about summary refs log tree commit diff
path: root/src
diff options
context:
space:
mode:
Diffstat (limited to 'src')
-rw-r--r--src/aio/aio.c32
-rw-r--r--src/aio/aio_suspend.c6
-rw-r--r--src/aio/lio_listio.c2
-rw-r--r--src/conf/sysconf.c13
-rw-r--r--src/dirent/alphasort.c2
-rw-r--r--src/dirent/readdir.c2
-rw-r--r--src/dirent/readdir_r.c2
-rw-r--r--src/dirent/scandir.c2
-rw-r--r--src/dirent/versionsort.c3
-rw-r--r--src/fcntl/creat.c2
-rw-r--r--src/fcntl/open.c2
-rw-r--r--src/fcntl/openat.c2
-rw-r--r--src/fcntl/posix_fadvise.c2
-rw-r--r--src/fcntl/posix_fallocate.c2
-rw-r--r--src/fenv/loongarch64/fenv.S78
-rw-r--r--src/fenv/riscv32/fenv-sf.c3
-rw-r--r--src/fenv/riscv32/fenv.S56
-rw-r--r--src/internal/atomic.h2
-rw-r--r--src/internal/dynlink.h10
-rw-r--r--src/internal/emulate_wait4.c55
-rw-r--r--src/internal/fork_impl.h4
-rw-r--r--src/internal/ksigaction.h5
-rw-r--r--src/internal/syscall.h18
-rw-r--r--src/ipc/semtimedop.c3
-rw-r--r--src/ldso/dlerror.c34
-rw-r--r--src/ldso/loongarch64/dlsym.s7
-rw-r--r--src/ldso/riscv32/dlsym.s6
-rw-r--r--src/ldso/riscv64/tlsdesc.s32
-rw-r--r--src/ldso/sh/dlsym.s2
-rw-r--r--src/legacy/ftw.c2
-rw-r--r--src/linux/cache.c3
-rw-r--r--src/linux/clone.c56
-rw-r--r--src/linux/epoll.c1
-rw-r--r--src/linux/fallocate.c3
-rw-r--r--src/linux/getdents.c2
-rw-r--r--src/linux/membarrier.c2
-rw-r--r--src/linux/preadv2.c17
-rw-r--r--src/linux/prlimit.c3
-rw-r--r--src/linux/pwritev2.c17
-rw-r--r--src/linux/sendfile.c2
-rw-r--r--src/linux/statx.c42
-rw-r--r--src/linux/wait4.c4
-rw-r--r--src/locale/iconv.c9
-rw-r--r--src/malloc/mallocng/free.c2
-rw-r--r--src/malloc/mallocng/glue.h2
-rw-r--r--src/math/acoshl.c10
-rw-r--r--src/math/logf.c2
-rw-r--r--src/math/powl.c34
-rw-r--r--src/math/riscv32/copysign.c15
-rw-r--r--src/math/riscv32/copysignf.c15
-rw-r--r--src/math/riscv32/fabs.c15
-rw-r--r--src/math/riscv32/fabsf.c15
-rw-r--r--src/math/riscv32/fma.c15
-rw-r--r--src/math/riscv32/fmaf.c15
-rw-r--r--src/math/riscv32/fmax.c15
-rw-r--r--src/math/riscv32/fmaxf.c15
-rw-r--r--src/math/riscv32/fmin.c15
-rw-r--r--src/math/riscv32/fminf.c15
-rw-r--r--src/math/riscv32/sqrt.c15
-rw-r--r--src/math/riscv32/sqrtf.c15
-rw-r--r--src/math/sqrtl.c4
-rw-r--r--src/misc/getopt.c3
-rw-r--r--src/misc/getrlimit.c2
-rw-r--r--src/misc/lockf.c2
-rw-r--r--src/misc/mntent.c46
-rw-r--r--src/misc/nftw.c6
-rw-r--r--src/misc/setrlimit.c2
-rw-r--r--src/misc/syslog.c3
-rw-r--r--src/mman/mmap.c2
-rw-r--r--src/mq/mq_notify.c47
-rw-r--r--src/multibyte/mbrtowc.c2
-rw-r--r--src/network/accept4.c4
-rw-r--r--src/network/dns_parse.c11
-rw-r--r--src/network/gai_strerror.c2
-rw-r--r--src/network/getaddrinfo.c5
-rw-r--r--src/network/gethostbyaddr.c2
-rw-r--r--src/network/gethostbyaddr_r.c5
-rw-r--r--src/network/gethostbyname2.c2
-rw-r--r--src/network/gethostbyname2_r.c6
-rw-r--r--src/network/getifaddrs.c14
-rw-r--r--src/network/getnameinfo.c9
-rw-r--r--src/network/getservbyport_r.c4
-rw-r--r--src/network/inet_pton.c1
-rw-r--r--src/network/lookup.h2
-rw-r--r--src/network/lookup_ipliteral.c4
-rw-r--r--src/network/lookup_name.c47
-rw-r--r--src/network/res_mkquery.c1
-rw-r--r--src/network/res_msend.c183
-rw-r--r--src/network/res_send.c10
-rw-r--r--src/network/sendmsg.c7
-rw-r--r--src/process/_Fork.c31
-rw-r--r--src/process/fork.c8
-rw-r--r--src/process/posix_spawn.c7
-rw-r--r--src/process/riscv64/vfork.s12
-rw-r--r--src/process/waitpid.c2
-rw-r--r--src/regex/glob.c5
-rw-r--r--src/search/hsearch.c4
-rw-r--r--src/select/poll.c9
-rw-r--r--src/select/ppoll.c (renamed from src/linux/ppoll.c)2
-rw-r--r--src/select/select.c1
-rw-r--r--src/setjmp/loongarch64/longjmp.S32
-rw-r--r--src/setjmp/loongarch64/setjmp.S34
-rw-r--r--src/setjmp/powerpc/longjmp.S8
-rw-r--r--src/setjmp/riscv32/longjmp.S42
-rw-r--r--src/setjmp/riscv32/setjmp.S41
-rw-r--r--src/signal/loongarch64/restore.s (renamed from src/signal/mips/restore.s)15
-rw-r--r--src/signal/loongarch64/sigsetjmp.s25
-rw-r--r--src/signal/mips64/restore.s11
-rw-r--r--src/signal/mipsn32/restore.s11
-rw-r--r--src/signal/riscv32/restore.s8
-rw-r--r--src/signal/riscv32/sigsetjmp.s23
-rw-r--r--src/signal/sh/sigsetjmp.s2
-rw-r--r--src/signal/sigaction.c5
-rw-r--r--src/stat/__xstat.c5
-rw-r--r--src/stat/fchmodat.c7
-rw-r--r--src/stat/fstat.c4
-rw-r--r--src/stat/fstatat.c5
-rw-r--r--src/stat/lstat.c4
-rw-r--r--src/stat/stat.c4
-rw-r--r--src/stat/statvfs.c6
-rw-r--r--src/stdio/fgetpos.c2
-rw-r--r--src/stdio/fgets.c5
-rw-r--r--src/stdio/fopen.c2
-rw-r--r--src/stdio/freopen.c4
-rw-r--r--src/stdio/fseek.c2
-rw-r--r--src/stdio/fsetpos.c2
-rw-r--r--src/stdio/ftell.c2
-rw-r--r--src/stdio/open_wmemstream.c6
-rw-r--r--src/stdio/pclose.c2
-rw-r--r--src/stdio/tmpfile.c2
-rw-r--r--src/stdio/vfprintf.c31
-rw-r--r--src/stdio/vfwprintf.c43
-rw-r--r--src/stdio/vsnprintf.c5
-rw-r--r--src/stdio/vswprintf.c4
-rw-r--r--src/string/strverscmp.c6
-rw-r--r--src/string/wcscmp.c2
-rw-r--r--src/string/wcsncmp.c2
-rw-r--r--src/string/wmemcmp.c2
-rw-r--r--src/temp/mkostemp.c2
-rw-r--r--src/temp/mkostemps.c1
-rw-r--r--src/temp/mkstemp.c2
-rw-r--r--src/temp/mkstemps.c2
-rw-r--r--src/thread/loongarch64/__set_thread_area.s7
-rw-r--r--src/thread/loongarch64/__unmapself.s7
-rw-r--r--src/thread/loongarch64/clone.s29
-rw-r--r--src/thread/loongarch64/syscall_cp.s29
-rw-r--r--src/thread/pthread_atfork.c8
-rw-r--r--src/thread/pthread_cancel.c9
-rw-r--r--src/thread/pthread_create.c16
-rw-r--r--src/thread/pthread_detach.c8
-rw-r--r--src/thread/pthread_key_create.c8
-rw-r--r--src/thread/riscv32/__set_thread_area.s6
-rw-r--r--src/thread/riscv32/__unmapself.s7
-rw-r--r--src/thread/riscv32/clone.s34
-rw-r--r--src/thread/riscv32/syscall_cp.s29
-rw-r--r--src/thread/sem_getvalue.c3
-rw-r--r--src/thread/sem_post.c12
-rw-r--r--src/thread/sem_timedwait.c10
-rw-r--r--src/thread/sem_trywait.c6
-rw-r--r--src/thread/synccall.c10
-rw-r--r--src/time/__year_to_secs.c4
-rw-r--r--src/time/clock_gettime.c3
-rw-r--r--src/time/strftime.c8
-rw-r--r--src/time/timer_create.c8
-rw-r--r--src/unistd/dup3.c6
-rw-r--r--src/unistd/faccessat.c2
-rw-r--r--src/unistd/ftruncate.c2
-rw-r--r--src/unistd/lseek.c1
-rw-r--r--src/unistd/mipsn32/lseek.c1
-rw-r--r--src/unistd/pipe2.c1
-rw-r--r--src/unistd/pread.c2
-rw-r--r--src/unistd/preadv.c2
-rw-r--r--src/unistd/pwrite.c2
-rw-r--r--src/unistd/pwritev.c2
-rw-r--r--src/unistd/setxid.c2
-rw-r--r--src/unistd/truncate.c2
-rw-r--r--src/unistd/x32/lseek.c1
177 files changed, 1588 insertions, 423 deletions
diff --git a/src/aio/aio.c b/src/aio/aio.c
index a1a3e791..d7e063bf 100644
--- a/src/aio/aio.c
+++ b/src/aio/aio.c
@@ -82,6 +82,8 @@ static size_t io_thread_stack_size;
 
 static struct aio_queue *__aio_get_queue(int fd, int need)
 {
+	sigset_t allmask, origmask;
+	int masked = 0;
 	if (fd < 0) {
 		errno = EBADF;
 		return 0;
@@ -93,6 +95,9 @@ static struct aio_queue *__aio_get_queue(int fd, int need)
 	if ((!map || !map[a] || !map[a][b] || !map[a][b][c] || !(q=map[a][b][c][d])) && need) {
 		pthread_rwlock_unlock(&maplock);
 		if (fcntl(fd, F_GETFD) < 0) return 0;
+		sigfillset(&allmask);
+		masked = 1;
+		pthread_sigmask(SIG_BLOCK, &allmask, &origmask);
 		pthread_rwlock_wrlock(&maplock);
 		if (!io_thread_stack_size) {
 			unsigned long val = __getauxval(AT_MINSIGSTKSZ);
@@ -119,6 +124,7 @@ static struct aio_queue *__aio_get_queue(int fd, int need)
 	if (q) pthread_mutex_lock(&q->lock);
 out:
 	pthread_rwlock_unlock(&maplock);
+	if (masked) pthread_sigmask(SIG_SETMASK, &origmask, 0);
 	return q;
 }
 
@@ -401,18 +407,26 @@ void __aio_atfork(int who)
 	if (who<0) {
 		pthread_rwlock_rdlock(&maplock);
 		return;
+	} else if (!who) {
+		pthread_rwlock_unlock(&maplock);
+		return;
 	}
-	if (who>0 && map) for (int a=0; a<(-1U/2+1)>>24; a++)
+	aio_fd_cnt = 0;
+	if (pthread_rwlock_tryrdlock(&maplock)) {
+		/* Obtaining lock may fail if _Fork was called nor via
+		 * fork. In this case, no further aio is possible from
+		 * child and we can just null out map so __aio_close
+		 * does not attempt to do anything. */
+		map = 0;
+		return;
+	}
+	if (map) for (int a=0; a<(-1U/2+1)>>24; a++)
 		if (map[a]) for (int b=0; b<256; b++)
 			if (map[a][b]) for (int c=0; c<256; c++)
 				if (map[a][b][c]) for (int d=0; d<256; d++)
 					map[a][b][c][d] = 0;
-	pthread_rwlock_unlock(&maplock);
+	/* Re-initialize the rwlock rather than unlocking since there
+	 * may have been more than one reference on it in the parent.
+	 * We are not a lock holder anyway; the thread in the parent was. */
+	pthread_rwlock_init(&maplock, 0);
 }
-
-weak_alias(aio_cancel, aio_cancel64);
-weak_alias(aio_error, aio_error64);
-weak_alias(aio_fsync, aio_fsync64);
-weak_alias(aio_read, aio_read64);
-weak_alias(aio_write, aio_write64);
-weak_alias(aio_return, aio_return64);
diff --git a/src/aio/aio_suspend.c b/src/aio/aio_suspend.c
index 1c1060e3..1f0c9aaa 100644
--- a/src/aio/aio_suspend.c
+++ b/src/aio/aio_suspend.c
@@ -9,7 +9,7 @@ int aio_suspend(const struct aiocb *const cbs[], int cnt, const struct timespec
 {
 	int i, tid = 0, ret, expect = 0;
 	struct timespec at;
-	volatile int dummy_fut, *pfut;
+	volatile int dummy_fut = 0, *pfut;
 	int nzcnt = 0;
 	const struct aiocb *cb = 0;
 
@@ -73,7 +73,3 @@ int aio_suspend(const struct aiocb *const cbs[], int cnt, const struct timespec
 		}
 	}
 }
-
-#if !_REDIR_TIME64
-weak_alias(aio_suspend, aio_suspend64);
-#endif
diff --git a/src/aio/lio_listio.c b/src/aio/lio_listio.c
index 0799c15d..a672812f 100644
--- a/src/aio/lio_listio.c
+++ b/src/aio/lio_listio.c
@@ -139,5 +139,3 @@ int lio_listio(int mode, struct aiocb *restrict const *restrict cbs, int cnt, st
 
 	return 0;
 }
-
-weak_alias(lio_listio, lio_listio64);
diff --git a/src/conf/sysconf.c b/src/conf/sysconf.c
index 3baaed32..60d3e745 100644
--- a/src/conf/sysconf.c
+++ b/src/conf/sysconf.c
@@ -4,6 +4,7 @@
 #include <sys/resource.h>
 #include <signal.h>
 #include <sys/sysinfo.h>
+#include <sys/auxv.h>
 #include "syscall.h"
 #include "libc.h"
 
@@ -19,6 +20,8 @@
 #define JT_AVPHYS_PAGES JT(9)
 #define JT_ZERO JT(10)
 #define JT_DELAYTIMER_MAX JT(11)
+#define JT_MINSIGSTKSZ JT(12)
+#define JT_SIGSTKSZ JT(13)
 
 #define RLIM(x) (-32768|(RLIMIT_ ## x))
 
@@ -165,6 +168,9 @@ long sysconf(int name)
 		[_SC_XOPEN_STREAMS] = JT_ZERO,
 		[_SC_THREAD_ROBUST_PRIO_INHERIT] = -1,
 		[_SC_THREAD_ROBUST_PRIO_PROTECT] = -1,
+
+		[_SC_MINSIGSTKSZ] = JT_MINSIGSTKSZ,
+		[_SC_SIGSTKSZ] = JT_SIGSTKSZ,
 	};
 
 	if (name >= sizeof(values)/sizeof(values[0]) || !values[name]) {
@@ -212,6 +218,13 @@ long sysconf(int name)
 		mem *= si.mem_unit;
 		mem /= PAGE_SIZE;
 		return (mem > LONG_MAX) ? LONG_MAX : mem;
+	case JT_MINSIGSTKSZ & 255:
+	case JT_SIGSTKSZ & 255: ;
+		long val = __getauxval(AT_MINSIGSTKSZ);
+		if (val < MINSIGSTKSZ) val = MINSIGSTKSZ;
+		if (values[name] == JT_SIGSTKSZ)
+			val += SIGSTKSZ - MINSIGSTKSZ;
+		return val;
 	case JT_ZERO & 255:
 		return 0;
 	}
diff --git a/src/dirent/alphasort.c b/src/dirent/alphasort.c
index bee672eb..ab2624e2 100644
--- a/src/dirent/alphasort.c
+++ b/src/dirent/alphasort.c
@@ -5,5 +5,3 @@ int alphasort(const struct dirent **a, const struct dirent **b)
 {
 	return strcoll((*a)->d_name, (*b)->d_name);
 }
-
-weak_alias(alphasort, alphasort64);
diff --git a/src/dirent/readdir.c b/src/dirent/readdir.c
index 569fc705..5a03b363 100644
--- a/src/dirent/readdir.c
+++ b/src/dirent/readdir.c
@@ -25,5 +25,3 @@ struct dirent *readdir(DIR *dir)
 	dir->tell = de->d_off;
 	return de;
 }
-
-weak_alias(readdir, readdir64);
diff --git a/src/dirent/readdir_r.c b/src/dirent/readdir_r.c
index e2a818f3..0d5de5f5 100644
--- a/src/dirent/readdir_r.c
+++ b/src/dirent/readdir_r.c
@@ -25,5 +25,3 @@ int readdir_r(DIR *restrict dir, struct dirent *restrict buf, struct dirent **re
 	*result = buf;
 	return 0;
 }
-
-weak_alias(readdir_r, readdir64_r);
diff --git a/src/dirent/scandir.c b/src/dirent/scandir.c
index 7ee195dd..7456b9b8 100644
--- a/src/dirent/scandir.c
+++ b/src/dirent/scandir.c
@@ -43,5 +43,3 @@ int scandir(const char *path, struct dirent ***res,
 	*res = names;
 	return cnt;
 }
-
-weak_alias(scandir, scandir64);
diff --git a/src/dirent/versionsort.c b/src/dirent/versionsort.c
index d4c48923..97696105 100644
--- a/src/dirent/versionsort.c
+++ b/src/dirent/versionsort.c
@@ -6,6 +6,3 @@ int versionsort(const struct dirent **a, const struct dirent **b)
 {
 	return strverscmp((*a)->d_name, (*b)->d_name);
 }
-
-#undef versionsort64
-weak_alias(versionsort, versionsort64);
diff --git a/src/fcntl/creat.c b/src/fcntl/creat.c
index 8f8aab64..c9c43910 100644
--- a/src/fcntl/creat.c
+++ b/src/fcntl/creat.c
@@ -4,5 +4,3 @@ int creat(const char *filename, mode_t mode)
 {
 	return open(filename, O_CREAT|O_WRONLY|O_TRUNC, mode);
 }
-
-weak_alias(creat, creat64);
diff --git a/src/fcntl/open.c b/src/fcntl/open.c
index 1d817a2d..4c3c8275 100644
--- a/src/fcntl/open.c
+++ b/src/fcntl/open.c
@@ -19,5 +19,3 @@ int open(const char *filename, int flags, ...)
 
 	return __syscall_ret(fd);
 }
-
-weak_alias(open, open64);
diff --git a/src/fcntl/openat.c b/src/fcntl/openat.c
index ad165ec3..83a9e0d0 100644
--- a/src/fcntl/openat.c
+++ b/src/fcntl/openat.c
@@ -15,5 +15,3 @@ int openat(int fd, const char *filename, int flags, ...)
 
 	return syscall_cp(SYS_openat, fd, filename, flags|O_LARGEFILE, mode);
 }
-
-weak_alias(openat, openat64);
diff --git a/src/fcntl/posix_fadvise.c b/src/fcntl/posix_fadvise.c
index 75b8e1ae..07346d21 100644
--- a/src/fcntl/posix_fadvise.c
+++ b/src/fcntl/posix_fadvise.c
@@ -14,5 +14,3 @@ int posix_fadvise(int fd, off_t base, off_t len, int advice)
 		__SYSCALL_LL_E(len), advice);
 #endif
 }
-
-weak_alias(posix_fadvise, posix_fadvise64);
diff --git a/src/fcntl/posix_fallocate.c b/src/fcntl/posix_fallocate.c
index c57a24ae..80a65cbf 100644
--- a/src/fcntl/posix_fallocate.c
+++ b/src/fcntl/posix_fallocate.c
@@ -6,5 +6,3 @@ int posix_fallocate(int fd, off_t base, off_t len)
 	return -__syscall(SYS_fallocate, fd, 0, __SYSCALL_LL_E(base),
 		__SYSCALL_LL_E(len));
 }
-
-weak_alias(posix_fallocate, posix_fallocate64);
diff --git a/src/fenv/loongarch64/fenv.S b/src/fenv/loongarch64/fenv.S
new file mode 100644
index 00000000..9c38599e
--- /dev/null
+++ b/src/fenv/loongarch64/fenv.S
@@ -0,0 +1,78 @@
+#ifndef __loongarch_soft_float
+
+#ifdef BROKEN_LOONGARCH_FCSR_ASM
+#define FCSR $r0
+#else
+#define FCSR $fcsr0
+#endif
+
+.global feclearexcept
+.type   feclearexcept,@function
+feclearexcept:
+	li.w    $t0, 0x1f0000
+	and     $a0, $a0, $t0
+	movfcsr2gr $t1, FCSR
+	andn    $t1, $t1, $a0
+	movgr2fcsr FCSR, $t1
+	li.w    $a0, 0
+	jr      $ra
+
+.global feraiseexcept
+.type   feraiseexcept,@function
+feraiseexcept:
+	li.w    $t0, 0x1f0000
+	and     $a0, $a0, $t0
+	movfcsr2gr $t1, FCSR
+	or      $t1, $t1, $a0
+	movgr2fcsr FCSR, $t1
+	li.w    $a0, 0
+	jr      $ra
+
+.global fetestexcept
+.type   fetestexcept,@function
+fetestexcept:
+	li.w    $t0, 0x1f0000
+	and     $a0, $a0, $t0
+	movfcsr2gr $t1, FCSR
+	and     $a0, $t1, $a0
+	jr      $ra
+
+.global fegetround
+.type   fegetround,@function
+fegetround:
+	movfcsr2gr $t0, FCSR
+	andi    $a0, $t0, 0x300
+	jr      $ra
+
+.global __fesetround
+.hidden __fesetround
+.type   __fesetround,@function
+__fesetround:
+	li.w    $t0, 0x300
+	and     $a0, $a0, $t0
+	movfcsr2gr $t1, FCSR
+	andn    $t1, $t1, $t0
+	or      $t1, $t1, $a0
+	movgr2fcsr FCSR, $t1
+	li.w    $a0, 0
+	jr      $ra
+
+.global fegetenv
+.type   fegetenv,@function
+fegetenv:
+	movfcsr2gr $t0, FCSR
+	st.w    $t0, $a0, 0
+	li.w    $a0, 0
+	jr      $ra
+
+.global fesetenv
+.type   fesetenv,@function
+fesetenv:
+	addi.d  $t0, $a0, 1
+	beq     $t0, $r0, 1f
+	ld.w    $t0, $a0, 0
+1:	movgr2fcsr FCSR, $t0
+	li.w    $a0, 0
+	jr      $ra
+
+#endif
diff --git a/src/fenv/riscv32/fenv-sf.c b/src/fenv/riscv32/fenv-sf.c
new file mode 100644
index 00000000..ecd3cb5c
--- /dev/null
+++ b/src/fenv/riscv32/fenv-sf.c
@@ -0,0 +1,3 @@
+#ifndef __riscv_flen
+#include "../fenv.c"
+#endif
diff --git a/src/fenv/riscv32/fenv.S b/src/fenv/riscv32/fenv.S
new file mode 100644
index 00000000..0ea78bf9
--- /dev/null
+++ b/src/fenv/riscv32/fenv.S
@@ -0,0 +1,56 @@
+#ifdef __riscv_flen
+
+.global feclearexcept
+.type feclearexcept, %function
+feclearexcept:
+	csrc fflags, a0
+	li a0, 0
+	ret
+
+.global feraiseexcept
+.type feraiseexcept, %function
+feraiseexcept:
+	csrs fflags, a0
+	li a0, 0
+	ret
+
+.global fetestexcept
+.type fetestexcept, %function
+fetestexcept:
+	frflags t0
+	and a0, t0, a0
+	ret
+
+.global fegetround
+.type fegetround, %function
+fegetround:
+	frrm a0
+	ret
+
+.global __fesetround
+.type __fesetround, %function
+__fesetround:
+	fsrm t0, a0
+	li a0, 0
+	ret
+
+.global fegetenv
+.type fegetenv, %function
+fegetenv:
+	frcsr t0
+	sw t0, 0(a0)
+	li a0, 0
+	ret
+
+.global fesetenv
+.type fesetenv, %function
+fesetenv:
+	li t2, -1
+	li t1, 0
+	beq a0, t2, 1f
+	lw t1, 0(a0)
+1:	fscsr t1
+	li a0, 0
+	ret
+
+#endif
diff --git a/src/internal/atomic.h b/src/internal/atomic.h
index 96c1552d..8f71c8cd 100644
--- a/src/internal/atomic.h
+++ b/src/internal/atomic.h
@@ -194,7 +194,7 @@ static inline void a_store(volatile int *p, int v)
 
