* [RFC 3/2] get rid of rx handler function pointers
@ 2008-06-19 12:38 Johannes Berg
2008-06-19 17:20 ` Harvey Harrison
0 siblings, 1 reply; 5+ messages in thread
From: Johannes Berg @ 2008-06-19 12:38 UTC (permalink / raw)
To: linux-wireless
Even better :)
add/remove: 0/12 grow/shrink: 1/0 up/down: 5532/-6296 (-764)
function old new delta
ieee80211_invoke_rx_handlers 724 6256 +5532
ieee80211_rx_handlers 52 - -52
ieee80211_rx_h_mgmt 116 - -116
ieee80211_rx_h_remove_qos_control 148 - -148
ieee80211_rx_h_passive_scan 168 - -168
ieee80211_rx_h_data 192 - -192
ieee80211_rx_h_ctrl 288 - -288
ieee80211_rx_h_ps_poll 440 - -440
ieee80211_rx_h_decrypt 508 - -508
ieee80211_rx_h_check 588 - -588
ieee80211_rx_h_amsdu 872 - -872
ieee80211_rx_h_sta_process 920 - -920
ieee80211_rx_h_defragment 2004 - -2004
All three patches together give:
add/remove: 1/23 grow/shrink: 1/2 up/down: 8924/-10404 (-1480)
function old new delta
ieee80211_invoke_rx_handlers 724 6256 +5532
invoke_tx_handlers - 3392 +3392
ieee80211_tx_handlers 48 - -48
ieee80211_rx_handlers 52 - -52
ieee80211_tx_h_sequence 116 - -116
ieee80211_rx_h_mgmt 116 - -116
ieee80211_get_buffered_bc 620 484 -136
ieee80211_master_start_xmit 1152 1008 -144
ieee80211_rx_h_remove_qos_control 148 - -148
ieee80211_tx_h_stats 168 - -168
ieee80211_tx_h_encrypt 168 - -168
ieee80211_rx_h_passive_scan 168 - -168
ieee80211_rx_h_data 192 - -192
ieee80211_tx_h_select_key 256 - -256
ieee80211_tx_h_calculate_duration 276 - -276
ieee80211_rx_h_ctrl 288 - -288
ieee80211_tx_h_rate_ctrl 344 - -344
ieee80211_tx_h_check_assoc 344 - -344
ieee80211_rx_h_ps_poll 440 - -440
ieee80211_rx_h_decrypt 508 - -508
ieee80211_rx_h_check 588 - -588
ieee80211_tx_h_fragment 640 - -640
ieee80211_tx_h_misc 724 - -724
ieee80211_tx_h_ps_buf 744 - -744
ieee80211_rx_h_amsdu 872 - -872
ieee80211_rx_h_sta_process 920 - -920
ieee80211_rx_h_defragment 2004 - -2004
---
net/mac80211/rx.c | 71 +++++++++++++++++++++++-------------------------------
1 file changed, 31 insertions(+), 40 deletions(-)
--- everything.orig/net/mac80211/rx.c 2008-06-19 14:30:13.000000000 +0200
+++ everything/net/mac80211/rx.c 2008-06-19 14:36:38.000000000 +0200
@@ -1732,66 +1732,57 @@ static void ieee80211_rx_cooked_monitor(
dev_kfree_skb(skb);
}
-typedef ieee80211_rx_result (*ieee80211_rx_handler)(struct ieee80211_rx_data *);
-static ieee80211_rx_handler ieee80211_rx_handlers[] =
-{
- ieee80211_rx_h_passive_scan,
- ieee80211_rx_h_check,
- ieee80211_rx_h_decrypt,
- ieee80211_rx_h_sta_process,
- ieee80211_rx_h_defragment,
- ieee80211_rx_h_ps_poll,
- ieee80211_rx_h_michael_mic_verify,
- /* this must be after decryption - so header is counted in MPDU mic
- * must be before pae and data, so QOS_DATA format frames
- * are not passed to user space by these functions
- */
- ieee80211_rx_h_remove_qos_control,
- ieee80211_rx_h_amsdu,
- ieee80211_rx_h_data,
- ieee80211_rx_h_ctrl,
- ieee80211_rx_h_mgmt,
- NULL
-};
static void ieee80211_invoke_rx_handlers(struct ieee80211_sub_if_data *sdata,
