[PATCH v3] arm64: fpsimd: improve stacking logic in non-interruptible context

Currently, we allow kernel mode NEON in softirq or hardirq context by
stacking and unstacking a slice of the NEON register file for each call
to kernel_neon_begin() and kernel_neon_end(), respectively.

Given that
a) a CPU typically spends most of its time in userland, during which time
   no kernel mode NEON in process context is in progress,
b) a CPU spends most of its time in the kernel doing other things than
   kernel mode NEON when it gets interrupted to perform kernel mode NEON
   in softirq context

the stacking and subsequent unstacking is only necessary if we are
interrupting a thread while it is performing kernel mode NEON in process
context, which means that in all other cases, we can simply preserve the
userland FPSIMD state once, and only restore it upon return to userland,
even if we are being invoked from softirq or hardirq context.

So instead of checking whether we are running in interrupt context, keep
track of the level of nested kernel mode NEON calls in progress, and only
perform the eager stack/unstack if the level exceeds 1.

Signed-off-by: Ard Biesheuvel <redacted>
---
v3:
- avoid corruption by concurrent invocations of kernel_neon_begin()/_end()

v2:
- BUG() on unexpected values of the nesting level
- relax the BUG() on num_regs>32 to a WARN, given that nothing actually
  breaks in that case

 arch/arm64/kernel/fpsimd.c | 47 ++++++++++++++------
 1 file changed, 33 insertions(+), 14 deletions(-)

diff --git a/arch/arm64/kernel/fpsimd.c b/arch/arm64/kernel/fpsimd.c
index 394c61db5566..536ddab9316d 100644
--- a/arch/arm64/kernel/fpsimd.c
+++ b/arch/arm64/kernel/fpsimd.c

@@ -220,20 +220,35 @@ void fpsimd_flush_task_state(struct task_struct *t)
 
 #ifdef CONFIG_KERNEL_MODE_NEON
 
-static DEFINE_PER_CPU(struct fpsimd_partial_state, hardirq_fpsimdstate);
-static DEFINE_PER_CPU(struct fpsimd_partial_state, softirq_fpsimdstate);
+/*
+ * Although unlikely, it is possible for three kernel mode NEON contexts to
+ * be live at the same time: process context, softirq context and hardirq
+ * context. So while the userland context is stashed in the thread's fpsimd
+ * state structure, we need two additional levels of storage.
+ */
+static DEFINE_PER_CPU(struct fpsimd_partial_state, nested_fpsimdstate[2]);
+static DEFINE_PER_CPU(atomic_t, kernel_neon_nesting_level);
 
 /*
  * Kernel-side NEON support functions
  */
 void kernel_neon_begin_partial(u32 num_regs)
 {
-	if (in_interrupt()) {
-		struct fpsimd_partial_state *s = this_cpu_ptr(
-			in_irq() ? &hardirq_fpsimdstate : &softirq_fpsimdstate);
+	struct fpsimd_partial_state *s;
+	int level;
+
+	preempt_disable();
+
+	level = atomic_inc_return(this_cpu_ptr(&kernel_neon_nesting_level));
+	BUG_ON(level > 3);
+
+	if (level > 1) {
+		s = this_cpu_ptr(nested_fpsimdstate);
 
-		BUG_ON(num_regs > 32);
-		fpsimd_save_partial_state(s, roundup(num_regs, 2));
+		WARN_ON_ONCE(num_regs > 32);
+		num_regs = min(roundup(num_regs, 2), 32U);
+
+		fpsimd_save_partial_state(&s[level - 2], num_regs);
 	} else {
 		/*
 		 * Save the userland FPSIMD state if we have one and if we

@@ -241,7 +256,6 @@ void kernel_neon_begin_partial(u32 num_regs)
 		 * that there is no longer userland FPSIMD state in the
 		 * registers.
 		 */
-		preempt_disable();
 		if (current->mm &&
 		    !test_and_set_thread_flag(TIF_FOREIGN_FPSTATE))
 			fpsimd_save_state(&current->thread.fpsimd_state);

@@ -252,13 +266,18 @@ EXPORT_SYMBOL(kernel_neon_begin_partial);
 
 void kernel_neon_end(void)
 {
-	if (in_interrupt()) {
-		struct fpsimd_partial_state *s = this_cpu_ptr(
-			in_irq() ? &hardirq_fpsimdstate : &softirq_fpsimdstate);
-		fpsimd_load_partial_state(s);
-	} else {
-		preempt_enable();
+	struct fpsimd_partial_state *s;
+	int level;
+
+	level = atomic_read(this_cpu_ptr(&kernel_neon_nesting_level));
+	BUG_ON(level < 1);
+
+	if (level > 1) {
+		s = this_cpu_ptr(nested_fpsimdstate);
+		fpsimd_load_partial_state(&s[level - 2]);
 	}
+	atomic_dec(this_cpu_ptr(&kernel_neon_nesting_level));
+	preempt_enable();
 }
 EXPORT_SYMBOL(kernel_neon_end);
 

-- 
2.7.4

On Thu, Dec 08, 2016 at 04:49:39PM +0000, Ard Biesheuvel wrote:

Come to think of it, should we use this_cpu_{read,add}() etc?

I'm not sure why we need the barrier semantics of atomic_t ops here.

[...]

Cheers
---Dave

On 9 December 2016 at 12:20, Dave Martin [off-list ref] wrote:

On Thu, Dec 08, 2016 at 04:49:39PM +0000, Ard Biesheuvel wrote:

quoted

Currently, we allow kernel mode NEON in softirq or hardirq context by
stacking and unstacking a slice of the NEON register file for each call
to kernel_neon_begin() and kernel_neon_end(), respectively.

