some of the files that provide stdio.h and wchar.h functions have a
filename prefixed with 'io', such as 'iovsprintf.c'. On platforms that
imply ldbl-128ibm-compat, these files must be compiled with the flag
-mabi=ibmlongdouble. This patch adds this flag to their compilation.
Notice that this is not required for the other files that provide
similar functions, because filenames that are not prefixed with 'io'
have ldbl-128ibm-compat counterparts in the Makefile, which already adds
-mabi=ibmlongdouble to them.
Reviewed-by: Gabriel F. T. Gomes <gabrielftg@linux.ibm.com>
Do not redirect calls to __GI_* symbols, when redirecting to *ieee128
On platforms where long double has IEEE binary128 format as a third
option (initially, only powerpc64le), many exported functions are
redirected to their __*ieee128 equivalents. This redirection is
provided by installed headers such as stdio-ldbl.h, and is supposed to
work correctly with user code.
However, during the build of glibc, similar redirections are employed,
in internal headers, such as include/stdio.h, in order to avoid extra
PLT entries. These redirections conflict with the redirections to
__*ieee128, and must be avoided during the build. This patch protects
the second redirections with a test for __LONG_DOUBLE_USES_FLOAT128, a
new macro that is defined to 1 when functions that deal with long double
typed values reuses the _Float128 implementation (this is currently only
true for powerpc64le).
Tested for powerpc64le, x86_64, and with build-many-glibcs.py.
Co-authored-by: Gabriel F. T. Gomes <gabrielftg@linux.ibm.com> Reviewed-by: Florian Weimer <fweimer@redhat.com>
The functions do not fail regardless of the argument value. Also, for
Linux the return value is not correct on some platforms due the missing
usage of INTERNAL_SYSCALL_ERROR_P / INTERNAL_SYSCALL_ERRNO macros.
Checked on x86_64-linux-gnu, i686-linux-gnu, and sparc64-linux-gnu.
Avoid compat symbols for totalorder in powerpc64le IEEE long double
On powerpc64le, the libm_alias_float128_other_r_ldbl macro is
used to create an alias between totalorderf128 and __totalorderlieee128,
as well as between the totalordermagf128 and __totalordermaglieee128.
However, the totalorder* and totalordermag* functions changed their
parameter type since commit ID 42760d764649 and got compat symbols for
their old versions. With this change, the aforementioned macro would
create two conflicting aliases for __totalorderlieee128 and
__totalordermaglieee128.
This patch avoids the creation of the alias between the IEEE long double
symbols (__totalorderl*ieee128) and the compat symbols, because the IEEE
long double functions have never been exported thus don't need such
compat symbol.
Tested for powerpc64le.
Reviewed-by: Joseph Myers <joseph@codesourcery.com>
This patch adds IEEE long double versions of q*cvt* functions for
powerpc64le. Unlike all other long double to/from string conversion
functions, these do not rely on internal functions that can take
floating-point numbers with different formats and act on them
accordingly, instead, the related files are rebuilt with the
-mabi=ieeelongdouble compiler flag set.
Having -mabi=ieeelongdouble passed to the compiler causes the object
files to be marked with a .gnu_attribute that is incompatible with the
.gnu_attribute in files built with -mabi=ibmlongdouble (the default).
The difference causes error messages similar to the following:
ld: libc_pic.a(s_isinfl.os) uses IBM long double,
libc_pic.a(ieee128-qefgcvt_r.os) uses IEEE long double.
collect2: error: ld returned 1 exit status
make[2]: *** [../Makerules:649: libc_pic.os] Error 1
Although this warning is useful in other situations, the library
actually needs to have functions with different long double formats, so
.gnu_attribute generation is explicitly disabled for these files with
the use of -mno-gnu-attribute.
Tested for powerpc64le on the branch that actually enables the
sysdeps/ieee754/ldbl-128ibm-compat for powerpc64le.
