Merge tag 'bsd-user-preen-2022q2-pull-request' of ssh://github.com/qemu-bsd-user/qemu-bsd-user into staging

bsd-user upstreaming: read, write and exit

This series of patches continues the effort to get system calls working
upstream. This series was cleaved off a prior series to give me time to rework
based on the feedback from the first time I posted these. read, write and exit
are implemented, along with a few helper functions and tracing.

# -----BEGIN PGP SIGNATURE-----
# Comment: GPGTools - https://gpgtools.org
#
# iQIzBAABCgAdFiEEIDX4lLAKo898zeG3bBzRKH2wEQAFAmKknqsACgkQbBzRKH2w
# EQDJFhAA3ARnphfxKUvKHzi6ARza8YRFNJs/X48l78xugW8Fx2+UFxn/oPzVYuM3
# TUIIRNIXQxYDWP5rNN4zNfYjn1F32qx4PrLzzHw4t/0pNvNDvbVKyw+OB1pNUVI9
# 6hNH7jTLczi4OwZFPC4bJdFFHI55t4PCHxPaMQQO76lEsFzz3+9OcNNj3/j5aTBH
# kY3FsmOIKFEFuqDANjdn13lBlkNNWj5WMIsEQd+k1TPTfG3EWDlfHVJVezx9j5uH
# KMUjRRqBANLVJuqhog7sVAdkR6o/aKuq16nPVAgkLeJrsjwljoO1bGL3DTFWzhAo
# v3+S3a5gCReLXA/Z+dGCuJysmimsm+XdJWHyNIwDCV/UUg9tXc9/BI/eBSK7MLbz
# hvURY3cRSSVY+xiu5ts1OCp1bbqMaYbpIrv1mLZMA/32Y1KNv0LwSI0ClGag0MJ1
# rSVC3i/TGYn8Dk7TlfKkrKn751ZCCDsxZGROEcPlTV84Mc7080lpgYZKlgLCPJNT
# WWaVSBs9YhHhmoG3R3GurfsHnnRr8Y8iZkmndoCJdeSlmZh/1oD31PWKhnydsPa9
# jQCENkAM0dOM/r+hV+paFxOPkgklu7KLtjqGnVRxgB4j37jA9KDyw+UXLmjipP0K
# IwPlsNIfk1LujID3o5nJVgf/K+TJpXS7JDcUsuB9+FvlkeEMWm8=
# =nOc2
# -----END PGP SIGNATURE-----
# gpg: Signature made Sat 11 Jun 2022 06:54:51 AM PDT
# gpg:                using RSA key 2035F894B00AA3CF7CCDE1B76C1CD1287DB01100
# gpg: Good signature from "Warner Losh <wlosh@netflix.com>" [unknown]
# gpg:                 aka "Warner Losh <imp@bsdimp.com>" [unknown]
# gpg:                 aka "Warner Losh <imp@freebsd.org>" [unknown]
# gpg:                 aka "Warner Losh <imp@village.org>" [unknown]
# gpg:                 aka "Warner Losh <wlosh@bsdimp.com>" [unknown]
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg:          There is no indication that the signature belongs to the owner.
# Primary key fingerprint: 2035 F894 B00A A3CF 7CCD  E1B7 6C1C D128 7DB0 1100

* tag 'bsd-user-preen-2022q2-pull-request' of ssh://github.com/qemu-bsd-user/qemu-bsd-user:
  bsd-user/freebsd/os-syscall.c: Implement exit
  bsd-user/bsd-file.h: Meat of the write system calls
  bsd-user/bsd-file.h: Add implementations for read, pread, readv and preadv
  bsd-user/freebsd/os-syscall.c: Tracing and error boilerplate
  bsd-user/freebsd/os-syscall.c: unlock_iovec
  bsd-user/freebsd/os-syscall.c: lock_iovec

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

1 files changed, 213 insertions, 4 deletions

diff --git a/bsd-user/freebsd/os-syscall.c b/bsd-user/freebsd/os-syscall.c
index d272478e7b..71aa0d38e0 100644
--- a/bsd-user/freebsd/os-syscall.c
+++ b/bsd-user/freebsd/os-syscall.c
@@ -41,6 +41,22 @@
 #include "user/syscall-trace.h"
 
