[Qemu-devel] [PATCH v2 00/16] linux-user fixes & improvements

[Qemu-devel] [PATCH v2 00/16] linux-user fixes & improvements

[Paul Burton]

From: Paul Burton <paul.burton@imgtec.com>

This series fixes a number of bugs in QEMUs linux-user support, some
specific to targetting the MIPS architecture but mostly generic. It also
adds support for some previously unsupported syscalls & {g,s}etsockopt
options.

Paul Burton (16):
  linux-user: translate the result of getsockopt SO_TYPE
  linux-user: support SO_ACCEPTCONN getsockopt option
  linux-user: support SO_{SND,RCV}BUFFORCE setsockopt options
  linux-user: support SO_PASSSEC setsockopt option
  linux-user: allow NULL arguments to mount
  linux-user: support strace of epoll_create1
  linux-user: fix struct target_epoll_event layout for MIPS
  linux-user: respect timezone for settimeofday
  linux-user: allow NULL tv argument for settimeofday
  linux-user: support timerfd_{create,gettime,settime} syscalls
  linux-user: support ioprio_{get,set} syscalls
  linux-user: support {name_to,open_by}_handle_at syscalls
  linux-user: support the setns syscall
  linux-user: support the unshare syscall
  linux-user: support the KDSIGACCEPT ioctl
  linux-user: support the SIOCGIFINDEX ioctl

 linux-user/ioctls.h       |   2 +
 linux-user/socket.h       |   5 +
 linux-user/strace.c       |  30 +++++
 linux-user/strace.list    |  21 ++++
 linux-user/syscall.c      | 271 +++++++++++++++++++++++++++++++++++++++++-----
 linux-user/syscall_defs.h |   9 +-
 6 files changed, 311 insertions(+), 27 deletions(-)

-- 
2.0.0

[Qemu-devel] [PATCH v2 01/16] linux-user: translate the result of getsockopt SO_TYPE

[Paul Burton]

QEMU previously passed the result of the host syscall directly to the
target program. This is a problem if the host & target have different
representations of socket types, as is the case when running a MIPS
target program on an x86 host. Introduce a host_to_target_sock_type
helper function mirroring the existing target_to_host_sock_type, and
call it to translate the value provided by getsockopt when called for
the SO_TYPE option.

Signed-off-by: Paul Burton <paul@archlinuxmips.org>
---
Changes in v2:
  - Remove indirection via a function pointer, and just call
    host_to_target_sock_type directly.
---
 linux-user/syscall.c | 31 +++++++++++++++++++++++++++++++
 1 file changed, 31 insertions(+)

diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index 7d74079..30610e7 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -592,6 +592,35 @@ char *target_strerror(int err)
     return strerror(target_to_host_errno(err));
 }
 
+static inline int host_to_target_sock_type(int host_type)
+{
+    int target_type;
+
+    switch (host_type & 0xf /* SOCK_TYPE_MASK */) {
+    case SOCK_DGRAM:
+        target_type = TARGET_SOCK_DGRAM;
+        break;
+    case SOCK_STREAM:
+        target_type = TARGET_SOCK_STREAM;
+        break;
+    default:
+        target_type = host_type & 0xf /* SOCK_TYPE_MASK */;
+        break;
+    }
+
+#if defined(SOCK_CLOEXEC)
+    if (host_type & SOCK_CLOEXEC)
+        target_type |= TARGET_SOCK_CLOEXEC;
+#endif
+
+#if defined(SOCK_NONBLOCK)
+    if (host_type & SOCK_NONBLOCK)
+        target_type |= TARGET_SOCK_NONBLOCK;
+#endif
+
+    return target_type;
+}
+
 static abi_ulong target_brk;
 static abi_ulong target_original_brk;
 static abi_ulong brk_page;
@@ -1636,6 +1665,8 @@ static abi_long do_getsockopt(int sockfd, int level, int optname,
         ret = get_errno(getsockopt(sockfd, level, optname, &val, &lv));
         if (ret < 0)
             return ret;
+        if (optname == SO_TYPE)
+            val = host_to_target_sock_type(val);
         if (len > lv)
             len = lv;
         if (len == 4) {
-- 
2.0.0

[Qemu-devel] [PATCH v2 02/16] linux-user: support SO_ACCEPTCONN getsockopt option

[Paul Burton]

Translate the SO_ACCEPTCONN option to the host value & execute the
syscall as expected.

Signed-off-by: Paul Burton <paul@archlinuxmips.org>
---
Changes in v2:
  - None.
---
 linux-user/syscall.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index 30610e7..dba5b4d 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -1650,6 +1650,9 @@ static abi_long do_getsockopt(int sockfd, int level, int optname,
         case TARGET_SO_RCVLOWAT:
             optname = SO_RCVLOWAT;
             goto int_case;
+        case TARGET_SO_ACCEPTCONN:
+            optname = SO_ACCEPTCONN;
+            goto int_case;
         default:
             goto int_case;
         }
-- 
2.0.0

[Qemu-devel] [PATCH v2 03/16] linux-user: support SO_{SND, RCV}BUFFORCE setsockopt options

[Paul Burton]

Translate the SO_SNDBUFFORCE & SO_RCVBUFFORCE options to setsockopt to
the host values & perform the syscall as expected, allowing use of those
options by target programs.

Signed-off-by: Paul Burton <paul@archlinuxmips.org>
---
Changes in v2:
  - None.
---
 linux-user/syscall.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index dba5b4d..ce1ea1f 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -1500,9 +1500,15 @@ set_timeout:
         case TARGET_SO_SNDBUF:
 		optname = SO_SNDBUF;
 		break;
+        case TARGET_SO_SNDBUFFORCE:
+		optname = SO_SNDBUFFORCE;
+		break;
         case TARGET_SO_RCVBUF:
 		optname = SO_RCVBUF;
 		break;
+        case TARGET_SO_RCVBUFFORCE:
+		optname = SO_RCVBUFFORCE;
+		break;
         case TARGET_SO_KEEPALIVE:
 		optname = SO_KEEPALIVE;
 		break;
-- 
2.0.0

[Qemu-devel] [PATCH v2 04/16] linux-user: support SO_PASSSEC setsockopt option

[Paul Burton]

Translate the SO_PASSSEC option to setsockopt to the host value &
perform the syscall as expected, allowing use of the option by target
programs.

