2020-08-03 12:09:28

by Maninder Singh

[permalink] [raw]
Subject: [PATCH 1/1] arm64: add support for PAGE_SIZE aligned kernel stack

currently THREAD_SIZE is always in power of 2, which will waste
memory in cases there is need to increase of stack size.

Thus adding support for PAGE_SIZE(not power of 2) stacks for arm64.
User can decide any value 12KB, 16KB, 20 KB etc. based on value
of THREAD_SHIFT. User can set any value which is PAGE_SIZE aligned for
PAGE_ALIGNED_STACK_SIZE config.

Value of THREAD_SIZE is defined as 12KB for now, since with irq stacks
it is enough and it will save 4KB per thread.

IRQ stack size is not changed and alignement of IRQ stack and kernel stack
is maintained same to catch stack overflow faults as earlier.

THREAD_SIZE masking in common files is changed to THREAD_SIZE_ALIGNED.

Co-developed-by: Vaneet narang <[email protected]>
Signed-off-by: Vaneet narang <[email protected]>
Signed-off-by: Maninder Singh <[email protected]>
---
arch/arm64/Kconfig | 9 +++++++++
arch/arm64/include/asm/memory.h | 29 +++++++++++++++++++++++++----
arch/arm64/kernel/entry.S | 4 ++--
arch/arm64/kernel/ptrace.c | 4 ++--
drivers/misc/lkdtm/stackleak.c | 2 +-
fs/proc/base.c | 4 ++--
include/linux/thread_info.h | 4 ++++
kernel/trace/trace_stack.c | 4 ++--
8 files changed, 47 insertions(+), 13 deletions(-)

diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
index c970171..301e068 100644
--- a/arch/arm64/Kconfig
+++ b/arch/arm64/Kconfig
@@ -977,6 +977,15 @@ config NODES_SHIFT
Specify the maximum number of NUMA Nodes available on the target
system. Increases memory reserved to accommodate various tables.

+config PAGE_ALIGNED_STACK_SIZE
+ int "set per thread stack size (THREAD_SIZE)"
+ default 12288
+ depends on VMAP_STACK && ARM64_4K_PAGES && !KASAN
+ help
+ Per Thread stack size, value must be PAGE_SIZE aligned.
+ make sure value should be less than (1 << THREAD_SHIFT),
+ otherwise increase THREAD_SHIFT also.
+
config USE_PERCPU_NUMA_NODE_ID
def_bool y
depends on NUMA
diff --git a/arch/arm64/include/asm/memory.h b/arch/arm64/include/asm/memory.h
index 5767836..597071e 100644
--- a/arch/arm64/include/asm/memory.h
+++ b/arch/arm64/include/asm/memory.h
@@ -93,6 +93,7 @@
*/
#if defined(CONFIG_VMAP_STACK) && (MIN_THREAD_SHIFT < PAGE_SHIFT)
#define THREAD_SHIFT PAGE_SHIFT
+#define THREAD_SIZE (UL(1) << THREAD_SHIFT)
#else
#define THREAD_SHIFT MIN_THREAD_SHIFT
#endif
@@ -101,7 +102,15 @@
#define THREAD_SIZE_ORDER (THREAD_SHIFT - PAGE_SHIFT)
#endif

-#define THREAD_SIZE (UL(1) << THREAD_SHIFT)
+#define THREAD_SIZE_ALIGNED (UL(1) << THREAD_SHIFT)
+
+#ifndef THREAD_SIZE
+#if defined(CONFIG_VMAP_STACK) && (CONFIG_PAGE_ALIGNED_STACK_SIZE)
+#define THREAD_SIZE CONFIG_PAGE_ALIGNED_STACK_SIZE
+#else
+#define THREAD_SIZE THREAD_SIZE_ALIGNED
+#endif
+#endif