Reviewed-by: Paul E. Murphy <murphyp@linux.ibm.com>
This patch refactors the *cvt functions implementation in a way that
makes it easier to re-use them for implementing the IEEE long double on
powerpc64le. By removing the macros that generate the function names
(APPEND combined with FUNC_PREFIX), the new code makes it easier to
define new function names, such as __qecvtieee128.
Tested that installed stripped binaries for all build-many-glibcs
targets remain identical before and after this patch. Also tested for
powerpc64le and x86_64.
Reviewed-by: Paul E. Murphy <murphyp@linux.ibm.com>
This patch refactors the *cvt functions implementation in a way that
makes it easier to re-use them for implementing the IEEE long double on
powerpc64le. By splitting the implementation per se in one file
(efgcvt-template.c) and the alias definitions in others (e.g. efgcvt.c),
the new code makes it easier to define new function names, such as
__qecvtieee128.
Tested that installed stripped binaries for all build-many-glibcs
targets remain identical before and after this patch. Also tested for
powerpc64le and x86_64.
Reviewed-by: Paul E. Murphy <murphyp@linux.ibm.com>
Xuelei Zhang [Thu, 19 Dec 2019 15:26:33 +0000 (15:26 +0000)]
aarch64: Optimized memset for Kunpeng processor.
Due to the branch prediction issue of Kunpeng processor, we found
memset_generic has poor performance on middle sizes setting, and so
we reconstructed the logic, expanded the loop by 4 times in set_long
to solve the problem, even when setting below 1K sizes have benefit.
Another change is that DZ_ZVA seems no work when setting zero, so we
discarded it and used set_long to set zero instead. Fewer branches and
predictions also make the zero case have slightly improvement.
Xuelei Zhang [Thu, 19 Dec 2019 13:41:40 +0000 (13:41 +0000)]
aarch64: Optimized strlen for strlen_asimd
Optimize the strlen implementation by using vector operations and
loop unrolling in main loop.Compared to __strlen_generic,it reduces
latency of cases in bench-strlen by 7%~18% when the length of src
is greater than 128 bytes, with gains throughout the benchmark.
Xuelei Zhang [Thu, 19 Dec 2019 14:53:29 +0000 (14:53 +0000)]
aarch64: Optimized implementation of memrchr
Considering the excellent performance of memchr.S on glibc 2.30, the
same algorithm is used to find chrin. Compared to memrchr.c, this
method with memrchr.S achieves an average performance improvement
of 58% based on benchtest and its extension cases.
Xuelei Zhang [Thu, 19 Dec 2019 13:49:46 +0000 (13:49 +0000)]
aarch64: Optimized implementation of strnlen
Optimize the strlen implementation by using vector operations and
loop unrooling in main loop. Compared to aarch64/strnlen.S, it
reduces latency of cases in bench-strnlen by 11%~24% when the length
of src is greater than 64 bytes, with gains throughout the benchmark.
Xuelei Zhang [Thu, 19 Dec 2019 13:08:11 +0000 (13:08 +0000)]
aarch64: Optimized implementation of strcpy
Optimize the strcpy implementation by using vector loads and operations
in main loop.Compared to aarch64/strcpy.S, it reduces latency of cases
in bench-strlen by 5%~18% when the length of src is greater than 64
bytes, with gains throughout the benchmark.
Xuelei Zhang [Thu, 19 Dec 2019 12:31:59 +0000 (12:31 +0000)]
aarch64: Optimized implementation of memcmp
The loop body is expanded from a 16-byte comparison to a 64-byte
comparison, and the usage of ldp is replaced by the Post-index
mode to the Base plus offset mode. Hence, compare can faster 18%
around > 128 bytes in all.
linux: Use waitid on wait4 if __NR_wait4 is not defined
If the wait4 syscall is not available (such as y2038 safe 32-bit
systems) waitid should be used instead. However prior Linux 5.4
waitid is not a full superset of other wait syscalls, since it
does not include support for waiting for the current process group.
It is possible to emulate wait4 by issuing an extra syscall to get
the current process group, but it is inherent racy: after the current
process group is received and before it is passed to waitid a signal
could arrive causing the current process group to change.