 #ifndef a_barrier
 #define a_barrier a_barrier
-static void a_barrier()
+static inline void a_barrier()
 {
 	volatile int tmp = 0;
 	a_cas(&tmp, 0, 0);
diff --git a/src/internal/dynlink.h b/src/internal/dynlink.h
index 51c0639f..40c743e2 100644
--- a/src/internal/dynlink.h
+++ b/src/internal/dynlink.h
@@ -73,6 +73,10 @@ struct fdpic_dummy_loadmap {
 #define DL_NOMMU_SUPPORT 0
 #endif
 
+#ifndef TLSDESC_BACKWARDS
+#define TLSDESC_BACKWARDS 0
+#endif
+
 #if !DL_FDPIC
 #define IS_RELATIVE(x,s) ( \
 	(R_TYPE(x) == REL_RELATIVE) || \
@@ -92,8 +96,12 @@ struct fdpic_dummy_loadmap {
 #define DT_DEBUG_INDIRECT 0
 #endif
 
+#ifndef DT_DEBUG_INDIRECT_REL
+#define DT_DEBUG_INDIRECT_REL 0
+#endif
+
 #define AUX_CNT 32
-#define DYN_CNT 32
+#define DYN_CNT 37
 
 typedef void (*stage2_func)(unsigned char *, size_t *);
 
diff --git a/src/internal/emulate_wait4.c b/src/internal/emulate_wait4.c
new file mode 100644
index 00000000..f6303412
--- /dev/null
+++ b/src/internal/emulate_wait4.c
@@ -0,0 +1,55 @@
+#include <sys/wait.h>
+#include "syscall.h"
+
+#ifndef SYS_wait4
+hidden long __emulate_wait4(int pid, int *status, int options, void *kru, int cp)
+{
+	idtype_t t;
+	int r;
+	siginfo_t info;
+
+	info.si_pid = 0;
+	if (pid < -1) {
+		t = P_PGID;
+		pid = -pid;
+	} else if (pid == -1) {
+		t = P_ALL;
+	} else if (pid == 0) {
+		t = P_PGID;
+	} else {
+		t = P_PID;
+	}
+
+	if (cp) r = __syscall_cp(SYS_waitid, t, pid, &info, options|WEXITED, kru);
+	else r = __syscall(SYS_waitid, t, pid, &info, options|WEXITED, kru);
+
+	if (r<0) return r;
+
+	if (info.si_pid && status) {
+		int sw=0;
+		switch (info.si_code) {
+		case CLD_CONTINUED:
+			sw = 0xffff;
+			break;
+		case CLD_DUMPED:
+			sw = info.si_status&0x7f | 0x80;
+			break;
+		case CLD_EXITED:
+			sw = (info.si_status&0xff) << 8;
+			break;
+		case CLD_KILLED:
+			sw = info.si_status&0x7f;
+			break;
+		case CLD_STOPPED:
+		case CLD_TRAPPED:
+			/* see ptrace(2); the high bits of si_status can contain */
+			/* PTRACE_EVENT_ values which must be preserved */
+			sw = (info.si_status << 8) + 0x7f;
+			break;
+		}
+		*status = sw;
+	}
+
+	return info.si_pid;
+}
+#endif
diff --git a/src/internal/fork_impl.h b/src/internal/fork_impl.h
index 5892c13b..f995fce2 100644
--- a/src/internal/fork_impl.h
+++ b/src/internal/fork_impl.h
@@ -2,7 +2,6 @@
 
 extern hidden volatile int *const __at_quick_exit_lockptr;
 extern hidden volatile int *const __atexit_lockptr;
-extern hidden volatile int *const __dlerror_lockptr;
 extern hidden volatile int *const __gettext_lockptr;
 extern hidden volatile int *const __locale_lockptr;
 extern hidden volatile int *const __random_lockptr;
@@ -17,3 +16,6 @@ extern hidden volatile int *const __vmlock_lockptr;
 
 hidden void __malloc_atfork(int);
 hidden void __ldso_atfork(int);
+hidden void __pthread_key_atfork(int);
+
+hidden void __post_Fork(int);
diff --git a/src/internal/ksigaction.h b/src/internal/ksigaction.h
index 8ebd5938..ef333f33 100644
--- a/src/internal/ksigaction.h
+++ b/src/internal/ksigaction.h
@@ -6,8 +6,13 @@
 struct k_sigaction {
 	void (*handler)(int);
 	unsigned long flags;
+#ifdef SA_RESTORER
 	void (*restorer)(void);
+#endif
 	unsigned mask[2];
+#ifndef SA_RESTORER
+	void *unused;
+#endif
 };
 
 hidden void __restore(), __restore_rt();
diff --git a/src/internal/syscall.h b/src/internal/syscall.h
index 4f41e1dc..33d981f9 100644
--- a/src/internal/syscall.h
+++ b/src/internal/syscall.h
@@ -58,7 +58,7 @@ hidden long __syscall_ret(unsigned long),
 #define __syscall_cp(...) __SYSCALL_DISP(__syscall_cp,__VA_ARGS__)
 #define syscall_cp(...) __syscall_ret(__syscall_cp(__VA_ARGS__))
 
-static inline long __alt_socketcall(int sys, int sock, int cp, long a, long b, long c, long d, long e, long f)
+static inline long __alt_socketcall(int sys, int sock, int cp, syscall_arg_t a, syscall_arg_t b, syscall_arg_t c, syscall_arg_t d, syscall_arg_t e, syscall_arg_t f)
 {
 	long r;
 	if (cp) r = __syscall_cp(sys, a, b, c, d, e, f);
@@ -71,9 +71,9 @@ static inline long __alt_socketcall(int sys, int sock, int cp, long a, long b, l
 	return r;
 }
 #define __socketcall(nm, a, b, c, d, e, f) __alt_socketcall(SYS_##nm, __SC_##nm, 0, \
-	(long)(a), (long)(b), (long)(c), (long)(d), (long)(e), (long)(f))
+	__scc(a), __scc(b), __scc(c), __scc(d), __scc(e), __scc(f))
 #define __socketcall_cp(nm, a, b, c, d, e, f) __alt_socketcall(SYS_##nm, __SC_##nm, 1, \
-	(long)(a), (long)(b), (long)(c), (long)(d), (long)(e), (long)(f))
+	__scc(a), __scc(b), __scc(c), __scc(d), __scc(e), __scc(f))
 
 /* fixup legacy 16-bit junk */
 
@@ -391,6 +391,18 @@ static inline long __alt_socketcall(int sys, int sock, int cp, long a, long b, l
 #define __sys_open_cp(...) __SYSCALL_DISP(__sys_open_cp,,__VA_ARGS__)
 #define sys_open_cp(...) __syscall_ret(__sys_open_cp(__VA_ARGS__))
 
+#ifdef SYS_wait4
+#define __sys_wait4(a,b,c,d) __syscall(SYS_wait4,a,b,c,d)
+#define __sys_wait4_cp(a,b,c,d) __syscall_cp(SYS_wait4,a,b,c,d)
+#else
+hidden long __emulate_wait4(int, int *, int, void *, int);
+#define __sys_wait4(a,b,c,d) __emulate_wait4(a,b,c,d,0)
+#define __sys_wait4_cp(a,b,c,d) __emulate_wait4(a,b,c,d,1)
+#endif
+
+#define sys_wait4(a,b,c,d) __syscall_ret(__sys_wait4(a,b,c,d))
+#define sys_wait4_cp(a,b,c,d) __syscall_ret(__sys_wait4_cp(a,b,c,d))
+
 hidden void __procfdname(char __buf[static 15+3*sizeof(int)], unsigned);
 
 hidden void *__vdsosym(const char *, const char *);
diff --git a/src/ipc/semtimedop.c b/src/ipc/semtimedop.c
index 1632e7b0..a104af21 100644
--- a/src/ipc/semtimedop.c
+++ b/src/ipc/semtimedop.c
@@ -7,7 +7,8 @@
 #define IS32BIT(x) !((x)+0x80000000ULL>>32)
 #define CLAMP(x) (int)(IS32BIT(x) ? (x) : 0x7fffffffU+((0ULL+(x))>>63))
 
-#if !defined(SYS_semtimedop) && !defined(SYS_ipc)
+#if !defined(SYS_semtimedop) && !defined(SYS_ipc) || \
+	SYS_semtimedop == SYS_semtimedop_time64
 #define NO_TIME32 1
 #else
 #define NO_TIME32 0
diff --git a/src/ldso/dlerror.c b/src/ldso/dlerror.c
index afe59253..dae0f3a9 100644
--- a/src/ldso/dlerror.c
+++ b/src/ldso/dlerror.c
@@ -3,8 +3,7 @@
 #include <stdarg.h>
 #include "pthread_impl.h"
 #include "dynlink.h"
-#include "lock.h"
-#include "fork_impl.h"
+#include "atomic.h"
 
 #define malloc __libc_malloc
 #define calloc __libc_calloc
@@ -23,28 +22,31 @@ char *dlerror()
 		return s;
 }
 
-static volatile int freebuf_queue_lock[1];
-static void **freebuf_queue;
-volatile int *const __dlerror_lockptr = freebuf_queue_lock;
+/* Atomic singly-linked list, used to store list of thread-local dlerror
+ * buffers for deferred free. They cannot be freed at thread exit time
+ * because, by the time it's known they can be freed, the exiting thread
+ * is in a highly restrictive context where it cannot call (even the
+ * libc-internal) free. It also can't take locks; thus the atomic list. */
+
+static void *volatile freebuf_queue;
 
 void __dl_thread_cleanup(void)
 {
 	pthread_t self = __pthread_self();
-	if (self->dlerror_buf && self->dlerror_buf != (void *)-1) {
-		LOCK(freebuf_queue_lock);
-		void **p = (void **)self->dlerror_buf;
-		*p = freebuf_queue;
-		freebuf_queue = p;
-		UNLOCK(freebuf_queue_lock);
-	}
+	if (!self->dlerror_buf || self->dlerror_buf == (void *)-1)
+		return;
+	void *h;
+	do {
+		h = freebuf_queue;
+		*(void **)self->dlerror_buf = h;
+	} while (a_cas_p(&freebuf_queue, h, self->dlerror_buf) != h);
 }
 
 hidden void __dl_vseterr(const char *fmt, va_list ap)
 {
-	LOCK(freebuf_queue_lock);
-	void **q = freebuf_queue;
-	freebuf_queue = 0;
-	UNLOCK(freebuf_queue_lock);
+	void **q;
+	do q = freebuf_queue;
+	while (q && a_cas_p(&freebuf_queue, q, 0) != q);
 
 	while (q) {
 		void **p = *q;
diff --git a/src/ldso/loongarch64/dlsym.s b/src/ldso/loongarch64/dlsym.s
new file mode 100644
index 00000000..26fabcdb
--- /dev/null
+++ b/src/ldso/loongarch64/dlsym.s
@@ -0,0 +1,7 @@
+.global dlsym
+.hidden __dlsym
+.type   dlsym,@function
+dlsym:
+	move      $a2, $ra
+	la.global $t0, __dlsym
+	jr        $t0
diff --git a/src/ldso/riscv32/dlsym.s b/src/ldso/riscv32/dlsym.s
new file mode 100644
index 00000000..2bafd72d
--- /dev/null
+++ b/src/ldso/riscv32/dlsym.s
@@ -0,0 +1,6 @@
+.global dlsym
+.hidden __dlsym
+.type dlsym, %function
+dlsym:
+	mv a2, ra
+	tail __dlsym
diff --git a/src/ldso/riscv64/tlsdesc.s b/src/ldso/riscv64/tlsdesc.s
new file mode 100644
index 00000000..bef8b322
--- /dev/null
+++ b/src/ldso/riscv64/tlsdesc.s
@@ -0,0 +1,32 @@
+.text
+.global __tlsdesc_static
+.hidden __tlsdesc_static
+.type __tlsdesc_static,%function
+__tlsdesc_static:
+	ld a0,8(a0)
+	jr t0
+
+.global __tlsdesc_dynamic
+.hidden __tlsdesc_dynamic
+.type __tlsdesc_dynamic,%function
+__tlsdesc_dynamic:
+	add sp,sp,-16
+	sd t1,(sp)
+	sd t2,8(sp)
+
+	ld t2,-8(tp) # t2=dtv
+
+	ld a0,8(a0)  # a0=&{modidx,off}
+	ld t1,8(a0)  # t1=off
+	ld a0,(a0)   # a0=modidx
+	sll a0,a0,3  # a0=8*modidx
+
+	add a0,a0,t2 # a0=dtv+8*modidx
+	ld a0,(a0)   # a0=dtv[modidx]
+	add a0,a0,t1 # a0=dtv[modidx]+off
+	sub a0,a0,tp # a0=dtv[modidx]+off-tp
+
+	ld t1,(sp)
+	ld t2,8(sp)
+	add sp,sp,16
+	jr t0
diff --git a/src/ldso/sh/dlsym.s b/src/ldso/sh/dlsym.s
index 11a6fff5..34f3c35c 100644
--- a/src/ldso/sh/dlsym.s
+++ b/src/ldso/sh/dlsym.s
@@ -5,7 +5,7 @@
 dlsym:
 	mov.l L1, r0
 1:	braf  r0
-	 mov.l @r15, r6
+	 sts pr, r6
 
 .align 2
 L1:	.long __dlsym@PLT-(1b+4-.)
diff --git a/src/legacy/ftw.c b/src/legacy/ftw.c
index 506bd29c..e757fc6f 100644
--- a/src/legacy/ftw.c
+++ b/src/legacy/ftw.c
@@ -7,5 +7,3 @@ int ftw(const char *path, int (*fn)(const char *, const struct stat *, int), int
 	 * actually undefined, but works on all real-world machines. */
 	return nftw(path, (int (*)())fn, fd_limit, FTW_PHYS);
 }
-
-weak_alias(ftw, ftw64);
diff --git a/src/linux/cache.c b/src/linux/cache.c
index 0eb051c2..e76f7812 100644
--- a/src/linux/cache.c
+++ b/src/linux/cache.c
@@ -21,7 +21,7 @@ weak_alias(__cachectl, cachectl);
 #ifdef SYS_riscv_flush_icache
 
 #define VDSO_FLUSH_ICACHE_SYM "__vdso_flush_icache"
-#define VDSO_FLUSH_ICACHE_VER "LINUX_4.5"
+#define VDSO_FLUSH_ICACHE_VER "LINUX_4.15"
 
 static void *volatile vdso_func;
 
@@ -45,6 +45,7 @@ int __riscv_flush_icache(void *start, void *end, unsigned long int flags)
 		if (!r) return r;
 		if (r != -ENOSYS) return __syscall_ret(r);
 	}
+	return syscall(SYS_riscv_flush_icache, start, end, flags);
 }
 weak_alias(__riscv_flush_icache, riscv_flush_icache);
 #endif
diff --git a/src/linux/clone.c b/src/linux/clone.c
index 8c1af7d3..257c1cec 100644
--- a/src/linux/clone.c
+++ b/src/linux/clone.c
@@ -4,18 +4,62 @@
 #include <sched.h>
 #include "pthread_impl.h"
 #include "syscall.h"
+#include "lock.h"
+#include "fork_impl.h"
+
+struct clone_start_args {
+	int (*func)(void *);
+	void *arg;
+	sigset_t sigmask;
+};
+
+static int clone_start(void *arg)
+{
+	struct clone_start_args *csa = arg;
+	__post_Fork(0);
+	__restore_sigs(&csa->sigmask);
+	return csa->func(csa->arg);
+}
 
 int clone(int (*func)(void *), void *stack, int flags, void *arg, ...)
 {
+	struct clone_start_args csa;
 	va_list ap;
-	pid_t *ptid, *ctid;
-	void  *tls;
+	pid_t *ptid = 0, *ctid = 0;
+	void  *tls = 0;
+
+	/* Flags that produce an invalid thread/TLS state are disallowed. */
+	int badflags = CLONE_THREAD | CLONE_SETTLS | CLONE_CHILD_CLEARTID;
+
+	if ((flags & badflags) || !stack)
+		return __syscall_ret(-EINVAL);
 
 	va_start(ap, arg);
-	ptid = va_arg(ap, pid_t *);
-	tls  = va_arg(ap, void *);
-	ctid = va_arg(ap, pid_t *);
+	if (flags & (CLONE_PIDFD | CLONE_PARENT_SETTID | CLONE_CHILD_SETTID))
+	 	ptid = va_arg(ap, pid_t *);
+	if (flags & CLONE_CHILD_SETTID) {
+		tls = va_arg(ap, void *);
+		ctid = va_arg(ap, pid_t *);
+	}
 	va_end(ap);
 
-	return __syscall_ret(__clone(func, stack, flags, arg, ptid, tls, ctid));
+	/* If CLONE_VM is used, it's impossible to give the child a consistent
+	 * thread structure. In this case, the best we can do is assume the
+	 * caller is content with an extremely restrictive execution context
+	 * like the one vfork() would provide. */
+	if (flags & CLONE_VM) return __syscall_ret(
+		__clone(func, stack, flags, arg, ptid, tls, ctid));
+
+	__block_all_sigs(&csa.sigmask);
+	LOCK(__abort_lock);
+
+	/* Setup the a wrapper start function for the child process to do
+	 * mimic _Fork in producing a consistent execution state. */
+	csa.func = func;
+	csa.arg = arg;
+	int ret = __clone(clone_start, stack, flags, &csa, ptid, tls, ctid);
+
+	__post_Fork(ret);
+	__restore_sigs(&csa.sigmask);
+	return __syscall_ret(ret);
 }
diff --git a/src/linux/epoll.c b/src/linux/epoll.c
index 93baa814..e56e8f4c 100644
--- a/src/linux/epoll.c
+++ b/src/linux/epoll.c
@@ -5,6 +5,7 @@
 
 int epoll_create(int size)
 {
+	if (size<=0) return __syscall_ret(-EINVAL);
 	return epoll_create1(0);
 }
 
