garbaged AP mac address in dmesg in wireless-testing?

garbaged AP mac address in dmesg in wireless-testing?

[Hin-Tak Leung]

Since a few days ago, my dmesg is looking like this - instead of the
older AA:BB:CC:DD:EE:FF form
with the mac address of the AP.
That ffff88* looks like a pointer/address to me? Is this change intentional?

--------------
wlan2 direct probe responded
wlan2: authenticate with AP ffff880070995aa0
wlan2: authenticated
wlan2: associate with AP ffff880070995aa0
wlan2: RX ReassocResp from ffff88001431f01a (capab=0x431 status=0 aid=1)
wlan2: associated
-------------

Re: garbaged AP mac address in dmesg in wireless-testing?

[John W. Linville]

On Fri, Jan 16, 2009 at 05:07:50PM +0000, Hin-Tak Leung wrote:
> Since a few days ago, my dmesg is looking like this - instead of the
> older AA:BB:CC:DD:EE:FF form
> with the mac address of the AP.
> That ffff88* looks like a pointer/address to me? Is this change intentional?
> 
> --------------
> wlan2 direct probe responded
> wlan2: authenticate with AP ffff880070995aa0
> wlan2: authenticated
> wlan2: associate with AP ffff880070995aa0
> wlan2: RX ReassocResp from ffff88001431f01a (capab=0x431 status=0 aid=1)
> wlan2: associated

Odd...some problem with the "%pM" printk conversion?  What kernel
are you running?

John
-- 
John W. Linville		Someday the world will need a hero, and you
linville@tuxdriver.com			might be all we have.  Be ready.

Re: garbaged AP mac address in dmesg in wireless-testing?

[Kalle Valo]

"John W. Linville" <linville@tuxdriver.com> writes:

> On Fri, Jan 16, 2009 at 05:07:50PM +0000, Hin-Tak Leung wrote:
>> Since a few days ago, my dmesg is looking like this - instead of the
>> older AA:BB:CC:DD:EE:FF form
>> with the mac address of the AP.
>> That ffff88* looks like a pointer/address to me? Is this change intentional?
>> 
>> --------------
>> wlan2 direct probe responded
>> wlan2: authenticate with AP ffff880070995aa0
>> wlan2: authenticated
>> wlan2: associate with AP ffff880070995aa0
>> wlan2: RX ReassocResp from ffff88001431f01a (capab=0x431 status=0 aid=1)
>> wlan2: associated

Works for me:

[37125.111672] wlan0: authenticate with AP 00:12:17:e7:98:de
[37125.113509] wlan0: authenticated
[37125.113514] wlan0: associate with AP 00:12:17:e7:98:de
[37125.115875] wlan0: RX AssocResp from 00:12:17:e7:98:de (capab=0x11 status=0 aid=1)
[37125.115881] wlan0: associated

This is on x86 32bit with iwl3945 and latest wireless-testing.

> Odd...some problem with the "%pM" printk conversion?  What kernel
> are you running?

And on what architecture?

-- 
Kalle Valo

Re: garbaged AP mac address in dmesg in wireless-testing?

[Hauke Mehrtens]

[-- Attachment #1: Type: text/plain, Size: 1564 bytes --]

Kalle Valo wrote:
> "John W. Linville" <linville@tuxdriver.com> writes:
> 
>> On Fri, Jan 16, 2009 at 05:07:50PM +0000, Hin-Tak Leung wrote:
>>> Since a few days ago, my dmesg is looking like this - instead of the
>>> older AA:BB:CC:DD:EE:FF form
>>> with the mac address of the AP.
>>> That ffff88* looks like a pointer/address to me? Is this change intentional?
>>>
>>> --------------
>>> wlan2 direct probe responded
>>> wlan2: authenticate with AP ffff880070995aa0
>>> wlan2: authenticated
>>> wlan2: associate with AP ffff880070995aa0
>>> wlan2: RX ReassocResp from ffff88001431f01a (capab=0x431 status=0 aid=1)
>>> wlan2: associated
> 
> Works for me:
> 
> [37125.111672] wlan0: authenticate with AP 00:12:17:e7:98:de
> [37125.113509] wlan0: authenticated
> [37125.113514] wlan0: associate with AP 00:12:17:e7:98:de
> [37125.115875] wlan0: RX AssocResp from 00:12:17:e7:98:de (capab=0x11 status=0 aid=1)
> [37125.115881] wlan0: associated
> 
> This is on x86 32bit with iwl3945 and latest wireless-testing.
> 
>> Odd...some problem with the "%pM" printk conversion?  What kernel
>> are you running?
> 
> And on what architecture?
> 