Signed-off-by: Paul Burton <paul@archlinuxmips.org>
---
Changes in v2:
  - Fix the value of TARGET_SO_PASSSEC for the sparc target.
---
 linux-user/socket.h  | 5 +++++
 linux-user/syscall.c | 3 +++
 2 files changed, 8 insertions(+)

diff --git a/linux-user/socket.h b/linux-user/socket.h
index ae17959..4dacae6 100644
--- a/linux-user/socket.h
+++ b/linux-user/socket.h
@@ -63,6 +63,7 @@
     #define TARGET_SO_PEERSEC              30
     #define TARGET_SO_SNDBUFFORCE          31
     #define TARGET_SO_RCVBUFFORCE          33
+    #define TARGET_SO_PASSSEC              34
 
     /** sock_type - Socket types
      *
@@ -242,6 +243,10 @@
 
     #define TARGET_SOCK_MAX (TARGET_SOCK_PACKET + 1)
     #define TARGET_SOCK_TYPE_MASK    0xf  /* Covers up to TARGET_SOCK_MAX-1. */
+
+    #define TARGET_SO_PASSSEC        31
+#else
+    #define TARGET_SO_PASSSEC        34
 #endif
 
     /* For setsockopt(2) */
diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index ce1ea1f..7da21c7 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -1529,6 +1529,9 @@ set_timeout:
         case TARGET_SO_PASSCRED:
 		optname = SO_PASSCRED;
 		break;
+        case TARGET_SO_PASSSEC:
+		optname = SO_PASSSEC;
+		break;
         case TARGET_SO_TIMESTAMP:
 		optname = SO_TIMESTAMP;
 		break;
-- 
2.0.0

[Qemu-devel] [PATCH v2 05/16] linux-user: allow NULL arguments to mount

[Paul Burton]

Calls to the mount syscall can legitimately provide NULL as the value
for the source of filesystemtype arguments, which QEMU would previously
reject & return -EFAULT to the target program. An example of this is
remounting an already mounted filesystem with different properties.

Instead of rejecting such syscalls with -EFAULT, pass NULL along to the
kernel as the target program expects.

Additionally this patch fixes a potential memory leak when DEBUG_REMAP
is enabled and lock_user_string fails on the target or filesystemtype
arguments but a prior argument was non-NULL and already locked.

Since the patch already touched most lines of the TARGET_NR_mount case,
it fixes the indentation for good measure.

Signed-off-by: Paul Burton <paul@archlinuxmips.org>
---
Changes in v2:
  - None.
---
 linux-user/syscall.c | 68 +++++++++++++++++++++++++++++++++++-----------------
 1 file changed, 46 insertions(+), 22 deletions(-)

diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index 7da21c7..3861f0f 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -5563,29 +5563,53 @@ abi_long do_syscall(void *cpu_env, int num, abi_long arg1,
         break;
 #endif
     case TARGET_NR_mount:
-		{
-			/* need to look at the data field */
-			void *p2, *p3;
-			p = lock_user_string(arg1);
-			p2 = lock_user_string(arg2);
-			p3 = lock_user_string(arg3);
-                        if (!p || !p2 || !p3)
-                            ret = -TARGET_EFAULT;
-                        else {
-                            /* FIXME - arg5 should be locked, but it isn't clear how to
-                             * do that since it's not guaranteed to be a NULL-terminated
-                             * string.
-                             */
-                            if ( ! arg5 )
-                                ret = get_errno(mount(p, p2, p3, (unsigned long)arg4, NULL));
-                            else
-                                ret = get_errno(mount(p, p2, p3, (unsigned long)arg4, g2h(arg5)));
-                        }
+        {
+            /* need to look at the data field */
+            void *p2, *p3;
+
+            if (arg1) {
+                p = lock_user_string(arg1);
+                if (!p)
+                    goto efault;
+            } else {
+                p = NULL;
+            }
+
+            p2 = lock_user_string(arg2);
+            if (!p2) {
+                if (arg1)
+                    unlock_user(p, arg1, 0);
+                    goto efault;
+            }
+
+            if (arg3) {
+                p3 = lock_user_string(arg3);
+                if (!p3) {
+                    if (arg1)
                         unlock_user(p, arg1, 0);
-                        unlock_user(p2, arg2, 0);
-                        unlock_user(p3, arg3, 0);
-			break;
-		}
+                    unlock_user(p2, arg2, 0);
+                    goto efault;
+                }
+            } else {
+                p3 = NULL;
+            }
+
+            /* FIXME - arg5 should be locked, but it isn't clear how to
+             * do that since it's not guaranteed to be a NULL-terminated
+             * string.
+             */
+             if (!arg5)
+                 ret = get_errno(mount(p, p2, p3, (unsigned long)arg4, NULL));
+             else
+                 ret = get_errno(mount(p, p2, p3, (unsigned long)arg4, g2h(arg5)));
+
+             if (arg1)
+                 unlock_user(p, arg1, 0);
+             unlock_user(p2, arg2, 0);
+             if (arg3)
+                 unlock_user(p3, arg3, 0);
+        }
+        break;
 #ifdef TARGET_NR_umount
     case TARGET_NR_umount:
         if (!(p = lock_user_string(arg1)))
-- 
2.0.0

[Qemu-devel] [PATCH v2 06/16] linux-user: support strace of epoll_create1

[Paul Burton]

Add the epoll_create1 syscall to strace.list in order to display that
syscall when it occurs, rather than a message about the syscall being
unknown despite QEMU already implementing support for it.

Signed-off-by: Paul Burton <paul@archlinuxmips.org>
---
Changes in v2:
  - None.
---
 linux-user/strace.list | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/linux-user/strace.list b/linux-user/strace.list
index cf5841a..fcb258d 100644
--- a/linux-user/strace.list
+++ b/linux-user/strace.list
@@ -114,6 +114,9 @@
 #ifdef TARGET_NR_epoll_create
 { TARGET_NR_epoll_create, "epoll_create" , NULL, NULL, NULL },
 #endif
+#ifdef TARGET_NR_epoll_create1
+{ TARGET_NR_epoll_create1, "epoll_create1" , NULL, NULL, NULL },
+#endif
 #ifdef TARGET_NR_epoll_ctl
 { TARGET_NR_epoll_ctl, "epoll_ctl" , NULL, NULL, NULL },
 #endif
-- 
2.0.0

[Qemu-devel] [PATCH v2 07/16] linux-user: fix struct target_epoll_event layout for MIPS

[Paul Burton]

MIPS requires the pad field to 64b-align the data field just as ARM
does.

