Hi,
I'm still working on an extension of the carl9170 firmware to enable TX
service time measurements. This implies the addition of a u32 field to
carl9170_tx_superdesc to record the TSF value at the time the frame is
received by the firmware.
The problem with WoWLAN, more specifically with changes introduced by
af915846d44d23adc02c7ded807282d607b46faa, is that adding that field makes
carl9170_sram_reserved::cmd larger than CARL9170_MAX_CMD_LEN.
I still don't see an easy way out other than disabling WoWLAN altogether. If
anyone has any suggestion, I'd be happy to read it.
BTW, I noticed something that may be a bug in wlan.c, in
wlan_wol_connection_monitor(void)'s definition:
memset(nullf, 0, sizeof(nullf));
Shouldn't that be
memset(nullf, 0, sizeof *nullf);
instead?
Cheers,
Ignacy
--
To err is human, to moo bovine.
On Thu, Jul 14, 2011 at 03:18:03PM +0200, thus spake Christian Lamparter:
> On Thursday, July 14, 2011 02:28:03 PM Ignacy Gawedzki wrote:
> > Hi,
> >
> > I'm still working on an extension of the carl9170 firmware to enable TX
> > service time measurements. This implies the addition of a u32 field to
> > carl9170_tx_superdesc to record the TSF value at the time the frame is
> > received by the firmware.
> >
> > The problem with WoWLAN, more specifically with changes introduced by
> > af915846d44d23adc02c7ded807282d607b46faa, is that adding that field makes
> > carl9170_sram_reserved::cmd larger than CARL9170_MAX_CMD_LEN.
> >
> > I still don't see an easy way out other than disabling WoWLAN altogether.
> > If anyone has any suggestion, I'd be happy to read it.
> Well, I think there's enough space in the rsp [fw->driver] buffer.
Okay, so I moved null to the rsp union and adapted wlan.c accordingly. It
compiles just fine indeed.
> After all in WoWLAN mode, this buffer will be unused as well?
It sounds like it makes sense. For the moment, I can only say that it works
for normal use. I would still have to check WoWLAN, though.
Thanks for your help.
Ignacy
--
The groove will take you through times without money
much better than money will take you through times without groove.
On Thursday, July 14, 2011 02:28:03 PM Ignacy Gawedzki wrote:
> Hi,
>
> I'm still working on an extension of the carl9170 firmware to enable TX
> service time measurements. This implies the addition of a u32 field to
> carl9170_tx_superdesc to record the TSF value at the time the frame is
> received by the firmware.
>
> The problem with WoWLAN, more specifically with changes introduced by
> af915846d44d23adc02c7ded807282d607b46faa, is that adding that field makes
> carl9170_sram_reserved::cmd larger than CARL9170_MAX_CMD_LEN.
>
> I still don't see an easy way out other than disabling WoWLAN altogether.
> If anyone has any suggestion, I'd be happy to read it.
Well, I think there's enough space in the rsp [fw->driver] buffer.
After all in WoWLAN mode, this buffer will be unused as well?
> BTW, I noticed something that may be a bug in wlan.c, in
> wlan_wol_connection_monitor(void)'s definition:
>
> memset(nullf, 0, sizeof(nullf));
>
> Shouldn't that be
>
> memset(nullf, 0, sizeof *nullf);
>
> instead?
Yes it should be. Thanks [fix pushed].
Regards,
Chr