/*
* By aligning VMAP'd stacks to 2 * THREAD_SIZE, we can detect overflow by
@@ -109,12 +118,24 @@
* assembly.
*/
#ifdef CONFIG_VMAP_STACK
-#define THREAD_ALIGN (2 * THREAD_SIZE)
+#define THREAD_ALIGN (2 * THREAD_SIZE_ALIGNED)
#else
-#define THREAD_ALIGN THREAD_SIZE
+#define THREAD_ALIGN THREAD_SIZE_ALIGNED
+#endif
+
+#ifdef CONFIG_PAGE_ALIGNED_STACK_SIZE
+
+#if (THREAD_SIZE_ALIGNED < THREAD_SIZE)
+#error "PAGE_ALIGNED_STACK_SIZE is more than THREAD_SIZE_ALIGNED, increase THREAD_SHIFT"
+#endif
+
+#if (THREAD_SIZE % PAGE_SIZE)
+#error "PAGE_ALIGNED_STACK_SIZE must be PAGE_SIZE align"
+#endif
+
#endif

-#define IRQ_STACK_SIZE THREAD_SIZE
+#define IRQ_STACK_SIZE THREAD_SIZE_ALIGNED

#define OVERFLOW_STACK_SIZE SZ_4K

diff --git a/arch/arm64/kernel/entry.S b/arch/arm64/kernel/entry.S
index 13458c2..5190573 100644
--- a/arch/arm64/kernel/entry.S
+++ b/arch/arm64/kernel/entry.S
@@ -444,12 +444,12 @@ alternative_insn eret, nop, ARM64_UNMAP_KERNEL_AT_EL0

/*
* Compare sp with the base of the task stack.
- * If the top ~(THREAD_SIZE - 1) bits match, we are on a task stack,
+ * If the top ~(THREAD_SIZE_ALIGNED - 1) bits match, we are on a task stack,
* and should switch to the irq stack.
*/
ldr x25, [tsk, TSK_STACK]
eor x25, x25, x19
- and x25, x25, #~(THREAD_SIZE - 1)
+ and x25, x25, #~(THREAD_SIZE_ALIGNED - 1)
cbnz x25, 9998f

ldr_this_cpu x25, irq_stack_ptr, x26
diff --git a/arch/arm64/kernel/ptrace.c b/arch/arm64/kernel/ptrace.c
index b82eb50..800bb84 100644
--- a/arch/arm64/kernel/ptrace.c
+++ b/arch/arm64/kernel/ptrace.c
@@ -120,8 +120,8 @@ int regs_query_register_offset(const char *name)
*/
static bool regs_within_kernel_stack(struct pt_regs *regs, unsigned long addr)
{
- return ((addr & ~(THREAD_SIZE - 1)) ==
- (kernel_stack_pointer(regs) & ~(THREAD_SIZE - 1))) ||
+ return ((addr & ~(THREAD_SIZE_ALIGNED - 1)) ==
+ (kernel_stack_pointer(regs) & ~(THREAD_SIZE_ALIGNED - 1))) ||
on_irq_stack(addr, NULL);
}

diff --git a/drivers/misc/lkdtm/stackleak.c b/drivers/misc/lkdtm/stackleak.c
index d1a5c07..f4ab60a 100644
--- a/drivers/misc/lkdtm/stackleak.c
+++ b/drivers/misc/lkdtm/stackleak.c
@@ -24,7 +24,7 @@ void lkdtm_STACKLEAK_ERASING(void)
*/
sp = PTR_ALIGN(&i, sizeof(unsigned long));

- left = ((unsigned long)sp & (THREAD_SIZE - 1)) / sizeof(unsigned long);
+ left = ((unsigned long)sp & (THREAD_SIZE_ALIGNED - 1)) / sizeof(unsigned long);
sp--;

/*
diff --git a/fs/proc/base.c b/fs/proc/base.c
index f3b6e12..f89e2c5 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -3135,9 +3135,9 @@ static int proc_stack_depth(struct seq_file *m, struct pid_namespace *ns,
struct pid *pid, struct task_struct *task)
{
unsigned long prev_depth = THREAD_SIZE -
- (task->prev_lowest_stack & (THREAD_SIZE - 1));
+ (task->prev_lowest_stack & (THREAD_SIZE_ALIGNED - 1));
unsigned long depth = THREAD_SIZE -
- (task->lowest_stack & (THREAD_SIZE - 1));
+ (task->lowest_stack & (THREAD_SIZE_ALIGNED - 1));

seq_printf(m, "previous stack depth: %lu\nstack depth: %lu\n",
prev_depth, depth);
diff --git a/include/linux/thread_info.h b/include/linux/thread_info.h
index e93e249..35a73b5 100644
--- a/include/linux/thread_info.h
+++ b/include/linux/thread_info.h
@@ -43,6 +43,10 @@ enum {
#define THREAD_ALIGN THREAD_SIZE
#endif