Signed-off-by: Paul Burton <paul@archlinuxmips.org>
---
Changes in v2:
  - Apply the same padding for the mips64 target.
---
 linux-user/syscall_defs.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/linux-user/syscall_defs.h b/linux-user/syscall_defs.h
index 69c3982..e379b45 100644
--- a/linux-user/syscall_defs.h
+++ b/linux-user/syscall_defs.h
@@ -2528,7 +2528,7 @@ typedef union target_epoll_data {
 
 struct target_epoll_event {
     uint32_t events;
-#ifdef TARGET_ARM
+#if defined(TARGET_ARM) || defined(TARGET_MIPS) || defined(TARGET_MIPS64)
     uint32_t __pad;
 #endif
     target_epoll_data_t data;
-- 
2.0.0

[Qemu-devel] [PATCH v2 08/16] linux-user: respect timezone for settimeofday

[Paul Burton]

The settimeofday syscall accepts a tz argument indicating the desired
timezone to the kernel. QEMU previously ignored any argument provided
by the target program & always passed NULL to the kernel. Instead,
translate the argument & pass along the data userland provided.

Although this argument is described by the settimeofday man page as
obsolete, it is used by systemd as of version 213.

Signed-off-by: Paul Burton <paul@archlinuxmips.org>
---
Changes in v2:
  - None.
---
 linux-user/syscall.c      | 27 ++++++++++++++++++++++++++-
 linux-user/syscall_defs.h |  5 +++++
 2 files changed, 31 insertions(+), 1 deletion(-)

diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index 3861f0f..ef04dc6 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -933,6 +933,22 @@ static inline abi_long copy_to_user_timeval(abi_ulong target_tv_addr,
     return 0;
 }
 
+static inline abi_long copy_from_user_timezone(struct timezone *tz,
+                                               abi_ulong target_tz_addr)
+{
+    struct target_timezone *target_tz;
+
+    if (!lock_user_struct(VERIFY_READ, target_tz, target_tz_addr, 1))
+        return -TARGET_EFAULT;
+
+    __get_user(tz->tz_minuteswest, &target_tz->tz_minuteswest);
+    __get_user(tz->tz_dsttime, &target_tz->tz_dsttime);
+
+    unlock_user_struct(target_tz, target_tz_addr, 0);
+
+    return 0;
+}
+
 #if defined(TARGET_NR_mq_open) && defined(__NR_mq_open)
 #include <mqueue.h>
 
@@ -6327,9 +6343,18 @@ abi_long do_syscall(void *cpu_env, int num, abi_long arg1,
     case TARGET_NR_settimeofday:
         {
             struct timeval tv;
+            struct timezone tz, *ptz = NULL;
+
             if (copy_from_user_timeval(&tv, arg1))
                 goto efault;
-            ret = get_errno(settimeofday(&tv, NULL));
+
+            if (arg2) {
+                if (copy_from_user_timezone(&tz, arg2))
+                    goto efault;
+                ptz = &tz;
+            }
+
+            ret = get_errno(settimeofday(&tv, ptz));
         }
         break;
 #if defined(TARGET_NR_select)
diff --git a/linux-user/syscall_defs.h b/linux-user/syscall_defs.h
index e379b45..a1f1fce 100644
--- a/linux-user/syscall_defs.h
+++ b/linux-user/syscall_defs.h
@@ -165,6 +165,11 @@ struct target_timespec {
     abi_long tv_nsec;
 };
 
+struct target_timezone {
+    abi_int tz_minuteswest;
+    abi_int tz_dsttime;
+};
+
 struct target_itimerval {
     struct target_timeval it_interval;
     struct target_timeval it_value;
-- 
2.0.0

[Qemu-devel] [PATCH v2 09/16] linux-user: allow NULL tv argument for settimeofday

[Paul Burton]

The tv argument to the settimeofday syscall is allowed to be NULL, if
the program only wishes to provide the timezone. QEMU previously
returned -EFAULT when tv was NULL. Instead, execute the syscall &
provide NULL to the kernel as the target program expected.

Signed-off-by: Paul Burton <paul@archlinuxmips.org>
---
Changes in v2:
  - None.
---
 linux-user/syscall.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index ef04dc6..cd0221b 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -6342,11 +6342,14 @@ abi_long do_syscall(void *cpu_env, int num, abi_long arg1,
         break;
     case TARGET_NR_settimeofday:
         {
-            struct timeval tv;
+            struct timeval tv, *ptv = NULL;
             struct timezone tz, *ptz = NULL;
 
-            if (copy_from_user_timeval(&tv, arg1))
-                goto efault;
+            if (arg1) {
+                if (copy_from_user_timeval(&tv, arg1))
+                    goto efault;
+                ptv = &tv;
+            }
 
             if (arg2) {
                 if (copy_from_user_timezone(&tz, arg2))
@@ -6354,7 +6357,7 @@ abi_long do_syscall(void *cpu_env, int num, abi_long arg1,
                 ptz = &tz;
             }
 
-            ret = get_errno(settimeofday(&tv, ptz));
+            ret = get_errno(settimeofday(ptv, ptz));
         }
         break;
 #if defined(TARGET_NR_select)
-- 
2.0.0

[Qemu-devel] [PATCH v2 10/16] linux-user: support timerfd_{create, gettime, settime} syscalls

[Paul Burton]

Adds support for the timerfd_create, timerfd_gettime & timerfd_settime
syscalls, allowing use of timerfds by target programs.

Signed-off-by: Paul Burton <paul@archlinuxmips.org>
---
Changes in v2:
  - None.
---
 linux-user/strace.list |  9 +++++++++
 linux-user/syscall.c   | 44 ++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 53 insertions(+)

diff --git a/linux-user/strace.list b/linux-user/strace.list
index fcb258d..8de972a 100644
--- a/linux-user/strace.list
+++ b/linux-user/strace.list
@@ -1404,6 +1404,15 @@
 #ifdef TARGET_NR_timer_settime
 { TARGET_NR_timer_settime, "timer_settime" , NULL, NULL, NULL },
 #endif
+#ifdef TARGET_NR_timerfd_create
+{ TARGET_NR_timerfd_create, "timerfd_create" , NULL, NULL, NULL },
+#endif
+#ifdef TARGET_NR_timerfd_gettime
+{ TARGET_NR_timerfd_gettime, "timerfd_gettime" , NULL, NULL, NULL },
+#endif
+#ifdef TARGET_NR_timerfd_settime
+{ TARGET_NR_timerfd_settime, "timerfd_settime" , NULL, NULL, NULL },
+#endif
 #ifdef TARGET_NR_times
 { TARGET_NR_times, "times" , NULL, NULL, NULL },
 #endif
diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index cd0221b..1c300be 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -58,6 +58,7 @@ int __clone2(int (*fn)(void *), void *child_stack_base,
 #include <sys/shm.h>
 #include <sys/sem.h>
 #include <sys/statfs.h>
+#include <sys/timerfd.h>
 #include <utime.h>
 #include <sys/sysinfo.h>
 //#include <sys/user.h>
@@ -9430,6 +9431,49 @@ abi_long do_syscall(void *cpu_env, int num, abi_long arg1,
     }
 #endif
 
