netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH net-next v2] openvswitch: reduce stack usage in do_execute_actions
@ 2023-09-21 19:42 Ilya Maximets
  2023-09-21 19:58 ` Eric Dumazet
                   ` (3 more replies)
  0 siblings, 4 replies; 5+ messages in thread
From: Ilya Maximets @ 2023-09-21 19:42 UTC (permalink / raw)
  To: netdev
  Cc: Jakub Kicinski, David S. Miller, Eric Dumazet, Paolo Abeni,
	linux-kernel, dev, Pravin B Shelar, Eelco Chaudron, Ilya Maximets

do_execute_actions() function can be called recursively multiple
times while executing actions that require pipeline forking or
recirculations.  It may also be re-entered multiple times if the packet
leaves openvswitch module and re-enters it through a different port.

Currently, there is a 256-byte array allocated on stack in this
function that is supposed to hold NSH header.  Compilers tend to
pre-allocate that space right at the beginning of the function:

     a88:       48 81 ec b0 01 00 00    sub    $0x1b0,%rsp

NSH is not a very common protocol, but the space is allocated on every
recursive call or re-entry multiplying the wasted stack space.

Move the stack allocation to push_nsh() function that is only used
if NSH actions are actually present.  push_nsh() is also a simple
function without a possibility for re-entry, so the stack is returned
right away.

With this change the preallocated space is reduced by 256 B per call:

     b18:       48 81 ec b0 00 00 00    sub    $0xb0,%rsp

Signed-off-by: Ilya Maximets <i.maximets@ovn.org>
---

Version 2:
  - Added 'noinline_for_stack' to avoid potential inlining. [Eric]

 net/openvswitch/actions.c | 23 +++++++++++------------
 1 file changed, 11 insertions(+), 12 deletions(-)

diff --git a/net/openvswitch/actions.c b/net/openvswitch/actions.c
index 5f8094acd056..6fcd7e2ca81f 100644
--- a/net/openvswitch/actions.c
+++ b/net/openvswitch/actions.c
@@ -311,11 +311,18 @@ static int push_eth(struct sk_buff *skb, struct sw_flow_key *key,
 	return 0;
 }
 
-static int push_nsh(struct sk_buff *skb, struct sw_flow_key *key,
-		    const struct nshhdr *nh)
+static noinline_for_stack int push_nsh(struct sk_buff *skb,
+				       struct sw_flow_key *key,
+				       const struct nlattr *a)
 {
+	u8 buffer[NSH_HDR_MAX_LEN];
+	struct nshhdr *nh = (struct nshhdr *)buffer;
 	int err;
 
+	err = nsh_hdr_from_nlattr(a, nh, NSH_HDR_MAX_LEN);
+	if (err)
+		return err;
+
 	err = nsh_push(skb, nh);
 	if (err)
 		return err;
@@ -1439,17 +1446,9 @@ static int do_execute_actions(struct datapath *dp, struct sk_buff *skb,
 			err = pop_eth(skb, key);
 			break;
 
-		case OVS_ACTION_ATTR_PUSH_NSH: {
-			u8 buffer[NSH_HDR_MAX_LEN];
-			struct nshhdr *nh = (struct nshhdr *)buffer;
-
-			err = nsh_hdr_from_nlattr(nla_data(a), nh,
-						  NSH_HDR_MAX_LEN);
-			if (unlikely(err))
-				break;
-			err = push_nsh(skb, key, nh);
+		case OVS_ACTION_ATTR_PUSH_NSH:
+			err = push_nsh(skb, key, nla_data(a));
 			break;
-		}
 
 		case OVS_ACTION_ATTR_POP_NSH:
 			err = pop_nsh(skb, key);
-- 
2.41.0


^ permalink raw reply related	[flat|nested] 5+ messages in thread

* Re: [PATCH net-next v2] openvswitch: reduce stack usage in do_execute_actions
  2023-09-21 19:42 [PATCH net-next v2] openvswitch: reduce stack usage in do_execute_actions Ilya Maximets
@ 2023-09-21 19:58 ` Eric Dumazet
  2023-09-22  6:43 ` Eelco Chaudron
                   ` (2 subsequent siblings)
  3 siblings, 0 replies; 5+ messages in thread
From: Eric Dumazet @ 2023-09-21 19:58 UTC (permalink / raw)
  To: Ilya Maximets
  Cc: netdev, Jakub Kicinski, David S. Miller, Paolo Abeni,
	linux-kernel, dev, Pravin B Shelar, Eelco Chaudron

On Thu, Sep 21, 2023 at 9:42 PM Ilya Maximets <i.maximets@ovn.org> wrote:
>
> do_execute_actions() function can be called recursively multiple
> times while executing actions that require pipeline forking or
> recirculations.  It may also be re-entered multiple times if the packet
> leaves openvswitch module and re-enters it through a different port.
>
> Currently, there is a 256-byte array allocated on stack in this
> function that is supposed to hold NSH header.  Compilers tend to
> pre-allocate that space right at the beginning of the function:
>
>      a88:       48 81 ec b0 01 00 00    sub    $0x1b0,%rsp
>
> NSH is not a very common protocol, but the space is allocated on every
> recursive call or re-entry multiplying the wasted stack space.
>