diff --git a/src/linux/fallocate.c b/src/linux/fallocate.c
index 7d68bc8f..9146350e 100644
--- a/src/linux/fallocate.c
+++ b/src/linux/fallocate.c
@@ -7,6 +7,3 @@ int fallocate(int fd, int mode, off_t base, off_t len)
 	return syscall(SYS_fallocate, fd, mode, __SYSCALL_LL_E(base),
 		__SYSCALL_LL_E(len));
 }
-
-#undef fallocate64
-weak_alias(fallocate, fallocate64);
diff --git a/src/linux/getdents.c b/src/linux/getdents.c
index 796c1e5c..97f76e14 100644
--- a/src/linux/getdents.c
+++ b/src/linux/getdents.c
@@ -8,5 +8,3 @@ int getdents(int fd, struct dirent *buf, size_t len)
 	if (len>INT_MAX) len = INT_MAX;
 	return syscall(SYS_getdents, fd, buf, len);
 }
-
-weak_alias(getdents, getdents64);
diff --git a/src/linux/membarrier.c b/src/linux/membarrier.c
index 343f7360..f64fe7e1 100644
--- a/src/linux/membarrier.c
+++ b/src/linux/membarrier.c
@@ -35,7 +35,7 @@ int __membarrier(int cmd, int flags)
 		__tl_lock();
 		sem_init(&barrier_sem, 0, 0);
 		struct sigaction sa = {
-			.sa_flags = SA_RESTART,
+			.sa_flags = SA_RESTART | SA_ONSTACK,
 			.sa_handler = bcast_barrier
 		};
 		memset(&sa.sa_mask, -1, sizeof sa.sa_mask);
diff --git a/src/linux/preadv2.c b/src/linux/preadv2.c
new file mode 100644
index 00000000..5e7ab70f
--- /dev/null
+++ b/src/linux/preadv2.c
@@ -0,0 +1,17 @@
+#define _GNU_SOURCE
+#include <sys/uio.h>
+#include <unistd.h>
+#include "syscall.h"
+
+ssize_t preadv2(int fd, const struct iovec *iov, int count, off_t ofs, int flags)
+{
+#ifdef SYS_preadv
+	if (!flags) {
+		if (ofs==-1) return readv(fd, iov, count);
+		return syscall_cp(SYS_preadv, fd, iov, count,
+			(long)(ofs), (long)(ofs>>32));
+	}
+#endif
+	return syscall_cp(SYS_preadv2, fd, iov, count,
+		(long)(ofs), (long)(ofs>>32), flags);
+}
diff --git a/src/linux/prlimit.c b/src/linux/prlimit.c
index 3df9ffba..fcf45aab 100644
--- a/src/linux/prlimit.c
+++ b/src/linux/prlimit.c
@@ -21,6 +21,3 @@ int prlimit(pid_t pid, int resource, const struct rlimit *new_limit, struct rlim
 	}
 	return r;
 }
-
-#undef prlimit64
-weak_alias(prlimit, prlimit64);
diff --git a/src/linux/pwritev2.c b/src/linux/pwritev2.c
new file mode 100644
index 00000000..ece90d7c
--- /dev/null
+++ b/src/linux/pwritev2.c
@@ -0,0 +1,17 @@
+#define _GNU_SOURCE
+#include <sys/uio.h>
+#include <unistd.h>
+#include "syscall.h"
+
+ssize_t pwritev2(int fd, const struct iovec *iov, int count, off_t ofs, int flags)
+{
+#ifdef SYS_pwritev
+	if (!flags) {
+		if (ofs==-1) return writev(fd, iov, count);
+		return syscall_cp(SYS_pwritev, fd, iov, count,
+			(long)(ofs), (long)(ofs>>32));
+	}
+#endif
+	return syscall_cp(SYS_pwritev2, fd, iov, count,
+		(long)(ofs), (long)(ofs>>32), flags);
+}
diff --git a/src/linux/sendfile.c b/src/linux/sendfile.c
index 9afe6dd6..fc1577d3 100644
--- a/src/linux/sendfile.c
+++ b/src/linux/sendfile.c
@@ -5,5 +5,3 @@ ssize_t sendfile(int out_fd, int in_fd, off_t *ofs, size_t count)
 {
 	return syscall(SYS_sendfile, out_fd, in_fd, ofs, count);
 }
-
-weak_alias(sendfile, sendfile64);
diff --git a/src/linux/statx.c b/src/linux/statx.c
new file mode 100644
index 00000000..4616bff4
--- /dev/null
+++ b/src/linux/statx.c
@@ -0,0 +1,42 @@
+#define _GNU_SOURCE
+#include <sys/stat.h>
+#include <string.h>
+#include <syscall.h>
+#include <sys/sysmacros.h>
+#include <errno.h>
+
+int statx(int dirfd, const char *restrict path, int flags, unsigned mask, struct statx *restrict stx)
+{
+	int ret = __syscall(SYS_statx, dirfd, path, flags, mask, stx);
+
+#ifndef SYS_fstatat
+	return __syscall_ret(ret);
+#endif
+
+	if (ret != -ENOSYS) return __syscall_ret(ret);
+
+	struct stat st;
+	ret = fstatat(dirfd, path, &st, flags);
+	if (ret) return ret;
+
+	stx->stx_dev_major = major(st.st_dev);
+	stx->stx_dev_minor = minor(st.st_dev);
+	stx->stx_ino = st.st_ino;
+	stx->stx_mode = st.st_mode;
+	stx->stx_nlink = st.st_nlink;
+	stx->stx_uid = st.st_uid;
+	stx->stx_gid = st.st_gid;
+	stx->stx_size = st.st_size;
+	stx->stx_blksize = st.st_blksize;
+	stx->stx_blocks = st.st_blocks;
+	stx->stx_atime.tv_sec = st.st_atim.tv_sec;
+	stx->stx_atime.tv_nsec = st.st_atim.tv_nsec;
+	stx->stx_mtime.tv_sec = st.st_mtim.tv_sec;
+	stx->stx_mtime.tv_nsec = st.st_mtim.tv_nsec;
+	stx->stx_ctime.tv_sec = st.st_ctim.tv_sec;
+	stx->stx_ctime.tv_nsec = st.st_ctim.tv_nsec;
+	stx->stx_btime = (struct statx_timestamp){.tv_sec=0, .tv_nsec=0};
+	stx->stx_mask = STATX_BASIC_STATS;
+
+	return 0;
+}
diff --git a/src/linux/wait4.c b/src/linux/wait4.c
index 83650e34..fb08c0d0 100644
--- a/src/linux/wait4.c
+++ b/src/linux/wait4.c
@@ -12,7 +12,7 @@ pid_t wait4(pid_t pid, int *status, int options, struct rusage *ru)
 	if (ru) {
 		long long kru64[18];
 		r = __syscall(SYS_wait4_time64, pid, status, options, kru64);
-		if (!r) {
+		if (r > 0) {
 			ru->ru_utime = (struct timeval)
 				{ .tv_sec = kru64[0], .tv_usec = kru64[1] };
 			ru->ru_stime = (struct timeval)
@@ -26,7 +26,7 @@ pid_t wait4(pid_t pid, int *status, int options, struct rusage *ru)
 	}
 #endif
 	char *dest = ru ? (char *)&ru->ru_maxrss - 4*sizeof(long) : 0;
-	r = __syscall(SYS_wait4, pid, status, options, dest);
+	r = __sys_wait4(pid, status, options, dest);
 	if (r>0 && ru && sizeof(time_t) > sizeof(long)) {
 		long kru[4];
 		memcpy(kru, dest, 4*sizeof(long));
diff --git a/src/locale/iconv.c b/src/locale/iconv.c
index 3047c27b..7fb2e1ef 100644
--- a/src/locale/iconv.c
+++ b/src/locale/iconv.c
@@ -49,10 +49,10 @@ static const unsigned char charmaps[] =
 "ucs4\0utf32\0\0\313"
 "ucs2\0\0\314"
 "eucjp\0\0\320"
-"shiftjis\0sjis\0\0\321"
+"shiftjis\0sjis\0cp932\0\0\321"
 "iso2022jp\0\0\322"
 "gb18030\0\0\330"
-"gbk\0\0\331"
+"gbk\0cp936\0windows936\0\0\331"
 "gb2312\0\0\332"
 "big5\0bigfive\0cp950\0big5hkscs\0\0\340"
 "euckr\0ksc5601\0ksx1001\0cp949\0\0\350"
@@ -340,6 +340,7 @@ size_t iconv(iconv_t cd, char **restrict in, size_t *restrict inb, char **restri
 				c++;
 				d -= 159;
 			}
+			if (c>=84) goto ilseq;
 			c = jis0208[c][d];
 			if (!c) goto ilseq;
 			break;
@@ -403,6 +404,10 @@ size_t iconv(iconv_t cd, char **restrict in, size_t *restrict inb, char **restri
 			if (c < 128) break;
 			if (c < 0xa1) goto ilseq;
 		case GBK:
+			if (c == 128) {
+				c = 0x20ac;
+				break;
+			}
 		case GB18030:
 			if (c < 128) break;
 			c -= 0x81;
diff --git a/src/malloc/mallocng/free.c b/src/malloc/mallocng/free.c
index 418a085c..43f32aad 100644
--- a/src/malloc/mallocng/free.c
+++ b/src/malloc/mallocng/free.c
@@ -119,7 +119,7 @@ void free(void *p)
 	if (((uintptr_t)(start-1) ^ (uintptr_t)end) >= 2*PGSZ && g->last_idx) {
 		unsigned char *base = start + (-(uintptr_t)start & (PGSZ-1));
 		size_t len = (end-base) & -PGSZ;
-		if (len) {
+		if (len && USE_MADV_FREE) {
 			int e = errno;
 			madvise(base, len, MADV_FREE);
 			errno = e;
diff --git a/src/malloc/mallocng/glue.h b/src/malloc/mallocng/glue.h
index 151c48b8..77f4c812 100644
--- a/src/malloc/mallocng/glue.h
+++ b/src/malloc/mallocng/glue.h
@@ -24,6 +24,8 @@
 #define realloc __libc_realloc
 #define free __libc_free
 
+#define USE_MADV_FREE 0
+
 #if USE_REAL_ASSERT
 #include <assert.h>
 #else
diff --git a/src/math/acoshl.c b/src/math/acoshl.c
index 8d4b43f6..943cec17 100644
--- a/src/math/acoshl.c
+++ b/src/math/acoshl.c
@@ -10,14 +10,18 @@ long double acoshl(long double x)
 long double acoshl(long double x)
 {
 	union ldshape u = {x};
-	int e = u.i.se & 0x7fff;
+	int e = u.i.se;
 
 	if (e < 0x3fff + 1)
-		/* |x| < 2, invalid if x < 1 or nan */
+		/* 0 <= x < 2, invalid if x < 1 */
 		return log1pl(x-1 + sqrtl((x-1)*(x-1)+2*(x-1)));
 	if (e < 0x3fff + 32)
-		/* |x| < 0x1p32 */
+		/* 2 <= x < 0x1p32 */
 		return logl(2*x - 1/(x+sqrtl(x*x-1)));
+	if (e & 0x8000)
+		/* x < 0 or x = -0, invalid */
+		return (x - x) / (x - x);
+	/* 0x1p32 <= x or nan */
 	return logl(x) + 0.693147180559945309417232121458176568L;
 }
 #elif LDBL_MANT_DIG == 113 && LDBL_MAX_EXP == 16384
diff --git a/src/math/logf.c b/src/math/logf.c
index 7ee5d7fe..e4c2237c 100644
--- a/src/math/logf.c
+++ b/src/math/logf.c
@@ -53,7 +53,7 @@ float logf(float x)
 	tmp = ix - OFF;
 	i = (tmp >> (23 - LOGF_TABLE_BITS)) % N;
 	k = (int32_t)tmp >> 23; /* arithmetic shift */
-	iz = ix - (tmp & 0x1ff << 23);
+	iz = ix - (tmp & 0xff800000);
 	invc = T[i].invc;
 	logc = T[i].logc;
 	z = (double_t)asfloat(iz);
diff --git a/src/math/powl.c b/src/math/powl.c
index 5b6da07b..6f64ea71 100644
--- a/src/math/powl.c
+++ b/src/math/powl.c
@@ -212,25 +212,33 @@ long double powl(long double x, long double y)
 	}
 	if (x == 1.0)
 		return 1.0; /* 1**y = 1, even if y is nan */
-	if (x == -1.0 && !isfinite(y))
-		return 1.0; /* -1**inf = 1 */
 	if (y == 0.0)
 		return 1.0; /* x**0 = 1, even if x is nan */
 	if (y == 1.0)
 		return x;
-	if (y >= LDBL_MAX) {
-		if (x > 1.0 || x < -1.0)
-			return INFINITY;
-		if (x != 0.0)
-			return 0.0;
-	}
-	if (y <= -LDBL_MAX) {
-		if (x > 1.0 || x < -1.0)
+	/* if y*log2(x) < log2(LDBL_TRUE_MIN)-1 then x^y uflows to 0
+	   if y*log2(x) > -log2(LDBL_TRUE_MIN)+1 > LDBL_MAX_EXP then x^y oflows
+	   if |x|!=1 then |log2(x)| > |log(x)| > LDBL_EPSILON/2 so
+	   x^y oflows/uflows if |y|*LDBL_EPSILON/2 > -log2(LDBL_TRUE_MIN)+1 */
+	if (fabsl(y) > 2*(-LDBL_MIN_EXP+LDBL_MANT_DIG+1)/LDBL_EPSILON) {
+		/* y is not an odd int */
+		if (x == -1.0)
+			return 1.0;
+		if (y == INFINITY) {
+			if (x > 1.0 || x < -1.0)
+				return INFINITY;
 			return 0.0;
-		if (x != 0.0 || y == -INFINITY)
+		}
+		if (y == -INFINITY) {
+			if (x > 1.0 || x < -1.0)
+				return 0.0;
 			return INFINITY;
+		}
+		if ((x > 1.0 || x < -1.0) == (y > 0))
+			return huge * huge;
+		return twom10000 * twom10000;
 	}
-	if (x >= LDBL_MAX) {
+	if (x == INFINITY) {
 		if (y > 0.0)
 			return INFINITY;
 		return 0.0;
@@ -253,7 +261,7 @@ long double powl(long double x, long double y)
 			yoddint = 1;
 	}
 
-	if (x <= -LDBL_MAX) {
+	if (x == -INFINITY) {
 		if (y > 0.0) {
 			if (yoddint)
 				return -INFINITY;
diff --git a/src/math/riscv32/copysign.c b/src/math/riscv32/copysign.c
new file mode 100644
index 00000000..c7854178
--- /dev/null
+++ b/src/math/riscv32/copysign.c
@@ -0,0 +1,15 @@
+#include <math.h>
+
+#if __riscv_flen >= 64
+
+double copysign(double x, double y)
+{
+	__asm__ ("fsgnj.d %0, %1, %2" : "=f"(x) : "f"(x), "f"(y));
+	return x;
+}
+
+#else
+
+#include "../copysign.c"
+
+#endif
diff --git a/src/math/riscv32/copysignf.c b/src/math/riscv32/copysignf.c
new file mode 100644
index 00000000..a125611a
--- /dev/null
+++ b/src/math/riscv32/copysignf.c
@@ -0,0 +1,15 @@
+#include <math.h>
+
+#if __riscv_flen >= 32
+
+float copysignf(float x, float y)
+{
+	__asm__ ("fsgnj.s %0, %1, %2" : "=f"(x) : "f"(x), "f"(y));
+	return x;
+}
+
+#else
+
+#include "../copysignf.c"
+
+#endif
diff --git a/src/math/riscv32/fabs.c b/src/math/riscv32/fabs.c
new file mode 100644
index 00000000..5290b6f0
--- /dev/null
+++ b/src/math/riscv32/fabs.c
@@ -0,0 +1,15 @@
+#include <math.h>
+
+#if __riscv_flen >= 64
+
+double fabs(double x)
+{
+	__asm__ ("fabs.d %0, %1" : "=f"(x) : "f"(x));
+	return x;
+}
+
+#else
+
+#include "../fabs.c"
+
+#endif
diff --git a/src/math/riscv32/fabsf.c b/src/math/riscv32/fabsf.c
new file mode 100644
index 00000000..f5032e35
--- /dev/null
+++ b/src/math/riscv32/fabsf.c
@@ -0,0 +1,15 @@
+#include <math.h>
+
+#if __riscv_flen >= 32
+
+float fabsf(float x)
+{
+	__asm__ ("fabs.s %0, %1" : "=f"(x) : "f"(x));
+	return x;
+}
+
+#else
+
+#include "../fabsf.c"
+
+#endif
diff --git a/src/math/riscv32/fma.c b/src/math/riscv32/fma.c
new file mode 100644
index 00000000..99b05713
--- /dev/null
+++ b/src/math/riscv32/fma.c
@@ -0,0 +1,15 @@
+#include <math.h>
+
+#if __riscv_flen >= 64
+
+double fma(double x, double y, double z)
+{
+	__asm__ ("fmadd.d %0, %1, %2, %3" : "=f"(x) : "f"(x), "f"(y), "f"(z));
+	return x;
+}
+
+#else
+
+#include "../fma.c"
+
+#endif
diff --git a/src/math/riscv32/fmaf.c b/src/math/riscv32/fmaf.c
new file mode 100644
index 00000000..f9dc47ed
--- /dev/null
+++ b/src/math/riscv32/fmaf.c
@@ -0,0 +1,15 @@
+#include <math.h>
+
+#if __riscv_flen >= 32
+
+float fmaf(float x, float y, float z)
+{
+	__asm__ ("fmadd.s %0, %1, %2, %3" : "=f"(x) : "f"(x), "f"(y), "f"(z));
+	return x;
+}
+
+#else
+
+#include "../fmaf.c"
+
+#endif
diff --git a/src/math/riscv32/fmax.c b/src/math/riscv32/fmax.c
new file mode 100644
index 00000000..023709cd
--- /dev/null
+++ b/src/math/riscv32/fmax.c
@@ -0,0 +1,15 @@
+#include <math.h>
+
+#if __riscv_flen >= 64
+
+double fmax(double x, double y)
+{
+	__asm__ ("fmax.d %0, %1, %2" : "=f"(x) : "f"(x), "f"(y));
+	return x;
+}
+
+#else
+
+#include "../fmax.c"
+
+#endif
diff --git a/src/math/riscv32/fmaxf.c b/src/math/riscv32/fmaxf.c
new file mode 100644
index 00000000..863d2bd1
--- /dev/null
+++ b/src/math/riscv32/fmaxf.c
@@ -0,0 +1,15 @@
+#include <math.h>
+
+#if __riscv_flen >= 32
+
+float fmaxf(float x, float y)
+{
+	__asm__ ("fmax.s %0, %1, %2" : "=f"(x) : "f"(x), "f"(y));
+	return x;
+}
+
+#else
+
+#include "../fmaxf.c"
+
+#endif
diff --git a/src/math/riscv32/fmin.c b/src/math/riscv32/fmin.c
new file mode 100644
index 00000000..a4e3b067
--- /dev/null
+++ b/src/math/riscv32/fmin.c
@@ -0,0 +1,15 @@
+#include <math.h>
+
+#if __riscv_flen >= 64
+
+double fmin(double x, double y)
+{
+	__asm__ ("fmin.d %0, %1, %2" : "=f"(x) : "f"(x), "f"(y));
+	return x;
+}
+
+#else
+
+#include "../fmin.c"
+
+#endif
diff --git a/src/math/riscv32/fminf.c b/src/math/riscv32/fminf.c
new file mode 100644
index 00000000..32156e80
--- /dev/null
+++ b/src/math/riscv32/fminf.c
@@ -0,0 +1,15 @@
+#include <math.h>
+
+#if __riscv_flen >= 32
+
+float fminf(float x, float y)
+{
+	__asm__ ("fmin.s %0, %1, %2" : "=f"(x) : "f"(x), "f"(y));
+	return x;
+}
+
+#else
+
+#include "../fminf.c"
+
+#endif
diff --git a/src/math/riscv32/sqrt.c b/src/math/riscv32/sqrt.c
new file mode 100644
index 00000000..867a504c
--- /dev/null
+++ b/src/math/riscv32/sqrt.c
@@ -0,0 +1,15 @@
+#include <math.h>
+
+#if __riscv_flen >= 64
+
+double sqrt(double x)
+{
+	__asm__ ("fsqrt.d %0, %1" : "=f"(x) : "f"(x));
+	return x;
+}
+
+#else
+
+#include "../sqrt.c"
+
+#endif
diff --git a/src/math/riscv32/sqrtf.c b/src/math/riscv32/sqrtf.c
new file mode 100644
index 00000000..610c2cf8
--- /dev/null
+++ b/src/math/riscv32/sqrtf.c
@@ -0,0 +1,15 @@
+#include <math.h>
+
+#if __riscv_flen >= 32
+
+float sqrtf(float x)
+{
+	__asm__ ("fsqrt.s %0, %1" : "=f"(x) : "f"(x));
+	return x;
+}
+
+#else
+
+#include "../sqrtf.c"
+
+#endif
diff --git a/src/math/sqrtl.c b/src/math/sqrtl.c
index 1b9f19c7..a231b3f2 100644
--- a/src/math/sqrtl.c
+++ b/src/math/sqrtl.c
@@ -205,7 +205,7 @@ long double sqrtl(long double x)
 	top = (top + 0x3fff) >> 1;
 
 	/* r ~ 1/sqrt(m) */
-	static const uint64_t three = 0xc0000000;
+	const uint64_t three = 0xc0000000;
 	uint64_t r, s, d, u, i;
 	i = (ix.hi >> 42) % 128;
 	r = (uint32_t)__rsqrt_tab[i] << 16;
@@ -227,7 +227,7 @@ long double sqrtl(long double x)
 	r = mul64(u, r) << 1;
 	/* |r sqrt(m) - 1| < 0x1.c001p-59, switch to 128bit */
 
-	static const u128 threel = {.hi=three<<32, .lo=0};
+	const u128 threel = {.hi=three<<32, .lo=0};
 	u128 rl, sl, dl, ul;
 	rl.hi = r;
 	rl.lo = 0;
diff --git a/src/misc/getopt.c b/src/misc/getopt.c
index c3f66995..b02b81c3 100644
--- a/src/misc/getopt.c
+++ b/src/misc/getopt.c
@@ -87,7 +87,8 @@ int getopt(int argc, char * const argv[], const char *optstring)
 	if (optstring[i] == ':') {
 		optarg = 0;
 		if (optstring[i+1] != ':' || optpos) {
-			optarg = argv[optind++] + optpos;
+			optarg = argv[optind++];
+			if (optpos) optarg += optpos;
 			optpos = 0;
 		}
 		if (optind > argc) {
diff --git a/src/misc/getrlimit.c b/src/misc/getrlimit.c
index bf676307..a5558d81 100644
--- a/src/misc/getrlimit.c
+++ b/src/misc/getrlimit.c
@@ -26,5 +26,3 @@ int getrlimit(int resource, struct rlimit *rlim)
 	return ret;
 #endif
 }
-
-weak_alias(getrlimit, getrlimit64);
diff --git a/src/misc/lockf.c b/src/misc/lockf.c
index 16a80bec..0162442b 100644
--- a/src/misc/lockf.c
+++ b/src/misc/lockf.c
@@ -28,5 +28,3 @@ int lockf(int fd, int op, off_t size)
 	errno = EINVAL;
 	return -1;
 }
-
-weak_alias(lockf, lockf64);
diff --git a/src/misc/mntent.c b/src/misc/mntent.c
index d404fbe3..78bf0cd0 100644
--- a/src/misc/mntent.c
+++ b/src/misc/mntent.c
@@ -20,6 +20,42 @@ int endmntent(FILE *f)
 	return 1;
 }
 
+static char *unescape_ent(char *beg)
+{
+	char *dest = beg;
+	const char *src = beg;
+	while (*src) {
+		const char *val;
+		unsigned char cval = 0;
+		if (*src != '\\') {
+			*dest++ = *src++;
+			continue;
+		}
+		if (src[1] == '\\') {
+			++src;
+			*dest++ = *src++;
+			continue;
+		}
+		val = src + 1;
+		for (int i = 0; i < 3; ++i) {
+			if (*val >= '0' && *val <= '7') {
+				cval <<= 3;
+				cval += *val++ - '0';
+			} else {
+				break;
+			}
+		}
+		if (cval) {
+			*dest++ = cval;
+			src = val;
+		} else {
+			*dest++ = *src++;
+		}
+	}
+	*dest = 0;
+	return beg;
+}
+
 struct mntent *getmntent_r(FILE *f, struct mntent *mnt, char *linebuf, int buflen)
 {
 	int n[8], use_internal = (linebuf == SENTINEL);
@@ -45,7 +81,7 @@ struct mntent *getmntent_r(FILE *f, struct mntent *mnt, char *linebuf, int bufle
 		len = strlen(linebuf);
 		if (len > INT_MAX) continue;
 		for (i = 0; i < sizeof n / sizeof *n; i++) n[i] = len;
-		sscanf(linebuf, " %n%*s%n %n%*s%n %n%*s%n %n%*s%n %d %d",
+		sscanf(linebuf, " %n%*[^ \t]%n %n%*[^ \t]%n %n%*[^ \t]%n %n%*[^ \t]%n %d %d",
 			n, n+1, n+2, n+3, n+4, n+5, n+6, n+7,
 			&mnt->mnt_freq, &mnt->mnt_passno);
 	} while (linebuf[n[0]] == '#' || n[1]==len);
@@ -55,10 +91,10 @@ struct mntent *getmntent_r(FILE *f, struct mntent *mnt, char *linebuf, int bufle
 	linebuf[n[5]] = 0;
 	linebuf[n[7]] = 0;
 
-	mnt->mnt_fsname = linebuf+n[0];
-	mnt->mnt_dir = linebuf+n[2];
-	mnt->mnt_type = linebuf+n[4];
-	mnt->mnt_opts = linebuf+n[6];
+	mnt->mnt_fsname = unescape_ent(linebuf+n[0]);
+	mnt->mnt_dir = unescape_ent(linebuf+n[2]);
+	mnt->mnt_type = unescape_ent(linebuf+n[4]);
+	mnt->mnt_opts = unescape_ent(linebuf+n[6]);
 
 	return mnt;
 }
diff --git a/src/misc/nftw.c b/src/misc/nftw.c
index 8dcff7fe..71bc62ee 100644
--- a/src/misc/nftw.c
+++ b/src/misc/nftw.c
@@ -31,6 +31,8 @@ static int do_nftw(char *path, int (*fn)(const char *, const struct stat *, int,
 	int err;
 	struct FTW lev;
 
+	st.st_dev = st.st_ino = 0;
+
 	if ((flags & FTW_PHYS) ? lstat(path, &st) : stat(path, &st) < 0) {
 		if (!(flags & FTW_PHYS) && errno==ENOENT && !lstat(path, &st))
 			type = FTW_SLN;
@@ -46,7 +48,7 @@ static int do_nftw(char *path, int (*fn)(const char *, const struct stat *, int,
 		type = FTW_F;
 	}
 