+#ifdef TARGET_NR_timerfd_create
+    case TARGET_NR_timerfd_create:
+        ret = get_errno(timerfd_create(arg1,
+                target_to_host_bitmask(arg2, fcntl_flags_tbl)));
+        break;
+#endif
+
+#ifdef TARGET_NR_timerfd_gettime
+    case TARGET_NR_timerfd_gettime:
+        {
+            struct itimerspec its_curr;
+
+            ret = get_errno(timerfd_gettime(arg1, &its_curr));
+
+            if (arg2 && host_to_target_itimerspec(arg2, &its_curr)) {
+                goto efault;
+            }
+        }
+        break;
+#endif
+
+#ifdef TARGET_NR_timerfd_settime
+    case TARGET_NR_timerfd_settime:
+        {
+            struct itimerspec its_new, its_old, *p_new;
+
+            if (arg3) {
+                if (target_to_host_itimerspec(&its_new, arg3))
+                    goto efault;
+                p_new = &its_new;
+            } else {
+                p_new = NULL;
+            }
+
+            ret = get_errno(timerfd_settime(arg1, arg2, p_new, &its_old));
+
+            if (arg4 && host_to_target_itimerspec(arg4, &its_old)) {
+                goto efault;
+            }
+        }
+        break;
+#endif
+
     default:
     unimplemented:
         gemu_log("qemu: Unsupported syscall: %d\n", num);
-- 
2.0.0

[Qemu-devel] [PATCH v2 11/16] linux-user: support ioprio_{get, set} syscalls

[Paul Burton]

Add support for the ioprio_get & ioprio_set syscalls, allowing their
use by target programs.

Signed-off-by: Paul Burton <paul@archlinuxmips.org>
---
Changes in v2:
  - Declare the host syscalls conditionally upon whether both host &
    target define their numbers, to avoid compile time warnings/errors
    about the host syscall functions being unused when built for targets
    which don't support them.
---
 linux-user/syscall.c | 18 ++++++++++++++++++
 1 file changed, 18 insertions(+)

diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index 1c300be..fb36f46 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -252,6 +252,12 @@ _syscall2(int, capget, struct __user_cap_header_struct *, header,
           struct __user_cap_data_struct *, data);
 _syscall2(int, capset, struct __user_cap_header_struct *, header,
           struct __user_cap_data_struct *, data);
+#if defined(TARGET_NR_ioprio_get) && defined(__NR_ioprio_get)
+_syscall2(int, ioprio_get, int, which, int, who)
+#endif
+#if defined(TARGET_NR_ioprio_set) && defined(__NR_ioprio_set)
+_syscall3(int, ioprio_set, int, which, int, who, int, ioprio)
+#endif
 
 static bitmask_transtbl fcntl_flags_tbl[] = {
   { TARGET_O_ACCMODE,   TARGET_O_WRONLY,    O_ACCMODE,   O_WRONLY,    },
@@ -9474,6 +9480,18 @@ abi_long do_syscall(void *cpu_env, int num, abi_long arg1,
         break;
 #endif
 
+#if defined(TARGET_NR_ioprio_get) && defined(__NR_ioprio_get)
+    case TARGET_NR_ioprio_get:
+        ret = get_errno(ioprio_get(arg1, arg2));
+        break;
+#endif
+
+#if defined(TARGET_NR_ioprio_set) && defined(__NR_ioprio_set)
+    case TARGET_NR_ioprio_set:
+        ret = get_errno(ioprio_set(arg1, arg2, arg3));
+        break;
+#endif
+
     default:
     unimplemented:
         gemu_log("qemu: Unsupported syscall: %d\n", num);
-- 
2.0.0

[Qemu-devel] [PATCH v2 12/16] linux-user: support {name_to, open_by}_handle_at syscalls

[Paul Burton]

Implement support for the name_to_handle_at and open_by_handle_at
syscalls, allowing their use by the target program.

Signed-off-by: Paul Burton <paul@archlinuxmips.org>
---
Changes in v2:
  - None.
---
 linux-user/strace.c    | 30 ++++++++++++++++++++++++++++++
 linux-user/strace.list |  6 ++++++
 linux-user/syscall.c   | 50 ++++++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 86 insertions(+)

diff --git a/linux-user/strace.c b/linux-user/strace.c
index ea6c1d2..c20ddf1 100644
--- a/linux-user/strace.c
+++ b/linux-user/strace.c
@@ -1552,6 +1552,36 @@ print_kill(const struct syscallname *name,
 }
 #endif
 
