Commit a5d8c310ee

@@ -101,3 +101,10 @@ static inline long __syscall6(long n, long a, long b, long c, long d, long e, lo
 #define SYSCALL_FADVISE_6_ARG
 
 #define SYSCALL_IPC_BROKEN_MODE
+
+#define VDSO_USEFUL
+#define VDSO_CGT32_SYM "__vdso_clock_gettime"
+#define VDSO_CGT32_VER "LINUX_2.6"
+#define VDSO_CGT_SYM "__vdso_clock_gettime64"
+#define VDSO_CGT_VER "LINUX_2.6"
+#define VDSO_CGT_WORKAROUND 1

@@ -1,1 +0,0 @@
-COMPAT_SRC_DIRS = compat/time32

@@ -1,1 +0,0 @@
-COMPAT_SRC_DIRS = compat/time32

@@ -66,7 +66,6 @@ typedef struct __ucontext {
 #define SA_RESTART    0x10000000
 #define SA_NODEFER    0x40000000
 #define SA_RESETHAND  0x80000000
-#define SA_RESTORER   0x04000000
 
 #undef SIG_BLOCK
 #undef SIG_UNBLOCK

@@ -1,1 +0,0 @@
-COMPAT_SRC_DIRS = compat/time32

@@ -4,10 +4,7 @@ struct k_sigaction {
 	unsigned flags;
 	void (*handler)(int);
 	unsigned long mask[4];
-	/* The following field is past the end of the structure the
-	 * kernel will read or write, and exists only to avoid having
-	 * mips-specific preprocessor conditionals in sigaction.c. */
-	void (*restorer)();
+	void *unused;
 };
 
 hidden void __restore(), __restore_rt();

@@ -29,6 +29,7 @@
 
 #define NEED_MIPS_GOT_RELOCS 1
 #define DT_DEBUG_INDIRECT DT_MIPS_RLD_MAP
+#define DT_DEBUG_INDIRECT_REL DT_MIPS_RLD_MAP_REL
 #define ARCH_SYM_REJECT_UND(s) (!((s)->st_other & STO_MIPS_PLT))
 
 #define CRTJMP(pc,sp) __asm__ __volatile__( \

@@ -85,7 +85,6 @@ typedef struct __ucontext {
 #define SA_RESTART    0x10000000
 #define SA_NODEFER    0x40000000
 #define SA_RESETHAND  0x80000000
-#define SA_RESTORER   0x04000000
 
 #undef SIG_BLOCK
 #undef SIG_UNBLOCK

@@ -4,7 +4,7 @@ struct k_sigaction {
 	unsigned flags;
 	void (*handler)(int);
 	unsigned long mask[2];
-	void (*restorer)();
+	void *unused;
 };
 
 hidden void __restore(), __restore_rt();

@@ -38,6 +38,7 @@
 
 #define NEED_MIPS_GOT_RELOCS 1
 #define DT_DEBUG_INDIRECT DT_MIPS_RLD_MAP
+#define DT_DEBUG_INDIRECT_REL DT_MIPS_RLD_MAP_REL
 #define ARCH_SYM_REJECT_UND(s) (!((s)->st_other & STO_MIPS_PLT))
 
 #define CRTJMP(pc,sp) __asm__ __volatile__( \

@@ -1,1 +0,0 @@
-COMPAT_SRC_DIRS = compat/time32

@@ -76,7 +76,6 @@ typedef struct __ucontext
 #define SA_RESTART   0x10000000
 #define SA_NODEFER   0x40000000
 #define SA_RESETHAND 0x80000000
-#define SA_RESTORER  0x04000000
 
 #endif
 

@@ -3,22 +3,22 @@
 
 struct stat32;
 
-int __fxstat64(int ver, int fd, struct stat32 *buf)
+int __fxstat(int ver, int fd, struct stat32 *buf)
 {
 	return __fstat_time32(fd, buf);
 }
 
-int __fxstatat64(int ver, int fd, const char *path, struct stat32 *buf, int flag)
+int __fxstatat(int ver, int fd, const char *path, struct stat32 *buf, int flag)
 {
 	return __fstatat_time32(fd, path, buf, flag);
 }
 
-int __lxstat64(int ver, const char *path, struct stat32 *buf)
+int __lxstat(int ver, const char *path, struct stat32 *buf)
 {
 	return __lstat_time32(path, buf);
 }
 
