On Wed, 17 Jun 2009, Josip Rodin wrote: > > > > So I thought - let me check that wwan_enable thing on 2.6.30: > > > > > > > > % cat /sys/devices/platform/thinkpad_acpi/wwan_enable > > > > 0
The firmware will power down the internal WWAN module if that thing is set to 0... so it will obviously not show up on the USB bus. > > > > thinkpad_acpi: deprecated sysfs attribute: access by process with PID > > > > 2968 > > > > thinkpad_acpi: WARNING: sysfs attribute wwan_enable is deprecated and > > > > will be removed. Please switch to generic rfkill before year 2010 I am likely going to underprecate it for 2.6.32, but whatever. I will cross that bridge when I get there. > > > > % sudo sh -c 'echo 1 > /sys/devices/platform/thinkpad_acpi/wwan_enable' > > > > thinkpad_acpi: deprecated sysfs attribute: access by process with PID > > > > 2968 > > > > thinkpad_acpi: WARNING: sysfs attribute wwan_enable is deprecated and > > > > will be removed. Please switch to generic rfkill before year 2010 > > > > usb 8-4: new high speed USB device using ehci_hcd and address 4 > > > > usb 8-4: configuration #1 chosen from 2 choices > > > > cdc_acm 8-4:1.1: ttyACM0: USB ACM device > > > > cdc_acm 8-4:1.3: ttyACM1: USB ACM device > > > > cdc_wdm 8-4:1.5: cdc-wdm0: USB WDM device > > > > cdc_wdm 8-4:1.6: cdc-wdm1: USB WDM device > > > > usb0: register 'cdc_ether' at usb-0000:00:1d.7-4, CDC Ethernet Device, > > > > 02:80:37:ec:02:00 > > > > cdc_acm 8-4:1.9: ttyACM2: USB ACM device Works as it should. Good. > > > > Yay. Now I see that I've been looking at a different scenario - while > > > > upgrading to 2.6.30, I noticed that iwlagn has RFKILL support, and > > > > enabled > > > > it. I definitely tested it at one point - maybe that tripped off the > > > > problematic situation. So now I tried using the little physical switch, > > > > on and off again, and everything seems fine. I'll try it again after a > > > > reboot, in a minute. > > > > > > Oddly enough, it works fine after a reboot - the wwan_enable is still set > > > to 1. thinkpad-acpi tries to retain state across suspend, reboots and shutdowns :P This is likely going to go weird on you in 2.6.31, though. Many rfkill core changes in that area. > > > So it looks like the remaining problem is that something somewhere seems > > > to have set that option in a way that disabled the hardware, but nothing > > > tells the user that there is a problem. Can this be fixed? > > > > I really have no idea, it doesn't sound like a USB problem, but a > > platform issue :( Well, if you guys would like that, I can certainly log to syslog the initial state of the rfkill switches. Which priority do you want? INFO? NOTICE? But please _do_ open a bug about it on bugzilla.kernel.org, just in case I have to send the fix after -rc1. > I'm moving this from linux-usb to the thinkpad-acpi maintainers, guessing > that you may know more. It is related to thinkpad-acpi, yes. > In the meantime I've rebooted as well as hibernated the laptop a few more > times, and I can't seem to establish a clear pattern by which wwan_enable > gets set or unset, because it unset itself at least twice since, without > me using the physical switch. It will be unset if something messes with the rfkill state, or if the firmware does something weird, which it just might if you press fn+f5 while the OS didn't come up yet :p And there is always the possibility of a thinkpad-acpi bug. > Help? Sure thing. Load thinkpad-acpi with the debug=0x0004 parameter, and it should be really annoying and tell you when anything changes rfkill state. Maybe compile the kernel with CONFIG_THINKPAD_ACPI_DEBUG set for extra logging. Then keep an eye on the logs, you want to check both the ones since boot for the session where you notice WWAN is disabled, as well as the ones since the boot _previous_ to that one, because something might have caused thinkpad-acpi to tell the firmware to store in NVRAM that WWAN should be disabled on the next power up... -- "One disk to rule them all, One disk to find them. One disk to bring them all and in the darkness grind them. In the Land of Redmond where the shadows lie." -- The Silicon Valley Tarot Henrique Holschuh ------------------------------------------------------------------------------ Crystal Reports - New Free Runtime and 30 Day Trial Check out the new simplified licensing option that enables unlimited royalty-free distribution of the report engine for externally facing server and web deployment. http://p.sf.net/sfu/businessobjects _______________________________________________ ibm-acpi-devel mailing list ibm-acpi-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ibm-acpi-devel