+#ifndef THREAD_SIZE_ALIGNED
+#define THREAD_SIZE_ALIGNED THREAD_SIZE
+#endif
+
#define THREADINFO_GFP (GFP_KERNEL_ACCOUNT | __GFP_ZERO)

/*
diff --git a/kernel/trace/trace_stack.c b/kernel/trace/trace_stack.c
index 5810fb8..ef3d442 100644
--- a/kernel/trace/trace_stack.c
+++ b/kernel/trace/trace_stack.c
@@ -159,7 +159,7 @@ static void check_stack(unsigned long ip, unsigned long *stack)
int frame_size = READ_ONCE(tracer_frame);
int i, x;

- this_size = ((unsigned long)stack) & (THREAD_SIZE-1);
+ this_size = ((unsigned long)stack) & (THREAD_SIZE_ALIGNED - 1);
this_size = THREAD_SIZE - this_size;
/* Remove the frame of the tracer */
this_size -= frame_size;
@@ -211,7 +211,7 @@ static void check_stack(unsigned long ip, unsigned long *stack)
x = 0;
start = stack;
top = (unsigned long *)
- (((unsigned long)start & ~(THREAD_SIZE-1)) + THREAD_SIZE);
+ (((unsigned long)start & ~(THREAD_SIZE_ALIGNED - 1)) + THREAD_SIZE);

/*
* Loop through all the entries. One of the entries may
--
1.9.1


2020-08-03 12:37:52

by Mark Rutland

[permalink] [raw]
Subject: Re: [PATCH 1/1] arm64: add support for PAGE_SIZE aligned kernel stack

On Sun, Aug 02, 2020 at 10:05:15PM +0530, Maninder Singh wrote:
> currently THREAD_SIZE is always in power of 2, which will waste
> memory in cases there is need to increase of stack size.

If you are seeing issues with the current stack size, can you please
explain that in more detail? Where are you seeing problems? Which
configuration options do you have selected?

I'm not keen on making kernel stack sizes configurable as it's not
currently possible for the person building the kernel to figure out a
safe size (and if this were possible, it's be better to handle this
automatically).

If the stack size is too small in some configurations I think we need to
ensure that it is appropriately sized regardless of whether the person
building the kernel believes they can identify a reasonable size.

> Thus adding support for PAGE_SIZE(not power of 2) stacks for arm64.
> User can decide any value 12KB, 16KB, 20 KB etc. based on value
> of THREAD_SHIFT. User can set any value which is PAGE_SIZE aligned for
> PAGE_ALIGNED_STACK_SIZE config.
>
> Value of THREAD_SIZE is defined as 12KB for now, since with irq stacks
> it is enough and it will save 4KB per thread.

How are you certain of this?

> IRQ stack size is not changed and alignement of IRQ stack and kernel stack
> is maintained same to catch stack overflow faults as earlier.
>
> THREAD_SIZE masking in common files is changed to THREAD_SIZE_ALIGNED.

This is definitely going to be confused with THREAD_ALIGN, so if we do
go ahead with this, the naming will need work.

Thanks,
Mark.