+#ifdef TARGET_NR_name_to_handle_at
+static void
+print_name_to_handle_at(const struct syscallname *name,
+    abi_long arg0, abi_long arg1, abi_long arg2,
+    abi_long arg3, abi_long arg4, abi_long arg5)
+{
+    print_syscall_prologue(name);
+    print_at_dirfd(arg0, 0);
+    print_string(arg1, 0);
+    print_pointer(arg2, 0);
+    print_pointer(arg3, 0);
+    print_raw_param("0x%x", arg4, 1);
+    print_syscall_epilogue(name);
+}
+#endif
+
+#ifdef TARGET_NR_open_by_handle_at
+static void
+print_open_by_handle_at(const struct syscallname *name,
+    abi_long arg0, abi_long arg1, abi_long arg2,
+    abi_long arg3, abi_long arg4, abi_long arg5)
+{
+    print_syscall_prologue(name);
+    print_raw_param("%d", arg0, 0);
+    print_pointer(arg2, 0);
+    print_open_flags(arg3, 1);
+    print_syscall_epilogue(name);
+}
+#endif
+
 /*
  * An array of all of the syscalls we know about
  */
diff --git a/linux-user/strace.list b/linux-user/strace.list
index 8de972a..147f579 100644
--- a/linux-user/strace.list
+++ b/linux-user/strace.list
@@ -582,6 +582,9 @@
 #ifdef TARGET_NR_munmap
 { TARGET_NR_munmap, "munmap" , NULL, print_munmap, NULL },
 #endif
+#ifdef TARGET_NR_name_to_handle_at
+{ TARGET_NR_name_to_handle_at, "name_to_handle_at" , NULL, print_name_to_handle_at, NULL },
+#endif
 #ifdef TARGET_NR_nanosleep
 { TARGET_NR_nanosleep, "nanosleep" , NULL, NULL, NULL },
 #endif
@@ -624,6 +627,9 @@
 #ifdef TARGET_NR_openat
 { TARGET_NR_openat, "openat" , NULL, print_openat, NULL },
 #endif
+#ifdef TARGET_NR_open_by_handle_at
+{ TARGET_NR_open_by_handle_at, "open_by_handle_at" , NULL, print_open_by_handle_at, NULL },
+#endif
 #ifdef TARGET_NR_osf_adjtime
 { TARGET_NR_osf_adjtime, "osf_adjtime" , NULL, NULL, NULL },
 #endif
diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index fb36f46..0495781 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -5347,6 +5347,56 @@ abi_long do_syscall(void *cpu_env, int num, abi_long arg1,
         unlock_user(p, arg2, 0);
         break;
 #endif
+#ifdef TARGET_NR_name_to_handle_at
+    case TARGET_NR_name_to_handle_at:
+        {
+            struct file_handle *fh;
+            uint32_t sz;
+            int mount_id;
+
+            if (!(p = lock_user_string(arg2)))
+                goto efault;
+
+            if (get_user_u32(sz, arg3)) {
+                unlock_user(p, arg2, 0);
+                goto efault;
+            }
+
+            if (!(fh = lock_user(VERIFY_WRITE, arg3, sizeof(*fh) + sz, 1))) {
+                unlock_user(p, arg2, 0);
+                goto efault;
+            }
+
+            ret = get_errno(name_to_handle_at(arg1, path(p), fh,
+                                              &mount_id, arg5));
+
+            unlock_user(p, arg2, 0);
+            unlock_user(p, arg3, sizeof(*fh) + sz);
+
+            if (put_user_s32(mount_id, arg4))
+                goto efault;
+        }
+        break;
+#endif
+#ifdef TARGET_NR_open_by_handle_at
+    case TARGET_NR_open_by_handle_at:
+        {
+            struct file_handle *fh;
+            uint32_t sz;
+
+            if (get_user_u32(sz, arg2))
+                goto efault;
+
+            if (!(fh = lock_user(VERIFY_WRITE, arg2, sizeof(*fh) + sz, 1)))
+                goto efault;
+
+            ret = get_errno(open_by_handle_at(arg1, fh,
+                    target_to_host_bitmask(arg3, fcntl_flags_tbl)));
+
+            unlock_user(p, arg2, sizeof(*fh) + sz);
+        }
+        break;
+#endif
     case TARGET_NR_close:
         ret = get_errno(close(arg1));
         break;
-- 
2.0.0

[Qemu-devel] [PATCH v2 13/16] linux-user: support the setns syscall

[Paul Burton]

Add support for the setns syscall, trivially passed through to the host.

Signed-off-by: Paul Burton <paul@archlinuxmips.org>
---
Changes in v2:
  - None.
---
 linux-user/strace.list | 3 +++
 linux-user/syscall.c   | 6 ++++++
 2 files changed, 9 insertions(+)

diff --git a/linux-user/strace.list b/linux-user/strace.list
index 147f579..d5b8033 100644
--- a/linux-user/strace.list
+++ b/linux-user/strace.list
@@ -1191,6 +1191,9 @@
 #ifdef TARGET_NR_set_mempolicy
 { TARGET_NR_set_mempolicy, "set_mempolicy" , NULL, NULL, NULL },
 #endif
+#ifdef TARGET_NR_setns
+{ TARGET_NR_setns, "setns" , NULL, NULL, NULL },
+#endif
 #ifdef TARGET_NR_setpgid
 { TARGET_NR_setpgid, "setpgid" , NULL, NULL, NULL },
 #endif
diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index 0495781..0bbdf61 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -9542,6 +9542,12 @@ abi_long do_syscall(void *cpu_env, int num, abi_long arg1,
         break;
 #endif
 
+#ifdef TARGET_NR_setns
+    case TARGET_NR_setns:
+        ret = get_errno(setns(arg1, arg2));
+        break;
+#endif
+
     default:
     unimplemented:
         gemu_log("qemu: Unsupported syscall: %d\n", num);
-- 
2.0.0

[Qemu-devel] [PATCH v2 14/16] linux-user: support the unshare syscall

[Paul Burton]

Add support for the unshare syscall, trivially passed through to the
host.

Signed-off-by: Paul Burton <paul@archlinuxmips.org>
---
Changes in v2:
  - None.
---
 linux-user/syscall.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index 0bbdf61..936064c 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -9548,6 +9548,12 @@ abi_long do_syscall(void *cpu_env, int num, abi_long arg1,
         break;
 #endif
 
+#ifdef TARGET_NR_unshare
+    case TARGET_NR_unshare:
+        ret = get_errno(unshare(arg1));
+        break;
+#endif
+
     default:
     unimplemented:
         gemu_log("qemu: Unsupported syscall: %d\n", num);
-- 
2.0.0

[Qemu-devel] [PATCH v2 15/16] linux-user: support the KDSIGACCEPT ioctl

[Paul Burton]

Add a definition of the KDSIGACCEPT ioctl & allow its use by target
programs.

Signed-off-by: Paul Burton <paul@archlinuxmips.org>
---
Changes in v2:
  - None.
---
 linux-user/ioctls.h       | 1 +
 linux-user/syscall_defs.h | 1 +
 2 files changed, 2 insertions(+)

diff --git a/linux-user/ioctls.h b/linux-user/ioctls.h
index 309fb21..cd21e64 100644
--- a/linux-user/ioctls.h
+++ b/linux-user/ioctls.h
@@ -64,6 +64,7 @@
      IOCTL(KDSKBLED, 0, TYPE_INT)
      IOCTL(KDGETLED, 0, TYPE_INT)
      IOCTL(KDSETLED, 0, TYPE_INT)
+     IOCTL(KDSIGACCEPT, 0, TYPE_INT)
 
      IOCTL(BLKROSET, IOC_W, MK_PTR(TYPE_INT))
      IOCTL(BLKROGET, IOC_R, MK_PTR(TYPE_INT))
diff --git a/linux-user/syscall_defs.h b/linux-user/syscall_defs.h
index a1f1fce..4adfd3a 100644
--- a/linux-user/syscall_defs.h
+++ b/linux-user/syscall_defs.h
@@ -831,6 +831,7 @@ struct target_pollfd {
 #define TARGET_KDSKBLED        0x4B65	/* set led flags (not lights) */
 #define TARGET_KDGETLED        0x4B31	/* return current led state */
 #define TARGET_KDSETLED        0x4B32	/* set led state [lights, not flags] */
+#define TARGET_KDSIGACCEPT     0x4B4E
 
 #define TARGET_SIOCATMARK      0x8905
 
-- 
2.0.0

[Qemu-devel] [PATCH v2 16/16] linux-user: support the SIOCGIFINDEX ioctl

[Paul Burton]

Add a definition of the SIOCGIFINDEX ioctl, allowing its use by target
programs.