-	if ((flags & FTW_MOUNT) && h && st.st_dev != h->dev)
+	if ((flags & FTW_MOUNT) && h && type != FTW_NS && st.st_dev != h->dev)
 		return 0;
 	
 	new.chain = h;
@@ -138,5 +140,3 @@ int nftw(const char *path, int (*fn)(const char *, const struct stat *, int, str
 	pthread_setcancelstate(cs, 0);
 	return r;
 }
-
-weak_alias(nftw, nftw64);
diff --git a/src/misc/setrlimit.c b/src/misc/setrlimit.c
index 5b713cf3..edb413fa 100644
--- a/src/misc/setrlimit.c
+++ b/src/misc/setrlimit.c
@@ -49,5 +49,3 @@ int setrlimit(int resource, const struct rlimit *rlim)
 	return __syscall_ret(ret);
 #endif
 }
-
-weak_alias(setrlimit, setrlimit64);
diff --git a/src/misc/syslog.c b/src/misc/syslog.c
index 7dc0c1be..710202f9 100644
--- a/src/misc/syslog.c
+++ b/src/misc/syslog.c
@@ -11,6 +11,7 @@
 #include <fcntl.h>
 #include "lock.h"
 #include "fork_impl.h"
+#include "locale_impl.h"
 
 static volatile int lock[1];
 static char log_ident[32];
@@ -99,7 +100,7 @@ static void _vsyslog(int priority, const char *message, va_list ap)
 
 	now = time(NULL);
 	gmtime_r(&now, &tm);
-	strftime(timebuf, sizeof timebuf, "%b %e %T", &tm);
+	strftime_l(timebuf, sizeof timebuf, "%b %e %T", &tm, C_LOCALE);
 
 	pid = (log_opt & LOG_PID) ? getpid() : 0;
 	l = snprintf(buf, sizeof buf, "<%d>%s %n%s%s%.0d%s: ",
diff --git a/src/mman/mmap.c b/src/mman/mmap.c
index eff88d82..43e5e029 100644
--- a/src/mman/mmap.c
+++ b/src/mman/mmap.c
@@ -37,5 +37,3 @@ void *__mmap(void *start, size_t len, int prot, int flags, int fd, off_t off)
 }
 
 weak_alias(__mmap, mmap);
-
-weak_alias(mmap, mmap64);
diff --git a/src/mq/mq_notify.c b/src/mq/mq_notify.c
index 221591c7..0e1e6c7a 100644
--- a/src/mq/mq_notify.c
+++ b/src/mq/mq_notify.c
@@ -4,11 +4,14 @@
 #include <sys/socket.h>
 #include <signal.h>
 #include <unistd.h>
+#include <semaphore.h>
 #include "syscall.h"
 
 struct args {
-	pthread_barrier_t barrier;
+	sem_t sem;
 	int sock;
+	mqd_t mqd;
+	int err;
 	const struct sigevent *sev;
 };
 
@@ -20,8 +23,19 @@ static void *start(void *p)
 	int s = args->sock;
 	void (*func)(union sigval) = args->sev->sigev_notify_function;
 	union sigval val = args->sev->sigev_value;
+	struct sigevent sev2;
+	static const char zeros[32];
+	int err;
+
+	sev2.sigev_notify = SIGEV_THREAD;
+	sev2.sigev_signo = s;
+	sev2.sigev_value.sival_ptr = (void *)&zeros;
+
+	args->err = err = -__syscall(SYS_mq_notify, args->mqd, &sev2);
+	sem_post(&args->sem);
+	if (err) return 0;
 
-	pthread_barrier_wait(&args->barrier);
+	pthread_detach(pthread_self());
 	n = recv(s, buf, sizeof(buf), MSG_NOSIGNAL|MSG_WAITALL);
 	close(s);
 	if (n==sizeof buf && buf[sizeof buf - 1] == 1)
@@ -35,8 +49,8 @@ int mq_notify(mqd_t mqd, const struct sigevent *sev)
 	pthread_attr_t attr;
 	pthread_t td;
 	int s;
-	struct sigevent sev2;
-	static const char zeros[32];
+	int cs;
+	sigset_t allmask, origmask;
 
 	if (!sev || sev->sigev_notify != SIGEV_THREAD)
 		return syscall(SYS_mq_notify, mqd, sev);
@@ -44,30 +58,35 @@ int mq_notify(mqd_t mqd, const struct sigevent *sev)
 	s = socket(AF_NETLINK, SOCK_RAW|SOCK_CLOEXEC, 0);
 	if (s < 0) return -1;
 	args.sock = s;
+	args.mqd = mqd;
 
 	if (sev->sigev_notify_attributes) attr = *sev->sigev_notify_attributes;
 	else pthread_attr_init(&attr);
-	pthread_attr_setdetachstate(&attr, PTHREAD_CREATE_DETACHED);
-	pthread_barrier_init(&args.barrier, 0, 2);
+	pthread_attr_setdetachstate(&attr, PTHREAD_CREATE_JOINABLE);
+	sem_init(&args.sem, 0, 0);
 
+	sigfillset(&allmask);
+	pthread_sigmask(SIG_BLOCK, &allmask, &origmask);
 	if (pthread_create(&td, &attr, start, &args)) {
 		__syscall(SYS_close, s);
+		pthread_sigmask(SIG_SETMASK, &origmask, 0);
 		errno = EAGAIN;
 		return -1;
 	}
+	pthread_sigmask(SIG_SETMASK, &origmask, 0);
 
-	pthread_barrier_wait(&args.barrier);
-	pthread_barrier_destroy(&args.barrier);
-
-	sev2.sigev_notify = SIGEV_THREAD;
-	sev2.sigev_signo = s;
-	sev2.sigev_value.sival_ptr = (void *)&zeros;
+	pthread_setcancelstate(PTHREAD_CANCEL_DISABLE, &cs);
+	sem_wait(&args.sem);
+	sem_destroy(&args.sem);
 
-	if (syscall(SYS_mq_notify, mqd, &sev2) < 0) {
-		pthread_cancel(td);
+	if (args.err) {
 		__syscall(SYS_close, s);
+		pthread_join(td, 0);
+		pthread_setcancelstate(cs, 0);
+		errno = args.err;
 		return -1;
 	}
 
+	pthread_setcancelstate(cs, 0);
 	return 0;
 }
diff --git a/src/multibyte/mbrtowc.c b/src/multibyte/mbrtowc.c
index c94819e7..7824997e 100644
--- a/src/multibyte/mbrtowc.c
+++ b/src/multibyte/mbrtowc.c
@@ -8,7 +8,7 @@ size_t mbrtowc(wchar_t *restrict wc, const char *restrict src, size_t n, mbstate
 	static unsigned internal_state;
 	unsigned c;
 	const unsigned char *s = (const void *)src;
-	const unsigned N = n;
+	const size_t N = n;
 	wchar_t dummy;
 
 	if (!st) st = (void *)&internal_state;
diff --git a/src/network/accept4.c b/src/network/accept4.c
index 59ab1726..765a38ed 100644
--- a/src/network/accept4.c
+++ b/src/network/accept4.c
@@ -9,6 +9,10 @@ int accept4(int fd, struct sockaddr *restrict addr, socklen_t *restrict len, int
 	if (!flg) return accept(fd, addr, len);
 	int ret = socketcall_cp(accept4, fd, addr, len, flg, 0, 0);
 	if (ret>=0 || (errno != ENOSYS && errno != EINVAL)) return ret;
+	if (flg & ~(SOCK_CLOEXEC|SOCK_NONBLOCK)) {
+		errno = EINVAL;
+		return -1;
+	}
 	ret = accept(fd, addr, len);
 	if (ret<0) return ret;
 	if (flg & SOCK_CLOEXEC)
diff --git a/src/network/dns_parse.c b/src/network/dns_parse.c
index e6ee19d9..09813112 100644
--- a/src/network/dns_parse.c
+++ b/src/network/dns_parse.c
@@ -1,7 +1,7 @@
 #include <string.h>
 #include "lookup.h"
 
-int __dns_parse(const unsigned char *r, int rlen, int (*callback)(void *, int, const void *, int, const void *), void *ctx)
+int __dns_parse(const unsigned char *r, int rlen, int (*callback)(void *, int, const void *, int, const void *, int), void *ctx)
 {
 	int qdcount, ancount;
 	const unsigned char *p;
@@ -12,21 +12,20 @@ int __dns_parse(const unsigned char *r, int rlen, int (*callback)(void *, int, c
 	p = r+12;
 	qdcount = r[4]*256 + r[5];
 	ancount = r[6]*256 + r[7];
-	if (qdcount+ancount > 64) return -1;
 	while (qdcount--) {
 		while (p-r < rlen && *p-1U < 127) p++;
-		if (*p>193 || (*p==193 && p[1]>254) || p>r+rlen-6)
+		if (p>r+rlen-6)
 			return -1;
 		p += 5 + !!*p;
 	}
 	while (ancount--) {
 		while (p-r < rlen && *p-1U < 127) p++;
-		if (*p>193 || (*p==193 && p[1]>254) || p>r+rlen-6)
+		if (p>r+rlen-12)
 			return -1;
 		p += 1 + !!*p;
 		len = p[8]*256 + p[9];
-		if (p+len > r+rlen) return -1;
-		if (callback(ctx, p[1], p+10, len, r) < 0) return -1;
+		if (len+10 > r+rlen-p) return -1;
+		if (callback(ctx, p[1], p+10, len, r, rlen) < 0) return -1;
 		p += 10 + len;
 	}
 	return 0;
diff --git a/src/network/gai_strerror.c b/src/network/gai_strerror.c
index 9596580e..56b71503 100644
--- a/src/network/gai_strerror.c
+++ b/src/network/gai_strerror.c
@@ -6,7 +6,7 @@ static const char msgs[] =
 	"Name does not resolve\0"
 	"Try again\0"
 	"Non-recoverable error\0"
-	"Unknown error\0"
+	"Name has no usable address\0"
 	"Unrecognized address family or invalid length\0"
 	"Unrecognized socket type\0"
 	"Unrecognized service\0"
diff --git a/src/network/getaddrinfo.c b/src/network/getaddrinfo.c
index 9df045f6..64ad259a 100644
--- a/src/network/getaddrinfo.c
+++ b/src/network/getaddrinfo.c
@@ -16,6 +16,7 @@ int getaddrinfo(const char *restrict host, const char *restrict serv, const stru
 	char canon[256], *outcanon;
 	int nservs, naddrs, nais, canon_len, i, j, k;
 	int family = AF_UNSPEC, flags = 0, proto = 0, socktype = 0;
+	int no_family = 0;
 	struct aibuf *out;
 
 	if (!host && !serv) return EAI_NONAME;
@@ -82,7 +83,7 @@ int getaddrinfo(const char *restrict host, const char *restrict serv, const stru
 			default:
 				return EAI_SYSTEM;
 			}
-			if (family == tf[i]) return EAI_NONAME;
+			if (family == tf[i]) no_family = 1;
 			family = tf[1-i];
 		}
 	}
@@ -93,6 +94,8 @@ int getaddrinfo(const char *restrict host, const char *restrict serv, const stru
 	naddrs = __lookup_name(addrs, canon, host, family, flags);
 	if (naddrs < 0) return naddrs;
 
+	if (no_family) return EAI_NODATA;
+
 	nais = nservs * naddrs;
 	canon_len = strlen(canon);
 	out = calloc(1, nais * sizeof(*out) + canon_len + 1);
diff --git a/src/network/gethostbyaddr.c b/src/network/gethostbyaddr.c
index 598e2241..c3cacaac 100644
--- a/src/network/gethostbyaddr.c
+++ b/src/network/gethostbyaddr.c
@@ -20,5 +20,5 @@ struct hostent *gethostbyaddr(const void *a, socklen_t l, int af)
 		err = gethostbyaddr_r(a, l, af, h,
 			(void *)(h+1), size-sizeof *h, &res, &h_errno);
 	} while (err == ERANGE);
-	return err ? 0 : h;
+	return res;
 }
diff --git a/src/network/gethostbyaddr_r.c b/src/network/gethostbyaddr_r.c
index 0f1e61aa..ceaf3935 100644
--- a/src/network/gethostbyaddr_r.c
+++ b/src/network/gethostbyaddr_r.c
@@ -54,10 +54,11 @@ int gethostbyaddr_r(const void *a, socklen_t l, int af,
 	case EAI_OVERFLOW:
 		return ERANGE;
 	default:
-	case EAI_MEMORY:
-	case EAI_SYSTEM:
 	case EAI_FAIL:
 		*err = NO_RECOVERY;
+		return EBADMSG;
+	case EAI_SYSTEM:
+		*err = NO_RECOVERY;
 		return errno;
 	case 0:
 		break;
diff --git a/src/network/gethostbyname2.c b/src/network/gethostbyname2.c
index dc9d6621..bd0da7f8 100644
--- a/src/network/gethostbyname2.c
+++ b/src/network/gethostbyname2.c
@@ -21,5 +21,5 @@ struct hostent *gethostbyname2(const char *name, int af)
 		err = gethostbyname2_r(name, af, h,
 			(void *)(h+1), size-sizeof *h, &res, &h_errno);
 	} while (err == ERANGE);
-	return err ? 0 : h;
+	return res;
 }
diff --git a/src/network/gethostbyname2_r.c b/src/network/gethostbyname2_r.c
index fc894877..a5eb67fe 100644
--- a/src/network/gethostbyname2_r.c
+++ b/src/network/gethostbyname2_r.c
@@ -22,7 +22,10 @@ int gethostbyname2_r(const char *name, int af,
 	if (cnt<0) switch (cnt) {
 	case EAI_NONAME:
 		*err = HOST_NOT_FOUND;
-		return ENOENT;
+		return 0;
+	case EAI_NODATA:
+		*err = NO_DATA;
+		return 0;
 	case EAI_AGAIN:
 		*err = TRY_AGAIN;
 		return EAGAIN;
@@ -30,7 +33,6 @@ int gethostbyname2_r(const char *name, int af,
 	case EAI_FAIL:
 		*err = NO_RECOVERY;
 		return EBADMSG;
-	case EAI_MEMORY:
 	case EAI_SYSTEM:
 		*err = NO_RECOVERY;
 		return errno;
diff --git a/src/network/getifaddrs.c b/src/network/getifaddrs.c
index fed75bd8..74df4d6c 100644
--- a/src/network/getifaddrs.c
+++ b/src/network/getifaddrs.c
@@ -39,8 +39,8 @@ struct ifaddrs_storage {
 };
 
 struct ifaddrs_ctx {
-	struct ifaddrs_storage *first;
-	struct ifaddrs_storage *last;
+	struct ifaddrs *first;
+	struct ifaddrs *last;
 	struct ifaddrs_storage *hash[IFADDRS_HASH_SIZE];
 };
 
@@ -195,9 +195,9 @@ static int netlink_msg_to_ifaddr(void *pctx, struct nlmsghdr *h)
 	}
 
 	if (ifs->ifa.ifa_name) {
-		if (!ctx->first) ctx->first = ifs;
-		if (ctx->last) ctx->last->ifa.ifa_next = &ifs->ifa;
-		ctx->last = ifs;
+		if (!ctx->first) ctx->first = &ifs->ifa;
+		if (ctx->last) ctx->last->ifa_next = &ifs->ifa;
+		ctx->last = &ifs->ifa;
 	} else {
 		free(ifs);
 	}
@@ -210,7 +210,7 @@ int getifaddrs(struct ifaddrs **ifap)
 	int r;
 	memset(ctx, 0, sizeof *ctx);
 	r = __rtnetlink_enumerate(AF_UNSPEC, AF_UNSPEC, netlink_msg_to_ifaddr, ctx);
-	if (r == 0) *ifap = &ctx->first->ifa;
-	else freeifaddrs(&ctx->first->ifa);
+	if (r == 0) *ifap = ctx->first;
+	else freeifaddrs(ctx->first);
 	return r;
 }
diff --git a/src/network/getnameinfo.c b/src/network/getnameinfo.c
index 949e1811..133c15b3 100644
--- a/src/network/getnameinfo.c
+++ b/src/network/getnameinfo.c
@@ -58,6 +58,7 @@ static void reverse_hosts(char *buf, const unsigned char *a, unsigned scopeid, i
 		if ((p=strchr(line, '#'))) *p++='\n', *p=0;
 
 		for (p=line; *p && !isspace(*p); p++);
+		if (!*p) continue;
 		*p++ = 0;
 		if (__lookup_ipliteral(&iplit, line, AF_UNSPEC)<=0)
 			continue;
@@ -108,10 +109,10 @@ static void reverse_services(char *buf, int port, int dgram)
 	__fclose_ca(f);
 }
 
-static int dns_parse_callback(void *c, int rr, const void *data, int len, const void *packet)
+static int dns_parse_callback(void *c, int rr, const void *data, int len, const void *packet, int plen)
 {
 	if (rr != RR_PTR) return 0;
-	if (__dn_expand(packet, (const unsigned char *)packet + 512,
+	if (__dn_expand(packet, (const unsigned char *)packet + plen,
 	    data, c, 256) <= 0)
 		*(char *)c = 0;
 	return 0;
@@ -161,8 +162,10 @@ int getnameinfo(const struct sockaddr *restrict sa, socklen_t sl,
 			query[3] = 0; /* don't need AD flag */
 			int rlen = __res_send(query, qlen, reply, sizeof reply);
 			buf[0] = 0;
-			if (rlen > 0)
+			if (rlen > 0) {
+				if (rlen > sizeof reply) rlen = sizeof reply;
 				__dns_parse(reply, rlen, dns_parse_callback, buf);
+			}
 		}
 		if (!*buf) {
 			if (flags & NI_NAMEREQD) return EAI_NONAME;
diff --git a/src/network/getservbyport_r.c b/src/network/getservbyport_r.c
index b7f21c6b..e4cc3079 100644
--- a/src/network/getservbyport_r.c
+++ b/src/network/getservbyport_r.c
@@ -26,7 +26,7 @@ int getservbyport_r(int port, const char *prots,
 	/* Align buffer */
 	i = (uintptr_t)buf & sizeof(char *)-1;
 	if (!i) i = sizeof(char *);
-	if (buflen < 3*sizeof(char *)-i)
+	if (buflen <= 3*sizeof(char *)-i)
 		return ERANGE;
 	buf += sizeof(char *)-i;
 	buflen -= sizeof(char *)-i;
@@ -46,6 +46,8 @@ int getservbyport_r(int port, const char *prots,
 	case EAI_MEMORY:
 	case EAI_SYSTEM:
 		return ENOMEM;
+	case EAI_OVERFLOW:
+		return ERANGE;
 	default:
 		return ENOENT;
 	case 0:
diff --git a/src/network/inet_pton.c b/src/network/inet_pton.c
index d36c3689..bcbdd9ef 100644
--- a/src/network/inet_pton.c
+++ b/src/network/inet_pton.c
@@ -54,6 +54,7 @@ int inet_pton(int af, const char *restrict s, void *restrict a0)
 			if (s[j]!='.' || (i<6 && brk<0)) return 0;
 			need_v4=1;
 			i++;
+			ip[i&7]=0;
 			break;
 		}
 		s += j+1;
diff --git a/src/network/lookup.h b/src/network/lookup.h
index ef662725..54b2f8b5 100644
--- a/src/network/lookup.h
+++ b/src/network/lookup.h
@@ -50,6 +50,6 @@ hidden int __lookup_ipliteral(struct address buf[static 1], const char *name, in
 hidden int __get_resolv_conf(struct resolvconf *, char *, size_t);
 hidden int __res_msend_rc(int, const unsigned char *const *, const int *, unsigned char *const *, int *, int, const struct resolvconf *);
 
-hidden int __dns_parse(const unsigned char *, int, int (*)(void *, int, const void *, int, const void *), void *);
+hidden int __dns_parse(const unsigned char *, int, int (*)(void *, int, const void *, int, const void *, int), void *);
 
 #endif
diff --git a/src/network/lookup_ipliteral.c b/src/network/lookup_ipliteral.c
index 2fddab73..1e766206 100644
--- a/src/network/lookup_ipliteral.c
+++ b/src/network/lookup_ipliteral.c
@@ -15,7 +15,7 @@ int __lookup_ipliteral(struct address buf[static 1], const char *name, int famil
 	struct in6_addr a6;
 	if (__inet_aton(name, &a4) > 0) {
 		if (family == AF_INET6) /* wrong family */
-			return EAI_NONAME;
+			return EAI_NODATA;
 		memcpy(&buf[0].addr, &a4, sizeof a4);
 		buf[0].family = AF_INET;
 		buf[0].scopeid = 0;
@@ -34,7 +34,7 @@ int __lookup_ipliteral(struct address buf[static 1], const char *name, int famil
 	if (inet_pton(AF_INET6, name, &a6) <= 0)
 		return 0;
 	if (family == AF_INET) /* wrong family */
-		return EAI_NONAME;
+		return EAI_NODATA;
 
 	memcpy(&buf[0].addr, &a6, sizeof a6);
 	buf[0].family = AF_INET6;
diff --git a/src/network/lookup_name.c b/src/network/lookup_name.c
index b5232ce8..35218185 100644
--- a/src/network/lookup_name.c
+++ b/src/network/lookup_name.c
@@ -79,7 +79,7 @@ static int name_from_hosts(struct address buf[static MAXADDRS], char canon[stati
 		case 0:
 			continue;
 		default:
-			badfam = EAI_NONAME;
+			badfam = EAI_NODATA;
 			break;
 		}
 
@@ -102,45 +102,50 @@ struct dpc_ctx {
 	struct address *addrs;
 	char *canon;
 	int cnt;
+	int rrtype;
 };
 
 #define RR_A 1
 #define RR_CNAME 5
 #define RR_AAAA 28
 
-static int dns_parse_callback(void *c, int rr, const void *data, int len, const void *packet)
+#define ABUF_SIZE 4800
+
+static int dns_parse_callback(void *c, int rr, const void *data, int len, const void *packet, int plen)
 {
 	char tmp[256];
+	int family;
 	struct dpc_ctx *ctx = c;
-	if (ctx->cnt >= MAXADDRS) return -1;
+	if (rr == RR_CNAME) {
+		if (__dn_expand(packet, (const unsigned char *)packet + plen,
+		    data, tmp, sizeof tmp) > 0 && is_valid_hostname(tmp))
+			strcpy(ctx->canon, tmp);
+		return 0;
+	}
+	if (ctx->cnt >= MAXADDRS) return 0;
+	if (rr != ctx->rrtype) return 0;
 	switch (rr) {
 	case RR_A:
 		if (len != 4) return -1;
-		ctx->addrs[ctx->cnt].family = AF_INET;
-		ctx->addrs[ctx->cnt].scopeid = 0;
-		memcpy(ctx->addrs[ctx->cnt++].addr, data, 4);
+		family = AF_INET;
 		break;
 	case RR_AAAA:
 		if (len != 16) return -1;
-		ctx->addrs[ctx->cnt].family = AF_INET6;
-		ctx->addrs[ctx->cnt].scopeid = 0;
-		memcpy(ctx->addrs[ctx->cnt++].addr, data, 16);
-		break;
-	case RR_CNAME:
-		if (__dn_expand(packet, (const unsigned char *)packet + 512,
-		    data, tmp, sizeof tmp) > 0 && is_valid_hostname(tmp))
-			strcpy(ctx->canon, tmp);
+		family = AF_INET6;
 		break;
 	}
+	ctx->addrs[ctx->cnt].family = family;
+	ctx->addrs[ctx->cnt].scopeid = 0;
+	memcpy(ctx->addrs[ctx->cnt++].addr, data, len);
 	return 0;
 }
 
 static int name_from_dns(struct address buf[static MAXADDRS], char canon[static 256], const char *name, int family, const struct resolvconf *conf)
 {
-	unsigned char qbuf[2][280], abuf[2][512];
+	unsigned char qbuf[2][280], abuf[2][ABUF_SIZE];
 	const unsigned char *qp[2] = { qbuf[0], qbuf[1] };
 	unsigned char *ap[2] = { abuf[0], abuf[1] };
-	int qlens[2], alens[2];
+	int qlens[2], alens[2], qtypes[2];
 	int i, nq = 0;
 	struct dpc_ctx ctx = { .addrs = buf, .canon = canon };
 	static const struct { int af; int rr; } afrr[2] = {
@@ -153,7 +158,8 @@ static int name_from_dns(struct address buf[static MAXADDRS], char canon[static
 			qlens[nq] = __res_mkquery(0, name, 1, afrr[i].rr,
 				0, 0, 0, qbuf[nq], sizeof *qbuf);
 			if (qlens[nq] == -1)
-				return EAI_NONAME;
+				return 0;
+			qtypes[nq] = afrr[i].rr;
 			qbuf[nq][3] = 0; /* don't need AD flag */
 			/* Ensure query IDs are distinct. */
 			if (nq && qbuf[nq][0] == qbuf[0][0])
@@ -171,11 +177,14 @@ static int name_from_dns(struct address buf[static MAXADDRS], char canon[static
 		if ((abuf[i][3] & 15) != 0) return EAI_FAIL;
 	}
 
