UML fixes for 3.1-rcX

UML fixes for 3.1-rcX

[Richard Weinberger]

[PATCH 1/9] um: disable CMPXCHG_DOUBLE as it breaks UML build
[PATCH 2/9] um: drivers/xterm.c: fix a file descriptor leak
[PATCH 3/9] um: Save FPU registers between task switches
[PATCH 4/9] um: fix oopsable race in line_close()
[PATCH 5/9] um: winch_interrupt() can happen inside of free_winch()
[PATCH 6/9] um: fix free_winch() mess
[PATCH 7/9] um: PTRACE_[GS]ETFPXREGS had been wired on the wrong
[PATCH 8/9] um: fix strrchr problems
[PATCH 9/9] um: clean arch_ptrace() up a bit

 arch/um/Kconfig.x86                       |    4 ++
 arch/um/Makefile                          |    2 +-
 arch/um/drivers/line.c                    |   61 +++++++++++++++++------------
 arch/um/drivers/xterm.c                   |    1 +
 arch/um/include/asm/ptrace-generic.h      |    4 --
 arch/um/include/shared/line.h             |    1 +
 arch/um/include/shared/registers.h        |    2 +-
 arch/um/kernel/process.c                  |    2 +-
 arch/um/kernel/ptrace.c                   |   28 -------------
 arch/um/os-Linux/registers.c              |    9 ++++-
 arch/um/os-Linux/skas/mem.c               |    2 +-
 arch/um/os-Linux/skas/process.c           |   19 ++++++++-
 arch/um/sys-i386/asm/ptrace.h             |    5 --
 arch/um/sys-i386/ptrace.c                 |   28 +++++++++++--
 arch/um/sys-i386/shared/sysdep/ptrace.h   |    1 +
 arch/um/sys-x86_64/ptrace.c               |   12 ++++--
 arch/um/sys-x86_64/shared/sysdep/ptrace.h |    1 +
 17 files changed, 105 insertions(+), 77 deletions(-)

[PATCH 1/9] um: disable CMPXCHG_DOUBLE as it breaks UML build

[Richard Weinberger]

From: Thadeu Lima de Souza Cascardo <cascardo@holoscopio.com>

Commit b789ef518b2a7231b0668c813f677cee528a9d3f tests for cmpxchg_double
support in the SLUB code and it breaks UML builds with SLUB. Since UML
does not support checking for CPU features, disable CMPXCHG_DOUBLE just
like CMPXCHG_LOCAL is disabled for UML.

Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@holoscopio.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Jeff Dike <jdike@addtoit.com>
Cc: Richard Weinberger <richard@nod.at>
Signed-off-by: Richard Weinberger <richard@nod.at>
---
 arch/um/Kconfig.x86 |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)

diff --git a/arch/um/Kconfig.x86 b/arch/um/Kconfig.x86
index d31ecf3..21bebe6 100644
--- a/arch/um/Kconfig.x86
+++ b/arch/um/Kconfig.x86
@@ -10,6 +10,10 @@ config CMPXCHG_LOCAL
 	bool
 	default n
 
+config CMPXCHG_DOUBLE
+	bool
+	default n
+
 source "arch/x86/Kconfig.cpu"
 
 endmenu
-- 
1.7.6.1

[PATCH 2/9] um: drivers/xterm.c: fix a file descriptor leak

From: Jonathan Neuschäfer <j.neuschaefer@gmx.net>

I could use out_close1, but that seems to be the code path to close the
fd returned by os_create_unix_socket, and using it to close the fd
returned by mkstemp might lead to some confusion, so I don't do it.

Signed-off-by: Jonathan Neuschäfer <j.neuschaefer@gmx.net>
Signed-off-by: Richard Weinberger <richard@nod.at>
---
 arch/um/drivers/xterm.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/arch/um/drivers/xterm.c b/arch/um/drivers/xterm.c
index 8ac7146..2e1de57 100644
--- a/arch/um/drivers/xterm.c
+++ b/arch/um/drivers/xterm.c
@@ -123,6 +123,7 @@ static int xterm_open(int input, int output, int primary, void *d,
 		err = -errno;
 		printk(UM_KERN_ERR "xterm_open : unlink failed, errno = %d\n",
 		       errno);
+		close(fd);
 		return err;
 	}
 	close(fd);
-- 
1.7.6.1

[uml-devel] [PATCH 2/9] um: drivers/xterm.c: fix a file descriptor leak

[Richard Weinberger]

From: Jonathan Neuschäfer <j.neuschaefer@gmx.net>

I could use out_close1, but that seems to be the code path to close the
fd returned by os_create_unix_socket, and using it to close the fd
returned by mkstemp might lead to some confusion, so I don't do it.

Signed-off-by: Jonathan Neuschäfer <j.neuschaefer@gmx.net>
Signed-off-by: Richard Weinberger <richard@nod.at>
---
 arch/um/drivers/xterm.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/arch/um/drivers/xterm.c b/arch/um/drivers/xterm.c
index 8ac7146..2e1de57 100644
--- a/arch/um/drivers/xterm.c
+++ b/arch/um/drivers/xterm.c
@@ -123,6 +123,7 @@ static int xterm_open(int input, int output, int primary, void *d,
 		err = -errno;
 		printk(UM_KERN_ERR "xterm_open : unlink failed, errno = %d\n",
 		       errno);
+		close(fd);
 		return err;
 	}
 	close(fd);
-- 
1.7.6.1


------------------------------------------------------------------------------
EMC VNX: the world's simplest storage, starting under $10K
The only unified storage solution that offers unified management 
Up to 160% more powerful than alternatives and 25% more efficient. 
Guaranteed. http://p.sf.net/sfu/emc-vnx-dev2dev
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

[PATCH 3/9] um: Save FPU registers between task switches

[Richard Weinberger]

From: Ingo van Lil <inguin@gmx.de>

Some time ago Jeff prepared a patch (42daba3) for UML to stop saving the
process FP state between task switches. The assumption was that since
with SKAS0 every guest process runs inside a host process context the
host OS will take care of keeping the proper FP state. Unfortunately
this is not true for multi-threaded applications, where all guest
threads share a single host process context yet all may use the FPU on
their own. Although I haven't verified it I suspect things to be even
worse in SKAS3 mode where all guest processes run inside a single host
process.

The patch below reintroduces the saving and restoring of the FP context
between task switches.

Signed-off-by: Ingo van Lil <inguin@gmx.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
Reported-by: gunnarlindroth@hotmail.com
Tested-by: gunnarlindroth@hotmail.com
CC: Stanislav Meduna <stano@meduna.org>

[richard@nod.at: Ingo posted this patch in 2009, sadly it was never applied
and got lost. Now in 2011 the problem was reported by Gunnar.]
---
 arch/um/include/shared/registers.h        |    2 +-
 arch/um/kernel/process.c                  |    2 +-
 arch/um/os-Linux/registers.c              |    9 ++++++++-
 arch/um/os-Linux/skas/mem.c               |    2 +-
 arch/um/os-Linux/skas/process.c           |   19 ++++++++++++++++++-
 arch/um/sys-i386/shared/sysdep/ptrace.h   |    1 +
 arch/um/sys-x86_64/shared/sysdep/ptrace.h |    1 +
 7 files changed, 31 insertions(+), 5 deletions(-)

diff --git a/arch/um/include/shared/registers.h b/arch/um/include/shared/registers.h
index b0b4589..f1e0aa5 100644
--- a/arch/um/include/shared/registers.h
+++ b/arch/um/include/shared/registers.h
@@ -16,7 +16,7 @@ extern int restore_fpx_registers(int pid, unsigned long *fp_regs);
 extern int save_registers(int pid, struct uml_pt_regs *regs);
 extern int restore_registers(int pid, struct uml_pt_regs *regs);
 extern int init_registers(int pid);
-extern void get_safe_registers(unsigned long *regs);
+extern void get_safe_registers(unsigned long *regs, unsigned long *fp_regs);
 extern unsigned long get_thread_reg(int reg, jmp_buf *buf);
 extern int get_fp_registers(int pid, unsigned long *regs);
 extern int put_fp_registers(int pid, unsigned long *regs);
diff --git a/arch/um/kernel/process.c b/arch/um/kernel/process.c
index fab4371..21c1ae7 100644
--- a/arch/um/kernel/process.c
+++ b/arch/um/kernel/process.c
@@ -202,7 +202,7 @@ int copy_thread(unsigned long clone_flags, unsigned long sp,
 		arch_copy_thread(&current->thread.arch, &p->thread.arch);
 	}
 	else {
-		get_safe_registers(p->thread.regs.regs.gp);
+		get_safe_registers(p->thread.regs.regs.gp, p->thread.regs.regs.fp);
 		p->thread.request.u.thread = current->thread.request.u.thread;
 		handler = new_thread_handler;
 	}
diff --git a/arch/um/os-Linux/registers.c b/arch/um/os-Linux/registers.c
index 830fe6a..b866b9e 100644
--- a/arch/um/os-Linux/registers.c
+++ b/arch/um/os-Linux/registers.c
@@ -8,6 +8,8 @@
 #include <string.h>
 #include <sys/ptrace.h>
 #include "sysdep/ptrace.h"
+#include "sysdep/ptrace_user.h"
+#include "registers.h"
 
 int save_registers(int pid, struct uml_pt_regs *regs)
 {
@@ -32,6 +34,7 @@ int restore_registers(int pid, struct uml_pt_regs *regs)
 /* This is set once at boot time and not changed thereafter */
 
 static unsigned long exec_regs[MAX_REG_NR];
+static unsigned long exec_fp_regs[FP_SIZE];
 
 int init_registers(int pid)
 {
@@ -42,10 +45,14 @@ int init_registers(int pid)
 		return -errno;
 
 	arch_init_registers(pid);
+	get_fp_registers(pid, exec_fp_regs);
 	return 0;
 }
 
-void get_safe_registers(unsigned long *regs)
+void get_safe_registers(unsigned long *regs, unsigned long *fp_regs)
 {
 	memcpy(regs, exec_regs, sizeof(exec_regs));
+
+	if (fp_regs)
+		memcpy(fp_regs, exec_fp_regs, sizeof(exec_fp_regs));
 }
diff --git a/arch/um/os-Linux/skas/mem.c b/arch/um/os-Linux/skas/mem.c
index d261f17..e771398 100644
--- a/arch/um/os-Linux/skas/mem.c
+++ b/arch/um/os-Linux/skas/mem.c
@@ -39,7 +39,7 @@ static unsigned long syscall_regs[MAX_REG_NR];
 
 static int __init init_syscall_regs(void)
 {
-	get_safe_registers(syscall_regs);
+	get_safe_registers(syscall_regs, NULL);
 	syscall_regs[REGS_IP_INDEX] = STUB_CODE +
 		((unsigned long) &batch_syscall_stub -
 		 (unsigned long) &__syscall_stub_start);
diff --git a/arch/um/os-Linux/skas/process.c b/arch/um/os-Linux/skas/process.c
index d6e0a22..dee0e8c 100644
--- a/arch/um/os-Linux/skas/process.c
+++ b/arch/um/os-Linux/skas/process.c
@@ -373,6 +373,9 @@ void userspace(struct uml_pt_regs *regs)
 		if (ptrace(PTRACE_SETREGS, pid, 0, regs->gp))
 			fatal_sigsegv();
 
+		if (put_fp_registers(pid, regs->fp))
+			fatal_sigsegv();
+
 		/* Now we set local_using_sysemu to be used for one loop */
 		local_using_sysemu = get_using_sysemu();
 
@@ -399,6 +402,12 @@ void userspace(struct uml_pt_regs *regs)
 			fatal_sigsegv();
 		}
 
+		if (get_fp_registers(pid, regs->fp)) {
+			printk(UM_KERN_ERR "userspace -  get_fp_registers failed, "
+			       "errno = %d\n", errno);
+			fatal_sigsegv();
+		}
+
 		UPT_SYSCALL_NR(regs) = -1; /* Assume: It's not a syscall */
 
 		if (WIFSTOPPED(status)) {
@@ -457,10 +466,11 @@ void userspace(struct uml_pt_regs *regs)
 }
 
 static unsigned long thread_regs[MAX_REG_NR];
+static unsigned long thread_fp_regs[FP_SIZE];
 
 static int __init init_thread_regs(void)
 {
-	get_safe_registers(thread_regs);
+	get_safe_registers(thread_regs, thread_fp_regs);
 	/* Set parent's instruction pointer to start of clone-stub */
 	thread_regs[REGS_IP_INDEX] = STUB_CODE +
 				(unsigned long) stub_clone_handler -
@@ -503,6 +513,13 @@ int copy_context_skas0(unsigned long new_stack, int pid)
 		return err;
 	}
 
+	err = put_fp_registers(pid, thread_fp_regs);
+	if (err < 0) {
+		printk(UM_KERN_ERR "copy_context_skas0 : put_fp_registers "
+		       "failed, pid = %d, err = %d\n", pid, err);
+		return err;
+	}
+
 	/* set a well known return code for detection of child write failure */
 	child_data->err = 12345678;
 
diff --git a/arch/um/sys-i386/shared/sysdep/ptrace.h b/arch/um/sys-i386/shared/sysdep/ptrace.h
index d50e62e..c398a50 100644
--- a/arch/um/sys-i386/shared/sysdep/ptrace.h
+++ b/arch/um/sys-i386/shared/sysdep/ptrace.h
@@ -53,6 +53,7 @@ extern int sysemu_supported;
 
 struct uml_pt_regs {
 	unsigned long gp[MAX_REG_NR];
+	unsigned long fp[HOST_FPX_SIZE];
 	struct faultinfo faultinfo;
 	long syscall;
 	int is_user;
diff --git a/arch/um/sys-x86_64/shared/sysdep/ptrace.h b/arch/um/sys-x86_64/shared/sysdep/ptrace.h
index fdba545..8ee8f8e 100644
--- a/arch/um/sys-x86_64/shared/sysdep/ptrace.h
+++ b/arch/um/sys-x86_64/shared/sysdep/ptrace.h
@@ -85,6 +85,7 @@
 
 struct uml_pt_regs {
 	unsigned long gp[MAX_REG_NR];
+	unsigned long fp[HOST_FP_SIZE];
 	struct faultinfo faultinfo;
 	long syscall;
 	int is_user;
-- 
1.7.6.1

