Tarantool development patches archive
 help / color / mirror / Atom feed
From: "Георгий Кириченко" <georgy@tarantool.org>
To: tarantool-patches@freelists.org
Subject: [tarantool-patches] Re: [RFC] fiber: Increase default stack size
Date: Fri, 22 Feb 2019 10:38:42 +0300	[thread overview]
Message-ID: <7936473.6eviP72C8M@home.lan> (raw)
In-Reply-To: <20190221212642.GT7198@uranus>

[-- Attachment #1: Type: text/plain, Size: 6239 bytes --]

Hi Cyrill!
Thanks for the patch, please see some comments bellow.
I like your approach but I think we should do it in a different manner: we 
could poison 64k page of a fiber stack on start and then check the poison mark 
when fiber finished. If watermark was overwritten by fiber activity we could use 
madvise in order to decrease RSS usage.


On Friday, February 22, 2019 12:26:42 AM MSK Cyrill Gorcunov wrote:
> The default 64K stack size used for years become too
> small for modern distors (Fedora 29 and etc) where third
> party libraries (such as ncurses) started to use 64K for
> own buffers and we get SIGSGV early without reaching
> interactive console phase.
> 
> To address this problem and hopefully eliminate such
> problems in future we increase default size up to 1M.
> Because this value may be too big for old distros or
> other libraries, which would never use such deep stack,
> we do a trick: put watermark at 64K offset of the stack
> and once fiber get scheduled out we test if the mark is
> still present. If we're lucky and noone touched the memory
> we use madvise() syscall to reduce RSS usage.
> 
> https://github.com/tarantool/tarantool/issues/3418
> 
> Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
> ---
> Please review the patch very very carefully since I'm
> not that familiar with the codebase.
> 
>  src/fiber.c |  110
> +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++- src/fiber.h | 
>   6 +++
>  2 files changed, 115 insertions(+), 1 deletion(-)
> 
> Index: tarantool.git/src/fiber.c
> ===================================================================
> --- tarantool.git.orig/src/fiber.c
> +++ tarantool.git/src/fiber.c
> @@ -91,11 +91,26 @@ pthread_t main_thread_id;
>  static size_t page_size;
>  static int stack_direction;
> 
> +/*
> + * A random unique value with help of uuidgen.
> + *
> + * 39ee5420-13f7-417b-9610-ea100c591ab6
> + */
> +static const char stack_wmark[] = {
> +	0x39, 0xee, 0x54, 0x20, 0x13, 0xf7, 0x41, 0x7b,
> +	0x96, 0x10, 0xea, 0x10, 0x0c, 0x59, 0x1a, 0xb6
> +};
> +
> +static void
> +stack_sched_out(struct fiber *fiber);
> +
>  enum {
>  	/* The minimum allowable fiber stack size in bytes */
>  	FIBER_STACK_SIZE_MINIMAL = 16384,
> +	/* Stack size for stack relaxed tasks */
> +	FIBER_STACK_MADVISE_LIMIT = 64536,
>  	/* Default fiber stack size in bytes */
> -	FIBER_STACK_SIZE_DEFAULT = 65536
> +	FIBER_STACK_SIZE_DEFAULT = 1048576
>  };
> 
>  /** Default fiber attributes */
> @@ -188,6 +203,7 @@ fiber_call_impl(struct fiber *callee)
>  	ASAN_START_SWITCH_FIBER(asan_state, 1,
>  				callee->stack,
>  				callee->stack_size);
> +	stack_sched_out(callee);
>  	coro_transfer(&caller->ctx, &callee->ctx);
>  	ASAN_FINISH_SWITCH_FIBER(asan_state);
>  }
> @@ -441,6 +457,7 @@ fiber_yield(void)
>  				(caller->flags & FIBER_IS_DEAD) == 0,
>  				callee->stack,
>  				callee->stack_size);
> +	stack_sched_out(callee);
>  	coro_transfer(&caller->ctx, &callee->ctx);
>  	ASAN_FINISH_SWITCH_FIBER(asan_state);
>  }
> @@ -710,6 +727,96 @@ page_align_up(void *ptr)
>  	return page_align_down(ptr + page_size - 1);
>  }
> 
> +static inline void *
> +stack_wmark_pos(struct fiber *fiber)
> +{
> +	void *pos;
> +
> +	assert(fiber->stack);
> +	assert(fiber->stack_size);
> +
> +	if (stack_direction < 0) {
