[PATCH] net: sctp: Don't use 64 kilobyte lookup table for four elements

[PATCH] net: sctp: Don't use 64 kilobyte lookup table for four elements

[Denys Vlasenko]

Seemingly innocuous sctp_trans_state_to_prio_map[] array
is way bigger than it looks, since
"[SCTP_UNKNOWN] = 2" expands into "[0xffff] = 2" !

This patch replaces it with switch() statement.

Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
CC: Vlad Yasevich <vyasevich@gmail.com>
CC: Neil Horman <nhorman@tuxdriver.com>
CC: linux-sctp@vger.kernel.org
CC: linux-kernel@vger.kernel.org
---
 net/sctp/associola.c | 20 +++++++++++---------
 1 file changed, 11 insertions(+), 9 deletions(-)

diff --git a/net/sctp/associola.c b/net/sctp/associola.c
index 197c3f5..dae51ac 100644
--- a/net/sctp/associola.c
+++ b/net/sctp/associola.c
@@ -1208,20 +1208,22 @@ void sctp_assoc_update(struct sctp_association *asoc,
  *   within this document.
  *
  * Our basic strategy is to round-robin transports in priorities
- * according to sctp_state_prio_map[] e.g., if no such
+ * according to sctp_trans_score() e.g., if no such
  * transport with state SCTP_ACTIVE exists, round-robin through
  * SCTP_UNKNOWN, etc. You get the picture.
  */
-static const u8 sctp_trans_state_to_prio_map[] = {
-	[SCTP_ACTIVE]	= 3,	/* best case */
-	[SCTP_UNKNOWN]	= 2,
-	[SCTP_PF]	= 1,
-	[SCTP_INACTIVE] = 0,	/* worst case */
-};
-
 static u8 sctp_trans_score(const struct sctp_transport *trans)
 {
-	return sctp_trans_state_to_prio_map[trans->state];
+	switch (trans->state) {
+	case SCTP_ACTIVE:
+		return 3; /* best case */
+	case SCTP_UNKNOWN:
+		return 2;
+	case SCTP_PF:
+		return 1;
+	default: /* case SCTP_INACTIVE */
+		return 0;	/* worst case */
+	}
 }
 
 static struct sctp_transport *sctp_trans_elect_tie(struct sctp_transport *trans1,
-- 
1.8.1.4

Re: [PATCH] net: sctp: Don't use 64 kilobyte lookup table for four elements

[Marcelo Ricardo Leitner]

On Sat, Sep 26, 2015 at 08:53:14PM +0200, Denys Vlasenko wrote:
> Seemingly innocuous sctp_trans_state_to_prio_map[] array
> is way bigger than it looks, since
> "[SCTP_UNKNOWN] = 2" expands into "[0xffff] = 2" !
> 
> This patch replaces it with switch() statement.
> 
> Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
> CC: Vlad Yasevich <vyasevich@gmail.com>
> CC: Neil Horman <nhorman@tuxdriver.com>
> CC: linux-sctp@vger.kernel.org
> CC: linux-kernel@vger.kernel.org
> ---
>  net/sctp/associola.c | 20 +++++++++++---------
>  1 file changed, 11 insertions(+), 9 deletions(-)
> 
> diff --git a/net/sctp/associola.c b/net/sctp/associola.c
> index 197c3f5..dae51ac 100644
> --- a/net/sctp/associola.c
> +++ b/net/sctp/associola.c
> @@ -1208,20 +1208,22 @@ void sctp_assoc_update(struct sctp_association *asoc,
>   *   within this document.
>   *
>   * Our basic strategy is to round-robin transports in priorities
> - * according to sctp_state_prio_map[] e.g., if no such
> + * according to sctp_trans_score() e.g., if no such
>   * transport with state SCTP_ACTIVE exists, round-robin through
>   * SCTP_UNKNOWN, etc. You get the picture.
>   */
> -static const u8 sctp_trans_state_to_prio_map[] = {
> -	[SCTP_ACTIVE]	= 3,	/* best case */
> -	[SCTP_UNKNOWN]	= 2,
> -	[SCTP_PF]	= 1,
> -	[SCTP_INACTIVE] = 0,	/* worst case */
> -};
> -
>  static u8 sctp_trans_score(const struct sctp_transport *trans)
>  {
> -	return sctp_trans_state_to_prio_map[trans->state];
> +	switch (trans->state) {
> +	case SCTP_ACTIVE:
> +		return 3; /* best case */
> +	case SCTP_UNKNOWN:
> +		return 2;
> +	case SCTP_PF:
> +		return 1;
> +	default: /* case SCTP_INACTIVE */
> +		return 0;	/* worst case */
> +	}
>  }

Nice finding, but please make above comments aligned to the column.
Then it looks good to me.

When posting v2, also post to netdev@vger.kernel.org, it's where Dave
collects sctp (amongst others) patches. Kernel patches are not collected
on linux-sctp@, though they are posted there too but for reviewing only.

Thanks,
Marcelo