[uml-devel] [PATCH 3/9] um: Save FPU registers between task switches

[Richard Weinberger]

From: Ingo van Lil <inguin@gmx.de>

Some time ago Jeff prepared a patch (42daba3) for UML to stop saving the
process FP state between task switches. The assumption was that since
with SKAS0 every guest process runs inside a host process context the
host OS will take care of keeping the proper FP state. Unfortunately
this is not true for multi-threaded applications, where all guest
threads share a single host process context yet all may use the FPU on
their own. Although I haven't verified it I suspect things to be even
worse in SKAS3 mode where all guest processes run inside a single host
process.

The patch below reintroduces the saving and restoring of the FP context
between task switches.

Signed-off-by: Ingo van Lil <inguin@gmx.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
Reported-by: gunnarlindroth@hotmail.com
Tested-by: gunnarlindroth@hotmail.com
CC: Stanislav Meduna <stano@meduna.org>

[richard@nod.at: Ingo posted this patch in 2009, sadly it was never applied
and got lost. Now in 2011 the problem was reported by Gunnar.]
---
 arch/um/include/shared/registers.h        |    2 +-
 arch/um/kernel/process.c                  |    2 +-
 arch/um/os-Linux/registers.c              |    9 ++++++++-
 arch/um/os-Linux/skas/mem.c               |    2 +-
 arch/um/os-Linux/skas/process.c           |   19 ++++++++++++++++++-
 arch/um/sys-i386/shared/sysdep/ptrace.h   |    1 +
 arch/um/sys-x86_64/shared/sysdep/ptrace.h |    1 +
 7 files changed, 31 insertions(+), 5 deletions(-)

diff --git a/arch/um/include/shared/registers.h b/arch/um/include/shared/registers.h
index b0b4589..f1e0aa5 100644
--- a/arch/um/include/shared/registers.h
+++ b/arch/um/include/shared/registers.h
@@ -16,7 +16,7 @@ extern int restore_fpx_registers(int pid, unsigned long *fp_regs);
 extern int save_registers(int pid, struct uml_pt_regs *regs);
 extern int restore_registers(int pid, struct uml_pt_regs *regs);
 extern int init_registers(int pid);
-extern void get_safe_registers(unsigned long *regs);
+extern void get_safe_registers(unsigned long *regs, unsigned long *fp_regs);
 extern unsigned long get_thread_reg(int reg, jmp_buf *buf);
 extern int get_fp_registers(int pid, unsigned long *regs);
 extern int put_fp_registers(int pid, unsigned long *regs);
diff --git a/arch/um/kernel/process.c b/arch/um/kernel/process.c
index fab4371..21c1ae7 100644
--- a/arch/um/kernel/process.c
+++ b/arch/um/kernel/process.c
@@ -202,7 +202,7 @@ int copy_thread(unsigned long clone_flags, unsigned long sp,
 		arch_copy_thread(&current->thread.arch, &p->thread.arch);
 	}
 	else {
-		get_safe_registers(p->thread.regs.regs.gp);
+		get_safe_registers(p->thread.regs.regs.gp, p->thread.regs.regs.fp);
 		p->thread.request.u.thread = current->thread.request.u.thread;
 		handler = new_thread_handler;
 	}
diff --git a/arch/um/os-Linux/registers.c b/arch/um/os-Linux/registers.c
index 830fe6a..b866b9e 100644
--- a/arch/um/os-Linux/registers.c
+++ b/arch/um/os-Linux/registers.c
@@ -8,6 +8,8 @@
 #include <string.h>
 #include <sys/ptrace.h>
 #include "sysdep/ptrace.h"
+#include "sysdep/ptrace_user.h"
+#include "registers.h"
 
 int save_registers(int pid, struct uml_pt_regs *regs)
 {
@@ -32,6 +34,7 @@ int restore_registers(int pid, struct uml_pt_regs *regs)
 /* This is set once at boot time and not changed thereafter */
 
 static unsigned long exec_regs[MAX_REG_NR];
+static unsigned long exec_fp_regs[FP_SIZE];
 
 int init_registers(int pid)
 {
@@ -42,10 +45,14 @@ int init_registers(int pid)
 		return -errno;
 
 	arch_init_registers(pid);
+	get_fp_registers(pid, exec_fp_regs);
 	return 0;
 }
 
-void get_safe_registers(unsigned long *regs)
+void get_safe_registers(unsigned long *regs, unsigned long *fp_regs)
 {
 	memcpy(regs, exec_regs, sizeof(exec_regs));
+
+	if (fp_regs)
+		memcpy(fp_regs, exec_fp_regs, sizeof(exec_fp_regs));
 }
diff --git a/arch/um/os-Linux/skas/mem.c b/arch/um/os-Linux/skas/mem.c
index d261f17..e771398 100644
--- a/arch/um/os-Linux/skas/mem.c
+++ b/arch/um/os-Linux/skas/mem.c
@@ -39,7 +39,7 @@ static unsigned long syscall_regs[MAX_REG_NR];
 
 static int __init init_syscall_regs(void)
 {
-	get_safe_registers(syscall_regs);
+	get_safe_registers(syscall_regs, NULL);
 	syscall_regs[REGS_IP_INDEX] = STUB_CODE +
 		((unsigned long) &batch_syscall_stub -
 		 (unsigned long) &__syscall_stub_start);
diff --git a/arch/um/os-Linux/skas/process.c b/arch/um/os-Linux/skas/process.c
index d6e0a22..dee0e8c 100644
--- a/arch/um/os-Linux/skas/process.c
+++ b/arch/um/os-Linux/skas/process.c
@@ -373,6 +373,9 @@ void userspace(struct uml_pt_regs *regs)
 		if (ptrace(PTRACE_SETREGS, pid, 0, regs->gp))
 			fatal_sigsegv();
 
+		if (put_fp_registers(pid, regs->fp))
+			fatal_sigsegv();
+
 		/* Now we set local_using_sysemu to be used for one loop */
 		local_using_sysemu = get_using_sysemu();
 
@@ -399,6 +402,12 @@ void userspace(struct uml_pt_regs *regs)
 			fatal_sigsegv();
 		}
 
