Re: device not attaching at boot
Hello, Hans, I have not tried a -STABLE kernel. My kernel is -p2, with a -p7 userland. I will give -STABLE a try in the morning. There are no relevant USB options in the BIOS. The only settings are to disable specific ports (for security purposes). I looked for any legacy options, but there were none. I will let the list know how it goes with -STABLE. Thanks! -jan- -- Jan L. Peterson http://www.peterson-tech.com/~jlp/ Sent from TypeApp On Feb 19, 2017, 02:52, at 02:52, Hans Petter Selasky wrote: >On 02/19/17 10:04, Jan L. Peterson wrote: >> Hi, folks, >> >> I have a weird problem... >> >> Hardware is an HP dc7900 small form factor PC. OS is FreeBSD >> 11.0-RELEASE-p7. I am running the GENERIC kernel. >> >> I have a USB ethernet adapter that contains a Realtek chip... the >vendor >> ID is 0x0bda and the product ID is 0x8153. >> >> If I plug this device into the machine while the system is already up >> and running, it is detected and attached as ue0 on cdce0 (it appears >as >> ugen3.2, no matter which of the several USB ports on the machine I >plug >> it into). >> >> If the device is plugged into the machine when it boots up, nothing I >> can do (short of unplugging it and replugging it) will make the >machine >> recognize that it is there. >> >> I have tried the following: >> >> 1. played with hw.usb.timings.* values >> 2. usbconfig -d 3.1 set_config 255; usbconfig -d 3.1 set_config 0 >> (which deactivates the hub that the device is connected to and then >> reactivates it) >> 3. did the same (set_config 255; set_config 0) on *all* of the usb >hubs >> (ugen*.1) >> 4. turned on various hw.usb.*.debug options (which produced copious >> debug info, but nothing containing the strings 8153, 0bda, or >Realtek) >> 5. tried various devctl rescan options (including pci0) >> 6. googled frantically for something similar to this problem to no >> avail (well, I did come across some people with similar problems, but >> their problems all seemed to be around the device coming up with the >> wrong config and needing a SET_CFG_1 quirk set, which appears to >already >> be in the kernel as of 2015 or so) >> >> I've looked for ways to force the USB bus to rescan (hence the >usbconfig >> stuff in number 2 above) but nothing I have tried seems to work. >> >> Again, the device works fine if I unplug it from the machine and then >> plug it back in. It is detected, probes, and attaches just fine and >> dandy. The problem is that the machine is destined to be locked in a >> cabinet at a remote location where there won't be anyone available to >go >> unplug and replug it any time the machine needs a reboot. >> >> To clarify... when I have the device plugged into the machine and I >> reboot, the device is not detected or noticed by the OS in any way. >> There is nothing relevant in dmesg or /var/log/messages. usbconfig >and >> devinfo do not see it at all. If I unplug the device and plug it >back >> in, it is recognized immediately and works fine. >> >> Any ideas? >> > >Hi, > >Have you tried a 11-stable kernel? > >Have you tried disabling USB in the BIOS? > >--HPS ___ freebsd-usb@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-usb To unsubscribe, send any mail to "freebsd-usb-unsubscr...@freebsd.org"
Re: device not attaching at boot
On 02/19/17 10:04, Jan L. Peterson wrote: Hi, folks, I have a weird problem... Hardware is an HP dc7900 small form factor PC. OS is FreeBSD 11.0-RELEASE-p7. I am running the GENERIC kernel. I have a USB ethernet adapter that contains a Realtek chip... the vendor ID is 0x0bda and the product ID is 0x8153. If I plug this device into the machine while the system is already up and running, it is detected and attached as ue0 on cdce0 (it appears as ugen3.2, no matter which of the several USB ports on the machine I plug it into). If the device is plugged into the machine when it boots up, nothing I can do (short of unplugging it and replugging it) will make the machine recognize that it is there. I have tried the following: 1. played with hw.usb.timings.* values 2. usbconfig -d 3.1 set_config 255; usbconfig -d 3.1 set_config 0 (which deactivates the hub that the device is connected to and then reactivates it) 3. did the same (set_config 255; set_config 0) on *all* of the usb hubs (ugen*.1) 4. turned on various hw.usb.*.debug options (which produced copious debug info, but nothing containing the strings 8153, 0bda, or Realtek) 5. tried various devctl rescan options (including pci0) 6. googled frantically for something similar to this problem to no avail (well, I did come across some people with similar problems, but their problems all seemed to be around the device coming up with the wrong config and needing a SET_CFG_1 quirk set, which appears to already be in the kernel as of 2015 or so) I've looked for ways to force the USB bus to rescan (hence the usbconfig stuff in number 2 above) but nothing I have tried seems to work. Again, the device works fine if I unplug it from the machine and then plug it back in. It is detected, probes, and attaches just fine and dandy. The problem is that the machine is destined to be locked in a cabinet at a remote location where there won't be anyone available to go unplug and replug it any time the machine needs a reboot. To clarify... when I have the device plugged into the machine and I reboot, the device is not detected or noticed by the OS in any way. There is nothing relevant in dmesg or /var/log/messages. usbconfig and devinfo do not see it at all. If I unplug the device and plug it back in, it is recognized immediately and works fine. Any ideas? Hi, Have you tried a 11-stable kernel? Have you tried disabling USB in the BIOS? --HPS ___ freebsd-usb@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-usb To unsubscribe, send any mail to "freebsd-usb-unsubscr...@freebsd.org"
device not attaching at boot
Hi, folks, I have a weird problem... Hardware is an HP dc7900 small form factor PC. OS is FreeBSD 11.0-RELEASE-p7. I am running the GENERIC kernel. I have a USB ethernet adapter that contains a Realtek chip... the vendor ID is 0x0bda and the product ID is 0x8153. If I plug this device into the machine while the system is already up and running, it is detected and attached as ue0 on cdce0 (it appears as ugen3.2, no matter which of the several USB ports on the machine I plug it into). If the device is plugged into the machine when it boots up, nothing I can do (short of unplugging it and replugging it) will make the machine recognize that it is there. I have tried the following: 1. played with hw.usb.timings.* values 2. usbconfig -d 3.1 set_config 255; usbconfig -d 3.1 set_config 0 (which deactivates the hub that the device is connected to and then reactivates it) 3. did the same (set_config 255; set_config 0) on *all* of the usb hubs (ugen*.1) 4. turned on various hw.usb.*.debug options (which produced copious debug info, but nothing containing the strings 8153, 0bda, or Realtek) 5. tried various devctl rescan options (including pci0) 6. googled frantically for something similar to this problem to no avail (well, I did come across some people with similar problems, but their problems all seemed to be around the device coming up with the wrong config and needing a SET_CFG_1 quirk set, which appears to already be in the kernel as of 2015 or so) I've looked for ways to force the USB bus to rescan (hence the usbconfig stuff in number 2 above) but nothing I have tried seems to work. Again, the device works fine if I unplug it from the machine and then plug it back in. It is detected, probes, and attaches just fine and dandy. The problem is that the machine is destined to be locked in a cabinet at a remote location where there won't be anyone available to go unplug and replug it any time the machine needs a reboot. To clarify... when I have the device plugged into the machine and I reboot, the device is not detected or noticed by the OS in any way. There is nothing relevant in dmesg or /var/log/messages. usbconfig and devinfo do not see it at all. If I unplug the device and plug it back in, it is recognized immediately and works fine. Any ideas? -jan- -- Jan L. Peterson http://www.peterson-tech.com/~jlp/ signature.asc Description: OpenPGP digital signature