-	for (i=0; i<nq; i++)
+	for (i=nq-1; i>=0; i--) {
+		ctx.rrtype = qtypes[i];
+		if (alens[i] > sizeof(abuf[i])) alens[i] = sizeof abuf[i];
 		__dns_parse(abuf[i], alens[i], dns_parse_callback, &ctx);
+	}
 
 	if (ctx.cnt) return ctx.cnt;
-	return EAI_NONAME;
+	return EAI_NODATA;
 }
 
 static int name_from_dns_search(struct address buf[static MAXADDRS], char canon[static 256], const char *name, int family)
diff --git a/src/network/res_mkquery.c b/src/network/res_mkquery.c
index 33f50cb9..614bf786 100644
--- a/src/network/res_mkquery.c
+++ b/src/network/res_mkquery.c
@@ -13,6 +13,7 @@ int __res_mkquery(int op, const char *dname, int class, int type,
 	int n;
 
 	if (l && dname[l-1]=='.') l--;
+	if (l && dname[l-1]=='.') return -1;
 	n = 17+l+!!l;
 	if (l>253 || buflen<n || op>15u || class>255u || type>255u)
 		return -1;
diff --git a/src/network/res_msend.c b/src/network/res_msend.c
index 3e018009..86c2fcf4 100644
--- a/src/network/res_msend.c
+++ b/src/network/res_msend.c
@@ -1,5 +1,6 @@
 #include <sys/socket.h>
 #include <netinet/in.h>
+#include <netinet/tcp.h>
 #include <netdb.h>
 #include <arpa/inet.h>
 #include <stdint.h>
@@ -16,17 +17,65 @@
 
 static void cleanup(void *p)
 {
-	__syscall(SYS_close, (intptr_t)p);
+	struct pollfd *pfd = p;
+	for (int i=0; pfd[i].fd >= -1; i++)
+		if (pfd[i].fd >= 0) __syscall(SYS_close, pfd[i].fd);
 }
 
 static unsigned long mtime()
 {
 	struct timespec ts;
-	clock_gettime(CLOCK_REALTIME, &ts);
+	if (clock_gettime(CLOCK_MONOTONIC, &ts) < 0 && errno == ENOSYS)
+		clock_gettime(CLOCK_REALTIME, &ts);
 	return (unsigned long)ts.tv_sec * 1000
 		+ ts.tv_nsec / 1000000;
 }
 
+static int start_tcp(struct pollfd *pfd, int family, const void *sa, socklen_t sl, const unsigned char *q, int ql)
+{
+	struct msghdr mh = {
+		.msg_name = (void *)sa,
+		.msg_namelen = sl,
+		.msg_iovlen = 2,
+		.msg_iov = (struct iovec [2]){
+			{ .iov_base = (uint8_t[]){ ql>>8, ql }, .iov_len = 2 },
+			{ .iov_base = (void *)q, .iov_len = ql } }
+	};
+	int r;
+	int fd = socket(family, SOCK_STREAM|SOCK_CLOEXEC|SOCK_NONBLOCK, 0);
+	pfd->fd = fd;
+	pfd->events = POLLOUT;
+	if (!setsockopt(fd, IPPROTO_TCP, TCP_FASTOPEN_CONNECT,
+	    &(int){1}, sizeof(int))) {
+		r = sendmsg(fd, &mh, MSG_FASTOPEN|MSG_NOSIGNAL);
+		if (r == ql+2) pfd->events = POLLIN;
+		if (r >= 0) return r;
+		if (errno == EINPROGRESS) return 0;
+	}
+	r = connect(fd, sa, sl);
+	if (!r || errno == EINPROGRESS) return 0;
+	close(fd);
+	pfd->fd = -1;
+	return -1;
+}
+
+static void step_mh(struct msghdr *mh, size_t n)
+{
+	/* Adjust iovec in msghdr to skip first n bytes. */
+	while (mh->msg_iovlen && n >= mh->msg_iov->iov_len) {
+		n -= mh->msg_iov->iov_len;
+		mh->msg_iov++;
+		mh->msg_iovlen--;
+	}
+	if (!mh->msg_iovlen) return;
+	mh->msg_iov->iov_base = (char *)mh->msg_iov->iov_base + n;
+	mh->msg_iov->iov_len -= n;
+}
+
+/* Internal contract for __res_msend[_rc]: asize must be >=512, nqueries
+ * must be sufficiently small to be safe as VLA size. In practice it's
+ * either 1 or 2, anyway. */
+
 int __res_msend_rc(int nqueries, const unsigned char *const *queries,
 	const int *qlens, unsigned char *const *answers, int *alens, int asize,
 	const struct resolvconf *conf)
@@ -44,7 +93,10 @@ int __res_msend_rc(int nqueries, const unsigned char *const *queries,
 	int next;
 	int i, j;
 	int cs;
-	struct pollfd pfd;
+	struct pollfd pfd[nqueries+2];
+	int qpos[nqueries], apos[nqueries];
+	unsigned char alen_buf[nqueries][2];
+	int r;
 	unsigned long t0, t1, t2;
 
 	pthread_setcancelstate(PTHREAD_CANCEL_DISABLE, &cs);
@@ -68,29 +120,22 @@ int __res_msend_rc(int nqueries, const unsigned char *const *queries,
 	}
 
 	/* Get local address and open/bind a socket */
-	sa.sin.sin_family = family;
 	fd = socket(family, SOCK_DGRAM|SOCK_CLOEXEC|SOCK_NONBLOCK, 0);
 
 	/* Handle case where system lacks IPv6 support */
 	if (fd < 0 && family == AF_INET6 && errno == EAFNOSUPPORT) {
+		for (i=0; i<nns && conf->ns[nns].family == AF_INET6; i++);
+		if (i==nns) {
+			pthread_setcancelstate(cs, 0);
+			return -1;
+		}
 		fd = socket(AF_INET, SOCK_DGRAM|SOCK_CLOEXEC|SOCK_NONBLOCK, 0);
 		family = AF_INET;
+		sl = sizeof sa.sin;
 	}
-	if (fd < 0 || bind(fd, (void *)&sa, sl) < 0) {
-		if (fd >= 0) close(fd);
-		pthread_setcancelstate(cs, 0);
-		return -1;
-	}
-
-	/* Past this point, there are no errors. Each individual query will
-	 * yield either no reply (indicated by zero length) or an answer
-	 * packet which is up to the caller to interpret. */
-
-	pthread_cleanup_push(cleanup, (void *)(intptr_t)fd);
-	pthread_setcancelstate(cs, 0);
 
 	/* Convert any IPv4 addresses in a mixed environment to v4-mapped */
-	if (family == AF_INET6) {
+	if (fd >= 0 && family == AF_INET6) {
 		setsockopt(fd, IPPROTO_IPV6, IPV6_V6ONLY, &(int){0}, sizeof 0);
 		for (i=0; i<nns; i++) {
 			if (ns[i].sin.sin_family != AF_INET) continue;
@@ -104,16 +149,38 @@ int __res_msend_rc(int nqueries, const unsigned char *const *queries,
 		}
 	}
 
+	sa.sin.sin_family = family;
+	if (fd < 0 || bind(fd, (void *)&sa, sl) < 0) {
+		if (fd >= 0) close(fd);
+		pthread_setcancelstate(cs, 0);
+		return -1;
+	}
+
+	/* Past this point, there are no errors. Each individual query will
+	 * yield either no reply (indicated by zero length) or an answer
+	 * packet which is up to the caller to interpret. */
+
+	for (i=0; i<nqueries; i++) pfd[i].fd = -1;
+	pfd[nqueries].fd = fd;
+	pfd[nqueries].events = POLLIN;
+	pfd[nqueries+1].fd = -2;
+
+	pthread_cleanup_push(cleanup, pfd);
+	pthread_setcancelstate(cs, 0);
+
 	memset(alens, 0, sizeof *alens * nqueries);
 
-	pfd.fd = fd;
-	pfd.events = POLLIN;
 	retry_interval = timeout / attempts;
 	next = 0;
 	t0 = t2 = mtime();
 	t1 = t2 - retry_interval;
 
 	for (; t2-t0 < timeout; t2=mtime()) {
+		/* This is the loop exit condition: that all queries
+		 * have an accepted answer. */
+		for (i=0; i<nqueries && alens[i]>0; i++);
+		if (i==nqueries) break;
+
 		if (t2-t1 >= retry_interval) {
 			/* Query all configured namservers in parallel */
 			for (i=0; i<nqueries; i++)
@@ -127,10 +194,20 @@ int __res_msend_rc(int nqueries, const unsigned char *const *queries,
 		}
 
 		/* Wait for a response, or until time to retry */
-		if (poll(&pfd, 1, t1+retry_interval-t2) <= 0) continue;
+		if (poll(pfd, nqueries+1, t1+retry_interval-t2) <= 0) continue;
 
-		while ((rlen = recvfrom(fd, answers[next], asize, 0,
-		  (void *)&sa, (socklen_t[1]){sl})) >= 0) {
+		while (next < nqueries) {
+			struct msghdr mh = {
+				.msg_name = (void *)&sa,
+				.msg_namelen = sl,
+				.msg_iovlen = 1,
+				.msg_iov = (struct iovec []){
+					{ .iov_base = (void *)answers[next],
+					  .iov_len = asize }
+				}
+			};
+			rlen = recvmsg(fd, &mh, 0);
+			if (rlen < 0) break;
 
 			/* Ignore non-identifiable packets */
 			if (rlen < 4) continue;
@@ -170,12 +247,72 @@ int __res_msend_rc(int nqueries, const unsigned char *const *queries,
 			else
 				memcpy(answers[i], answers[next], rlen);
 
-			if (next == nqueries) goto out;
+			/* Ignore further UDP if all slots full or TCP-mode */
+			if (next == nqueries) pfd[nqueries].events = 0;
+
+			/* If answer is truncated (TC bit), fallback to TCP */
+			if ((answers[i][2] & 2) || (mh.msg_flags & MSG_TRUNC)) {
+				alens[i] = -1;
+				pthread_setcancelstate(PTHREAD_CANCEL_DISABLE, 0);
+				r = start_tcp(pfd+i, family, ns+j, sl, queries[i], qlens[i]);
+				pthread_setcancelstate(cs, 0);
+				if (r >= 0) {
+					qpos[i] = r;
+					apos[i] = 0;
+				}
+				continue;
+			}
+		}
+
+		for (i=0; i<nqueries; i++) if (pfd[i].revents & POLLOUT) {
+			struct msghdr mh = {
+				.msg_iovlen = 2,
+				.msg_iov = (struct iovec [2]){
+					{ .iov_base = (uint8_t[]){ qlens[i]>>8, qlens[i] }, .iov_len = 2 },
+					{ .iov_base = (void *)queries[i], .iov_len = qlens[i] } }
+			};
+			step_mh(&mh, qpos[i]);
+			r = sendmsg(pfd[i].fd, &mh, MSG_NOSIGNAL);
+			if (r < 0) goto out;
+			qpos[i] += r;
+			if (qpos[i] == qlens[i]+2)
+				pfd[i].events = POLLIN;
+		}
+
+		for (i=0; i<nqueries; i++) if (pfd[i].revents & POLLIN) {
+			struct msghdr mh = {
+				.msg_iovlen = 2,
+				.msg_iov = (struct iovec [2]){
+					{ .iov_base = alen_buf[i], .iov_len = 2 },
+					{ .iov_base = answers[i], .iov_len = asize } }
+			};
+			step_mh(&mh, apos[i]);
+			r = recvmsg(pfd[i].fd, &mh, 0);
+			if (r <= 0) goto out;
+			apos[i] += r;
+			if (apos[i] < 2) continue;
+			int alen = alen_buf[i][0]*256 + alen_buf[i][1];
+			if (alen < 13) goto out;
+			if (apos[i] < alen+2 && apos[i] < asize+2)
+				continue;
+			int rcode = answers[i][3] & 15;
+			if (rcode != 0 && rcode != 3)
+				goto out;
+
+			/* Storing the length here commits the accepted answer.
+			 * Immediately close TCP socket so as not to consume
+			 * resources we no longer need. */
+			alens[i] = alen;
+			__syscall(SYS_close, pfd[i].fd);
+			pfd[i].fd = -1;
 		}
 	}
 out:
 	pthread_cleanup_pop(1);
 
+	/* Disregard any incomplete TCP results */
+	for (i=0; i<nqueries; i++) if (alens[i]<0) alens[i] = 0;
+
 	return 0;
 }
 
diff --git a/src/network/res_send.c b/src/network/res_send.c
index ee4abf1f..9593164d 100644
--- a/src/network/res_send.c
+++ b/src/network/res_send.c
@@ -1,8 +1,16 @@
 #include <resolv.h>
+#include <string.h>
 
 int __res_send(const unsigned char *msg, int msglen, unsigned char *answer, int anslen)
 {
-	int r = __res_msend(1, &msg, &msglen, &answer, &anslen, anslen);
+	int r;
+	if (anslen < 512) {
+		unsigned char buf[512];
+		r = __res_send(msg, msglen, buf, sizeof buf);
+		if (r >= 0) memcpy(answer, buf, r < anslen ? r : anslen);
+		return r;
+	}
+	r = __res_msend(1, &msg, &msglen, &answer, &anslen, anslen);
 	return r<0 || !anslen ? -1 : anslen;
 }
 
diff --git a/src/network/sendmsg.c b/src/network/sendmsg.c
index 80cc5f41..acdfdf29 100644
--- a/src/network/sendmsg.c
+++ b/src/network/sendmsg.c
@@ -8,13 +8,16 @@ ssize_t sendmsg(int fd, const struct msghdr *msg, int flags)
 {
 #if LONG_MAX > INT_MAX
 	struct msghdr h;
-	struct cmsghdr chbuf[1024/sizeof(struct cmsghdr)+1], *c;
+	/* Kernels before 2.6.38 set SCM_MAX_FD to 255, allocate enough
+	 * space to support an SCM_RIGHTS ancillary message with 255 fds.
+	 * Kernels since 2.6.38 set SCM_MAX_FD to 253. */
+	struct cmsghdr chbuf[CMSG_SPACE(255*sizeof(int))/sizeof(struct cmsghdr)+1], *c;
 	if (msg) {
 		h = *msg;
 		h.__pad1 = h.__pad2 = 0;
 		msg = &h;
 		if (h.msg_controllen) {
-			if (h.msg_controllen > 1024) {
+			if (h.msg_controllen > sizeof chbuf) {
 				errno = ENOMEM;
 				return -1;
 			}
diff --git a/src/process/_Fork.c b/src/process/_Fork.c
index da063868..9c07792d 100644
--- a/src/process/_Fork.c
+++ b/src/process/_Fork.c
@@ -5,25 +5,16 @@
 #include "lock.h"
 #include "pthread_impl.h"
 #include "aio_impl.h"
+#include "fork_impl.h"
 
 static void dummy(int x) { }
 weak_alias(dummy, __aio_atfork);
 
-pid_t _Fork(void)
+void __post_Fork(int ret)
 {
-	pid_t ret;
-	sigset_t set;
-	__block_all_sigs(&set);
-	__aio_atfork(-1);
-	LOCK(__abort_lock);
-#ifdef SYS_fork
-	ret = __syscall(SYS_fork);
-#else
-	ret = __syscall(SYS_clone, SIGCHLD, 0);
-#endif
 	if (!ret) {
 		pthread_t self = __pthread_self();
-		self->tid = __syscall(SYS_gettid);
+		self->tid = __syscall(SYS_set_tid_address, &__thread_list_lock);
 		self->robust_list.off = 0;
 		self->robust_list.pending = 0;
 		self->next = self->prev = self;
@@ -32,7 +23,21 @@ pid_t _Fork(void)
 		if (libc.need_locks) libc.need_locks = -1;
 	}
 	UNLOCK(__abort_lock);
-	__aio_atfork(!ret);
+	if (!ret) __aio_atfork(1);
+}
+
+pid_t _Fork(void)
+{
+	pid_t ret;
+	sigset_t set;
+	__block_all_sigs(&set);
+	LOCK(__abort_lock);
+#ifdef SYS_fork
+	ret = __syscall(SYS_fork);
+#else
+	ret = __syscall(SYS_clone, SIGCHLD, 0);
+#endif
+	__post_Fork(ret);
 	__restore_sigs(&set);
 	return __syscall_ret(ret);
 }
diff --git a/src/process/fork.c b/src/process/fork.c
index 54bc2892..56f19313 100644
--- a/src/process/fork.c
+++ b/src/process/fork.c
@@ -9,7 +9,6 @@ static volatile int *const dummy_lockptr = 0;
 
 weak_alias(dummy_lockptr, __at_quick_exit_lockptr);
 weak_alias(dummy_lockptr, __atexit_lockptr);
-weak_alias(dummy_lockptr, __dlerror_lockptr);
 weak_alias(dummy_lockptr, __gettext_lockptr);
 weak_alias(dummy_lockptr, __locale_lockptr);
 weak_alias(dummy_lockptr, __random_lockptr);
@@ -24,7 +23,6 @@ weak_alias(dummy_lockptr, __vmlock_lockptr);
 static volatile int *const *const atfork_locks[] = {
 	&__at_quick_exit_lockptr,
 	&__atexit_lockptr,
-	&__dlerror_lockptr,
 	&__gettext_lockptr,
 	&__locale_lockptr,
 	&__random_lockptr,
@@ -38,6 +36,8 @@ static volatile int *const *const atfork_locks[] = {
 static void dummy(int x) { }
 weak_alias(dummy, __fork_handler);
 weak_alias(dummy, __malloc_atfork);
+weak_alias(dummy, __aio_atfork);
+weak_alias(dummy, __pthread_key_atfork);
 weak_alias(dummy, __ldso_atfork);
 
 static void dummy_0(void) { }
@@ -52,6 +52,8 @@ pid_t fork(void)
 	int need_locks = libc.need_locks > 0;
 	if (need_locks) {
 		__ldso_atfork(-1);
+		__pthread_key_atfork(-1);
+		__aio_atfork(-1);
 		__inhibit_ptc();
 		for (int i=0; i<sizeof atfork_locks/sizeof *atfork_locks; i++)
 			if (*atfork_locks[i]) LOCK(*atfork_locks[i]);
@@ -77,6 +79,8 @@ pid_t fork(void)
 				if (ret) UNLOCK(*atfork_locks[i]);
 				else **atfork_locks[i] = 0;
 		__release_ptc();
+		if (ret) __aio_atfork(0);
+		__pthread_key_atfork(!ret);
 		__ldso_atfork(!ret);
 	}
 	__restore_sigs(&set);
diff --git a/src/process/posix_spawn.c b/src/process/posix_spawn.c
index 728551b3..8294598b 100644
--- a/src/process/posix_spawn.c
+++ b/src/process/posix_spawn.c
@@ -4,6 +4,7 @@
 #include <unistd.h>
 #include <signal.h>
 #include <fcntl.h>
+#include <errno.h>
 #include <sys/wait.h>
 #include "syscall.h"
 #include "lock.h"
@@ -156,7 +157,11 @@ static int child(void *args_vp)
 fail:
 	/* Since sizeof errno < PIPE_BUF, the write is atomic. */
 	ret = -ret;
-	if (ret) while (__syscall(SYS_write, p, &ret, sizeof ret) < 0);
+	if (ret) {
+		int r;
+		do r = __syscall(SYS_write, p, &ret, sizeof ret);
+		while (r<0 && r!=-EPIPE);
+	}
 	_exit(127);
 }
 
diff --git a/src/process/riscv64/vfork.s b/src/process/riscv64/vfork.s
new file mode 100644
index 00000000..c93dca23
--- /dev/null
+++ b/src/process/riscv64/vfork.s
@@ -0,0 +1,12 @@
+.global vfork
+.type vfork,@function
+vfork:
+	/* riscv does not have SYS_vfork, so we must use clone instead */
+	/* note: riscv's clone = clone(flags, sp, ptidptr, tls, ctidptr) */
+	li a7, 220
+	li a0, 0x100 | 0x4000 | 17 /* flags = CLONE_VM | CLONE_VFORK | SIGCHLD */
+	mv a1, sp
+	/* the other arguments are ignoreable */
+	ecall
+	.hidden __syscall_ret
+	j __syscall_ret
diff --git a/src/process/waitpid.c b/src/process/waitpid.c
index 1b65bf05..80231862 100644
--- a/src/process/waitpid.c
+++ b/src/process/waitpid.c
@@ -3,5 +3,5 @@
 
 pid_t waitpid(pid_t pid, int *status, int options)
 {
-	return syscall_cp(SYS_wait4, pid, status, options, 0);
+	return sys_wait4_cp(pid, status, options, 0);
 }
diff --git a/src/regex/glob.c b/src/regex/glob.c
index 9de080ed..87bae084 100644
--- a/src/regex/glob.c
+++ b/src/regex/glob.c
@@ -265,7 +265,7 @@ int glob(const char *restrict pat, int flags, int (*errfunc)(const char *path, i
 			if (append(&tail, pat, strlen(pat), 0))
 				return GLOB_NOSPACE;
 			cnt++;
-		} else
+		} else if (!error)
 			return GLOB_NOMATCH;
 	}
 