 #include "bsd-file.h"
+#include "bsd-proc.h"
+
+/* I/O */
+safe_syscall3(ssize_t, read, int, fd, void *, buf, size_t, nbytes);
+safe_syscall4(ssize_t, pread, int, fd, void *, buf, size_t, nbytes, off_t,
+    offset);
+safe_syscall3(ssize_t, readv, int, fd, const struct iovec *, iov, int, iovcnt);
+safe_syscall4(ssize_t, preadv, int, fd, const struct iovec *, iov, int, iovcnt,
+    off_t, offset);
+
+safe_syscall3(ssize_t, write, int, fd, void *, buf, size_t, nbytes);
+safe_syscall4(ssize_t, pwrite, int, fd, void *, buf, size_t, nbytes, off_t,
+    offset);
+safe_syscall3(ssize_t, writev, int, fd, const struct iovec *, iov, int, iovcnt);
+safe_syscall4(ssize_t, pwritev, int, fd, const struct iovec *, iov, int, iovcnt,
+    off_t, offset);
 
 void target_set_brk(abi_ulong new_brk)
 {
@@ -74,16 +90,209 @@ bool is_error(abi_long ret)
 }
 
 /*
- * do_syscall() should always have a single exit point at the end so that
- * actions, such as logging of syscall results, can be performed.  All errnos
- * that do_syscall() returns must be -TARGET_<errcode>.
+ * Unlocks a iovec. Unlike unlock_iovec, it assumes the tvec array itself is
+ * already locked from target_addr. It will be unlocked as well as all the iovec
+ * elements.
+ */
+static void helper_unlock_iovec(struct target_iovec *target_vec,
+                                abi_ulong target_addr, struct iovec *vec,
+                                int count, int copy)
+{
+    for (int i = 0; i < count; i++) {
+        abi_ulong base = tswapal(target_vec[i].iov_base);
+
+        if (vec[i].iov_base) {
+            unlock_user(vec[i].iov_base, base, copy ? vec[i].iov_len : 0);
+        }
+    }
+    unlock_user(target_vec, target_addr, 0);
+}
+
+struct iovec *lock_iovec(int type, abi_ulong target_addr,
+        int count, int copy)
+{
+    struct target_iovec *target_vec;
+    struct iovec *vec;
+    abi_ulong total_len, max_len;
+    int i;
+    int err = 0;
+
+    if (count == 0) {
+        errno = 0;
+        return NULL;
+    }
+    if (count < 0 || count > IOV_MAX) {
+        errno = EINVAL;
+        return NULL;
+    }
+
+    vec = g_try_new0(struct iovec, count);
+    if (vec == NULL) {
+        errno = ENOMEM;
+        return NULL;
+    }
+
+    target_vec = lock_user(VERIFY_READ, target_addr,
+                           count * sizeof(struct target_iovec), 1);
+    if (target_vec == NULL) {
+        err = EFAULT;
+        goto fail2;
+    }
+
+    max_len = 0x7fffffff & MIN(TARGET_PAGE_MASK, PAGE_MASK);
+    total_len = 0;
+
+    for (i = 0; i < count; i++) {
+        abi_ulong base = tswapal(target_vec[i].iov_base);
+        abi_long len = tswapal(target_vec[i].iov_len);
+
+        if (len < 0) {
+            err = EINVAL;
+            goto fail;
+        } else if (len == 0) {
+            /* Zero length pointer is ignored. */
+            vec[i].iov_base = 0;
+        } else {
+            vec[i].iov_base = lock_user(type, base, len, copy);
+            /*
+             * If the first buffer pointer is bad, this is a fault.  But
+             * subsequent bad buffers will result in a partial write; this is
+             * realized by filling the vector with null pointers and zero
+             * lengths.
+             */
+            if (!vec[i].iov_base) {
+                if (i == 0) {
+                    err = EFAULT;
+                    goto fail;
+                } else {
+                    /*
+                     * Fail all the subsequent addresses, they are already
+                     * zero'd.
+                     */
+                    goto out;
+                }
+            }
+            if (len > max_len - total_len) {