-int __xstat64(int ver, const char *path, struct stat32 *buf)
+int __xstat(int ver, const char *path, struct stat32 *buf)
 {
 	return __stat_time32(path, buf);
 }

@@ -7,5 +7,3 @@ int __aio_suspend_time32(const struct aiocb *const cbs[], int cnt, const struct
 	return aio_suspend(cbs, cnt, ts32 ? (&(struct timespec){
 		.tv_sec = ts32->tv_sec, .tv_nsec = ts32->tv_nsec}) : 0);
 }
-
-weak_alias(aio_suspend, aio_suspend64);

@@ -13,5 +13,3 @@ int __fstat_time32(int fd, struct stat32 *restrict st32)
 	if (!r) memcpy(st32, &st, offsetof(struct stat, st_atim));
 	return r;
 }
-
-weak_alias(fstat, fstat64);

@@ -13,5 +13,3 @@ int __fstatat_time32(int fd, const char *restrict path, struct stat32 *restrict
 	if (!r) memcpy(st32, &st, offsetof(struct stat, st_atim));
 	return r;
 }
-
-weak_alias(fstatat, fstatat64);

@@ -13,5 +13,3 @@ int __lstat_time32(const char *restrict path, struct stat32 *restrict st32)
 	if (!r) memcpy(st32, &st, offsetof(struct stat, st_atim));
 	return r;
 }
-
-weak_alias(lstat, lstat64);

@@ -13,5 +13,3 @@ int __stat_time32(const char *restrict path, struct stat32 *restrict st32)
 	if (!r) memcpy(st32, &st, offsetof(struct stat, st_atim));
 	return r;
 }
-
-weak_alias(stat, stat64);

@@ -188,6 +188,36 @@ typedef enum __ns_type {
 	ns_t_sink = 40,
 	ns_t_opt = 41,
 	ns_t_apl = 42,
+	ns_t_ds = 43,
+	ns_t_sshfp = 44,
+	ns_t_ipseckey = 45,
+	ns_t_rrsig = 46,
+	ns_t_nsec = 47,
+	ns_t_dnskey = 48,
+	ns_t_dhcid = 49,
+	ns_t_nsec3 = 50,
+	ns_t_nsec3param = 51,
+	ns_t_tlsa = 52,
+	ns_t_smimea = 53,
+	ns_t_hip = 55,
+	ns_t_ninfo = 56,
+	ns_t_rkey = 57,
+	ns_t_talink = 58,
+	ns_t_cds = 59,
+	ns_t_cdnskey = 60,
+	ns_t_openpgpkey = 61,
+	ns_t_csync = 62,
+	ns_t_spf = 99,
+	ns_t_uinfo = 100,
+	ns_t_uid = 101,
+	ns_t_gid = 102,
+	ns_t_unspec = 103,
+	ns_t_nid = 104,
+	ns_t_l32 = 105,
+	ns_t_l64 = 106,
+	ns_t_lp = 107,
+	ns_t_eui48 = 108,
+	ns_t_eui64 = 109,
 	ns_t_tkey = 249,
 	ns_t_tsig = 250,
 	ns_t_ixfr = 251,
@@ -196,6 +226,11 @@ typedef enum __ns_type {
 	ns_t_maila = 254,
 	ns_t_any = 255,
 	ns_t_zxfr = 256,
+	ns_t_uri = 256,
+	ns_t_caa = 257,
+	ns_t_avc = 258,
+	ns_t_ta = 32768,
+	ns_t_dlv = 32769,
 	ns_t_max = 65536
 } ns_type;
 
