[PATCH v2 0/2] context tracker support for arm64

[PATCH v2 0/2] context tracker support for arm64

[Larry Bassel]

Implement and enable context tracking for arm64 (which is
a prerequisite for FULL_NOHZ support). This patchset
builds upon earlier work by Kevin Hilman and is based on 3.15-rc2.

Larry Bassel (2):
  arm64: adjust el0_sync so that a function can be called
  arm64: enable context tracking

 arch/arm64/Kconfig                   |  1 +
 arch/arm64/include/asm/thread_info.h |  1 +
 arch/arm64/kernel/entry.S            | 36 +++++++++++++++++++++++++++++++-----
 3 files changed, 33 insertions(+), 5 deletions(-)

-- 
1.8.3.2

[PATCH v2 1/2] arm64: adjust el0_sync so that a function can be called

[Larry Bassel]

To implement the context tracker properly on arm64,
a function call needs to be made after debugging and
interrupts are turned on, but before the lr is changed
to point to ret_from_exception(). If the function call
is made after the lr is changed the function will not
return to the correct place.

For similar reasons, defer the setting of x0 so that
it doesn't need to be saved around the function call
(save far_el1 in x26 temporarily instead).

Signed-off-by: Larry Bassel <larry.bassel@linaro.org>
Reviewed-by: Kevin Hilman <khilman@linaro.org>
---
 arch/arm64/kernel/entry.S | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)

diff --git a/arch/arm64/kernel/entry.S b/arch/arm64/kernel/entry.S
index 39ac630..d920d7f 100644
--- a/arch/arm64/kernel/entry.S
+++ b/arch/arm64/kernel/entry.S
@@ -349,11 +349,11 @@ el0_sync:
 	lsr	x24, x25, #ESR_EL1_EC_SHIFT	// exception class
 	cmp	x24, #ESR_EL1_EC_SVC64		// SVC in 64-bit state
 	b.eq	el0_svc
-	adr	lr, ret_from_exception
 	cmp	x24, #ESR_EL1_EC_DABT_EL0	// data abort in EL0
 	b.eq	el0_da
 	cmp	x24, #ESR_EL1_EC_IABT_EL0	// instruction abort in EL0
 	b.eq	el0_ia
+	adr	lr, ret_from_exception
 	cmp	x24, #ESR_EL1_EC_FP_ASIMD	// FP/ASIMD access
 	b.eq	el0_fpsimd_acc
 	cmp	x24, #ESR_EL1_EC_FP_EXC64	// FP/ASIMD exception
@@ -378,11 +378,11 @@ el0_sync_compat:
 	lsr	x24, x25, #ESR_EL1_EC_SHIFT	// exception class
 	cmp	x24, #ESR_EL1_EC_SVC32		// SVC in 32-bit state
 	b.eq	el0_svc_compat
-	adr	lr, ret_from_exception
 	cmp	x24, #ESR_EL1_EC_DABT_EL0	// data abort in EL0
 	b.eq	el0_da
 	cmp	x24, #ESR_EL1_EC_IABT_EL0	// instruction abort in EL0
 	b.eq	el0_ia
+	adr	lr, ret_from_exception
 	cmp	x24, #ESR_EL1_EC_FP_ASIMD	// FP/ASIMD access
 	b.eq	el0_fpsimd_acc
 	cmp	x24, #ESR_EL1_EC_FP_EXC32	// FP/ASIMD exception
@@ -421,28 +421,32 @@ el0_da:
 	/*
 	 * Data abort handling
 	 */
-	mrs	x0, far_el1
-	bic	x0, x0, #(0xff << 56)
+	mrs	x26, far_el1
 	disable_step x1
 	isb
 	enable_dbg
 	// enable interrupts before calling the main handler
 	enable_irq
+	mov	x0, x26
+	bic	x0, x0, #(0xff << 56)
 	mov	x1, x25
 	mov	x2, sp
+	adr	lr, ret_from_exception
 	b	do_mem_abort
 el0_ia:
 	/*
 	 * Instruction abort handling
 	 */
-	mrs	x0, far_el1
+	mrs	x26, far_el1
 	disable_step x1
 	isb
 	enable_dbg
 	// enable interrupts before calling the main handler
 	enable_irq
+	mov	x0, x26
 	orr	x1, x25, #1 << 24		// use reserved ISS bit for instruction aborts
 	mov	x2, sp
+	adr	lr, ret_from_exception
 	b	do_mem_abort
 el0_fpsimd_acc:
 	/*
-- 
1.8.3.2

[PATCH v2 2/2] arm64: enable context tracking

[Larry Bassel]

Make calls to ct_user_enter when the kernel is exited
and ct_user_exit when the kernel is entered (in el0_da,
el0_ia, el0_svc, el0_irq).