+                len = max_len - total_len;
+            }
+        }
+        vec[i].iov_len = len;
+        total_len += len;
+    }
+out:
+    unlock_user(target_vec, target_addr, 0);
+    return vec;
+
+fail:
+    helper_unlock_iovec(target_vec, target_addr, vec, i, copy);
+fail2:
+    g_free(vec);
+    errno = err;
+    return NULL;
+}
+
+void unlock_iovec(struct iovec *vec, abi_ulong target_addr,
+        int count, int copy)
+{
+    struct target_iovec *target_vec;
+
+    target_vec = lock_user(VERIFY_READ, target_addr,
+                           count * sizeof(struct target_iovec), 1);
+    if (target_vec) {
+        helper_unlock_iovec(target_vec, target_addr, vec, count, copy);
+    }
+
+    g_free(vec);
+}
+
+/*
+ * All errnos that freebsd_syscall() returns must be -TARGET_<errcode>.
+ */
+static abi_long freebsd_syscall(void *cpu_env, int num, abi_long arg1,
+                                abi_long arg2, abi_long arg3, abi_long arg4,
+                                abi_long arg5, abi_long arg6, abi_long arg7,
+                                abi_long arg8)
+{
+    abi_long ret;
+
+    switch (num) {
+        /*
+         * process system calls
+         */
+    case TARGET_FREEBSD_NR_exit: /* exit(2) */
+        ret = do_bsd_exit(cpu_env, arg1);
+        break;
+
+        /*
+         * File system calls.
+         */
+    case TARGET_FREEBSD_NR_read: /* read(2) */
+        ret = do_bsd_read(arg1, arg2, arg3);
+        break;
+
+    case TARGET_FREEBSD_NR_pread: /* pread(2) */
+        ret = do_bsd_pread(cpu_env, arg1, arg2, arg3, arg4, arg5, arg6);
+        break;
+
+    case TARGET_FREEBSD_NR_readv: /* readv(2) */
+        ret = do_bsd_readv(arg1, arg2, arg3);
+        break;
+
+    case TARGET_FREEBSD_NR_preadv: /* preadv(2) */
+        ret = do_bsd_preadv(cpu_env, arg1, arg2, arg3, arg4, arg5, arg6);
+
+    case TARGET_FREEBSD_NR_write: /* write(2) */
+        ret = do_bsd_write(arg1, arg2, arg3);
+        break;
+
+    case TARGET_FREEBSD_NR_pwrite: /* pwrite(2) */
+        ret = do_bsd_pwrite(cpu_env, arg1, arg2, arg3, arg4, arg5, arg6);
+        break;
+
+    case TARGET_FREEBSD_NR_writev: /* writev(2) */
+        ret = do_bsd_writev(arg1, arg2, arg3);
+        break;
+
+    case TARGET_FREEBSD_NR_pwritev: /* pwritev(2) */
+        ret = do_bsd_pwritev(cpu_env, arg1, arg2, arg3, arg4, arg5, arg6);
+        break;
+
+    default:
+        qemu_log_mask(LOG_UNIMP, "Unsupported syscall: %d\n", num);
+        ret = -TARGET_ENOSYS;
+        break;
+    }
+
+    return ret;
+}
+
+/*
+ * do_freebsd_syscall() should always have a single exit point at the end so
+ * that actions, such as logging of syscall results, can be performed. This
+ * as a wrapper around freebsd_syscall() so that actually happens. Since
+ * that is a singleton, modern compilers will inline it anyway...
  */
 abi_long do_freebsd_syscall(void *cpu_env, int num, abi_long arg1,
                             abi_long arg2, abi_long arg3, abi_long arg4,
                             abi_long arg5, abi_long arg6, abi_long arg7,
                             abi_long arg8)
 {
-    return 0;
+    CPUState *cpu = env_cpu(cpu_env);
+    int ret;
+
+    trace_guest_user_syscall(cpu, num, arg1, arg2, arg3, arg4, arg5, arg6, arg7, arg8);
+    if (do_strace) {
+        print_freebsd_syscall(num, arg1, arg2, arg3, arg4, arg5, arg6);
+    }
+
+    ret = freebsd_syscall(cpu_env, num, arg1, arg2, arg3, arg4, arg5, arg6,
+                          arg7, arg8);
+    if (do_strace) {
+        print_freebsd_syscall_ret(num, ret);
+    }
+    trace_guest_user_syscall_ret(cpu, num, ret);
+
+    return ret;
 }
 
 void syscall_init(void)