struct ieee80211_rx_data *rx,
struct sk_buff *skb)
{
- ieee80211_rx_handler *handler;
ieee80211_rx_result res = RX_DROP_MONITOR;
rx->skb = skb;
rx->sdata = sdata;
rx->dev = sdata->dev;
- for (handler = ieee80211_rx_handlers; *handler != NULL; handler++) {
- res = (*handler)(rx);
+#define CALL_RXH(rxh) if ((res = rxh(rx)) != RX_CONTINUE) goto rxh_done;
+ CALL_RXH(ieee80211_rx_h_passive_scan);
+ CALL_RXH(ieee80211_rx_h_check);
+ CALL_RXH(ieee80211_rx_h_decrypt);
+ CALL_RXH(ieee80211_rx_h_sta_process);
+ CALL_RXH(ieee80211_rx_h_defragment);
+ CALL_RXH(ieee80211_rx_h_ps_poll);
+ CALL_RXH(ieee80211_rx_h_michael_mic_verify);
+ /*
+ * this must be after decryption - so header is counted in MPDU mic
+ * must be before pae and data); so QOS_DATA format frames
+ * are not passed to user space by these functions
+ */
+ CALL_RXH(ieee80211_rx_h_remove_qos_control);
+ CALL_RXH(ieee80211_rx_h_amsdu);
+ CALL_RXH(ieee80211_rx_h_data);
+ CALL_RXH(ieee80211_rx_h_ctrl);
+ CALL_RXH(ieee80211_rx_h_mgmt);
- switch (res) {
- case RX_CONTINUE:
- continue;
- case RX_DROP_UNUSABLE:
- case RX_DROP_MONITOR:
- I802_DEBUG_INC(sdata->local->rx_handlers_drop);
- if (rx->sta)
- rx->sta->rx_dropped++;
- break;
- case RX_QUEUED:
- I802_DEBUG_INC(sdata->local->rx_handlers_queued);
- break;
- }
- break;
- }
+#undef CALL_RXH
+ rxh_done:
switch (res) {
- case RX_CONTINUE:
case RX_DROP_MONITOR:
+ I802_DEBUG_INC(sdata->local->rx_handlers_drop);
+ if (rx->sta)
+ rx->sta->rx_dropped++;
+ /* fall through */
+ case RX_CONTINUE:
ieee80211_rx_cooked_monitor(rx);
break;
case RX_DROP_UNUSABLE:
+ I802_DEBUG_INC(sdata->local->rx_handlers_drop);
+ if (rx->sta)
+ rx->sta->rx_dropped++;
dev_kfree_skb(rx->skb);
break;
+ case RX_QUEUED:
+ I802_DEBUG_INC(sdata->local->rx_handlers_queued);
+ break;
}
}
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [RFC 3/2] get rid of rx handler function pointers
2008-06-19 12:38 [RFC 3/2] get rid of rx handler function pointers Johannes Berg
@ 2008-06-19 17:20 ` Harvey Harrison
2008-06-19 17:24 ` Johannes Berg
2008-06-19 17:44 ` Michael Buesch
0 siblings, 2 replies; 5+ messages in thread
From: Harvey Harrison @ 2008-06-19 17:20 UTC (permalink / raw)
To: Johannes Berg; +Cc: linux-wireless
On Thu, 2008-06-19 at 14:38 +0200, Johannes Berg wrote:
> Even better :)
>=20
> add/remove: 0/12 grow/shrink: 1/0 up/down: 5532/-6296 (-764)
> function old new delta
> ieee80211_invoke_rx_handlers 724 6256 +5532
> ieee80211_rx_handlers 52 - -52
> ieee80211_rx_h_mgmt 116 - -116
> ieee80211_rx_h_remove_qos_control 148 - -148
> ieee80211_rx_h_passive_scan 168 - -168
> ieee80211_rx_h_data 192 - -192
> ieee80211_rx_h_ctrl 288 - -288
> ieee80211_rx_h_ps_poll 440 - -440
> ieee80211_rx_h_decrypt 508 - -508
> ieee80211_rx_h_check 588 - -588
> ieee80211_rx_h_amsdu 872 - -872
> ieee80211_rx_h_sta_process 920 - -920
> ieee80211_rx_h_defragment 2004 - -2004
>=20
>=20
> All three patches together give:
>=20
> add/remove: 1/23 grow/shrink: 1/2 up/down: 8924/-10404 (-1480)
> function old new delta
