Created attachment 191221 [details] The kernel configuration file of my system. ==Summary== Apple Thunderbolt Ethernet Adapter not Working After Suspend & Resume ==Overview== On a Retina MacBook Pro 11,3, with an Apple Thunderbolt Ethernet Adapter plugged in, the previously working adapter does not work after suspend & resume. ==Steps to Reproduce== 1) Boot the Retina MacBook Pro 11,3. Plug in the Apple Thunderbolt Ethernet Adapter. (Thunderbolt hot-plugging works on my setup.) The adapter is usable now (Link is reported up, lspci as in attachment 2) Suspend the machine (I close the lid which works when disabling 'XHCI' in `/proc/acpi/wakeup`) 3) Resume the machine (Open the lid) 4) The adapter does not work now. It is reported DOWN by `ip link`. lspci output shows different lables (see attachments). Many error messages in log output. ==Actual Result== As described above: The Thunderbolt Ethernet Adapter does not work anymore after resume. Unplug and re-plugin does not resolve the issue. (See additional information for my current workarounds) ==Expected Result== The Thunderbolt Ethernet Adapter should continue to work link before suspend. ==Build Data & Hardware== Linux 4.3.0-rc7 https://www.kernel.org/pub/linux/kernel/v4.x/testing/linux-4.3-rc7.tar.xz Kernel Config: See attachment Userland: Arch Linux Retina MacBook Pro 11,3 http://www.everymac.com/systems/apple/macbook_pro/specs/macbook-pro-core-i7-2.3-15-dual-graphics-late-2013-retina-display-specs.html Apple Thunderbolt Ethernet Adapter (labeled BCM2590 on the backside, can't explain why this identifier differs from lspci output. I think this is not of critical importance for this bug report.) ==Addtitional Information== Workaround 1: The obvious workaround is to unplug the adapter before suspending. Workaround 2: If the machine is in the state described in "Steps to Reproduce", unplug the adapter in normal operation, the suspend and resume, then plug-in again. The adapter works again. ==Attachments== kernel_config The kernel configuration file of my system. thunderbolt_ethernet_lspci_before_suspend.txt lspci -k output before suspending the system (as in "Steps to Reproduce") thunderbolt_ethernet_after_resume.log lspci -k output after resuming the system (as in "Steps to Reproduce") thunderbolt_ethernet_normal_plugin_unplug.log journalctl output of what happens when just plugging in and shortly after unplugging the adapter. All this works as expected. (meant as a possible reference) thunderbolt_ethernet_workaround_2_wakeup.log journalctl output of what happens when resuming like in workaround 2 thunderbolt_ethernet_workaround_2_replugin_thunderbolt_adapter.log journalctl output of what happens when plugging in the adapter again after resume like in workaround 2
Created attachment 191231 [details] lspci -k output before suspending the system (as in "Steps to Reproduce")
Created attachment 191241 [details] lspci -k output after resuming the system (as in "Steps to Reproduce")
Created attachment 191251 [details] Log output after normal working plugin and unplug (reference)
Created attachment 191261 [details] lspci -k output after resuming the system (as in "Steps to Reproduce")
Created attachment 191271 [details] Log output after resuming as in "Steps to Reproduce"
Created attachment 191281 [details] Log output of what happens after resuming as in "Workaround 2"
The two workaround logs have been merged into one. Sorry for the flood of attachments: this is my first bug report and I could not find an option to add multiple attachments before initial submission.
On Tue, Oct 27, 2015 at 10:23:03AM +0000, bugzilla-daemon@bugzilla.kernel.org wrote: > https://bugzilla.kernel.org/show_bug.cgi?id=106681 > > Bug ID: 106681 > Summary: Apple Thunderbolt Ethernet Adapter not Working After > Suspend & Resume Can you send all of this information to the netdev@vger.kernel.org mailing list so the network driver developers can help you out with this? thanks, greg k-h
(In reply to Greg Kroah-Hartman from comment #8) > Can you send all of this information to the netdev@vger.kernel.org > mailing list so the network driver developers can help you out with > this? Should I send them a link to this bug or create an e-mail containing the entire bug description + attachments? Can't this bug be moved to the netdev section if this is more appropriate? Thanks for your patience!
On Wed, Oct 28, 2015 at 12:36:15AM +0000, bugzilla-daemon@bugzilla.kernel.org wrote: > https://bugzilla.kernel.org/show_bug.cgi?id=106681 > > --- Comment #9 from me@cschwarz.com --- > (In reply to Greg Kroah-Hartman from comment #8) > > Can you send all of this information to the netdev@vger.kernel.org > > mailing list so the network driver developers can help you out with > > this? > > Should I send them a link to this bug or create an e-mail containing the > entire > bug description + attachments? An email please. > Can't this bug be moved to the netdev section if this is more appropriate? The network developers (and most driver subsystems), do not use bugzilla, they use the mailing list instead. thanks, greg k-h