Signed-off-by: Paul Burton <paul@archlinuxmips.org>
---
Changes in v2:
  - None.
---
 linux-user/ioctls.h       | 1 +
 linux-user/syscall_defs.h | 1 +
 2 files changed, 2 insertions(+)

diff --git a/linux-user/ioctls.h b/linux-user/ioctls.h
index cd21e64..20551a8 100644
--- a/linux-user/ioctls.h
+++ b/linux-user/ioctls.h
@@ -118,6 +118,7 @@
   IOCTL(SIOCSIFMEM, IOC_W, MK_PTR(MK_STRUCT(STRUCT_ptr_ifreq)))
   IOCTL(SIOCADDMULTI, IOC_W, MK_PTR(MK_STRUCT(STRUCT_sockaddr_ifreq)))
   IOCTL(SIOCDELMULTI, IOC_W, MK_PTR(MK_STRUCT(STRUCT_sockaddr_ifreq)))
+  IOCTL(SIOCGIFINDEX, IOC_W | IOC_R, MK_PTR(MK_STRUCT(STRUCT_sockaddr_ifreq)))
   IOCTL(SIOCSIFLINK, 0, TYPE_NULL)
   IOCTL_SPECIAL(SIOCGIFCONF, IOC_W | IOC_R, do_ioctl_ifconf,
                 MK_PTR(MK_STRUCT(STRUCT_ifconf)))
diff --git a/linux-user/syscall_defs.h b/linux-user/syscall_defs.h
index 4adfd3a..8563027 100644
--- a/linux-user/syscall_defs.h
+++ b/linux-user/syscall_defs.h
@@ -865,6 +865,7 @@ struct target_pollfd {
 #define TARGET_SIOCSIFSLAVE    0x8930
 #define TARGET_SIOCADDMULTI    0x8931          /* Multicast address lists      */
 #define TARGET_SIOCDELMULTI    0x8932
+#define TARGET_SIOCGIFINDEX    0x8933
 
 /* Bridging control calls */
 #define TARGET_SIOCGIFBR       0x8940          /* Bridging support             */
-- 
2.0.0

Re: [Qemu-devel] [PATCH v2 15/16] linux-user: support the KDSIGACCEPT ioctl

[Peter Maydell]

On 21 June 2014 23:53, Paul Burton <paul@archlinuxmips.org> wrote:
> Add a definition of the KDSIGACCEPT ioctl & allow its use by target
> programs.
>
> Signed-off-by: Paul Burton <paul@archlinuxmips.org>
> ---
> Changes in v2:
>   - None.
> ---
>  linux-user/ioctls.h       | 1 +
>  linux-user/syscall_defs.h | 1 +
>  2 files changed, 2 insertions(+)
>
> diff --git a/linux-user/ioctls.h b/linux-user/ioctls.h
> index 309fb21..cd21e64 100644
> --- a/linux-user/ioctls.h
> +++ b/linux-user/ioctls.h
> @@ -64,6 +64,7 @@
>       IOCTL(KDSKBLED, 0, TYPE_INT)
>       IOCTL(KDGETLED, 0, TYPE_INT)
>       IOCTL(KDSETLED, 0, TYPE_INT)
> +     IOCTL(KDSIGACCEPT, 0, TYPE_INT)
>
>       IOCTL(BLKROSET, IOC_W, MK_PTR(TYPE_INT))
>       IOCTL(BLKROGET, IOC_R, MK_PTR(TYPE_INT))
> diff --git a/linux-user/syscall_defs.h b/linux-user/syscall_defs.h
> index a1f1fce..4adfd3a 100644
> --- a/linux-user/syscall_defs.h
> +++ b/linux-user/syscall_defs.h
> @@ -831,6 +831,7 @@ struct target_pollfd {
>  #define TARGET_KDSKBLED        0x4B65  /* set led flags (not lights) */
>  #define TARGET_KDGETLED        0x4B31  /* return current led state */
>  #define TARGET_KDSETLED        0x4B32  /* set led state [lights, not flags] */
> +#define TARGET_KDSIGACCEPT     0x4B4E
>
>  #define TARGET_SIOCATMARK      0x8905

The argument to this ioctl is a signal number, right?
Surely we need to translate it to a host signal number,
not just pass it through as TYPE_INT...

thanks
-- PMM

Re: [Qemu-devel] [PATCH v2 08/16] linux-user: respect timezone for settimeofday

[Peter Maydell]

On 21 June 2014 23:53, Paul Burton <paul@archlinuxmips.org> wrote:
> The settimeofday syscall accepts a tz argument indicating the desired
> timezone to the kernel. QEMU previously ignored any argument provided
> by the target program & always passed NULL to the kernel. Instead,
> translate the argument & pass along the data userland provided.
>
> Although this argument is described by the settimeofday man page as
> obsolete, it is used by systemd as of version 213.
>
> Signed-off-by: Paul Burton <paul@archlinuxmips.org>
> ---
> Changes in v2:
>   - None.
> ---
>  linux-user/syscall.c      | 27 ++++++++++++++++++++++++++-
>  linux-user/syscall_defs.h |  5 +++++
>  2 files changed, 31 insertions(+), 1 deletion(-)
>
> diff --git a/linux-user/syscall.c b/linux-user/syscall.c
> index 3861f0f..ef04dc6 100644
> --- a/linux-user/syscall.c
> +++ b/linux-user/syscall.c
> @@ -933,6 +933,22 @@ static inline abi_long copy_to_user_timeval(abi_ulong target_tv_addr,
>      return 0;
>  }
>
> +static inline abi_long copy_from_user_timezone(struct timezone *tz,
> +                                               abi_ulong target_tz_addr)
> +{
> +    struct target_timezone *target_tz;
> +
> +    if (!lock_user_struct(VERIFY_READ, target_tz, target_tz_addr, 1))
> +        return -TARGET_EFAULT;

Coding style mandates braces even on single-line if()s; checkpatch.pl
will catch this usually.