I have got the same problem on x86_64 with rtl8187, kernel 2.6.27 and
compat-wireless-2008-01-16.

[ 3064.480766] wlan0: authenticate with AP ffff8801291dfa20
[ 3064.482271] wlan0: authenticated
[ 3064.482275] wlan0: associate with AP ffff8801291dfa20
[ 3064.485021] wlan0: RX AssocResp from ffff88011128c01a (capab=0x411
status=0 aid=1)
[ 3064.485025] wlan0: associated

Hauke


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 898 bytes --]

Re: garbaged AP mac address in dmesg in wireless-testing?

[John W. Linville]

On Fri, Jan 16, 2009 at 08:13:28PM +0100, Hauke Mehrtens wrote:
> Kalle Valo wrote:
> > "John W. Linville" <linville@tuxdriver.com> writes:
> > 
> >> On Fri, Jan 16, 2009 at 05:07:50PM +0000, Hin-Tak Leung wrote:
> >>> Since a few days ago, my dmesg is looking like this - instead of the
> >>> older AA:BB:CC:DD:EE:FF form
> >>> with the mac address of the AP.
> >>> That ffff88* looks like a pointer/address to me? Is this change intentional?
> >>>
> >>> --------------
> >>> wlan2 direct probe responded
> >>> wlan2: authenticate with AP ffff880070995aa0
> >>> wlan2: authenticated
> >>> wlan2: associate with AP ffff880070995aa0
> >>> wlan2: RX ReassocResp from ffff88001431f01a (capab=0x431 status=0 aid=1)
> >>> wlan2: associated
> > 
> > Works for me:
> > 
> > [37125.111672] wlan0: authenticate with AP 00:12:17:e7:98:de
> > [37125.113509] wlan0: authenticated
> > [37125.113514] wlan0: associate with AP 00:12:17:e7:98:de
> > [37125.115875] wlan0: RX AssocResp from 00:12:17:e7:98:de (capab=0x11 status=0 aid=1)
> > [37125.115881] wlan0: associated
> > 
> > This is on x86 32bit with iwl3945 and latest wireless-testing.
> > 
> >> Odd...some problem with the "%pM" printk conversion?  What kernel
> >> are you running?
> > 
> > And on what architecture?
> > 
> 
> I have got the same problem on x86_64 with rtl8187, kernel 2.6.27 and
> compat-wireless-2008-01-16.
> 
> [ 3064.480766] wlan0: authenticate with AP ffff8801291dfa20
> [ 3064.482271] wlan0: authenticated
> [ 3064.482275] wlan0: associate with AP ffff8801291dfa20
> [ 3064.485021] wlan0: RX AssocResp from ffff88011128c01a (capab=0x411
> status=0 aid=1)
> [ 3064.485025] wlan0: associated

compat-wireless users are probably using kernels that have printk
implementations that do not understand "%pM".  QED. :-)

-- 
John W. Linville		Someday the world will need a hero, and you
linville@tuxdriver.com			might be all we have.  Be ready.

Re: garbaged AP mac address in dmesg in wireless-testing?

