Re: ipw2200 and suspend (again)

Rob Miller wrote:
> hi,
>
> I just recently got an ipw2200 card, and can use it happily except after a 
> suspend-to-disk.  I'm using gentoo, ipw2200 1.1.3, kernel 2.6.17.  I've 
> tried this both with the in-kernel suspend code and the suspend2 patches 
> (both the gentoo versions).  I was also seeing the same failures with the 
> in-kernel ipw2200 code with 2.6.17.
>
> basically there's a problem loading the firmware after resuming, but of course it is 
> confusing as there is a complete power cycle during the suspend.  I've tried the 
> echo 100 > /sys/class/firmware/timeout fix and the /sbin/udev_run_hotplugd in 
> /etc/udev/rules that I've read about, neither with any effect on this problem (tho 
> I tried these before getting into the debug messages).
>
> my debug traces look like:
>
> fail:
>
> Aug 25 18:17:05 vista ipw2200: Intel(R) PRO/Wireless 2200/2915 Network Driver, 1.1.3mprq
> Aug 25 18:17:05 vista ipw2200: Copyright(c) 2003-2006 Intel Corporation
> Aug 25 18:17:05 vista ACPI: PCI Interrupt 0000:02:0a.0[A] -> Link [LNKH] -> GSI 11 (level, low) -> IRQ 11
> Aug 25 18:17:05 vista ipw2200: U ipw_pci_probe pci_resource_len = 0x00001000
> Aug 25 18:17:05 vista ipw2200: U ipw_pci_probe pci_resource_base = f9000000
> Aug 25 18:17:05 vista ipw2200: U ipw_sw_reset Hardware crypto [off]
> Aug 25 18:17:05 vista ipw2200: Detected Intel PRO/Wireless 2200BG Network Connection
> Aug 25 18:17:05 vista ipw2200: U ipw_get_fw Read firmware 'ipw2200-bss.fw' image v3.0 (191126 bytes)
> Aug 25 18:17:05 vista ipw2200: U ipw_load initial device response after 0ms
> Aug 25 18:17:05 vista ipw2200: U ipw_stop_master stop master 0ms
> Aug 25 18:17:05 vista ipw2200: U ipw_load_ucode Microcode is not alive
> Aug 25 18:17:05 vista ipw2200: Unable to load ucode: -22
> Aug 25 18:17:05 vista ipw2200: Unable to load firmware: -22
> Aug 25 18:17:05 vista ipw2200: failed to register network device
> Aug 25 18:17:05 vista ACPI: PCI interrupt for device 0000:02:0a.0 disabled
> Aug 25 18:17:05 vista ipw2200: probe of 0000:02:0a.0 failed with error -5
>
> -------------------
>
> succeed:
>
> Aug 25 18:21:47 vista ACPI: PCI interrupt for device 0000:02:0a.0 disabled
> Aug 25 18:21:47 vista ipw2200: Intel(R) PRO/Wireless 2200/2915 Network Driver, 1.1.3mprq
> Aug 25 18:21:47 vista ipw2200: Copyright(c) 2003-2006 Intel Corporation
> Aug 25 18:21:47 vista ACPI: PCI Interrupt 0000:02:0a.0[A] -> Link [LNKH] -> GSI 11 (level, low) -> IRQ 11
> Aug 25 18:21:47 vista ipw2200: U ipw_pci_probe pci_resource_len = 0x00001000
> Aug 25 18:21:47 vista ipw2200: U ipw_pci_probe pci_resource_base = f8860000
> Aug 25 18:21:47 vista ipw2200: U ipw_sw_reset Hardware crypto [off]
> Aug 25 18:21:47 vista ipw2200: Detected Intel PRO/Wireless 2200BG Network Connection
> Aug 25 18:21:47 vista ipw2200: U ipw_get_fw Read firmware 'ipw2200-bss.fw' image v3.0 (191126 bytes)
> Aug 25 18:21:47 vista ipw2200: U ipw_load initial device response after 10ms
> Aug 25 18:21:47 vista ipw2200: U ipw_stop_master stop master 0ms
> Aug 25 18:21:47 vista ipw2200: U ipw_load_ucode Microcode OK, rev. 53594 (0xd15a) dev. 3 (0x3) of 11/22/04 20:27
> Aug 25 18:21:47 vista ipw2200: U ipw_load device response after 50ms
> Aug 25 18:21:47 vista ipw2200: U ipw_eeprom_init_sram Writing EEPROM data into SRAM
> Aug 25 18:21:47 vista 00000000 0D 02 01 14 02 14 03 14  04 14 05 14 06 14 07 14   ........ ........
> (snip)
>
> as you may see, the first thing that looks suspicious is the immediate return from polling
> the fw_init bit ("initial response after 0ms") instead of 10ms like it should take when it is working.
>
> I can rmmod/modprobe when it is working repeatedly and it still works, I can do a full reboot (not suspend reboot-only,
> haven't tried that yet) and it works (that is, no power down).  I just can't suspend-off-resume. I am rmmod'ing all
> modules before suspending normally, though I did try this once with that turned off in the hibernate script.
>
> any ideas or pointers to try?  I have the code together where I can modify / reload /etc. so happy to test
> ideas on it.
>
> thanks,
>
> rob.
>   
It works here, however, i remember seeing a post on this mailing list
saying that there is a race condition in the kernel code for loading the
firmware which might cause what you're seeing. Search the mailing list
archives. I think the poster was pjones (@redhat.com maybe).

You could try puting a sleep 1 somewhere in the resume scripts before
loading the ipw2200 module, i don't know.
Hope this helps.
--
Pedro Ramalhais

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642