+		if (get_fp_registers(pid, regs->fp)) {
+			printk(UM_KERN_ERR "userspace -  get_fp_registers failed, "
+			       "errno = %d\n", errno);
+			fatal_sigsegv();
+		}
+
 		UPT_SYSCALL_NR(regs) = -1; /* Assume: It's not a syscall */
 
 		if (WIFSTOPPED(status)) {
@@ -457,10 +466,11 @@ void userspace(struct uml_pt_regs *regs)
 }
 
 static unsigned long thread_regs[MAX_REG_NR];
+static unsigned long thread_fp_regs[FP_SIZE];
 
 static int __init init_thread_regs(void)
 {
-	get_safe_registers(thread_regs);
+	get_safe_registers(thread_regs, thread_fp_regs);
 	/* Set parent's instruction pointer to start of clone-stub */
 	thread_regs[REGS_IP_INDEX] = STUB_CODE +
 				(unsigned long) stub_clone_handler -
@@ -503,6 +513,13 @@ int copy_context_skas0(unsigned long new_stack, int pid)
 		return err;
 	}
 
+	err = put_fp_registers(pid, thread_fp_regs);
+	if (err < 0) {
+		printk(UM_KERN_ERR "copy_context_skas0 : put_fp_registers "
+		       "failed, pid = %d, err = %d\n", pid, err);
+		return err;
+	}
+
 	/* set a well known return code for detection of child write failure */
 	child_data->err = 12345678;
 
diff --git a/arch/um/sys-i386/shared/sysdep/ptrace.h b/arch/um/sys-i386/shared/sysdep/ptrace.h
index d50e62e..c398a50 100644
--- a/arch/um/sys-i386/shared/sysdep/ptrace.h
+++ b/arch/um/sys-i386/shared/sysdep/ptrace.h
@@ -53,6 +53,7 @@ extern int sysemu_supported;
 
 struct uml_pt_regs {
 	unsigned long gp[MAX_REG_NR];
+	unsigned long fp[HOST_FPX_SIZE];
 	struct faultinfo faultinfo;
 	long syscall;
 	int is_user;
diff --git a/arch/um/sys-x86_64/shared/sysdep/ptrace.h b/arch/um/sys-x86_64/shared/sysdep/ptrace.h
index fdba545..8ee8f8e 100644
--- a/arch/um/sys-x86_64/shared/sysdep/ptrace.h
+++ b/arch/um/sys-x86_64/shared/sysdep/ptrace.h
@@ -85,6 +85,7 @@
 
 struct uml_pt_regs {
 	unsigned long gp[MAX_REG_NR];
+	unsigned long fp[HOST_FP_SIZE];
 	struct faultinfo faultinfo;
 	long syscall;
 	int is_user;
-- 
1.7.6.1


------------------------------------------------------------------------------
EMC VNX: the world's simplest storage, starting under $10K
The only unified storage solution that offers unified management 
Up to 160% more powerful than alternatives and 25% more efficient. 
Guaranteed. http://p.sf.net/sfu/emc-vnx-dev2dev
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

[PATCH 4/9] um: fix oopsable race in line_close()

[Richard Weinberger]

From: Al Viro <viro@ftp.linux.org.uk>

tty->count is decremented only after ->close() had been called and several
tasks can hit it in parallel.  As the result, using tty->count to check if
you are the last one is broken.  We end up leaving line->tty not reset to
NULL and the next IRQ on that sucker will blow up trying to dereference
pointers from kfree'd struct tty.

Fix is obvious: we need to use a counter of our own.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Richard Weinberger <richard@nod.at>
---
 arch/um/drivers/line.c        |   25 ++++++++++++-------------
 arch/um/include/shared/line.h |    1 +
 2 files changed, 13 insertions(+), 13 deletions(-)

diff --git a/arch/um/drivers/line.c b/arch/um/drivers/line.c
index d51c404..c5bff1d 100644
--- a/arch/um/drivers/line.c
+++ b/arch/um/drivers/line.c
@@ -399,8 +399,8 @@ int line_setup_irq(int fd, int input, int output, struct line *line, void *data)
  * is done under a spinlock.  Checking whether the device is in use is
  * line->tty->count > 1, also under the spinlock.
  *
- * tty->count serves to decide whether the device should be enabled or
- * disabled on the host.  If it's equal to 1, then we are doing the
+ * line->count serves to decide whether the device should be enabled or
+ * disabled on the host.  If it's equal to 0, then we are doing the
  * first open or last close.  Otherwise, open and close just return.
  */
 
@@ -414,16 +414,16 @@ int line_open(struct line *lines, struct tty_struct *tty)
 		goto out_unlock;
 
 	err = 0;
-	if (tty->count > 1)
+	if (line->count++)
 		goto out_unlock;
 
-	spin_unlock(&line->count_lock);
-
+	BUG_ON(tty->driver_data);
 	tty->driver_data = line;
 	line->tty = tty;
 
+	spin_unlock(&line->count_lock);
 	err = enable_chan(line);
-	if (err)
+	if (err) /* line_close() will be called by our caller */
 		return err;
 
 	INIT_DELAYED_WORK(&line->task, line_timer_cb);
@@ -436,7 +436,7 @@ int line_open(struct line *lines, struct tty_struct *tty)
 	chan_window_size(&line->chan_list, &tty->winsize.ws_row,
 			 &tty->winsize.ws_col);
 
-	return err;
+	return 0;
 
 out_unlock:
 	spin_unlock(&line->count_lock);
@@ -460,17 +460,16 @@ void line_close(struct tty_struct *tty, struct file * filp)
 	flush_buffer(line);
 
 	spin_lock(&line->count_lock);
-	if (!line->valid)
-		goto out_unlock;
+	BUG_ON(!line->valid);
 
-	if (tty->count > 1)
+	if (--line->count)
 		goto out_unlock;
 
-	spin_unlock(&line->count_lock);
-
 	line->tty = NULL;
 	tty->driver_data = NULL;
 
+	spin_unlock(&line->count_lock);
+
 	if (line->sigio) {
 		unregister_winch(tty);
 		line->sigio = 0;
@@ -498,7 +497,7 @@ static int setup_one_line(struct line *lines, int n, char *init, int init_prio,
 
 	spin_lock(&line->count_lock);
 
-	if (line->tty != NULL) {
+	if (line->count) {
 		*error_out = "Device is already open";
 		goto out;
 	}
diff --git a/arch/um/include/shared/line.h b/arch/um/include/shared/line.h
index 72f4f25..63df3ca 100644
--- a/arch/um/include/shared/line.h
+++ b/arch/um/include/shared/line.h
@@ -33,6 +33,7 @@ struct line_driver {
 struct line {
 	struct tty_struct *tty;
 	spinlock_t count_lock;
+	unsigned long count;
 	int valid;
 
 	char *init_str;
-- 
1.7.6.1

[uml-devel] [PATCH 4/9] um: fix oopsable race in line_close()

[Richard Weinberger]

From: Al Viro <viro@ftp.linux.org.uk>

tty->count is decremented only after ->close() had been called and several
tasks can hit it in parallel.  As the result, using tty->count to check if
you are the last one is broken.  We end up leaving line->tty not reset to
NULL and the next IRQ on that sucker will blow up trying to dereference
pointers from kfree'd struct tty.