[Hin-Tak Leung]

On Fri, Jan 16, 2009 at 7:20 PM, John W. Linville
<linville@tuxdriver.com> wrote:
> On Fri, Jan 16, 2009 at 08:13:28PM +0100, Hauke Mehrtens wrote:
>> Kalle Valo wrote:
>> > "John W. Linville" <linville@tuxdriver.com> writes:
>> >
>> >> On Fri, Jan 16, 2009 at 05:07:50PM +0000, Hin-Tak Leung wrote:
>> >>> Since a few days ago, my dmesg is looking like this - instead of the
>> >>> older AA:BB:CC:DD:EE:FF form
>> >>> with the mac address of the AP.
>> >>> That ffff88* looks like a pointer/address to me? Is this change intentional?
>> >>>
>> >>> --------------
>> >>> wlan2 direct probe responded
>> >>> wlan2: authenticate with AP ffff880070995aa0
>> >>> wlan2: authenticated
>> >>> wlan2: associate with AP ffff880070995aa0
>> >>> wlan2: RX ReassocResp from ffff88001431f01a (capab=0x431 status=0 aid=1)
>> >>> wlan2: associated
>> >
>> > Works for me:
>> >
>> > [37125.111672] wlan0: authenticate with AP 00:12:17:e7:98:de
>> > [37125.113509] wlan0: authenticated
>> > [37125.113514] wlan0: associate with AP 00:12:17:e7:98:de
>> > [37125.115875] wlan0: RX AssocResp from 00:12:17:e7:98:de (capab=0x11 status=0 aid=1)
>> > [37125.115881] wlan0: associated
>> >
>> > This is on x86 32bit with iwl3945 and latest wireless-testing.
>> >
>> >> Odd...some problem with the "%pM" printk conversion?  What kernel
>> >> are you running?
>> >
>> > And on what architecture?
>> >
>>
>> I have got the same problem on x86_64 with rtl8187, kernel 2.6.27 and
>> compat-wireless-2008-01-16.
>>
>> [ 3064.480766] wlan0: authenticate with AP ffff8801291dfa20
>> [ 3064.482271] wlan0: authenticated
>> [ 3064.482275] wlan0: associate with AP ffff8801291dfa20
>> [ 3064.485021] wlan0: RX AssocResp from ffff88011128c01a (capab=0x411
>> status=0 aid=1)
>> [ 3064.485025] wlan0: associated
>
> compat-wireless users are probably using kernels that have printk
> implementations that do not understand "%pM".  QED. :-)
>
> --
> John W. Linville                Someday the world will need a hero, and you
> linville@tuxdriver.com                  might be all we have.  Be ready.
>

2.6.28.1-9.rc2.fc10.x86_64 and yes, I am using compat-wireless ...
what puzzles me is
that the as-shipped modules have the mac addresses correctly.

Okay, sorry for the noise. I did look at git blame and, when I saw the
date of the lines with the
%pM commits was two months ago, I didn't bother to look further, since
the problem is very recent...
I was looking in the right direction but just didn't look far enough.

commit 0c68ae2605dbcf67414d8d1f19af93be44b355fb
Author: Johannes Berg <johannes@sipsolutions.net>
Date:   Mon Oct 27 15:56:10 2008 -0700

    mac80211: convert to %pM away from print_mac

    Also remove a few stray DECLARE_MAC_BUF that were no longer
    used at all.

    Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
    Signed-off-by: David S. Miller <davem@davemloft.net>

--------------------------------
commit dd45c9cf687682c9ce256ab14bd8914db77410bb
Author: Harvey Harrison <harvey.harrison@gmail.com>
Date:   Mon Oct 27 15:47:12 2008 -0700

    printk: add %pM format specifier for MAC addresses
--------------

and it is in lib/vsprintf.c ...

But I guess ideally compat-wireless should patch this in to somewhere
under net/mac80211? (that's the whole point
of compat-wireless, to run with older kernels?)

Re: garbaged AP mac address in dmesg in wireless-testing?