@@ -306,6 +306,3 @@ void globfree(glob_t *g)
 	g->gl_pathc = 0;
 	g->gl_pathv = NULL;
 }
-
-weak_alias(glob, glob64);
-weak_alias(globfree, globfree64);
diff --git a/src/search/hsearch.c b/src/search/hsearch.c
index b3ac8796..2634a67f 100644
--- a/src/search/hsearch.c
+++ b/src/search/hsearch.c
@@ -41,9 +41,9 @@ static int resize(size_t nel, struct hsearch_data *htab)
 {
 	size_t newsize;
 	size_t i, j;
+	size_t oldsize = htab->__tab->mask + 1;
 	ENTRY *e, *newe;
 	ENTRY *oldtab = htab->__tab->entries;
-	ENTRY *oldend = htab->__tab->entries + htab->__tab->mask + 1;
 
 	if (nel > MAXSIZE)
 		nel = MAXSIZE;
@@ -56,7 +56,7 @@ static int resize(size_t nel, struct hsearch_data *htab)
 	htab->__tab->mask = newsize - 1;
 	if (!oldtab)
 		return 1;
-	for (e = oldtab; e < oldend; e++)
+	for (e = oldtab; e < oldtab + oldsize; e++)
 		if (e->key) {
 			for (i=keyhash(e->key),j=1; ; i+=j++) {
 				newe = htab->__tab->entries + (i & htab->__tab->mask);
diff --git a/src/select/poll.c b/src/select/poll.c
index c84c8a99..7883dfab 100644
--- a/src/select/poll.c
+++ b/src/select/poll.c
@@ -8,8 +8,13 @@ int poll(struct pollfd *fds, nfds_t n, int timeout)
 #ifdef SYS_poll
 	return syscall_cp(SYS_poll, fds, n, timeout);
 #else
+#if SYS_ppoll_time64 == SYS_ppoll
+	typedef long long ppoll_ts_t[2];
+#else
+	typedef long ppoll_ts_t[2];
+#endif
 	return syscall_cp(SYS_ppoll, fds, n, timeout>=0 ?
-		&((struct timespec){ .tv_sec = timeout/1000,
-		.tv_nsec = timeout%1000*1000000 }) : 0, 0, _NSIG/8);
+		((ppoll_ts_t){ timeout/1000, timeout%1000*1000000 }) : 0,
+		0, _NSIG/8);
 #endif
 }
diff --git a/src/linux/ppoll.c b/src/select/ppoll.c
index e614600a..9a0bf929 100644
--- a/src/linux/ppoll.c
+++ b/src/select/ppoll.c
@@ -1,4 +1,4 @@
-#define _GNU_SOURCE
+#define _BSD_SOURCE
 #include <poll.h>
 #include <signal.h>
 #include <errno.h>
diff --git a/src/select/select.c b/src/select/select.c
index 8a786884..f1d72863 100644
--- a/src/select/select.c
+++ b/src/select/select.c
@@ -33,6 +33,7 @@ int select(int n, fd_set *restrict rfds, fd_set *restrict wfds, fd_set *restrict
 			((syscall_arg_t[]){ 0, _NSIG/8 }));
 	if (SYS_pselect6 == SYS_pselect6_time64 || r!=-ENOSYS)
 		return __syscall_ret(r);
+	s = CLAMP(s);
 #endif
 #ifdef SYS_select
 	return syscall_cp(SYS_select, n, rfds, wfds, efds,
diff --git a/src/setjmp/loongarch64/longjmp.S b/src/setjmp/loongarch64/longjmp.S
new file mode 100644
index 00000000..896d2e26
--- /dev/null
+++ b/src/setjmp/loongarch64/longjmp.S
@@ -0,0 +1,32 @@
+.global _longjmp
+.global longjmp
+.type   _longjmp,@function
+.type   longjmp,@function
+_longjmp:
+longjmp:
+	ld.d    $ra, $a0, 0
+	ld.d    $sp, $a0, 8
+	ld.d    $r21,$a0, 16
+	ld.d    $fp, $a0, 24
+	ld.d    $s0, $a0, 32
+	ld.d    $s1, $a0, 40
+	ld.d    $s2, $a0, 48
+	ld.d    $s3, $a0, 56
+	ld.d    $s4, $a0, 64
+	ld.d    $s5, $a0, 72
+	ld.d    $s6, $a0, 80
+	ld.d    $s7, $a0, 88
+	ld.d    $s8, $a0, 96
+#ifndef __loongarch_soft_float
+	fld.d   $fs0, $a0, 104
+	fld.d   $fs1, $a0, 112
+	fld.d   $fs2, $a0, 120
+	fld.d   $fs3, $a0, 128
+	fld.d   $fs4, $a0, 136
+	fld.d   $fs5, $a0, 144
+	fld.d   $fs6, $a0, 152
+	fld.d   $fs7, $a0, 160
+#endif
+	sltui   $a0, $a1, 1
+	add.d   $a0, $a0, $a1
+	jr      $ra
diff --git a/src/setjmp/loongarch64/setjmp.S b/src/setjmp/loongarch64/setjmp.S
new file mode 100644
index 00000000..d158a3d2
--- /dev/null
+++ b/src/setjmp/loongarch64/setjmp.S
@@ -0,0 +1,34 @@
+.global __setjmp
+.global _setjmp
+.global setjmp
+.type __setjmp,@function
+.type _setjmp,@function
+.type setjmp,@function
+__setjmp:
+_setjmp:
+setjmp:
+	st.d    $ra, $a0, 0
+	st.d    $sp, $a0, 8
+	st.d    $r21,$a0, 16
+	st.d    $fp, $a0, 24
+	st.d    $s0, $a0, 32
+	st.d    $s1, $a0, 40
+	st.d    $s2, $a0, 48
+	st.d    $s3, $a0, 56
+	st.d    $s4, $a0, 64
+	st.d    $s5, $a0, 72
+	st.d    $s6, $a0, 80
+	st.d    $s7, $a0, 88
+	st.d    $s8, $a0, 96
+#ifndef __loongarch_soft_float
+	fst.d   $fs0, $a0, 104
+	fst.d   $fs1, $a0, 112
+	fst.d   $fs2, $a0, 120
+	fst.d   $fs3, $a0, 128
+	fst.d   $fs4, $a0, 136
+	fst.d   $fs5, $a0, 144
+	fst.d   $fs6, $a0, 152
+	fst.d   $fs7, $a0, 160
+#endif
+	move    $a0, $zero
+	jr      $ra
diff --git a/src/setjmp/powerpc/longjmp.S b/src/setjmp/powerpc/longjmp.S
index 611389fe..465e4cd7 100644
--- a/src/setjmp/powerpc/longjmp.S
+++ b/src/setjmp/powerpc/longjmp.S
@@ -42,10 +42,10 @@ longjmp:
 	bl 1f
 	.hidden __hwcap
 	.long __hwcap-.
-1:	mflr 4
-	lwz 5, 0(4)
-	lwzx 4, 4, 5
-	andis. 4, 4, 0x80
+1:	mflr 6
+	lwz 5, 0(6)
+	lwzx 6, 6, 5
+	andis. 6, 6, 0x80
 	beq 1f
 	.long 0x11c35b01 /* evldd 14,88(3) */
 	.long 0x11e36301 /* ... */
diff --git a/src/setjmp/riscv32/longjmp.S b/src/setjmp/riscv32/longjmp.S
new file mode 100644
index 00000000..f9cb3318
--- /dev/null
+++ b/src/setjmp/riscv32/longjmp.S
@@ -0,0 +1,42 @@
+.global __longjmp
+.global _longjmp
+.global longjmp
+.type __longjmp, %function
+.type _longjmp,  %function
+.type longjmp,   %function
+__longjmp:
+_longjmp:
+longjmp:
+	lw s0,    0(a0)
+	lw s1,    4(a0)
+	lw s2,    8(a0)
+	lw s3,    12(a0)
+	lw s4,    16(a0)
+	lw s5,    20(a0)
+	lw s6,    24(a0)
+	lw s7,    28(a0)
+	lw s8,    32(a0)
+	lw s9,    36(a0)
+	lw s10,   40(a0)
+	lw s11,   44(a0)
+	lw sp,    48(a0)
+	lw ra,    52(a0)
+
+#ifndef __riscv_float_abi_soft
+	fld fs0,  56(a0)
+	fld fs1,  64(a0)
+	fld fs2,  72(a0)
+	fld fs3,  80(a0)
+	fld fs4,  88(a0)
+	fld fs5,  96(a0)
+	fld fs6,  104(a0)
+	fld fs7,  112(a0)
+	fld fs8,  120(a0)
+	fld fs9,  128(a0)
+	fld fs10, 136(a0)
+	fld fs11, 144(a0)
+#endif
+
+	seqz a0, a1
+	add a0, a0, a1
+	ret
diff --git a/src/setjmp/riscv32/setjmp.S b/src/setjmp/riscv32/setjmp.S
new file mode 100644
index 00000000..8a75cf55
--- /dev/null
+++ b/src/setjmp/riscv32/setjmp.S
@@ -0,0 +1,41 @@
+.global __setjmp
+.global _setjmp
+.global setjmp
+.type __setjmp, %function
+.type _setjmp,  %function
+.type setjmp,   %function
+__setjmp:
+_setjmp:
+setjmp:
+	sw s0,    0(a0)
+	sw s1,    4(a0)
+	sw s2,    8(a0)
+	sw s3,    12(a0)
+	sw s4,    16(a0)
+	sw s5,    20(a0)
+	sw s6,    24(a0)
+	sw s7,    28(a0)
+	sw s8,    32(a0)
+	sw s9,    36(a0)
+	sw s10,   40(a0)
+	sw s11,   44(a0)
+	sw sp,    48(a0)
+	sw ra,    52(a0)
+
+#ifndef __riscv_float_abi_soft
+	fsd fs0,  56(a0)
+	fsd fs1,  64(a0)
+	fsd fs2,  72(a0)
+	fsd fs3,  80(a0)
+	fsd fs4,  88(a0)
+	fsd fs5,  96(a0)
+	fsd fs6,  104(a0)
+	fsd fs7,  112(a0)
+	fsd fs8,  120(a0)
+	fsd fs9,  128(a0)
+	fsd fs10, 136(a0)
+	fsd fs11, 144(a0)
+#endif
+
+	li a0, 0
+	ret
diff --git a/src/signal/mips/restore.s b/src/signal/loongarch64/restore.s
index b6dadce0..d90a8ebb 100644
--- a/src/signal/mips/restore.s
+++ b/src/signal/loongarch64/restore.s
@@ -1,15 +1,10 @@
-.set noreorder
-
 .global __restore_rt
-.hidden __restore_rt
-.type   __restore_rt,@function
-__restore_rt:
-	li $2, 4193
-	syscall
-
 .global __restore
+.hidden __restore_rt
 .hidden __restore
+.type   __restore_rt,@function
 .type   __restore,@function
+__restore_rt:
 __restore:
-	li $2, 4119
-	syscall
+	li.w    $a7, 139
+	syscall 0
diff --git a/src/signal/loongarch64/sigsetjmp.s b/src/signal/loongarch64/sigsetjmp.s
new file mode 100644
index 00000000..9c0e3ae2
--- /dev/null
+++ b/src/signal/loongarch64/sigsetjmp.s
@@ -0,0 +1,25 @@
+.global sigsetjmp
+.global __sigsetjmp
+.type sigsetjmp,@function
+.type __sigsetjmp,@function
+sigsetjmp:
+__sigsetjmp:
+	beq     $a1, $zero, 1f
+	st.d    $ra, $a0, 184
+	st.d    $s0, $a0, 200  #184+8+8
+	move    $s0, $a0
+
+	la.global  $t0, setjmp
+	jirl       $ra, $t0, 0
+
+	move    $a1, $a0        # Return from 'setjmp' or 'longjmp'
+	move    $a0, $s0
+	ld.d    $ra, $a0, 184
+	ld.d    $s0, $a0, 200 #184+8+8
+
+.hidden __sigsetjmp_tail
+	la.global  $t0, __sigsetjmp_tail
+	jr         $t0
+1:
+	la.global  $t0, setjmp
+	jr         $t0
diff --git a/src/signal/mips64/restore.s b/src/signal/mips64/restore.s
deleted file mode 100644
index 401f8e73..00000000
--- a/src/signal/mips64/restore.s
+++ /dev/null
@@ -1,11 +0,0 @@
-.set	noreorder
-.global	__restore_rt
-.global	__restore
-.hidden __restore_rt
-.hidden __restore
-.type	__restore_rt,@function
-.type	__restore,@function
-__restore_rt:
-__restore:
-	li	$2,5211
-	syscall
diff --git a/src/signal/mipsn32/restore.s b/src/signal/mipsn32/restore.s
deleted file mode 100644
index 4cd4e1b4..00000000
--- a/src/signal/mipsn32/restore.s
+++ /dev/null
@@ -1,11 +0,0 @@
-.set	noreorder
-.global	__restore_rt
-.global	__restore
-.hidden __restore_rt
-.hidden __restore
-.type	__restore_rt,@function
-.type	__restore,@function
-__restore_rt:
-__restore:
-	li	$2,6211
-	syscall
diff --git a/src/signal/riscv32/restore.s b/src/signal/riscv32/restore.s
new file mode 100644
index 00000000..40012c75
--- /dev/null
+++ b/src/signal/riscv32/restore.s
@@ -0,0 +1,8 @@
+.global __restore
+.type __restore, %function
+__restore:
+.global __restore_rt
+.type __restore_rt, %function
+__restore_rt:
+	li a7, 139 # SYS_rt_sigreturn
+	ecall
diff --git a/src/signal/riscv32/sigsetjmp.s b/src/signal/riscv32/sigsetjmp.s
new file mode 100644
index 00000000..c1caeab1
--- /dev/null
+++ b/src/signal/riscv32/sigsetjmp.s
@@ -0,0 +1,23 @@
+.global sigsetjmp
+.global __sigsetjmp
+.type sigsetjmp, %function
+.type __sigsetjmp, %function
+sigsetjmp:
+__sigsetjmp:
+	bnez a1, 1f
+	tail setjmp
+1:
+
+	sw ra, 152(a0)
+	sw s0, 164(a0)
+	mv s0, a0
+
+	call setjmp
+
+	mv a1, a0
+	mv a0, s0
+	lw s0, 164(a0)
+	lw ra, 152(a0)
+
+.hidden __sigsetjmp_tail
+	tail __sigsetjmp_tail
diff --git a/src/signal/sh/sigsetjmp.s b/src/signal/sh/sigsetjmp.s
index 1e2270be..f0f604e2 100644
--- a/src/signal/sh/sigsetjmp.s
+++ b/src/signal/sh/sigsetjmp.s
@@ -27,7 +27,7 @@ __sigsetjmp:
 
 	mov.l 3f, r0
 4:	braf r0
-	 mov.l @(4+8,r4), r8
+	 mov.l @(4+8,r6), r8
 
 9:	mov.l 5f, r0
 6:	braf r0
diff --git a/src/signal/sigaction.c b/src/signal/sigaction.c
index 2203471b..e45308fa 100644
--- a/src/signal/sigaction.c
+++ b/src/signal/sigaction.c
@@ -44,8 +44,11 @@ int __libc_sigaction(int sig, const struct sigaction *restrict sa, struct sigact
 			}
 		}
 		ksa.handler = sa->sa_handler;
-		ksa.flags = sa->sa_flags | SA_RESTORER;
+		ksa.flags = sa->sa_flags;
+#ifdef SA_RESTORER
+		ksa.flags |= SA_RESTORER;
 		ksa.restorer = (sa->sa_flags & SA_SIGINFO) ? __restore_rt : __restore;
+#endif
 		memcpy(&ksa.mask, &sa->sa_mask, _NSIG/8);
 	}
 	int r = __syscall(SYS_rt_sigaction, sig, sa?&ksa:0, old?&ksa_old:0, _NSIG/8);
diff --git a/src/stat/__xstat.c b/src/stat/__xstat.c
index 630936a0..b4560df7 100644
--- a/src/stat/__xstat.c
+++ b/src/stat/__xstat.c
@@ -22,11 +22,6 @@ int __xstat(int ver, const char *path, struct stat *buf)
 	return stat(path, buf);
 }
 
-weak_alias(__fxstat, __fxstat64);
-weak_alias(__fxstatat, __fxstatat64);
-weak_alias(__lxstat, __lxstat64);
-weak_alias(__xstat, __xstat64);
-
 #endif
 
 int __xmknod(int ver, const char *path, mode_t mode, dev_t *dev)
diff --git a/src/stat/fchmodat.c b/src/stat/fchmodat.c
index bc581050..92c9d1b0 100644
--- a/src/stat/fchmodat.c
+++ b/src/stat/fchmodat.c
@@ -5,13 +5,16 @@
 
 int fchmodat(int fd, const char *path, mode_t mode, int flag)
 {
-	if (!flag) return syscall(SYS_fchmodat, fd, path, mode, flag);
+	if (!flag) return syscall(SYS_fchmodat, fd, path, mode);
+
+	int ret = __syscall(SYS_fchmodat2, fd, path, mode, flag);
+	if (ret != -ENOSYS) return __syscall_ret(ret);
 
 	if (flag != AT_SYMLINK_NOFOLLOW)
 		return __syscall_ret(-EINVAL);
 
 	struct stat st;
-	int ret, fd2;
+	int fd2;
 	char proc[15+3*sizeof(int)];
 
 	if (fstatat(fd, path, &st, flag))
diff --git a/src/stat/fstat.c b/src/stat/fstat.c
index 27db0ccb..fd28b8ac 100644
--- a/src/stat/fstat.c
+++ b/src/stat/fstat.c
@@ -11,7 +11,3 @@ int __fstat(int fd, struct stat *st)
 }
 
 weak_alias(__fstat, fstat);
-
-#if !_REDIR_TIME64
-weak_alias(fstat, fstat64);
-#endif
diff --git a/src/stat/fstatat.c b/src/stat/fstatat.c
index 74c51cf5..9eed063b 100644
--- a/src/stat/fstatat.c
+++ b/src/stat/fstatat.c
@@ -36,6 +36,7 @@ static int fstatat_statx(int fd, const char *restrict path, struct stat *restric
 {
 	struct statx stx;
 
+	flag |= AT_NO_AUTOMOUNT;
 	int ret = __syscall(SYS_statx, fd, path, flag, 0x7ff, &stx);
 	if (ret) return ret;
 
@@ -151,7 +152,3 @@ int __fstatat(int fd, const char *restrict path, struct stat *restrict st, int f
 }
 
 weak_alias(__fstatat, fstatat);
-
-#if !_REDIR_TIME64
-weak_alias(fstatat, fstatat64);
-#endif
diff --git a/src/stat/lstat.c b/src/stat/lstat.c
index 6fe004de..6822fcae 100644
--- a/src/stat/lstat.c
+++ b/src/stat/lstat.c
@@ -5,7 +5,3 @@ int lstat(const char *restrict path, struct stat *restrict buf)
 {
 	return fstatat(AT_FDCWD, path, buf, AT_SYMLINK_NOFOLLOW);
 }
-
-#if !_REDIR_TIME64
-weak_alias(lstat, lstat64);
-#endif
diff --git a/src/stat/stat.c b/src/stat/stat.c
index ea70efc4..23570e7a 100644
--- a/src/stat/stat.c
+++ b/src/stat/stat.c
@@ -5,7 +5,3 @@ int stat(const char *restrict path, struct stat *restrict buf)
 {
 	return fstatat(AT_FDCWD, path, buf, 0);
 }
-
-#if !_REDIR_TIME64
-weak_alias(stat, stat64);
-#endif
diff --git a/src/stat/statvfs.c b/src/stat/statvfs.c
index f65d1b54..bc12da8b 100644
--- a/src/stat/statvfs.c
+++ b/src/stat/statvfs.c
@@ -39,6 +39,7 @@ static void fixup(struct statvfs *out, const struct statfs *in)
 	out->f_fsid = in->f_fsid.__val[0];
 	out->f_flag = in->f_flags;
 	out->f_namemax = in->f_namelen;
+	out->f_type = in->f_type;
 }
 
 int statvfs(const char *restrict path, struct statvfs *restrict buf)
@@ -56,8 +57,3 @@ int fstatvfs(int fd, struct statvfs *buf)
 	fixup(buf, &kbuf);
 	return 0;
 }
-
-weak_alias(statvfs, statvfs64);
-weak_alias(statfs, statfs64);
-weak_alias(fstatvfs, fstatvfs64);
-weak_alias(fstatfs, fstatfs64);
diff --git a/src/stdio/fgetpos.c b/src/stdio/fgetpos.c
index 50813d2c..392f7323 100644
--- a/src/stdio/fgetpos.c
+++ b/src/stdio/fgetpos.c
@@ -7,5 +7,3 @@ int fgetpos(FILE *restrict f, fpos_t *restrict pos)
 	*(long long *)pos = off;
 	return 0;
 }
-
-weak_alias(fgetpos, fgetpos64);
diff --git a/src/stdio/fgets.c b/src/stdio/fgets.c
index 6171f398..4a100b39 100644
--- a/src/stdio/fgets.c
+++ b/src/stdio/fgets.c
@@ -12,13 +12,14 @@ char *fgets(char *restrict s, int n, FILE *restrict f)
 
 	FLOCK(f);
 
-	if (n--<=1) {
+	if (n<=1) {
 		f->mode |= f->mode-1;
 		FUNLOCK(f);
-		if (n) return 0;
+		if (n<1) return 0;
 		*s = 0;
 		return s;
 	}
+	n--;
 
 	while (n) {
 		if (f->rpos != f->rend) {
diff --git a/src/stdio/fopen.c b/src/stdio/fopen.c
index e1b91e12..80bc341e 100644
--- a/src/stdio/fopen.c
+++ b/src/stdio/fopen.c
@@ -29,5 +29,3 @@ FILE *fopen(const char *restrict filename, const char *restrict mode)
 	__syscall(SYS_close, fd);
 	return 0;
 }
-
-weak_alias(fopen, fopen64);
diff --git a/src/stdio/freopen.c b/src/stdio/freopen.c
index 615d4b47..1641a4c5 100644
--- a/src/stdio/freopen.c
+++ b/src/stdio/freopen.c
@@ -40,6 +40,8 @@ FILE *freopen(const char *restrict filename, const char *restrict mode, FILE *re
 		fclose(f2);
 	}
 
+	f->mode = 0;
+	f->locale = 0;
 	FUNLOCK(f);
 	return f;
 
@@ -49,5 +51,3 @@ fail:
 	fclose(f);
 	return NULL;
 }
-
-weak_alias(freopen, freopen64);
diff --git a/src/stdio/fseek.c b/src/stdio/fseek.c
index c07f7e95..c7425802 100644
--- a/src/stdio/fseek.c
+++ b/src/stdio/fseek.c
@@ -46,5 +46,3 @@ int fseek(FILE *f, long off, int whence)
 }
 
 weak_alias(__fseeko, fseeko);
-
-weak_alias(fseeko, fseeko64);
diff --git a/src/stdio/fsetpos.c b/src/stdio/fsetpos.c
index 77ab8d82..779cb3cc 100644
--- a/src/stdio/fsetpos.c
+++ b/src/stdio/fsetpos.c
@@ -4,5 +4,3 @@ int fsetpos(FILE *f, const fpos_t *pos)
 {
 	return __fseeko(f, *(const long long *)pos, SEEK_SET);
 }
-
-weak_alias(fsetpos, fsetpos64);
diff --git a/src/stdio/ftell.c b/src/stdio/ftell.c
index 1a2afbbc..1e1a08d8 100644
--- a/src/stdio/ftell.c
+++ b/src/stdio/ftell.c
@@ -37,5 +37,3 @@ long ftell(FILE *f)
 }
 
 weak_alias(__ftello, ftello);
-
-weak_alias(ftello, ftello64);
diff --git a/src/stdio/open_wmemstream.c b/src/stdio/open_wmemstream.c
index ed1b561d..b8ae4a79 100644
--- a/src/stdio/open_wmemstream.c
+++ b/src/stdio/open_wmemstream.c
@@ -40,8 +40,12 @@ fail:
 static size_t wms_write(FILE *f, const unsigned char *buf, size_t len)
 {
 	struct cookie *c = f->cookie;
-	size_t len2;
+	size_t len2 = f->wpos - f->wbase;
 	wchar_t *newbuf;
+	if (len2) {
+		f->wpos = f->wbase;
+		if (wms_write(f, f->wbase, len2) < len2) return 0;
+	}
 	if (len + c->pos >= c->space) {
 		len2 = 2*c->space+1 | c->pos+len+1;
 		if (len2 > SSIZE_MAX/4) return 0;
diff --git a/src/stdio/pclose.c b/src/stdio/pclose.c
index 080a4262..c64da405 100644
--- a/src/stdio/pclose.c
+++ b/src/stdio/pclose.c
@@ -7,7 +7,7 @@ int pclose(FILE *f)
 	int status, r;
 	pid_t pid = f->pipe_pid;
 	fclose(f);
-	while ((r=__syscall(SYS_wait4, pid, &status, 0, 0)) == -EINTR);
+	while ((r=__sys_wait4(pid, &status, 0, 0)) == -EINTR);
 	if (r<0) return __syscall_ret(r);
 	return status;
 }
diff --git a/src/stdio/tmpfile.c b/src/stdio/tmpfile.c
index ae493987..2fa8803f 100644
--- a/src/stdio/tmpfile.c
+++ b/src/stdio/tmpfile.c
@@ -27,5 +27,3 @@ FILE *tmpfile(void)
 	}
 	return 0;
 }
-
-weak_alias(tmpfile, tmpfile64);
diff --git a/src/stdio/vfprintf.c b/src/stdio/vfprintf.c
index 9b961e7f..497c5e19 100644
--- a/src/stdio/vfprintf.c
+++ b/src/stdio/vfprintf.c
@@ -52,7 +52,7 @@ static const unsigned char states[]['z'-'A'+1] = {
 		S('o') = UINT, S('u') = UINT, S('x') = UINT, S('X') = UINT,
 		S('e') = DBL, S('f') = DBL, S('g') = DBL, S('a') = DBL,
 		S('E') = DBL, S('F') = DBL, S('G') = DBL, S('A') = DBL,
-		S('c') = CHAR, S('C') = INT,
+		S('c') = INT, S('C') = UINT,
 		S('s') = PTR, S('S') = PTR, S('p') = UIPTR, S('n') = PTR,
 		S('m') = NOARG,
 		S('l') = LPRE, S('h') = HPRE, S('L') = BIGLPRE,
@@ -62,7 +62,7 @@ static const unsigned char states[]['z'-'A'+1] = {
 		S('o') = ULONG, S('u') = ULONG, S('x') = ULONG, S('X') = ULONG,
 		S('e') = DBL, S('f') = DBL, S('g') = DBL, S('a') = DBL,
 		S('E') = DBL, S('F') = DBL, S('G') = DBL, S('A') = DBL,
-		S('c') = INT, S('s') = PTR, S('n') = PTR,
+		S('c') = UINT, S('s') = PTR, S('n') = PTR,
 		S('l') = LLPRE,
 	}, { /* 2: ll-prefixed */
 		S('d') = LLONG, S('i') = LLONG,
@@ -132,7 +132,7 @@ static void pop_arg(union arg *arg, int type, va_list *ap)
 
 static void out(FILE *f, const char *s, size_t l)
 {
-	if (!(f->flags & F_ERR)) __fwritex((void *)s, l, f);
+	if (!ferror(f)) __fwritex((void *)s, l, f);
 }
 
 static void pad(FILE *f, char c, int w, int l, int fl)
@@ -437,7 +437,7 @@ static int printf_core(FILE *f, const char *fmt, va_list *ap, union arg *nl_arg,
 	unsigned st, ps;
 	int cnt=0, l=0;
 	size_t i;
-	char buf[sizeof(uintmax_t)*3+3+LDBL_MANT_DIG/4];
+	char buf[sizeof(uintmax_t)*3];
 	const char *prefix;
 	int t, pl;
 	wchar_t wc[2], *ws;
@@ -478,8 +478,8 @@ static int printf_core(FILE *f, const char *fmt, va_list *ap, union arg *nl_arg,
 		if (*s=='*') {
 			if (isdigit(s[1]) && s[2]=='$') {
 				l10n=1;
-				nl_type[s[1]-'0'] = INT;
-				w = nl_arg[s[1]-'0'].i;
+				if (!f) nl_type[s[1]-'0'] = INT, w = 0;
+				else w = nl_arg[s[1]-'0'].i;
 				s+=3;
 			} else if (!l10n) {
 				w = f ? va_arg(*ap, int) : 0;
@@ -491,8 +491,8 @@ static int printf_core(FILE *f, const char *fmt, va_list *ap, union arg *nl_arg,
 		/* Read precision */
 		if (*s=='.' && s[1]=='*') {
 			if (isdigit(s[2]) && s[3]=='$') {
-				nl_type[s[2]-'0'] = INT;
-				p = nl_arg[s[2]-'0'].i;
+				if (!f) nl_type[s[2]-'0'] = INT, p = 0;
+				else p = nl_arg[s[2]-'0'].i;
 				s+=4;
 			} else if (!l10n) {
 				p = f ? va_arg(*ap, int) : 0;
@@ -521,13 +521,18 @@ static int printf_core(FILE *f, const char *fmt, va_list *ap, union arg *nl_arg,
 		if (st==NOARG) {
 			if (argpos>=0) goto inval;
 		} else {
-			if (argpos>=0) nl_type[argpos]=st, arg=nl_arg[argpos];
-			else if (f) pop_arg(&arg, st, ap);
+			if (argpos>=0) {
+				if (!f) nl_type[argpos]=st;
+				else arg=nl_arg[argpos];
+			} else if (f) pop_arg(&arg, st, ap);
 			else return 0;
 		}
 
 		if (!f) continue;
 
+		/* Do not process any new directives once in error state. */
+		if (ferror(f)) return -1;
+
 		z = buf + sizeof(buf);
 		prefix = "-+   0X0x";
 		pl = 0;
@@ -583,6 +588,7 @@ static int printf_core(FILE *f, const char *fmt, va_list *ap, union arg *nl_arg,
 			}
 			p = MAX(p, z-a + !arg.i);
 			break;
+		narrow_c:
 		case 'c':
 			*(a=z-(p=1))=arg.i;
 			fl &= ~ZERO_PAD;
@@ -597,6 +603,7 @@ static int printf_core(FILE *f, const char *fmt, va_list *ap, union arg *nl_arg,
 			fl &= ~ZERO_PAD;
 			break;
 		case 'C':
+			if (!arg.i) goto narrow_c;
 			wc[0] = arg.i;
 			wc[1] = 0;
 			arg.p = wc;
@@ -672,7 +679,7 @@ int vfprintf(FILE *restrict f, const char *restrict fmt, va_list ap)
 
 	FLOCK(f);
 	olderr = f->flags & F_ERR;
-	if (f->mode < 1) f->flags &= ~F_ERR;
+	f->flags &= ~F_ERR;
 	if (!f->buf_size) {
 		saved_buf = f->buf;
 		f->buf = internal_buf;
@@ -688,7 +695,7 @@ int vfprintf(FILE *restrict f, const char *restrict fmt, va_list ap)
 		f->buf_size = 0;
 		f->wpos = f->wbase = f->wend = 0;
 	}
-	if (f->flags & F_ERR) ret = -1;
+	if (ferror(f)) ret = -1;
 	f->flags |= olderr;
 	FUNLOCK(f);
 	va_end(ap2);
diff --git a/src/stdio/vfwprintf.c b/src/stdio/vfwprintf.c
index 85b036c3..59d5471b 100644
--- a/src/stdio/vfwprintf.c
+++ b/src/stdio/vfwprintf.c
@@ -45,7 +45,7 @@ static const unsigned char states[]['z'-'A'+1] = {
 		S('o') = UINT, S('u') = UINT, S('x') = UINT, S('X') = UINT,
 		S('e') = DBL, S('f') = DBL, S('g') = DBL, S('a') = DBL,
 		S('E') = DBL, S('F') = DBL, S('G') = DBL, S('A') = DBL,
-		S('c') = CHAR, S('C') = INT,
+		S('c') = INT, S('C') = UINT,
 		S('s') = PTR, S('S') = PTR, S('p') = UIPTR, S('n') = PTR,
 		S('m') = NOARG,
 		S('l') = LPRE, S('h') = HPRE, S('L') = BIGLPRE,
@@ -55,7 +55,7 @@ static const unsigned char states[]['z'-'A'+1] = {
 		S('o') = ULONG, S('u') = ULONG, S('x') = ULONG, S('X') = ULONG,
 		S('e') = DBL, S('f') = DBL, S('g') = DBL, S('a') = DBL,
 		S('E') = DBL, S('F') = DBL, S('G') = DBL, S('A') = DBL,
-		S('c') = INT, S('s') = PTR, S('n') = PTR,
+		S('c') = UINT, S('s') = PTR, S('n') = PTR,
 		S('l') = LLPRE,
 	}, { /* 2: ll-prefixed */
 		S('d') = LLONG, S('i') = LLONG,
@@ -125,7 +125,13 @@ static void pop_arg(union arg *arg, int type, va_list *ap)
 
 static void out(FILE *f, const wchar_t *s, size_t l)
 {
-	while (l-- && !(f->flags & F_ERR)) fputwc(*s++, f);
+	while (l-- && !ferror(f)) fputwc(*s++, f);
+}
+
+static void pad(FILE *f, int n, int fl)
+{
+	if ((fl & LEFT_ADJ) || !n || ferror(f)) return;
+	fprintf(f, "%*s", n, "");
 }
 
 static int getint(wchar_t **s) {
@@ -242,6 +248,10 @@ static int wprintf_core(FILE *f, const wchar_t *fmt, va_list *ap, union arg *nl_
 		}
 
 		if (!f) continue;
+
+		/* Do not process any new directives once in error state. */
+		if (ferror(f)) return -1;
+
 		t = s[-1];
 		if (ps && (t&15)==3) t&=~32;
 
@@ -258,25 +268,22 @@ static int wprintf_core(FILE *f, const wchar_t *fmt, va_list *ap, union arg *nl_
 			}
 			continue;
 		case 'c':
+		case 'C':
 			if (w<1) w=1;
-			if (w>1 && !(fl&LEFT_ADJ)) fprintf(f, "%*s", w-1, "");
-			fputwc(btowc(arg.i), f);
-			if (w>1 && (fl&LEFT_ADJ)) fprintf(f, "%*s", w-1, "");
+			pad(f, w-1, fl);
+			out(f, &(wchar_t){t=='C' ? arg.i : btowc(arg.i)}, 1);
+			pad(f, w-1, fl^LEFT_ADJ);
 			l = w;
 			continue;
-		case 'C':
-			fputwc(arg.i, f);
-			l = 1;
-			continue;
 		case 'S':
 			a = arg.p;
 			z = a + wcsnlen(a, p<0 ? INT_MAX : p);
 			if (p<0 && *z) goto overflow;
 			p = z-a;
 			if (w<p) w=p;
-			if (!(fl&LEFT_ADJ)) fprintf(f, "%*s", w-p, "");
+			pad(f, w-p, fl);
 			out(f, a, p);
-			if ((fl&LEFT_ADJ)) fprintf(f, "%*s", w-p, "");
+			pad(f, w-p, fl^LEFT_ADJ);
 			l=w;
 			continue;
 		case 'm':
@@ -289,14 +296,14 @@ static int wprintf_core(FILE *f, const wchar_t *fmt, va_list *ap, union arg *nl_
 			if (p<0 && *bs) goto overflow;
 			p=l;
 			if (w<p) w=p;
-			if (!(fl&LEFT_ADJ)) fprintf(f, "%*s", w-p, "");
+			pad(f, w-p, fl);
 			bs = arg.p;
 			while (l--) {
 				i=mbtowc(&wc, bs, MB_LEN_MAX);
 				bs+=i;
-				fputwc(wc, f);
+				out(f, &wc, 1);
 			}
-			if ((fl&LEFT_ADJ)) fprintf(f, "%*s", w-p, "");
+			pad(f, w-p, fl^LEFT_ADJ);
 			l=w;
 			continue;
 		}
@@ -340,8 +347,8 @@ overflow:
 int vfwprintf(FILE *restrict f, const wchar_t *restrict fmt, va_list ap)
 {
 	va_list ap2;
-	int nl_type[NL_ARGMAX] = {0};
-	union arg nl_arg[NL_ARGMAX];
+	int nl_type[NL_ARGMAX+1] = {0};
+	union arg nl_arg[NL_ARGMAX+1];
 	int olderr;
 	int ret;
 
@@ -357,7 +364,7 @@ int vfwprintf(FILE *restrict f, const wchar_t *restrict fmt, va_list ap)
 	olderr = f->flags & F_ERR;
 	f->flags &= ~F_ERR;
 	ret = wprintf_core(f, fmt, &ap2, nl_arg, nl_type);
-	if (f->flags & F_ERR) ret = -1;
+	if (ferror(f)) ret = -1;
 	f->flags |= olderr;
 	FUNLOCK(f);
 	va_end(ap2);
diff --git a/src/stdio/vsnprintf.c b/src/stdio/vsnprintf.c
index b3510a63..409b9c85 100644
--- a/src/stdio/vsnprintf.c
+++ b/src/stdio/vsnprintf.c
@@ -45,11 +45,6 @@ int vsnprintf(char *restrict s, size_t n, const char *restrict fmt, va_list ap)
 		.cookie = &c,
 	};
 
-	if (n > INT_MAX) {
-		errno = EOVERFLOW;
-		return -1;
-	}
-
 	*c.s = 0;
 	return vfprintf(&f, fmt, ap);
 }
diff --git a/src/stdio/vswprintf.c b/src/stdio/vswprintf.c
index 7f98c5c9..5e9a4dad 100644
--- a/src/stdio/vswprintf.c
+++ b/src/stdio/vswprintf.c
@@ -18,6 +18,7 @@ static size_t sw_write(FILE *f, const unsigned char *s, size_t l)
 	if (s!=f->wbase && sw_write(f, f->wbase, f->wpos-f->wbase)==-1)
 		return -1;
 	while (c->l && l && (i=mbtowc(c->ws, (void *)s, l))>=0) {
+		if (!i) i=1;
 		s+=i;
 		l-=i;
 		c->l--;
@@ -50,9 +51,6 @@ int vswprintf(wchar_t *restrict s, size_t n, const wchar_t *restrict fmt, va_lis
 
 	if (!n) {
 		return -1;
-	} else if (n > INT_MAX) {
-		errno = EOVERFLOW;
-		return -1;
 	}
 	r = vfwprintf(&f, fmt, ap);
 	sw_write(&f, 0, 0);
diff --git a/src/string/strverscmp.c b/src/string/strverscmp.c
index 4daf276d..16c1da22 100644
--- a/src/string/strverscmp.c
+++ b/src/string/strverscmp.c
@@ -18,9 +18,9 @@ int strverscmp(const char *l0, const char *r0)
 		else if (c!='0') z=0;
 	}
 
-	if (l[dp]!='0' && r[dp]!='0') {
-		/* If we're not looking at a digit sequence that began
-		 * with a zero, longest digit string is greater. */
+	if (l[dp]-'1'<9U && r[dp]-'1'<9U) {
+		/* If we're looking at non-degenerate digit sequences starting
+		 * with nonzero digits, longest digit string is greater. */
 		for (j=i; isdigit(l[j]); j++)
 			if (!isdigit(r[j])) return 1;
 		if (isdigit(r[j])) return -1;
diff --git a/src/string/wcscmp.c b/src/string/wcscmp.c
index 26eeee70..286ec3ea 100644
--- a/src/string/wcscmp.c
+++ b/src/string/wcscmp.c
@@ -3,5 +3,5 @@
 int wcscmp(const wchar_t *l, const wchar_t *r)
 {
 	for (; *l==*r && *l && *r; l++, r++);
-	return *l - *r;
+	return *l < *r ? -1 : *l > *r;
 }
diff --git a/src/string/wcsncmp.c b/src/string/wcsncmp.c
index 4ab32a92..2b3558bf 100644
--- a/src/string/wcsncmp.c
+++ b/src/string/wcsncmp.c
@@ -3,5 +3,5 @@
 int wcsncmp(const wchar_t *l, const wchar_t *r, size_t n)
 {
 	for (; n && *l==*r && *l && *r; n--, l++, r++);
-	return n ? *l - *r : 0;
+	return n ? (*l < *r ? -1 : *l > *r) : 0;
 }
diff --git a/src/string/wmemcmp.c b/src/string/wmemcmp.c
index 2a193263..717d77b1 100644
--- a/src/string/wmemcmp.c
+++ b/src/string/wmemcmp.c
@@ -3,5 +3,5 @@
 int wmemcmp(const wchar_t *l, const wchar_t *r, size_t n)
 {
 	for (; n && *l==*r; n--, l++, r++);
-	return n ? *l-*r : 0;
+	return n ? (*l < *r ? -1 : *l > *r) : 0;
 }
diff --git a/src/temp/mkostemp.c b/src/temp/mkostemp.c
index d8dcb805..e3dfdd91 100644
--- a/src/temp/mkostemp.c
+++ b/src/temp/mkostemp.c
@@ -5,5 +5,3 @@ int mkostemp(char *template, int flags)
 {
 	return __mkostemps(template, 0, flags);
 }
-
-weak_alias(mkostemp, mkostemp64);
diff --git a/src/temp/mkostemps.c b/src/temp/mkostemps.c
index ef24eeae..093d2380 100644
--- a/src/temp/mkostemps.c
+++ b/src/temp/mkostemps.c
@@ -26,4 +26,3 @@ int __mkostemps(char *template, int len, int flags)
 }
 
 weak_alias(__mkostemps, mkostemps);
-weak_alias(__mkostemps, mkostemps64);
diff --git a/src/temp/mkstemp.c b/src/temp/mkstemp.c
index 166b8afe..76c835bb 100644
--- a/src/temp/mkstemp.c
+++ b/src/temp/mkstemp.c
@@ -4,5 +4,3 @@ int mkstemp(char *template)
 {
 	return __mkostemps(template, 0, 0);
 }
-
-weak_alias(mkstemp, mkstemp64);
diff --git a/src/temp/mkstemps.c b/src/temp/mkstemps.c
index 6b7531b5..f8eabfec 100644
--- a/src/temp/mkstemps.c
+++ b/src/temp/mkstemps.c
@@ -5,5 +5,3 @@ int mkstemps(char *template, int len)
 {
 	return __mkostemps(template, len, 0);
 }
-
-weak_alias(mkstemps, mkstemps64);
diff --git a/src/thread/loongarch64/__set_thread_area.s b/src/thread/loongarch64/__set_thread_area.s
new file mode 100644
index 00000000..021307fc
--- /dev/null
+++ b/src/thread/loongarch64/__set_thread_area.s
@@ -0,0 +1,7 @@
+.global __set_thread_area
+.hidden __set_thread_area
+.type   __set_thread_area,@function
+__set_thread_area:
+	move $tp, $a0
+	move $a0, $zero
+	jr   $ra
diff --git a/src/thread/loongarch64/__unmapself.s b/src/thread/loongarch64/__unmapself.s
new file mode 100644
index 00000000..719ad056
--- /dev/null
+++ b/src/thread/loongarch64/__unmapself.s
@@ -0,0 +1,7 @@
+.global __unmapself
+.type   __unmapself, @function
+__unmapself:
+	li.d    $a7, 215   # call munmap
+	syscall 0
+	li.d    $a7, 93    # call exit
+	syscall 0
diff --git a/src/thread/loongarch64/clone.s b/src/thread/loongarch64/clone.s
new file mode 100644
index 00000000..a165b365
--- /dev/null
+++ b/src/thread/loongarch64/clone.s
@@ -0,0 +1,29 @@
+#__clone(func, stack, flags, arg, ptid, tls, ctid)
+#         a0,    a1,   a2,    a3,  a4,  a5,   a6
+# sys_clone(flags, stack, ptid, ctid, tls)
+#            a0,    a1,   a2,    a3,  a4
+
+.global __clone
+.hidden __clone
+.type __clone,@function
+__clone:
+	bstrins.d $a1, $zero, 3, 0   #stack to 16 align
+	# Save function pointer and argument pointer on new thread stack
+	addi.d  $a1, $a1, -16
+	st.d    $a0, $a1, 0     # save function pointer
+	st.d    $a3, $a1, 8     # save argument pointer
+	or      $a0, $a2, $zero
+	or      $a2, $a4, $zero
+	or      $a3, $a6, $zero
+	or      $a4, $a5, $zero
+	ori     $a7, $zero, 220
+	syscall 0               # call clone
+
+	beqz    $a0, 1f         # whether child process
+	jirl    $zero, $ra, 0   # parent process return
+1:
+	ld.d    $t8, $sp, 0     # function pointer
+	ld.d    $a0, $sp, 8     # argument pointer
+	jirl    $ra, $t8, 0     # call the user's function
+	ori     $a7, $zero, 93
+	syscall 0               # child process exit
diff --git a/src/thread/loongarch64/syscall_cp.s b/src/thread/loongarch64/syscall_cp.s
new file mode 100644
index 00000000..c057a97b
--- /dev/null
+++ b/src/thread/loongarch64/syscall_cp.s
@@ -0,0 +1,29 @@
+.global __cp_begin
+.hidden __cp_begin
+.global __cp_end
+.hidden __cp_end
+.global __cp_cancel
+.hidden __cp_cancel
+.hidden __cancel
+.global __syscall_cp_asm
+.hidden __syscall_cp_asm
+.type   __syscall_cp_asm,@function
+
+__syscall_cp_asm:
+__cp_begin:
+	ld.w $a0, $a0, 0
+	bnez $a0, __cp_cancel
+	move $t8, $a1    # reserve system call number
+	move $a0, $a2
+	move $a1, $a3
+	move $a2, $a4
+	move $a3, $a5
+	move $a4, $a6
+	move $a5, $a7
+	move $a7, $t8
+	syscall 0
+__cp_end:
+	jr $ra
+__cp_cancel:
+	la.local $t8, __cancel
+	jr $t8
diff --git a/src/thread/pthread_atfork.c b/src/thread/pthread_atfork.c
index 76497401..26d32543 100644
--- a/src/thread/pthread_atfork.c
+++ b/src/thread/pthread_atfork.c
@@ -1,7 +1,13 @@
 #include <pthread.h>
+#include <errno.h>
 #include "libc.h"
 #include "lock.h"
 
+#define malloc __libc_malloc
+#define calloc undef
+#define realloc undef
+#define free undef
+
 static struct atfork_funcs {
 	void (*prepare)(void);
 	void (*parent)(void);
@@ -34,7 +40,7 @@ void __fork_handler(int who)
 int pthread_atfork(void (*prepare)(void), void (*parent)(void), void (*child)(void))
 {
 	struct atfork_funcs *new = malloc(sizeof *new);
-	if (!new) return -1;
+	if (!new) return ENOMEM;
 
 	LOCK(lock);
 	new->next = funcs;
diff --git a/src/thread/pthread_cancel.c b/src/thread/pthread_cancel.c
index 2f9d5e97..139a6fc8 100644
--- a/src/thread/pthread_cancel.c
+++ b/src/thread/pthread_cancel.c
@@ -56,7 +56,12 @@ static void cancel_handler(int sig, siginfo_t *si, void *ctx)
 
 	_sigaddset(&uc->uc_sigmask, SIGCANCEL);
 
-	if (self->cancelasync || pc >= (uintptr_t)__cp_begin && pc < (uintptr_t)__cp_end) {
+	if (self->cancelasync) {
+		pthread_sigmask(SIG_SETMASK, &uc->uc_sigmask, 0);
+		__cancel();
+	}
+
+	if (pc >= (uintptr_t)__cp_begin && pc < (uintptr_t)__cp_end) {
 		uc->uc_mcontext.MC_PC = (uintptr_t)__cp_cancel;
 #ifdef CANCEL_GOT
 		uc->uc_mcontext.MC_GOT = CANCEL_GOT;
@@ -77,7 +82,7 @@ void __testcancel()
 static void init_cancellation()
 {
 	struct sigaction sa = {
-		.sa_flags = SA_SIGINFO | SA_RESTART,
+		.sa_flags = SA_SIGINFO | SA_RESTART | SA_ONSTACK,
 		.sa_sigaction = cancel_handler
 	};
 	memset(&sa.sa_mask, -1, _NSIG/8);
diff --git a/src/thread/pthread_create.c b/src/thread/pthread_create.c
index 6f187ee8..087f6206 100644
--- a/src/thread/pthread_create.c
+++ b/src/thread/pthread_create.c
@@ -107,6 +107,16 @@ _Noreturn void __pthread_exit(void *result)
 
 	/* At this point we are committed to thread termination. */
 
+	/* After the kernel thread exits, its tid may be reused. Clear it
+	 * to prevent inadvertent use and inform functions that would use
+	 * it that it's no longer available. At this point the killlock
+	 * may be released, since functions that use it will consistently
+	 * see the thread as having exited. Release it now so that no
+	 * remaining locks (except thread list) are held if we end up
+	 * resetting need_locks below. */
+	self->tid = 0;
+	UNLOCK(self->killlock);
+
 	/* Process robust list in userspace to handle non-pshared mutexes
 	 * and the detached thread case where the robust list head will
 	 * be invalid when the kernel would process it. */
@@ -159,12 +169,6 @@ _Noreturn void __pthread_exit(void *result)
 	a_store(&self->detach_state, DT_EXITED);
 	__wake(&self->detach_state, 1, 1);
 
-	/* After the kernel thread exits, its tid may be reused. Clear it
-	 * to prevent inadvertent use and inform functions that would use
-	 * it that it's no longer available. */
-	self->tid = 0;
-	UNLOCK(self->killlock);
-
 	for (;;) __syscall(SYS_exit, 0);
 }
 
diff --git a/src/thread/pthread_detach.c b/src/thread/pthread_detach.c
index 77772af2..d73a500e 100644
--- a/src/thread/pthread_detach.c
+++ b/src/thread/pthread_detach.c
@@ -5,8 +5,12 @@ static int __pthread_detach(pthread_t t)
 {
 	/* If the cas fails, detach state is either already-detached
 	 * or exiting/exited, and pthread_join will trap or cleanup. */
-	if (a_cas(&t->detach_state, DT_JOINABLE, DT_DETACHED) != DT_JOINABLE)
-		return __pthread_join(t, 0);
+	if (a_cas(&t->detach_state, DT_JOINABLE, DT_DETACHED) != DT_JOINABLE) {
+		int cs;
+		__pthread_setcancelstate(PTHREAD_CANCEL_DISABLE, &cs);
+		__pthread_join(t, 0);
+		__pthread_setcancelstate(cs, 0);
+	}
 	return 0;
 }
 
diff --git a/src/thread/pthread_key_create.c b/src/thread/pthread_key_create.c
index d1120941..39770c7a 100644
--- a/src/thread/pthread_key_create.c
+++ b/src/thread/pthread_key_create.c
@@ -1,4 +1,5 @@
 #include "pthread_impl.h"
+#include "fork_impl.h"
 
 volatile size_t __pthread_tsd_size = sizeof(void *) * PTHREAD_KEYS_MAX;
 void *__pthread_tsd_main[PTHREAD_KEYS_MAX] = { 0 };
@@ -20,6 +21,13 @@ static void dummy_0(void)
 weak_alias(dummy_0, __tl_lock);
 weak_alias(dummy_0, __tl_unlock);
 
+void __pthread_key_atfork(int who)
+{
+	if (who<0) __pthread_rwlock_rdlock(&key_lock);
+	else if (!who) __pthread_rwlock_unlock(&key_lock);
+	else key_lock = (pthread_rwlock_t)PTHREAD_RWLOCK_INITIALIZER;
+}
+
 int __pthread_key_create(pthread_key_t *k, void (*dtor)(void *))
 {
 	pthread_t self = __pthread_self();
diff --git a/src/thread/riscv32/__set_thread_area.s b/src/thread/riscv32/__set_thread_area.s
new file mode 100644
index 00000000..828154d2
--- /dev/null
+++ b/src/thread/riscv32/__set_thread_area.s
@@ -0,0 +1,6 @@
+.global __set_thread_area
+.type   __set_thread_area, %function
+__set_thread_area:
+	mv tp, a0
+	li a0, 0
+	ret
diff --git a/src/thread/riscv32/__unmapself.s b/src/thread/riscv32/__unmapself.s
new file mode 100644
index 00000000..2849119c
--- /dev/null
+++ b/src/thread/riscv32/__unmapself.s
@@ -0,0 +1,7 @@
+.global __unmapself
+.type __unmapself, %function
+__unmapself:
+	li a7, 215 # SYS_munmap
+	ecall
+	li a7, 93  # SYS_exit
+	ecall
diff --git a/src/thread/riscv32/clone.s b/src/thread/riscv32/clone.s
new file mode 100644
index 00000000..3102239d
--- /dev/null
+++ b/src/thread/riscv32/clone.s
@@ -0,0 +1,34 @@
+# __clone(func, stack, flags, arg, ptid, tls, ctid)
+#           a0,    a1,    a2,  a3,   a4,  a5,   a6
+
+# syscall(SYS_clone, flags, stack, ptid, tls, ctid)
+#                a7     a0,    a1,   a2,  a3,   a4
+
+.global __clone
+.type  __clone, %function
+__clone:
+	# Save func and arg to stack
+	addi a1, a1, -16
+	sw a0, 0(a1)
+	sw a3, 4(a1)
+
+	# Call SYS_clone
+	mv a0, a2
+	mv a2, a4
+	mv a3, a5
+	mv a4, a6
+	li a7, 220 # SYS_clone
+	ecall
+
+	beqz a0, 1f
+	# Parent
+	ret
+
+	# Child
+1:      lw a1, 0(sp)
+	lw a0, 4(sp)
+	jalr a1
+
+	# Exit
+	li a7, 93 # SYS_exit
+	ecall
diff --git a/src/thread/riscv32/syscall_cp.s b/src/thread/riscv32/syscall_cp.s
new file mode 100644
index 00000000..079d1ba0
--- /dev/null
+++ b/src/thread/riscv32/syscall_cp.s
@@ -0,0 +1,29 @@
+.global __cp_begin
+.hidden __cp_begin
+.global __cp_end
+.hidden __cp_end
+.global __cp_cancel
+.hidden __cp_cancel
+.hidden __cancel
+.global __syscall_cp_asm
+.hidden __syscall_cp_asm
+.type __syscall_cp_asm, %function
+__syscall_cp_asm:
+__cp_begin:
+	lw t0, 0(a0)
+	bnez t0, __cp_cancel
+
+	mv t0, a1
+	mv a0, a2
+	mv a1, a3
+	mv a2, a4
+	mv a3, a5
+	mv a4, a6
+	mv a5, a7
+	lw a6, 0(sp)
+	mv a7, t0
+	ecall
+__cp_end:
+	ret
+__cp_cancel:
+	tail __cancel
diff --git a/src/thread/sem_getvalue.c b/src/thread/sem_getvalue.c
index d9d83071..c0b7762d 100644
--- a/src/thread/sem_getvalue.c
+++ b/src/thread/sem_getvalue.c
@@ -1,8 +1,9 @@
 #include <semaphore.h>
+#include <limits.h>
 
 int sem_getvalue(sem_t *restrict sem, int *restrict valp)
 {
 	int val = sem->__val[0];
-	*valp = val < 0 ? 0 : val;
+	*valp = val & SEM_VALUE_MAX;
 	return 0;
 }
diff --git a/src/thread/sem_post.c b/src/thread/sem_post.c
index 31e3293d..5c2517f2 100644
--- a/src/thread/sem_post.c
+++ b/src/thread/sem_post.c
@@ -1,17 +1,21 @@
 #include <semaphore.h>
+#include <limits.h>
 #include "pthread_impl.h"
 
 int sem_post(sem_t *sem)
 {
-	int val, waiters, priv = sem->__val[2];
+	int val, new, waiters, priv = sem->__val[2];
 	do {
 		val = sem->__val[0];
 		waiters = sem->__val[1];
-		if (val == SEM_VALUE_MAX) {
+		if ((val & SEM_VALUE_MAX) == SEM_VALUE_MAX) {
 			errno = EOVERFLOW;
 			return -1;
 		}
-	} while (a_cas(sem->__val, val, val+1+(val<0)) != val);
-	if (val<0 || waiters) __wake(sem->__val, 1, priv);
+		new = val + 1;
+		if (waiters <= 1)
+			new &= ~0x80000000;
+	} while (a_cas(sem->__val, val, new) != val);
+	if (val<0) __wake(sem->__val, waiters>1 ? 1 : -1, priv);
 	return 0;
 }
diff --git a/src/thread/sem_timedwait.c b/src/thread/sem_timedwait.c
index 58d3ebfe..aa67376c 100644
--- a/src/thread/sem_timedwait.c
+++ b/src/thread/sem_timedwait.c
@@ -1,4 +1,5 @@
 #include <semaphore.h>
+#include <limits.h>
 #include "pthread_impl.h"
 
 static void cleanup(void *p)
@@ -13,14 +14,15 @@ int sem_timedwait(sem_t *restrict sem, const struct timespec *restrict at)
 	if (!sem_trywait(sem)) return 0;
 
 	int spins = 100;
-	while (spins-- && sem->__val[0] <= 0 && !sem->__val[1]) a_spin();
+	while (spins-- && !(sem->__val[0] & SEM_VALUE_MAX) && !sem->__val[1])
+		a_spin();
 
 	while (sem_trywait(sem)) {
-		int r;
+		int r, priv = sem->__val[2];
 		a_inc(sem->__val+1);
-		a_cas(sem->__val, 0, -1);
+		a_cas(sem->__val, 0, 0x80000000);
 		pthread_cleanup_push(cleanup, (void *)(sem->__val+1));
-		r = __timedwait_cp(sem->__val, -1, CLOCK_REALTIME, at, sem->__val[2]);
+		r = __timedwait_cp(sem->__val, 0x80000000, CLOCK_REALTIME, at, priv);
 		pthread_cleanup_pop(1);
 		if (r) {
 			errno = r;
diff --git a/src/thread/sem_trywait.c b/src/thread/sem_trywait.c
index 04edf46b..beb435da 100644
--- a/src/thread/sem_trywait.c
+++ b/src/thread/sem_trywait.c
@@ -1,12 +1,12 @@
 #include <semaphore.h>
+#include <limits.h>
 #include "pthread_impl.h"
 
 int sem_trywait(sem_t *sem)
 {
 	int val;
-	while ((val=sem->__val[0]) > 0) {
-		int new = val-1-(val==1 && sem->__val[1]);
-		if (a_cas(sem->__val, val, new)==val) return 0;
+	while ((val=sem->__val[0]) & SEM_VALUE_MAX) {
+		if (a_cas(sem->__val, val, val-1)==val) return 0;
 	}
 	errno = EAGAIN;
 	return -1;
diff --git a/src/thread/synccall.c b/src/thread/synccall.c
index d58c851f..38597254 100644
--- a/src/thread/synccall.c
+++ b/src/thread/synccall.c
@@ -11,7 +11,7 @@ weak_alias(dummy_0, __tl_unlock);
 
 static int target_tid;
 static void (*callback)(void *), *context;
-static sem_t target_sem, caller_sem;
+static sem_t target_sem, caller_sem, exit_sem;
 
 static void dummy(void *p)
 {
@@ -33,7 +33,7 @@ static void handler(int sig)
 	/* Inform caller we've complered the callback and wait
 	 * for the caller to release us to return. */
 	sem_post(&caller_sem);
-	sem_wait(&target_sem);
+	sem_wait(&exit_sem);
 
 	/* Inform caller we are returning and state is destroyable. */
 	sem_post(&caller_sem);
@@ -45,7 +45,7 @@ void __synccall(void (*func)(void *), void *ctx)
 {
 	sigset_t oldmask;
 	int cs, i, r;
-	struct sigaction sa = { .sa_flags = SA_RESTART, .sa_handler = handler };
+	struct sigaction sa = { .sa_flags = SA_RESTART | SA_ONSTACK, .sa_handler = handler };
 	pthread_t self = __pthread_self(), td;
 	int count = 0;
 
@@ -62,6 +62,7 @@ void __synccall(void (*func)(void *), void *ctx)
 
 	sem_init(&target_sem, 0, 0);
 	sem_init(&caller_sem, 0, 0);
+	sem_init(&exit_sem, 0, 0);
 
 	if (!libc.threads_minus_1 || __syscall(SYS_gettid) != self->tid)
 		goto single_threaded;
@@ -107,12 +108,13 @@ single_threaded:
 	/* Only release the caught threads once all threads, including the
 	 * caller, have returned from the callback function. */
 	for (i=0; i<count; i++)
-		sem_post(&target_sem);
+		sem_post(&exit_sem);
 	for (i=0; i<count; i++)
 		sem_wait(&caller_sem);
 
 	sem_destroy(&caller_sem);
 	sem_destroy(&target_sem);
+	sem_destroy(&exit_sem);
 
 	pthread_setcancelstate(cs, 0);
 	__tl_unlock();
diff --git a/src/time/__year_to_secs.c b/src/time/__year_to_secs.c
index 2824ec6d..b42f5a6d 100644
--- a/src/time/__year_to_secs.c
+++ b/src/time/__year_to_secs.c
@@ -10,9 +10,9 @@ long long __year_to_secs(long long year, int *is_leap)
 		return 31536000*(y-70) + 86400*leaps;
 	}
 
-	int cycles, centuries, leaps, rem;
+	int cycles, centuries, leaps, rem, dummy;
 
-	if (!is_leap) is_leap = &(int){0};
+	if (!is_leap) is_leap = &dummy;
 	cycles = (year-100) / 400;
 	rem = (year-100) % 400;
 	if (rem < 0) {
diff --git a/src/time/clock_gettime.c b/src/time/clock_gettime.c
index c7e66a51..4d2ec22f 100644
--- a/src/time/clock_gettime.c
+++ b/src/time/clock_gettime.c
@@ -42,6 +42,9 @@ static int cgt_init(clockid_t clk, struct timespec *ts)
 			p = cgt_time32_wrap;
 		}
 	}
+#ifdef VDSO_CGT_WORKAROUND
+	if (!__vdsosym(VDSO_CGT32_VER, VDSO_CGT32_SYM)) p = 0;
+#endif
 #endif
 	int (*f)(clockid_t, struct timespec *) =
 		(int (*)(clockid_t, struct timespec *))p;
diff --git a/src/time/strftime.c b/src/time/strftime.c
index cc53d536..c40246db 100644
--- a/src/time/strftime.c
+++ b/src/time/strftime.c
@@ -3,6 +3,7 @@
 #include <string.h>
 #include <langinfo.h>
 #include <locale.h>
+#include <ctype.h>
 #include <time.h>
 #include <limits.h>
 #include "locale_impl.h"
@@ -233,7 +234,12 @@ size_t __strftime_l(char *restrict s, size_t n, const char *restrict f, const st
 		pad = 0;
 		if (*f == '-' || *f == '_' || *f == '0') pad = *f++;
 		if ((plus = (*f == '+'))) f++;
-		width = strtoul(f, &p, 10);
+		if (isdigit(*f)) {
+			width = strtoul(f, &p, 10);
+		} else {
+			width = 0;
+			p = (void *)f;
+		}
 		if (*p == 'C' || *p == 'F' || *p == 'G' || *p == 'Y') {
 			if (!width && p!=f) width = 1;
 		} else {
diff --git a/src/time/timer_create.c b/src/time/timer_create.c
index 4bef2390..9216b3ab 100644
--- a/src/time/timer_create.c
+++ b/src/time/timer_create.c
@@ -43,6 +43,8 @@ static void *start(void *arg)
 	union sigval val = args->sev->sigev_value;
 
 	pthread_barrier_wait(&args->b);
+	if (self->cancel)
+		return 0;
 	for (;;) {
 		siginfo_t si;
 		while (sigwaitinfo(SIGTIMER_SET, &si) < 0);
@@ -59,7 +61,7 @@ static void *start(void *arg)
 
 int timer_create(clockid_t clk, struct sigevent *restrict evp, timer_t *restrict res)
 {
-	volatile static int init = 0;
+	static volatile int init = 0;
 	pthread_t td;
 	pthread_attr_t attr;
 	int r;
@@ -113,8 +115,10 @@ int timer_create(clockid_t clk, struct sigevent *restrict evp, timer_t *restrict
 		ksev.sigev_signo = SIGTIMER;
 		ksev.sigev_notify = SIGEV_THREAD_ID;
 		ksev.sigev_tid = td->tid;
-		if (syscall(SYS_timer_create, clk, &ksev, &timerid) < 0)
+		if (syscall(SYS_timer_create, clk, &ksev, &timerid) < 0) {
 			timerid = -1;
+			td->cancel = 1;
+		}
 		td->timer_id = timerid;
 		pthread_barrier_wait(&args.b);
 		if (timerid < 0) return -1;
diff --git a/src/unistd/dup3.c b/src/unistd/dup3.c
index f919f791..40798bde 100644
--- a/src/unistd/dup3.c
+++ b/src/unistd/dup3.c
@@ -9,12 +9,14 @@ int __dup3(int old, int new, int flags)
 	int r;
 #ifdef SYS_dup2
 	if (old==new) return __syscall_ret(-EINVAL);
-	if (flags & O_CLOEXEC) {
+	if (flags) {
 		while ((r=__syscall(SYS_dup3, old, new, flags))==-EBUSY);
 		if (r!=-ENOSYS) return __syscall_ret(r);
+		if (flags & ~O_CLOEXEC) return __syscall_ret(-EINVAL);
 	}
 	while ((r=__syscall(SYS_dup2, old, new))==-EBUSY);
-	if (flags & O_CLOEXEC) __syscall(SYS_fcntl, new, F_SETFD, FD_CLOEXEC);
+	if (r >= 0 && (flags & O_CLOEXEC))
+		__syscall(SYS_fcntl, new, F_SETFD, FD_CLOEXEC);
 #else
 	while ((r=__syscall(SYS_dup3, old, new, flags))==-EBUSY);
 #endif
diff --git a/src/unistd/faccessat.c b/src/unistd/faccessat.c
index 557503eb..43052dd7 100644
--- a/src/unistd/faccessat.c
+++ b/src/unistd/faccessat.c
@@ -53,7 +53,7 @@ int faccessat(int fd, const char *filename, int amode, int flag)
 	if (pid<0 || __syscall(SYS_read, p[0], &ret, sizeof ret) != sizeof(ret))
 		ret = -EBUSY;
 	__syscall(SYS_close, p[0]);
-	__syscall(SYS_wait4, pid, &status, __WCLONE, 0);
+	__sys_wait4(pid, &status, __WCLONE, 0);
 
 	__restore_sigs(&set);
 
diff --git a/src/unistd/ftruncate.c b/src/unistd/ftruncate.c
index b41be0fa..54ff34bc 100644
--- a/src/unistd/ftruncate.c
+++ b/src/unistd/ftruncate.c
@@ -5,5 +5,3 @@ int ftruncate(int fd, off_t length)
 {
 	return syscall(SYS_ftruncate, fd, __SYSCALL_LL_O(length));
 }
-
-weak_alias(ftruncate, ftruncate64);
diff --git a/src/unistd/lseek.c b/src/unistd/lseek.c
index b4984f3e..f5b66682 100644
--- a/src/unistd/lseek.c
+++ b/src/unistd/lseek.c
@@ -12,4 +12,3 @@ off_t __lseek(int fd, off_t offset, int whence)
 }
 
 weak_alias(__lseek, lseek);
-weak_alias(__lseek, lseek64);
diff --git a/src/unistd/mipsn32/lseek.c b/src/unistd/mipsn32/lseek.c
index 60e74a51..0f6cbcaa 100644
--- a/src/unistd/mipsn32/lseek.c
+++ b/src/unistd/mipsn32/lseek.c
@@ -17,4 +17,3 @@ off_t __lseek(int fd, off_t offset, int whence)
 }
 
 weak_alias(__lseek, lseek);
-weak_alias(__lseek, lseek64);
diff --git a/src/unistd/pipe2.c b/src/unistd/pipe2.c
index f24f74fb..a096990b 100644
--- a/src/unistd/pipe2.c
+++ b/src/unistd/pipe2.c
@@ -8,6 +8,7 @@ int pipe2(int fd[2], int flag)
 	if (!flag) return pipe(fd);
 	int ret = __syscall(SYS_pipe2, fd, flag);
 	if (ret != -ENOSYS) return __syscall_ret(ret);
+	if (flag & ~(O_CLOEXEC|O_NONBLOCK)) return __syscall_ret(-EINVAL);
 	ret = pipe(fd);
 	if (ret) return ret;
 	if (flag & O_CLOEXEC) {
diff --git a/src/unistd/pread.c b/src/unistd/pread.c
index 5681b045..b03fb0ad 100644
--- a/src/unistd/pread.c
+++ b/src/unistd/pread.c
@@ -5,5 +5,3 @@ ssize_t pread(int fd, void *buf, size_t size, off_t ofs)
 {
 	return syscall_cp(SYS_pread, fd, buf, size, __SYSCALL_LL_PRW(ofs));
 }
-
-weak_alias(pread, pread64);
diff --git a/src/unistd/preadv.c b/src/unistd/preadv.c
index 8376d60f..890ab403 100644
--- a/src/unistd/preadv.c
+++ b/src/unistd/preadv.c
@@ -8,5 +8,3 @@ ssize_t preadv(int fd, const struct iovec *iov, int count, off_t ofs)
 	return syscall_cp(SYS_preadv, fd, iov, count,
 		(long)(ofs), (long)(ofs>>32));
 }
-
-weak_alias(preadv, preadv64);
diff --git a/src/unistd/pwrite.c b/src/unistd/pwrite.c
index ca376576..869b69f0 100644
--- a/src/unistd/pwrite.c
+++ b/src/unistd/pwrite.c
@@ -5,5 +5,3 @@ ssize_t pwrite(int fd, const void *buf, size_t size, off_t ofs)
 {
 	return syscall_cp(SYS_pwrite, fd, buf, size, __SYSCALL_LL_PRW(ofs));
 }
-
-weak_alias(pwrite, pwrite64);
diff --git a/src/unistd/pwritev.c b/src/unistd/pwritev.c
index f5a612c4..becf9deb 100644
--- a/src/unistd/pwritev.c
+++ b/src/unistd/pwritev.c
@@ -8,5 +8,3 @@ ssize_t pwritev(int fd, const struct iovec *iov, int count, off_t ofs)
 	return syscall_cp(SYS_pwritev, fd, iov, count,
 		(long)(ofs), (long)(ofs>>32));
 }
-
-weak_alias(pwritev, pwritev64);
diff --git a/src/unistd/setxid.c b/src/unistd/setxid.c
index 487c1a16..a629ed4b 100644
--- a/src/unistd/setxid.c
+++ b/src/unistd/setxid.c
@@ -30,5 +30,5 @@ int __setxid(int nr, int id, int eid, int sid)
 	 * trigger the safety kill above. */
 	struct ctx c = { .nr = nr, .id = id, .eid = eid, .sid = sid, .ret = 1 };
 	__synccall(do_setxid, &c);
-	return __syscall_ret(c.ret);
+	return __syscall_ret(c.ret > 0 ? -EAGAIN : c.ret);
 }
diff --git a/src/unistd/truncate.c b/src/unistd/truncate.c
index 97296800..077351e1 100644
--- a/src/unistd/truncate.c
+++ b/src/unistd/truncate.c
@@ -5,5 +5,3 @@ int truncate(const char *path, off_t length)
 {
 	return syscall(SYS_truncate, path, __SYSCALL_LL_O(length));
 }
-
-weak_alias(truncate, truncate64);
diff --git a/src/unistd/x32/lseek.c b/src/unistd/x32/lseek.c
index 32636429..5f93292f 100644
--- a/src/unistd/x32/lseek.c
+++ b/src/unistd/x32/lseek.c
@@ -12,4 +12,3 @@ off_t __lseek(int fd, off_t offset, int whence)
 }
 
 weak_alias(__lseek, lseek);
-weak_alias(__lseek, lseek64);