> +		pos  = fiber->stack + fiber->stack_size;
> +		pos -= FIBER_STACK_MADVISE_LIMIT;
> +		return page_align_up(pos);
> +	} else {
> +		pos  = fiber->stack - fiber->stack_size;
> +		pos += FIBER_STACK_MADVISE_LIMIT;
> +		return page_align_down(pos);
> +	}
> +}
> +
> +/*
> + * Set watermark to the predefined place thus on
> + * fiber sched-out procedure we may detect if
> + * a task was too eager for stack usage.
> + */
> +static inline void
> +stack_set_wmark(struct fiber *fiber)
> +{
> +	void *pos = stack_wmark_pos(fiber);
> +	memcpy(pos, stack_wmark, sizeof(stack_wmark));
> +}
> +
> +static inline bool
> +stack_has_wmark(struct fiber *fiber)
> +{
> +	void *pos = stack_wmark_pos(fiber);
> +	return !!!memcmp(pos, stack_wmark, sizeof(stack_wmark));
> +}
> +
> +/*
> + * Process stack switching: we try to eliminate unneeded
> + * physical page usage for stack. If fiber never touched
> + * area after/before the watermark, the OS get ralaxed in
> + * RSS usage.
> + */
> +static void
> +stack_sched_out(struct fiber *fiber)
> +{
> +	if (!fiber->stack || (fiber->flags & FIBER_CUSTOM_STACK))
> +		return;
> +
> +	/*
> +	 * If fiber ecxeed a watermark, just clear the
> +	 * flag and don't waste time on it in future.
> +	 */
> +	if (!stack_has_wmark(fiber)) {
> +		fiber->flags &= ~FIBER_MADVISED_STACK;
> +		return;
> +	}
> +
> +	/*
> +	 * This is a good one, we simply notify
> +	 * OS about unused stack tail, so associated
> +	 * pages would be put back into a page pool.
> +	 *
> +	 * Note though the fiber still can use
> +	 * remaining space, simply won't be handled
> +	 * that fast on _first_ #pf.
> +	 */
> +	if (!(fiber->flags & FIBER_MADVISED_STACK)) {
> +		size_t size;
> +		void *tail;
> +
> +		if (stack_direction < 0) {
> +			tail = stack_wmark_pos(fiber) - page_size;
> +			size = tail - fiber->stack;
> +		} else {
> +			tail = stack_wmark_pos(fiber) + page_size;
> +			size = fiber->stack - tail;
> +		}
> +
> +		/*
> +		 * Set the flag iif've successed,
> +		 * otherwise will try on the next
> +		 * round.
> +		 */
> +		if (!madvise(fiber->stack, size, MADV_DONTNEED))
> +			fiber->flags |= FIBER_MADVISED_STACK;
> +	}
> +}
> +
>  static int
>  fiber_stack_create(struct fiber *fiber, size_t stack_size)
>  {
> @@ -751,6 +858,7 @@ fiber_stack_create(struct fiber *fiber,
>  						  fiber->stack_size);
> 
>  	mprotect(guard, page_size, PROT_NONE);
> +	stack_set_wmark(fiber);
>  	return 0;
>  }
> 
> Index: tarantool.git/src/fiber.h
> ===================================================================
> --- tarantool.git.orig/src/fiber.h
> +++ tarantool.git/src/fiber.h
> @@ -89,6 +89,12 @@ enum {
>  	 * This flag is set when fiber uses custom stack size.
>  	 */
>  	FIBER_CUSTOM_STACK	= 1 << 5,
> +	/**
> +	 * This flag is set when fiber didn't exceed stack's
> +	 * size watermark and considered being small enough.
> +	 */
> +	FIBER_MADVISED_STACK	= 1 << 6,
> +
>  	FIBER_DEFAULT_FLAGS = FIBER_IS_CANCELLABLE
>  };


[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

  reply	other threads:[~2019-02-22  7:36 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-02-21 21:26 Cyrill Gorcunov
2019-02-22  7:38 ` Георгий Кириченко [this message]
2019-02-22  7:46   ` [tarantool-patches] " Cyrill Gorcunov
2019-02-22  9:24     ` Георгий Кириченко
2019-02-22  9:45       ` Cyrill Gorcunov
2019-02-22 10:48 ` [tarantool-patches] " Alexander Turenko
2019-02-22 10:57   ` Cyrill Gorcunov

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=7936473.6eviP72C8M@home.lan \
    --to=georgy@tarantool.org \
    --cc=tarantool-patches@freelists.org \
    --subject='[tarantool-patches] Re: [RFC] fiber: Increase default stack size' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox