qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: riku.voipio@linaro.org
To: qemu-devel@nongnu.org
Cc: Peter Maydell <peter.maydell@linaro.org>
Subject: [Qemu-devel] [PULL 04/24] linux-user: Don't use sigfillset() on uc->uc_sigmask
Date: Fri, 24 Jun 2016 12:52:59 +0300	[thread overview]
Message-ID: <0494c67c6caa499e759bdea3ae7ddc7ea52da643.1466760944.git.riku.voipio@linaro.org> (raw)
In-Reply-To: <cover.1466760944.git.riku.voipio@linaro.org>

From: Peter Maydell <peter.maydell@linaro.org>

The kernel and libc have different ideas about what a sigset_t
is -- for the kernel it is only _NSIG / 8 bytes in size (usually
8 bytes), but for libc it is much larger, 128 bytes. In most
situations the difference doesn't matter, because if you pass a
pointer to a libc sigset_t to the kernel it just acts on the first
8 bytes of it, but for the ucontext_t* argument to a signal handler
it trips us up. The kernel allocates this ucontext_t on the stack
according to its idea of the sigset_t type, but the type of the
ucontext_t defined by the libc headers uses the libc type, and
so do the manipulator functions like sigfillset(). This means that
 (1) sizeof(uc->uc_sigmask) is much larger than the actual
     space used on the stack
 (2) sigfillset(&uc->uc_sigmask) will write garbage 0xff bytes
     off the end of the structure, which can trash data that
     was on the stack before the signal handler was invoked,
     and may result in a crash after the handler returns

To avoid this, we use a memset() of the correct size to fill
the signal mask rather than using the libc function.

This fixes a problem where we would crash at least some of the
time on an i386 host when a signal was taken.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
---
 linux-user/qemu.h    |  5 +++++
 linux-user/signal.c  | 10 +++++++++-
 linux-user/syscall.c |  5 -----
 3 files changed, 14 insertions(+), 6 deletions(-)

diff --git a/linux-user/qemu.h b/linux-user/qemu.h
index 56f29c3..e8a5aed 100644
--- a/linux-user/qemu.h
+++ b/linux-user/qemu.h
@@ -20,6 +20,11 @@
 
 #define THREAD __thread
 
+/* This is the size of the host kernel's sigset_t, needed where we make
+ * direct system calls that take a sigset_t pointer and a size.
+ */
+#define SIGSET_T_SIZE (_NSIG / 8)
+
 /* This struct is used to hold certain information about the image.
  * Basically, it replicates in user space what would be certain
  * task_struct fields in the kernel
diff --git a/linux-user/signal.c b/linux-user/signal.c
index e2d55ff..9d98045 100644
--- a/linux-user/signal.c
+++ b/linux-user/signal.c
@@ -636,8 +636,16 @@ static void host_signal_handler(int host_signum, siginfo_t *info,
      * code in case the guest code provokes one in the window between
      * now and it getting out to the main loop. Signals will be
      * unblocked again in process_pending_signals().
+     *
+     * WARNING: we cannot use sigfillset() here because the uc_sigmask
+     * field is a kernel sigset_t, which is much smaller than the
+     * libc sigset_t which sigfillset() operates on. Using sigfillset()
+     * would write 0xff bytes off the end of the structure and trash
+     * data on the struct.
+     * We can't use sizeof(uc->uc_sigmask) either, because the libc
+     * headers define the struct field with the wrong (too large) type.
      */
-    sigfillset(&uc->uc_sigmask);
+    memset(&uc->uc_sigmask, 0xff, SIGSET_T_SIZE);
     sigdelset(&uc->uc_sigmask, SIGSEGV);
     sigdelset(&uc->uc_sigmask, SIGBUS);
 
diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index 8dc8c7a..95eafeb 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -123,11 +123,6 @@ int __clone2(int (*fn)(void *), void *child_stack_base,
 #define	VFAT_IOCTL_READDIR_BOTH		_IOR('r', 1, struct linux_dirent [2])
 #define	VFAT_IOCTL_READDIR_SHORT	_IOR('r', 2, struct linux_dirent [2])
 
-/* This is the size of the host kernel's sigset_t, needed where we make
- * direct system calls that take a sigset_t pointer and a size.
- */
-#define SIGSET_T_SIZE (_NSIG / 8)
-
 #undef _syscall0
 #undef _syscall1
 #undef _syscall2
-- 
2.1.4

  parent reply	other threads:[~2016-06-24  9:53 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-06-24  9:52 [Qemu-devel] [PULL 00/24] linux-user changes riku.voipio
2016-06-24  9:52 ` [Qemu-devel] [PULL 01/24] linux-user: Avoid possible misalignment in host_to_target_siginfo() riku.voipio
2016-06-24  9:52 ` [Qemu-devel] [PULL 02/24] linux-user: Use __get_user() and __put_user() to handle structs in do_fcntl() riku.voipio
2016-06-24  9:52 ` [Qemu-devel] [PULL 03/24] linux-user: Use safe_syscall wrapper for fcntl riku.voipio
2016-06-24  9:52 ` riku.voipio [this message]
2016-06-24  9:53 ` [Qemu-devel] [PULL 05/24] configure: Don't override ARCH=unknown if enabling TCI riku.voipio
2016-06-24  9:53 ` [Qemu-devel] [PULL 06/24] configure: Don't allow user-only targets for unknown CPU architectures riku.voipio
2016-06-24  9:53 ` [Qemu-devel] [PULL 07/24] user-exec: Delete now-unused hppa and m68k cpu_signal_handler() code riku.voipio
2016-06-24  9:53 ` [Qemu-devel] [PULL 08/24] user-exec: Remove unused code for OSX hosts riku.voipio
2016-06-24  9:53 ` [Qemu-devel] [PULL 09/24] linux-user: Create a hostdep.h for each host architecture riku.voipio
2016-06-24  9:53 ` [Qemu-devel] [PULL 10/24] linux-user: Fix wrong type used for argument to rt_sigqueueinfo riku.voipio
2016-06-24  9:53 ` [Qemu-devel] [PULL 11/24] linux-user: Support F_GETPIPE_SZ and F_SETPIPE_SZ fcntls riku.voipio
2016-06-24  9:53 ` [Qemu-devel] [PULL 12/24] linux-user: add socketcall() strace riku.voipio
2016-06-24  9:53 ` [Qemu-devel] [PULL 13/24] linux-user: add socket() strace riku.voipio
2016-06-24  9:53 ` [Qemu-devel] [PULL 14/24] linux-user: fix clone() strace riku.voipio
2016-06-24  9:53 ` [Qemu-devel] [PULL 15/24] linux-user: update get_thread_area/set_thread_area strace riku.voipio
2016-06-24  9:53 ` [Qemu-devel] [PULL 16/24] linux-user: add missing return in netlink switch statement riku.voipio
2016-06-24  9:53 ` [Qemu-devel] [PULL 17/24] linux-user: fd_trans_host_to_target_data() must process only received data riku.voipio
2016-06-24  9:53 ` [Qemu-devel] [PULL 18/24] linux-user: don't swap NLMSG_DATA() fields riku.voipio
2016-06-24  9:53 ` [Qemu-devel] [PULL 19/24] linux-user: fix x86_64 safe_syscall riku.voipio
2016-06-24  9:53 ` [Qemu-devel] [PULL 20/24] linux-user: Provide safe_syscall for i386 riku.voipio
2016-06-24  9:53 ` [Qemu-devel] [PULL 21/24] linux-user: Provide safe_syscall for arm riku.voipio
2016-06-24  9:53 ` [Qemu-devel] [PULL 22/24] linux-user: Provide safe_syscall for aarch64 riku.voipio
2016-06-24  9:53 ` [Qemu-devel] [PULL 23/24] linux-user: Provide safe_syscall for s390x riku.voipio
2016-06-24  9:53 ` [Qemu-devel] [PULL 24/24] linux-user: Provide safe_syscall for ppc64 riku.voipio
2016-06-24 13:36 ` [Qemu-devel] [PULL 00/24] linux-user changes Peter Maydell
2016-06-26 10:34   ` Riku Voipio

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=0494c67c6caa499e759bdea3ae7ddc7ea52da643.1466760944.git.riku.voipio@linaro.org \
    --to=riku.voipio@linaro.org \
    --cc=peter.maydell@linaro.org \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).