Fix is obvious: we need to use a counter of our own.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Richard Weinberger <richard@nod.at>
---
 arch/um/drivers/line.c        |   25 ++++++++++++-------------
 arch/um/include/shared/line.h |    1 +
 2 files changed, 13 insertions(+), 13 deletions(-)

diff --git a/arch/um/drivers/line.c b/arch/um/drivers/line.c
index d51c404..c5bff1d 100644
--- a/arch/um/drivers/line.c
+++ b/arch/um/drivers/line.c
@@ -399,8 +399,8 @@ int line_setup_irq(int fd, int input, int output, struct line *line, void *data)
  * is done under a spinlock.  Checking whether the device is in use is
  * line->tty->count > 1, also under the spinlock.
  *
- * tty->count serves to decide whether the device should be enabled or
- * disabled on the host.  If it's equal to 1, then we are doing the
+ * line->count serves to decide whether the device should be enabled or
+ * disabled on the host.  If it's equal to 0, then we are doing the
  * first open or last close.  Otherwise, open and close just return.
  */
 
@@ -414,16 +414,16 @@ int line_open(struct line *lines, struct tty_struct *tty)
 		goto out_unlock;
 
 	err = 0;
-	if (tty->count > 1)
+	if (line->count++)
 		goto out_unlock;
 
-	spin_unlock(&line->count_lock);
-
+	BUG_ON(tty->driver_data);
 	tty->driver_data = line;
 	line->tty = tty;
 
+	spin_unlock(&line->count_lock);
 	err = enable_chan(line);
-	if (err)
+	if (err) /* line_close() will be called by our caller */
 		return err;
 
 	INIT_DELAYED_WORK(&line->task, line_timer_cb);
@@ -436,7 +436,7 @@ int line_open(struct line *lines, struct tty_struct *tty)
 	chan_window_size(&line->chan_list, &tty->winsize.ws_row,
 			 &tty->winsize.ws_col);
 
-	return err;
+	return 0;
 
 out_unlock:
 	spin_unlock(&line->count_lock);
@@ -460,17 +460,16 @@ void line_close(struct tty_struct *tty, struct file * filp)
 	flush_buffer(line);
 
 	spin_lock(&line->count_lock);
-	if (!line->valid)
-		goto out_unlock;
+	BUG_ON(!line->valid);
 
-	if (tty->count > 1)
+	if (--line->count)
 		goto out_unlock;
 
-	spin_unlock(&line->count_lock);
-
 	line->tty = NULL;
 	tty->driver_data = NULL;
 
+	spin_unlock(&line->count_lock);
+
 	if (line->sigio) {
 		unregister_winch(tty);
 		line->sigio = 0;
@@ -498,7 +497,7 @@ static int setup_one_line(struct line *lines, int n, char *init, int init_prio,
 
 	spin_lock(&line->count_lock);
 
-	if (line->tty != NULL) {
+	if (line->count) {
 		*error_out = "Device is already open";
 		goto out;
 	}
diff --git a/arch/um/include/shared/line.h b/arch/um/include/shared/line.h
index 72f4f25..63df3ca 100644
--- a/arch/um/include/shared/line.h
+++ b/arch/um/include/shared/line.h
@@ -33,6 +33,7 @@ struct line_driver {
 struct line {
 	struct tty_struct *tty;
 	spinlock_t count_lock;
+	unsigned long count;
 	int valid;
 
 	char *init_str;
-- 
1.7.6.1


------------------------------------------------------------------------------
EMC VNX: the world's simplest storage, starting under $10K
The only unified storage solution that offers unified management 
Up to 160% more powerful than alternatives and 25% more efficient. 
Guaranteed. http://p.sf.net/sfu/emc-vnx-dev2dev
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

[PATCH 5/9] um: winch_interrupt() can happen inside of free_winch()

[Richard Weinberger]

From: Al Viro <viro@ftp.linux.org.uk>

... so set winch->fd to -1 before doing free_irq(), to avoid having
winch_interrupt() come from/during the latter and attempt to do
reactivate_fd() on something that's already gone.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Richard Weinberger <richard@nod.at>
---
 arch/um/drivers/line.c |    6 ++++--
 1 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/arch/um/drivers/line.c b/arch/um/drivers/line.c
index c5bff1d..91bf1894 100644
--- a/arch/um/drivers/line.c
+++ b/arch/um/drivers/line.c
@@ -725,6 +725,8 @@ struct winch {
 
 static void free_winch(struct winch *winch, int free_irq_ok)
 {
+	int fd = winch->fd;
+	winch->fd = -1;
 	if (free_irq_ok)
 		free_irq(WINCH_IRQ, winch);
 
@@ -732,8 +734,8 @@ static void free_winch(struct winch *winch, int free_irq_ok)
 
 	if (winch->pid != -1)
 		os_kill_process(winch->pid, 1);
-	if (winch->fd != -1)
-		os_close_file(winch->fd);
+	if (fd != -1)
+		os_close_file(fd);
 	if (winch->stack != 0)
 		free_stack(winch->stack, 0);
 	kfree(winch);
-- 
1.7.6.1

[uml-devel] [PATCH 5/9] um: winch_interrupt() can happen inside of free_winch()

[Richard Weinberger]

From: Al Viro <viro@ftp.linux.org.uk>

... so set winch->fd to -1 before doing free_irq(), to avoid having
winch_interrupt() come from/during the latter and attempt to do
reactivate_fd() on something that's already gone.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Richard Weinberger <richard@nod.at>
---
 arch/um/drivers/line.c |    6 ++++--
 1 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/arch/um/drivers/line.c b/arch/um/drivers/line.c
index c5bff1d..91bf1894 100644
--- a/arch/um/drivers/line.c
+++ b/arch/um/drivers/line.c
@@ -725,6 +725,8 @@ struct winch {
 
 static void free_winch(struct winch *winch, int free_irq_ok)
 {
+	int fd = winch->fd;
+	winch->fd = -1;
 	if (free_irq_ok)
 		free_irq(WINCH_IRQ, winch);
 
@@ -732,8 +734,8 @@ static void free_winch(struct winch *winch, int free_irq_ok)
 
 	if (winch->pid != -1)
 		os_kill_process(winch->pid, 1);
-	if (winch->fd != -1)
-		os_close_file(winch->fd);
+	if (fd != -1)
+		os_close_file(fd);
 	if (winch->stack != 0)
 		free_stack(winch->stack, 0);
 	kfree(winch);
-- 
1.7.6.1


------------------------------------------------------------------------------
EMC VNX: the world's simplest storage, starting under $10K
The only unified storage solution that offers unified management 
Up to 160% more powerful than alternatives and 25% more efficient. 
Guaranteed. http://p.sf.net/sfu/emc-vnx-dev2dev
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

[PATCH 6/9] um: fix free_winch() mess

[Richard Weinberger]