So waitid is used if wait4 is not defined iff the build is
enabled with a minimum kernel if 5.4+. The new assume
__ASSUME_WAITID_PID0_P_PGID is added and an error is issued if waitid
can not be implemented by either __NR_wait4 or
__NR_waitid && __ASSUME_WAITID_PID0_P_PGID.
Checked on x86_64-linux-gnu and i686-linux-gnu.
Co-authored-by: Alistair Francis <alistair.francis@wdc.com>
The POSIX implementation is used as default and both BSD and Linux
version are removed. It simplifies the implementation for
architectures that do not provide either __NR_waitpid or
__NR_wait4.
Checked on x86_64-linux-gnu and powerpc64le-linux-gnu.
It enables and disables cancellation with pthread_setcancelstate
before calling the waitpid. It simplifies the waitpid implementation
for architectures that do not provide either __NR_waitpid or
__NR_wait4.
Florian Weimer [Thu, 19 Dec 2019 13:45:50 +0000 (14:45 +0100)]
Fix test isolation for elf/tst-ifunc-fault-lazy, elf/tst-ifunc-fault-bindnow
Previously, ld.so was invoked only with the elf subdirectory on the
library search path. Since the soname link for libc.so only exists in
the top-level build directory, this leaked the system libc into the
test.
Another possible fix is use a static const object. Although there
should not be a difference between a const compound literal and a static
const object, the gcc C99 status page [1] has a note stating that this
optimization is not implemented:
"const-qualified compound literals could share storage with each
other and with string literals, but currently don't.".
This patch fixes it by moving both sigset_t that represent the
signal sets to static const data object. It generates slight better
code where the object reference is used directly instead of a stack
allocation plus the content materialization.
Checked on x86_64-linux-gnu, i686-linux-gnu, and sparc64-linux-gnu.
Florian Weimer [Wed, 18 Dec 2019 09:27:10 +0000 (10:27 +0100)]
hurd: Do not make sigprocmask available in ld.so
After commit f7649d5780aa4682393b9daedd653e4d9c12784c ("dlopen: Do not
block signals"), the dynamic linker no longer uses sigprocmask, which
means that it does not have to be made available explicitly on hurd.
James Clarke [Tue, 17 Dec 2019 18:29:29 +0000 (18:29 +0000)]
hurd: Make getrandom honour GRND_NONBLOCK
* sysdeps/mach/hurd/getrandom.c (__getrandom): Open the random source
with O_NONBLOCK when the GRND_NONBLOCK flag is provided.
Message-Id: <20191217182929.90989-1-jrtc27@jrtc27.com>
mips: Do not include hi and lo in __SYSCALL_CLOBBERS for R6
GCC 10 (PR 91233) won't silently allow registers that are not architecturally
available to be present in the clobber list anymore, resulting in build failure
for mips*r6 targets in form of:
...
.../sysdep.h:146:2: error: the register ‘lo’ cannot be clobbered in ‘asm’ for the current target
146 | __asm__ volatile ( \
| ^~~~~~~
This is because base R6 ISA doesn't define hi and lo registers w/o DSP extension.
This patch provides the alternative definitions of __SYSCALL_CLOBBERS for r6
targets that won't include those registers.
* sysdeps/unix/sysv/linux/mips/mips32/sysdep.h (__SYSCALL_CLOBBERS): Exclude
hi and lo from the clobber list for __mips_isa_rev >= 6.
* sysdeps/unix/sysv/linux/mips/mips64/n32/sysdep.h (__SYSCALL_CLOBBERS): Likewise.
* sysdeps/unix/sysv/linux/mips/mips64/n64/sysdep.h (__SYSCALL_CLOBBERS): Likewise.
ldbl-128ibm-compat: Add ISO C99 versions of scanf functions
In the format string for *scanf functions, the '%as', '%aS', and '%a[]'
modifiers behave differently depending on ISO C99 compatibility. When
_GNU_SOURCE is defined and -std=c89 is passed to the compiler, these
functions behave like ascanf, and the modifiers allocate memory for the
output. Otherwise, the ISO C99 compliant version of these functions is
used, and the modifiers consume a floating-point argument. This patch
adds the IEEE binary128 variant of ISO C99 compliant functions for the
third long double format on powerpc64le.