Code looks OK otherwise.

thanks
-- PMM

Re: [Qemu-devel] [PATCH v2 15/16] linux-user: support the KDSIGACCEPT ioctl

[Paul Burton]

[-- Attachment #1: Type: text/plain, Size: 666 bytes --]

On Sun, Jun 22, 2014 at 12:13:27AM +0100, Peter Maydell wrote:
> > diff --git a/linux-user/ioctls.h b/linux-user/ioctls.h
> > index 309fb21..cd21e64 100644
> > --- a/linux-user/ioctls.h
> > +++ b/linux-user/ioctls.h
> > @@ -64,6 +64,7 @@
> >       IOCTL(KDSKBLED, 0, TYPE_INT)
> >       IOCTL(KDGETLED, 0, TYPE_INT)
> >       IOCTL(KDSETLED, 0, TYPE_INT)
> > +     IOCTL(KDSIGACCEPT, 0, TYPE_INT)
...snip...
> 
> The argument to this ioctl is a signal number, right?
> Surely we need to translate it to a host signal number,
> not just pass it through as TYPE_INT...

True, not sure how I missed that - the clue is in the name!

Thanks,
    Paul

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

Re: [Qemu-devel] [PATCH v2 08/16] linux-user: respect timezone for settimeofday

[Paul Burton]

[-- Attachment #1: Type: text/plain, Size: 494 bytes --]

On Sun, Jun 22, 2014 at 12:18:02AM +0100, Peter Maydell wrote:
> > +    if (!lock_user_struct(VERIFY_READ, target_tz, target_tz_addr, 1))
> > +        return -TARGET_EFAULT;
> 
> Coding style mandates braces even on single-line if()s; checkpatch.pl
> will catch this usually.

I copied that style from the equivalent timeval functions, but yeah I'll
make sure to get into the habit of running checkpatch for QEMU patches.

> Code looks OK otherwise.

Thanks for looking,
    Paul

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

Re: [Qemu-devel] [PATCH v2 12/16] linux-user: support {name_to, open_by}_handle_at syscalls

[Riku Voipio]

Hi,

Comments inline:

On Sat, Jun 21, 2014 at 11:53:07PM +0100, Paul Burton wrote:
> Implement support for the name_to_handle_at and open_by_handle_at
> syscalls, allowing their use by the target program.
> 
> Signed-off-by: Paul Burton <paul@archlinuxmips.org>
> ---
> Changes in v2:
>   - None.
> ---
>  linux-user/strace.c    | 30 ++++++++++++++++++++++++++++++
>  linux-user/strace.list |  6 ++++++
>  linux-user/syscall.c   | 50 ++++++++++++++++++++++++++++++++++++++++++++++++++
>  3 files changed, 86 insertions(+)
> 
> diff --git a/linux-user/strace.c b/linux-user/strace.c
> index ea6c1d2..c20ddf1 100644
> --- a/linux-user/strace.c
> +++ b/linux-user/strace.c
> @@ -1552,6 +1552,36 @@ print_kill(const struct syscallname *name,
>  }
>  #endif
>  
> +#ifdef TARGET_NR_name_to_handle_at
> +static void
> +print_name_to_handle_at(const struct syscallname *name,
> +    abi_long arg0, abi_long arg1, abi_long arg2,
> +    abi_long arg3, abi_long arg4, abi_long arg5)
> +{
> +    print_syscall_prologue(name);
> +    print_at_dirfd(arg0, 0);
> +    print_string(arg1, 0);
> +    print_pointer(arg2, 0);
> +    print_pointer(arg3, 0);
> +    print_raw_param("0x%x", arg4, 1);
> +    print_syscall_epilogue(name);
> +}
> +#endif
> +
> +#ifdef TARGET_NR_open_by_handle_at
> +static void
> +print_open_by_handle_at(const struct syscallname *name,
> +    abi_long arg0, abi_long arg1, abi_long arg2,
> +    abi_long arg3, abi_long arg4, abi_long arg5)
> +{
> +    print_syscall_prologue(name);
> +    print_raw_param("%d", arg0, 0);
> +    print_pointer(arg2, 0);
> +    print_open_flags(arg3, 1);
> +    print_syscall_epilogue(name);
> +}
> +#endif
> +
>  /*
>   * An array of all of the syscalls we know about
>   */
> diff --git a/linux-user/strace.list b/linux-user/strace.list
> index 8de972a..147f579 100644
> --- a/linux-user/strace.list
> +++ b/linux-user/strace.list
> @@ -582,6 +582,9 @@
>  #ifdef TARGET_NR_munmap
>  { TARGET_NR_munmap, "munmap" , NULL, print_munmap, NULL },
>  #endif
> +#ifdef TARGET_NR_name_to_handle_at
> +{ TARGET_NR_name_to_handle_at, "name_to_handle_at" , NULL, print_name_to_handle_at, NULL },
> +#endif
>  #ifdef TARGET_NR_nanosleep
>  { TARGET_NR_nanosleep, "nanosleep" , NULL, NULL, NULL },
>  #endif
> @@ -624,6 +627,9 @@
>  #ifdef TARGET_NR_openat
>  { TARGET_NR_openat, "openat" , NULL, print_openat, NULL },
>  #endif
> +#ifdef TARGET_NR_open_by_handle_at
> +{ TARGET_NR_open_by_handle_at, "open_by_handle_at" , NULL, print_open_by_handle_at, NULL },
> +#endif
>  #ifdef TARGET_NR_osf_adjtime
>  { TARGET_NR_osf_adjtime, "osf_adjtime" , NULL, NULL, NULL },
>  #endif
> diff --git a/linux-user/syscall.c b/linux-user/syscall.c
> index fb36f46..0495781 100644
> --- a/linux-user/syscall.c
> +++ b/linux-user/syscall.c
> @@ -5347,6 +5347,56 @@ abi_long do_syscall(void *cpu_env, int num, abi_long arg1,
>          unlock_user(p, arg2, 0);
>          break;
>  #endif
> +#ifdef TARGET_NR_name_to_handle_at

+#if defined(TARGET_NR_name_to_handle_at) && defined(__NR_name_to_handle_at)

This is something that all other syscall definitions need - else qemu won't compile
on old distributions anymore.