> ieee80211_invoke_rx_handlers 724 6256 +5532
> invoke_tx_handlers - 3392 +3392
> ieee80211_tx_handlers 48 - -48
> ieee80211_rx_handlers 52 - -52
> ieee80211_tx_h_sequence 116 - -116
> ieee80211_rx_h_mgmt 116 - -116
> ieee80211_get_buffered_bc 620 484 -136
> ieee80211_master_start_xmit 1152 1008 -144
> ieee80211_rx_h_remove_qos_control 148 - -148
> ieee80211_tx_h_stats 168 - -168
> ieee80211_tx_h_encrypt 168 - -168
> ieee80211_rx_h_passive_scan 168 - -168
> ieee80211_rx_h_data 192 - -192
> ieee80211_tx_h_select_key 256 - -256
> ieee80211_tx_h_calculate_duration 276 - -276
> ieee80211_rx_h_ctrl 288 - -288
> ieee80211_tx_h_rate_ctrl 344 - -344
> ieee80211_tx_h_check_assoc 344 - -344
> ieee80211_rx_h_ps_poll 440 - -440
> ieee80211_rx_h_decrypt 508 - -508
> ieee80211_rx_h_check 588 - -588
> ieee80211_tx_h_fragment 640 - -640
> ieee80211_tx_h_misc 724 - -724
> ieee80211_tx_h_ps_buf 744 - -744
> ieee80211_rx_h_amsdu 872 - -872
> ieee80211_rx_h_sta_process 920 - -920
> ieee80211_rx_h_defragment 2004 - -2004
>=20
>=20
> ---
> net/mac80211/rx.c | 71 +++++++++++++++++++++++--------------------=
-----------
> 1 file changed, 31 insertions(+), 40 deletions(-)
>=20
> --- everything.orig/net/mac80211/rx.c 2008-06-19 14:30:13.000000000 +=
0200
> +++ everything/net/mac80211/rx.c 2008-06-19 14:36:38.000000000 +0200
> @@ -1732,66 +1732,57 @@ static void ieee80211_rx_cooked_monitor(
> dev_kfree_skb(skb);
> }
> =20
> -typedef ieee80211_rx_result (*ieee80211_rx_handler)(struct ieee80211=
_rx_data *);
> -static ieee80211_rx_handler ieee80211_rx_handlers[] =3D
> -{
> - ieee80211_rx_h_passive_scan,
> - ieee80211_rx_h_check,
> - ieee80211_rx_h_decrypt,
> - ieee80211_rx_h_sta_process,
> - ieee80211_rx_h_defragment,
> - ieee80211_rx_h_ps_poll,
> - ieee80211_rx_h_michael_mic_verify,
> - /* this must be after decryption - so header is counted in MPDU mic
> - * must be before pae and data, so QOS_DATA format frames
> - * are not passed to user space by these functions
> - */
> - ieee80211_rx_h_remove_qos_control,
> - ieee80211_rx_h_amsdu,
> - ieee80211_rx_h_data,
> - ieee80211_rx_h_ctrl,
> - ieee80211_rx_h_mgmt,
> - NULL
> -};
> =20
> static void ieee80211_invoke_rx_handlers(struct ieee80211_sub_if_dat=
a *sdata,
> struct ieee80211_rx_data *rx,
> struct sk_buff *skb)
> {
> - ieee80211_rx_handler *handler;
> ieee80211_rx_result res =3D RX_DROP_MONITOR;
> =20
> rx->skb =3D skb;
> rx->sdata =3D sdata;
> rx->dev =3D sdata->dev;
> =20
> - for (handler =3D ieee80211_rx_handlers; *handler !=3D NULL; handler=
++) {
> - res =3D (*handler)(rx);
> =EF=BB=BF#define CALL_RXH(rxh) if ((res =3D rxh(rx)) !=3D RX_CONTINUE=
) goto rxh_done;
Would it really be so bad to just open-code them rather than the macro =
approach?
res =3D =EF=BB=BFieee80211_rx_h_passive_scan(rx);
if (res !=3D RX_CONTINUE) goto rxh_done;
res =3D =EF=BB=BFieee80211_rx_h_check(rx);
if (res !=3D RX_CONTINUE) goto rxh_done;
...
Similar comment for the transmit handler patch. Either way, the approa=
ch looks good,
and the size reduction is nice.
Cheers,
Harvey
--
To unsubscribe from this list: send the line "unsubscribe linux-wireles=
s" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [RFC 3/2] get rid of rx handler function pointers
2008-06-19 17:20 ` Harvey Harrison
@ 2008-06-19 17:24 ` Johannes Berg
2008-06-19 17:44 ` Michael Buesch
1 sibling, 0 replies; 5+ messages in thread
From: Johannes Berg @ 2008-06-19 17:24 UTC (permalink / raw)
To: Harvey Harrison; +Cc: linux-wireless
[-- Attachment #1: Type: text/plain, Size: 603 bytes --]
> > #define CALL_RXH(rxh) if ((res = rxh(rx)) != RX_CONTINUE) goto rxh_done;
>
> Would it really be so bad to just open-code them rather than the macro approach?
>
> res = ieee80211_rx_h_passive_scan(rx);
> if (res != RX_CONTINUE) goto rxh_done;
>
> res = ieee80211_rx_h_check(rx);
> if (res != RX_CONTINUE) goto rxh_done;
>
> ...
>
>
> Similar comment for the transmit handler patch.
Well, I was lazy, and figured it'd be easier to add new ones that way if
needed.
> Either way, the approach looks good,
> and the size reduction is nice.
:)
johannes
[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 836 bytes --]
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [RFC 3/2] get rid of rx handler function pointers
2008-06-19 17:20 ` Harvey Harrison
2008-06-19 17:24 ` Johannes Berg
@ 2008-06-19 17:44 ` Michael Buesch
2008-06-19 17:46 ` Johannes Berg
1 sibling, 1 reply; 5+ messages in thread
From: Michael Buesch @ 2008-06-19 17:44 UTC (permalink / raw)
To: Harvey Harrison; +Cc: Johannes Berg, linux-wireless
On Thursday 19 June 2008 19:20:16 Harvey Harrison wrote:
> > =EF=BB=BF#define CALL_RXH(rxh) if ((res =3D rxh(rx)) !=3D RX_CONTIN=
UE) goto rxh_done;
>=20
> Would it really be so bad to just open-code them rather than the macr=
o approach?
I think the macro makes it readable. And as it is defined locally it
does not obfuscate it.
> res =3D =EF=BB=BFieee80211_rx_h_passive_scan(rx);
> if (res !=3D RX_CONTINUE) goto rxh_done;
This would most likely need to look like this, if opencoded:
res =3D =EF=BB=BFieee80211_rx_h_passive_scan(rx);
if (res !=3D RX_CONTINUE)
goto rxh_done;
>=20
> res =3D =EF=BB=BFieee80211_rx_h_check(rx);
> if (res !=3D RX_CONTINUE) goto rxh_done;
>=20
> ...
>
--=20
Greetings Michael.
--
To unsubscribe from this list: send the line "unsubscribe linux-wireles=
s" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [RFC 3/2] get rid of rx handler function pointers
2008-06-19 17:44 ` Michael Buesch
@ 2008-06-19 17:46 ` Johannes Berg
0 siblings, 0 replies; 5+ messages in thread
From: Johannes Berg @ 2008-06-19 17:46 UTC (permalink / raw)
To: Michael Buesch; +Cc: Harvey Harrison, linux-wireless
[-- Attachment #1: Type: text/plain, Size: 548 bytes --]
On Thu, 2008-06-19 at 19:44 +0200, Michael Buesch wrote:
> On Thursday 19 June 2008 19:20:16 Harvey Harrison wrote:
> > > #define CALL_RXH(rxh) if ((res = rxh(rx)) != RX_CONTINUE) goto rxh_done;
> >
> > Would it really be so bad to just open-code them rather than the macro approach?
>
> I think the macro makes it readable. And as it is defined locally it
> does not obfuscate it.
We can always make the macro more readable:
#define CALL_RXH(rxh) \
res = rxh(rx); \
if (res != RX_CONTINUE) \
goto rxh_done;
johannes
[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 836 bytes --]
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2008-06-19 17:47 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2008-06-19 12:38 [RFC 3/2] get rid of rx handler function pointers Johannes Berg
2008-06-19 17:20 ` Harvey Harrison
2008-06-19 17:24 ` Johannes Berg
2008-06-19 17:44 ` Michael Buesch
2008-06-19 17:46 ` Johannes Berg
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.