Tested for powerpc64le.
Reviewed-by: Paul E. Murphy <murphyp@linux.ibm.com>
Use C99-compliant scanf under _GNU_SOURCE with modern compilers.
the selection of the GNU versions of scanf functions requires both
_GNU_SOURCE and -std=c89. This patch changes the tests in
ldbl-128ibm-compat so that they actually test the GNU versions (without
this change, the redirection to the ISO C99 version always happens, so
GNU versions of the new implementation (e.g. __scanfieee128) were left
untested).
Tested for powerpc64le.
Reviewed-by: Paul E. Murphy <murphyp@linux.ibm.com>
Florian Weimer [Fri, 13 Dec 2019 09:23:10 +0000 (10:23 +0100)]
dlopen: Do not block signals
Blocking signals causes issues with certain anti-malware solutions
which rely on an unblocked SIGSYS signal for system calls they
intercept.
This reverts commit a2e8aa0d9ea648068d8be52dd7b15f1b6a008e23
("Block signals during the initial part of dlopen") and adds
comments related to async signal safety to active_nodelete and
its caller.
Note that this does not make lazy binding async-signal-safe with regards
to dlopen. It merely avoids introducing new async-signal-safety hazards
as part of the NODELETE changes.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org> Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Fri, 13 Dec 2019 09:18:46 +0000 (10:18 +0100)]
dlopen: Rework handling of pending NODELETE status
Commit a2e8aa0d9ea648068d8be52dd7b15f1b6a008e23 ("Block signals during
the initial part of dlopen") was deemed necessary because of
read-modify-write operations like the one in add_dependency in
elf/dl-lookup.c. In the old code, we check for any kind of NODELETE
status and bail out:
/* Redo the NODELETE check, as when dl_load_lock wasn't held
yet this could have changed. */
if (map->l_nodelete != link_map_nodelete_inactive)
goto out;
If a signal arrives during relocation and the signal handler, through
lazy binding, adds a global scope dependency on the same map, it will
set map->l_nodelete to link_map_nodelete_active. This will be
overwritten with link_map_nodelete_pending by the dlopen relocation
code.
To avoid such problems in relation to the l_nodelete member, this
commit introduces two flags for active NODELETE status (irrevocable)
and pending NODELETE status (revocable until activate_nodelete is
invoked). As a result, NODELETE processing in dlopen does not
introduce further reasons why lazy binding from signal handlers
is unsafe during dlopen, and a subsequent commit can remove signal
blocking from dlopen.
This does not address pre-existing issues (unrelated to the NODELETE
changes) which make lazy binding in a signal handler during dlopen
unsafe, such as the use of malloc in both cases.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org> Reviewed-by: Carlos O'Donell <carlos@redhat.com>
It can happen that an already-loaded object that is in the local
scope is promoted to NODELETE status, via binding to a unique
symbol.
Similarly, it is possible that such NODELETE promotion occurs to
an already-loaded object from the global scope. This is why the
loop in activate_nodelete has to cover all objects in the namespace
of the new object.
In do_lookup_unique, it could happen that the NODELETE status of
an already-loaded object was overwritten with a pending NODELETE
status. As a result, if dlopen fails, this could cause a loss of
the NODELETE status of the affected object, eventually resulting
in an incorrect unload.
Stefan Liebler [Wed, 11 Dec 2019 14:09:33 +0000 (15:09 +0100)]
S390: Implement roundtoint and converttoint and define TOINT_INTRINSICS.
This patch implements roundtoint and convertoint for s390
by using the load-fp-integer and convert-to-fixed instructions.
Both functions are using "round to nearest with ties away from zero"
rounding mode and do not raise inexact exceptions.
projects / thirdparty / glibc.git / log
