several discrepancies with new -2139 kernel.

Tue Jun 27 03:32:11 UTC 2006

Lonni J Friedman wrote:
> On 6/26/06, Gene Heskett <gene.heskett at verizon.net> wrote:
>> Lonni J Friedman wrote:
>> > On 6/26/06, Gene Heskett <gene.heskett at verizon.net> wrote:
>> >> Greetings;
>> >>
>> >> HP DV5320US lappy, with amd64 turion, but running 32 bit FC5.
>> >>
>> >> Since 2139 has been out for about a week now, I thought I'd see how 
>> much
>> >> of the system would die if I booted to it, so I tried.
>> >>
>> >> First, from /var/log/messages:
>> >>
>> >> Jun 26 18:09:32 diablo kernel: Kernel command line: ro root=LABEL=/
>> >> irqpoll noapic noapci pci=assign-busses lapic
>> >> Jun 26 18:09:32 diablo kernel: Misrouted IRQ fixup and polling support
>> >> enabled
>> >> Jun 26 18:09:32 diablo kernel: This may significantly impact system
>> >> performance
>> >>
>> >> Thats probably been there before, and yes, this box does seem to be 
>> slow
>> >> for a 1.8GHZ processor.  Can anyone comment?
>> >
>> > That looks like a BIOS bug.  The BIOS routes the IRQs most of the
>> > time. Have you verified that you're using the latest BIOS?
>> >
>>
>> Hp seems to have those rather well hidden, url please?
> 
> Someone else seems to have already done your homework for you in another 
> post.
> 
>> >> Then later:
>> >> Jun 26 18:09:34 diablo kernel: NET: Registered protocol family 16
>> >> Jun 26 18:09:34 diablo kernel: ACPI: bus type pci registered
>> >> Jun 26 18:09:34 diablo kernel: PCI: Using MMCONFIG
>> >> Jun 26 18:09:34 diablo kernel: PCI: No mmconfig possible on 0:18
>> >>
>> >> No idea what 0:18 might represent in the hardware.  How do I define 
>> this?
>> >
>> > Its the PCI bus ID of the device.  See lspci output.
>>
>> Ok, that would be:
>>
>> 00:18.0 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron]
>> HyperTransport Technology Configuration
>> 00:18.1 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron]
>> Address Map
>> 00:18.2 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron]
>> DRAM Controller
>> 00:18.3 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron]
>> Miscellaneous Control
>>
>> which sounds pretty important to me.
> 
> That's the memory controller.  I could be wrong, but I think this is
> yet another BIOS issue.
> 
>> >> Then a few lines later:
>> >>
>> >> Jun 26 18:09:40 diablo kernel: bcm43xx driver
>> >> Jun 26 18:09:40 diablo kernel: ACPI: PCI Interrupt Link [LNKF] enabled
>> >> at IRQ 10
>> >> Jun 26 18:09:40 diablo kernel: ACPI: PCI Interrupt 0000:03:02.0[A] ->
>> >> Link [LNKF] -> GSI 10 (level, low) -> IRQ 10
>> >> Jun 26 18:09:40 diablo kernel: bcm43xx: Chip ID 0x4318, rev 0x2
>> >> Jun 26 18:09:40 diablo kernel: bcm43xx: Number of cores: 4
>> >> Jun 26 18:09:40 diablo kernel: bcm43xx: Core 0: ID 0x800, rev 0xd,
>> >> vendor 0x4243, enabled
>> >> Jun 26 18:09:40 diablo kernel: bcm43xx: Core 1: ID 0x812, rev 0x9,
>> >> vendor 0x4243, disabled
>> >> Jun 26 18:09:40 diablo kernel: bcm43xx: Core 2: ID 0x804, rev 0xc,
>> >> vendor 0x4243, enabled
>> >> Jun 26 18:09:40 diablo kernel: bcm43xx: Core 3: ID 0x80d, rev 0x7,
>> >> vendor 0x4243, enabled
>> >> Jun 26 18:09:40 diablo kernel: bcm43xx: PHY connected
>> >> Jun 26 18:09:40 diablo kernel: bcm43xx: Detected PHY: Version: 3, Type
>> >> 2, Revision 7
>> >> Jun 26 18:09:40 diablo kernel: bcm43xx: Detected Radio: ID: 8205017f
>> >> (Manuf: 17f Ver: 2050 Rev: 8)
>> >> Jun 26 18:09:40 diablo kernel: bcm43xx: Radio turned off
>> >> Jun 26 18:09:40 diablo kernel: bcm43xx: Radio turned off
>> >>
>> >> And quite a bit later:
>> >>
>> >> Jun 26 18:09:41 diablo kernel: bcm43xx: set security called
>> >> Jun 26 18:09:41 diablo kernel: bcm43xx:    .level = 0
>> >> Jun 26 18:09:41 diablo kernel: bcm43xx:    .enabled = 0
>> >> Jun 26 18:09:41 diablo kernel: bcm43xx:    .encrypt = 0
>> >> Jun 26 18:09:41 diablo kernel: SoftMAC: Associate: Scanning for 
>> networks
>> >> first.
>> >> Jun 26 18:09:41 diablo kernel: SoftMAC: Associate: failed to initiate
>> >> scan. Is device up?
>> >> Jun 26 18:09:41 diablo kernel: bcm43xx: PHY connected
>> >> Jun 26 18:09:41 diablo kernel: bcm43xx: Error: Microcode
>> >> "bcm43xx_microcode5.fw" not available or load failed.
>> >>
>> >> Now I've asked whats the procedure to switch from ndiswrapper to
>> >> actually using the bmc43xx.ko driver thats now part of the kernel 
>> tree,
>> >> and have been ignored.  Hell, even a link to rtfm would be fine.
>> >>
>> >> So I had modified my /etc/modprobe.conf
>> >>
>> >> options ndiswrapper if_name=wlan0
>> >> alias wlan0 ndiswrapper
>> >> alias eth1 tg3
>> >> #alias wlan0 bcm43xx
>> >>
>> >> to
>> >> #options ndiswrapper if_name=wlan0
>> >> #alias wlan0 ndiswrapper
>> >> alias eth1 tg3
>> >> alias wlan0 bcm43xx
>> >>
>> >> Which of course didn't work.
>> >
>> > Perhaps it wants eth0 or wifi0 ?   What does "doesn't work" really
>> > mean?  Did you get an error?
>> >
>> Those errors that occured are posted above.  And of course bringing
>> wlan0 up failed twice, as seen in the bootup screen scrolling by, once
>> when bringing up the interface, and once when dhcp tried to obtain and
>> IP address for the machine from the motels WRG54 router.
>>
>> If the name of the device has been changed from wlan0 to wifi0, why was
>> there not a notice of that change ON THIS LIST?  And what would I need
> 
> huh?  You have a rather bizarre sense of self-entitlement.  You
> changed drivers.  You assumed that the interface used by both was the

and this goddamned touchpad on this lappy cannot be disabled, so parts 
of the above line got overwritten when it decided to move the cursor 
without bothering to advise me.  If someone can advise me on howto 
disable it, I'll send them a jug of a local wine we think is pretty good 
for a commodity wine.

   Is this list responsible for providing you with personalized
> notification of everything that is relevant to your own circumstances?

Well since it makes a huge difference in all users of the broadcom 
radios being able to get onto the net, I would have thought it might 
have been worth an honorable mention.  Particularly since I did ask at 
least twice.

>> to change besides the name and contents of ifcfg-wlan0 file, made triply
>> difficult by that file existing in 3 seperate locations via hard links.
>>   It took me about a week to discover this and figure out a way to
>> actually make a change in the MAC address to take globally.  What I
>> think of that isn't printable...  But it would be nice if one knew which
>> file of the 3 was the actual master copy so we don't waste hours, even
>> days redoing such an edit & then looking at the log to see the old MAC
>> address still being used.
> 
> I'm not sure which 3 separate locations you're talking about.  The
> only place I ever touch is /etc/sysconfig/network-scripts/ifcfg-*
> You'd also need to update the alias in /etc/modprobe.conf.
> 

and that would take the form of?

As for the 3 hardlinked copies of ifcfg-wlan0, Std fc5 install, with 
nearly all updates, so I'll let this speak for me:

[root at diablo bcm43xx]# ls -li `locate ifcfg-wlan0`
12900218 -rw-r--r-- 3 root root 303 May 22 21:34 
/etc/sysconfig/networking/devices/ifcfg-wlan0
12900218 -rw-r--r-- 3 root root 303 May 22 21:34 
/etc/sysconfig/networking/profiles/default/ifcfg-wlan0
12900218 -rw-r--r-- 3 root root 303 May 22 21:34 
/etc/sysconfig/network-scripts/ifcfg-wlan0

Its a known problem, or feature, and I fail to find the feature when the 
only way to edit it is to edit all 3 copies BEFORE next accessing it 
with a 'service network restart'.  IMO its a pita, not a feature.

-- 
Cheers, Gene