> +    case TARGET_NR_name_to_handle_at:
> +        {
> +            struct file_handle *fh;
> +            uint32_t sz;
> +            int mount_id;
> +
> +            if (!(p = lock_user_string(arg2)))
> +                goto efault;
> +
> +            if (get_user_u32(sz, arg3)) {
> +                unlock_user(p, arg2, 0);
> +                goto efault;
> +            }
> +
> +            if (!(fh = lock_user(VERIFY_WRITE, arg3, sizeof(*fh) + sz, 1))) {
> +                unlock_user(p, arg2, 0);
> +                goto efault;
> +            }
> +
> +            ret = get_errno(name_to_handle_at(arg1, path(p), fh,
> +                                              &mount_id, arg5));
> +
> +            unlock_user(p, arg2, 0);
> +            unlock_user(p, arg3, sizeof(*fh) + sz);

I think:
+            unlock_user(fh, arg3, sizeof(*fh) + sz);

> +
> +            if (put_user_s32(mount_id, arg4))
> +                goto efault;
> +        }
> +        break;
> +#endif
> +#ifdef TARGET_NR_open_by_handle_at
> +    case TARGET_NR_open_by_handle_at:
> +        {
> +            struct file_handle *fh;
> +            uint32_t sz;
> +
> +            if (get_user_u32(sz, arg2))
> +                goto efault;
> +
> +            if (!(fh = lock_user(VERIFY_WRITE, arg2, sizeof(*fh) + sz, 1)))
> +                goto efault;
> +
> +            ret = get_errno(open_by_handle_at(arg1, fh,
> +                    target_to_host_bitmask(arg3, fcntl_flags_tbl)));
> +
> +            unlock_user(p, arg2, sizeof(*fh) + sz);

And here too:

+            unlock_user(fh, arg2, sizeof(*fh) + sz);

> +        }
> +        break;
> +#endif
>      case TARGET_NR_close:
>          ret = get_errno(close(arg1));
>          break;
> -- 
> 2.0.0

Re: [Qemu-devel] [PATCH v2 00/16] linux-user fixes & improvements

[Riku Voipio]

On Sat, Jun 21, 2014 at 11:52:55PM +0100, Paul Burton wrote:
> From: Paul Burton <paul.burton@imgtec.com>
> 
> This series fixes a number of bugs in QEMUs linux-user support, some
> specific to targetting the MIPS architecture but mostly generic. It also
> adds support for some previously unsupported syscalls & {g,s}etsockopt
> options.

Looks mostly good - unfortunately I missed some issues last round. 
iI gave comments to the {name_to,open_by}_handle_at patch - which also
caused the build error for Peter.

Also, compile test on older distro showed that new syscalls need both
target and host test:

#ifdef TARGET_NR_setns

needs to be replaced with

#if defined(TARGET_NR_setns) && defined(__NR_setns) 

Riku

> Paul Burton (16):
>   linux-user: translate the result of getsockopt SO_TYPE
>   linux-user: support SO_ACCEPTCONN getsockopt option
>   linux-user: support SO_{SND,RCV}BUFFORCE setsockopt options
>   linux-user: support SO_PASSSEC setsockopt option
>   linux-user: allow NULL arguments to mount
>   linux-user: support strace of epoll_create1
>   linux-user: fix struct target_epoll_event layout for MIPS
>   linux-user: respect timezone for settimeofday
>   linux-user: allow NULL tv argument for settimeofday
>   linux-user: support timerfd_{create,gettime,settime} syscalls
>   linux-user: support ioprio_{get,set} syscalls
>   linux-user: support {name_to,open_by}_handle_at syscalls
>   linux-user: support the setns syscall
>   linux-user: support the unshare syscall
>   linux-user: support the KDSIGACCEPT ioctl
>   linux-user: support the SIOCGIFINDEX ioctl
> 
>  linux-user/ioctls.h       |   2 +
>  linux-user/socket.h       |   5 +
>  linux-user/strace.c       |  30 +++++
>  linux-user/strace.list    |  21 ++++
>  linux-user/syscall.c      | 271 +++++++++++++++++++++++++++++++++++++++++-----
>  linux-user/syscall_defs.h |   9 +-
>  6 files changed, 311 insertions(+), 27 deletions(-)
> 
> -- 
> 2.0.0

Re: [Qemu-devel] [PATCH v2 00/16] linux-user fixes & improvements

[Riku Voipio]

Hi Paul,

On Fri, Jun 27, 2014 at 04:30:34PM +0300, Riku Voipio wrote:
> On Sat, Jun 21, 2014 at 11:52:55PM +0100, Paul Burton wrote:
> > From: Paul Burton <paul.burton@imgtec.com>
> > 
> > This series fixes a number of bugs in QEMUs linux-user support, some
> > specific to targetting the MIPS architecture but mostly generic. It also
> > adds support for some previously unsupported syscalls & {g,s}etsockopt
> > options.
 
> Looks mostly good - unfortunately I missed some issues last round. 
> iI gave comments to the {name_to,open_by}_handle_at patch - which also
> caused the build error for Peter.
 
> Also, compile test on older distro showed that new syscalls need both
> target and host test:
 
> #ifdef TARGET_NR_setns
 
> needs to be replaced with
 
> #if defined(TARGET_NR_setns) && defined(__NR_setns) 

Do you have time to to make the needed ifdef changes and send the remaining
patches again? I can also make the changes myself, if you prefer. 

Riku

> Riku
> 
> > Paul Burton (16):
> >   linux-user: translate the result of getsockopt SO_TYPE
> >   linux-user: support SO_ACCEPTCONN getsockopt option
> >   linux-user: support SO_{SND,RCV}BUFFORCE setsockopt options
> >   linux-user: support SO_PASSSEC setsockopt option
> >   linux-user: allow NULL arguments to mount
> >   linux-user: support strace of epoll_create1
> >   linux-user: fix struct target_epoll_event layout for MIPS
> >   linux-user: respect timezone for settimeofday
> >   linux-user: allow NULL tv argument for settimeofday
> >   linux-user: support timerfd_{create,gettime,settime} syscalls
> >   linux-user: support ioprio_{get,set} syscalls
> >   linux-user: support {name_to,open_by}_handle_at syscalls
> >   linux-user: support the setns syscall
> >   linux-user: support the unshare syscall
> >   linux-user: support the KDSIGACCEPT ioctl
> >   linux-user: support the SIOCGIFINDEX ioctl
> > 
> >  linux-user/ioctls.h       |   2 +
> >  linux-user/socket.h       |   5 +
> >  linux-user/strace.c       |  30 +++++
> >  linux-user/strace.list    |  21 ++++
> >  linux-user/syscall.c      | 271 +++++++++++++++++++++++++++++++++++++++++-----
> >  linux-user/syscall_defs.h |   9 +-
> >  6 files changed, 311 insertions(+), 27 deletions(-)
> > 
> > -- 
> > 2.0.0