[Hin-Tak Leung]

On Fri, Jan 16, 2009 at 10:06 PM, Hin-Tak Leung <hintak.leung@gmail.com> wrote:
> On Fri, Jan 16, 2009 at 7:20 PM, John W. Linville
> <linville@tuxdriver.com> wrote:
>> On Fri, Jan 16, 2009 at 08:13:28PM +0100, Hauke Mehrtens wrote:
>>> Kalle Valo wrote:
>>> > "John W. Linville" <linville@tuxdriver.com> writes:
>>> >
>>> >> On Fri, Jan 16, 2009 at 05:07:50PM +0000, Hin-Tak Leung wrote:
>>> >>> Since a few days ago, my dmesg is looking like this - instead of the
>>> >>> older AA:BB:CC:DD:EE:FF form
>>> >>> with the mac address of the AP.
>>> >>> That ffff88* looks like a pointer/address to me? Is this change intentional?
>>> >>>
>>> >>> --------------
>>> >>> wlan2 direct probe responded
>>> >>> wlan2: authenticate with AP ffff880070995aa0
>>> >>> wlan2: authenticated
>>> >>> wlan2: associate with AP ffff880070995aa0
>>> >>> wlan2: RX ReassocResp from ffff88001431f01a (capab=0x431 status=0 aid=1)
>>> >>> wlan2: associated
>>> >
>>> > Works for me:
>>> >
>>> > [37125.111672] wlan0: authenticate with AP 00:12:17:e7:98:de
>>> > [37125.113509] wlan0: authenticated
>>> > [37125.113514] wlan0: associate with AP 00:12:17:e7:98:de
>>> > [37125.115875] wlan0: RX AssocResp from 00:12:17:e7:98:de (capab=0x11 status=0 aid=1)
>>> > [37125.115881] wlan0: associated
>>> >
>>> > This is on x86 32bit with iwl3945 and latest wireless-testing.
>>> >
>>> >> Odd...some problem with the "%pM" printk conversion?  What kernel
>>> >> are you running?
>>> >
>>> > And on what architecture?
>>> >
>>>
>>> I have got the same problem on x86_64 with rtl8187, kernel 2.6.27 and
>>> compat-wireless-2008-01-16.
>>>
>>> [ 3064.480766] wlan0: authenticate with AP ffff8801291dfa20
>>> [ 3064.482271] wlan0: authenticated
>>> [ 3064.482275] wlan0: associate with AP ffff8801291dfa20
>>> [ 3064.485021] wlan0: RX AssocResp from ffff88011128c01a (capab=0x411
>>> status=0 aid=1)
>>> [ 3064.485025] wlan0: associated
>>
>> compat-wireless users are probably using kernels that have printk
>> implementations that do not understand "%pM".  QED. :-)
>>
>> --
>> John W. Linville                Someday the world will need a hero, and you
>> linville@tuxdriver.com                  might be all we have.  Be ready.
>>
>
> 2.6.28.1-9.rc2.fc10.x86_64 and yes, I am using compat-wireless ...
> what puzzles me is
> that the as-shipped modules have the mac addresses correctly.
>
> Okay, sorry for the noise. I did look at git blame and, when I saw the
> date of the lines with the
> %pM commits was two months ago, I didn't bother to look further, since
> the problem is very recent...
> I was looking in the right direction but just didn't look far enough.
>
> commit 0c68ae2605dbcf67414d8d1f19af93be44b355fb
> Author: Johannes Berg <johannes@sipsolutions.net>
> Date:   Mon Oct 27 15:56:10 2008 -0700
>
>    mac80211: convert to %pM away from print_mac
>
>    Also remove a few stray DECLARE_MAC_BUF that were no longer
>    used at all.
>
>    Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
>    Signed-off-by: David S. Miller <davem@davemloft.net>
>
> --------------------------------
> commit dd45c9cf687682c9ce256ab14bd8914db77410bb
> Author: Harvey Harrison <harvey.harrison@gmail.com>
> Date:   Mon Oct 27 15:47:12 2008 -0700
>
>    printk: add %pM format specifier for MAC addresses
> --------------
>
> and it is in lib/vsprintf.c ...
>
> But I guess ideally compat-wireless should patch this in to somewhere
> under net/mac80211? (that's the whole point
> of compat-wireless, to run with older kernels?)
>

Argh, a few days ago I went from kernel-2.6.27.10-169.fc10 (heavily
wireless-testing patched from John) to
kernel-2.6.28-X.fc10 (mostly vanilla), as a routine tracking of what's
on koji...