@@ -430,12 +465,48 @@ typedef struct {
 #define T_NAPTR		ns_t_naptr
 #define T_A6		ns_t_a6
 #define T_DNAME		ns_t_dname
+#define T_DS		ns_t_ds
+#define T_SSHFP		ns_t_sshfp
+#define T_IPSECKEY	ns_t_ipseckey
+#define T_RRSIG		ns_t_rrsig
+#define T_NSEC		ns_t_nsec
+#define T_DNSKEY	ns_t_dnskey
+#define T_DHCID		ns_t_dhcid
+#define T_NSEC3		ns_t_nsec3
+#define T_NSEC3PARAM	ns_t_nsec3param
+#define T_TLSA		ns_t_tlsa
+#define T_SMIMEA	ns_t_smimea
+#define T_HIP		ns_t_hip
+#define T_NINFO		ns_t_ninfo
+#define T_RKEY		ns_t_rkey
+#define T_TALINK	ns_t_talink
+#define T_CDS		ns_t_cds
+#define T_CDNSKEY	ns_t_cdnskey
+#define T_OPENPGPKEY	ns_t_openpgpkey
+#define T_CSYNC		ns_t_csync
+#define T_SPF		ns_t_spf
+#define T_UINFO		ns_t_uinfo
+#define T_UID		ns_t_uid
+#define T_GID		ns_t_gid
+#define T_UNSPEC	ns_t_unspec
+#define T_NID		ns_t_nid
+#define T_L32		ns_t_l32
+#define T_L64		ns_t_l64
+#define T_LP		ns_t_lp
+#define T_EUI48		ns_t_eui48
+#define T_EUI64		ns_t_eui64
+#define T_TKEY		ns_t_tkey
 #define	T_TSIG		ns_t_tsig
 #define	T_IXFR		ns_t_ixfr
 #define T_AXFR		ns_t_axfr
 #define T_MAILB		ns_t_mailb
 #define T_MAILA		ns_t_maila
 #define T_ANY		ns_t_any
+#define T_URI		ns_t_uri
+#define T_CAA		ns_t_caa
+#define T_AVC		ns_t_avc
+#define T_TA		ns_t_ta
+#define T_DLV		ns_t_dlv
 
 #define C_IN		ns_c_in
 #define C_CHAOS		ns_c_chaos

@@ -141,7 +141,7 @@ int mincore (void *, size_t, unsigned char *);
 int shm_open (const char *, int, mode_t);
 int shm_unlink (const char *);
 
-#if defined(_LARGEFILE64_SOURCE) || defined(_GNU_SOURCE)
+#if defined(_LARGEFILE64_SOURCE)
 #define mmap64 mmap
 #define off64_t off_t
 #endif

@@ -95,7 +95,7 @@ int prlimit(pid_t, int, const struct rlimit *, struct rlimit *);
 
 #define RLIM_NLIMITS RLIMIT_NLIMITS
 
-#if defined(_LARGEFILE64_SOURCE) || defined(_GNU_SOURCE)
+#if defined(_LARGEFILE64_SOURCE)
 #define RLIM64_INFINITY RLIM_INFINITY
 #define RLIM64_SAVED_CUR RLIM_SAVED_CUR
 #define RLIM64_SAVED_MAX RLIM_SAVED_MAX

@@ -10,7 +10,7 @@ extern "C" {
 
 ssize_t sendfile(int, int, off_t *, size_t);
 
-#if defined(_LARGEFILE64_SOURCE) || defined(_GNU_SOURCE)
+#if defined(_LARGEFILE64_SOURCE)
 #define sendfile64 sendfile
 #define off64_t off_t
 #endif

@@ -98,7 +98,7 @@ int lchmod(const char *, mode_t);
 #define S_IEXEC S_IXUSR
 #endif
 
-#if defined(_LARGEFILE64_SOURCE) || defined(_GNU_SOURCE)
+#if defined(_LARGEFILE64_SOURCE)
 #define stat64 stat
 #define fstat64 fstat
 #define lstat64 lstat

@@ -18,7 +18,7 @@ typedef struct __fsid_t {
 int statfs (const char *, struct statfs *);
 int fstatfs (int, struct statfs *);
 
-#if defined(_LARGEFILE64_SOURCE) || defined(_GNU_SOURCE)
+#if defined(_LARGEFILE64_SOURCE)
 #define statfs64 statfs
 #define fstatfs64 fstatfs
 #define fsblkcnt64_t fsblkcnt_t

@@ -42,7 +42,7 @@ int fstatvfs (int, struct statvfs *);
 #define ST_NODIRATIME  2048
 #define ST_RELATIME    4096
 
-#if defined(_LARGEFILE64_SOURCE) || defined(_GNU_SOURCE)
+#if defined(_LARGEFILE64_SOURCE)
 #define statvfs64 statvfs
 #define fstatvfs64 fstatvfs
 #define fsblkcnt64_t fsblkcnt_t

@@ -71,7 +71,7 @@ typedef unsigned long long u_quad_t;
 #include <sys/select.h>
 #endif
 
