linux-toradex.git/arch/x86/kernel/irq_64.c, branch v3.8.2

x86: Add stack top margin for stack overflow checking

2011-12-07T08:27:11+00:00

It seems that a margin for stack overflow checking is added to
top of a kernel stack but is not added to IRQ and exception
stacks in stack_overflow_check(). Therefore, the overflows of
IRQ and exception stacks are always detected only after they
actually occurred and data corruption might occur due to them.

This patch adds the margin to top of IRQ and exception stacks
as well as a kernel stack to enhance reliability.

Signed-off-by: Mitsuo Hayasaka 
Cc: yrl.pp-manager.tt@hitachi.com
Cc: Linus Torvalds 
Cc: Andrew Morton 
Link: http://lkml.kernel.org/r/20111207082910.9847.3359.stgit@ltc219.sdl.hitachi.co.jp
[ removed the #undef - we typically don't do that for uncommon names ]
Signed-off-by: Ingo Molnar

x86: Clean up the range of stack overflow checking

2011-12-05T10:37:48+00:00

The overflow checking of kernel stack checks if the stack
pointer points to the available kernel stack range, which is
derived from the original overflow checking.

It is clear that curbase address is always less than low
boundary of available kernel stack. So, this patch removes the
first condition that checks if the pointer is higher than
curbase.

Signed-off-by: Mitsuo Hayasaka 
Cc: yrl.pp-manager.tt@hitachi.com
Cc: Randy Dunlap 
Link: http://lkml.kernel.org/r/20111129060845.11076.40916.stgit@ltc219.sdl.hitachi.co.jp
Signed-off-by: Ingo Molnar 
Cc: "H. Peter Anvin"

x86: Panic on detection of stack overflow

2011-12-05T10:37:47+00:00

Currently, messages are just output on the detection of stack
overflow, which is not sufficient for systems that need a
high reliability. This is because in general the overflow may
corrupt data, and the additional corruption may occur due to
reading them unless systems stop.

This patch adds the sysctl parameter
kernel.panic_on_stackoverflow and causes a panic when detecting
the overflows of kernel, IRQ and exception stacks except user
stack according to the parameter. It is disabled by default.

Signed-off-by: Mitsuo Hayasaka 
Cc: yrl.pp-manager.tt@hitachi.com
Cc: Randy Dunlap 
Cc: "H. Peter Anvin" 
Link: http://lkml.kernel.org/r/20111129060836.11076.12323.stgit@ltc219.sdl.hitachi.co.jp
Signed-off-by: Ingo Molnar

x86: Check stack overflow in detail

2011-12-05T10:37:45+00:00

Currently, only kernel stack is checked for the overflow, which
is not sufficient for systems that need a high reliability. To
enhance it, it is required to check the IRQ and exception
stacks, as well.

This patch checks all the stack types and will cause messages of
stacks in detail when free stack space drops below a certain
limit except user stack.

Signed-off-by: Mitsuo Hayasaka 
Cc: yrl.pp-manager.tt@hitachi.com
Cc: Randy Dunlap 
Link: http://lkml.kernel.org/r/20111129060829.11076.51733.stgit@ltc219.sdl.hitachi.co.jp
Signed-off-by: Ingo Molnar 
Cc: "H. Peter Anvin"

x86: Add user_mode_vm check in stack_overflow_check

2011-12-05T10:28:25+00:00

The kernel stack overflow is checked in stack_overflow_check(),
which may wrongly detect the overflow if the stack pointer in
user space points to the kernel stack intentionally or
accidentally. So, the actual overflow is never detected after
this misdetection because WARN_ONCE() is used on the detection
of it.

This patch adds user-mode-vm checking before it to avoid this
problem and bails out early if the user stack is used.

Signed-off-by: Mitsuo Hayasaka 
Cc: yrl.pp-manager.tt@hitachi.com
Cc: Randy Dunlap 
Link: http://lkml.kernel.org/r/20111129060821.11076.55315.stgit@ltc219.sdl.hitachi.co.jp
Signed-off-by: Ingo Molnar 
Cc: "H. Peter Anvin"

x86: Unify fixup_irqs() for 32-bit and 64-bit kernels

2009-11-02T14:56:34+00:00

There is no reason to have different fixup_irqs() for 32-bit and
64-bit kernels. Unify by using the superior 64-bit version for
both the kernels.

Signed-off-by: Suresh Siddha 
Signed-off-by: Gary Hade 
Cc: Eric W. Biederman 
LKML-Reference: <20091026230001.562512739@sbs-t61.sc.intel.com>
Signed-off-by: Ingo Molnar

x86: unify do_IRQ()

2009-02-09T11:16:05+00:00

With the differences in interrupt handling hoisted into handle_irq(),
do_IRQ is more or less identical between 32 and 64 bit, so unify it.

Signed-off-by: Jeremy Fitzhardinge 
Signed-off-by: Ingo Molnar

x86: add handle_irq() to allow interrupt injection

2009-02-09T11:15:57+00:00

Xen uses a different interrupt path, so introduce handle_irq() to
allow interrupts to be inserted into the normal interrupt path.  This
is handled slightly differently on 32 and 64-bit.

Signed-off-by: Jeremy Fitzhardinge 
Signed-off-by: Ingo Molnar

x86: remove include of apic.h from hardirq_64.h

2009-01-23T02:03:29+00:00

Impact: cleanup

APIC definitions aren't needed here.  Remove the include and fix
up the fallout.

tj: added include to mce_intel_64.c.

Signed-off-by: Brian Gerst 
Signed-off-by: Tejun Heo

x86: merge irq_regs.h

2009-01-21T08:26:06+00:00

Impact: cleanup, better irq_regs code generation for x86_64

Make 64-bit use the same optimizations as 32-bit.

Signed-off-by: Brian Gerst 
Signed-off-by: Tejun Heo