>
> Co-developed-by: Vaneet narang <[email protected]>
> Signed-off-by: Vaneet narang <[email protected]>
> Signed-off-by: Maninder Singh <[email protected]>
> ---
> arch/arm64/Kconfig | 9 +++++++++
> arch/arm64/include/asm/memory.h | 29 +++++++++++++++++++++++++----
> arch/arm64/kernel/entry.S | 4 ++--
> arch/arm64/kernel/ptrace.c | 4 ++--
> drivers/misc/lkdtm/stackleak.c | 2 +-
> fs/proc/base.c | 4 ++--
> include/linux/thread_info.h | 4 ++++
> kernel/trace/trace_stack.c | 4 ++--
> 8 files changed, 47 insertions(+), 13 deletions(-)
>
> diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
> index c970171..301e068 100644
> --- a/arch/arm64/Kconfig
> +++ b/arch/arm64/Kconfig
> @@ -977,6 +977,15 @@ config NODES_SHIFT
> Specify the maximum number of NUMA Nodes available on the target
> system. Increases memory reserved to accommodate various tables.
>
> +config PAGE_ALIGNED_STACK_SIZE
> + int "set per thread stack size (THREAD_SIZE)"
> + default 12288
> + depends on VMAP_STACK && ARM64_4K_PAGES && !KASAN
> + help
> + Per Thread stack size, value must be PAGE_SIZE aligned.
> + make sure value should be less than (1 << THREAD_SHIFT),
> + otherwise increase THREAD_SHIFT also.
> +
> config USE_PERCPU_NUMA_NODE_ID
> def_bool y
> depends on NUMA
> diff --git a/arch/arm64/include/asm/memory.h b/arch/arm64/include/asm/memory.h
> index 5767836..597071e 100644
> --- a/arch/arm64/include/asm/memory.h
> +++ b/arch/arm64/include/asm/memory.h
> @@ -93,6 +93,7 @@
> */
> #if defined(CONFIG_VMAP_STACK) && (MIN_THREAD_SHIFT < PAGE_SHIFT)
> #define THREAD_SHIFT PAGE_SHIFT
> +#define THREAD_SIZE (UL(1) << THREAD_SHIFT)
> #else
> #define THREAD_SHIFT MIN_THREAD_SHIFT
> #endif
> @@ -101,7 +102,15 @@
> #define THREAD_SIZE_ORDER (THREAD_SHIFT - PAGE_SHIFT)
> #endif
>
> -#define THREAD_SIZE (UL(1) << THREAD_SHIFT)
> +#define THREAD_SIZE_ALIGNED (UL(1) << THREAD_SHIFT)
> +
> +#ifndef THREAD_SIZE
> +#if defined(CONFIG_VMAP_STACK) && (CONFIG_PAGE_ALIGNED_STACK_SIZE)
> +#define THREAD_SIZE CONFIG_PAGE_ALIGNED_STACK_SIZE
> +#else
> +#define THREAD_SIZE THREAD_SIZE_ALIGNED
> +#endif
> +#endif
>
> /*
> * By aligning VMAP'd stacks to 2 * THREAD_SIZE, we can detect overflow by
> @@ -109,12 +118,24 @@
> * assembly.
> */
> #ifdef CONFIG_VMAP_STACK
> -#define THREAD_ALIGN (2 * THREAD_SIZE)
> +#define THREAD_ALIGN (2 * THREAD_SIZE_ALIGNED)
> #else
> -#define THREAD_ALIGN THREAD_SIZE
> +#define THREAD_ALIGN THREAD_SIZE_ALIGNED
> +#endif
> +
> +#ifdef CONFIG_PAGE_ALIGNED_STACK_SIZE
> +
> +#if (THREAD_SIZE_ALIGNED < THREAD_SIZE)
> +#error "PAGE_ALIGNED_STACK_SIZE is more than THREAD_SIZE_ALIGNED, increase THREAD_SHIFT"
> +#endif
> +
> +#if (THREAD_SIZE % PAGE_SIZE)
> +#error "PAGE_ALIGNED_STACK_SIZE must be PAGE_SIZE align"
> +#endif
> +
> #endif
>
> -#define IRQ_STACK_SIZE THREAD_SIZE
> +#define IRQ_STACK_SIZE THREAD_SIZE_ALIGNED
>
> #define OVERFLOW_STACK_SIZE SZ_4K
>
> diff --git a/arch/arm64/kernel/entry.S b/arch/arm64/kernel/entry.S
> index 13458c2..5190573 100644
> --- a/arch/arm64/kernel/entry.S
> +++ b/arch/arm64/kernel/entry.S
> @@ -444,12 +444,12 @@ alternative_insn eret, nop, ARM64_UNMAP_KERNEL_AT_EL0
>
> /*
> * Compare sp with the base of the task stack.
> - * If the top ~(THREAD_SIZE - 1) bits match, we are on a task stack,