-#if defined(_LARGEFILE64_SOURCE) || defined(_GNU_SOURCE)
+#if defined(_LARGEFILE64_SOURCE)
 #define blkcnt64_t blkcnt_t
 #define fsblkcnt64_t fsblkcnt_t
 #define fsfilcnt64_t fsfilcnt_t

@@ -29,7 +29,7 @@ ssize_t writev (int, const struct iovec *, int);
 #if defined(_GNU_SOURCE) || defined(_BSD_SOURCE)
 ssize_t preadv (int, const struct iovec *, int, off_t);
 ssize_t pwritev (int, const struct iovec *, int, off_t);
-#if defined(_LARGEFILE64_SOURCE) || defined(_GNU_SOURCE)
+#if defined(_LARGEFILE64_SOURCE)
 #define preadv64 preadv
 #define pwritev64 pwritev
 #define off64_t off_t

@@ -50,7 +50,7 @@ pid_t wait4 (pid_t, int *, int, struct rusage *);
 #define WSTOPSIG(s) WEXITSTATUS(s)
 #define WCOREDUMP(s) ((s) & 0x80)
 #define WIFEXITED(s) (!WTERMSIG(s))
-#define WIFSTOPPED(s) ((short)((((s)&0xffff)*0x10001)>>8) > 0x7f00)
+#define WIFSTOPPED(s) ((short)((((s)&0xffff)*0x10001U)>>8) > 0x7f00)
 #define WIFSIGNALED(s) (((s)&0xffff)-1U < 0xffu)
 #define WIFCONTINUED(s) ((s) == 0xffff)
 

@@ -49,7 +49,7 @@ int aio_fsync(int, struct aiocb *);
 
 int lio_listio(int, struct aiocb *__restrict const *__restrict, int, struct sigevent *__restrict);
 
-#if defined(_LARGEFILE64_SOURCE) || defined(_GNU_SOURCE)
+#if defined(_LARGEFILE64_SOURCE)
 #define aiocb64 aiocb
 #define aio_read64 aio_read
 #define aio_write64 aio_write

@@ -56,7 +56,7 @@ int getdents(int, struct dirent *, size_t);
 int versionsort(const struct dirent **, const struct dirent **);
 #endif
 
-#if defined(_LARGEFILE64_SOURCE) || defined(_GNU_SOURCE)
+#if defined(_LARGEFILE64_SOURCE)
 #define dirent64 dirent
 #define readdir64 readdir
 #define readdir64_r readdir_r

@@ -385,7 +385,8 @@ typedef struct {
 #define SHT_PREINIT_ARRAY 16
 #define SHT_GROUP	  17
 #define SHT_SYMTAB_SHNDX  18
-#define	SHT_NUM		  19
+#define SHT_RELR	  19
+#define	SHT_NUM		  20
 #define SHT_LOOS	  0x60000000
 #define SHT_GNU_ATTRIBUTES 0x6ffffff5
 #define SHT_GNU_HASH	  0x6ffffff6
@@ -436,6 +437,7 @@ typedef struct {
 } Elf64_Chdr;
 
 #define ELFCOMPRESS_ZLIB	1
+#define ELFCOMPRESS_ZSTD	2
 #define ELFCOMPRESS_LOOS	0x60000000
 #define ELFCOMPRESS_HIOS	0x6fffffff
 #define ELFCOMPRESS_LOPROC	0x70000000
@@ -754,7 +756,10 @@ typedef struct {
 #define DT_PREINIT_ARRAY 32
 #define DT_PREINIT_ARRAYSZ 33
 #define DT_SYMTAB_SHNDX	34
-#define	DT_NUM		35
+#define DT_RELRSZ	35
+#define DT_RELR		36
+#define DT_RELRENT	37
+#define	DT_NUM		38
 #define DT_LOOS		0x6000000d
 #define DT_HIOS		0x6ffff000
 #define DT_LOPROC	0x70000000

@@ -195,7 +195,7 @@ ssize_t tee(int, int, size_t, unsigned);
 #define loff_t off_t
 #endif
 
-#if defined(_LARGEFILE64_SOURCE) || defined(_GNU_SOURCE)
+#if defined(_LARGEFILE64_SOURCE)
 #define F_GETLK64 F_GETLK
 #define F_SETLK64 F_SETLK
 #define F_SETLKW64 F_SETLKW