Given that
a) a CPU typically spends most of its time in userland, during which time
   no kernel mode NEON in process context is in progress,
b) a CPU spends most of its time in the kernel doing other things than
   kernel mode NEON when it gets interrupted to perform kernel mode NEON
   in softirq context

the stacking and subsequent unstacking is only necessary if we are
interrupting a thread while it is performing kernel mode NEON in process
context, which means that in all other cases, we can simply preserve the
userland FPSIMD state once, and only restore it upon return to userland,
even if we are being invoked from softirq or hardirq context.

So instead of checking whether we are running in interrupt context, keep
track of the level of nested kernel mode NEON calls in progress, and only
perform the eager stack/unstack if the level exceeds 1.

Signed-off-by: Ard Biesheuvel <redacted>
---
v3:
- avoid corruption by concurrent invocations of kernel_neon_begin()/_end()

v2:
- BUG() on unexpected values of the nesting level
- relax the BUG() on num_regs>32 to a WARN, given that nothing actually
  breaks in that case

 arch/arm64/kernel/fpsimd.c | 47 ++++++++++++++------
 1 file changed, 33 insertions(+), 14 deletions(-)

diff --git a/arch/arm64/kernel/fpsimd.c b/arch/arm64/kernel/fpsimd.c
index 394c61db5566..536ddab9316d 100644
--- a/arch/arm64/kernel/fpsimd.c
+++ b/arch/arm64/kernel/fpsimd.c

@@ -220,20 +220,35 @@ void fpsimd_flush_task_state(struct task_struct *t)

 #ifdef CONFIG_KERNEL_MODE_NEON

-static DEFINE_PER_CPU(struct fpsimd_partial_state, hardirq_fpsimdstate);
-static DEFINE_PER_CPU(struct fpsimd_partial_state, softirq_fpsimdstate);
+/*
+ * Although unlikely, it is possible for three kernel mode NEON contexts to
+ * be live at the same time: process context, softirq context and hardirq
+ * context. So while the userland context is stashed in the thread's fpsimd
+ * state structure, we need two additional levels of storage.
+ */
+static DEFINE_PER_CPU(struct fpsimd_partial_state, nested_fpsimdstate[2]);
+static DEFINE_PER_CPU(atomic_t, kernel_neon_nesting_level);

Come to think of it, should we use this_cpu_{read,add}() etc?

I wasn't aware those existed ...

I'm not sure why we need the barrier semantics of atomic_t ops here.

I don't think we do. We only need the atomic increment/decrement, afaict

On Fri, Dec 09, 2016 at 12:24:03PM +0000, Ard Biesheuvel wrote:

On 9 December 2016 at 12:20, Dave Martin [off-list ref] wrote:

quoted

On Thu, Dec 08, 2016 at 04:49:39PM +0000, Ard Biesheuvel wrote:

quoted

Currently, we allow kernel mode NEON in softirq or hardirq context by
stacking and unstacking a slice of the NEON register file for each call
to kernel_neon_begin() and kernel_neon_end(), respectively.

Given that
a) a CPU typically spends most of its time in userland, during which time
   no kernel mode NEON in process context is in progress,
b) a CPU spends most of its time in the kernel doing other things than
   kernel mode NEON when it gets interrupted to perform kernel mode NEON
   in softirq context

the stacking and subsequent unstacking is only necessary if we are
interrupting a thread while it is performing kernel mode NEON in process
context, which means that in all other cases, we can simply preserve the
userland FPSIMD state once, and only restore it upon return to userland,
even if we are being invoked from softirq or hardirq context.

So instead of checking whether we are running in interrupt context, keep
track of the level of nested kernel mode NEON calls in progress, and only
perform the eager stack/unstack if the level exceeds 1.

Signed-off-by: Ard Biesheuvel <redacted>
---
v3:
- avoid corruption by concurrent invocations of kernel_neon_begin()/_end()

v2:
- BUG() on unexpected values of the nesting level
- relax the BUG() on num_regs>32 to a WARN, given that nothing actually
  breaks in that case

 arch/arm64/kernel/fpsimd.c | 47 ++++++++++++++------
 1 file changed, 33 insertions(+), 14 deletions(-)

diff --git a/arch/arm64/kernel/fpsimd.c b/arch/arm64/kernel/fpsimd.c
index 394c61db5566..536ddab9316d 100644
--- a/arch/arm64/kernel/fpsimd.c
+++ b/arch/arm64/kernel/fpsimd.c

@@ -220,20 +220,35 @@ void fpsimd_flush_task_state(struct task_struct *t)

[...]

quoted

quoted

+ * state structure, we need two additional levels of storage.
+ */
+static DEFINE_PER_CPU(struct fpsimd_partial_state, nested_fpsimdstate[2]);
+static DEFINE_PER_CPU(atomic_t, kernel_neon_nesting_level);

Come to think of it, should we use this_cpu_{read,add}() etc?

I wasn't aware those existed ...

Actually, nor was I ... I asked Will about local atomics and he
suggested to use those ... then he remembered about the this_cpu_
variants.

The local atomic might still have some barrier semantics, apparently.
For arm64, they're just implemented using the non-local atomics anyway.

quoted

I'm not sure why we need the barrier semantics of atomic_t ops here.

I don't think we do. We only need the atomic increment/decrement, afaict

That's what I was thinking.

Otherwise, the patch looked reasonable to me, but I'll take another look
if you wanted to twiddle the atomics.

Cheers
---Dave