These macros expand to function calls which will only work
properly if el0_sync and related code has been rearranged
(in a previous patch of this series).

In order to avoid saving registers, the slow syscall path
is forced (as x86 does).

The calls to ct_user_exit are made after hw debugging has been
enabled (enable_dbg).

The call to ct_user_enter is made at the beginning of the
kernel_exit macro.

This patch is based on earlier work by Kevin Hilman.

Signed-off-by: Kevin Hilman <khilman@linaro.org>
Signed-off-by: Larry Bassel <larry.bassel@linaro.org>
---
 arch/arm64/Kconfig                   |  1 +
 arch/arm64/include/asm/thread_info.h |  1 +
 arch/arm64/kernel/entry.S            | 22 ++++++++++++++++++++++
 3 files changed, 24 insertions(+)

diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
index e6e4d37..152d92b 100644
--- a/arch/arm64/Kconfig
+++ b/arch/arm64/Kconfig
@@ -55,6 +55,7 @@ config ARM64
 	select RTC_LIB
 	select SPARSE_IRQ
 	select SYSCTL_EXCEPTION_TRACE
+	select HAVE_CONTEXT_TRACKING
 	help
 	  ARM 64-bit (AArch64) Linux support.
 
diff --git a/arch/arm64/include/asm/thread_info.h b/arch/arm64/include/asm/thread_info.h
index 720e70b..301ea6a 100644
--- a/arch/arm64/include/asm/thread_info.h
+++ b/arch/arm64/include/asm/thread_info.h
@@ -108,6 +108,7 @@ static inline struct thread_info *current_thread_info(void)
 #define TIF_SINGLESTEP		21
 #define TIF_32BIT		22	/* 32bit process */
 #define TIF_SWITCH_MM		23	/* deferred switch_mm */
+#define TIF_NOHZ                24
 
 #define _TIF_SIGPENDING		(1 << TIF_SIGPENDING)
 #define _TIF_NEED_RESCHED	(1 << TIF_NEED_RESCHED)
diff --git a/arch/arm64/kernel/entry.S b/arch/arm64/kernel/entry.S
index d920d7f..5fe447c 100644
--- a/arch/arm64/kernel/entry.S
+++ b/arch/arm64/kernel/entry.S
@@ -30,6 +30,22 @@
 #include <asm/unistd32.h>
 
 /*
+ * Context tracking subsystem.  Used to instrument transitions
+ * between user and kernel mode.
+ */
+	.macro ct_user_exit
+#ifdef CONFIG_CONTEXT_TRACKING
+	bl	context_tracking_user_exit
+#endif
+	.endm
+
+	.macro ct_user_enter
+#ifdef CONFIG_CONTEXT_TRACKING
+	bl	context_tracking_user_enter
+#endif
+	.endm
+
+/*
  * Bad Abort numbers
  *-----------------
  */
@@ -88,6 +104,7 @@
 	.macro	kernel_exit, el, ret = 0
 	ldp	x21, x22, [sp, #S_PC]		// load ELR, SPSR
 	.if	\el == 0
+	ct_user_enter
 	ldr	x23, [sp, #S_SP]		// load return stack pointer
 	.endif
 	.if	\ret
@@ -427,6 +444,7 @@ el0_da:
 	enable_dbg
 	// enable interrupts before calling the main handler
 	enable_irq
+	ct_user_exit
 	mov	x0, x26
 	bic	x0, x0, #(0xff << 56)
 	mov	x1, x25
@@ -443,6 +461,7 @@ el0_ia:
 	enable_dbg
 	// enable interrupts before calling the main handler
 	enable_irq
+	ct_user_exit
 	mov	x0, x26
 	orr	x1, x25, #1 << 24		// use reserved ISS bit for instruction aborts
 	mov	x2, sp
@@ -511,6 +530,7 @@ el0_irq_naked:
 	bl	trace_hardirqs_off
 #endif
 
+	ct_user_exit
 	irq_handler
 	get_thread_info tsk
 
@@ -633,10 +653,12 @@ el0_svc_naked:					// compat entry point
 	isb
 	enable_dbg
 	enable_irq
+	ct_user_exit
 
 	get_thread_info tsk
 	ldr	x16, [tsk, #TI_FLAGS]		// check for syscall tracing
 	tbnz	x16, #TIF_SYSCALL_TRACE, __sys_trace // are we tracing syscalls?
+	tbnz	x16, #TIF_NOHZ, __sys_trace
 	adr	lr, ret_fast_syscall		// return address
 	cmp     scno, sc_nr                     // check upper syscall limit
 	b.hs	ni_sys
-- 
1.8.3.2

[PATCH v2 2/2] arm64: enable context tracking

[Will Deacon]

On Wed, May 07, 2014 at 12:32:29AM +0100, Larry Bassel wrote:
> Make calls to ct_user_enter when the kernel is exited
> and ct_user_exit when the kernel is entered (in el0_da,
> el0_ia, el0_svc, el0_irq).

Why only these entry points? I can reschedule after any exception from EL0,
so I'd expect all exceptions from userspace to need annotating, no?

> These macros expand to function calls which will only work
> properly if el0_sync and related code has been rearranged
> (in a previous patch of this series).
> 
> In order to avoid saving registers, the slow syscall path
> is forced (as x86 does).

... and if you decide to handle undef exceptions, I think you'll need
the register saving too, in case the kernel needs to perform emulation.

Will

[PATCH v2 2/2] arm64: enable context tracking

[Larry Bassel]

On 07 May 14 11:17, Will Deacon wrote:
> On Wed, May 07, 2014 at 12:32:29AM +0100, Larry Bassel wrote:
> > Make calls to ct_user_enter when the kernel is exited
> > and ct_user_exit when the kernel is entered (in el0_da,
> > el0_ia, el0_svc, el0_irq).
> 
> Why only these entry points? I can reschedule after any exception from EL0,
> so I'd expect all exceptions from userspace to need annotating, no?
> 
> > These macros expand to function calls which will only work
> > properly if el0_sync and related code has been rearranged
> > (in a previous patch of this series).
> > 
> > In order to avoid saving registers, the slow syscall path
> > is forced (as x86 does).
> 
> ... and if you decide to handle undef exceptions, I think you'll need
> the register saving too, in case the kernel needs to perform emulation.

These are excellent points, I will rework the patch and submit v3.

Thanks for the feedback.

> 
> Will

Larry

[PATCH v2 2/2] arm64: enable context tracking

[Kevin Hilman]

Hi Will,

Will Deacon <will.deacon@arm.com> writes:

> On Wed, May 07, 2014 at 12:32:29AM +0100, Larry Bassel wrote:
>> Make calls to ct_user_enter when the kernel is exited
>> and ct_user_exit when the kernel is entered (in el0_da,
>> el0_ia, el0_svc, el0_irq).
>
> Why only these entry points? I can reschedule after any exception from EL0,
> so I'd expect all exceptions from userspace to need annotating, no?

In my initial approach to this, you might recall (though it was over a
year ago now) was to just instrument kernel_enter rather than sprinkle
the instrumentaion in cl0_*.  However, your concern at the time was that
since it was before debugging was enabled it would complicate debugging
these paths.

Any chance you have any other suggestion on how we might do this in
kernel_entry rather than sprinkling them all over cl0_*?  or is the
sprinkling the only good way to handle this.

Kevin

[PATCH v2 2/2] arm64: enable context tracking

[Will Deacon]

On Thu, May 08, 2014 at 12:49:04AM +0100, Kevin Hilman wrote:
> Hi Will,

Hello Kevin,

> Will Deacon <will.deacon@arm.com> writes:
> > On Wed, May 07, 2014 at 12:32:29AM +0100, Larry Bassel wrote:
> >> Make calls to ct_user_enter when the kernel is exited
> >> and ct_user_exit when the kernel is entered (in el0_da,
> >> el0_ia, el0_svc, el0_irq).
> >
> > Why only these entry points? I can reschedule after any exception from EL0,
> > so I'd expect all exceptions from userspace to need annotating, no?
> 
> In my initial approach to this, you might recall (though it was over a
> year ago now) was to just instrument kernel_enter rather than sprinkle
> the instrumentaion in cl0_*.  However, your concern at the time was that
> since it was before debugging was enabled it would complicate debugging
> these paths.
> 
> Any chance you have any other suggestion on how we might do this in
> kernel_entry rather than sprinkling them all over cl0_*?  or is the
> sprinkling the only good way to handle this.

Unfortunately, different exceptions do subtly different things before
invoking the main handler. For example:

 - Stashing the far
 - Enabling IRQs
 - Enabling debug
 - All the stuff on the syscall path

so putting the logic in kernel_entry isn't really do-able, unfortunately.

Will