Reviewed-by: Eric Dumazet <edumazet@google.com>

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH net-next v2] openvswitch: reduce stack usage in do_execute_actions
  2023-09-21 19:42 [PATCH net-next v2] openvswitch: reduce stack usage in do_execute_actions Ilya Maximets
  2023-09-21 19:58 ` Eric Dumazet
@ 2023-09-22  6:43 ` Eelco Chaudron
  2023-09-27 20:24 ` [ovs-dev] " Aaron Conole
  2023-10-01 18:10 ` patchwork-bot+netdevbpf
  3 siblings, 0 replies; 5+ messages in thread
From: Eelco Chaudron @ 2023-09-22  6:43 UTC (permalink / raw)
  To: netdev



On 21 Sep 2023, at 21:42, Ilya Maximets wrote:

> do_execute_actions() function can be called recursively multiple
> times while executing actions that require pipeline forking or
> recirculations.  It may also be re-entered multiple times if the packet
> leaves openvswitch module and re-enters it through a different port.
>
> Currently, there is a 256-byte array allocated on stack in this
> function that is supposed to hold NSH header.  Compilers tend to
> pre-allocate that space right at the beginning of the function:
>
>      a88:       48 81 ec b0 01 00 00    sub    $0x1b0,%rsp
>
> NSH is not a very common protocol, but the space is allocated on every
> recursive call or re-entry multiplying the wasted stack space.
>
> Move the stack allocation to push_nsh() function that is only used
> if NSH actions are actually present.  push_nsh() is also a simple
> function without a possibility for re-entry, so the stack is returned
> right away.
>
> With this change the preallocated space is reduced by 256 B per call:
>
>      b18:       48 81 ec b0 00 00 00    sub    $0xb0,%rsp
>
> Signed-off-by: Ilya Maximets <i.maximets@ovn.org>

One more time to the list only, as for some reason I had HTML reply turned on :(

Thanks Ilya for optimizing this.

Reviewed-by: Eelco Chaudron echaudro@redhat.com


^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [ovs-dev] [PATCH net-next v2] openvswitch: reduce stack usage in do_execute_actions
  2023-09-21 19:42 [PATCH net-next v2] openvswitch: reduce stack usage in do_execute_actions Ilya Maximets
  2023-09-21 19:58 ` Eric Dumazet
  2023-09-22  6:43 ` Eelco Chaudron
@ 2023-09-27 20:24 ` Aaron Conole
  2023-10-01 18:10 ` patchwork-bot+netdevbpf
  3 siblings, 0 replies; 5+ messages in thread
From: Aaron Conole @ 2023-09-27 20:24 UTC (permalink / raw)
  To: Ilya Maximets
  Cc: netdev, dev, linux-kernel, Eric Dumazet, Jakub Kicinski,
	Paolo Abeni, David S. Miller

Ilya Maximets <i.maximets@ovn.org> writes:

> do_execute_actions() function can be called recursively multiple
> times while executing actions that require pipeline forking or
> recirculations.  It may also be re-entered multiple times if the packet
> leaves openvswitch module and re-enters it through a different port.
>
> Currently, there is a 256-byte array allocated on stack in this
> function that is supposed to hold NSH header.  Compilers tend to
> pre-allocate that space right at the beginning of the function:
>
>      a88:       48 81 ec b0 01 00 00    sub    $0x1b0,%rsp
>
> NSH is not a very common protocol, but the space is allocated on every
> recursive call or re-entry multiplying the wasted stack space.
>
> Move the stack allocation to push_nsh() function that is only used
> if NSH actions are actually present.  push_nsh() is also a simple
> function without a possibility for re-entry, so the stack is returned
> right away.
>
> With this change the preallocated space is reduced by 256 B per call:
>
>      b18:       48 81 ec b0 00 00 00    sub    $0xb0,%rsp
>
> Signed-off-by: Ilya Maximets <i.maximets@ovn.org>
> ---

Thanks, Ilya.

Reviewed-by: Aaron Conole <aconole@redhat.com>


^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH net-next v2] openvswitch: reduce stack usage in do_execute_actions
  2023-09-21 19:42 [PATCH net-next v2] openvswitch: reduce stack usage in do_execute_actions Ilya Maximets
                   ` (2 preceding siblings ...)
  2023-09-27 20:24 ` [ovs-dev] " Aaron Conole
@ 2023-10-01 18:10 ` patchwork-bot+netdevbpf
  3 siblings, 0 replies; 5+ messages in thread
From: patchwork-bot+netdevbpf @ 2023-10-01 18:10 UTC (permalink / raw)
  To: Ilya Maximets
  Cc: netdev, kuba, davem, edumazet, pabeni, linux-kernel, dev, pshelar,
	echaudro

Hello:

This patch was applied to netdev/net-next.git (main)
by David S. Miller <davem@davemloft.net>:

On Thu, 21 Sep 2023 21:42:35 +0200 you wrote:
> do_execute_actions() function can be called recursively multiple
> times while executing actions that require pipeline forking or
> recirculations.  It may also be re-entered multiple times if the packet
> leaves openvswitch module and re-enters it through a different port.
> 
> Currently, there is a 256-byte array allocated on stack in this
> function that is supposed to hold NSH header.  Compilers tend to
> pre-allocate that space right at the beginning of the function:
> 
> [...]

Here is the summary with links:
  - [net-next,v2] openvswitch: reduce stack usage in do_execute_actions
    https://git.kernel.org/netdev/net-next/c/06bc3668cc2a

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html



^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2023-10-01 18:10 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2023-09-21 19:42 [PATCH net-next v2] openvswitch: reduce stack usage in do_execute_actions Ilya Maximets
2023-09-21 19:58 ` Eric Dumazet
2023-09-22  6:43 ` Eelco Chaudron
2023-09-27 20:24 ` [ovs-dev] " Aaron Conole
2023-10-01 18:10 ` patchwork-bot+netdevbpf

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).