@@ -29,7 +29,7 @@ struct FTW {
 int ftw(const char *, int (*)(const char *, const struct stat *, int), int);
 int nftw(const char *, int (*)(const char *, const struct stat *, int, struct FTW *), int, int);
 
-#if defined(_LARGEFILE64_SOURCE) || defined(_GNU_SOURCE)
+#if defined(_LARGEFILE64_SOURCE)
 #define ftw64 ftw
 #define nftw64 nftw
 #endif

@@ -39,7 +39,7 @@ void globfree(glob_t *);
 #define GLOB_NOMATCH 3
 #define GLOB_NOSYS   4
 
-#if defined(_LARGEFILE64_SOURCE) || defined(_GNU_SOURCE)
+#if defined(_LARGEFILE64_SOURCE)
 #define glob64 glob
 #define globfree64 globfree
 #define glob64_t glob_t

@@ -44,6 +44,7 @@ struct addrinfo {
 #define EAI_NONAME     -2
 #define EAI_AGAIN      -3
 #define EAI_FAIL       -4
+#define EAI_NODATA     -5
 #define EAI_FAMILY     -6
 #define EAI_SOCKTYPE   -7
 #define EAI_SERVICE    -8

@@ -124,7 +124,7 @@ __CPU_op_func_S(XOR, ^)
 #define CPU_ALLOC(n) ((cpu_set_t *)calloc(1,CPU_ALLOC_SIZE(n)))
 #define CPU_FREE(set) free(set)
 
-#define CPU_SETSIZE 128
+#define CPU_SETSIZE 1024
 
 #define CPU_SET(i, set) CPU_SET_S(i,sizeof(cpu_set_t),set)
 #define CPU_CLR(i, set) CPU_CLR_S(i,sizeof(cpu_set_t),set)

@@ -205,7 +205,7 @@ typedef struct _IO_cookie_io_functions_t {
 FILE *fopencookie(void *, const char *, cookie_io_functions_t);
 #endif
 
-#if defined(_LARGEFILE64_SOURCE) || defined(_GNU_SOURCE)
+#if defined(_LARGEFILE64_SOURCE)
 #define tmpfile64 tmpfile
 #define fopen64 fopen
 #define freopen64 freopen

@@ -95,7 +95,7 @@ size_t __ctype_get_mb_cur_max(void);
 #define WTERMSIG(s) ((s) & 0x7f)
 #define WSTOPSIG(s) WEXITSTATUS(s)
 #define WIFEXITED(s) (!WTERMSIG(s))
-#define WIFSTOPPED(s) ((short)((((s)&0xffff)*0x10001)>>8) > 0x7f00)
+#define WIFSTOPPED(s) ((short)((((s)&0xffff)*0x10001U)>>8) > 0x7f00)
 #define WIFSIGNALED(s) (((s)&0xffff)-1U < 0xffu)
 
 int posix_memalign (void **, size_t, size_t);
@@ -163,7 +163,7 @@ double strtod_l(const char *__restrict, char **__restrict, struct __locale_struc
 long double strtold_l(const char *__restrict, char **__restrict, struct __locale_struct *);
 #endif
 
-#if defined(_LARGEFILE64_SOURCE) || defined(_GNU_SOURCE)
+#if defined(_LARGEFILE64_SOURCE)
 #define mkstemp64 mkstemp
 #define mkostemp64 mkostemp
 #if defined(_GNU_SOURCE) || defined(_BSD_SOURCE)

@@ -73,6 +73,7 @@ char *strsignal(int);
 char *strerror_l (int, locale_t);
 int strcoll_l (const char *, const char *, locale_t);
 size_t strxfrm_l (char *__restrict, const char *__restrict, size_t, locale_t);
+void *memmem(const void *, size_t, const void *, size_t);
 #endif
 
 #if defined(_XOPEN_SOURCE) || defined(_GNU_SOURCE) \
@@ -92,7 +93,6 @@ void explicit_bzero (void *, size_t);
 int strverscmp (const char *, const char *);
 char *strchrnul(const char *, int);
 char *strcasestr(const char *, const char *);
-void *memmem(const void *, size_t, const void *, size_t);
 void *memrchr(const void *, int, size_t);
 void *mempcpy(void *, const void *, size_t);
 #ifndef __cplusplus