> + * If the top ~(THREAD_SIZE_ALIGNED - 1) bits match, we are on a task stack,
> * and should switch to the irq stack.
> */
> ldr x25, [tsk, TSK_STACK]
> eor x25, x25, x19
> - and x25, x25, #~(THREAD_SIZE - 1)
> + and x25, x25, #~(THREAD_SIZE_ALIGNED - 1)
> cbnz x25, 9998f
>
> ldr_this_cpu x25, irq_stack_ptr, x26
> diff --git a/arch/arm64/kernel/ptrace.c b/arch/arm64/kernel/ptrace.c
> index b82eb50..800bb84 100644
> --- a/arch/arm64/kernel/ptrace.c
> +++ b/arch/arm64/kernel/ptrace.c
> @@ -120,8 +120,8 @@ int regs_query_register_offset(const char *name)
> */
> static bool regs_within_kernel_stack(struct pt_regs *regs, unsigned long addr)
> {
> - return ((addr & ~(THREAD_SIZE - 1)) ==
> - (kernel_stack_pointer(regs) & ~(THREAD_SIZE - 1))) ||
> + return ((addr & ~(THREAD_SIZE_ALIGNED - 1)) ==
> + (kernel_stack_pointer(regs) & ~(THREAD_SIZE_ALIGNED - 1))) ||
> on_irq_stack(addr, NULL);
> }
>
> diff --git a/drivers/misc/lkdtm/stackleak.c b/drivers/misc/lkdtm/stackleak.c
> index d1a5c07..f4ab60a 100644
> --- a/drivers/misc/lkdtm/stackleak.c
> +++ b/drivers/misc/lkdtm/stackleak.c
> @@ -24,7 +24,7 @@ void lkdtm_STACKLEAK_ERASING(void)
> */
> sp = PTR_ALIGN(&i, sizeof(unsigned long));
>
> - left = ((unsigned long)sp & (THREAD_SIZE - 1)) / sizeof(unsigned long);
> + left = ((unsigned long)sp & (THREAD_SIZE_ALIGNED - 1)) / sizeof(unsigned long);
> sp--;
>
> /*
> diff --git a/fs/proc/base.c b/fs/proc/base.c
> index f3b6e12..f89e2c5 100644
> --- a/fs/proc/base.c
> +++ b/fs/proc/base.c
> @@ -3135,9 +3135,9 @@ static int proc_stack_depth(struct seq_file *m, struct pid_namespace *ns,
> struct pid *pid, struct task_struct *task)
> {
> unsigned long prev_depth = THREAD_SIZE -
> - (task->prev_lowest_stack & (THREAD_SIZE - 1));
> + (task->prev_lowest_stack & (THREAD_SIZE_ALIGNED - 1));
> unsigned long depth = THREAD_SIZE -
> - (task->lowest_stack & (THREAD_SIZE - 1));
> + (task->lowest_stack & (THREAD_SIZE_ALIGNED - 1));
>
> seq_printf(m, "previous stack depth: %lu\nstack depth: %lu\n",
> prev_depth, depth);
> diff --git a/include/linux/thread_info.h b/include/linux/thread_info.h
> index e93e249..35a73b5 100644
> --- a/include/linux/thread_info.h
> +++ b/include/linux/thread_info.h
> @@ -43,6 +43,10 @@ enum {
> #define THREAD_ALIGN THREAD_SIZE
> #endif
>
> +#ifndef THREAD_SIZE_ALIGNED
> +#define THREAD_SIZE_ALIGNED THREAD_SIZE
> +#endif
> +
> #define THREADINFO_GFP (GFP_KERNEL_ACCOUNT | __GFP_ZERO)
>
> /*
> diff --git a/kernel/trace/trace_stack.c b/kernel/trace/trace_stack.c
> index 5810fb8..ef3d442 100644
> --- a/kernel/trace/trace_stack.c
> +++ b/kernel/trace/trace_stack.c
> @@ -159,7 +159,7 @@ static void check_stack(unsigned long ip, unsigned long *stack)
> int frame_size = READ_ONCE(tracer_frame);
> int i, x;
>
> - this_size = ((unsigned long)stack) & (THREAD_SIZE-1);
> + this_size = ((unsigned long)stack) & (THREAD_SIZE_ALIGNED - 1);
> this_size = THREAD_SIZE - this_size;
> /* Remove the frame of the tracer */
> this_size -= frame_size;
> @@ -211,7 +211,7 @@ static void check_stack(unsigned long ip, unsigned long *stack)
> x = 0;
> start = stack;
> top = (unsigned long *)
> - (((unsigned long)start & ~(THREAD_SIZE-1)) + THREAD_SIZE);
> + (((unsigned long)start & ~(THREAD_SIZE_ALIGNED - 1)) + THREAD_SIZE);
>
> /*
> * Loop through all the entries. One of the entries may
> --
> 1.9.1
>

