Re: kern/44060: iwn wireless works for awhile, then fails

[Donald Allen <donaldcallen%gmail.com@localhost>]

I have experimented with various values of iwn_debug and while I am able to reproduce this problem reliably, I am not seeing anything in the debugging output that is noticeably different when the wireless is working vs. not working.

I have also tried diff'ing the netbsd driver and the openbsd driver (which does not exhibit this problem), which, as I've observed, have common ancestry, but the differences are extensive enough that the crucial difference is not obvious to me, especially since I don't understand the code or the hardware.

I really think that the best way to debug this is for someone who knows the iwn driver to get in touch with me so that we can arrange a session together where I supply the problem and fingers and (s)he supplies the knowledge. Either that or someone who understands this driver needs to reproduce the problem herself.

On Mon, Nov 8, 2010 at 4:25 PM, Donald Allen <donaldcallen%gmail.com@localhost> wrote:

The following reply was made to PR kern/44060; it has been noted by GNATS.

From: Donald Allen <donaldcallen%gmail.com@localhost>

To: gnats-bugs%netbsd.org@localhost

Cc: kern-bug-people%netbsd.org@localhost, gnats-admin%netbsd.org@localhost, netbsd-bugs%netbsd.org@localhost

Subject: Re: kern/44060: iwn wireless works for awhile, then fails

Date: Mon, 8 Nov 2010 16:22:21 -0500

 --20cf303dd91a425c2c04949138bd

 Content-Type: text/plain; charset=ISO-8859-1

 On Mon, Nov 8, 2010 at 4:15 PM, David Holland <dholland-bugs%netbsd.org@localhost>wrote:

 > The following reply was made to PR kern/44060; it has been noted by GNATS.
 >
 > From: David Holland <dholland-bugs%netbsd.org@localhost>
 > To: gnats-bugs%netbsd.org@localhost
 > Cc:
 > Subject: Re: kern/44060: iwn wireless works for awhile, then fails
 > Date: Mon, 8 Nov 2010 21:10:50 +0000
 >
 >  On Mon, Nov 08, 2010 at 08:55:03PM +0000, Donald Allen wrote:
 >  >  Adding options IWN_DEBUG produces
 >  >
 >  >  ../../../../dev/pci/if_iwn.c:190:1: error: "IWN_DEBUG" redefined
 >  >
 >  >  during make depend.
 >  >
 >  >  IWN_DEBUG is already #define'ed in if_iwn.c, but the variable iwn_debug
 > is
 >  >  set to 0.
 >
 >  Oops, I was looking at a -current tree, and it looks like there's been
 >  a lot of changes between netbsd-5 and current.
 >
 >  >  Looks like the right thing is to set it non-zero, e.g., to 5 for
 >  >  the maximum amount of chatter, by changing line 195 to
 >  >
 >  >  int iwn_debug = 5;
 >  >
 >  >  I will give this a try.
 >
 >  That will hopefully do the trick...
 >

 5 is much too enthusiastic. It talks so much it reminds me of some of my
 relatives. Couldn't even get the system to come up, with all the chatter. I
 hit the power button, turned off the iwn stuff in the bios, and rebuilt the
 kernel with iwn_debug=1. I'm running with that now. We'll see ...


 >
 >  --
 >  David A. Holland
 >  dholland%netbsd.org@localhost
 >
 >

 --20cf303dd91a425c2c04949138bd

 Content-Type: text/html; charset=ISO-8859-1
 Content-Transfer-Encoding: quoted-printable

 <br><br><div class=3D"gmail_quote">On Mon, Nov 8, 2010 at 4:15 PM, David Ho=

 lland <span dir=3D"ltr">&lt;<a href="" href="mailto:dholland-bugs%netbsd.org@localhost">dholland-bugs%netbsd.org@localhost">dho=
 lland-bugs%netbsd.org@localhost</a>&gt;</span> wrote:<br><blockquote class=3D"gmail_q=
 uote" style=3D"margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 2=
 04, 204); padding-left: 1ex;">
 <div class=3D"im">The following reply was made to PR kern/44060; it has bee=
 n noted by GNATS.<br>
 <br>

 From: David Holland &lt;<a href="" href="mailto:dholland-bugs%netbsd.org@localhost">dholland-bugs%netbsd.org@localhost">dhollan=
 d-bugs%netbsd.org@localhost</a>&gt;<br>
 </div>To: <a href="" href="mailto:gnats-bugs%netbsd.org@localhost">gnats-bugs%netbsd.org@localhost">gnats-bugs%netbsd.org@localhost</a=

 ><br>
 <div class=3D"im">Cc:<br>

 Subject: Re: kern/44060: iwn wireless works for awhile, then fails<br>
 </div>Date: Mon, 8 Nov 2010 21:10:50 +0000<br>

 <div class=3D"im"><br>

 =A0On Mon, Nov 08, 2010 at 08:55:03PM +0000, Donald Allen wrote:<br>
 =A0&gt; =A0Adding options IWN_DEBUG produces<br>
 =A0&gt;<br>
 =A0&gt; =A0../../../../dev/pci/if_iwn.c:190:1: error: &quot;IWN_DEBUG&quot=
 ; redefined<br>
 =A0&gt;<br>
 =A0&gt; =A0during make depend.<br>
 =A0&gt;<br>
 =A0&gt; =A0IWN_DEBUG is already #define&#39;ed in if_iwn.c, but the variab=
 le iwn_debug is<br>
 =A0&gt; =A0set to 0.<br>
 <br>
 </div>=A0Oops, I was looking at a -current tree, and it looks like there&#3=
 9;s been<br>
 =A0a lot of changes between netbsd-5 and current.<br>

 <div class=3D"im"><br>

 =A0&gt; =A0Looks like the right thing is to set it non-zero, e.g., to 5 fo=
 r<br>
 =A0&gt; =A0the maximum amount of chatter, by changing line 195 to<br>
 =A0&gt;<br>
 =A0&gt; =A0int iwn_debug =3D 5;<br>
 =A0&gt;<br>
 =A0&gt; =A0I will give this a try.<br>
 <br>
 </div>=A0That will hopefully do the trick...<br></blockquote><div><br>5 is =
 much too enthusiastic. It talks so much it reminds me of some of my relativ=
 es. Couldn&#39;t even get the system to come up, with all the chatter. I hi=
 t the power button, turned off the iwn stuff in the bios, and rebuilt the k=
 ernel with iwn_debug=3D1. I&#39;m running with that now. We&#39;ll see ...<=
 br>
 =A0<br></div><blockquote class=3D"gmail_quote" style=3D"margin: 0pt 0pt 0pt=

 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
 <br>
 =A0--<br>

 <div><div></div><div class=3D"h5">=A0David A. Holland<br>

 =A0<a href="" href="mailto:dholland%netbsd.org@localhost">dholland%netbsd.org@localhost">dholland%netbsd.org@localhost</a><br>
 <br>

 </div></div></blockquote></div><br>

 --20cf303dd91a425c2c04949138bd--