@@ -5,6 +5,7 @@
 extern "C" {
 #endif
 
+#include <features.h>
 
 #define __NEED_size_t
 #define __NEED_locale_t

@@ -198,7 +198,7 @@ ssize_t copy_file_range(int, off_t *, int, off_t *, size_t, unsigned);
 pid_t gettid(void);
 #endif
 
-#if defined(_LARGEFILE64_SOURCE) || defined(_GNU_SOURCE)
+#if defined(_LARGEFILE64_SOURCE)
 #define lseek64 lseek
 #define pread64 pread
 #define pwrite64 pwrite
@@ -425,6 +425,8 @@ pid_t gettid(void);
 #define _SC_XOPEN_STREAMS	246
 #define _SC_THREAD_ROBUST_PRIO_INHERIT	247
 #define _SC_THREAD_ROBUST_PRIO_PROTECT	248
+#define _SC_MINSIGSTKSZ	249
+#define _SC_SIGSTKSZ	250
 
 #define _CS_PATH	0
 #define _CS_POSIX_V6_WIDTH_RESTRICTED_ENVS	1
@@ -467,6 +469,8 @@ pid_t gettid(void);
 #define _CS_POSIX_V7_LPBIG_OFFBIG_LINTFLAGS	1147
 #define _CS_V6_ENV	1148
 #define _CS_V7_ENV	1149
+#define _CS_POSIX_V7_THREADS_CFLAGS	1150
+#define _CS_POSIX_V7_THREADS_LDFLAGS	1151
 
 #ifdef __cplusplus
 }

@@ -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);

@@ -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

@@ -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);

@@ -7,7 +7,7 @@ size_t confstr(int name, char *buf, size_t len)
 	const char *s = "";
 	if (!name) {
 		s = "/bin:/usr/bin";
-	} else if ((name&~4U)!=1 && name-_CS_POSIX_V6_ILP32_OFF32_CFLAGS>33U) {
+	} else if ((name&~4U)!=1 && name-_CS_POSIX_V6_ILP32_OFF32_CFLAGS>35U) {
 		errno = EINVAL;
 		return 0;
 	}

@@ -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;
 	}

@@ -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);

@@ -25,5 +25,3 @@ struct dirent *readdir(DIR *dir)
 	dir->tell = de->d_off;
 	return de;
 }
-
-weak_alias(readdir, readdir64);

@@ -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);

@@ -43,5 +43,3 @@ int scandir(const char *path, struct dirent ***res,
 	*res = names;
 	return cnt;
 }
-
-weak_alias(scandir, scandir64);

@@ -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);

@@ -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);

@@ -19,5 +19,3 @@ int open(const char *filename, int flags, ...)
 
 	return __syscall_ret(fd);
 }
-
-weak_alias(open, open64);

@@ -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);

@@ -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);

@@ -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);

@@ -0,0 +1,9 @@
+#ifndef SYS_STAT_H
+#define SYS_STAT_H
+
+#include "../../../include/sys/stat.h"
+
+hidden int __fstat(int, struct stat *);
+hidden int __fstatat(int, const char *restrict, struct stat *restrict, int);
+
+#endif

@@ -92,8 +92,12 @@ 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 *);
 

@@ -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 +16,4 @@ extern hidden volatile int *const __vmlock_lockptr;
 
 hidden void __malloc_atfork(int);
 hidden void __ldso_atfork(int);
+hidden void __pthread_key_atfork(int);

@@ -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();

@@ -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 */
 
@@ -201,43 +201,43 @@ static inline long __alt_socketcall(int sys, int sock, int cp, long a, long b, l
 #define SYS_sendfile SYS_sendfile64
 #endif
 
-#ifndef SYS_timer_settime
+#ifdef SYS_timer_settime32
 #define SYS_timer_settime SYS_timer_settime32
 #endif
 
-#ifndef SYS_timer_gettime
+#ifdef SYS_timer_gettime32
 #define SYS_timer_gettime SYS_timer_gettime32
 #endif
 
-#ifndef SYS_timerfd_settime
+#ifdef SYS_timerfd_settime32
 #define SYS_timerfd_settime SYS_timerfd_settime32
 #endif
 
-#ifndef SYS_timerfd_gettime
+#ifdef SYS_timerfd_gettime32
 #define SYS_timerfd_gettime SYS_timerfd_gettime32
 #endif
 
-#ifndef SYS_clock_settime
+#ifdef SYS_clock_settime32
 #define SYS_clock_settime SYS_clock_settime32
 #endif
 
-#ifndef SYS_clock_gettime
+#ifdef SYS_clock_gettime32
 #define SYS_clock_gettime SYS_clock_gettime32
 #endif
 
-#ifndef SYS_clock_getres
+#ifdef SYS_clock_getres_time32
 #define SYS_clock_getres SYS_clock_getres_time32
 #endif
 
-#ifndef SYS_clock_nanosleep
+#ifdef SYS_clock_nanosleep_time32
 #define SYS_clock_nanosleep SYS_clock_nanosleep_time32
 #endif
 
-#ifndef SYS_gettimeofday
+#ifdef SYS_gettimeofday_time32
 #define SYS_gettimeofday SYS_gettimeofday_time32
 #endif
 
-#ifndef SYS_settimeofday
+#ifdef SYS_settimeofday_time32
 #define SYS_settimeofday SYS_settimeofday_time32
 #endif
 

@@ -1,1 +1,1 @@
-#define VERSION "1.2.3"
+#define VERSION "1.2.4"

@@ -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

@@ -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;

@@ -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);

@@ -5,6 +5,7 @@
 
 int epoll_create(int size)
 {
+	if (size<=0) return __syscall_ret(-EINVAL);
 	return epoll_create1(0);
 }
 

@@ -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);

@@ -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);

@@ -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);

@@ -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);

@@ -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);

@@ -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)

@@ -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;

@@ -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

@@ -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);

@@ -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) {

@@ -6,12 +6,13 @@
 
 int getrlimit(int resource, struct rlimit *rlim)
 {
-	unsigned long k_rlim[2];
 	int ret = syscall(SYS_prlimit64, 0, resource, 0, rlim);
 	if (!ret) {
 		FIX(rlim->rlim_cur);
 		FIX(rlim->rlim_max);
 	}
+#ifdef SYS_getrlimit
+	unsigned long k_rlim[2];
 	if (!ret || errno != ENOSYS)
 		return ret;
 	if (syscall(SYS_getrlimit, resource, k_rlim) < 0)
@@ -21,6 +22,7 @@ int getrlimit(int resource, struct rlimit *rlim)
 	FIX(rlim->rlim_cur);
 	FIX(rlim->rlim_max);
 	return 0;
+#else
+	return ret;
+#endif
 }
-
-weak_alias(getrlimit, getrlimit64);

@@ -28,5 +28,3 @@ int lockf(int fd, int op, off_t size)
 	errno = EINVAL;
 	return -1;
 }
-
-weak_alias(lockf, lockf64);

@@ -2,6 +2,7 @@
 #include <string.h>
 #include <mntent.h>
 #include <errno.h>
+#include <limits.h>
 
 static char *internal_buf;
 static size_t internal_bufsize;
@@ -21,7 +22,8 @@ int endmntent(FILE *f)
 
 struct mntent *getmntent_r(FILE *f, struct mntent *mnt, char *linebuf, int buflen)
 {
-	int cnt, n[8], use_internal = (linebuf == SENTINEL);
+	int n[8], use_internal = (linebuf == SENTINEL);
+	size_t len, i;
 
 	mnt->mnt_freq = 0;
 	mnt->mnt_passno = 0;
@@ -39,10 +41,14 @@ struct mntent *getmntent_r(FILE *f, struct mntent *mnt, char *linebuf, int bufle
 			errno = ERANGE;
 			return 0;
 		}
-		cnt = sscanf(linebuf, " %n%*s%n %n%*s%n %n%*s%n %n%*s%n %d %d",
+
+		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",
 			n, n+1, n+2, n+3, n+4, n+5, n+6, n+7,
 			&mnt->mnt_freq, &mnt->mnt_passno);
-	} while (cnt < 2 || linebuf[n[0]] == '#');
+	} while (linebuf[n[0]] == '#' || n[1]==len);
 
 	linebuf[n[1]] = 0;
 	linebuf[n[3]] = 0;

@@ -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);

@@ -12,12 +12,14 @@ struct ctx {
 	int err;
 };
 
+#ifdef SYS_setrlimit
 static void do_setrlimit(void *p)
 {
 	struct ctx *c = p;
 	if (c->err>0) return;
 	c->err = -__syscall(SYS_setrlimit, c->res, c->lim);
 }
+#endif
 
 int setrlimit(int resource, const struct rlimit *rlim)
 {
@@ -29,6 +31,7 @@ int setrlimit(int resource, const struct rlimit *rlim)
 		rlim = &tmp;
 	}
 	int ret = __syscall(SYS_prlimit64, 0, resource, rlim, 0);
+#ifdef SYS_setrlimit
 	if (ret != -ENOSYS) return __syscall_ret(ret);
 
 	struct ctx c = {
@@ -42,6 +45,7 @@ int setrlimit(int resource, const struct rlimit *rlim)
 		return -1;
 	}
 	return 0;
+#else
+	return __syscall_ret(ret);
+#endif
 }
-
-weak_alias(setrlimit, setrlimit64);

@@ -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);

@@ -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;
 }

@@ -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)

@@ -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;
@@ -15,18 +15,18 @@ int __dns_parse(const unsigned char *r, int rlen, int (*callback)(void *, int, c
 	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 || *p>193 || (*p==193 && p[1]>254))
 			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 || *p>193 || (*p==193 && p[1]>254))
 			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;

@@ -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"

@@ -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;
@@ -66,9 +67,11 @@ int getaddrinfo(const char *restrict host, const char *restrict serv, const stru
 				pthread_setcancelstate(
 					PTHREAD_CANCEL_DISABLE, &cs);
 				int r = connect(s, ta[i], tl[i]);
+				int saved_errno = errno;
 				pthread_setcancelstate(cs, 0);
 				close(s);
 				if (!r) continue;
+				errno = saved_errno;
 			}
 			switch (errno) {
 			case EADDRNOTAVAIL:
@@ -80,7 +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];
 		}
 	}
@@ -91,6 +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);

@@ -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;
 }

@@ -54,9 +54,10 @@ 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:

@@ -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;
 }

@@ -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;

@@ -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;
 }

@@ -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;

@@ -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:

@@ -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;

@@ -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

@@ -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;

@@ -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 768
+
+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] = {
@@ -153,8 +158,12 @@ 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])
+				qbuf[nq][0]++;
 			nq++;
 		}
 	}
@@ -168,11 +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)

@@ -86,7 +86,7 @@ struct ifaddrmsg {
 #define RTA_DATALEN(rta)	((rta)->rta_len-sizeof(struct rtattr))
 #define RTA_DATAEND(rta)	((char*)(rta)+(rta)->rta_len)
 #define RTA_NEXT(rta)		(struct rtattr*)((char*)(rta)+NETLINK_ALIGN((rta)->rta_len))
-#define RTA_OK(nlh,end)		((char*)(end)-(char*)(rta) >= sizeof(struct rtattr))
+#define RTA_OK(rta,end)		((char*)(end)-(char*)(rta) >= sizeof(struct rtattr))
 
 #define NLMSG_RTA(nlh,len)	((void*)((char*)(nlh)+sizeof(struct nlmsghdr)+NETLINK_ALIGN(len)))
 #define NLMSG_RTAOK(rta,nlh)	RTA_OK(rta,NLMSG_DATAEND(nlh))

@@ -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;

@@ -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;
 }
 

@@ -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;
 }
 

@@ -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;
 			}

@@ -0,0 +1,9 @@
+.global vfork
+.type vfork,%function
+vfork:
+	mov x8, 220    // SYS_clone
+	mov x0, 0x4111 // SIGCHLD | CLONE_VM | CLONE_VFORK
+	mov x1, 0
+	svc 0
+	.hidden __syscall_ret
+	b __syscall_ret

@@ -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

@@ -14,7 +14,6 @@ pid_t _Fork(void)
 	pid_t ret;
 	sigset_t set;
 	__block_all_sigs(&set);
-	__aio_atfork(-1);
 	LOCK(__abort_lock);
 #ifdef SYS_fork
 	ret = __syscall(SYS_fork);
@@ -32,7 +31,7 @@ pid_t _Fork(void)
 		if (libc.need_locks) libc.need_locks = -1;
 	}
 	UNLOCK(__abort_lock);
-	__aio_atfork(!ret);
+	if (!ret) __aio_atfork(1);
 	__restore_sigs(&set);
 	return __syscall_ret(ret);
 }

@@ -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);