From: Al Viro <viro@ftp.linux.org.uk>

while not doing free_irq() from irq handler is commendable, kfree() on
the data passed to said handler before free_irq() is Not Good(tm).
Freeing the stack it's being run on is also not nice...  Solution:
delay actually freeing stuff.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Richard Weinberger <richard@nod.at>
---
 arch/um/drivers/line.c |   38 ++++++++++++++++++++++++--------------
 1 files changed, 24 insertions(+), 14 deletions(-)

diff --git a/arch/um/drivers/line.c b/arch/um/drivers/line.c
index 91bf1894..364c8a1 100644
--- a/arch/um/drivers/line.c
+++ b/arch/um/drivers/line.c
@@ -721,43 +721,53 @@ struct winch {
 	int pid;
 	struct tty_struct *tty;
 	unsigned long stack;
+	struct work_struct work;
 };
 
-static void free_winch(struct winch *winch, int free_irq_ok)
+static void __free_winch(struct work_struct *work)
 {
-	int fd = winch->fd;
-	winch->fd = -1;
-	if (free_irq_ok)
-		free_irq(WINCH_IRQ, winch);
-
-	list_del(&winch->list);
+	struct winch *winch = container_of(work, struct winch, work);
+	free_irq(WINCH_IRQ, winch);
 
 	if (winch->pid != -1)
 		os_kill_process(winch->pid, 1);
-	if (fd != -1)
-		os_close_file(fd);
 	if (winch->stack != 0)
 		free_stack(winch->stack, 0);
 	kfree(winch);
 }
 
+static void free_winch(struct winch *winch)
+{
+	int fd = winch->fd;
+	winch->fd = -1;
+	if (fd != -1)
+		os_close_file(fd);
+	list_del(&winch->list);
+	__free_winch(&winch->work);
+}
+
 static irqreturn_t winch_interrupt(int irq, void *data)
 {
 	struct winch *winch = data;
 	struct tty_struct *tty;
 	struct line *line;
+	int fd = winch->fd;
 	int err;
 	char c;
 
-	if (winch->fd != -1) {
-		err = generic_read(winch->fd, &c, NULL);
+	if (fd != -1) {
+		err = generic_read(fd, &c, NULL);
 		if (err < 0) {
 			if (err != -EAGAIN) {
+				winch->fd = -1;
+				list_del(&winch->list);
+				os_close_file(fd);
 				printk(KERN_ERR "winch_interrupt : "
 				       "read failed, errno = %d\n", -err);
 				printk(KERN_ERR "fd %d is losing SIGWINCH "
 				       "support\n", winch->tty_fd);
-				free_winch(winch, 0);
+				INIT_WORK(&winch->work, __free_winch);
+				schedule_work(&winch->work);
 				return IRQ_HANDLED;
 			}
 			goto out;
@@ -829,7 +839,7 @@ static void unregister_winch(struct tty_struct *tty)
 	list_for_each_safe(ele, next, &winch_handlers) {
 		winch = list_entry(ele, struct winch, list);
 		if (winch->tty == tty) {
-			free_winch(winch, 1);
+			free_winch(winch);
 			break;
 		}
 	}
@@ -845,7 +855,7 @@ static void winch_cleanup(void)
 
 	list_for_each_safe(ele, next, &winch_handlers) {
 		winch = list_entry(ele, struct winch, list);
-		free_winch(winch, 1);
+		free_winch(winch);
 	}
 
 	spin_unlock(&winch_handler_lock);
-- 
1.7.6.1

[uml-devel] [PATCH 6/9] um: fix free_winch() mess

[Richard Weinberger]

From: Al Viro <viro@ftp.linux.org.uk>

while not doing free_irq() from irq handler is commendable, kfree() on
the data passed to said handler before free_irq() is Not Good(tm).
Freeing the stack it's being run on is also not nice...  Solution:
delay actually freeing stuff.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Richard Weinberger <richard@nod.at>
---
 arch/um/drivers/line.c |   38 ++++++++++++++++++++++++--------------
 1 files changed, 24 insertions(+), 14 deletions(-)

diff --git a/arch/um/drivers/line.c b/arch/um/drivers/line.c
index 91bf1894..364c8a1 100644
--- a/arch/um/drivers/line.c
+++ b/arch/um/drivers/line.c
@@ -721,43 +721,53 @@ struct winch {
 	int pid;
 	struct tty_struct *tty;
 	unsigned long stack;
+	struct work_struct work;
 };
 
-static void free_winch(struct winch *winch, int free_irq_ok)
+static void __free_winch(struct work_struct *work)
 {
-	int fd = winch->fd;
-	winch->fd = -1;
-	if (free_irq_ok)
-		free_irq(WINCH_IRQ, winch);
-
-	list_del(&winch->list);
+	struct winch *winch = container_of(work, struct winch, work);
+	free_irq(WINCH_IRQ, winch);
 
 	if (winch->pid != -1)
 		os_kill_process(winch->pid, 1);
-	if (fd != -1)
-		os_close_file(fd);
 	if (winch->stack != 0)
 		free_stack(winch->stack, 0);
 	kfree(winch);
 }
 
+static void free_winch(struct winch *winch)
+{
+	int fd = winch->fd;
+	winch->fd = -1;
+	if (fd != -1)
+		os_close_file(fd);
+	list_del(&winch->list);
+	__free_winch(&winch->work);
+}
+
 static irqreturn_t winch_interrupt(int irq, void *data)
 {
 	struct winch *winch = data;
 	struct tty_struct *tty;
 	struct line *line;
+	int fd = winch->fd;
 	int err;
 	char c;
 
-	if (winch->fd != -1) {
-		err = generic_read(winch->fd, &c, NULL);
+	if (fd != -1) {
+		err = generic_read(fd, &c, NULL);
 		if (err < 0) {
 			if (err != -EAGAIN) {
+				winch->fd = -1;
+				list_del(&winch->list);
+				os_close_file(fd);
 				printk(KERN_ERR "winch_interrupt : "
 				       "read failed, errno = %d\n", -err);
 				printk(KERN_ERR "fd %d is losing SIGWINCH "
 				       "support\n", winch->tty_fd);
-				free_winch(winch, 0);
+				INIT_WORK(&winch->work, __free_winch);
+				schedule_work(&winch->work);
 				return IRQ_HANDLED;
 			}
 			goto out;
@@ -829,7 +839,7 @@ static void unregister_winch(struct tty_struct *tty)
 	list_for_each_safe(ele, next, &winch_handlers) {
 		winch = list_entry(ele, struct winch, list);
 		if (winch->tty == tty) {
-			free_winch(winch, 1);
+			free_winch(winch);
 			break;
 		}
 	}
@@ -845,7 +855,7 @@ static void winch_cleanup(void)
 
 	list_for_each_safe(ele, next, &winch_handlers) {
 		winch = list_entry(ele, struct winch, list);
-		free_winch(winch, 1);
+		free_winch(winch);
 	}
 
 	spin_unlock(&winch_handler_lock);
-- 
1.7.6.1


------------------------------------------------------------------------------
EMC VNX: the world's simplest storage, starting under $10K
The only unified storage solution that offers unified management 
Up to 160% more powerful than alternatives and 25% more efficient. 
Guaranteed. http://p.sf.net/sfu/emc-vnx-dev2dev
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

[PATCH 7/9] um: PTRACE_[GS]ETFPXREGS had been wired on the wrong subarch

[Richard Weinberger]

From: Al Viro <viro@ftp.linux.org.uk>

It's 32bit-only, not 64bit-only...  And while we are at it, it's
set_fpxregs(), not set_fpregs()...

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Richard Weinberger <richard@nod.at>
---
 arch/um/sys-i386/ptrace.c   |   14 +++++++++++++-
 arch/um/sys-x86_64/ptrace.c |   14 +-------------
 2 files changed, 14 insertions(+), 14 deletions(-)

diff --git a/arch/um/sys-i386/ptrace.c b/arch/um/sys-i386/ptrace.c
index d23b2d3..49fd25a 100644
--- a/arch/um/sys-i386/ptrace.c
+++ b/arch/um/sys-i386/ptrace.c
@@ -206,5 +206,17 @@ int set_fpxregs(struct user_fxsr_struct __user *buf, struct task_struct *child)
 long subarch_ptrace(struct task_struct *child, long request,
 		    unsigned long addr, unsigned long data)
 {
-	return -EIO;
+	int ret = -EIO;
+	void __user *datap = (void __user *) data;
+
+	switch (request) {
+	case PTRACE_GETFPXREGS: /* Get the child FPU state. */
+		ret = get_fpxregs(datap, child);
+		break;
+	case PTRACE_SETFPXREGS: /* Set the child FPU state. */
+		ret = set_fpxregs(datap, child);
+		break;
+	}
+
+	return ret;
 }
diff --git a/arch/um/sys-x86_64/ptrace.c b/arch/um/sys-x86_64/ptrace.c
index f436136..7ed49ac 100644
--- a/arch/um/sys-x86_64/ptrace.c
+++ b/arch/um/sys-x86_64/ptrace.c
@@ -178,17 +178,5 @@ int set_fpregs(struct user_i387_struct __user *buf, struct task_struct *child)
 long subarch_ptrace(struct task_struct *child, long request,
 		    unsigned long addr, unsigned long data)
 {
-	int ret = -EIO;
-	void __user *datap = (void __user *) data;
-
-	switch (request) {
-	case PTRACE_GETFPXREGS: /* Get the child FPU state. */
-		ret = get_fpregs(datap, child);
-		break;
-	case PTRACE_SETFPXREGS: /* Set the child FPU state. */
-		ret = set_fpregs(datap, child);
-		break;
-	}
-
-	return ret;
+	return -EIO;
 }
-- 
1.7.6.1

[uml-devel] [PATCH 7/9] um: PTRACE_[GS]ETFPXREGS had been wired on the wrong subarch

[Richard Weinberger]

From: Al Viro <viro@ftp.linux.org.uk>

It's 32bit-only, not 64bit-only...  And while we are at it, it's
set_fpxregs(), not set_fpregs()...

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Richard Weinberger <richard@nod.at>
---
 arch/um/sys-i386/ptrace.c   |   14 +++++++++++++-
 arch/um/sys-x86_64/ptrace.c |   14 +-------------
 2 files changed, 14 insertions(+), 14 deletions(-)

diff --git a/arch/um/sys-i386/ptrace.c b/arch/um/sys-i386/ptrace.c
index d23b2d3..49fd25a 100644
--- a/arch/um/sys-i386/ptrace.c
+++ b/arch/um/sys-i386/ptrace.c
@@ -206,5 +206,17 @@ int set_fpxregs(struct user_fxsr_struct __user *buf, struct task_struct *child)
 long subarch_ptrace(struct task_struct *child, long request,
 		    unsigned long addr, unsigned long data)
 {
-	return -EIO;
+	int ret = -EIO;
+	void __user *datap = (void __user *) data;
+
+	switch (request) {
+	case PTRACE_GETFPXREGS: /* Get the child FPU state. */
+		ret = get_fpxregs(datap, child);
+		break;
+	case PTRACE_SETFPXREGS: /* Set the child FPU state. */
+		ret = set_fpxregs(datap, child);
+		break;
+	}
+
+	return ret;
 }
diff --git a/arch/um/sys-x86_64/ptrace.c b/arch/um/sys-x86_64/ptrace.c
index f436136..7ed49ac 100644
--- a/arch/um/sys-x86_64/ptrace.c
+++ b/arch/um/sys-x86_64/ptrace.c
@@ -178,17 +178,5 @@ int set_fpregs(struct user_i387_struct __user *buf, struct task_struct *child)
 long subarch_ptrace(struct task_struct *child, long request,
 		    unsigned long addr, unsigned long data)
 {
-	int ret = -EIO;
-	void __user *datap = (void __user *) data;
-
-	switch (request) {
-	case PTRACE_GETFPXREGS: /* Get the child FPU state. */
-		ret = get_fpregs(datap, child);
-		break;
-	case PTRACE_SETFPXREGS: /* Set the child FPU state. */
-		ret = set_fpregs(datap, child);
-		break;
-	}
-
-	return ret;
+	return -EIO;
 }
-- 
1.7.6.1


------------------------------------------------------------------------------
EMC VNX: the world's simplest storage, starting under $10K
The only unified storage solution that offers unified management 
Up to 160% more powerful than alternatives and 25% more efficient. 
Guaranteed. http://p.sf.net/sfu/emc-vnx-dev2dev
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

[PATCH 3/9] um: Save FPU registers between task switches

[Richard Weinberger]

From: Ingo van Lil <inguin@gmx.de>

Some time ago Jeff prepared a patch (42daba3) for UML to stop saving the
process FP state between task switches. The assumption was that since
with SKAS0 every guest process runs inside a host process context the
host OS will take care of keeping the proper FP state. Unfortunately
this is not true for multi-threaded applications, where all guest
threads share a single host process context yet all may use the FPU on
their own. Although I haven't verified it I suspect things to be even
worse in SKAS3 mode where all guest processes run inside a single host
process.

The patch below reintroduces the saving and restoring of the FP context
between task switches.

Signed-off-by: Ingo van Lil <inguin@gmx.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
Reported-by: gunnarlindroth@hotmail.com
Tested-by: gunnarlindroth@hotmail.com
CC: Stanislav Meduna <stano@meduna.org>

[richard@nod.at: Ingo posted this patch in 2009, sadly it was never applied
and got lost. Now in 2011 the problem was reported by Gunnar.]
---
 arch/um/include/shared/registers.h        |    2 +-
 arch/um/kernel/process.c                  |    2 +-
 arch/um/os-Linux/registers.c              |    9 ++++++++-
 arch/um/os-Linux/skas/mem.c               |    2 +-
 arch/um/os-Linux/skas/process.c           |   19 ++++++++++++++++++-
 arch/um/sys-i386/shared/sysdep/ptrace.h   |    1 +
 arch/um/sys-x86_64/shared/sysdep/ptrace.h |    1 +
 7 files changed, 31 insertions(+), 5 deletions(-)

diff --git a/arch/um/include/shared/registers.h b/arch/um/include/shared/registers.h
index b0b4589..f1e0aa5 100644
--- a/arch/um/include/shared/registers.h
+++ b/arch/um/include/shared/registers.h
@@ -16,7 +16,7 @@ extern int restore_fpx_registers(int pid, unsigned long *fp_regs);
 extern int save_registers(int pid, struct uml_pt_regs *regs);
 extern int restore_registers(int pid, struct uml_pt_regs *regs);
 extern int init_registers(int pid);
-extern void get_safe_registers(unsigned long *regs);
+extern void get_safe_registers(unsigned long *regs, unsigned long *fp_regs);
 extern unsigned long get_thread_reg(int reg, jmp_buf *buf);
 extern int get_fp_registers(int pid, unsigned long *regs);
 extern int put_fp_registers(int pid, unsigned long *regs);
diff --git a/arch/um/kernel/process.c b/arch/um/kernel/process.c
index fab4371..21c1ae7 100644
--- a/arch/um/kernel/process.c
+++ b/arch/um/kernel/process.c
@@ -202,7 +202,7 @@ int copy_thread(unsigned long clone_flags, unsigned long sp,
 		arch_copy_thread(&current->thread.arch, &p->thread.arch);
 	}
 	else {
-		get_safe_registers(p->thread.regs.regs.gp);
+		get_safe_registers(p->thread.regs.regs.gp, p->thread.regs.regs.fp);
 		p->thread.request.u.thread = current->thread.request.u.thread;
 		handler = new_thread_handler;
 	}
diff --git a/arch/um/os-Linux/registers.c b/arch/um/os-Linux/registers.c
index 830fe6a..b866b9e 100644
--- a/arch/um/os-Linux/registers.c
+++ b/arch/um/os-Linux/registers.c
@@ -8,6 +8,8 @@
 #include <string.h>
 #include <sys/ptrace.h>
 #include "sysdep/ptrace.h"
+#include "sysdep/ptrace_user.h"
+#include "registers.h"
 
 int save_registers(int pid, struct uml_pt_regs *regs)
 {
@@ -32,6 +34,7 @@ int restore_registers(int pid, struct uml_pt_regs *regs)
 /* This is set once at boot time and not changed thereafter */
 
 static unsigned long exec_regs[MAX_REG_NR];
+static unsigned long exec_fp_regs[FP_SIZE];
 
 int init_registers(int pid)
 {
@@ -42,10 +45,14 @@ int init_registers(int pid)
 		return -errno;
 
 	arch_init_registers(pid);
+	get_fp_registers(pid, exec_fp_regs);
 	return 0;
 }
 
-void get_safe_registers(unsigned long *regs)
+void get_safe_registers(unsigned long *regs, unsigned long *fp_regs)
 {
 	memcpy(regs, exec_regs, sizeof(exec_regs));
+
+	if (fp_regs)
+		memcpy(fp_regs, exec_fp_regs, sizeof(exec_fp_regs));
 }
diff --git a/arch/um/os-Linux/skas/mem.c b/arch/um/os-Linux/skas/mem.c
index d261f17..e771398 100644
--- a/arch/um/os-Linux/skas/mem.c
+++ b/arch/um/os-Linux/skas/mem.c
@@ -39,7 +39,7 @@ static unsigned long syscall_regs[MAX_REG_NR];
 
 static int __init init_syscall_regs(void)
 {
-	get_safe_registers(syscall_regs);
+	get_safe_registers(syscall_regs, NULL);
 	syscall_regs[REGS_IP_INDEX] = STUB_CODE +
 		((unsigned long) &batch_syscall_stub -
 		 (unsigned long) &__syscall_stub_start);
diff --git a/arch/um/os-Linux/skas/process.c b/arch/um/os-Linux/skas/process.c
index d6e0a22..dee0e8c 100644
--- a/arch/um/os-Linux/skas/process.c
+++ b/arch/um/os-Linux/skas/process.c
@@ -373,6 +373,9 @@ void userspace(struct uml_pt_regs *regs)
 		if (ptrace(PTRACE_SETREGS, pid, 0, regs->gp))
 			fatal_sigsegv();
 
+		if (put_fp_registers(pid, regs->fp))
+			fatal_sigsegv();
+
 		/* Now we set local_using_sysemu to be used for one loop */
 		local_using_sysemu = get_using_sysemu();
 
@@ -399,6 +402,12 @@ void userspace(struct uml_pt_regs *regs)
 			fatal_sigsegv();
 		}
 
+		if (get_fp_registers(pid, regs->fp)) {
+			printk(UM_KERN_ERR "userspace -  get_fp_registers failed, "
+			       "errno = %d\n", errno);
+			fatal_sigsegv();
+		}
+
 		UPT_SYSCALL_NR(regs) = -1; /* Assume: It's not a syscall */
 
 		if (WIFSTOPPED(status)) {
@@ -457,10 +466,11 @@ void userspace(struct uml_pt_regs *regs)
 }
 
 static unsigned long thread_regs[MAX_REG_NR];
+static unsigned long thread_fp_regs[FP_SIZE];
 
 static int __init init_thread_regs(void)
 {
-	get_safe_registers(thread_regs);
+	get_safe_registers(thread_regs, thread_fp_regs);
 	/* Set parent's instruction pointer to start of clone-stub */
 	thread_regs[REGS_IP_INDEX] = STUB_CODE +
 				(unsigned long) stub_clone_handler -
@@ -503,6 +513,13 @@ int copy_context_skas0(unsigned long new_stack, int pid)
 		return err;
 	}
 
+	err = put_fp_registers(pid, thread_fp_regs);
+	if (err < 0) {
+		printk(UM_KERN_ERR "copy_context_skas0 : put_fp_registers "
+		       "failed, pid = %d, err = %d\n", pid, err);
+		return err;
+	}
+
 	/* set a well known return code for detection of child write failure */
 	child_data->err = 12345678;
 
diff --git a/arch/um/sys-i386/shared/sysdep/ptrace.h b/arch/um/sys-i386/shared/sysdep/ptrace.h
index d50e62e..c398a50 100644
--- a/arch/um/sys-i386/shared/sysdep/ptrace.h
+++ b/arch/um/sys-i386/shared/sysdep/ptrace.h
@@ -53,6 +53,7 @@ extern int sysemu_supported;
 
 struct uml_pt_regs {
 	unsigned long gp[MAX_REG_NR];
+	unsigned long fp[HOST_FPX_SIZE];
 	struct faultinfo faultinfo;
 	long syscall;
 	int is_user;
diff --git a/arch/um/sys-x86_64/shared/sysdep/ptrace.h b/arch/um/sys-x86_64/shared/sysdep/ptrace.h
index fdba545..8ee8f8e 100644
--- a/arch/um/sys-x86_64/shared/sysdep/ptrace.h
+++ b/arch/um/sys-x86_64/shared/sysdep/ptrace.h
@@ -85,6 +85,7 @@
 
 struct uml_pt_regs {
 	unsigned long gp[MAX_REG_NR];
+	unsigned long fp[HOST_FP_SIZE];
 	struct faultinfo faultinfo;
 	long syscall;
 	int is_user;
-- 
1.7.6.1