2020-08-04 06:16:28

by Vaneet Narang

[permalink] [raw]
Subject: RE: [PATCH 1/1] arm64: add support for PAGE_SIZE aligned kernel stack

Hi Mark,

>> currently THREAD_SIZE is always in power of 2, which will waste
>> memory in cases there is need to increase of stack size.
>
>If you are seeing issues with the current stack size, can you please
>explain that in more detail? Where are you seeing problems? Which
>configuration options do you have selected?
>
>I'm not keen on making kernel stack sizes configurable as it's not
>currently possible for the person building the kernel to figure out a
>safe size (and if this were possible, it's be better to handle this
>automatically).
>
>If the stack size is too small in some configurations I think we need to
>ensure that it is appropriately sized regardless of whether the person
>building the kernel believes they can identify a reasonable size.

Motivation behind these changes is saving memory on our system.
Our system runs around 2500 threads concurrently so saving 4Kb
might help us in saving 10MB memory.

To ensure 12KB is sufficient for our system we have used stack tracing and
realised maximum stack used is not more than 9KB.
/sys/kernel/tracing/stack_max_size

Tracing interface defined by kernel to track maximum stack size can be used
by others to decide appropriate stack size.

>> Thus adding support for PAGE_SIZE(not power of 2) stacks for arm64.
>> User can decide any value 12KB, 16KB, 20 KB etc. based on value
>> of THREAD_SHIFT. User can set any value which is PAGE_SIZE aligned for
>> PAGE_ALIGNED_STACK_SIZE config.
>>
>> Value of THREAD_SIZE is defined as 12KB for now, since with irq stacks
>> it is enough and it will save 4KB per thread.
>
>How are you certain of this?

Prev ARM64 uses 16KB stack size to store IRQ stack and thread on the same kernel stack.
Now since these are stored seperately so maximum kernel stack requirement is also reduced.
ARM still uses 8KB stack to store both SVC and IRQ mode stack. So ARM64 shouldn't
have requirement more than double when compared to ARM.
So considering these points we realised 12KB stack might be sufficient
for our system. Analyzing stack using stack tracer confirmed max stack requirement.

This is still configurable, if some system has higher stack requirement then user can
go with 16KB but there should be option to change it.


Regards,
Vaneet Narang

2020-08-07 11:26:40

by Maninder Singh

[permalink] [raw]
Subject: RE: [PATCH 1/1] arm64: add support for PAGE_SIZE aligned kernel stack


Hi Mark,

>>If you are seeing issues with the current stack size, can you please
>>explain that in more detail? Where are you seeing problems? Which
>>configuration options do you have selected?
>>

We checked on our system with netflix and youtube 4K videos running
max stack consumption was 7 KB:

sh-3.2# cat /sys/kernel/debug/tracing/stack_max_size
7312

Thus we thought 16KB wil waste a lot of memory for our ARM64 target,
and 8KB can be less in some exceptional case.
So we decided for 12 KB change.


>>I'm not keen on making kernel stack sizes configurable as it's not
>>currently possible for the person building the kernel to figure out a
>>safe size (and if this were possible, it's be better to handle this
>>automatically).
>>

Ok, can we do something else?, like not making it configurable,
but do code changes to support it for PAGE_SIZE aligned stacks
rather than power of 2 alignement.

Although with configurable also, it depends on VMAP stack,
so user will have stack exhaust exception and can increase acccordingly also.

>
>Motivation behind these changes is saving memory on our system.
>Our system runs around 2500 threads concurrently so saving 4Kb
>might help us in saving 10MB memory.
>
>To ensure 12KB is sufficient for our system we have used stack tracing and
>realised maximum stack used is not more than 9KB.
> /sys/kernel/tracing/stack_max_size
>
>Tracing interface defined by kernel to track maximum stack size can be used
>by others to decide appropriate stack size.
>

Thanks